article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
systems of differential equations in both finite- and infinite - dimensional settings close to an ahb have been subject to intense research due to their dynamical complexity and importance in applications .
the latter range from models in fluid dynamics @xcite to those in the life sciences , in particular , in computational neuroscience @xcite .
when the proximity to the ahb coincides with certain global properties of the vector field , it may result in a very complex dynamics @xcite .
the formation of the smale horseshoes in systems with a homoclinic orbit to a saddle - focus equilibrium provides one of the most representative examples of this type @xcite .
canard explosion in relaxation systems affords another example @xcite .
recent studies of relaxation systems , motivated mainly by applications in the life sciences , have revealed that the proximity to an ahb has a significant impact on the system dynamics .
it manifests itself as a family of multimodal periodic solution that are composed of large - amplitude relaxation oscillations ( generated by the global structure of the vector field ) and small - amplitude nearly harmonic oscillations ( generated by the vector field near the equilibrium close to the ahb ) @xcite ( see figure [ f.1 ] ) .
these families of solutions possess rich bifurcation structure .
a remarkable example of an infinite - dimensional system close to the ahb has been recently studied by frankel and roytburd @xcite .
they derived and systematically studied a model of solid fuel combustion in the form of a free boundary problem for a @xmath1 heat equation with nonlinear conditions imposed at the free boundary modeling the interface between solid fuel mixture and a solid product .
the investigations of this model revealed a wealth of spatial - temporal patterns ranging from a uniform front propagation to periodic and aperiodic front oscillations .
the transitions between different dynamical regimes involve a variety of nontrivial bifurcation phenomena including period - doubling cascades , period - adding sequences , and windows of chaotic dynamics . to elucidate the mechanisms responsible for different dynamical regimes and transitions between them , frankel and roytburd employed pseudo - spectral techniques to derive a finite - dimensional approximation for the interface dynamics in the free boundary problem @xcite . as shown in @xcite , a system of three ordinary differential equations captured the essential features of the bifurcation structure of the infinite - dimensional problem .
the numerical bifurcation analysis of the finite - dimensional approximation revealed a rich family of multimodal periodic solutions similar to those reported in the context of relaxation systems near the ahb @xcite .
the bifurcation diagrams presented in @xcite and in @xcite share a striking similarity , despite the absence of any apparent common structures in the underlying models ( except to the proximity to the ahb ) . in particular , in both models ,
topologically distinct multimodal periodic solutions are located on isolas , closed curves in the parameter space .
the methods of analysis of the mixed - mode solutions in @xcite used in an essential way the relaxation structure present in these problems .
these approaches can not be applied directly to analyzing the model in @xcite , because it is not a priori clear what creates the separation of the time scales in this model , in spite of the evident fast - slow character of the numerical solutions .
this is partly due to the spectral method , which was used to derive the system of equations in @xcite : while it has captured well the finite - dimensional attractor of the interface dynamics , it has disguised the structure of the physical model .
one of the goals of the present paper is to identify the structure responsible for the generation of the multimodal oscillations in a finite - dimensional model for the interface dynamics and to relate it to those studied in the context of relaxation oscillations .
the family of flows in @xcite includes in a natural way two types of the ahbs .
depending on the parameter values , the equilibrium of the system of ordinary differential equations in @xcite undergoes either a sub- or a supercritical ahb .
a similar situation is encountered in certain neuronal models ( see , e.g. , @xcite ) . in either case , the global multimodal periodic solutions are created after the ahb . however , in the case of a supercritical bifurcation , they are preceded by a series of period - doubling bifurcations of small amplitude limit cycles , arising from the ahb . on the other hand , in the subcritical case
, the ahb gives rise to multimodal solutions , whose lengths and time intervals between successive large amplitude oscillations can be very long . in the present paper
, we perform a detailed asymptotic analysis of the trajectories in a class of systems motivated by the problem in @xcite .
our analysis includes both cases of the sub- and supercritical ahbs .
we also investigate the dynamical regimes arising near the border between the regions of sub- and supercritical ahb .
this region in the parameter space contains a number of nontrivial oscillatory patterns including multimodal trajectories with substantial time intervals between successive spikes , irregular , and possibly chaotic oscillations , as well as a family of periodic orbits undergoing a cascade of period - doubling bifurcations .
our analysis shows that these dynamical patterns and the order in which they appear under the variation of the control parameters are independent on the details of the model , but are characteristic to the transition from sub- to supercritical ahb .
the outline of the paper is as follows . after introducing the model and rewriting it in the normal
coordinates , we present a set of the numerical experiments to be explained in the remainder of the paper .
then we state our results for each of the following cases : supercritical ahb , subcritical ahb , and the transition layer between the regions of sub- and supercritical ahb . in section 3
, we analyze the local behavior of trajectories near a weakly unstable saddle - focus .
the local expansions used in this section are similar to those used in @xcite for analyzing the ahb using the method of averaging .
however , rather than establishing existence of periodic solutions , the goal of the present section is to approximate the trajectories near the saddle - focus .
for this , after rescaling the variables and recasting the system into cylindrical coordinates , we reduce the system dynamics to a @xmath2 slow manifold . by integrating the leading order approximation of the reduced system ,
we obtain necessary information about the local behavior of trajectories .
the results of this section are summarized in theorem 3.1 . in section 4
, we study oscillatory patterns generated by the class of systems under investigation .
it is divided into three subsections devoted to the oscillations triggered by the supercritical ahb ( 4.1 ) , subcritical ahb ( 4.2 ) , and those found in the transition region between sub- and supercritical ahb ( 4.3 ) . in the supercritical case ,
we show that the oscillations just after the ahb are already bimodal .
generically , one needs to use two harmonics to describe the limit cycle born at the supercritical ahb in @xmath0 .
we give a geometric interpretation of this effect , which we call the frequency doubling , due to the fact that the second frequency is twice as large as that predicted by the hopf bifurcation theorem @xcite .
we also compute the curvature and the torsion of the periodic orbit as a curve in @xmath0 .
the latter is useful for the geometric explanation of the frequency doubling . in subsection 4.2 , we study the subcritical case . we show that under two general assumptions on the global vector field , the presence of the return mechanism ( * g1 * ) and the strong contraction property ( * g2 * ) , the subcritical ahb results in sustained multimodal oscillations . even though the oscillations may not be periodic ( our assumptions do not warrant periodicity ) , the time intervals between consecutive spikes of the resultant motion comply to the uniform bounds given in theorem 4.1
this subsection also contains the definitions of the multimodal oscillations and the precise formulation of the assumptions on the global vector field .
the proof of theorem 4.1 is relegated to section 5 .
subsection 4.3 describes the transition between sub- and supercritical ahb .
this transition contains a distinct bifurcation scenario , when a small positive real part of the complex conjugate pair of eigenvalues is positive and fixed and the first lyapunov coefficient changes sign .
the numerical simulations show that this results in the formation of the chaotic attractor followed by the reverse period - doubling cascade . to explain this bifurcation sequence , we derive a @xmath1 first return map .
the asymptotic analysis in this subsection is complemented by numerical extension of the map to the region inaccessible by local asymptotic expansions .
the first return map obtained by the combination of the analytic and numerical techniques reveals the principal traits of the transition from sub- to supercritical ahb .
finally , the discussion of the results of the present paper and their relation to the previous work is given in section 6 .
in the present section , we formulate the model and present a set of numerical results , which motivated our study .
the following system of three odes was derived in @xcite , as a finite - dimensional approximation for the interface dynamics in a free boundary problem modeling solid fuel combustion : @xmath3 here , @xmath4 approximates the velocity of the interface between the solid fuel and the burnt material .
functions @xmath5 and @xmath6 are the first two coefficients in the series expansion for the spatial temperature profile with respect to the basis of chebyshev - laguerre polynomials . equations ( [ 6.1])-([6.3 ] ) are obtained by projecting the original infinite - dimensional problem onto a finite - dimensional function space and using the method of collocations for determining the unknown coefficients .
an important ingredient of the model , nonlinear kinetic function @xmath7 is given by k(v)=(1-v)^p-(1-v)^-1p+1 .
* a * + * c * + * e * + * g * it reflects the dependence of the velocity of propagation on the temperature in the front .
this relation is very complex and is not completely understood at present . to account for a range of possible kinetic mechanisms , the model includes two control parameters : @xmath8 and @xmath9 .
both the interface velocity and the temperature profile are calculated in the uniformly moving frame of reference .
therefore , system of equations ( [ 6.1])-([6.3 ] ) describes the deviations of the interface dynamics from that of a front traveling with constant speed .
in particular , periodic and aperiodic oscillations generated by ( [ 6.1])-([6.3 ] ) correspond to complex spatio - temporal patterns in the infinite - dimensional model .
the numerical results presented in @xcite show a remarkable similarity between the complex patterns generated by the infinite- and finite - dimensional models and between the scenarios for transitions between different regimes in both models . for more information about the derivation of the free - boundary problem for solid fuel combustion and its finite - dimensional approximation , we refer the reader to @xcite and bibliography therein .
a simple inspection of ( [ 6.1])-([6.3 ] ) shows that it has an equilibrium at the origin @xmath10 for all values of @xmath8 and @xmath11 .
we linearize ( [ 6.1])-([6.3 ] ) about the equilibrium at the origin @xmath12 near @xmath13 , jacobian matrix @xmath14 has a negative eigenvalue @xmath15 and a pair of complex conjugate eigenvalues @xmath16 and @xmath17 .
the latter crosses the imaginary axis transversally at @xmath18 : @xmath19 therefore , the equilibrium of ( [ 6.1])-([6.3 ] ) undergoes an ahb . in the neighborhood of @xmath20 , @xmath21 defines a smooth invertible function .
we shall use @xmath22 as a new control parameter . after a linear coordinate transformation , system of equations ( [ 6.1])-([6.3 ] )
has the following form : x = ( ccc -1 & 0 & 0 + 0 & & -b ( ) + 0 & b ( ) & ) x + h(x , ) , x=(x_1 , x_2 , x_3)^t , h=(h_1 , h_2 , h_3)^t . here ,
by @xmath23 we denote the imaginary part of @xmath24 as a function of the new control parameter .
nonlinear function @xmath25 is a smooth function such that @xmath26 and @xmath27 for values of @xmath22 close to @xmath28 .
more precisely , @xmath29 stands for a family of functions parametrized by @xmath9 ( see ( [ 6.4 ] ) ) . to keep the notation simple
, we omit the dependence of @xmath29 on the second control parameter @xmath9 . *
a * it turns out that in the range of parameters of interest , the ahb of the equilibrium for @xmath30 can be either subcritical , or supercritical , or degenerate depending on the value of @xmath9 .
this gives rise to several qualitatively distinct oscillatory regimes generated by ( [ 6.5 ] ) for small values of @xmath31 .
the corresponding bifurcation scenarios for ( [ 6.1])-([6.3 ] ) for fixed values of @xmath9 and varying @xmath8 were studied numerically in @xcite .
below , we reproduce some of these numerics for the system in new coordinates and supplement them with a set of new numerical experiments relevant to the analysis of this paper . after that we state our results for each of the following cases : supercritical ahb , subcritical ahb , and the transition layer between the regions of sub- and supercritical bifurcations .
* a * we start with discussing the supercitical case .
it is well known that a supercritical ahb produces a stable periodic orbit in a small neighborhood of the bifurcating equilibrium .
the period of the nascent orbit is approximately equal to @xmath32 where @xmath33 is the imaginary part of the pair of complex conjugate eigenvalues of the matrix of the linearized system at the bifurcation .
the numerical time series of @xmath34 and @xmath35 show that , while @xmath35 oscillate with the frequency prescribed by the andronov - hopf bifurcation theorem @xcite , the former oscillates with twice that frequency ( see figure [ f.1]a , b ) .
the asymptotic analysis of section 3 explains this counter - intuitive effect and shows that this is , in fact , a generic property of the periodic orbits born from a supercritical bifurcation in @xmath36 , @xmath37 .
note that the difference in frequencies of oscillations in @xmath34 and @xmath35 can not be understood from the topological normal form from the ahb @xcite , because the latter does not contain the information about the geometry of the periodic orbit . in section 4.1
, we compute two geometric invariants of the periodic orbits as a curve in @xmath0 : the curvature and the torsion . the latter shows that generically the orbit born from a supercritical ahb is not planar .
this together with certain symmetry properties of the orbit explains the frequency doubling of the oscillations in @xmath34 .
as was noted in @xcite , for increasing values of @xmath38 , the small periodic orbit born from the supercritical ahb undergoes a cascade of period - doubling bifurcations , the first of which is shown in figure [ f.1]c , d . already at the first period - doubling bifurcation ,
the periodic orbit lies outside the region of validity of the local power series expansions , developed in the present paper .
therefore , our analysis does not explain the period - doubling bifurcations for increasing values of @xmath39 .
in section 4.3 we complement our analytical results with the numerical construction of the @xmath1 first return map .
the latter explains the mechanism for period - doubling cascade and the window of complex dynamics reported in @xcite . *
a * * a * we next turn to the subcritical case .
the dynamics resulting from the subcritical ahb depends on the properties of the vector field outside of a small neighborhood of the equilibrium .
we conducted a series of numerical experiments to study the global properties of the vector field ( see caption of figure [ f.2 ] for details ) .
based on these numerical observations , we identify two essential features of the vector field : * ( g1 ) * the return mechanism and * ( g2 ) * the strong contraction property . specifically , ( g1 ) : : for a suitably chosen cylindrical crossection , @xmath40 , placed sufficiently close to the origin , and another crossection , @xmath41 , transverse to @xmath42 ( see figure [ f.2]a ) , the flow - defined map @xmath43 is well - defined for @xmath44 and depends smoothly on the parameters of the system .
( g2 ) : : there is a region @xmath45 , adjacent to @xmath41 and containing a subset of @xmath42 ( see figure [ f.2]a ) , in which the projection of the vector field in direction transverse to @xmath42 is sufficiently stronger than that in the tangential direction . in section 4.2 , conditions
* ( g1 ) * and * ( g2 ) * are made precise .
properties * ( g1 ) * and * ( g2 ) * guarantee that for small values of @xmath39 , the trajectories leaving a small neighborhood of the saddle - focus , after some relatively short time enter @xmath45
. in @xmath45 , the trajectories approach @xmath42 closely and then follow it to a sufficiently small neighborhood of the unstable equilibrium . for small @xmath39 ,
the trajectories starting from a sufficiently small neighborhood of the saddle - focus , remain in some larger ( but still small ) neighborhood of the origin for a long time .
eventually , they hit @xmath40 and the dynamics described above repeats .
therefore , under conditions * ( g1 ) * and * ( g2 ) * , the subcritical ahb bifurcation results in a sustained motion consisting of long intervals of time that the trajectory spends near a weakly unstable saddle - focus and relatively brief excursions outside of a small neighborhood of the origin ( see figure [ f.2]a ) . in terms of the time series , the dynamical variables undergo series of small amplitude oscillations alternating with large spikes ( figure [ f.1]e - h ) . for increasing values of @xmath38 , the number of small amplitude oscillations decreases and so do the time intervals between consecutive spikes , so - called interspike intervals ( isis ) .
the multimodal solutions arising near the subcritical ahb are not necessarily periodic .
nonetheless , the isis may be characterized in terms of the control parameters present in the system , regardless of whether the underlying trajectories are periodic or not .
specifically , in section 4.2 , we derive an asymptotic relation for the duration of the isis : ~12 ( 1 + ( p)o(^- ) ) , 4 , where the first lyapunov coefficient @xmath46 reflects the distance of the system from the transition from sub- to supercritical ahb , with @xmath46 being positive when the ahb is subcritical and equal to zero at the transition point . finally , @xmath47 is a small parameter and positive @xmath48 .
the role of @xmath49 will become clear later .
we illustrate ( [ 6.6 ] ) with numerically computed plots of the isis under the variation of control parameters in figure [ f.3 ] .
the isis provide a convenient and important characteristics of the oscillatory patterns involving pronounced spikes .
for example , it is widely used in both theoretical and experimental neuroscience for description of patterns of electric activity in neural cells .
an important aspect of the isis , is that the dependence of the isis on the control parameters in the system ( which often can be directly established experimentally ) reveals the bifurcation structure of the system .
therefore , the analytical characterization of the isis in terms of the bifurcation parameters , such as given in ( [ 6.6 ] ) , is important in applications .
for the problem at hand , the isis depend on the interplay of the two parameters @xmath22 and @xmath46 , which reflect the proximity of the system to a codimension 2 bifurcation .
the dependence of the isis on @xmath22 was studied in @xcite for a model problem near a hopf - homoclinic bifurcation .
our analysis extends the formula for the isis obtained in @xcite to a wide class of systems and emphasizes the role of @xmath46 , which is important for the present problem .
we also note that the structure of the global vector field suggests that ( [ 6.5 ] ) is also close to a homoclinic bifurcation , which can also influence the isis . however , under the variation of the control parameters @xmath22 and @xmath9 the system remains bounded away from the homoclinic bifurcation , so that the latter effectively does not affect the isis .
in fact , ( [ 6.6 ] ) can be easily extended to include the distance from the homoclinic bifurcation as well . as follows from ( [ 6.6 ] ) , the isis increase for decreasing values of @xmath50 ( see figure [ f.3]b ) .
this observation prompted our interest in investigating the transition in the system dynamics as @xmath51 crosses @xmath28 .
numerical simulations show that as @xmath51 approaches @xmath28 from the positive side , the oscillations become very irregular and very likely to be chaotic ( figure [ f.4 ] ) .
when @xmath52 is fixed and @xmath51 crosses 0 , the irregular dynamics is followed by the reverse period - doubling cascade and terminates with the creation of a regular small amplitude periodic orbit , which can be then tracked down to a nondegenerate supercritical ahb ( figure [ f.5]a - h ) .
this scenario presents a certain interest as it suggests the formation of the chaotic attractor in a well - defined bifurcation setting .
namely , the transition from the region of subcritical to supercritical ahb for small @xmath52 leads to the formation of a chaotic attractor and a reverse period - doubling cascade . as in the case of the complex dynamics appearing for increasing values of @xmath22 for supercritical ahb
, the irregular oscillations in the present case can not be understood using the local analysis alone .
therefore , we complemented our analysis with the study of numerically constructed first return map .
the latter gives a clear geometric picture of the origins of the chaotic dynamics and the period - doubling cascade of periodic orbits in this parameter regime .
in the present section , we study trajectories of ( [ 6.5 ] ) in a small neighborhood of the equilibrium at the origin .
assume that @xmath53 is a smooth function in a small neighborhood of @xmath54 , so that all power series expansions below are justified .
we are interested in the dynamics of ( [ 6.5 ] ) for values of @xmath22 close to @xmath28 .
the role of @xmath22 in our analysis is twofold . on the one hand
, @xmath22 is a control parameter in this problem , on the other hand , the smallness of @xmath22 is used in the asymptotic analysis below . to separate these two roles , we use the following rescaling = ^2 , where @xmath47 is a fixed sufficiently small constant and @xmath55 varies in a certain interval of size @xmath56 .
we introduce cylindrical coordinates in @xmath0 ( x_1,x_2,x_3)(x_1 , , ) ^+s^1,x_2=,x_3= , and define d=\ { ( x_1 , , ): | x_1 | m^2 ^ 2 , ( 0,2 m ] } , d_0=\ { ( x_1 , , ): |
x_1 to characterize the trajectories of ( [ 6.5 ] ) in @xmath57 we will use an exponentially stable slow manifold , @xmath58 , whose leading order approximation is given by s_0= \ { ( x_1 , , ) : x_1=u()^2 , 02 m } , where u()= a + a(2- ) . and @xmath59 and @xmath60 are computable constants ( see appendix a ) . below
we show that the trajectories of ( [ 6.5 ] ) with initial data from @xmath61 approach an @xmath62 neighborhood of @xmath63 in time @xmath64 and stay in this neighborhood as long as they remain in @xmath57 .
the reduction of the system dynamics to @xmath58 yields a complete description of the trajectories of ( [ 6.5 ] ) in @xmath57 .
the qualitative character of the solution behavior in @xmath57 depends on the sign of _ the first lyapunov coefficient _
= 12_0 ^ 2 ( u ( ) |q_2()+ |q_3 ( ) ) d , where @xmath65periodic trigonometric polynomial @xmath66 are given in appendix a. finally , by @xmath33 we denote the absolute value of the imaginary parts of the complex conjugate pair of eigenvalues at the ahb ( see ( [ 6.5 ] ) ) .
the results of the analysis of this section are summarized in * theorem 3.1 * suppose @xmath67 and @xmath68 .
let @xmath47 denote a small parameter .
then for @xmath48 and for sufficiently large ( independent of @xmath49 ) @xmath69 , the trajectories with initial conditions in @xmath61 enter an @xmath62 neighborhood of @xmath63 in time @xmath70 and remain there as long as they stay in @xmath57 . in the neighborhood of @xmath63 , the trajectories can be uniformly approximated on any interval of time @xmath71 $ ] of length @xmath56 : x_1(t)=^2 u()+o(^3 ) , x_2(t)=+o(^2),x_3(t)=+o(^2 ) , where @xmath72 and @xmath51 are defined in ( [ 5.4 ] ) and ( [ 5.5 ] ) , respectively , and @xmath73 * remark 3.1 * a ) : : from theorem 3.1 , one can easily deduce the existence of a periodic orbit @xmath74 , when @xmath75 .
@xmath74 is stable ( unstable ) if @xmath39 ( @xmath76 ) .
the leading order approximation for @xmath74 follows from ( [ 5.7 ] ) : x_1=|^2 u()+o(^3 ) , x_2=|+o(^2 ) , x_3=|+o(^2 ) , [ 0 , 2 ) , where @xmath77 .
equations ( [ 5.4 ] ) and ( [ 5.8 ] ) imply that the frequency of oscillations in @xmath34 is twice that of oscillations in @xmath35 , provided @xmath78 in ( [ 5.4 ] ) ( see figure [ f.1]a , b ) . in section 4.1
, we give a geometric interpretation of this frequency doubling effect .
b ) : : if @xmath39 and @xmath50 , theorem 3.1 describes the trajectories with initial conditions near a weakly unstable equilibrium .
it shows that along these trajectories , @xmath35 undergo approximately harmonic oscillations , whose amplitude grows at the rate @xmath79 . in addition
, they satisfy the following scaling relation : |x_1(t ) ~a^2(t ) , where @xmath80 stands for the average value of @xmath81 over one cycle of oscillations .
c ) : : the domain of validity of the asymptotic expansions in ( [ 5.7 ] ) extends much farther than @xmath57 .
as will follow from the proof of theorem 3.1 , @xmath82 in the definition of @xmath57 may be taken up to @xmath83 .
this implies that , with the error term @xmath84 , ( [ 5.7 ] ) remains valid for the ranges of @xmath34 and @xmath35 up to @xmath85 and @xmath86 , respectively . in the remainder of this section , we prove theorem 3.1 . by expanding @xmath53 into finite taylor sum with the reminder term ,
we have @xmath87 where @xmath88 to study system of equations ( [ 5.1 ] ) in a small neighborhood of the origin , we rewrite it in cylindrical coordinates ( [ cl ] ) and rescale variables x_1=^2 = r. in new coordinates , we have @xmath89 where @xmath90 and @xmath91 are homogeneous trigonometric polynomials . using @xmath92 as a new variable , from ( [ 5.11 ] ) , we obtain @xmath93 where @xmath94 and @xmath95 are trigonometric polynomials of period @xmath96
. functions @xmath97 and @xmath98 are bounded and continuous in @xmath99 $ ] . here
@xmath100 stands for the domain of the rescaled variable @xmath101 when the original variable @xmath102 : d^= \ { ( , r ):
| | 4m^2 , 0<r2 m } .
similarly , we define d^_0=\ { ( , r ) : | | m^2 , 0<rm}. to determine the slow manifold for ( [ 5.1 ] ) , we introduce the following notation . by @xmath103 and @xmath104 we denote the mean value and the oscillating part of the trigonometric polynomial @xmath95 in ( [ 5.11a ] ) , respectively : p_1()-p_1^()1 + 4 ^ 2 .
we first prove an auxiliary lemma , which characterizes the trajectories of ( [ 5.11a ] ) and ( [ 5.12 ] ) in @xmath105 . in the following two lemmas , we will keep track of how the remainder terms in the asymptotic expansions depend on the size of @xmath105 .
this will be used later to determine the domain of validity for the asymptotic approximation of the slow manifold .
below @xmath106 means that there exist positive constants @xmath107 and @xmath108 ( independent of @xmath49 ) such that @xmath109 for @xmath110 $ ] . in the remainder of this section ,
all estimates are valid for sufficiently large @xmath69 independent of @xmath49 ( as stated in the theorem 3.1 ) , but also allow the possibility that @xmath111 grows as @xmath112 .
* lemma 3.1 * suppose @xmath113 remains in @xmath105 for @xmath114 $ ] .
then ( ) = _ 0()r^2 ( ) + e^-(-_0)\ { ( _ 0)-_0(_0 ) r^2(_0 ) } + o(m^3 ) , , where @xmath115 is defined in ( [ 5.16 ] ) . *
proof : * using ( [ 5.13 ] ) , we integrate ( [ 5.11a ] ) over @xmath116 $ ] e^()=e^_0(_0)+ |p_1 _ _ 0^e^s r^2(s)ds + _ _ 0^e^s
r^2(s ) p_1(s ) ds+ _ _ 0^e^s ( ( s ) , r(s),s , ) ds . using integration by parts
, we have @xmath117 from ( [ 5.12 ] ) and ( [ dprime ] ) , we find that for @xmath118 @xmath119 where @xmath120 is independent of @xmath49 . therefore , _
_ 0^e^s r^2(s)ds = e^r^2()-e^_0 r^2(_0 ) + e^-_0o(m^3 ) .
similarly , using integration by parts in the second integral on the right hand side of ( [ 5.17 ] ) , we obtain @xmath121 applying integration by parts in the last integral on the right hand side of ( [ a.2 ] ) , we have @xmath122 by direct verification , one finds that @xmath123 the integrals involving @xmath124 on the right hand sides of ( [ a.2 ] ) and ( [ a.3 ] ) are bounded by @xmath125 since , by ( [ 5.12 ] ) , @xmath126 . using these observations , from ( [ a.2 ] ) and ( [ a.3 ] )
we obtain _ _
0^e^s r^2(s ) p_1(s)ds = e^r^2()p_1()-e^_0 r^2(_0)p_1(_0)+ e^-_0o(m^3 ) , where @xmath127 is given by ( [ 5.16 ] ) .
the combination of ( [ 5.17 ] ) , ( [ 5.18 ] ) , and ( [ a.4 ] ) yields ( [ 5.15 ] ) .
+ @xmath128 the following lemma shows that the trajectories of ( [ 5.11a ] ) and ( [ 5.12 ] ) , which stay in @xmath105 for sufficiently long time enter a small neighborhood of @xmath58 no later than in time @xmath129 and remain in this neighborhood as long as they stay in @xmath105 ( see ( [ dprime ] ) ) .
* lemma 3.2 * let @xmath130 denote a trajectory of ( [ 5.11a ] ) and ( [ 5.12 ] ) , which stays in @xmath105 for @xmath131 $ ] , @xmath132 , @xmath133 .
then ( ) = _ 0 ( ) r^2 ( ) + o(m^3 ) , for @xmath134 and as long as @xmath135 .
* proof : * denote @xmath133 . by plugging in @xmath136 into ( [ 5.15 ] ) , we have |(_1 ) - _ 0(_1)r^2(_1)|c_2 , for some @xmath137 independent of @xmath47 . using lemma 3.1 again with @xmath138
, we obtain @xmath139 the expression in the curly brackets is @xmath62 by ( [ 5.21 ] ) .
+ @xmath128 * remark 3.2 * lemmas 3.1 and 3.2 show that the trajectories of ( [ 5.11a ] ) and ( [ 5.12 ] ) converge to an exponentially stable manifold @xmath58 , whose leading order approximation is given in the definition of @xmath63 .
the method , which we used in lemmas 3.1 and 3.2 to obtain the leading order approximation of the slow manifold , can be extended to calculate the higher order terms in the expansion for @xmath58 . having shown that the trajectories approach an @xmath62 neighborhood of @xmath63 in time @xmath70 , next we reduce the dynamics of ( [ 5.11a ] ) and ( [ 5.12 ] ) to the slow manifold .
for this , we define i()=(1r()+q_1())^2 , where @xmath140 and @xmath141 is a trigonometric polynomial on the right hand side of ( [ 5.12 ] ) . for the sake of definiteness
, we choose @xmath142 such that @xmath143 the following lemma provides the desired reduction .
* lemma 3.3 * for @xmath134 and as long as @xmath144 , @xmath145 satisfies the following system of equations @xmath146 where @xmath147 is a trigonometric polynomial with period @xmath148 and zero mean @xmath149 .
the expression for @xmath51 is given in ( [ 5.5 ] ) .
* proof : * the change of variables r()=1j()-q_1(),q_1^()=q_1 ( ) , in equation ( [ 5.12 ] ) yields j = -^2j((+q_2 ( ) ) j^2 + q_3())+o(m^2 ^ 3 ) .
after another change of variables , @xmath150 , we obtain = -2 ^ 2 ( ( + q_2 ( ) ) i + q_3 ( ) ) + o(m^3 ^ 3 ) , by lemma 3.2 , for @xmath134 ( ) = _ 0()r^2 ( ) + o(m^3 ) . by plugging in ( [ 5.29 ] ) into ( [ 5.28 ] ) , we obtain = -2 ^ 2 ( i + _ 0()q_2()+q_3())+ o(m^3 ^ 3 ) .
equation ( [ 5.24 ] ) follows by rewriting ( [ 5.30 ] ) = -2 ^ 2 ( i
+ + q())+ o(m^3 ^ 3 ) , where = 1_0^(_0()q_2()+q_3())d _ 0^ q()d=0 , and @xmath115 is given in ( [ 5.16 ] ) .
equation ( [ 5.25 ] ) follows from the last equation in ( [ 5.11 ] ) and the definition of @xmath151 .
+ @xmath128 the statements in theorem 3.1 can now be deduced from equations ( [ 5.24 ] ) and ( [ 5.25 ] ) .
we only need to show that a trajectory with initial condition in @xmath152 remains in @xmath105 for times longer than @xmath70 .
this follows from the fact that @xmath153 in @xmath105 .
consequently , it takes time @xmath154 for @xmath155 to undergo @xmath56 change necessary for leaving @xmath105 from @xmath156 .
we omit any further details . in conclusion
, we note that the domain of validity of the asymptotic analysis of this section extends much further beyond @xmath105 .
indeed , from ( [ 5.20 ] ) we observe that the remainder term tends to @xmath28 with @xmath112 provided @xmath157 .
this means that the expansions for @xmath158 and @xmath159 can be controlled in regions of size up to @xmath160 and @xmath161 respectively .
for the original variables @xmath162 @xmath34 , these estimates translate into @xmath163 and @xmath164 respectively .
we end this section by deriving a useful estimate for the time of flight of trajectories passing close to the saddle - focus . for this , consider an initial value problem for system of equations ( [ 5.11a ] ) and ( [ 5.12 ] ) .
suppose that the initial condition implies that @xmath165 .
we would like to know how long it takes for @xmath155 to reach a given value @xmath166 , @xmath167 .
we assume that @xmath168 and @xmath169 are sufficiently separated , e.g. , @xmath170 . to estimate @xmath171 , note that = _
1+_2 , where @xmath172 is time necessary to reach an @xmath62 neighborhood of @xmath63 from @xmath168 . denote i_1:= i(_1)=i_0+o(^2|| ) , _ 1=_0+_1 .
the second term on the right hand side of ( [ 5.51 ] ) can be estimated by integrating ( [ 5.24 ] ) over @xmath173:$ ] |i= i_1 e^-2_2+(1- e^-2_2 ) -2 ^ 2_0^_2 e^-2(_2-s ) q(_0+s)ds + o(^3 ) .
the integral on the right hand side of ( [ 5.53 ] ) is bounded uniformly for @xmath174 .
this observation combined with ( [ 5.52 ] ) implies |i= i_0
e^-2_2-(1- e^-2_2 ) + o(^2 ) . note that the contribution of @xmath175 term in ( [ 5.52 ] ) to ( [ 5.54 ] ) is negligible provided @xmath176 is sufficiently large . from ( [ 5.54 ] )
, we obtain the desired estimate = 12 ^ 2(i_0 + |i + + o(^2 ) ) .
in the present section , we use the local analysis of section 3 to study certain oscillatory patterns arising in the model of solid fuel combustion ( [ 6.1])-([6.3 ] ) . by plugging in the values of the parameters of ( [ 6.1])-([6.3 ] ) into the expression for @xmath51 ( [ 5.5 ] ) , we find that @xmath46 is a quadratic function with two zeros at @xmath177 and @xmath178 .
these values are in a good agreement with those obtained by numerical bifurcation analysis in @xcite .
the quadratic character of @xmath46 is explained by the fact that @xmath51 is determined by the second order terms in the taylor expansions of @xmath179 .
we concentrate on the parameter region around @xmath180 , where @xmath46 changes its sign . for small positive @xmath22 , there are two dynamical regimes : for values of @xmath9 lying to the left and to the right of some @xmath62 neighborhood of @xmath181 corresponding to the supercritical and subcritical ahbs .
the transition region between these two parameter regimes adds to the repertoire of qualitatively distinct dynamical behaviors .
below , we study these three cases in more detail .
it follows from theorem 3.1 that for @xmath182 and sufficiently small @xmath39 , ( [ 6.5 ] ) has a stable limit cycle @xmath74 , whose leading order approximation is given by x()=(|^2(a+a 2 ) , | , | ) , |=,[0 , 2 ) ( see figure [ f.6]a ) .
moreover , the analysis of section 3 shows that all trajectories starting from a sufficiently small neighborhood of the origin and not belonging to @xmath42 , converge to @xmath74 .
the leading order approximation of @xmath74 in ( [ 4.1 ] ) reveals a remarkable property of the oscillations generated by the limit cycle born from the supercritical ahb : the frequency of oscillations in @xmath34 is twice as large as that of the oscillations in @xmath35 . to explain the frequency doubling effect
, we recall that = + o ( ) .
therefore , ( [ 4.1 ] ) implies that , unless @xmath183 , the frequency of oscillations in @xmath34 is @xmath184 , while that of oscillations in @xmath35 is @xmath185 .
the latter coincides with the frequency of the bifurcating periodic solution .
below , we complement the analytical explanation of the frequency doubling with the geometric interpretation . the oscillations in @xmath35 can be well understood using the topological normal form of the ahb @xcite . indeed
, since at the bifurcation the center manifold at the origin is tangent to the @xmath186 plane , a standard treatment of the ahb using the center manifold reduction @xcite shows that the projection of the bifurcating limit cycle onto @xmath186 plane to leading order is a circle ( figure [ f.6]b ) and the projection of the vector field is given by the equation of the angular variable ( [ 4.2 ] ) .
therefore , the oscillations in @xmath35 are approximately harmonic with the period equal approximately to @xmath187 .
the topological normal from , however , does not describe the oscillations in @xmath34 . for this , one needs to take into account the geometry of the bifurcating periodic orbit as a curve in @xmath0 .
the geometry of @xmath74 is fully determined by the two geometric invariants : the curvature , @xmath188 , and the torsion , @xmath189 .
for the purposes of the present discussion , we need only the latter , but we compute both invariants for completeness . after some algebra ,
the parametric equation for the periodic orbit ( [ 4.1 ] ) yields @xmath190\right|\over \left|\dot x\right|^3}= \sqrt{1 + 2a^2\left(3\cos 4\theta + 5\right)}+\mbox{h.o.t},\\ \lbl{4.3 } \kappa(\theta ) & = & { \left(\dot x , \ddot x , \dddot x\right)\over\left|\left[\dot x , \ddot x\right]\right|^2}= { -\gamma\over\alpha } { 6a\sin 2\theta\over 1 + 2a^2\left(3\cos 4\theta + 5\right ) } + \mbox{h.o.t},\quad \theta\in[0 , 2\pi).\end{aligned}\ ] ] the geometry of the leading order approximation of the periodic orbit is determined by the three parameters @xmath191 @xmath192 and @xmath193 .
the former two parameters are the same as the parameters in the topological normal form of the nondegenerate ahb @xcite ; while the latter captures the geometry of the slow manifold ( or unstable manifold ) near the origin . from the geometrical viewpoint
the bifurcation is degenerate if either @xmath51 or @xmath193 is equal to zero .
the latter condition holds if and only if the slow manifold is either a plane or a circular paraboloid near the origin . in this case ,
equation ( [ 4.3 ] ) implies that ( to leading order ) the bifurcating orbit lies in a plane .
generically , @xmath74 is not planar ( see figure [ f.6]a ) .
the fact that the periodic orbit is generically not planar combined with the symmetry of the orbit about the @xmath194axis implies that its projection onto any plane containing @xmath194axis has to have a self - intersection ( see figure [ f.6]c ) . from figure [ f.6]c ,
it is clear that as the phase point goes around @xmath74 once , @xmath34 has to trace its range at least twice .
therefore , the frequency of oscillations in @xmath34 has to be at least twice as high as that of the oscillations in @xmath35 .
the above discussion implies that the frequency doubling of oscillations in @xmath34 is a generic geometric property of the ahb . *
a * if the ahb is subcritical ( @xmath50 ) the loss of stability of the equilibrium at the origin of ( [ 6.5 ] ) results in the creation of multimodal trajectories which spend a considerable amount of time near a weakly unstable equilibrium . to describe the resultant dynamics
we give the following definition . *
definition 4.1 * we say that a trajectory of ( [ 6.5 ] ) undergoes multimodal oscillations for @xmath195 if there exist positive constants @xmath196 and @xmath197 independent of @xmath49 and an unbounded sequence of times @xmath198 such that a. : : @xmath199 , b. : : @xmath200\;\;\forall t^\prime \in \left(t_{2i-1 } , t_{2i+1}\right).$ ] the time intervals @xmath201 are called isis . the proximity of ( [ 6.5 ] ) to the ahb alone is clearly not sufficient to account for the appearance of the multimodal oscillations in ( [ 6.5 ] ) .
below , we formulate two additional assumptions on the vector field outside of the small neighborhood of the equilibrium , which are relevant to ( [ 6.1])-([6.3 ] ) .
under these conditions , we show that the subcritical ahb results in sustained multimodal oscillations .
in addition , we determine the asymptotics of the isis for positive @xmath48 .
for ( [ 6.5 ] ) to generate multimodal oscillations , it is necessary that the trajectories leaving an @xmath79 neighborhood of the origin reenter it after some interval of time .
therefore , the vector field in @xmath56 neighborhood of the origin must provide a return mechanism .
our second assumption on the global vector field of ( [ 6.5 ] ) is that the trajectories approaching the origin along a @xmath1 stable manifold , @xmath42 , are subject to a strong contraction toward @xmath42 , i.e. , in a small neighborhood of an @xmath56 segment of @xmath42 , the projection of the vector field onto a plane transversal to @xmath42 is sufficiently stronger than that along @xmath42 ( figure [ f.2]a ) .
this guarantees that the trajectories entering such region of strong contraction approach @xmath42 very closely and follow it to an @xmath79 neighborhood of the unstable equilibrium , from where they are propelled away along the unstable manifold , @xmath202 . due to the proximity of ( [ 6.5 ] ) to the ahb
, the motion away from the origin is very slow .
this results in the pronounced intervals of the small amplitude oscillations .
below , we summarize these observations into two formal assumptions on the global vector field * ( g1 ) * and * ( g2)*. for this , we first need to introduce some auxiliary notation . for analytical convenience ,
we assume that in an @xmath56 neighborhood of the origin , @xmath42 can be and has been straightened by a smooth change of coordinates .
more specifically , the nonlinear terms in ( [ 6.5 ] ) satisfy h_2,3 ( x_1,0,0,)=0,w(x_1,)=-x_1+h_1(x_1,0,0,)>0 , x_1[d_1,0 ) , for some @xmath203 .
note that @xmath204 is assumed to be sufficiently far away from the origin ( see figure [ f.8 ] ) . to describe the mechanism of return ,
we introduce two crossections : ^+=[-c_1 ^ 2 , c_1 ^ 2]d_^-=\{d_1}d__1 , where @xmath205 denotes a disk of radius @xmath206 centered at the origin : d_=\ { y^2:|y|}. positive constants @xmath206 and @xmath207 are sufficiently small ( see figure [ f.2]a ) and @xmath208 is chosen so that @xmath40 intersects the slow manifold @xmath58 transversally .
in addition , we require that @xmath209 to guarantee that @xmath40 belongs to the region of validity of the local analysis of section 3 ( see remark 3.1c ) .
let @xmath210 and consider a trajectory of ( [ 6.5 ] ) starting from @xmath211 .
we assume that for sufficiently small @xmath47 and @xmath48 , every such trajectory intersects @xmath41 from the left .
denote the point of the first intersection by @xmath212 .
we assume that * ( g1 ) * the first return map @xmath213 depends smoothly on @xmath49 and @xmath214 .
to measure the rate of contraction toward @xmath42 , we consider a @xmath215 matrix a(x_1,)= ( cc f_2x_2 & f_2x_3 + f_3x_2 & f_3x_3 ) _ ( x_1,0,0 , ) .
let @xmath216 denote the eigenvalues of the symmetric matrix @xmath217 denote @xmath218 and @xmath219 .
we assume that for sufficiently small @xmath47 and @xmath48 , * ( g2 ) * @xmath220;&\\ \lbl{4.7 } \exists\ ; d_2 \in ( d_1,0):\ ; \bar\lambda(x_1,\alpha)>0 , & x_1\in[d_1,d_2]&\quad \ & \quad \min_{x_1\in[d_1 , d_2 ] } { \bar\lambda(x_1,\alpha)\over w(x_1,\alpha ) } = o\left(\left|\ln\epsilon\right|\right),\end{aligned}\ ] ] * ( g3 ) * @xmath221 under these conditions , we have * theorem 4.1 * let @xmath50 , conditions in ( [ 4.3a ] ) , * ( g1 ) * and * ( g2 ) * hold
. then for sufficiently small @xmath47 and @xmath222 a trajectory of ( [ 6.5 ] ) with initial condition from an @xmath79 neighborhood of the origin and not belonging to @xmath42 undergoes multimodal oscillations .
the isis are uniformly bounded from below _
i^-=12(1+^4c^- ) , i=2,3 , .
if , in addition , * ( g3 ) * holds then the isis satisfy two - sided bounds ^-_i ^+=12(1+^4+c^+ ) , i=2,3 ,
, for some @xmath223 .
positive constants @xmath224 do not depend on @xmath225 @xmath191 and @xmath51 .
* remark 4.1 * ( a ) : : the principal assumptions on the global vector field are formulated in * ( g1 ) * and ( [ 4.7 ] ) . the condition in ( [ 4.6 ] ) makes the derivation of certain estimates in the proof of the theorem easier and is used for analytical convenience .
likewise , * ( g3 ) * is not essential for the proposed mechanism .
however , without this condition obtaining the two - sided estimates for the isis requires an additional argument in the proof .
condition * ( g3 ) * means that the two eigenvalues of @xmath193 have the same order of magnitude .
this condition is not restrictive .
( b ) : : for fixed @xmath50 , the estimates in ( [ 4.8a ] ) and ( [ 4.8b ] ) can be rewrittten as _ i , with constants @xmath226 independent from @xmath22 .
similarly , for fixed @xmath22 and varying @xmath50 inequalities in ( [ 4.8a ] ) and ( [ 4.8b ] ) imply |c^- -12_i |c^+ -12 , where constants @xmath227 do not depend on @xmath50 .
the numerical experiments presented in section 2 show that the transition from subcritical to supercritical ahb contains a distinct bifurcation scenario involving the formation of chaotic attractor via a period - doubling cascade .
the analytical explanation of these phenomena is outside the scope of the present paper .
below we comment on the difficulties arising in the analytical treatment of this problem . in the present section
, we use a combination of the analytic and numerical techniques to elucidate the origins of the complex dynamics arising in the parameter regime near the border between regions of sub- and supercritical ahb .
the principal features of this bifurcation scenario are summarized in figures [ f.4 ] and [ f.5 ] . in these numerical experiments
, we kept @xmath22 fixed at a small positive value and varied @xmath51 . by taking progressively smaller values of @xmath50
, one first observes that the multimodal patterns exhibit an increase in the isis ( figure [ f.4]a , b ; see also figure [ f.3]b ) .
we consider these oscillatory patterns regular ( even if they are not periodic ) , because the timings of the spikes remain within narrow bounds in accord with ( [ 4.8a ] ) . for smaller values of @xmath50 ,
the oscillatory patterns become irregular and are characterized by long intervals of oscillations of small and intermediate amplitudes between successive spikes ( figure [ f.4 ] c , d ) . as
@xmath51 becomes negative , the trajectories lose spikes and consist of irregular oscillations ( figure [ f.5 ] a , b ) .
further decrease of @xmath51 leads the system through the reverse cascade of period - doubling bifurcations ( figure [ f.5 ] c - h ) .
the period - doubling cascade terminates with the creation of the limit cycle , which can be followed to a nondegenerate supercritical ahb by letting @xmath228 ( figure [ f.5 ] g , h ) . to account for the bifurcation scenario described above
, we construct the first return map .
since near the origin the trajectories spend most of the time in the vicinity of the @xmath2 slow manifold , the first return map is effectively one - dimensional .
the distinct unimodal structure of the @xmath1 first return map affords a lucid geometric interpretation for the bifurcation scenario in the @xmath229 systems of differential equations near the transition from sub- to supercritical ahb .
specifically , we show that the mechanism for generating complex dynamics in the continuous system in this parameter regime is the same as in the classical scenario of the period - doubling transition to chaos in the one - parameter families of unimodal maps @xcite .
* a * + * c * we next turn to the derivation of the first return map .
we start with extending the asymptotic analysis of section 3 to cover the case of @xmath230 .
this requires computing additional terms on the right hand side of the reduced equation ( [ 5.24 ] ) , because already for @xmath231 , the term involving @xmath51 on the right hand side of ( [ 5.24 ] ) is comparable with the @xmath232 remainder term .
we then use the more accurate approximation for the equation for @xmath155 to compute a @xmath1 mapping : g : i()i(+ 2 ^ -1 ) , which describes how @xmath155 changes after one cycle of oscillations . as noted above , the reduced equation ( [ 5.24 ] ) is not suitable for analyzing the case of small @xmath233 and one needs to include more terms in the expansion on the right hand side of ( [ 5.24 ] ) . in analogy with the topological normal form for the degenerate ahb @xcite
, we expect that terms up to @xmath234 ) are needed in ( [ 5.24 ] ) to resolve the dynamics for small @xmath233 .
this can be achieved by a straightforward albeit tedious calculation .
below we describe the formal procedure of obtaining the required expansions .
the justification of these expansions can be done in complete analogy with the analysis of section 3 .
first , we compute terms on the right hand sides of ( [ 5.11a ] ) and ( [ 5.12 ] ) up to @xmath232 and @xmath235 respectively
. then we look for a solution of ( [ 5.11a ] ) in the following form : ( ) = _ 0(r,)+_1(r,)+^2_2(r,)+o(^3 ) . by taking an initial condition from @xmath61 and using anzats ( [ 9.1 ] ) in ( [ 5.11a ] ) ,
we recover @xmath236 ( see ( [ 5.16 ] ) ) and find the next two terms in the expansion of @xmath237 : @xmath238 and @xmath239 .
thus , we obtain the approximation of the slow manifold with the accuracy @xmath232 .
next , we plug in ( [ 9.1 ] ) into ( [ 5.12 ] ) and collect terms multiplying equal powers of @xmath49 to obtain r = r_1(r,)+
+^5 r_5(r,)+o(^6 ) , where @xmath240 are @xmath187 periodic functions of the second argument . finally , by rewriting ( [ 9.2 ] ) in terms of @xmath155 ( see ( 3.25 ) ) , integrating it over one period of oscillations , @xmath241 , and disregarding @xmath234 terms , we obtain a map for the change in @xmath155 after one cycle of oscillations i_n+1=g(i_n ) , g(i)i-2 ^ 2(i + + ^2ci ) , = 2 ^ -1 . except for the last term in the definition of @xmath242 , the map in ( [ 9.4 ] )
follows from the reduced system ( [ 5.24 ] ) and ( [ 5.25 ] ) .
the last term is only needed if the value of @xmath233 does not exceed @xmath62 .
our calculations show that the second lyapunov coefficient , @xmath243 , is negative for the values of parameters used in ( [ 6.1])-([6.3 ] ) .
therefore , for sufficiently small @xmath47 , ( [ 9.4 ] ) defines a unimodal map .
away from an @xmath79 neighborhood of @xmath28 , the graph of @xmath244 is almost linear with a weakly attracting slope @xmath245 for @xmath246 ( figure [ f.7 ] ) . for @xmath246
, @xmath242 has a unique fixed point : |i=\ { cc - + c^2+o(^3 ) , & 0 . . for negative values of @xmath247
, @xmath248 is a stable fixed point of @xmath242 , which corresponds to a stable limit cycle born from a supercritical ahb ( see figure [ f.5 ] g , h ) .
it follows from ( [ 9.4 ] ) that for increasing values of @xmath51 , the graph of the map moves down .
consequently , the fixed point moves to the left and eventually looses stability via a period - doubling bifurcation ( figure [ f.5 ] e , f ) .
further increase in @xmath51 , yields more period - doubling bifurcations ( figure [ f.5 ] c , d ) . given the unimodal character of @xmath242 , one expects that this sequence of period - doubling bifurcations eventually leads to the formation of a chaotic attractor ( figure [ f.5 ] a , b ) .
the unimodal character of the map combined with the manner of its dependence on @xmath51 suggest a clear geometric mechanism for the formation of the chaotic attractor and the subsequent period - doubling cascade arising near the transition from sub- to supercritical ahb ( figure [ f.5]a - h ) .
we now outline the limitations of the asymptotic analysis of this section and difficulties arising in the justification of ( [ 9.4 ] ) . according to ( [ 9.5 ] ) already at the moment of the first period - doubling bifurcation , fixed point @xmath169 belongs to an @xmath79 neighborhood of @xmath28 . in this neighborhood , @xmath249 ( see ( [ 5.22 ] ) ) and , therefore , @xmath169 lies outside of the region of validity of the asymptotic analysis . in this region , ( [ 9.4 ] ) may only be considered as a formal asymptotic expression .
a rigorous justification of the from of the map in the boundary layer meets substantial analytical difficulties : it requires introducing an additional set of intermediate asymptotic expansions and matching them with those obtained in section 3 .
we do not address this problem in the present work .
below , we resort to using numerical techniques to verify the principal features of the first return map suggested by the asymptotic analysis : the unimodality of @xmath242 and its dependence on @xmath51 .
the numerics confirmed our predictions about the form of the first return map and it also revealed certain additional features . for numerical construction of the first - return map
, we fix angle @xmath250 and define a crossection in cylindrical coordinates : @xmath251 let @xmath252 be a trajectory of ( [ 6.5 ] ) and @xmath253 denote a sequence of times , at which @xmath254 .
then , we define : ( t_k ) ( t_k+1)(t ) = . by a suitable choice of @xmath255
, we can achieve that a trajectory of ( [ 6.5 ] ) , intersects @xmath256 once during one cycle of oscillations .
therefore , for a trajectory lying in a @xmath2 slow manifold , ( [ 9.6 ] ) defines a @xmath1 first - return map .
note that maps ( [ 9.4 ] ) and ( [ 9.6 ] ) are related via rescaling of the coefficients , since to leading order @xmath257 .
therefore , the graphs of @xmath258 and @xmath259 are similar in the domain , where ( [ 9.4 ] ) is valid .
we computed the first return maps for fixed @xmath260 and for several values of @xmath51 .
the representative plots are shown in figure [ f.8 ] a - c .
the numerically computed maps confirmed our expectations about the graph of @xmath261 in the boundary layer near @xmath28 : in this region , the map is decreasing with a strongly expanding slope . away from @xmath28 , the graph of @xmath261
contains an almost linear branch , whose slope is very close to @xmath262 as predicted by ( [ 9.4 ] ) ( see lower branches in figure [ f.8]b , c ) .
these numerics also confirm that under the variation of @xmath51 , the graph of the first return map is translated in vertical direction ( figure [ f.8 ] a - c ) .
therefore , the family of the first return maps possesses two principal ingredients ( the unimodality and the additive dependence on @xmath51 ) , which are necessary for the qualitative explanation of the bifurcation scenario given in figure [ f.7 ] .
in addition , our numerical experiments reveal a new feature of the first return map : for small @xmath263 , the graph of the map has another almost linear branch in the outer region away from the origin ( see figure [ f.8]b ) .
this branch of the graph of the map also has an attracting slope .
the presence of this branch in the first return map indicates the existence of another ( branch of the ) slow manifold different from that described in section 3 .
a possible explanation for the appearance of the upper branch in the first return map is due to the unstable manifold of the periodic orbit . for small @xmath263 , the first - return map ( [ 9.6 ] )
is multivalued in the outer region . however , since both the upper and the lower branches have positive slopes less than @xmath264 , the qualitative dynamics for the map shown in figure [ f.8]b does not depend on the exact mechanism for the selection between the branches in ( [ 9.6 ] ) .
although explaining the multivaluedness of the first return maps shown in figures [ f.8]b , c presents an interesting problem , it is not critical for the qualitative explanation of the bifurcation scenario arising during the transition from the subcritical to supercritical ahb .
the latter was the main goal of the present subsection .
* remark 4.2 * in addition , to the region in the parameter space containing the border between the regions of sub- and supercritical ahb , there is another parameter regime resulting in the complex dynamics . in @xcite
, it was shown numerically that the limit cycle born from the supercritical ahb in ( [ 6.5 ] ) undergoes a period - doubling cascade leading to the formation of the chaotic attractor for increasing values of @xmath22 .
the first period - doubling bifurcation in this cascade is shown in figure [ f.1]c , d .
this bifurcation scenario is consistent with the form of the first - return map constructed in this subsection .
indeed , it is easy to see from ( [ 9.5 ] ) that for increasing values of @xmath55 the fixed point , @xmath169 , moves to the left .
therefore , the explanation given above for the period - doubling cascade resulting from the variation of @xmath51 near @xmath28 also applies to the case of increasing @xmath22 and fixed @xmath265 ( see figure [ f.8]d ) .
in the present section , we show that under the assumptions of theorem 4.1 the trajectories of ( [ 2.5 ] ) with initial conditions in @xmath41 enter @xmath61 .
the local analysis in section 3 describes the behavior of trajectories from the moment they reach @xmath61 and until they leave @xmath57 . in particular , it shows that the dynamics near the origin has two phases : the fast approach to the slow manifold and the slow drift away from the origin along the slow manifold . upon leaving @xmath57 ,
the trajectories are reinjected back to @xmath41 by the return mechanism postulated in * ( g1)*. this scenario implies that the system undergoes multimodal oscillations as stated in theorem 4.1 .
we start with presenting several auxiliary estimates , which will be needed for the proof . by ( [ 4.3a ] )
, one can choose @xmath120 such that w(x_1 , ) c_1 |x_1|,x_1 and for sufficiently small @xmath44 .
next , we note that @xmath266 .
this equation and ( [ 4.6 ] ) , by the implicit function theorem , imply that there exists @xmath267 such that @xmath268 .
\end{array } \right.\ ] ] using ( [ 4.6 ] ) and ( [ 7.1 ] ) , we have @xmath269,\ ] ] where @xmath270 is defined in ( [ 4.3a ] ) .
let @xmath271 be such that ( * g1 * ) and ( * g2 * ) hold for @xmath272 $ ] . in the remainder of this section ,
unless stated otherwise , it is assumed that @xmath273 $ ] . to simplify notation
, we will often omit the dependence of various functions on @xmath22 . using ( [ 4.3a ] ) ,
we rewrite ( [ 6.5 ] ) in the following form : @xmath274 where @xmath275 and @xmath276 are defined in ( [ 4.3a ] ) and ( [ 4.5 ] ) respectively ; @xmath277 , and w(0)=0 , ( x_1,y)=o(|y|^2 ) , _ 1,2(x_1,y)=_1,2(x_1)+o(|y|),x .
let d_3=-m^2 ^ 2 denote the @xmath194coordinate on the left lateral boundary of @xmath61 ( see ( [ 5.2 ] ) ) and @xmath278\times d_{\delta_1}$ ] ( see figure [ f.9 ] ) . from ( [ 6.5 ] ) and
( [ 4.3a ] ) , one can see that for @xmath279 sufficiently small ( independent of @xmath49 ) , the right hand side of ( [ 7.2 ] ) @xmath280 therefore , in @xmath281 , ( [ 7.2 ] ) and ( [ 7.3 ] ) may be rewritten as follows = a(x_1 ) y + ( x_1,y ) , where @xmath282 by ( [ 7.1 ] ) and ( [ 7.4 ] ) , we have |(x_1,y)|,(x_1,y)__1 , for some @xmath137 independent of @xmath49 . to follow the trajectories from @xmath41 to @xmath61 , we introduce two regions : @xmath283\times d_{\delta_1}\quad\mbox{and}\quad\pi_{\delta_2}^2=[d_2,d_3]\times d_{\delta_2}\;\;\mbox{(see figure \ref{f.9})}.\ ] ] recall that @xmath284 denotes the disk of radius @xmath285 ( see ( [ 4.4a ] ) ) .
positive constant @xmath207 is the same as in ( [ 4.4 ] ) and @xmath286 will be specified later ( figure [ f.9 ] ) .
recall that @xmath287 denotes the value of the @xmath194coordinate on the the left lateral boundary of @xmath61 ( see ( [ 5.2 ] ) ) . by taking @xmath69 large enough
, we can arrange @xmath288 for @xmath289 $ ] . by * ( g1 ) * , the vector field in @xmath290 is sufficiently strong so that the trajectories entering @xmath291 through @xmath41 get into a narrow domain @xmath292 and remain there until they reach @xmath61 .
the following lemma allows to control the trajectories in @xmath278\times d_{\delta_1}$ ] .
* lemma 5.1 * let @xmath293\subseteq[d_1,d_3],\ ; 0<\bar\delta\le { \delta_1},$ ] and -=_x_1_|y|=1(a(x_1)y , y ) < 0 .
then |y(_1)|||y(x_1)|2|e^-(x_1-_1 ) , x_1 , provided 1 .
* proof : * we may assume that @xmath294 $ ] , since , otherwise , by the uniqueness of solution of the initial value problem for ( [ 7.5 ] ) , @xmath295 @xmath296 $ ] , and ( [ 7.8 ] ) holds . from ( [ 7.5 ] ) , ( [ 7.6 ] ) , and ( [ 7.7 ] ) , we have -|y(x_1)| + @xmath297 is defined in ( [ 7.7 ] ) , and @xmath298 denotes the euclidean norm of @xmath299 .
let @xmath300 denote the solution of the initial value problem |y^=-|y + |(x_1,|y ) , |y(_1)=| .
it is sufficient to show that ( [ 7.8 ] ) holds for @xmath300 .
we represent @xmath300 as the limit of the sequence of successive approximations |y^(n+1)(x_1)=|e^-(x_1-_1)+__1^x_1 e^-(x_1-s ) @xmath301 we use induction to show ||y^(n)(x_1)| 2 |e^-(x_1-_1),x , n=0,1,2, . inequality ( [ 7.13 ] ) holds for @xmath302 .
we show that @xmath303 .
using the definition of @xmath304 in ( [ 7.10 ] ) and @xmath305 , we have @xmath306 using ( [ 7.14 ] ) and ( [ 7.9 ] ) , from @xmath305 we obtain @xmath307 . by induction , ( [ 7.13 ] )
holds . by taking @xmath308 in ( [ 7.13 ] ) , we have ( [ 7.15 ] ) and the standard theory for differential inequalities @xcite .
+ @xmath128 in the following lemma , we determine the size of @xmath309 .
* lemma 5.2 * there exists positive @xmath310 , such that the trajectories of ( [ 7.2 ] ) and ( [ 7.3 ] ) entering @xmath309 from @xmath311 remain in @xmath309 until they reach @xmath61 ( see figure [ f.9 ] . ) * proof : * denote @xmath312}\sup_{\left|y\right|=1}\left(\tilde a(x_1)y , y\right).\ ] ] recall @xmath313 therefore , by ( [ 4.6 ] ) , @xmath314 is positive . in addition , @xmath315 } \sup_{\left|y\right|=1}\left(\tilde a(x_1)y , y\right)= \min_{x_1\in[d_2,d_3]}{-\sup_{\left|y\right|=1}\left ( a(x_1)y , y\right)\over w(x_1)}\\ & = & { \min_{x_1\in[d_2,d_3 ] } \bar\lambda(x_1)\over\max_{x_1\in[d_2,d_3 ] } w(x_1)}= { \bar\lambda(d_3)\over\max_{x_1\in[d_2,d_3 ] } w(x_1)}= o(\epsilon^2 ) .
\end{aligned}\ ] ] and @xmath316 .
next we apply lemma 5.1 with @xmath317 @xmath318 , and @xmath319 .
note that for small @xmath320 the inequality ( [ 7.9 ] ) can be rewritten as 1 . since @xmath321 ( see ( [ 7.4a ] ) ) and @xmath322
, one can choose @xmath310 so that ( [ 7.16 ] ) holds .
+ @xmath128 having found the size of @xmath309 , we now determine the rate of contraction in @xmath290 sufficient to funnel the trajectories entering @xmath290 through @xmath41 to @xmath309 .
* lemma 5.3 * the trajectories of ( [ 7.2 ] ) and ( [ 7.3 ] ) entering @xmath290 through @xmath41 remain in @xmath290 until they reach @xmath309 .
* proof : * denote @xmath323}\sup_{\left|y\right|=1}\left(\tilde a(x_1)y , y\right).\ ] ] from * ( g2 ) * , one finds that @xmath324 is positive .
lemma 5.3 now follows from lemma 5.1 with @xmath325 @xmath326 , and @xmath327 .
indeed , for @xmath328 , we have _ 1 . inequality ( [ 7.17 ] ) is sufficient for ( [ 7.9 ] ) to hold . by lemma 5.1
, we have |y(d_2)|2_1 e^-_1(d_2-d_1 ) . with @xmath328 , by ( [ 7.19 ] ) , we can achieve @xmath329 .
+ @xmath128 lemmas 5.2 and 5.3 imply that the trajectories entering @xmath290 through @xmath41 stay in @xmath330 until they reach @xmath61 .
moreover , the inequality in * ( g1 ) * guarantees that such trajectories are bounded away from @xmath42 .
thus , we can use the analysis of section 3 to describe the evolution of trajectories from the moment they reach @xmath61 until they leave @xmath57 . by remark 3.1c
, this description extends to any region where @xmath331 i.e. , the trajectories can be controlled until they hit @xmath40 .
this is followed by the return to @xmath291 according to * ( g1 ) * , and the next cycle of the multimodal oscillations begins .
the analysis of this section applies to any trajectory starting from @xmath61 and not belonging to @xmath42 .
it remains to estimate the isis .
for this , we compute the time needed for the trajectory starting in an @xmath79 neighborhood to return back to this neighborhood after making one global excursion .
since the time of flight of the trajectory outside a small neighborhood of the origin depends regularly on the control parameters , the duration of the very long isis is determined by the time spent in the neighborhood of the origin . to estimate the latter
, we note that at the moment a trajectory enters @xmath61 , we have |y(d_3)|c_3 ^2 .
this follows from lemma 5.3 , since @xmath332 and @xmath333 to obtain the lower bound on @xmath334 , recall that by * ( g1 ) * , we have |y(d_1)|>0 . as follows from * ( g3 ) * , the maximal rate of contraction in @xmath281 does not exceed @xmath70 in absolute value .
this combined with ( [ 7.20b ] ) implies that |y(d_3)|c_4 ^ 2 + 2 for some @xmath335 and @xmath336 independent of @xmath49 .
we omit the proof of ( [ 7.20c ] ) , because it is completely analogous to that of lemma 5.1 .
let @xmath337 denote the moment of time when a trajectory of ( [ 6.5 ] ) enters @xmath61 from @xmath309 .
after switching back to the original parametrization of @xmath338 by time , @xmath339 , we rewrite ( [ 7.20a ] ) and ( [ 7.20c ] ) : c_3 ^ 2 + 2|y(t^-)|c_4 ^ 2 .
for @xmath340 the trajectory approaches and remains close to the slow manifold as long as @xmath341 ( see remark 3.1c ) .
let @xmath342 and denote t^+=\{t > t^-:|y(t)|=^j}. from ( [ 7.20 ] ) and ( [ 7.21 ] ) , we have the following bounds for @xmath343 and @xmath344 : c_6 ^ -2-i_0c_7 ^ -2 i_1=o(^2(1-j ) ) , j(23 , 1 ) .
for these ranges of values of @xmath168 and @xmath345 , from ( [ 5.55 ] ) we have _ in = t^+-t^-=12 ^ 2((1+i_0 ) ( 1+o(1 ) ) ) . the combination of ( [ 7.22 ] ) and ( [ 7.23 ] ) yields two - sided bounds for @xmath346 : ^-_in_in_in^+ , where @xmath347 where positive constants @xmath348 can be chosen independent of @xmath49
. inequalities ( [ 7.24 ] ) provide bounds for the time that a multimodal trajectory spends near the unstable equilibrium . on the other hand , the time of flight outside a small neighborhood of the origin , @xmath349 , depends regularly on @xmath49 , by * ( g2)*. therefore , by adjusting constants @xmath350 in ( [ 7.24 ] ) if necessary , one can obtain uniform bounds on the isis @xmath351 for sufficiently small @xmath47 , positive @xmath51 and @xmath55 from bounded intervals , as stated in theorem 4.1 .
in the present paper , we investigated a mechanism for generation of multimodal oscillations in a class of systems of differential equations close to an ahb .
our analysis covers both cases of sub- and supercritical ahb . for the supercritical case
, we identified a novel geometric feature of the bifurcating limit cycle , the frequency doubling effect .
it turns out that generically in the normal system of coordinates the oscillations in one of the variables are twice faster than in the remaining two variables .
therefore , the leading order approximation of the limit cycle bifurcating from the supercritical ahb requires two harmonics .
the asymptotic analysis of the present paper explains the frequency - doubling .
in addition , we provide a complementary geometric interpretation to this counterintuitive effect .
in particular , we showed that it is a consequence of the geometry of the limit cycle .
the latter is not captured by the topological normal form of the ahb .
the analysis of the multimodal oscillations arising from the subcritical ahb requires additional assumptions on the global behavior of trajectories .
we identified two principal properties of the global vector field : the mechanism of return and the strong contraction property . in the presence of this global structure
, the subcritical ahb produces sustained multimodal oscillations combining the small amplitude oscillations near the unstable equilibrium with large amplitude spikes .
the resultant motion is recurrent in a weak sense : it may not be periodic but nevertheless the timings of the spikes possess certain regularity .
we have shown that the isis have well - defined asymptotics near the ahb and comply to the two - sided bounds , which depend on the principal bifurcation parameters .
our estimates show that near the ahb , the isis can be extremely long and can change greatly under relatively small variation of the bifurcation parameters . the ability of the system to exhibit such extreme variability in the isi duration is important in many applications , in particular , in the context of neuronal dynamics .
previous studies investigated different possible mechanisms for generating multimodal patterns with very long isis @xcite .
for the finite dimensional approximation of the model of solid - fuel combustion ( [ 6.1])-([6.3 ] ) , the main motivating example for our work , the proximity to the homoclinic bifurcation was suggested in @xcite as a possible mechanism for prolonged isis .
our conclusions confirm the importance of the proximity to the homoclinic bifurcation for explaining the oscillatory patterns in ( [ 6.1])-([6.3 ] ) .
the proximity to the homoclinic bifurcation is implicitly reflected in our assumptions on the global vector field .
however , we emphasize the critical role of the ahb : the duration of the isis can be effectively controlled by the parameters associated with the ahb without changing the distance of the system to the homoclinic bifurcation . in all our numerical experiments , the system remained bounded away from the homoclinic bifurcation , nevertheless it exhibited patterns with very long isis whose duration was amenable to control .
our analysis extends the estimate for the isis obtained in @xcite to a wide class of problems .
it also emphasizes the proximity of the system to the border between sub- and supercritical ahb , as another factor in creating oscillatory patterns with long isis .
we show that as this border is approached from the subcritical side the isi grow logarithmically .
this observation is important for explaining the oscillatory patterns generated by ( [ 6.1])-([6.3 ] ) since in this model the ahb can change its type under the variation of the second control parameter @xmath9 .
this situation is not specific to the model of solid fuel combustion .
recent studies suggest that there is a class of neuronal models close to the ahb whose type may change with the values of parameters @xcite .
therefore , it is important to understand the principal features of the transition from sub- to supercritical ahb .
we found that when the border between the regions of sub- and supercritical bifurcation is approached from the subcritical side ( while the distance from the ahb remains fixed ) , the oscillations become chaotic .
the regime of irregular oscillations is then followed by the reverse period doubling cascade . to understand the nature of this bifurcation scenario we used a combination of analytic and numerical techniques . using the insights gained from the asymptotic analysis , we constructed a @xmath1 first - return map .
the map provides a clear geometric interpretation for the bifurcation scenario near the transition from sub- to supercritical ahb .
our study suggests that the formation of the chaotic attractor via a period - doubling cascade is a universal feature of this transition .
for example , the bifurcation scenarios reported for the hodgkin - huxley model in @xcite are very similar to those studied in the present paper and are likely to share the same mechanism .
mixed - mode oscillations similar to those studied in the present paper , have been studied in for a class of the slow - fast systems in @xmath0 near the ahb .
although , the work toward developing a complete mathematical theory for such oscillations is still in progress , the general mechanism for their generation and the bifurcation structure of the problem have been greatly elucidated recently @xcite .
the present paper shows the relation between the mechanisms for the mixed - mode oscillations in the model in @xcite and for those in the slow - fast systems .
the latter possess a well - defined structure of the global vector field due to the presence of the disparate timescales in the governing equations @xcite .
the analyses of the mixed - mode oscillations in @xcite use in an essential way the relaxation structure of the problem .
the model in @xcite is an example of the mixed - mode generating system , which does not possess an explicit relaxation structure .
in fact , it is hard to expect such structure in a system obtained via projecting an infinite - dimensional system onto a finite - dimensional subspace . in formulating the assumptions on the global vector - field ( * g1 * ) and ( * g2 * ) , we were looking for the minimal requirements on the system near an ahb that guarantee the existence of the mixed - mode solutions . due to the lack of the information about the global vector field of ( [ 6.1])-([6.3 ] )
, it appears impossible to verify these conditions analytically .
however , the numerical simulations clearly show that system of equations ( [ 6.1])-([6.3 ] ) possesses the qualitative structure required by ( * g1 * ) and ( * g2 * ) ( figure [ f.2]b ) . on the other hand
, we expect that conditions ( * g1 * ) and ( * g2 * ) should be possible to verify analytically for a wide class of slow - fast systems
. therefore , we believe that our results will be useful for understanding mixed - mode oscillations in such systems . in particular , it would be interesting to apply this approach to the modified hodgkin - huxley system @xcite .
the numerical results reported in @xcite strongly suggest that the mechanism proposed in the present paper is responsible for the generation of the very slow rhythms and chaotic dynamics in the hodgkin - huxley model . *
acknowledgments . *
we thank victor roytburd for introducing us to this problem and to michael frankel and victor roytburd for helpful conversations .
this work was partially supported through national science foundation award no 0417624 .
[ sec : c ] in this appendix , we list explicit expressions of various constants and trigonometric polynomials , which appear in the definitions in of the slow manifold , @xmath58 , and the first lyapunov coefficient , @xmath51 .
all expressions are given in terms of the coefficients of the power expansions on the right hand side of ( [ 5.1 ] ) . by @xmath352
we denote @xmath353 @xmath354 .
the following constants are used to define the leading order approximation of the slow manifold in ( [ 5.3 ] ) and ( [ 5.4 ] ) : @xmath355 the following trigonometric polynomials enter the right hand sides of ( [ 5.11a ] ) and ( [ 5.12 ] ) : @xmath356 functions @xmath357 are used in the calculation of the first lyapunov coefficient @xmath51 ( see ( [ 5.5 ] ) ) .
j. drover , j. rubin , j. su , and b. ermentrout , analysis of a canard mechanism by which excitatory synaptic coupling can synchronize neurons at low firing frequencies , _ siam j. appl .
_ , @xmath360 , 6992 , 2004 . e.i .
volkov and d.v .
volkov , multirhythmicity generated by slow variable diffusion in a ring of relaxation oscillators and noise - induced abnormal interspike variability , _ phys .
e _ , @xmath360 , 2002 . | one- and two - parameter families of flows in @xmath0 near an andronov - hopf bifurcation ( ahb ) are investigated in this work .
we identify conditions on the global vector field , which yield a rich family of multimodal orbits passing close to a weakly unstable saddle - focus and perform a detailed asymptotic analysis of the trajectories in the vicinity of the saddle - focus .
our analysis covers both cases of sub- and supercritical ahb . for the supercritical case , we find that the periodic orbits born from the ahb are bimodal when viewed in the frame of coordinates generated by the linearization about the bifurcating equilibrium .
if the ahb is subcritical , it is accompanied by the appearance of multimodal orbits , which consist of long series of nearly harmonic oscillations separated by large amplitude spikes .
we analyze the dependence of the interspike intervals ( which can be extremely long ) on the control parameters .
in particular , we show that the interspike intervals grow logarithmically as the boundary between regions of sub- and supercritical ahb is approached in the parameter space .
we also identify a window of complex and possibly chaotic oscillations near the boundary between the regions of sub- and supercritical ahb and explain the mechanism generating these oscillations .
this work is motivated by the numerical results for a finite - dimensional approximation of a free boundary problem modeling solid fuel combustion . |
the classical theory by hodgkin and huxley ( hh ) describes nerve impulses ( spikes ) that manifest communication between nerve cells .
the underlying mechanism of a single spike is excitability , i.e. , a small disturbance triggers a large excursion that reverts without further input to the original state .
a spike lasts a 1/1000 second and even though during this period ions are exchanged across the nerve cell membrane , the change in the corresponding ion concentrations can become significant only in series of such spikes . under certain pathological conditions
changes in ion concentrations become massive and last minutes to hours before they recover .
this establishes a new type of excitability underlying communication failure between nerve cells during migraine and stroke . to clarify this mechanism and to recognize the relevant factors that determine the slow time scales of ion changes
, we use an extended version of the classical hh theory .
we identify one variable of particular importance , the potassium ion gain or loss through some reservoirs provided by the nerve cell surroundings .
we suggest to describe the new excitability as a sequence of two fast processes with constant total ion content separated by two slow processes of ion clearance ( loss ) and re uptake ( re gain ) .
in this paper we study ion dynamics in ion based neuron models . in comparison to classical hh type membrane models
this introduces dynamics on much slower time scales . while spiking activity is in the order of milliseconds , the time scales of ion dynamics range from seconds to minutes and even hours depending on the process ( transmembrane fluxes , glial buffering , backward buffering ) .
the slow dynamics leads to new phenomena .
slow burst modulation as in seizure like activity ( sla ) emerges from moderate changes in the ion concentrations .
phase space excursions with large changes in the ionic variables establish a new type of ionic excitability as observed in cortical spreading depression ( sd ) during stroke and in migraine with aura @xcite .
such newly emerging dynamics can be understood from the phase space structure of the ion based models .
mathematical models of neural ion dynamics can be divided into two classes . on the one hand
the discovery of sd by leo in 1944 @xcite a severe perturbation of neural ion homeostasis associated with a huge changes in the potassium , sodium and chloride ion concentrations in the extracellular space ( ecs)@xcite that spreads through the tissue has attracted many modelling approaches dealing with the propagation of large ion concentration variations in tissue . in 1963 grafstein described spatial potassium dynamics during sd in a reaction diffusion framework with a phenomenological cubic rate function for the local potassium release by the neurons@xcite .
reshodko and burs proposed an even simpler cellular automata model for sd propagation@xcite . in 1978
tuckwell and miura developed a sd model that is amenable to a more direct interpretation in terms of biophysical quantities @xcite .
it contains ion movements across the neural membrane and ion diffusion in the ecs . in more recent studies
dahlem et al
. suggested certain refinements of the spatial coupling mechanisms , e.g. , the inclusion of nonlocal and time delayed feedback terms to explain very specific patterns of sd propagation in pathological situations like migraine with aura and stroke @xcite . on the other hand single cell ion dynamics
were studied in hh like membrane models that were extended to include ion changes in the intracellular space ( ics ) and the ecs since the 1980s . while the first extensions of this type were developed for cardiac cells by difranceso and noble@xcite , the first cortical model in this spirit
was developed by kager , wadman and somjen ( kws)@xcite only in 2000 .
their model contains abundant physiological detail in terms of morphology and ion channels , and was in fact designed for seizure like activity ( sla ) and local sd dynamics .
it succeeded spectacularly in reproducing the experimentally known phenomenology .
an even more detailed model was proposed by shapiro at the same time@xcite who like yao , huang and miura for kws@xcite also investigated sd propagation with a spatial continuum ansatz . in the following hh like models of intermediate complexity were developed by frhlich , bazhenov et al . to describe potassium dynamics during epileptiform bursting@xcite .
the simplest hh like model of cortical ion dynamics was developed by barreto , cressman et al.@xcite who describe the effect of ion dynamics in epileptiform bursting modulation in a single compartment model that is based on the classical hh ion channels .
interestingly in none of these considerably simpler than shapiro and kws models extreme ion dynamics like in sd or stroke was studied .
to our knowledge the only exception is a study by zandt et al . who describe in the framework of cressman et al . what they call the `` wave of death '' that follows the anoxic depolarization after decapitation as measured in experiments with rats@xcite
] in this study we systematically analyze the entire phase space of such local ion based neuron models containing the full dynamical repertoire ranging from fast action potentials to slow changes in ion concentrations .
we start with the simplest possible model for sd dynamics a variation of the barreto , cressman et al .
model and reproduce most of the results for the kws model . our analysis covers sla and sd .
three situations should be distinguished : isolated , closed , and open systems , which is reminiscent of a thermodynamic viewpoint ( see fig . [
fig : system ] ) . an isolated system without transfer of metabolic energy for the atpase
driven @xmath0 pumps will attain its thermodynamic equilibrium , i.e. , its donnan equilibrium . a closed neuron system with functioning pumps but without ion regulation by glia cells or the vascular system is generally bistable@xcite .
there is a stable state of free energy
starvation ( fes ) that is close to the donnan equilibrium and coexists with the physiological resting state .
the ion pumps can not recover the physiological resting state from fes .
we will now develop a novel phase space perspective on the dynamics in open neuron systems .
we describe the first slow fast decomposition of local sd dynamics , in which the ion gain and loss through external reservoirs is identified as the crucial quantity whose essential importance was not realized in earlier studies . treating this slow variable as a parameter allows us to derive thresholds for sd ignition and the abrupt , subsequent repolarization of the membrane in a bifurcation analysis for the first time .
moreover we analyze oscillatory dynamics in open systems and thereby relate sla and sd to different so called torus bifurcations .
this categorizes sla and sd as genuinely different though they are ` sibling ' dynamics as they both bifurcate from the same ` parent ' limit cycle in a supercritical and subcritical manner , respectively , which also explains the all or none nature of sd .
sla is gradual in contrast .
local ion dynamics of neurons has been studied in models of various complexity . reduced model types consist of an electrically excitable membrane containing gated ion channels and ion concentrations in an intra and an extracellular compartment @xcite .
transmembrane currents must be converted to ion fluxes that lead to changes in the compartmental ion concentrations .
such an extension requires ion pumps to prevent the differences between ics and ecs ion concentrations that are present under physiological resting conditions from depleting .
we consider a model containing sodium , potassium and chloride ions . the hh like membrane dynamics
is described by the membrane potential @xmath1 and the potassium activation variable @xmath2 .
the sodium activation @xmath3 is approximated adiabatically and the sodium inactivation @xmath4 follows from an assumed functional relation between @xmath4 and @xmath2 .
the ics and ecs concentrations of sodium , potassium and chloride ions are denoted by @xmath5 , @xmath6 and @xmath7 , respectively . in a closed system mass conservation
holds , i.e. , @xmath8 with @xmath9 and the ics / ecs volumes @xmath10 . together with the electroneutrality of ion fluxes across the membrane , i.e. , @xmath11 only two of the six ion concentrations are independent dynamical variables .
the full list of rate equations then reads @xmath12 they are complemented by six constraints on gating variables and ion concentrations : @xmath13 superscript 0 of indicates ion concentrations in the physiological resting state .
unless otherwise stated @xmath14 and @xmath15 are used as initial conditions in the simulations . constrained ion concentrations ( eqs .
( [ eq:5])([eq:8 ] ) ) then also take their physiological resting state values .
these ion concentrations , the membrane capacitance @xmath16 , the gating time scale parameter @xmath17 , the conversion factor @xmath18 from currents to ion fluxes , and the ics and ecs volumes @xmath10 are listed in tab .
[ tab:1 ] .
the conversion factor @xmath18 is an expression of the membrane surface area @xmath19 and faraday s constant @xmath20 ( both given in tab .
[ tab:1 ] , too ) : @xmath21 we remark that all parameters in tab .
[ tab:1 ] are given in typical units of the respective quantities .
the numerical values in these units can directly be used for simulations .
time is then given in msec , the membrane potential in mv and ion concentrations in @xmath22 .
the electroneutrality of the total transmembrane ion flux as expressed in eqs .
( [ eq:0 ] ) and ( [ eq:5 ] ) is a consequence of the large time scale separation between the membrane dynamics and the ion dynamics ( cf . ref .
@xcite and the below discussion of time scales ) .
this constraint is the reason why the thermodynamic equilibrium of the system must be understood as a donnan equilibrium .
this is the electrochemical equilibrium of a system with a membrane that is impermeable to some charged particles , which can be reached in an electroneutral fashion , i.e. , without separating charges .
we do not include this impermeant matter explicitly , because it does not influence the dynamics as long as osmosis is not considered .
one should however keep in mind that the initial ion concentrations in tab .
[ tab:1 ] do not imply zero charge in the ics or ecs and hence impermeant matter to compensate for this must be present .
the gating functions @xmath23 , @xmath24 and @xmath25 are given by @xmath26 here @xmath23 and @xmath25 are the asymptotic values and @xmath24 is potassium activation time scale .
they are expressed in terms of the hodgkin
huxley exponential functions@xcite @xmath27 the three ion currents are @xmath28 they are given in terms of the leak and gated conductances @xmath29 ( with @xmath30 ) and the nernst potentials @xmath31 which are computed from the ( dynamical ) ion concentrations @xmath32 : @xmath33 @xmath34 denotes the valence of the particular ion species .
the pump current modelling the atpase driven exchange of intracellular sodium with extracellular potassium at a @xmath35ratio is given by @xmath36 where @xmath37 is the maximal pump rate@xcite .
the pump current increases with @xmath38 and @xmath39 .
the values for the conductances and pump rate are also given in tab .
[ tab:1 ] .
let us remark that in comparisons with ref.@xcite , we have mildly increased the maximal pump rate and decreased the chloride conductance to obtain a sd threshold in agreement with experiments ( see sect .
results ) .
( [ eq:1])([eq:10 ] ) describe a closed system in which ion pumps are the only mechanism maintaining ion homeostasis and in which mass conservation holds for each ion species .
a remark on terminology is due at this point : a ` closed ' system refers exclusively to the conservation of the ion species that we model .
we do not directly model other mass transfer that occurs in real neural systems .
yet it is indirectly included . the ion pumps use energy released by hydrolysis of atp , a molecule whose components ( glucose and oxygen or lactate ) therefore have to pass the system boundaries . in thermodynamics
, it is customary to call systems that exchange energy but not matter with their environment closed . since atp is in this framework only considered as an energy source , we can describe the system as closed , if ions can not be transferred across its boundaries . as mentioned above the closed system
is bistable .
superthreshold stimulations cause a transition from physiological resting conditions to fes . to resolve this and change the behaviour to local sd dynamics
it is necessary to include further regulation mechanisms@xcite .
since sd is in particular characterized by an extreme elevation of potassium in the ecs we will only discuss potassium regulation .
if ecs potassium ions are subject to a regulation mechanism which is independent of the membrane dynamics , then the symmetry between ics and ecs potassium dynamics is broken and eq .
( [ eq:7 ] ) for the potassium conservation does not hold .
let us represent changes of the potassium content of the system by a variable @xmath40 which is defined by the following relation : @xmath41 changes of the potassium content , i.e. , changes of @xmath40 , can be of different physiological origin .
if glial buffering is at work the potassium content will be reduced by the amount of buffered potassium @xmath42 .
an initial potassium elevation @xmath43 simply leads to an accordingly increased @xmath40 : @xmath44 for the coupling to an extracellular potassium bath or to the vasculature @xmath40 is a measure for the amount of potassium that has diffused into ( positive @xmath40 ) or out of ( negative @xmath40 ) the system .
we are going to discuss two regulation schemes coupling to an extracellular bath and glial buffering .
they could be implemented simultaneously , but for our purpose it will suffice to apply only one scheme at a time . in the second subsection of sect .
results , the dynamics of @xmath40 is given by glial buffering , while in the third subsection we will discuss the oscillatory regimes one finds for bath coupling with elevated bath concentrations .
to implement glial buffering we assume a phenomenological chemical reaction of the following type@xcite : @xmath45 the buffer concentration is denoted by @xmath46 .
we are using the buffer model from ref .
@xcite in which the potassium
dependent buffering rate @xmath47 is given as @xmath48 the parameter @xmath49 is normally assumed to have the same numerical value as the constant backward buffering rate @xmath50 which is hence an overall parameter for the buffering strength .
however , the parameters should be denoted differently as they have different units ( cf . tab .
[ tab:1 ] ) . this chemical reaction scheme
together with the mass conservation constraint @xmath51 where @xmath52 is the initial buffer concentration , leads to the following differential equation for @xmath42 : @xmath53 eq .
( [ eq:25 ] ) the implies the following rate equation for @xmath40 @xmath54 where @xmath42 and @xmath39 are given by eqs .
( [ eq:25 ] ) and ( [ eq:24 ] ) , respectively . to model the coupling to
a potassium bath one normally includes an explicit rate equation for the ecs potassium concentration @xmath55 where the diffusive coupling flux @xmath56 is defined by its coupling strength @xmath57 and the potassium bath concentration @xmath58 .
( [ eq:24 ] ) implies that eq .
( [ eq:31 ] ) can be rewritten in terms of @xmath40 as follows : @xmath59 note that we have chosen to formulate ion regulation in terms of @xmath40 rather than @xmath39 which would be completely equivalent .
this is crucial , because the dynamics of @xmath40 happens on a time scale that is only defined by the buffering or the diffusive process , while @xmath39 dynamics involves transmembrane fluxes and reservoir coupling dynamics at different time scales ( cf .
the last paragraph of this section ) .
this can be seen from eq .
( [ eq:31 ] ) .
both regulation schemes glial buffering given by eq . ( [ eq:30 ] ) and coupling to a bath with a physiological bath concentration as in eq .
( [ eq:33])can be used to change the system dynamics from bistable to ionically excitable , i.e. , excitable with large excursions in the ionic variables . like all other system parameters the regulation parameters @xmath50 and @xmath57
are given in tab .
[ tab:1 ] .
they have been adjusted so that the duration of the depolarized phase is in agreement with experimental data on spreading depression .
note that the parameters we have chosen are up to almost one order of magnitude lower than intact brain values like the ones used in refs .
while this does not affect the general time scale separation between glial or vascular ion regulation and ion fluxes across the cellular membrane , the duration of sd depends crucially on these parameters .
however , during sd oxygen deprivation will weaken glial buffering , and the swelling of glia cells and blood vessel constriction will restrict diffusion to the vasculature .
such processes can be included to ion based neuron models and make ion regulation during sd much slower@xcite . for our purpose
it is however sufficient to assume smaller values from the beginning .
we remark that the ion regulation schemes in our model only refer to vascular coupling and glial buffering .
lateral ion movement between the ecs of nearby neurons is a different diffusive process that determines the velocity of a travelling sd wave in tissue .
this is not described in our framework . in the following section we will demonstrate in detail how
@xmath40 can be understood as the inhibitory variable of this excitation process .
the above presented model is indeed the simplest ion based neuron model that exhibits local sd dynamics .
model simplicity is an appealing feature in its own right , but one might doubt the physiological relevance of such a reduced model .
our hypothesis is that it captures very general dynamical features of neuronal ion dynamics , and to confirm this we will compare the results obtained with the reduced model to the physiologically much more detailed kws model@xcite .
this detailed model contains five different gated ion channels ( transient and persistent sodium , delayed rectifier and transient potassium , and nmda receptor gated currents ) and has been used intensively to study sd and sla .
in fact , one modification is required so that we can replicate the results obtained from the reduced model .
the kws model contains an unphysical so
called fixed leak current @xmath60 that has a constant reversal potential of @xmath61 mv and no associated ion species .
this current only enters the rate equation for the membrane potential @xmath1 .
the effect on the model dynamics is dramatic . to see this note that the electroneutrality constraint eq .
( [ eq:6 ] ) reflects a model degeneracy @xmath62 that occurs when @xmath63 is modelled explicitly with @xmath64 ( for details see ref .
@xcite ) . with a fixed leak current eq .
( [ eq:1_new ] ) becomes @xmath65 which implies that @xmath66 mv is a necessary fixed point condition for the system .
in other words , the type of bistability with a second depolarized fixed point that we normally find in closed systems is ruled out by this unphysical current . if we , however , replace it with a chloride leak current as in our model ( cf.eqs .
( [ eq:4 ] ) and ( [ eq:21 ] ) ) , i.e. , a current with a dynamically adjusting reversal potential by virtue of eq .
( [ eq:22 ] ) , we find the same type of bistability for the closed system and monostability for the system that is buffered or coupled to a potassium bath . the morphological parameters ( compartmental volumes @xmath10 and membrane surface area @xmath19 ) are the same as for the reduced model .
in fact in ref .
@xcite the kws model was used without additional ion regulation for a reaction diffusion study of sd , and the only recovery mechanism of the local system seems to be this unphysical current . theoretically sd could be a travelling wave in a reaction diffusion system with bistable local dynamics , but unpublished results show that the propagation properties in the bistable system are dramatically different from standard sd dynamics with wave fronts and backs travelling at different velocities . we hence suppose that a local potassium clearing mechanism is crucially involved in sd .
we conclude this section with a discussion of the time scales of the model . to this end , it is helpful to keep in mind that the phenomenon of excitability requires a separation of time scales .
we have electrical and ionic excitability and these dynamics themselves are separated by no fewer than three orders of magnitude .
dynamics of @xmath1 happens on a scale that is faster than milliseconds .
this follows from the gating time scale @xmath24 which is given explicitly in eq .
( [ eq:13 ] ) and the time scale of @xmath67 of @xmath1 which can be computed from the membrane capacitance @xmath16 ( given in tab .
[ tab:1 ] ) and the resistance @xmath68 of the ion channels ( for details see ref .
@xcite ) : @xmath69 with @xmath70 if we approximate the products of gating variables in the above expression with 0.1 this gives @xmath71 .
dynamics of @xmath2 happens on a scale in the order of milliseconds .
the time scale of ion dynamics is more explicit in the goldman
katz ( ghk ) formalism than in the nernst formalism used in this paper .
the nernst currents in eqs .
( [ eq:19])([eq:21 ] ) are an approximation of the physically more accurate ghk currents , but in ref .
@xcite we have shown that ion dynamics of ghk models and nernst models are very similar .
that is why the latter may be used for studies like this . for time
scale considerations , however , we will now switch to the ghk description .
the ghk current of ions with concentrations @xmath32 across a membrane is given by @xmath72 where @xmath73 is the permeability of the membrane to the considered ion species and @xmath74 is the dimensionless membrane potential with @xmath75 this expression contains the ideal gas constant @xmath76 , the temperature @xmath77 , ion valence @xmath78 and faraday s constant @xmath20 . if we now write down the ghk analogue of the ion rate eqs .
( [ eq:3 ] ) and ( [ eq:4 ] ) we obtain @xmath79 for the conversion factor @xmath18 we have inserted the expression eq .
( [ eq:11 ] ) .
the fraction term is of the order of the ion concentrations , @xmath80 is a dimensionless quantity and hence of order one . with the ion dynamics time scale
@xmath81 we can thus group the parameters as follows @xmath82 permeabilities of ion channels can be found in refs@xcite .
similar as for the resistance @xmath68 the permeability @xmath83 of a gated channel involves a product of gating variables .
approximating such terms again with 0.1 a typical value for the permeability is @xmath84 .
together with the values for the membrane surface are and the cell volume from tab .
[ tab:1 ] the time scale of transmembrane ion dynamics is @xmath85 .
the slowest time scales are related to potassium regulation , i.e. , to @xmath40 dynamics . the glia scheme from eq .
( [ eq:26 ] ) and eq .
( [ eq:30 ] ) contains a forward buffering process that reduces @xmath40 at a time scale @xmath86 and a backward buffering process with time scale @xmath87 with the parameters from tab .
[ tab:1 ] this leads to @xmath88 and @xmath89 .
so backward buffering is much slower .
this is an important property , because in the following section we will see that recovery from fes requires a strong reduction of the potassium content .
if buffering and backward buffering would happen on the same time scale the required potassium reduction would not be possible .
backward buffering could well happen at a considerably faster scale than eq .
( [ eq:43 ] ) , but as soon as @xmath90 is comparable to @xmath91 the buffer can not re establish physiological conditions after fes .
the glia scheme here is phenomenological .
a more biophysically detailed model would describe a glial cell as a third compartment .
an elevation of ecs potassium leads to glial uptake .
spatial buffering , i.e. , the fast transfer of potassium ion between glia cells with elevated concentrations to regions of lower concentrations maintains an almost constant potassium concentrations in the glial cells . in sd
potassium in the ecs is strongly elevated during an about 80 sec lasting phase of fes and is continuously buffered during this time .
after 80 sec the concentration quickly reduces to slightly less than the normal physiological level . still there is a local potassium deficit and what we call backward buffering , i.e. , the release of potassium from the glial cells sets in .
it is much slower than the uptake , because it is driven by a far smaller deviation of the potassium concentration from physiological resting conditions of the glial cell .
so as for diffusion the forward and the backward process do not actually happen simultaneously .
similar to the above explanation of slow backward buffering in the glia scheme , an extremely slow backward time scale follows naturally in diffusive coupling . for diffusion
the potassium content is reduced at a time scale @xmath92 if extracellular potassium is greater than @xmath58 .
backward diffusion , however , only occurs in the final recovery phase that sets in after the neuron has returned from the transient fes state and is repolarized .
while @xmath93 is still far from the resting state level , @xmath39 is comparable to normal physiological conditions ( see the below bifurcation diagrams in figs .
[ fig:1]b and [ fig:2]b ) and hence the driving force @xmath94 during the final recovery phase is very small for a bath concentration close to the physiological resting state level .
consequently backward diffusion is much slower than forward diffusion .
note that this argument for different slow regulation time scales only relies on the values of the ecs potassium concentration along the physiological fixed point branch , and is not a feature of the particular regulation scheme we apply .
.[tab:1 ] parameters for ion based model [ cols="<,<,<",options="header " , ]
the results are presented in three parts that describe ( i ) the stability of closed models , where we treat the change @xmath40 of the potassium content as a bifurcation parameter , ( ii ) open models , i.e. , @xmath40 becomes a dynamical variable , with glial buffering and ( iii ) oscillations in ion concentrations in open models for bath coupling with the bath concentration @xmath95 as a bifurcation parameter . at first we will not treat the change @xmath40 of the potassium content as a dynamical variable , but as a parameter whose influence on the system s stability we investigate .
so the model we consider is defined by the rate eqs .
( [ eq:1])([eq:4 ] ) and the constraint eqs .
( [ eq:5 ] ) , ( [ eq:6 ] ) , ( [ eq:8])([eq:10 ] ) and ( [ eq:24 ] )
. its stability will be important for the full system with dynamical ion exchange between the neuron and a bath or glial reservoir to be discussed in the next two subsections .
the phenomenon of ionic excitability as in sd only occurs for dynamical @xmath40 .
we will see that a slow fast decomposition of ionic excitability is possible .
the fast ion dynamics is governed by the transmembrane dynamics that we discuss now and happens at the time scale @xmath96 .
the dynamics of @xmath40 is much slower ( @xmath88 and @xmath89 ) .
fast ion dynamics of the full system can hence be understood by assuming @xmath40 as a parameter that determines the level at which fast ( transmembrane ) ion dynamics occurs .
this implies a direct physiological relevance of the closed system bifurcation structure with respect to potassium content variation for transition thresholds in the full ( open ) system . as the bifurcation parameter ( purely transmembrane dynamics ) showing * ( a ) * the membrane potential of fixed points ( fp ) and limit cycles ( lc ) , and * ( b ) * potassium concentrations . the fixed point continuation yields the black curves .
solid sections are fully stable , dashed sections are unstable . the stability of the fixed point changes in hbs and lps .
the initial physiological condition is marked by a black square .
the limit cycle is represented by the extremal values of the dynamical variables during one oscillation .
the continuation yields the green lines with the same stability convention for solid and dashed sections .
the stability of the limit cycle changes either in a lp@xmath97 or in a period doubling bifurcation ( pd ) . in * ( b ) * the maximal and minimal extracellular potassium concentration of the limit cycle never differs by more than @xmath98 mm .
the values can hence not be distinguished on the scale of this figure and therefore only the maximal value is drawn .
the bifurcations are marked by full circles and labelled by the type , i.e. , hb , lp or @xmath99 , and a counter ( cf . also the insets with blow
ups , in particular the rightmost one showing @xmath100 and @xmath101 on a very small horizontal scale )
. the vertical and diagonal arrows labelled ` m ' and ` r ' indicate the direction of extracellular potassium changes due to ion fluxes across the membrane ( ` m ' ) and changes only due to @xmath40 , i.e. , because of ion exchange with a reservoir ( ` r ' ) . note that along the horizontal directions only the ics potassium concentration changes by a precise mixture of fluxes across the membrane and ion exchange with a reservoir .
[ fig:1 ] ] the bifurcation diagram of the reduced model is presented in fig .
[ fig:1 ] .
it is shown in the @xmath102plane ( fig .
[ fig:1]a ) and in the @xmath103plane ( fig .
[ fig:1]b ) to display membrane and ion dynamics , respectively . a pair of arrows pointing in the direction of extracellular potassium changes only due to fluxes across the membrane ( vertical ` m ' direction ) and only due to exchange with a reservoir ( diagonal ` r ' direction ) is added to fig . [ fig:1]b .
the fixed point continuation yields a branch ( black line ) where fully stable sections are solid and unstable sections are dashed .
stability changes occur in saddle node bifurcations ( also called limit point bifurcation , lp ) and hopf bifurcations ( hb ) . in a lp the stability changes in one direction ( zero eigenvalue bifurcation ) , in a hb
it changes in two directions and a limit cycle is created ( complex eigenvalue bifurcation ) . a limit cycle is usually represented by the maximal and minimal value of the dynamical variables . however , the oscillation amplitude of the ionic variables is almost zero for the limit cycles in our model .
maximal and minimal values can not be distinguished on the figure scale .
hence in the @xmath103plane the limit cycle continuation appears only as a single line ( green ) .
stability changes of limit cycles occur in saddle node bifurcations of limit cycles ( lp@xmath97 ) .
the limit cycles in our model disappear in homoclinic bifurcations . in this bifurcation
a limit cycle collides with a saddle .
when it reaches the saddle it becomes a homoclinic cycle of infinite period . as a reference point the initial physiological condition is marked by a black square .
we will call the entire stable fixed point branch that contains this point the physiological branch @xmath104 , because the conditions are comparable to the normal functioning physiological state in particular , action potential dynamics is possible when the system is on this branch .
let us discuss the bifurcation diagram starting from this reference point and follow the fixed point curve in the right direction , i.e. , for increasing @xmath40 .
the physiological fixed point loses its stability in the first ( supercritical ) hopf bifurcation ( hb1 ) at @xmath105 mm .
the extracellular potassium concentration is then at @xmath106 mm . in other word ,
much of the added potassium has been taken up by the cell .
the limit cycle associated with hb1 loses its stability in a period doubling bifurcation ( pd ) and remains unstable .
finally it disappears in a homoclinic bifurcation shortly after its creation ( cf .
right inset in fig .
[ fig:1]a ) .
the stable limit cycle emanating from the pd point becomes unstable in a @xmath107 and vanishes in a homoclinic bifurcation , too .
the parameter range of these bifurcations is extremely small ( @xmath108 ) .
such fine parameter scales will not play a role for the interpretation of ion dynamics .
ion concentrations are stationary and physiological up to @xmath109 , but for practical purposes it is irrelevant if we identify @xmath110 or @xmath109 as the end of the physiological branch @xmath104 .
the first hb is followed by four more bifurcations ( lp1 , hb2 , lp2 , hb3 ) that all neither restore the fixed point stability nor create any stable limit cycles .
the limit cycles for hb2 and hb3 are hence not plotted either .
it is only the fourth hopf bifurcation ( hb4 ) at @xmath111 mm in which the fixed point becomes stable again and in which a stable limit cycle is created .
the limit cycle branch loses its stability in lp1@xmath97 and regains it in lp2@xmath97 .
it becomes unstable again and even more unstable in lp3@xmath97 and lp4@xmath97 . shortly after that
( not resolved on the scales in fig .
[ fig:1 ] ) it ends in a homoclinic bifurcation with the saddle between hb1 and lp2 . at
hb4 the stable free energy
starved branch @xmath112 begins .
it is generally characterized by a strong increase in the ecs potassium compared to physiological resting conditions ( fig .
[ fig:1]b ) , and a significant membrane depolarization ( fig .
[ fig:1]a ) .
corresponding to the extracellular elevation intracellular potassium is significantly lowered .
this goes along with inverse changes of the compartmental sodium concentrations ( all not shown ) .
@xmath112 is hence characterized by largely reduced ion gradients and strong membrane depolarization .
in fact , at this membrane potential the sodium channels are inactivated which is normally called depolarization block in hh like membrane models without ion dynamics .
depolarization block is , however , only one feature of fes .
the closeness of fes to the thermodynamic equilibrium of the system is more importantly manifested in the reduced ion gradients . on @xmath112
no more bifurcations occur and it remains stable for increasing @xmath40 .
the interpretation of this bifurcation diagram should be as follows .
the end of @xmath104 defines the maximal potassium content compatible with a physiological state of a neuron . for larger @xmath40
it will be inevitably driven to the fes .
in other words the end of @xmath104 marks the threshold value for a slow , gradual elevation of the potassium content to cause the transition from physiological resting conditions to fes . in a buffered system
it is the threshold for sd ignition . on the other hand
stable fes like conditions require a minimal potassium content which marks the end of @xmath112 .
it is given by @xmath113 mm .
below this value the only stable fixed point is physiological .
again there is a narrow range , namely @xmath40 between @xmath114 and @xmath111 mm , in which stable oscillations can occur .
when glial buffering is at work the end of @xmath112 defines the threshold for potassium buffering , i.e. , for the potassium reduction that is required to return from fes to physiological conditions ( cf.eq .
( [ eq:25 ] ) ) . in the second subsection of sect .
results , we will see that this is exactly how ion regulation facilitates recovery in sd models .
there is another way the bifurcation diagram in fig .
[ fig:1]b can be read . as we have remarked above the limit cycles of the model
are characterized by large oscillation amplitudes in the membrane variables @xmath2 ( not shown ) and @xmath1 , but almost constant ionic variables @xmath6 , @xmath5 and @xmath7 ( only @xmath39 shown ) .
so fig . [ fig:1]b tells us which extracellular potassium concentrations can possibly be stable and which ones can not .
values below the end of @xmath104 at @xmath106 mm , values between @xmath115 mm and @xmath116 mm and finally concentrations in the range of @xmath112 starting at @xmath117 mm can be stable .
any other extracellular potassium concentration is unstable and the system will evolve towards a stable ion configuration that is present in the phase space .
the highest stable potassium concentration below fes values is @xmath118 .
if potassium in the ecs is increased instantaneously , this value indicates the threshold for sd ignition or the transition to fes .
panel * ( a ) * shows the membrane potential and panel * ( b ) * shows the extracellular potassium concentration of the invariant sets , i.e. , fixed points and limit cycles .
the line style convention ( solid for stable , dashed for unstable ) and bifurcation labels are the same as in fig .
[ fig:1 ] .
note the similar shape to fig .
[ fig:1 ] , but also the different scale of the two figures .
[ fig:2 ] ] performing the same type of bifurcation analysis with the physiologically more detailed model from kager et al.@xcite ( cf .
last paragraph of sect .
models ) leads to the diagram in fig .
[ fig:2 ] .
it has been shown before that also in this model there is stable fes@xcite .
we do not find the same bifurcations as in the reduced model , but only two lps and one hb .
however , the physiological implications are very similar . like in the reduced model
there is an upper limit of the potassium content @xmath40 for stable physiological conditions ( @xmath119 mm ) and a lower limit for stable fes ( @xmath120 mm ) . also the downward snaking and the stability changes of the limit cycle that starts from hb1 are very similar to fig . [ fig:1 ]
this leads to the same type of conclusion concerning possible stable extracellular potassium concentrations .
while numerical values of the stability limits in terms of @xmath40 are specific to each model , the topological similarity of the bifurcation diagrams suggests a generality of results : there is a stable physiological branch @xmath104 that ends at some maximal value @xmath40 of the potassium content . beyond this point
the neuron can not maintain physiological conditions , but will face fes .
on the other hand the stable fes branch @xmath112 ends for a sufficiently reduced potassium content the neuron will return to physiological conditions .
the new bifurcation diagrams presented in this section confirm our results from ref .
@xcite : neuron models whose ionic homeostasis is only provided by atpase driven pumps , but without diffusive coupling or glial buffering , will have a highly unphysiological fixed point that is characterized by free energy starvation and membrane depolarization .
however , the here presented bifurcation diagrams contain additional information of great importance .
using the new bifurcation parameter @xmath40 crucially extends our results from ref .
@xcite by uncovering the threshold concentrations in extracellular potassium concentration .
these are completely novel insights . in * ( a ) * , * ( b ) * the @xmath102plane and * ( c ) * , * ( d ) * the @xmath103plane .
the black curves are the stable fes branches that lose their stability in hopf bifurcations ( black circles ) .
starting from the leftmost fixed point curves the fixed @xmath121 values are 8 , 12 , 16 , 20 , 24 , 28 and 32 mm for the reduced model and 9 , 13 , 17 , 21 , 25 , 29 and 33 mm for the detailed model .
the hopf bifurcations for different chloride concentrations lead to the blue hopf line . as a reference the fixed point curves from figs .
[ fig:1 ] and [ fig:2 ] are also included to the diagram and drawn in grey .
[ fig:3 ] ] in the next subsection the bifurcation diagrams of the unbuffered ( closed ) systems shall facilitate a phase space understanding of the activation and inhibition process of ionic excitability as observed in sd in the buffered ( open ) systems .
we are aiming for an interpretation of ionic excitability where neuronal discharge and recovery are fast dynamics that are governed by the bistable structure discussed above , whereas additional ion regulation takes the role of slowly changing @xmath40 .
however , only the gated ion dynamics , i.e. , dynamics of sodium and potassium is fast compared to that of @xmath40 , chloride is similarly slow . by electroneutrality
this means that the overall concentration of positively charged ions in the ics , i.e. , the sum of sodium and potassium ion concentrations changes on the same slow time scale as the chloride concentration . to describe this slow process
not dynamically but
like @xmath40in terms of a parameter we simply investigate the stability for a given distribution of non dynamic , i.e. , impermeant chloride . to determine this stability we set the chloride current to zero and vary @xmath121 in a certain range ( from 8 to 32 mm for the reduced model , and from 9 to 33 mm for the detailed model ) .
this affects the system only through the electroneutrality constraint eq .
( [ eq:5 ] ) which sets the intracellular charge concentration to be shared by sodium and potassium . for each value of @xmath121
we perform a fixed point continuation as in figs .
[ fig:1 ] and [ fig:2 ] which yields similarly folded s shaped curves .
the result is shown in fig .
[ fig:3 ] . for our analysis of sd
it is only relevant where @xmath112 ends .
that is why the plot does not contain the whole fixed point curve , but only @xmath112 and a part of the unstable branch for a selection of @xmath121 values . as a reference
the diagrams also contain the fixed point curves from figs .
[ fig:1 ] and [ fig:2 ] which include chloride dynamics .
the fes branches in fig .
[ fig:3 ] end in hopf bifurcations .
the bifurcation points for different chloride concentrations yield the blue hopf line .
it marks the threshold for recovery from fes when dynamics of chloride and @xmath40 is slow . in the previous subsection
we have analyzed the phase space structure of ion based neuron models without contact to a reservoir , i.e. , without glial buffering or diffusive coupling .
these models have only transmembrane ion dynamics and obey mass conservation of each ion species .
hence they describe a closed system .
the bistability of a physiological state and fes that we found in these closed models is not experimentally observed , because real neurons are always open systems not merely in the sense that they consume energy a necessary prerequisite for being far from thermodynamic equilibrium
but they also can lose or gain ions through reservoirs or buffers .
we will now include glial buffering and show how it facilitates recovery from fes , a condition which in contrast to the physiological state is close to a thermodynamic equilibrium , namely the donnan equilibrium ( cf .
@xcite ) .
when glial buffering is at work , @xmath40 becomes a dynamical variable whose dynamics is given by the buffering rate eq .
( [ eq:30 ] ) . in previous subsection
we have explained that the bifurcation diagrams in figs .
[ fig:1 ] and [ fig:2 ] imply thresholds for an elevation of extracellular potassium to trigger the transition from physiological resting conditions to fes .
this is in agreement with computational and experimental sd studies in which high extracellular potassium concentrations are often used to trigger sd . another physiologically relevant way of sd ignition is the disturbance or temporary interruption of ion pump activity . as we have shown in ref .
@xcite there is a minimal pump rate required for normal physiological conditions in a neuron . below this rate
the neuron will go into a fes state and remain in that state even when the pump activity is back to normal .
mm after 20 sec ( vertical line ) . in * ( a ) * and * ( b ) * the time series of the membrane potentials ( black lines ) are shown .
nernst potentials for all ion species are included to the diagrams as a reference .
ion dynamics are shown in * ( c ) * and * ( d ) * where extracellular ion concentrations are in lighter color . [ fig:4 ] ] for the simulations in fig .
[ fig:4 ] we have interrupted the pump activity for about 10 sec in the reduced model , and we have elevated the extracellular potassium concentration by @xmath122 mm in the detailed model to trigger sd . both stimulation types work for both models , but only the two examples are shown .
the phase of pump interruption ( fig .
[ fig:4]a and [ fig:4]c ) is indicated by the shaded region in the plots , the time of potassium elevation is marked by the vertical grey line .
the dynamics of the two models is very similar : in response to the stimulation the neuron strongly depolarizes and remains in that depolarized state for about 70 sec ( fig .
[ fig:4]a and [ fig:4]b ) . after that the neurons repolarize
abruptly and asymptotically return to their initial state .
in addition to the membrane potential ( black curve ) the potential plots also contain the nernst potentials for sodium ( red line ) , potassium ( blue line ) and chloride ( green line ) that change with the ion concentrations according to the definition of the nernst potentials in eq .
( [ eq:22 ] ) . in fig .
[ fig:4]c and [ fig:4]d we see that the potential dynamics goes along with great changes in the ion concentrations .
in particular , extracellular potassium is strongly increased in the depolarized phase .
these conditions are very similar to the type of fes states discussed in the previous subsection .
the recovery of ion concentrations sets in with the abrupt repolarization , but it is a very slow asymptotic process that is not shown in fig .
[ fig:4 ] . in both models
the neuron is capable of producing spiking activity again right after the repolarization .
all these aspects of ion dynamics during sd are well known from several studies@xcite .
we remark that the time series are almost identical if glial buffering is replaced by the coupling to a potassium bath .
both , the strength of glial buffering and of diffusive coupling have been adjusted so that the depolarized phase lasts about 70 sec which is the experimentally determined time .
we will focus on bath coupling in last subsection of sec . results .
if neither buffering nor a potassium bath is included the neuron does not repolarize ( for time series plots of terminal transitions to fes see ref .
@xcite ) . .
as in fig .
[ fig:3 ] panels * ( a ) * and * ( b ) * contain plots of the membrane potentials , in panels * ( c ) * and * ( d ) * extracellular potassium is shown . *
( a ) * and * ( c ) * are for the reduced model , * ( b ) * and * ( d ) * for the detailed model .
the trajectories of the reduced model are represented as red curves , those of the detailed model are magenta .
the sections of the trajectories that belong to times before and during the stimulation are dashed .
the fixed point curves from fig .
[ fig:3 ] are added to the plots as shaded lines whereas the fixed point continuations for the unbuffered models with dynamical chloride are slightly darker .
the pair of arrows in the extracellular potassium plots indicates the direction of pure transmembrane ( vertical ) and pure buffering dynamics ( diagonal ) .
[ fig:5 ] ] the time series in fig .
[ fig:4 ] are useful to confirm that the neuron models we investigate have the desired phenomenology and indeed show sd like dynamics .
yet the nature of the different phases of this ionic excitation process
the fast depolarization , the prolonged fes phase and the abrupt repolarization
remains enigmatic@xcite . in a phase space plot
the picture becomes much clearer and the entire process can be directly related to the two stable branches , @xmath104 and @xmath112 , that we found for the closed and therefore pure transmembrane models in the previous subsection . in fig .
[ fig:5 ] the time series from fig . [ fig:4 ] for a simulation time of 50 min are shown in the @xmath102 and the @xmath103plane .
the parts of the trajectories during the stimulation ( pump interruption and potassium elevation ) are dashed . in the chosen planes vertical lines belong to dynamics of constant potassium contents that can be understood in terms of the models we analyzed in the previous subsection .
that is why fig .
[ fig:5 ] contains the fixed point curves from fig .
[ fig:3 ] as shaded lines as a guide to the eye . in fig .
[ fig:5]c and [ fig:5]d buffering dynamics is diagonal as indicated by the pair of arrows added to the plot . for both trajectories the stimulation
is followed by a vertical activation process that leads to the transition from @xmath104 to @xmath112 .
the verticality means that this is a process almost purely due to transmembrane dynamics .
it is governed by the bistable phase space structure that we discussed in the previous section and also in ref .
buffering dynamics is too slow to inhibit the activation .
the types of stimulation we applied are related to bifurcations of the transmembrane system : the potassium elevation is beyond the end of @xmath104 which is marked by the first hopf bifurcation ( hb1 ) in fig .
[ fig:1 ] .
the interruption of pump activity means that we go below a pump rate threshold that is defined by a saddle node bifurcation ( cf .
@xcite ) . more generally , to initiate an ionic excitation it is necessary to stimulate the system until it enters the basin of attraction derived in the unbuffered system of the fes state .
the activation is followed by a phase of both , slow transient transmembrane dynamics mostly due to chloride , and potassium buffering .
it is the latter that bends the trajectories in the diagonal direction so that they go along the fes branches from fig .
[ fig:3 ] .
the trajectories slowly approach the repolarization threshold given by the hopf line .
the duration of this fes phase is determined by how long it takes the system to reach the hopf line .
this process is a mixture of buffering and transient transmembrane dynamics for the reduced model and more buffering dominated in the detailed model .
the duration of the fes phase is consequently a result of both types of dynamics .
however , the main insight we gain from this plot is : glial buffering is the necessary inhibitory mechanism that takes the system to the hopf line so that it can repolarize .
we remark that the time series and phase space plots for bath coupling instead of buffering are almost identical and the same interpretation holds .
the more general conclusion is then : ion dynamics beyond transmembrane processes is necessary to take the system to the hopf line so that it can repolarize .
this can , of course , be a combination of bath coupling and buffering . when the hopf line is reached that neuron repolarizes abruptly which is the second almost purely vertical process .
the repolarization is followed by slow asymptotic recovery dynamics of ion concentrations that takes the neuron back to the initial state which is at @xmath123 mm .
the neuron regains the electrical excitability that is lost during fes already right after the repolarization .
so the system is back to physiological function long before the ion gradients are fully restored .
let us summarize the results from this subsection . by relating the sd time series from fig .
[ fig:4 ] to the bifurcation structure of the unbuffered models from the first subsection of sect .
results and in particular to the two stable branches @xmath104 and @xmath112 we have succeeded to understand ionic excitability as a sequence of different dynamical phases .
the initial depolarization and the later repolarization are membrane mediated fast processes that obey the bistable dynamics of unbuffered systems .
the fes phase is buffering dominated and lasts until buffering has taken the system to a well defined repolarization threshold .
the recovery phase is dominated by backward buffering .
the full excursion time is the sum of the durations of each phase . for the de and repolarization process
this duration mainly depends on the time scale of the transmembrane dynamics and is hence comparably short .
the duration of the fes phase is a result of both , the transient transmembrane dynamics and glial ion regulation at a much slower time scale .
the final recovery phase is mainly backward buffering dominated which is the slowest process .
hence the duration of an sd excursion is mainly determined by the slow buffering and backward buffering time scales .
this conclusion that relies on our novel understanding of the different thresholds involved in sd is in fact in agreement with recent experimental data suggesting vascular clearance of extracellular potassium as the central recovery mechanism in sd@xcite .
the dynamics of excitable systems can often be changed to self sustained oscillations by a suitable parameter variation .
the type of bifurcation that leads to the oscillations and the shape of the limit cycle in the oscillatory regime determine excitation properties like threshold sharpness and latency@xcite .
the oscillatory dynamics that is related to ionic excitability can be obtained for bath coupling with an elevated bath concentration @xmath95 .
so in this section we replace the buffering dynamics for @xmath40 with the diffusive coupling given by eq .
( [ eq:33 ] ) .
this coupling is used in experimental in vitro studies of sd@xcite and has also been applied in computational models that are very similar to our reduced one@xcite . . * ( a ) * and * ( b ) * , * ( c ) * and * ( d ) * , and * ( e ) * and * ( f ) * are simulations for @xmath124 , @xmath125 and @xmath126 , respectively .
the dynamics is typical for * ( a ) * and * ( b ) * seizure like activity , * ( c ) * and * ( d ) * tonic firing , * ( e ) * and * ( f ) * periodic sd .
note the different time scales of sla , tonic firing and period sd and also the different oscillation amplitudes in the ionic variables .
[ fig:6 ] ] depending on the level of the bath concentration , we find three qualitatively different types of oscillatory dynamics that are shown in fig .
[ fig:6 ] .
the top row ( a ) shows the time series of seizure like activity for @xmath124 .
it is characterized by repetitive bursting and low amplitude ion oscillations .
the other types of oscillatory dynamics are tonic firing at @xmath127 with almost constant ion concentrations ( fig .
[ fig:6]b ) and periodic sd at @xmath128 with large ionic amplitudes ( fig .
[ fig:6]c ) .
we see that sla and periodic sd exhibit slow oscillations of the ion concentrations and fast spiking activity , which hints at the toroidal nature of these dynamics . below we will relate sla and periodic sd to torus bifurcations of the tonic firing limit cycle .
color and line style conventions for fixed points and limit cycles are figs .
[ fig:1 ] and [ fig:2 ] : black and green lines are fixed point and limit cycles , solid and dashed line styles mean stable and unstable sections
. stable solution on invariant tori are blue .
they were obtained by direct simulations .
the fixed point changes stability in hbs and lps .
the bifurcation types limit cycle undergoes are @xmath107 , period doubling ( pd ) and torus bifurcation ( tr ) .
some physiologically irrelevant unstable limit cycles are omitted ( cf .
panel * ( a ) * shows the membrane potential , panel * ( b ) * shows the extracellular potassium concentration . *
( b ) * does not contain the limit cycle , because it can hardly be distinguished from the fixed point line.[fig:7 ] ] the examples in fig .
[ fig:6 ] show that our model contains a variety of physiologically distinct and clinically important dynamical regimes . a great richness of oscillatory dynamics , in fact
, under the simultaneous variation of @xmath129 and the glial buffering strength has already been reported in refs .
@xcite for a very similar model . in ref .
@xcite the authors even give a bifurcation analysis of ionic oscillations for @xmath129 elevation . to investigate dynamical changes and the transitions between the dynamical regimes in our model we perform a similar bifurcation analysis and vary @xmath95 , too .
two important differences should be noted though .
first , ref .
@xcite uses an approximation of the multi time scale model in which the fast spiking dynamics is averaged over time , while our analysis does not rely on such an approximation .
second , our analysis covers a bigger range of @xmath95 values which allows us to compare sla and sd , while ref .
@xcite exclusively deals with sla .
[ fig:7 ] shows the bifurcation diagram for @xmath95 variation in the @xmath130plane and in the @xmath131plane .
in addition to fixed points ( black ) and limit cycles ( green ) also quasiperiodic torus solutions ( blue ) are contained in the diagram . in comparison to fig .
[ fig:1 ] this model contains a new type of bifurcation , namely the neimark
sacker bifurcation , also called torus bifurcation ( tr ) .
a torus bifurcation is a secondary hopf bifurcation of the radius of a limit cycle in which an invariant torus is created .
if this torus is stable , nearby trajectories will be asymptotically bound to its surface .
however , we can not follow such solutions with standard continuation techniques , because these require an algebraic formulation in terms of the oscillation period .
this is not possible for torus solutions , because on a torus the motion is quasiperiodic , i.e. , characterized by two incommensurate frequencies .
we can hence only track the stable solutions by integrating the equations of motion and slowly varying @xmath95 .
it is due to this numerically expensive method that in this section we will only analyze oscillatory dynamics of the reduced hh model with time
dependent ion concentrations .
the result of this bifurcation analysis in fig.[fig:7 ] shows us that there is a maximal level @xmath132 of the bath concentration compatible with physiological conditions .
it is identified with the subcritical hopf bifurcation hb1 in which the fixed point loses its stability .
the related limit cycle is omitted , because it stays unstable and terminates in a homoclinic bifurcation with the unstable fixed point branch .
the fixed point undergoes further bifurcations ( lp1 , lp2 , hb2 , hb3 ) which all leave it unstable and do not create stable limit cycles .
it is in hb4 that the fixed point becomes stable again and also a stable limit cycle is created .
this is the last fixed point bifurcation of the model .
the limit cycle that is created in hb4 changes its stability in several bifurcations .
the physiologically most relevant ones are the four torus bifurcations .
the bifurcation labels indicate the order of detection for the continuation that starts at hb4 . initially the limit cycle is characterized by fast low amplitude oscillations .
it becomes unstable in the subcritical torus bifurcation tr1 .
it regains and again loses its stability in the subcritical torus bifurcations tr2 and tr3 .
the last torus bifurcation , the restabilizing supercritical tr4 , is directly followed by a pd after which no stable limit cycles exist any more .
again we have omitted in the diagram the unstable branch after pd and the limit cycle that is created in pd , which remains unstable .
physiologically it is more intuitive to discuss the diagram for increasing @xmath95 starting from the initial physiological conditions marked by the black square .
normal physiological conditions become unstable at @xmath132 and above this value the neuron spikes continuously according to the stable limit cycle branch between pd and tr4 . when @xmath133 is reached the dynamics changes from stationary spiking to seizure like activity on an invariant torus .
the beginning of sla is hence due to a supercritical torus bifurcation and the related ionic oscillation sets in with finite period and zero amplitude . from @xmath134 on tonic spiking activity is stable again and there is a small @xmath95range of bistability between sla and this tonic firing . as we mentioned above solutions on an invariant torus can not be followed with normal continuation tools like auto , so only stable branches are detected .
the details of the bifurcation scenario at tr3 are hence not totally clear , but we suspect that the unstable invariant torus that must exist near tr3 collides with the right end of the stable torus sla branch in a saddle node bifurcation of tori .
tonic spiking then remains stable until tr2 .
this bifurcation is related to the period sd that already exist well below @xmath135 .
in fact , the threshold value @xmath135 is in agreement with experiments@xcite . again
the unstable torus near tr2 is not detected , but we suppose that a similar scenario as in tr3 occurs .
the dynamics on the torus branch related to tr2 ( and tr1 where it seems to end ) is very different from the first torus branch . while the periods of the slow oscillations during sla are 1645 sec the ion oscillations of periodic sds are much slower with periods of 350550 sec .
another crucial difference is obvious from fig .
[ fig:7]b which shows the bifurcation diagram in the @xmath131plane .
the fixed point is just a straight line , because the diffusive coupling eq .
( [ eq:33 ] ) makes @xmath136 a necessary fixed point condition .
the limit cycle is always extremely close to this line . on the chosen scale
it can not be distinguished from the fixed point and is hence not contained in the plot . only the torus solutions of sd and sla
attain @xmath39 values that differ significantly from the regulation level .
the ionic amplitudes of sd are one order of magnitude larger than those of sla .
this has to do with the fact that the peak of sd as described above must be understood as a metastable fes state that exists due to the bistability of the transmembrane dynamics .
the dynamics of sla is clearly of a different nature .
note that the bifurcation diagram reveals a bistability of tonic firing and full blown sd between the left end of the sd branch at about 11 mm and tr2 .
this means that there is no gradual increase in the ionic amplitudes that slowly leads to sd , but instead it implies that sd is a manifest all or none process . .
panel * ( a ) * shows the extracellular sodium concentration and includes an inset around tr4 and pd .
panel * ( b ) * presents the potassium gain / loss.[fig:8 ] ] in fig .
[ fig:8 ] we look at the same bifurcation diagram in the @xmath137 and the @xmath138plane . while in fig .
[ fig:7 ] most of the ionic phase space structure is hidden , because @xmath139 for fixed points and limit cycles , the @xmath137presentation in fig .
[ fig:8]a provides further insights into the ion dynamics .
we see that the stable fixed point branch before hb1 has extracellular sodium concentrations close to the physiological value @xmath140 .
the stable branch after hb4 , however , has an extremely reduced extracellular sodium level and is indeed fes like .
the stable limit cycles between pd and tr4 and between tr3 and tr2 , and also sla are rather close to the physiological sodium level . on the other hand ,
periodic sd is an oscillation between fes and normal physiological conditions , which is an expected confirmation of the findings from the previous section .
[ fig:8]b is useful in connecting the phase space structure of the bath coupled system to that of the transmembrane model of the first subsection of sect
. results .
if we interchange the @xmath95 and the @xmath40axis in the diagram it looks very similar to fig .
[ fig:1]b .
the torus bifurcations tr1 , tr2 and tr3 are very close to the limit point bifurcations @xmath141 , @xmath142 and @xmath143 of the transmembrane model .
the fixed point curves are topologically identical .
this striking similarity has to do with the fact that the limit cycle in fig .
[ fig:1 ] has almost constant ion concentrations .
we have pointed out in the first subsection of sect .
results that fig .
[ fig:1 ] tells us which extracellular potassium concentrations are stable for pure transmembrane dynamics .
diffusive coupling with bath concentrations at such potassium levels leads to negligibly small values of @xmath144 ( cf .
( [ eq:33 ] ) ) .
therefore the limit cycle is still present in the bath coupled model and also the stability changes can be related to those in the transmembrane model . again
this can be seen as a confirmation of the results from the previous section : the transmembrane phase space plays a central role for models that are coupled to external reservoirs .
we can interpret the ionic oscillations from fig .
[ fig:6 ] and the bifurcations leading to them with respect to this phase space . .
only extracellular potassium is shown .
the limit cycle and fixed point curves from figs . [ fig:1 ] and [ fig:3 ] are superimposed to the plots as shaded lines whereas the limit cycle and fixed point from fig .
[ fig:1 ] ( dynamical chloride ) are darker .
the limit cycle and fixed point are not graphically distinguished , but comparison with fig .
[ fig:1 ] should avoid confusion .
[ fig:9 ] ] last we consider the dynamics of sla and periodic sd in a phase space projection . in fig .
[ fig:9 ] the trajectories for sla and periodic sd are plotted in the @xmath103plane together with the underlying fixed point and limit cycles from the transmembrane model ( cf .
[ fig:3 ] ) .
the periodic sd trajectory has a very similar shape to the single sd excursion from fig .
[ fig:5 ] and is clearly guided by the stable fixed point branches @xmath104 and @xmath112 . on the other hand sla is a qualitatively very different phenomenon . rather than relating to the fes branch
, it is an oscillation between physiological conditions and those stable limit cycles that exist for moderately elevated extracellular potassium concentrations .
the ion concentrations remain far from fes .
so sla and sd are not only related to distinct bifurcations , though of similar toroidal nature and branching from the same limit cycle , but they are also located far from each other in the phase space .
this completes our phase space analysis of local ion dynamics in open neuron systems
in this paper we have analyzed dynamics at different time scales in a hh model that includes time
dependent ion concentrations .
such models are also called second generation hodgkin huxley models .
they exhibit two types of excitability , electrical and ionic excitability , which are based on fast and slow dynamics .
the time scales of these types of excitability are themselves separated by four to five orders of magnitude .
the dynamics ranges from high frequency bursts of about 100 hz with short interburst periods of the order of 10 msec ( fig .
[ fig:6]a ) to the slow periodic sd with frequencies of about @xmath145 hz and periods of about 7:30 min ( fig .
[ fig:6]c ) .
the slow sd dynamics in our model is classified as ultra slow or near dc ( direct current ) activity and can not normally be observed by electroencephalography ( eeg ) recordings , because of artifacts due to the resistance of the dura ( thick outermost layer of the meninges that surrounds the brain ) .
however , recently subdural eeg recordings provided evidence that sds occur in abundance in people with structural brain damage@xcite .
indirect evidence was provided already earlier by functional magnetic resonance imaging ( fmri)@xcite and patient s symptom reports combined with fmri@xcite that sd also occurs in migraine with aura@xcite .
the slowest dynamics that can be accurately measured by eeg , i.e. , the delta band , with frequencies about 0.5 to 4 hz , has attracted modelling approaches much more than sd , which was doubted to occur in human brain until the first direct measurements were reported .
it is interesting to compare the origin of slow time scales in such delta band models to our slow dynamics .
models of the delta band essentially come in two types . on the one hand thalamo
cortical network and mean field models of hh neurons with fixed ion concentrations have been studied@xcite . in this case , a slow time scale emerges because the cells are interconnected via synaptic connections using metabotropic receptors that are slow , because they act through second messengers . on the other hand , single neuron models with currents that are not contained in hh , namely a hyperpolarization activated depolarizing current , @xmath146dependent sodium and potassium currents , and a persistent sodium current , were suggested .
the interplay between these currents gives rise to oscillations at a frequency of about 23 hz@xcite .
it is therefore hardly surprising that these currents , in particular the persistent sodium and the @xmath146dependent sodium and potassium currents , have also been proposed to play an essential role in sd@xcite .
furthermore , bursting as another example of slow modulating dynamics was studied in a pure conductance based model with a dendritic and an axo somatic compartment @xcite .
in contrast to those approaches our results show that already dynamics in a hh framework with time
dependent ion concentrations and buffer reservoirs range from seconds to hours even with the original set of voltage gated ion currents .
time scales from milliseconds ( membrane dynamics ) to seconds ( ion dynamics ) and even minutes to hours ( ion exchange with reservoirs ) can be directly computed from the model parameters ( cf.sect .
the interplay of membrane dynamics , ion dynamics and coupling to external reservoirs ( glia or vasculature ) naturally leads to dynamics typical of sla and sd . and @xmath112 ( see sec . results ) and the directions ( arrows ) of two paths of ` pure ' flux condition : fluxes exclusively across the membrane and fluxes exclusively from ( or to ) reservoirs .
a horizontal path is caused by a particular mixture of these fluxes that induces potassium ion concentration changes exclusively to the intracellular compartment .
ionic excitability can be understood as a cyclic process in this diagram ( see text ) .
[ fig : bifdiagram ] ] in particular sd is explained by a bistability of neuronal ion dynamics that occurs in the absence of external reservoirs . the potassium gain or loss @xmath40 through reservoirs provided by an extracellular bath ,
the vasculature or the glial cells is identified as a bifurcation parameter whose essential importance was not realized in earlier studies ( see fig .
[ fig : bifdiagram ] ) . using this bifurcation parameter and the extracellular potassium concentration as the order parameter
, we obtain a folded fixed point curve with the two outer stable branches corresponding to states with normal physiological function , hence named physiological branch @xmath104 , and to states being free energy starved ( @xmath112 ) .
the definition of the bifurcation parameter implies that exchange with ion reservoirs happens along the diagonal direction labelled by ` r ' .
mediated dynamics is in the vertical ` m ' direction . in the full system where the ion exchange is a dynamical variable our unconventional choice of variables , i.e. modelling @xmath40 instead of @xmath39 , makes it obvious that the time scales of diagonal and vertical dynamics is separated by at least two orders of magnitude .
slow dynamics is along @xmath104 and @xmath112 , and the fast dynamics describes the jumps between these branches .
we remark that dynamics along @xmath147 is slower than along @xmath148 , because the branch is almost horizontal which leads to a very small gradient driving the diffusive coupling .
similarly the release of buffered potassium from the glia cells is only weakly driven ( cf .
the discussion of buffering time scales in sect .
model ) . in the closed system sufficiently strong stimulations lead to the transition from the physiological resting state located on @xmath147 to fes . in the full system with dynamical ion exchange with the reservoirs ,
physiological conditions are restored after a large phase space excursion to the the before stable fes state .
we refer to this process as ionic excitability .
in contrast to the electrical excitability of the membrane potential this process involves large changes in the ion concentrations . the entire phase space excursion of this excitation process can be explained through the specific transits between and along @xmath104 and @xmath112 .
we observe ion changes on three slow time scales .
( i ) vertical transits between @xmath104 and @xmath112 caused by transmembrane dynamics in the order of seconds .
the time scale is determined by the volume
surface area ratio and the membrane permeability to the ions .
( ii ) diagonal dynamics along @xmath112 in the order of tens of seconds caused by contact to ion reservoirs .
this time scale is determined by buffer time constants or vascular coupling strength .
( iii ) dynamics on @xmath104 again caused by contact to ion reservoirs , but at the slower backward buffering time scale in the order of minutes to hours determined by the slower backward rate of the buffer @xcite . during this long refractory phase of ionic excitability the spiking dynamics based on electrical excitability separated by seven orders of magnitude seems fully functional .
the right end of @xmath147 and the left end of @xmath148 are marked by bifurcations that occur for an accordingly elevated or reduced potassium content .
this is the first explanation of thresholds for local sd dynamics in terms of bifurcations .
we remark , however , that for sd ignition the important question is not where @xmath147 ends , but instead where the basin of attraction of @xmath148 begins .
this new understanding of sd dynamics suggests a method to investigate the sd susceptibility of a given neuron model . one should consider the closed model without coupling to external reservoirs and check if shows the typical bistability between a physiological resting state and fes .
we remark that unphysical so called ` fixed leak ' currents must be replaced by proper leak currents with associated leaking ions .
thresholds for the transition between @xmath148 and @xmath147 translate to thresholds for sd ignition and repolarization , i.e. , recovery from fes in the full open model .
knowledge of the potassium reduction needed to reach the repolarization threshold and knowledge about the buffer capacity could then tell us if recovery from fes can be expected ( such as in migraine with aura ) or if the depolarization is terminal ( such as in stroke ) .
although our model does not contain all important processes involved in sd , our phase space explanation appears to be valid also for certain model extensions .
for example , considering only diffusive regulation of potassium is physically inconsistent , but adding an analoguous regulation term for sodium turns out not to alter the dynamics qualitatively .
moreover osmosis driven cell swelling normally regarded as a key indicator of sd is not included in our model , but can be added easily@xcite .
unpublished results confirm that also with such cell swelling dynamics the fundamental bifurcation structure of fig .
[ fig : bifdiagram ] is preserved . as a clinical application of our framework ,
we have linked a genetic defect , which affects the inactivation gate @xmath4 and which is present in a rare subtype of migraine with aura , to sd .
our simulations show that such mutations render neurons more vulnerable to sd@xcite .
the interesting point , however , is that on the level of the fast time scale the firing rate is decreased , which in a mean field approach ( as done for the delta band ) translates to decreased activity .
this effect seemingly contradicts the increased sd susceptibility and hence illustrates the pitfalls in trying to neglect ion dynamics in the brain and to bridge the gap in time scales by population models .
the authors are grateful for discussions with steven j. schiff and bas jan zandt .
nh thanks prof .
dr . eckehard schll for continuous support , fruitful discussions , and critically reading the manuscript .
36ifxundefined [ 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 * * , ( ) in @noop _ _ , ( , , ) pp . @noop * * , ( ) @noop * * , ( ) \doibase doi:10.1016/j.physd.2009.08.009 [ * * , ( ) ] link:\doibase 10.1186/2190 - 8567 - 3 - 7 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1007/s11538 - 011 - 9647 - 3 [ * * , ( ) ] @noop * * ( ) link:\doibase 10.1177/1073858408317955 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1007/s10867 - 010 - 9212 - 6 [ * * , ( ) ] \doibase doi:10.1371/journal.pone.0022127 [ * * , ( ) ] link:\doibase 10.1371/journal.pcbi.1003551 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1038/jcbfm.2009.285 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) \doibase doi:10.1371/journal.pone.0005007 [ * * , ( ) ] @noop * * ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/j.neuroimage.2010.02.076 [ * * , ( ) ] link:\doibase 10.7717/peerj.379 [ * * ( ) , 10.7717/peerj.379 ] | the classical hodgkin huxley ( hh ) model neglects the time dependence of ion concentrations in spiking dynamics .
the dynamics is therefore limited to a time scale of milliseconds , which is determined by the membrane capacitance multiplied by the resistance of the ion channels , and by the gating time constants .
we study slow dynamics in an extended hh framework that includes time
dependent ion concentrations , pumps , and buffers .
fluxes across the neuronal membrane change intra and extracellular ion concentrations , whereby the latter can also change through contact to reservoirs in the surroundings .
ion gain and loss of the system is identified as a bifurcation parameter whose essential importance was not realized in earlier studies .
our systematic study of the bifurcation structure and thus the phase space structure helps to understand activation and inhibition of a new excitability in ion homeostasis which emerges in such extended models .
also modulatory mechanisms that regulate the spiking rate can be explained by bifurcations .
the dynamics on three distinct slow times scales is determined by the cell volume to surface area ratio and the membrane permeability ( seconds ) , the buffer time constants ( tens of seconds ) , and the slower backward buffering ( minutes to hours ) .
the modulatory dynamics and the newly emerging excitable dynamics corresponds to pathological conditions observed in epileptiform burst activity , and spreading depression in migraine aura and stroke , respectively . |
the a15 superconductor v@xmath0si has been quite well known over the years both for its interesting normal state @xcite and superconducting properties @xcite and the correlation between the two states @xcite .
there is some renewed interest in v@xmath0si in recent years , first due to the observation of de haas - van alphen ( dhva ) effect @xcite , and lately due to the suggestion of the magnetic field induced phase transition in the flux - lattice structure @xcite .
the observation of dhva effect in a superconductor is quite puzzling to start with , since the superconducting energy gap is likely to eliminate quantum oscillations .
recent works on v@xmath0si @xcite and various other superconductors like nbse@xmath1 @xcite , nb@xmath0sn @xcite , ceru@xmath1 @xcite , uru@xmath1si@xmath1 @xcite have provided more interesting results , but both the experimental and theoretical situations are yet to be understood completely @xcite . in a very recent neutron measurement
it has been observed that in certain field direction , the hexagonal flux - line lattice ( fll ) of v@xmath0si distorts with the increase in magnetic field and abruptly becomes of square symmetry @xcite .
it is suggested that , this transition from the hexagonal to square symmetry may be a first order transition @xcite .
phase transition in flux - line lattice ( fll ) or vortex state in general , has been a subject of much interest in recent years both from theoretical @xcite and experimental @xcite points of view . in clean samples of type - ii superconductors with weak pinning properties , various topological phase transitions from a quasi - ordered fll ( or elastic solid or bragg - glass ) to a flux - line liquid , or from a quasi - ordered fll to a disordered fll ( or plastic solid or vortex - glass ) and then to a flux - line liquid were predicted theoretically ( see ref . 15 and references cited therein ) , and have subsequently been observed experimentally @xcite .
it is not clear at this moment whether there is an underlying correlation between the phase transitions associated with the change in the fll structure ( from hexagonal to square symmetry or vice versa ) and the field - induced change in topological character ( from ordered / quasi - ordered fll to disordered fll and/or flux - line liquid ) of the fll . for this purpose
it is important to identify various macroscopic as well as microscopic observable associated with the proposed phase transitions in the fll and study those in details .
peak - effect ( pe ) is an important observable for tracking the topological phase transitions ( from ordered fll to disordered fll ) in various high - t@xmath2 superconductors ( htsc ) @xcite .
pe is actually a generic term used to describe a peak or local maximum in the field variation of the critical current density ( j@xmath2(h ) ) in various type - ii superconductors @xcite . in dc magnetization study pe gives rise to a second peak in the field dependence of magnetization @xcite .
pe and its associated features have been used extensively in recent years to understand the exact nature of fll phase transitions in various classes of superconductors including both htsc @xcite as well as low t@xmath2 @xcite materials . in a recent theoretical study @xcite
it is suggested that pe in htsc materials may be explained by the softening of the fll due to an underlying structural phase transition from one fll symmetry to other .
this suggestion along with the experimental observation of the structural transition in the fll of v@xmath0si ( ref.7 ) have motivated us to study the superconducting mixed state properties of v@xmath0si in detail .
although there exist reports of pe in the dc magnetization @xcite and transport measurements @xcite of v@xmath0si , to our knowledge there exists no suggestion as yet of any topological phase transition ( from quasi - ordered elastic fll to disordered plastic fll ) associated with this pe .
the requirement of a detail study of pe in v@xmath0si has now become important in the light of various recent developments mentioned above . in this paper
we shall present results of our transport properties measurements in a good quality single crystal sample of v@xmath0si , focussing on pe and various interesting features associated with it .
our results will highlight the field - temperature history dependence of pe and associated metastable behaviour in v@xmath0si . based on our present results and other relevant experimental information from the existing literature , we shall also discuss the possibility of a phase transition in the fll of v@xmath0si .
the v@xmath0si single crystal used in our present study was prepared by dr .
a. menovski and it is cut from the same mother ingot , part of which was used earlier in de haas - van alphen study @xcite . while the residual resistivity ratio of the original sample ( from which the present sample is cut ) was reported to be 47 ( ref.6 ) , our measurements on the present sample yield a residual resistivity ratio of 42 .
the electrical transport measurements in our present study are performed using standard four probe technique .
we have used a superconducting magnet and cryostat system ( oxford instruments , uk ) to obtain the required temperature ( t ) and field ( h ) environment . in the configuration of our measurement the current ( i@xmath3 ) is passed along the @xmath4 direction of the sample and h is applied perpendicular to i@xmath3 .
the superconducting transition temperature ( t@xmath2 ) ( obtained from our zero field resistance measurement ) is 16.5k .
we have measured the magnetic field dependence of the resistance r(h ) within the following experimental protocols : 1 . cool the sample below t@xmath2 to various t of measurement in absence of any applied field h and then increase h isothermally above the upper critical field h@xmath5 .
this is zero - field cool ( zfc ) field - ascending mode .
2 . after the above step , decrease h isothermally from above h@xmath5 .
this is zfc field - descending mode .
field cool the sample in fixed h from a temperature well above t@xmath2 to various temperatures t ( @xmath6t@xmath2 ) of measurement .
this step is repeated for various h at each t. this is field - cool ( fc ) mode .
it is well known that for a sample of type - ii superconductor with pinning , the critical current i@xmath2 decreases monotonically with the increase in h and goes to zero at the irreversibility field ( h@xmath7h@xmath5 ) .
this is shown schematically in fig .
1 ( a ) . however , for superconductors showing pe , i@xmath2(h ) shows a peak or local maximum at an intermediate h value before finally going to zero at h@xmath8 ( see fig .
1(a ) ) . the signature of pe will appear in the field dependence of r(h ) , depending on the magnitude of the measuring current i@xmath3 . if i@xmath9i@xmath10 ( see fig .
1(a ) ) , r(h ) will show zero value up to ( depending on the exact value of i@xmath3 ) almost h@xmath11 ( see fig .
1(b ) ) . for h@xmath12 h@xmath13 , r(h ) will show flux - flow resistivity leading to the normal state behaviour for h@xmath14h@xmath5 . on the other hand if i@xmath15i@xmath16(h ) ( see fig .
1(a ) ) , flux - flow resistivity will start at lower field . in either of these cases , the r(h ) will not bear any signature of the pe .
the signature of pe will appear in r(h ) if i@xmath17 ( see fig .
1(a ) ) . in such a situation
a flux - flow resistance is observed with increasing h where i@xmath15i@xmath2(h ) but r(h ) will fall back to zero in the pe regime where i@xmath18i@xmath2(h ) before increasing again at higher field ( see fig .
1(b ) ) . while a direct measurement of the field dependence of i@xmath2 would have been very illuminating for the present study of pe in v@xmath0si , the lack of a suitable current source ( with i@xmath19 100 ma ) constrained us to the study of r(h ) only . adjusting the measuring current i@xmath3 accordingly , we present in fig .
2 r vs h plots for v@xmath0si at various t showing a distinct signature of pe .
the ( h , t ) regime where pe is observed , roughly agrees with that obtained earlier in magnetic measurements @xcite .
the finer quantitative discrepancy can be attributed to the different residual resistivity ratio of the samples used in the earlier measurement @xcite which leads to a small but perceptible change in h@xmath5(t ) .
( it should be noted here that the normal state resistance of the present v@xmath0si sample above h@xmath5(t ) is @xmath20 40 micro ohm as shown in figs . 2 - 5 . )
the distinct signature in pe disappears in the h dependence of r(h ) for t@xmath1415k .
the r vs h plots do not show the _ reentrant _ zero - resistance behaviour for any pre - decided value of i@xmath3 between 10ma and 100ma at t=15.5k , although a subtle minimum in r(h ) is observed for i@xmath3=20ma ( see fig.3).it should be noted here that the existence of pe in v@xmath0si was not at all clear in this t regime in the magnetic measurements as well @xcite .
the pe and the associated features have been very well studied in recent years in the c15-laves phase superconductor ceru@xmath1 @xcite . a very distinct history dependence of pe while cycling h or t was observed in ceru@xmath1 and it was suggested that at the onset of the pe regime , the field cooled ( fc ) flux - line lattice or the superconducting vortex state of ceru@xmath1 was more disordered than the corresponding zfc state @xcite .
we shall now investigate the possibility of the same in the present sample of v@xmath0si .
for this , we measure r vs h at various t following the fc protocol ( described above in the experimental section ) , and the results of such a study are shown in fig.4 .
the r(h ) measured with the fc protocol is found to be zero even for h - values at which measuring current i@xmath3 gave rise to the intermediate h flux - flow regime in the measurements following the zfc protocol ( see fig.2 ) .
this clearly indicates that in the fc mode i@xmath2(h)is greater than i@xmath3 in this intermediate h regime , while the opposite is true in the zfc mode .
thus i@xmath21 is higher in the fc mode than in the zfc mode and this is represented schematically in fig.1c . to the best of our knowledge
this history dependent property of the flux - line lattice of v@xmath0si has not been reported in the literature so far .
the field cooled fll at the onset of the pe regime in ceru@xmath1 was reported to be quite metastable in nature @xcite .
we shall now focus on the fc fll of v@xmath0si and check for the metastable behaviour in various h regime .
we subject the sample to small field cycling subsequent to the initial fc measurement with an applied h at a particular t. the results of such experiments are shown in fig.5 .
we find that within the intermediate h regime , the zero resistance state obtained in the fc mode is destroyed readily on field cycling and the flux - flow resistance corresponding to zfc state is recovered .
this clearly shows the metastability of fc state in the intermediate h regime just below the pe regime .
we also note that the minimum value of the cycling field ( @xmath22 ) to destroy the metastable fc state , decreases as we move away from the pe regime towards lower field . while we required @xmath23h=25 mt at h=3.2 t to reach the flux - flow resistance of the corresponding zfc state ( see fig .
5 ) , the recorded value of @xmath22 at h=2.4 t is 5 mt . below h=1.8 t where both zfc and fc r(h ) show zero resistance ,
the fc state is quite stable and not sensitive to any external field cycling .
inside the pe regime also the field cycling does not have any effect on the fc state .
so , it is clear that in the intermediate h regime ( where the fc fll is inferred to be more disordered than the zfc fll ) , the fc state is metastable in nature .
the main results of our present study are the following : 1 .
we see distinct signature of pe in the electrical resistance measurements of v@xmath0si .
2 . there is a clear path dependence in the electrical resistance in the field regime at the onset of pe .
this path dependence suggests that the fll prepared following the fc protocol is relatively more disordered , and can carry more critical current than the fll prepared in the zfc protocol .
the fc fll in the intermediate h regime of interest is quite metastable in nature and is sensitive to external fluctuations in the form of field cycling .
a related history dependence of the transport properties in polycrystalline samples of ceru@xmath1 , has earlier been reported by dilley et al @xcite .
these results of transport properties measurements are correlated to the history effects observed in the pe regime in various magnetic measurements of ceru@xmath1 @xcite .
we shall now make a comparative study of the history dependence of pe in v@xmath0si with the history effects in ceru@xmath1 and nbse@xmath1 where it was argued @xcite that the onset of pe marked a field induced first order transition from a relatively ordered fll to a disordered fll .
while the origin of this transition as well as the microscopic nature of the high field - high temperature phase remain a matter of debate ( see ref.20 ) , it is generally agreed that on reduction of h and t the system supercools across the first - order transition line @xcite .
experimentally observed field - temperature history effects were associated with this supercooling effect ( see ref .
this picture of first order transition is further strengthened with the recent theoretical argument ( ref.30 ) that the range of supercooling while reducing t ( i.e. in the fc mode ) is more than that obtained through the isothermal variation of h. this is observed experimentally both in ceru@xmath1 ( ref .
27 ) and nbse@xmath1 ( ref .
21 ) . on comparison ,
the pe and the associated history dependence in v@xmath0si turn out to be very similar to ceru@xmath1 and nbse@xmath1 .
however , it should be noted that the history dependence of the resistance on isothermal reduction of h ( from above h@xmath5 ) is relatively subtle in v@xmath0si ( see inset of fig.2 ) , and detailed magnetic measurements in the line of those in ceru@xmath1 @xcite are required here .
( this will be one of our future projects ) .
a similar field - temperature history dependence of pe in the transport properties measurements has also been observed in single crystal nb @xcite and thick film samples of nb@xmath0ge and mo@xmath0si @xcite . in these systems
the history dependence was associated with the pinning properties of the fll dislocations @xcite . however , while comparing the history effects observed in ceru@xmath1 , we argued ( ref .
20 and 27 ) that the rather fragile nature of the history dependent fll s can not be explained easily within such a picture .
it would imply that small field excursions anneal out fll dislocations or any other source of enhanced pinning . and this contradicts the conjecture ( ref .
32 ) that annealing occurs only when the field is reduced below the peak - effect regime . the fragile nature of fc fll is clearly shown in our present study on v@xmath0si .
we have also shown that the field cycling @xmath23h required to destabilize the concerned fc fll decreases rapidly as one goes away from the pe regime .
this is consistent with the theoretical picture that the energy barrier between the higher energy metastable state and the lower energy stable state in the supercooled regime diminishes rapidly on variation of ( t , h ) ; supercooling ceases to exist after ( t@xmath24 ) , where barrier height goes to zero and the metastable state is destroyed with infinitesimal fluctuation @xcite .
we believe that the idea of supercooling associated with a first - order transition probably has an edge over the depinning or annealing of fll dislocations in explaining the metastable behaviour of fc fll .
possibility of some kind of a field induced glass - like transition should also be considered here , especially when it is known that the low diffusivity of a glass can cause metastablities .
such metastabilities , however , are associated with hindered kinetics and not with local minima in free energy .
if the metastability arises due to reduced diffusivity , then naive arguments suggest that the metastability will be more persistent when larger motions of particles in configuration space are involved ; larger motions are involved when density is varied , rather than when temperature is varied . in the case of fll
a much larger rearrangement of fll is involved when an ( h , t ) point is reached by varying h isothermally than when the same ( h , t ) point is reached by varying t at constant h. hysteresis would thus be lower in the fc case than in the case of isothermal h variation .
this is in contrast of what we have actually observed in the fll of v@xmath0si in the vicinity of pe regime and thus negate the possibility of a glass - like transition to be associated with the pe in v@xmath0si .
we would also like to introduce a third possibility , namely an analogy ( if not exact similarity ) with the random - field ising systems ( rfim ) where similar field - temperature history effects are well known @xcite .
most experimental information in rfim systems has been obtained from studies on various diluted antiferromagnets . in the zero field
cooled mode , the diluted antiferromagnet is cooled through the zero - field neel transition temperature .
the resultant antiferromagnetic order thus formed , is preserved when an external magnetic field is subsequently switched on at low temperatures .
the long range order , however , gets reduced and ultimately goes to zero on heating the system to the high temperature paramagnetic phase . on cooling back now from the paramagnetic phase in presence of the applied field ( i.e. in the field cooled mode ) , the sample develops a short range ordered domain state @xcite in contrast to the long range order zfc state . the similarity with the fll state around the pe regime of v@xmath0si is apparent here , namely the fc fll is relatively more disordered . like
the fc state of rfims , the fc fll of v@xmath0si in the vicinity of pe , also becomes unstable on field and temperature cycling @xcite ; the fc states tend to develop long range order through such cycling process @xcite .
it is also interesting to note here that the question regarding the underlying phase transition in rfims whether it is a first order phase transition or a continuous second order phase transition is yet to be settled @xcite .
we have mentioned in the beginning that a very recent neutron study has shown the existence of a field induced structural transition from hexagonal to square symmetry in the fll of v@xmath0si ( ref.7 ) .
such a structural transition is capable of softening the fll and can give rise to a pe @xcite .
moreover , if the structural transition in the fll of v@xmath0si is indeed a first order transition as was speculated in the aforementioned neutron study @xcite , then all the experimentally observed history dependence of pe in v@xmath0si may find a natural explanation in terms of this first order phase transition .
however , we must note that the reported structural transition in fll takes place at a rather low field value of about 1.3 t at 1.8k .
unless there is a strong nonmonotonic temperature dependence of the onset field of pe in magnetic and transport measurements as is observed in the case of untwinned single crystals of ybco ( ref.38 ) , any simple correlation between pe and structural transition will not be valid .
future measurements of magnetic and transport properties down to 1.8k and/or neutrom measurements in the temperature regime 12k and above will settle this question .
softening of the fll in a type - ii superconductor may also arise due to a change in the character of the underlying superconducting order parameter .
it was argued by fulde and ferrel @xcite , and independently by larkin and ovchinnikov @xcite that in a superconductor with substantial normal state paramagnetism , a partially depaired superconducting state was energetically more favorable than the isotropic bcs state .
the superconducting order parameter in this high field state ( which is often termed as fulde - ferrel - larkin - ovchinnikov ( fflo ) state ) vanishes periodically ( with a period of 10 - 40 times of the coherence length @xmath25 ) along the direction of the applied field h @xcite .
these nodes in the order parameter can transform the three dimensional rigid flux - lines at lower h into a quasi - two dimensional structure at the onset of the fflo state@xcite at higher fields .
a fll with such a quasi - two dimensional structure is relatively more soft and can give rise to a pe .
moreover , since the transition to the fflo state is a first order transition , any system undergoing a fflo transition should show along with the pe , the history effects associated with supercooling . although in mid nineties ceru@xmath1 and the heavy fermion superconductor upd@xmath1al@xmath0 were thought to be prime candidates for the fflo state @xcite , there remain several arguments against the existence of such a state in those systems ( see ref .
19 and 44 ) .
it should be quite instructive now to check whether v@xmath0si meets the following necessary conditions for the existence of the fflo state : 1 .
the first condition that such a system should be in the clean limit is easily met for the present single crystal sample of v@xmath0si ( ref.6 ) as well as in the other samples of v@xmath0si where pe was observed earlier @xcite .
the second question , whether the upper critical field h@xmath5 of v@xmath0si is clearly pauli limited or not , is difficult to answer .
the h@xmath26 18.5 t @xcite is lower than the pauli limiting field h@xmath27=1.84
@xmath28 t@xmath2 tesla / k @xmath2030 t .
however , the effect of pauli paramagnetic limiting process is thought to be quite important for the superconducting properties of v@xmath0si @xcite .
it should be noted here the expression for h@xmath27 was derived for a system with spherical fermi surface and its application to a system with apparently complicated fermi surface should be made with some caution .
it is now known that the nesting properties of the fermi surface actually helps to stabilize the fflo state in a system @xcite .
also electron correlation effect is known to influence the value of h@xmath27 in v@xmath0si @xcite .
while in earlier theoretical works @xcite the fflo state ceased to exist for [email protected]@xmath2 , the extent of fflo state in various systems is found to be up to @xmath200.9t@xmath2 ( ref .
while pe was not observed in v@xmath0si for t@xmath1415k in magnetic measurements @xcite , in our present transport study the field induced zero resistance state associated with the pe is observed up to 15k only .
however , we have observed a local minimum ( but not the re - entrant zero resistance state ) in the r vs h plot for t@xmath1415k .
this resistance minimum , unlike the re - entrant zero resistance in the regime t@xmath3015k , is path independent .
the above discussion should by no means be taken as a support for the existence of the fflo state in v@xmath0si , but emphasizes that more studies , especially measurements probing the microscopic properties of the superconducting states , are required to reach a definite conclusion .
we must mention that a relatively simpler argument for pe is also possible in v@xmath0si in terms of minute microscopic inhomogeneity present even in the single crystals of v@xmath0si @xcite .
regions with slightly different superconducting properties can act as additional pinning centers at high fields , in a similar manner as the oxygen deficient centers in the htsc material ybco @xcite , and give rise to pe .
however , the magnetic history dependence of pe can not have any simple explanation within such a picture . in conclusion ,
our study of the peak effect in v@xmath0si has revealed interesting history effects and metastable features associated with it .
these results can be explained in terms of a first order phase transition in the flux - line lattice of v@xmath0si .
it is interesting to note here that very recently history effects and associated metastable features have been observed in the crossover regime from bragg - glass to vortex - glass in single crystal samples of ybco @xcite .
very recent magneto - optic studies in single crystal samples of bscco claim the presence of phase co - existance @xcite and supercooling across the bragg - glass to vortex - glass phase transition @xcite .
these , in turn , suggest the possibility of a first order transition .
josephson plasma resonance study in single crystal samples of bscco also support this picture @xcite . in this context ,
our present results on the well known a15 superconductor v@xmath0si are likely to provide useful information on the universality of the problem .
l. r. testardi , rev .
47 * , 637 ( 1975 ) .
p. w. anderson , a career in theoretical physics ( world scientific , 1994 ) p464 .
s. v. vonsovsky , yu . a. izyumov and e. z. kurmaev , superconductivity of transition metals ( springer verlag , 1982 ) .
g. bilbro and w. l. mcmillan , phys . rev . *
b14 * , 1887 ( 1976 ) .
f. m. mueller , d. h. lowndes , y. y. chang , a. j. arko and r. s. list , phys .
lett . * 68 * , 3928 ( 1992 ) .
r. corcoran , n. harrison , s. m. hayden , p. meeson , m. springford and p. j. van der wel , phys .
lett . * 72 * , 701 ( 1994 ) .
m. yethiraj , d. k. christen , d. mck .
paul , p. miranovic and j. r. thompson , phys .
lett . * 82 * , 5112 ( 1999 ) . t. j.
b. m. janssen , c. haworth , s. m. hayden , p. meeson , m. springford , a. wasserman , phys . rev . * b57 * 11698 ( 1998 ) .
r. corcoran , p. meeson , y. onuki , p. a. probst , m. springford , k. takita , h. harima , g. y. guo and b. l. gyorffy , j. phys.:condens matter * 6 * 4479 ( 1994 ) .
n. harrison , s. m. hayden , p. meeson , m. springford , p. j. van der wel and a. menovsky , phys .
rev . * b50 * , 4208 ( 1994 ) .
m. hedo , y. inada , t. ishida , e. yamamoto , y. haga , y. onuki , m. higuchi and a. hasegawa , j. phys .
. jpn . * 64 * , 4535 ( 1995 ) .
h. ohkuni , t. ishida , y. inada , y. haga , e. yamamoto , y. onuki and s. takahashi , j. phys .
* 66 * 945 ( 1997 ) .
g. blatter , m. v. feigekman , v. b. geshkenbein , a. i. larkin , and v. m. vinokur , rev .
* 66 * 1125 ( 1994 ) e. zeldov , d. majer , m. konczykowski , v. b. geshkenbein , v. m. vinokur and h. shtrikman , nature * 375 * 373 ( 1995 ) t. giamarchi and p. le doussal , phys
. rev . * b52 * , 1242 ( 1995 ) b. khaykovich , e. zeldov , d. majer , t. w. li , p. h. kes , and m. konczykowski , phys
lett . * 76 * , 2555 ( 1996 ) ; k. deligiannis , p. a. j. de groot , m. oussena , s. pinfold , r. langan , r. gagnon and l. taillefer , phys .
lett . * 79 * 2121 ( 1997 ) ; d. giller , a. shaulov , r. prozorov , y. abulafia , y. wolfus , l. burlachkov , y. yeshurun , e. zeldov , v. m. vinokur , j. l. peng and r. l. greene , phys . rev . lett . * 79 * 2542 ( 1997 ) .
a. m. campbell and j. e. evetts , adv .
* 21 * 327 ( 1972 )
. s. b. roy and p. chaddah , physica * c279 * 70 ( 1997 ) .
s. b. roy and p. chaddah , j. phys.:condensed matter * 9 * ( 1997 ) l625 .
s. b. roy , p. chaddah and s. chaudhary , j. phys.:condensed matter , * 10 * 4885 ( 1998 ) .
g. ravikumar , p. k. mishra , v. c. sahni , s. s. banerjee , s. ramakrishnan , a. k. grover , p. l. gammel , d. j. bishop , e. bucher , m. j. higgins and s. bhattacharya , physica c * 322 * 145 1999 . b. rosenstein and a. knigavko , phys .
lett . * 83 * , 844 ( 1999 ) m. isino , t. kobayashi , n. toyota , t. fukase and y. muto , phys . rev . *
b38 * 4457 ( 1988 ) .
m.pulver , z. physik * 257 * 22 ( 1972 ) .
r. meier - hirmer , h. kupfer and h. scheurer , phys . rev . *
b31 * 183 ( 1985 ) .
g. ravikumar , v. c. sahni , p. k. mishra , t. v. c. rao , s. s. banerjee , a. k. grover , s. ramakrishnan , s. bhattacharya , m. j. higgins , e. yamamoto , y. haga , m. hedo , y. inada and y. onuki , phys . rev . * b57 * r11069 ( 1998 )
. s. b. roy , p. chaddah and s. chaudhary , j. phys.:condensed matter , * 10 * 8327 ( 1998 ) .
n. r. dilley , j. hermann , s. h. han and m. b. maple , phys . rev . *
b56 * 2379 ( 1997 ) .
p. chaddah and s. b. roy , bull .
* 22 * 275 ( 1999 ) .
p. chaddah and s. b. roy , phys
. rev . * b60 * 11926 ( 1999 )
. m. steingart , a. g. putz and e. j. kramer , j. appl . phys . * 44 * 5580 ( 1973 ) .
r. wordenweber , p. h. kes and c. c. tsuei , phys . rev . *
b33 * 3172 ( 1986 ) .
d. p. belanger , phase transitions * 11 * 53 ( 1988 ) .
j. a. mydosh , spin - glasses ( taylor and francis , 1993 ) . r. j. birgeneau , j. magn .
magn . mater . * 177 * 1 ( 1998 ) .
s. chaudhary , a. k. rajarajan , k. j. singh , s. b. roy and p. chaddah , solid st .
( in press ) .
j. p. hill , q. feng , q. j. harris , r. j. birgeneau , a. p. ramirez and a. cassanho , phys .
* b55 * 356 ( 1997 ) ; q. feng , q. j.harris , r. j. birgeneau and j. p. hil , phys . rev . * 55 * 370 ( 1997 ) .
d. giller , a. shaulov , y. yeshurun , j. giapintzakis , phys . rev . *
b60 * 106 ( 1999 ) .
p. fulde and r. a. ferrel , phys . rev . * 135 * a550 ( 1964 ) .
a. i. larkin and y. n. ovchinnikov , sov .
jetp * 20 * 7 62 ( 1965 )
. m. tachiki , s. takahashi , p. gegenwart , m. weiden , c. geibel , f. steglich , r. modler , c. paulsen and y. onuki , z. phys . * b100 * 369 ( 1996 ) .
r. modler , p. gegenwart , m. lang , m. deppe , m. weiden , t. luhmann , c. geibel , f. steglich , c. paulsen , j. l. tholence , n. sato , t. komatsubara , y. onuki , m. tachiki , and s. takahashi , phys .
lett . * 76 * 1292 ( 1996 ) .
f. steglich , r. modler , p. gegenwart , m. deppe , m. weiden , m. lang , c. geibel , t. luhmann , c. paulsen , j. l. tholence , y. onuki , m. tachiki and s. takahashi , physica * c263 * 498 ( 1996 ) . n. dilley and m. b. maple , physica * c278 * 207 ( 1997 ) . t. p. orlando , e. j. mc niff jr . ,
s. foner and m. r. beasley , phys .
* b19 * 4545 ( 1979 ) h. simahara , phys . rev .
* b50 * 12760 ( 1994 ) .
orlando and m. r. beasley , phys .
lett . * 46 * 1598 ( 1981 ) l. w. gruenberg and l. gunther , phys .
* 16 * 996 ( 1966 ) p. chaddah and r. o. simmons , phys
b27 * 119 ( 1983 ) .
m. daumling , j. m. seuntjens and d. c.larbelstier , nature , * 346 * 332 ( 1990 ) .
s. kokkaliaris , p. a. j. de groot , s. n. gordeev , a. a. zhukov , r. gagnon , l. taillefer , phys .
lett . * 82 * 5116 ( 1999 ) ; s. kokaliaris , a. a. zhukov , p. a. j. de groot , r. gagnon , l. taillefer and t. wolf , phys . rev . *
b61 * 3655 ( 2000 ) ; s. o. valenzuela and v. bekeris , phys . rev
. lett , * 84 * 4200 ( 2000 ) ; y. redzyner , s. b. roy , d. giller , y. wolfus , a. shaulov , p. chaddah and y. yeshurun , phys . rev . * b * to appear in 1 june 2000 issue .
d. giller , a. shaulov , t. tamegai and y. yeshurun , phys .
lett . * 84 * 3698 ( 2000 ) . c. j. van der beek , s. colson , m. v. indenbohm and m. konczykowski , phys .
lett . * 84 * 4196 ( 2000 ) .
m. b. gaifullin , y. matsuda , n. chikumoto , j. shimoyama and k. kishio , phys .
* 84 * 2945 ( 2000 ) . | we present results of transport properties measurement on a single crystal of v@xmath0si showing distinct signature of peak - effect in its vortex state .
the field variation of the electrical resistance in the flux - line lattice prepared by different experimental path , namely zero field cooling ( zfc ) and field cooling ( fc ) , shows a distinct path dependence in the vicinity of the peak - effect regime . in the field
cooled state , small cycling of magnetic field modifies the resistance drastically around the peak - effect regime , highlighting the metastable nature of that state in the concerned regime . |
a galaxy is a self - gravitating system where stellar dynamics is governed by newton s law .
it could be naively described as a set of @xmath2 coupled , second - order , non - linear ordinary differential equations , where @xmath3 is the number of stars , which ranges between @xmath4 and @xmath5 @xcite .
solving such an equation set numerically is practically only possible at the very low end of the @xmath3-range , and even so very challenging with current computer hardware .
thus , various techniques are used to simplify the mathematical description of the system ; these are often designed to fit a particular problem in stellar dynamics and yield unphysical results when applied to another problem .
direct @xmath6-body simulation is one of the main techniques used to study gravitational systems in general and galaxies in particular . in this technique ,
the distribution function is sampled at @xmath7 points in a monte - carlo fashion .
this @xmath6 depends on the computational capabilities , and an astrophysical system with @xmath8 stars might be represented numerically by a sample of just @xmath4 `` supermassive '' particles .
this seems to be allowed because of the equivalence principle and the fact that gravitation is scale free , unlike , for example , in molecular dynamics .
however in gravity too this simplification can cause problems , as some dynamical effects depend on number density rather than just mass density .
the most well known @xmath6-dependent effect in stellar dynamics is two - body relaxation .
the relaxation time , the characteristic time for a particle s velocity to change by order of itself due to encounters with other particles , scales with the crossing time roughly as @xmath9 .
thus , the ratio between the relaxation times in a real and a simulated system is of similar order of magnitude to the undersampling factor .
this could be taken into account when interpreting the result of an undersampled simulation , but a poorly sampled distribution function might have other , unexpected , consequences .
galaxies are often described as collisionless stellar systems , which means that the relaxation time is much larger than the timescale of interest ( except perhaps at the very center ) .
this property could be very useful : since a particle s orbit is basically what it would be if it were moving in a smooth gravitational field , we could evaluate the field instead of calculating all stellar interactions , this is cheaper computationally .
another useful property is that galaxies are often spheroidal in shape .
even highly flattened galaxies will have a spherical dark halo component .
thus , a spherical shape could be used as a zeroth order approximation for the gravitational field , and higher order terms could be written using spherical harmonics .
the goal of this paper is to examine two techniques that utilize both these facts .
these are the multipole expansion ( mex ) and the self - consistent field ( scf ) methods .
they historically come from different ideas , and as explained below in detail , they are mathematically distinct . in the context of numerical simulations , however , they serve a similar function : to evaluate the gravitational force on all @xmath6 particles generated by this same collection of particles , in a way that discards spurious small scale structure ( in other words , smooths the field ) .
mex was born of the need to ease the computational burden .
the idea is that given spherical symmetry , gauss s law says that the gravitational force on a particle at radius @xmath10 from the center is simply @xmath11 , towards the center , where @xmath12 is the enclosed ( internal ) mass .
the gravitational constant , @xmath13 , will be omitted in the following text .
this idea was used by @xcite who simulated clusters with up to 100 particles to study phase mixing due to spherical collapse .
this `` spherical shells '' methods is mex of order zero and was also used for the same purpose by @xcite .
the extension of this this idea is that when spherical symmetry breaks , corrections to the force can be expressed by summing higher multiples ( dipole , quadruple , etc . ) of the other particles , both internal and external to @xmath10 .
@xcite used such a code to study a stellar cluster of a 1000 stars embedded in a galactic potential , truncating the expansion at @xmath14 .
@xcite used a variation of this method to study galaxy collision .
these authors employed a grid and conducted simulations of also up to @xmath15 and @xmath14 .
they additionally assumed azimuthal symmetry which reduced the number of terms in the expansion .
@xcite , @xcite , @xcite and @xcite all use variations of this method , with additional features which are partly discussed in secion [ sec : discussion - mex ] .
see also @xcite for a review .
the prehistory of scf is rooted in the problem of estimating a disk galaxy s mass distribution from its rotation curve .
@xcite proposed a mathematical method to generate a surface density profile and a corresponding rotation curve ( related to the potential ) by means of a hankel transform , and introduced a family of such pairs .
@xcite used toomre s idea , but in reverse : to calculate the gravitational field from an arbitrary 2d density , he generated an orthogonal set of density profiles and their corresponding potentials .
this solved two problems ( 1 ) with his orthogonal set it was possible to represent any flat galaxy as a finite linear combination of basis functions , and ( 2 ) unwanted collisional relaxation was curbed due to the smooth nature of the reconstructed gravitational field .
cf . a related method by @xcite .
@xcite introduced a 3d extension of his method , which was called scf by ( * ? ? ?
* hereafter ho92 ) by analogy to a similar technique used in stellar physics @xcite ; further historical developments are discussed in section [ sec : radial - basis ] . to exploit recent developments in the world of general purpose computing on gpus
, we implemented both scf and mex routines in a code called _ etics _
( acronym for _ expansion techniques in collisionless systems _ ) . in section [ sec : formalism ] we explain the mathematical formalism of both methods and highlight the differences between them . in section 3 we explain the unique challenges in a gpu implementation and measure the code s performance . in section 4
we discuss the accuracy of expansion and direct techniques . in section 5 we present a general discussion and finally summarize in section 6 .
here we clarify some terms used throughout this work : expansion methods : : a way to get potential and force by summing a series of terms ; in this paper either mex or scf .
mex : : multipole expansion method ( sometimes known in the literature as the spherical harmonics method ) ; expansion of the angular part .
scf : : self - consistent field method ; a `` pure '' expansion method since both angular and radial parts are expanded .
_ etics _ : : expansion techniques in collisionless systems ; the name of the code we wrote , which can calculate the force using both mex and scf , using a gpu .
gpu : : graphics processing unit ; a chip with highly parallel computing capabilities , originally designed to accelerate image rendering but is also used for general - purpose computing .
it often lies on a video card that can be inserted into an expansion slot on a computer motherboard .
both mex and scf methods are ways to solve the poisson equation : @xmath16 the formal solution of which is given by the integral : @xmath17 the expression @xmath18 is the green s function of the laplace operator in three dimensions and in free space ( no boundary conditions ) , and the integral is over the whole domain of definition of @xmath19 . in an @xmath6-body simulation , the density field @xmath19
is sampled at @xmath6 discrete points @xmath20 , such that @xmath21 where @xmath22 is the 3d dirac delta function .
direct @xmath6-body techniques evaluate integral ( [ eq : pot-3d ] ) _ directly _ : @xmath23 and thus at each point @xmath24 where the potential is evaluated , require @xmath6 calculations of inverse distance , or @xmath25 if @xmath26 , since there is no self - interaction . in practice , we are interested in evaluating the potential at the same points in which the density field is sampled , and thus a `` full '' solution of the poisson equation requires @xmath27 inverse distance calculations . in both mex and scf the integrand in equation ( [ eq : pot-3d ] ) is expanded as a series of terms , each of which more easily numerically integrable ; this is done in two different ways , lending the two methods quite different properties . in both methods , the reduction in numerical effort comes at the expense of accuracy compared to direct - summation , but this statement is arguable since in practice direct @xmath6-body techniques use a very small number of particle to sample the phase space . to demonstrate the difference between the two approaches in the following section ,
let us consider a 1d version of integral ( [ eq : pot-3d ] ) ; let us further assume that the density exists in the interval @xmath28 : @xmath29 note that this is not a solution for a 1d poisson equation ( hence the notation @xmath30 instead of @xmath31 ) , but just a simplification we will use to illustrate the properties of each method .
we will conveniently ignore the fact that this integral is generally divergent in 1d , as it does not affect the following discussion . in brief
, mex is a _
taylor_-like expansion of the green s function , while scf is a _
fourier_-like expansion of the density .
this already hints at the most critical difference between the mex and scf : while the former , like a taylor series , is local in nature , the latter is global .
another way to look at it is that in both methods the integrand is written as a series of functions ( of @xmath32 ) with coefficients : in mex one uses the given density to evaluate the functions , while their coefficients are known in advance ; in scf one evaluates coefficients , while the functions are known in advance .
let us define @xmath33 and expand the green s function equivalent in equation ( [ eq : pot-1d ] ) around @xmath34 , we get that for @xmath35 or @xmath36 : @xmath37 while for @xmath38 or @xmath39 we can expand around @xmath40 : @xmath41 the first and second terms of integral ( [ eq : pot-1d ] ) define the functions @xmath42 and @xmath43 ( utilizing the commutativity of the sum and integral operations ) : @xmath44 and thus @xmath45.\ ] ] while seemingly we made things worse ( instead of one integral to evaluate , we now have a series of integrals ) , the fact that @xmath32 has moved from the integrand to the integral s limit greatly simplifies things .
let the density @xmath46 be sampled at @xmath47 discrete and ordered points @xmath48 ; it is easy to show that @xmath49 in other words , each of these functions is a cumulative sum of simple terms and can be evaluated at all @xmath50 in just one pass , but a sorting operation is required .
let us leave the green s function as it is , and instead expand the density as a generalized fourier series : @xmath51 where @xmath52 is a complete set of real or complex functions ( the basis functions ) ; orthonormality of the basis functions is assumed above .
the integral ( [ eq : pot-1d ] ) becomes : @xmath53 the function set @xmath54 is defined by the above integral .
in essence , we replaced the integral over an arbitrary density @xmath46 with an integral over some predefined ` densities ' @xmath55 .
the advantage is that we can calculate the corresponding potentials , @xmath56 in advance , and then the problem is reduced to numerically determining the coefficients @xmath57 .
the choice of the basis is not unique , and an efficient scf scheme requires that the following : 1 .
the functions @xmath55 and @xmath56 are easy to evaluate numerically .
the sum ( [ eq : scf - sum ] ) convergence quickly , or in other words , @xmath58 is already close to @xmath46 .
the standard form of mex in 3d is @xmath59y_{lm}(\theta,\phi)\label{eq : mex - phi}\\ q_{lm}(r ) & = & \int_{r'<r}r^{\prime l}\rho({\bf r}')y_{lm}^{*}(\theta',\phi'){\rm d}^{3}r'\label{eq : mex - qlm}\\ p_{lm}(r ) & = & \int_{r
< r'}r^{\prime-(l+1)}\rho({\bf r}')y_{lm}^{*}(\theta',\phi'){\rm d}^{3}r'.\label{eq : mex - plm}\end{aligned}\ ] ] all together there are @xmath60 complex function pairs ( not counting negative @xmath61 , which are complex conjugates of the others ) that need be calculated from the density . since in practice the density field
is made of @xmath6 discrete points , they must be sorted by @xmath10 in order for the above integrals to be evaluated in one pass . the standard form for scf is : @xmath62 all together there are @xmath63 complex coefficients ( not counting negative @xmath61 ) that need be calculated from the density .
a typical choice is @xmath64 , for which there would be 308 coefficients .
the radial basis functions and coefficients for scf are discussed in the next section .
spherical harmonics are used in both cases to expand the angular part , but alternatives exist , such as spherical wavelets ( e.g. @xcite ) .
mex has two sums ( one infinite ) while scf has three sums ( two infinite ) . in practice
, the radial and angular infinite sums must be cut off at @xmath65 and @xmath66 , respectively .
the finite sum could in principle also be truncated to discard azimuthal information .
simply equating the expressions gives the relation between the two methods : @xmath67=\sum_{n=0}^{\infty}a_{nlm}\phi_{nl}(r),\ ] ] where @xmath68 is the @xmath69-pole . in case
the system is azimuthally symmetric , @xmath70 for all @xmath71 .
also , the same azimuthal information is carried in positive and negative @xmath61 terms , and they are related to each other by complex conjugation . if one decompose the density to a spherical average @xmath72 and the non - spherical deviation @xmath73 , then it is easy to show that @xmath74 depends only on the spherical average , while all other term depend only on the deviation . in a spherically symmetric system only @xmath75
is nonzero , and setting @xmath76 yields an accurate result .
while the choice of @xmath66 depends only on the deviation of the system from spherical symmetry , the choice of @xmath65 in scf depends on how well the system is described by the the zeroth radial basis function , and is usually determined by trial and error ( see section [ sec : inifinite - particle ] ) .
it is interesting to note a nontrivial mathematical difference between the two methods .
one can show that the laplacian of equation ( [ eq : mex - phi ] ) is zero when substituting the appropriate expressions for @xmath77 and @xmath78 from the equations ( [ eq : mex - qlm ] ) and ( [ eq : mex - plm ] ) ; the proof is mathematically cumbersome and will not be brought here .
this is surprising , since according to the poisson equation the result should be proportional to the density .
one can not appeal to series truncation to resolve this apparent contradiction ; indeed each term in the formally non - truncated infinite series yields a zero density , despite representing the multipoles as continuous functions .
the solution is that the potential at point @xmath79 has contributions from all internal ( i.e. at @xmath80 ) particles ( represented by @xmath77 ) and all external particles ( represented by @xmath78 ) , but no information about the density at @xmath79 itself .
this is the case also when the potential is constructed by a direct - summation of all gravitational point sources , so one may say mex is similar to direct methods in this sense . in scf , by construction , taking the laplacian of equation ( [ eq : scf - phi ] ) leads right back to the density field ( equation [ eq : poisson ] ) .
one can thus use the coefficients @xmath81 to represent a smoothened field .
one can also use mex for this purpose , if the derivatives of @xmath68 are calculated on a grid or with a spline .
a key difference between mex and scf is the freedom of choice of _ radial _ basis .
there are in fact two function sets : the radial densities @xmath82 and the radial potentials @xmath83 ; they are related via the poisson equation @xmath84 ( in this case @xmath85 only contains derivatives with respect to @xmath10 ) .
the choice of basis is not unique , and the basis functions themselves need not represent physical densities and potentials ( i.e. @xmath86 could be negative ) .
however it is convenient to take the zeroth term ( @xmath87 ) to represent some physical system , and to construct the rest of the set by some orthogonalization method , such as the gram
schmidt process .
the idea of @xcite was to use a @xcite model as the zeroth term and construct the next orders using the gegenbauer ( ultraspherical ) polynomials and spherical harmonics ( cf .
@xcite who developed a virtually identical method for finite stellar systems using spherical bessel functions for the radial part ) .
ho92 constructed a new radial basis ( also using gegenbauer polynomials ) which zeroth order was a @xcite model ; this is the basis set we adopt in _ etics_. they argued that this basis was more well suited to study galaxies .
more basis sets followed .
@xcite used the idea of @xcite , that the basis does not have to be biorthogonal , to construct as set which zeroth order was oblate . @xcite gave a radial basis set for the more general @xmath88-model ( of which both plummer and hernquist are special cases ) and @xcite introduced a basis for thick disks in cylindrical coordinates .
@xcite , @xcite describe numerical derivation of the radial basis set so that the lowest order matches any initial spherically - symmetric model , so called `` designer basis functions '' . @xcite introduced an analytical set which zeroth order is the perfect sphere of @xcite .
there are several levels of task parallelism available when writing computer code . at one level
, tasks are performed in parallel on different computational units ( such as cpus ) but only one copy of the data exists , which is accessed by all tasks ; this is called a _ shared memory _ scheme .
the tasks are called `` threads '' , and they are generally managed within one `` process '' of the program . a higher level of parallelism is called _ distributed memory _
scheme , where tasks are performed on different units ( often called `` nodes '' ) , but each unit has access only to its own memory ; thus data must be copied and passed . in this case
the parallel tasks are different processes , and cooperation between them is facilitated by a message passing interface ( mpi ) .
the parallel programming model is different between shared and distributed memory ; the former is considered easier since threads can faster and more easily cooperate . a high - performance supercomputer will generally enable parallelism on both levels : these machines are made of multiple nodes , each of which has its own memory and multiple computational units .
graphics processing units ( gpus ) are powerful and cost - effective devices for high performance parallel computing .
they are used to accelerate many scientific calculations , especially in astrophysics , such as dynamics of dense star clusters and galaxy centers ( @xcite ; @xcite ; @xcite ; @xcite ; see review by @xcite ) .
the gpu contains its own memory and many computational units , thus it is a shared memory device .
scf force calculation is particularly easy to parallelize , since the contribution of each particle to the coefficients @xmath81 is completely independent of all other particles .
particle data can be split to smaller chunks ( each could be on a different node ) , from each chunk partial @xmath81-s are calculated . then the partial coefficients summed up and the result communicated to all the nodes .
this was done by ( * ? ? ?
* hereafter h95 ) , whose code used the mpi call ` mpi_allreduce ` to combine the partial coefficients .
this parallelization scheme , however , is not suitable for gpus , as discussed in section [ sec : implementation - scf ] .
mex force calculation is harder to parallelize since the contribution of each particle depends on its position in a sorted list ( by radius ) .
however , in a shared memory scheme this too could be achieved relatively easily as explained in the following section .
the current implementation of the mex method relies on _ thrust _ @xcite , a c++ template library of parallel algorithms which is part of the cuda framework .
it makes parallel programming on a shared memory device ( either a gpu or a multicore cpu ) transparent , meaning that the task is performed in parallel with a single subroutine call , and the device setup and even choice of algorithm is performed by the library . _
thrust _ provides a sorting routine that selects one of two algorithms depending on input type . in the current version of mex and using version 1.6 of _ thrust _ , a general merge sort algorithm @xcite is used .
a flowchart of the entire mex routine is shown in fig .
[ fig : flowchart - mex ] .
the flow is controlled by the cpu , and boxes with double - struck vertical edges indicate a gpu - accelerated operation .
the blue double - struck boxes represent _ thrust _ calls , while the black ones are regular cuda kernel calls .
when a gpu operation is in progress , the cpu flow is paused .
[ fig : memory - mex ] shows the four main memory structures of the program and how the _ thrust _ subroutines and kernels in the program operate on them .
the particle array contains all particle coordinates and also the distance square from the center , which needs to reside in this structure for the sorting operation ( in practice the particle array contains additional data such as i d and velocity , but this is not used by the mex routine ) ; the cache structure contains functions of particle coordinates which are needed to calculate the multipoles .
kernel1 , which is executed once , reads the coordinates , calculates those functions and fills the cache structure .
kernel2 calculates the spherical harmonics at the current @xmath89-level and from that the contribution of the particle to @xmath77 and @xmath78 , which are saved in global memory .
when this kernel returns , the _ thrust _ subroutines are dispatched to perform the cumulative sum .
the `` scan '' ( forward cumulative sum ) and `` r. scan '' ( reverse scan ) are both in fact calls to the ` exclusive_scan ` subroutine , but to perform the reverse scan , we wrap @xmath78 with a special _ thrust _ structure called ` reverse_iterator ` . not shown in the flowchart , the two scan subroutines have to be called @xmath89 times at each @xmath89-level since they work on one @xmath61 value at a time .
kernel3 has both cache and compute operations : it calculates the partial forces in spherical coordinates ( i.e. the @xmath89-order correction to the force ) and/or potentials by evaluating all the spherical harmonics again ( and their derivatives with respect to spherical coordinates ) . later it advances @xmath90 and @xmath91 to the next @xmath89-level ( except at the last iteration ) . finally , the last kernel operates on the force structure and transforms it to cartesian coordinates .
[ fig : pie - mex ] shows the relative time it takes to do the internal operation .
we note that the potential could be calculated at the same time as the force ( in kernel3 ) and stored at another memory structure ( not shown in fig .
[ fig : memory - mex ] ) but is skipped if only forces are needed . alternatively
, only the potential could be calculated ( this is faster since the derivatives of the special functions are not calculated ) .
the choice between calculating force , potential or both is done with c++ templates .
we first briefly explain the serial algorithm used by ho92 .
the force ( and potential ) calculation had two parts : ( 1 ) calculation of all the @xmath81-s ( the plural suffix ` -s ' to emphasize that there are hundreds of coefficients in this 3d structure ) and ( 2 ) calculation of all the forces using the coefficients . in both parts ,
the particle loop ( the @xmath92-loop ) was the external one , inside of which there are again two main steps . in step ( 1a )
all necessary special functions were calculated using recursion relations .
step ( 2a ) was identical but additionally , the derivatives of those functions were calculated . in step ( 1b )
there was a nested loop ( @xmath93-@xmath89-@xmath61 structure ) in which a particle s contribution to every @xmath81was calculated and added serially . in step ( 2b ) there was also such a loop , which used all the @xmath81-s to calculate the force on each particle . in the parallel algorithm used by h95
, another part was added between the two parts mentioned above : communicating all partial @xmath81-s from the various mpi processes , adding them up and distributing the results . in practice
it was achieved using just one command , ` mpi_allreduce ` .
there are two main reasons why this algorithm could not be used effectively on a gpu , both are related to the difference between how the gpu and cpu access and cache memory . the first difficulty is performing the sum .
the partial sums from the different parallel threads could in principle be stored on a part of the gpu memory called _ global memory _ , and then summed in parallel .
however a modern gpu can execute tens of thousands of threads per kernel ( note that the concept of a thread in cuda is abstract , and the number of threads by far exceed the number of thread processors on the gpu chip ) , and every partial @xmath81 is @xmath94 kilobyte in size ( depending on @xmath95 and @xmath96 ) .
thus , writing and summing the partial coefficients would require extensive access to global memory , which is slow compared to the actual calculation part .
the second difficulty is that if one thread uses too much memory , for example to store all necessary legendre and gegenbauer polynomials as well as complex exponent ( as is done in the ho92 code ) , this may lead to an issue called _ register spilling _ , where instead of using the very fast register memory , the thread will store the values on the slow global memory , which again we wish to avoid on performance grounds . to tackle those issues we utilized another type of gpu memory called _
_ shared memory__. this memory is `` on chip '' ( on the multiprocessor circuit rather than elsewhere on the video card ) and has @xmath97 lower latency than global memory .
threads in a cuda program are grouped into blocks , threads in the same block share this fast memory ( hence the name ) .
it is also much less abundant than global memory .
the nvidia tesla k20 gpus have just 64 kilobytes of shared memory per block , while they have 5 gigabytes of global memory .
in order to use shared memory to calculate the coefficients , each thread would serially add contributions from particles to the partial @xmath81-s on shared memory ; then they would be summed up in parallel in each block .
however , there are usually hundreds of different @xmath81-s , as well as tens or hundreds of threads per block ( depending on hardware ; which is required for efficient loading of the gpu ) ; there is not enough shared memory for that ( by far ) . to solve this
, we changed the order of the loops : the external loop is the @xmath89-loop , then comes the @xmath93-loop . for each @xmath98 pair
, a cuda kernel is executed where the @xmath92-loop is performed in parallel on different threads , inside of which the @xmath61-loop is done .
now each threads has to deal with far fewer @xmath81-s ( no more than @xmath99 ) , for which there is usually enough shared memory .
but for the scf routine .
] a flowchart of the entire scf routine is shown in fig .
[ fig : flowchart - scf ] .
the flow is controlled by the cpu , and boxes with double - struck vertical edges indicate a cuda kernel call .
when a gpu operation is in progress , the cpu flow is paused .
[ fig : memory - scf ] shows the four main memory structures of the program and how the five kernels in the program operate on them .
the particle array contains all particle coordinates ( in practice it contains additional data such as i d and velocity , but this is not used by the scf routine ) ; the cache structure contains functions of particle coordinates which are needed to calculate the basis functions .
kernel1 , which is executed once , reads the coordinates , calculates those functions and fills the cache structure .
kernel2 only operates on the cache structure , it has just one function which is to advance @xmath100 by one level ; thus it needs to be executed at the beginning of each iteration of the @xmath89-loop .
as shown in the flowchart , it is skipped for @xmath101 because kernel1 calculates and caches @xmath102 .
kernel3 has both cache and compute operations : it calculates the current @xmath103 using recursion relations from the cached @xmath104 and @xmath105 and then updates the cache .
later it calculates the spherical harmonics and from that the contribution of the particle to the all @xmath81 in the current ( @xmath93,@xmath89)-level , which are saved in shared memory .
when all threads in the block have finished calculating contributions of the particles assigned to them , they are synchronized and a parallel reduction is performed . since threads from different blocks can not share memory , the data from each block must be transfered to the host machine s memory and the cpu finishes the summation process . for the force calculation , just a reading the @xmath81-s is required .
the gpu has yet another type of memory which is ideal for storing of coefficient or constant parameters .
it is fittingly called `` constant memory '' , and is as fast as shared memory when every thread in a warp accesses the same memory element .
it is also very limited ( usually to 64 kilobytes per device ) , but the @xmath81structure could still fit there nicely .
once calculation of all the coefficients is complete , it is transferred back to the gpu constant memory to be used to calculate the forces . since only reading the coefficient is required , in kernel4 which calculates the forces and/or potentials by evaluating all the basis functions again ( and their derivatives with respect to spherical coordinates ) , the @xmath92-loop is the external one . to avoid register spilling
we keep the internal loop structure as @xmath89-@xmath93-@xmath61 , and thus we only need to recalculate the complex exponents , which is relatively cheap .
finally , the last kernel operates on the force structure and transforms it to cartesian coordinates . fig .
[ fig : pie - scf ] shows the relative time it takes to do the internal operation .
we note that the potential could be calculated at the same time as the force ( in kernel4 ) and stored at another memory structure ( not shown in fig .
[ fig : memory - scf ] ) but is skipped if only forces are needed . alternatively , only the potential could be calculated ( this is faster since the derivatives of the special functions are not calculated ) . the choice between calculating force , potential or both is done with c++ templates .
we tested the performance of _ etics _ ( both mex and scf ) on a single nvidia tesla k20 gpus on the laohu supercomputer at the naoc in beijing . for comparison
, we also tested the fortran cpu scf code by lars hernquist on the accre cluster at vanderbilt university in nashville , tennessee ( we used a node with intel xeon e5520 cpu ) .
if the initial conditions are not sorted by @xmath10 in advance , the first mex force calculation is more costly than all the following , since the sorting of an already nearly - sorted particle list is faster .
thus , all measurements of the mex code are done after the system is evolved one very short leapfrog time step . fig .
[ fig : scaling ] shows the time it takes to do one full force calculation as a function of @xmath6 , @xmath96 and @xmath95 .
each point represents the mean time of 10 different calculations .
the dispersion is generally very low , with the exception of _ etics_-mex with @xmath106 ; only for which we show error bars .
note that the timing only depends on the number of particles ( and expansion cutoffs ) and not on their spatial distribution .
the cpu and gpu scf codes are both theoretically @xmath107 . at low @xmath6
the gpu is not fully loaded , and _ etics _ performance seems superlinear with @xmath6 .
_ etics_-mex is theoretically @xmath108 , but this again is an asymptotic behavior which is not observed .
the lack of good gpu load for @xmath109 is much more evident than the @xmath110 nature of the algorithm .
the gpu global memory was the limiting factor in how many particles could be used with both methods .
the dotted lines show the performance of _ etics _ using single - precision instead of double .
the speed increase is 61% for scf and 65% for mex , but there is a price to pay in accuracy as noted in section [ sec : single ] .
the speedup factor could be very different for different gpus .
all codes should scale quadratically with @xmath96 , but as the middle panel of fig .
[ fig : scaling ] shows , this behavior is not so clear for _ etics_-mex .
this is due to the extensive memory access this code requires , which rivals the calculation time .
memory latency on gpus is not easy to predict ; due to caching and the way memory is copied in blocks , and the latency depends not only on the amount of memory accessed but also on the memory access pattern .
scf codes theoretically scale linearly with @xmath95 .
a strange behavior of the cpu code is noted : it seems that the time increases with @xmath95 in a `` zigzag '' fashion ( the measurement error of the times is much smaller than this effect , and it is reproducible ) .
this is paradoxical : it takes a shorter time to calculate with @xmath111 than with @xmath112 , even though more operations are required .
it is not simple to understand why this is , but it seems that the compiler performs some optimization on the first @xmath92-loop ( coefficient computation ) that only help when @xmath95 is odd but not when it is even .
the comparison between _ etics_-gpu and hernquist s code is not exactly fair since they use different types of hardware . specifically for hardware we tested , _
etics_-gpu outperforms hernquist s code by a factor of about 20 ( which depends little on all parameters ) .
however , hernquist s code can utilize a multicore cpu ( using mpi ) .
the xeon cpu we used has 4 cores , and two such cpus are mounted on a single accre node .
we could use the fortran code in mpi mode on all 8 effective cores with almost no overhead , and the calculation is accelerated by a factor of 8 .
also , hernquist s code calculate the jerk ( force derivative ) , which _
etics_-gpu does not ; this takes @xmath113 percent of the total time . and @xmath114 ) ; the total time is 0.15 sec on nvidia tesla k20 gpu .
the results may differ significantly on different hardware and if single - precision is used instead .
the first operation is sort , followed counter - clockwise by initialization of the cache arrays , the @xmath89-loop where each iteration is divided to ( a ) summand calculation , ( b ) cumulative sum and ( c ) partial force calculation .
the final operation is coordinate transformation from spherical to cartesian ] for a full scf force calculation with _ etics _
( @xmath115 particles , @xmath114 and @xmath116 ) ; the total time is 0.16 sec on nvidia tesla k20 gpu .
the first operation is initialization , followed counter - clockwise by the @xmath89-loop ( in which the @xmath93-loop is nested ) .
the partial force calculation is a single cuda kernel , inside of which all the loops are performed . ]
[ fig : pie - mex ] and [ fig : pie - scf ] show the fraction of time it takes to perform the internal operations for the force calculation for _ etics_-mex and -scf , respectively , both use @xmath115 , @xmath114 and for scf , @xmath116 . for mex , operations inside each iteration of the @xmath89-loop are shown in different shades ( also denoted by letters corresponding to stages 3a , 3b and 3c as explained in section [ sec : implementation - mex ] ) .
the most costly operations are the ones we entrust to _ thrust _ , namely the sorting and cumulative sum . in fig .
[ fig : pie - scf ] the internal structure of each @xmath89-iteration is not shown ( since there are too many internal operations , including the @xmath93-loop ) .
the force calculation is executed as one operation ( a single cuda kernel call ) , and includes the @xmath89-loop nested inside it ( unlike mex where only a partial force was calculated at every @xmath89-iteration , step 3c ) .
two separate questions come up when discussing the accuracy of expansion methods : how well the expansion approximates the @xmath6-body force ( i.e. direct - summation ) , and how well it approximates the smooth force in the limit of infinite particles ( which we will refer to as the `` real '' force in the following discussion ) .
both questions depend on @xmath6 , @xmath96 and ( for scf ) @xmath95 .
a related question is how well the @xmath6-body force approximates the real force , as a function of @xmath6 .
all these questions depend not only on the expansion cutoff and @xmath6 , but on the stellar distribution as well ( e.g. global shape , central concentration , fractality , etc . ) ; this will not be fully explored in this work .
there are two types of error when considering the expansion methods versus the real force , analogous to systematic and random errors .
the first , systematic - like error , comes from the expansion cutoff , this is called the _
bias_. for example , a system which is highly flattened could not be described by keeping just the quadrupole moment , so both mex and scf cut off at @xmath117 would exhibit this type of error , regardless of @xmath6 ( see @xcite ; @xcite for discussion about bias due to softening ) .
the second , random - like error , comes from the finite number of particle and their coarse grainy distribution ; it is the equivalent of @xmath6-body noise ( also referred to as particle noise or sampling noise ) .
ho92 attempted to estimate accuracy of scf by showing convergence of the coefficient amplitudes with increasing @xmath93 for the density profiles of some well known stellar models .
they showed that @xmath118 decayed exponentially or like a power law with @xmath93 , depending on the model .
this analysis was not satisfactory because it applied to the limit of infinite @xmath6 , thus ignoring the random - like error .
furthermore , showing convergence of the coefficients does not give information about the force error .
the bias and the random error are not easy to distinguish .
the bias could be calculated , in principle , only if the true mass density @xmath19 is known , which is not generally the case ; however , it is still useful to look at some particular examples where it is known . to test the accuracy of the expansions techniques , we used two simple models for the mass density . both our models are ferrers ellipsoids @xcite ( often called ferrers bars ) with index . ] @xmath119 : ` model1 ` is a mildly oblate spheroid with axis ratio of 10:10:9 , ` model2 ` is triaxial with axis ratio of 3:2:1 .
ferrers ellipsoids are often used in stellar dynamics , especially in the modeling of bars ( e.g. @xcite ) .
they have a very simple mass density : @xmath120^{n } & \mu\leq a\\ 0 & \mu > a \end{cases}\ ] ] where @xmath121 , @xmath122 and @xmath123 are the axes , @xmath124 is the central density , @xmath125 is the index and @xmath126 is the ellipsoidal radius , defined by : @xmath127 the potential due to this family of models is simply a polynomial in @xmath128 if @xmath93 is an integer . the coefficients could be calculated numerically ( also analytically for some cases ) by solving a 1d integral ( * ? ? ? * chapter 2.5 ) . for both our models we used the mathematical software _
sage _ to calculate the coefficients to better than @xmath129 .
the force vector components are trivially derived from the potential polynomial ; this is the `` real '' force .
we created many realization of these two models , ranging from just 100 particles to @xmath130 .
the goal is to compare for each realization the force calculated using mex , scf and direct - summation ( no softening ) , with the real force .
all calculations performed using double - precision , and the direct - summation force is not softened . for each realization
we get a distribution of @xmath6 values of the relative force error , @xmath131 where @xmath132 is the particle s index .
it is not practical to show to full distribution for all cases , so in figs .
[ fig : force - error1 ] and [ fig : force - error2 ] we show the mean , and the full distribution for only selected cases . the left panel of fig .
[ fig : force - error1 ] shows the mean relative force error @xmath133 in ` model1 ` for direct - summation and mex with even @xmath96 between @xmath134 and @xmath135 ; odd terms are in principle zero if the expansion center coincides with the center of mass , and in practice very small . for this model
, @xmath133 is decreasing with @xmath6 for all cases but @xmath136 ( monopole only ) .
the smallest error is for @xmath137 ( monopole and quadrupole only ) .
unintuitively , adding correction terms _ increases _ the error ( for constant @xmath6 ) , this is because the model s deviation from sphericity is so mild , that the quadruple describes it well enough ; the following terms just capture some of the @xmath6-body noise in the realization and make more harm than good . in the right panel
we show the full log - distribution for selected cases .
the histograms for @xmath138 are made by stacking of @xmath139 realizations , so there are @xmath130 values of @xmath140 in all histograms . in all cases the @xmath140 distributions are close to log - normal
; the logarithmic horizontal axis hides the fact that the distributions on the right are much wider in terms of standard deviation due to a very long and fat tail when viewed in linear space .
note that while the number of particles increased by 1000 , in all cases the error distribution shifted down by just a factor of @xmath141 .
[ fig : force - error2 ] is the same but for the triaxial ` model2 ` . while in the @xmath6-body cases
the @xmath140 distributions are much the same , mex shows a different behavior . the most prominent feature is the bump on right side of the @xmath115 , @xmath114 error distribution , which demonstrates the issue of bias .
most of the particles which make up this bump are located in the lobes of the ellipsoid , where many angular terms are required . when @xmath96 is increased to @xmath135 ,
this bump disappears .
it also is not present in the @xmath138 , @xmath114 case , probably because it is overwhelmed by the random error .
this bump causes the mean error to saturate with particle number , as the left panel shows .
increasing @xmath6 will shift the bulk of the bell curve to the left ( zero ) , but will not quench the bump . at much larger @xmath6 ,
` model1 ` will show the same behavior as the random error becomes smaller than the bias , and the high-@xmath96 cases would outperform @xmath137 .
we repeat this exercise for scf , which has an additional source of bias due to the radial expansion cutoff . fig .
[ fig : force - scf - error2 ] illustrates that point by showing the relative force error distribution in ` model2 ` for scf compared to mex with the same number of particles ( @xmath115 ) and same angular cutoff ( @xmath142 ) . with increasing @xmath95 , the scf error distribution approaches that of mex , demonstrating the point made in section [ sec:3d ] that mex is equivalent to scf with @xmath143 .
it must be noted that the basis set we programmed in _ etics _ is not at all suitable for ferrers models ( which are finite and have a flat core ) , and the apparently slow convergence should not disparage one from using scf , even if it is not known in advance what basis to choose .
the overlap between the relative force error distributions of scf at @xmath116 and mex is 77% . a more intelligent choice of basis function is discussed by @xcite , who used a similar methodology to choose the best basis set for triaxial @xcite @xmath144-models @xcite with @xmath145 from a family of basis sets similar to the ho92 .
the results presented in this section suggest that there is some optimal expansion cutoff , which is different for different models and depends on the number of particles @xcite .
this is analogous to optimal softening in direct - summation force calculations @xcite .
if not enough terms are used , there is a large bias ; if too many terms are used , the particle noise dominates .
@xcite addressed this issue by calculating the variance of each scf coefficient among several realizations of the same triaxial dehnen model , found that for @xmath115 particles , angular terms beyond @xmath146 are dominated by noise ( and that only the first few @xmath93 , @xmath61 terms at that l - level are reliable ) . the force error discussed above is not directly related to energy diffusion or relaxation , which are reduced due to the smoothing , but not absent .
the mechanism for energy ( and angular momentum ) diffusion in both expansion methods is temporal fluctuation of the multipoles or coefficients ( due to the particle noise ) .
this is somewhat analogous to two - body relaxation in that the potential felt by every particle fluctuates ( although in this case there is no spatial graininess ) .
* in prep . )
examined energy diffusion in a plummer sphere with @xmath147 particles using scf and direct @xmath6-body codes , and found that scf demonstrated a diffusion rate only several times lower , which was close to the rate in a direct technique using near - optimal softening for this @xmath6 .
further reduction was achieved by discarding of expansion terms which are nominally zero in any triaxial system centered around the expansion center .
finally , vasiliev used temporal softening ( ho92 ) , where the coefficients ( and thus the potential ) are updated in longer intervals than the dynamical time step ; this procedure however introduces a global energy errors unless some measures are taken to amend this . .
the green histogram is also shown in the right panel of fig .
[ fig : force - error2 ] and represent a mex expansion with @xmath142 .
the other histograms represent scf expansions with @xmath142 and varying @xmath95 values as shown . with increasing @xmath95 , the scf error distribution approaches that of mex with the same @xmath96 .
in this case , the model differs greatly from the zeroth order function of the basis set , showing relatively slow convergence . ]
due to their original intended use , gpus are not optimized for double - precision arithmetic ( indeed early gpus completely lacked a double - precision floating - point type ) . in cards that do support double - precision , arithmetic operations
could still be significantly slower than for single . as noted before , in our test we measured a 6065% speed increase when using single - precision .
the nvidia tesla k20 gpus we used have enhanced double - precision performance with respect to other gpus , for which using double - precision may be significantly slower .
those devices are somewhat specialized for scientific use and are thus more expensive ( albeit in many applications still superior to parallel cpu architectures in terms of price / performance ratio due to the low energy consumption ) .
cpus usually take the same time to perform an arithmetic operation in either single- or double - precision , but a program s general performance could be faster in single - precision due to smaller memory load . for the hernquist - scf cpu code , we measured a 6% improvement in speed .
using single - precision however inevitably reduces the accuracy of the calculated force ; here we examine how bad this _ performance - accuracy trade - off _ is . fig .
[ fig : single ] show the relative force error distributions of single - precision calculations , compared to double .
the relative force error on particle @xmath132 is now defined as : @xmath148 we testes an @xmath115 realization of a hernquist sphere with characteristic scale of one unit .
the top panel shows two scf force calculations : the green histogram ( on the left ) is a low order expansion up to @xmath149 , retaining 36 coefficients ; the red histogram is an expansion up to @xmath150 , retaining 308 coefficients .
the bottom panel similarly shows two mex expansions . in both cases ,
the higher order expansion has relatively large errors . while it is still smaller than the error with respect to the `` real '' force discussed in section [ sec : inifinite - particle ] , its nature is numeric and it could hinder energy conservation .
the relatively large error is not remedied by usual methods to improve accuracy of floating point arithmetic such as kahan summation algorithm @xcite , because the error does not come from accumulation of round off errors . instead , the accuracy bottle neck is the calculation of the spherical harmonics and/or the gegenbauer polynomials .
particles for which those special functions are calculated with large numerical error will have a large force error , but additionally they contribute erroneously to _ all _ the coefficient or multipoles , thus causing some error in the force calculation of all other particles as well .
there are two groups of particles with large relative error in this implementation : particles that are very far away from the center , and particles which happen to lie very close to the @xmath151-axis .
the former group is not so problematic since the absolute force is very small as well as their contribution to the coefficients or multipoles .
the latter group causes large error because the recursion relation used to calculate the associated legendre polynomials : @xmath152 is not upwardly stable because of the @xmath153 factor , which diverges when the polar angle @xmath154 is very small or very close to @xmath155 ( although the polynomials themselves approach zero in these limits ) .
the distributions shown in fig .
[ fig : single ] may vary significantly depending on the model .
for example , ferrers ellipsoids are finite and flat at the center , thus they do not contain the problematic particles described above and have much smaller error in single - precision .
a hernquist sphere is more representative of the general case in galaxies , being infinite and relatively centrally concentrated .
one could conceivably improve the accuracy at single - precision in several ways . in the test
described above everything was calculated in single - precision , apart from some constant coefficients that were only calculated once , in double - precision , and then cast to single .
it may be possible to identify the most sensitive parts of the force calculation and use double - precision just for those , or use pseudo - double - precision ( as in @xcite ) for part of or the entire force calculation routine .
another possibility is to keep using single - precision for everything but prescribe special treatment to those orbits close to the @xmath151-axis . ) .
the model used is a hernquist sphere with @xmath115 .
the top and bottom panels shows scf and mex force calculations , respectively . in both cases
the left ( green ) histogram is a lower - order expansion as indicated in the legend . ]
expansion techniques , on their own , are best geared to simulate systems with a dominant single center , where it is important to minimize the effects of two - body relaxation , and where the system potential does not change radically ( quickly ) with time .
an ideal class would be long - term secular evolution in a near - equilibrium galaxy .
both methods presented in this work can be used to quickly calculate the gravitational potential and force on each particle in a many - body system , while discarding small scale structure .
mex comes from a taylor - like expansion of the green s function in the formal solution of the poisson equation , while scf is a fourier - like expansion of the density .
both methods are important tools for collisionless dynamics and has been used extensively in astrophysics as discussed in the following sections .
they are comparable in terms of both accuracy and performance . in both methods , there are free parameters to be set : the choice of length unit ( or model scaling ) affects the accuracy of scf expansion because the zeroth order of the radial basis functions corresponds to a model of a particular scale .
for example , the basis set offered by _ etics
_ corresponds to a @xcite model with scale length @xmath156 .
the main difference for the end - user is that scf smooths the radial direction as well .
this could be an advantage when @xmath6 is very small , since scf will still provide a rather smooth potential , although it might not represent the real potential well at all due to random error . in mex
, particles are not completely unaware of each other , and every time two particles cross each other s shell , there is a discontinuity in the force , which may lead to large energy error when @xmath6 is small .
this shell crossing occurs when two particles change places in the @xmath10-sorted list , and the particles need not be close to each other at all .
both methods have some problems close to the center . in scf ,
the limitation comes from both radial and angular expansions .
the radial expansion cutoff induces a bias if the central density profile does not match the zeroth basis function , and a very non - spherical model would cause force bias at the center as well as the lobes .
the latter is also a problem for mex , which has two additional problems : the discrete nature and inevitably small number of particles when one gets arbitrarily close to the center , as well as the numerical error ( and/or small step size required ) due to having to calculate @xmath157 ( the scf basis function we use are completely regular at the center ) .
it is clear from the literature that scf has been by far more popular . but
despite the above , we do not think that most authors intentionally avoided mex , and that scf was better publicized and became the standard .
mex is rarely used in its full form , but more frequently in the spherically symmetric version , sometimes called the `` spherical shells '' method ; in this case just the monopole term is kept ( @xmath136 ) .
for example , it is used in the poisson solvers of @xcite monte carlo method .
this hints that it might be easy to extends codes like mocca @xcite to non - spherical cases using our version of mex .
this monopole approximation has also been used to study dark and stellar halo growth @xcite .
the extension of the spherical case using spherical harmonics exists in several variations , divided roughly to two classes : grid and gridless codes .
the mex version presented in this work is gridless and follows from @xcite and @xcite .
these authors used cartesian instead of spherical coordinates , and softened the potential at the center .
this softening , albeit similar mathematically , is not equivalent to particle - particle softening in direct @xmath6-body simulations and was just used to prevent divergence at the center .
the first mex code however is by @xcite , who divided the simulation volume into thick shells , and the force on a particle was calculated by summing the multipoles of all shells except its own ( own - shell correction was added ) .
similarly , @xcite used a mex code with @xmath158 to explore galaxy correlation functions ; in their version each shell had six particles , and softened newtonian interaction was used within a shell .
as noted in the introduction , @xcite used a variation with axial symmetry ( up to @xmath14 but with no azimuthal terms , namely @xmath159 ) , with a grid in both @xmath10 and @xmath154 . in a follow up work @xcite the method was extended to 3d geometry .
finally , @xcite used a grid in @xmath10 only , with logarithmic spacing .
he argues that softening sacrifices the higher resolution near the center ( which is one of the primary advantages of the method ) and that a radial grid smooths the potential and prevents shell crossing .
recently , @xcite presented a similar potential solver with a spline instead of a grid .
we note that a virtually identical mathematical treatment to the mex method has been applied to solve the fokker - planck equation under the local approximation ( neglecting diffusion in position ) .
the collisional terms of the fokker - planck equation can be written by means of the rosenbluth potentials @xcite , which are integrals in velocity space very similar in form to equation ( [ eq : pot-3d ] ) .
@xcite assumed azimuthal symmetry and wrote the rosenbluth potentials using the legendre polynomials up to @xmath160 in a way exactly analogous to our equation ( [ eq : mex - phi ] ) .
this treatment was expanded to @xmath161 by @xcite .
as noted in the previous section , scf gained much more popularity than mex .
the scf formalism has had wide use on galaxy - scale problems .
it has been used to model the effect of black hole growth or adiabatic contraction on the structure ( density profile ) of the dark matter halo ( e.g. * ? ? ?
scf is also an appropriate tool to model the growth of the stellar and dark matter halos ( e.g. * ? ? ?
* ; * ? ? ?
* ) as well as the mass evolution of infalling satellite galaxies ( e.g. * ? ? ?
* ; * ? ? ?
one of the clearest uses of the scf technique is when the stability of the orbit matters such as in the study of chaos in galactic potentials ( e.g * ? ? ?
* ; * ? ? ?
* ) , and in the exchange of energy and angular momentum by mean resonances ( e.g * ? ? ?
* ; * ? ? ?
@xcite compare a number of methods and show that scf is superior for stability work . the initial motivation for _ this _ work was to follow up on @xcite , who studied supermassive black hole binaries using a restricted technique . in their method ,
the stellar potential was held constant while the black holes were treated separately as collisional particles ; it was thus not self - consistent in terms of the potential .
this class of problems , where there is a small subset of particles that need to be treated collisionally , has already been attempted using an extension of the expansion technique which hybridizes scf and direct aarseth - type gravitational force calculation ; in these extensions , either the black holes are the only collisional particle ( e.g * ? ? ?
* ; * ? ? ?
* ) , or all centrophilic particles are treated collisionally @xcite .
mex has not been applied to this particular problem to our knowledge , although it is as well suited as scf .
our scf implementation on gpu outperformed the serial hernquist cpu version by a factor of @xmath162 ( for double - precision ) , but this number depends on the particular gpu and cpu hardware compared .
the cpu code is definitely competitive on multi - core cpus .
intel recently introduced the many integrated core architecture ( known as intel mic ) , which are shared memory boards with the equivalent of tens of cpus . in principle , the fortran scf code for cpus could be adapted for this architecture with little modification , and it will most likely outperform the gpu version . on the other hand , next generation gpus ( such as nvidia s maxwell architecture ) would also deliver performance improving features , and it is not clear which one would win .
the goal of this project is to ultimately enable simulations of @xmath163 , and to perform them fast enough so that many could be performed , exploring large parameter space rather than making a few such large-@xmath6 simulations . to do that
, the code will be adapted to a multi - gpu and multi - node machines using mpi . as noted in section [ sec : implementation - parallel ] , this is easy for scf but not so much for mex . simultaneously we will attempt to improve the per - gpu performance .
we spent a lot of time trying to optimize this first version of _ etics _
, by no means we guarantee that out implementation is flawless .
some improvement might come from tweaking of the implementation .
for example , we decided not to cache @xmath164 but rather recalculate it in - kernel before every @xmath61-loop ( as a starting point for the recursion relation ) . since the legendre polynomials are `` hard - coded '' and computed very efficiently , it is not immediately clear if caching is a more efficient approach ( it is probably worth while at very high @xmath96 ) .
likewise , we chose to separate the caching operations that are performed once per routine or once per external loop , and execute them as independent kernels , while in principle they could be executed as statements inside the inner kernels ( so called `` kernel fusion '' , which would save kernel execution overhead ) , with an if - statement making sure that the cache operations are performed only if needed .
some possible more fundamental changes include trying to get rid of the sorting operation in mex ; while the most basic approach requires the particle list to be sorted and a cumulative sum performed over the multiples , some alternatives exist such as logarithmic grid ( as in @xcite ) or spline ( as in @xcite ) . also we might find a more sophisticated way to perform the cumulative sum , since we suspect that the _ thrust _ routines are not optimal for our uses .
another improvement might come from the integration side rather than force - calculations , such as implementation of higher order integrator instead of the leapfrog .
hernquist s scf code already contains a 4th order hermite scheme @xcite , which is not hard to implement for gpus , but mex has a fundamental problem with this scheme due to shell crossing , which causes the force derivatives to be discontinuous .
_ etics _ is a powerful code , but as with any computer program , one should understand its limitation .
the code in its current form should not be used for highly flattened system , or where two - body interactions are significant .
the code is momentarily available upon request from the authors , but we plan to make it public , including a module to integrate it with the amuse framework @xcite .
we thank peter berczik , adi nusser , marcel zemp and eugene vasiliev for the interesting and helpful discussions and the referee for useful comments .
ym is grateful for support from the china postdoctoral science foundation through grant no .
2013m530471 . the special gpu accelerated supercomputer laohu at the center of information and computing at national astronomical observatories , chinese academy of sciences , has been used for the simulations ; it was funded by ministry of finance of people s republic of china , under the grant zdy z2008 - 2 , and by the `` recruitment program of global experts '' ( qianren ) for rainer spurzem ( 2013 ) .
rs has been partially supported by nsfc ( national natural science foundation of china ) , grant no . | we present gpu implementations of two fast force calculation methods , based on series expansions of the poisson equation .
one is the self - consistent field ( scf ) method , which is a fourier - like expansion of the density field in some basis set ; the other is the multipole expansion ( mex ) method , which is a taylor - like expansion of the green s function .
mex , which has been advocated in the past , has not gained as much popularity as scf .
both are particle - field method and optimized for collisionless galactic dynamics , but while scf is a `` pure '' expansion , mex is an expansion in just the angular part ; it is thus capable of capturing radial structure easily , where scf needs a large number of radial terms . we show that despite the expansion bias , these methods are more accurate than direct techniques for the same number of particles . the performance of our gpu code , which we call _ etics _ , is profiled and compared to a cpu implementation . on the tested gpu hardware , a full force calculation for one million particles took @xmath0 seconds ( depending on expansion cutoff ) , making simulations with as many as @xmath1 particles fast on a comparatively small number of nodes . |
traditional methods for visual cryptography have been established , are consistent and easily understood .
unfortunately , these methods exist for black and white pictures only , leaving the encryption of colour images wanting . while there are a handful of attempts at bringing colour to visual cryptography
, it is still an open field with implementations of varying efficiency . in this paper
, we will establish some basic standards for encrypting colour pictures , as well as a simple , yet efficient method for encryption based upon those rules . while black and white pictures are fairly easy to work with due to their simple nature , colour pictures contain much more information and possibly details that may not be lost in the process .
this leads to a need for the static that normally appears in the decryption of black and white pictures to have to be absent[1 ] .
however , a partial reconstruction of the picture ( such as having less than all the necessary parts for full restoration ) may not hint at what the final image is meant to be .
in addition to the standard of security for the traditional black and white pictures , we set the following points as mandatory for encryption of colour images . * full restoration upon decryption . * no indication as to the original image , whether by eye or any other method , when combining a subset of all available parts .
* usability for any type of image , whether that image contains a mixture of colours , is black and white or is simply one single colour . *
destruction of intermediary steps in the encryption process .
our method of encryption is , due to the process of creating images within visual cryptography and the expanse of computer use , based around the rgb colour model for computers and other , similar devices .
this does not prevent implementation of this process with any other model as long as the information is stored as bits .
the encryption is fairly straightforward and can be easily understood as well as implemented .
although it is easy to handle , it is efficient , provides the necessary security and fulfills every point previously stated . for the process , the thought was to work on the bitwise level that represents the colours themselves ; in this case , the rgb values are used . as each pixel
is processed , two random values are generated ; the first one is compared to the rgb value of the current pixel and it is then separated into two values : an rgb value with the bits that were set in both the original as well as the first random value , and another one with the set bits left over from the original . next , the second random value is compared to the two new results from the previous step .
if both values are not set while the bit at the same position in the random value is , those bits for the values from the previous step are set .
these steps are repeated for every pixel and then the encryption is finished .
decryption is easily done via a bitwise xor of the rgb values of the two resulting pictures and we effectively have a one - time pad implementation on the colour values of an image .
the random values from the process of encryption are discarded alongside any other values we might have produced .
the encryption algorithm will `` split '' the original image into two so called _ shadow images_. let @xmath0 denote one pixel in the original image , and let @xmath1 and @xmath2 denote the corresponding pixels in the shadow images , respectively . here
@xmath0 , @xmath1 , and @xmath2 are vectors , representing the channels used , e.g. red , green and blue in the rgb colour model . the calculations during the encryption are carried out both bitwise and channel - wise .
@xmath3 @xmath4 @xmath5 @xmath6 we illustrate the encryption with an example .
assume that we want to encrypt the image in figure [ fig : ex1 ] .
then if we apply the algorithm , it may result in the two shadow images in figure [ fig : ex2]remember that the algorithm is probabilistic . to restore the original image , we compute @xmath7 for each pixel .
the result will give us back the original image , without any loss of quality , that is , figure [ fig : ex1 ] .
as with both visual cryptography and the one time pad , this method offers complete security as there is neither a repetition to be found , nor is a brute force attack possible as every possible result within the picture s resolution will show up .
this method does not only fulfill the standards we previously set , it even leaves the resolution of the original image intact .
of course , it should be noted that for this encryption to work to its full potential , the results must be saved as a lossless image type . in the event that one would wish to separate the original into more than two pictures , reapplying this process to the results until a satisfactory amount is reached is all one needs to do .
it could be argued that all this encryption would need is random values generated and applied with a bitwise xor to the original image , leaving the randomly generated sequence as one of the resulting images and the result of the bitwise xor as the other .
while this could be done , let us take a look at the absolute worst case scenario , disregarding the possibilities that an attacker knows the original or has access to all parts of the picture .
the scenario in mind would be in the highly unlikely event that an attacker would have an intimate enough knowledge of the encryption process to know exactly how it is implemented as well as knowing exactly which random values were generated .
if the method of encryption would be nothing more than the simplified version suggested , then having the result of the bitwise xor in your possession would be enough to get the original as being able to predict the exact pseudo - random values would mean that you effectively have the key .
this is not the case for the method we propose as depending on the random values and how they interact with the original data , only some bits could possibly be decided , as shown in table [ tab : values ] .
there are four possible cases for the pair @xmath8 .
each case can be divided into two subcases , where each corresponds to the possible value @xmath9 of @xmath0 . without loss of generality , we consider the cases on bit - level . * 5c case & @xmath10 & @xmath11 & @xmath9 & @xmath0 + 1 & 0 & 0 & 0 & + 2 & 0 & 0 & 1 & 1 + 3 & 0 & 1 & 0 & 1 + 4 & 0 & 1 & 1 &
+ 5 & 1 & 0 & 0 & + 6 & 1 & 0 & 1 & 1 + 7 & 1 & 1 & 0 & 1 + 8 & 1 & 1 & 1 & + as seen , even in the case of knowing the generated pseudo - random values , half the possible combinations lack a value , thus mitigating such an attack .
the method of encryption presented here could also be applied to the one time pad due to the likeness of them , so providing the extra bit of security by mitigating this most unlikely of scenarios .
the reasoning for the received original value presented in the above table is shown with simple boolean algebra .
the equations used are the ones presented in the pseudocode , also presented here for ease of following .
keep in mind that we do not know which one of the values we have and thus not what the other one could be .
[ [ case-1-and-2 ] ] * case 1 and 2 * + + + + + + + + + + + + + + here is @xmath12 and @xmath13 . while we know for sure that at least one of the values must be @xmath14 , the other could be either possible value .
this means that if our available result ( @xmath9 ) is @xmath14 , we have no way of knowing which value we have in our possession , which in turn means that @xmath1 and @xmath2 are equal in terms of possibility .
thus , since the original value could be either for @xmath1 , it remains undetermined .
if , on the other hand , our available value is @xmath15 , then we can logically conclude that it is @xmath2 that we are dealing with and we only have to complete the following equation : @xmath16 and therefore @xmath17 .
[ [ case-3-and-4 ] ] * case 3 and 4 * + + + + + + + + + + + + + + here is @xmath18 and @xmath19 . as per the same logic as in the previous one : if the available result is the same as the one we know must exist ( in this case one ) , then we can not tell which equation it is we have .
if it is the opposing value , then we can in this case rule out @xmath2 and calculate the equation : @xmath20 and hence @xmath17 .
[ [ case-5-and-6 ] ] * case 5 and 6 * + + + + + + + + + + + + + + here is @xmath21 and @xmath22 .
following the previous logical steps gives us an unknown for the available result of @xmath14 , but the following result for @xmath15 is @xmath17 .
[ [ case-7-and-8 ] ] * case 7 and 8 * + + + + + + + + + + + + + + here is @xmath23 and @xmath24 .
once again we can not say for sure which case is before us if our available result is the same as the one we know with certainty to exist , but in the case of a differing result , we know the following : @xmath25 and @xmath17 .
we would like to thank robert nyqvist for : | while strictly black and white images have been the basis for visual cryptography , there has been a lack of an easily implemented format for colour images .
this paper establishes a simple , yet secure way of implementing visual cryptography with colour , assuming a binary data representation . |
the concept of convex geometries was introduced by edelman @xcite and @xcite , see also edelman and jamison @xcite , adaricheva , gorbunov , and tumanov @xcite , and armstrong @xcite .
_ convex geometries _ are combinatorial structures : finite sets with anti - exchange closures such that the emptyset is closed .
they are equivalent to antimatroids , which are particular greedoids , and also to meet - distributive lattices . actually , the concept of convex geometries has many equivalent variants .
the first of these variants is due to dilworth @xcite , and the early ones were surveyed in monjardet @xcite .
since it would wander to far if we overviewed the rest , more than a dozen approaches , we only mention adaricheva @xcite , abels @xcite , caspard and monjardet @xcite , avann @xcite , jamison - waldner @xcite , and ward @xcite for additional sources , and stern @xcite , adaricheva and czdli @xcite , and czdli @xcite for some recent overviews . however , we need only a small part of the theory of convex geometries , and the present paper is intended to be self - contained for those who know the rudiments of lattice theory up to , say , the concept of semimodularity . from combinatorial point of view
, the finite convex geometries are the interesting ones . hence , and also because the tools we use are elaborated only for the finite case , the present paper is restricted to _ finite _ convex geometries . postponing the exact definition of convex geometries to section [ secketto ] , we present an important finite example as follows .
let @xmath1 , and let @xmath2 be a finite subset of the @xmath3-dimensional space @xmath4 .
the set of all subsets of @xmath2 is denoted by @xmath5 . for @xmath6 , we define @xmath7 , where @xmath8 denotes the usual convex hull of @xmath9 in @xmath4 .
map @xmath10 is a closure operator , and @xmath11 is a finite convex geometry .
convex geometries of this form are called _ geometries of relatively convex sets _ , and they ( not only the finite ones ) were studied by adaricheva @xcite and @xcite , bergman @xcite , and huhn @xcite .
we know from bergman @xcite that each finite convex geometry @xmath12 can be into a geometry @xmath11 of relative convex sets ; @xmath3 depends on @xmath12 .
however , even if a finite convex geometry is of convex dimension 2 , it is not necessarily isomorphic to some @xmath11 .
we start from a finite set @xmath0 of circles in the plane @xmath13 , and we define a convex geometry @xmath14 with the help of forming usual convex hulls in the plane , analogously to the geometries of relative convex sets .
the structures @xmath14 , called _ convex geometries of circles _ , are very close to the usual closure @xmath15 and , as opposed to geometries of relatively convex sets , we can prove that each finite convex geometry of convex dimension at most 2 is isomorphic to some convex geometry @xmath14 of circles . actually , our representation theorem , the main result of the paper , will assert more by imposing some conditions on @xmath0 ; see figure [ fig1 ] ( without @xmath16 , @xmath17 , the grey - colored plane shape @xmath18 , and the dotted curves ) for a first impression . until the proof of lemma [ fromtolemma ] ) [ fig1 ] ] this paper uses some lattices as auxiliary tools in proving the main result .
in fact , the theory of slim semimodular lattices has rapidly developed recently , as witnessed by czdli @xcite , @xcite , and @xcite , czdli , dkny , ozsvrt , szakcs , and udvari @xcite , czdli and grtzer @xcite , czdli , l. ozsvrt , and udvari @xcite , czdli and schmidt @xcite , @xcite , @xcite , and @xcite , grtzer and knapp @xcite , @xcite , and @xcite , and schmidt @xcite .
even if only a part of the results in these papers are eventually needed here , this progress provides the background of the present work . on the other hand , some auxiliary statements we prove here ,
namely some of the propositions , seem to be of some interest in the theory of slim semimodular lattices , while some other propositions could be interesting in the theory of convex geometries .
section [ secketto ] gives the basic concepts and formulates two results .
proposition [ prop1 ] asserts that , for every finite set @xmath0 of circles in the , @xmath14 is a convex geometry , while our main result , theorem [ thmmain ] , is the converse statement for the case of convex dimension @xmath19 .
the results of section [ secketto ] are proved in section [ proofsection ] , where several auxiliary statements are cited or proved , and some more concepts are recalled .
section [ oddsendch ] contains examples , statements , and open problems to indicate that although theorem [ thmmain ] gives a satisfactory representation of finite convex geometries of convex dimension at most 2 , we are far from settling the case of higher convex dimensions .
as mentioned already , the reader is only assumed a little knowledge of lattices .
besides the first few pages of any book on lattices or particular lattices , including grtzer @xcite , nation @xcite , and stern @xcite , even the first chapter of burris and sankappanavar @xcite , which does not even focus on lattice theory , is sufficient . note that @xcite and @xcite are freely downloadable
assume that we are given a set @xmath20 and a map @xmath21 .
if @xmath22 holds for all @xmath23 , then @xmath24 is a _ closure operator _ on @xmath20 . if @xmath24 is a closure operator on @xmath20 , @xmath25 , and @xmath24 satisfies the so - called _ anti - exchange property _
@xmath26 then @xmath27 is a _ convex geometry_. given a convex geometry @xmath27 , we use the notation @xmath28 to denote the _ set of closed sets _ of @xmath27 . actually , the structure @xmath29 is a _ lattice _ , and it is a complete meet - subsemilattice of the powerset lattice @xmath30 .
it is well - known that @xmath31 determines @xmath27 since we have @xmath32 .
hence , it is natural to say that @xmath27 can be _ embedded _ into or _
isomorphic _ to a convex geometry @xmath33 , if the lattice @xmath31 can be embedded into or isomorphic @xmath34 , respectively . as usual , a _
circle _ is a set @xmath35 , where @xmath36 and @xmath37 . a circle of radius 0 consists of a single point . since @xmath38 , the operator of forming convex hulls ,
is defined for subsets of @xmath13 rather than for sets of circles , we introduce a shorthand notation for `` _ _ p__oint__s _ _ of '' ( or `` _ _
p__oint _ _ s__et of '' ) as follows . for a set @xmath39 of circles in @xmath13 , the _ set of points _ belonging to some member of @xmath39
is denoted by @xmath40 . in other words , @xmath41 we define @xmath42 ( the superscript circle in the notation will remind us that @xmath43 is defined on a set of circles . )
the structure @xmath14 will be called the _ convex geometry _ of @xmath0 , and we call it a _ convex geometry of circles _ if @xmath0 is not specified .
( we shall soon prove that it is a convex geometry . )
note that if all circles of @xmath0 are of radius 0 and @xmath44 is the set of their centers , then @xmath14 is obviously isomorphic to @xmath45 .
we are now in the position to state our first observation ; the statements of this section will be proved in section [ proofsection ] .
[ prop1 ] for every finite set @xmath0 of circles in @xmath13 , @xmath14 is a convex geometry .
if there is a line containing the centers of all members of @xmath0 above , then @xmath0 is a _ set of collinear circles_. for simplicity , we will always assume that the line in question is the @xmath46 axis .
that is , in case of a set of collinear circles , all the centers are of the form @xmath47 .
for example , @xmath0 in figure [ fig1 ] is a set of collinear circles ; note that the dotted curves and the @xmath46-axis do not belong to @xmath0 .
the label of a circle in our figures is either below the center ( inside or outside but close to the circle ) , or we use an arrow .
the _ radius _ of a circle @xmath48 is denoted by @xmath49 .
if the center of @xmath48 is @xmath47 , then @xmath50 will denote the _ leftmost point _ and the _ rightmost point _ of @xmath48 , respectively .
since we allow that two distinct circles have the same leftmost point or the same rightmost point , we also need the following concept .
although circles are usually treated as endless figures , in case the center of a circle @xmath48 lies on the @xmath46 axis , we define the _ left end _ and the _ right end _ of @xmath48 as follows : @xmath51 left and right ends are ordered lexicographically ; this order is denoted by @xmath52 .
thus for later reference , note the obvious rules : @xmath53 let @xmath0 be a set of collinear circles .
we say that @xmath0 is a _ concave set of collinear circles _ if for all @xmath54 , @xmath55 for illustration , see @xmath56 , @xmath57 , and @xmath58 , or @xmath59 , @xmath60 , and @xmath16 , in figure [ fig1 ] .
note that since each @xmath61 is uniquely determined by @xmath62 and also by @xmath63 , `` @xmath52 '' in can be replaced by `` @xmath64 '' . if @xmath65 for any two distinct @xmath56 and @xmath57 in a set @xmath0 of collinear circles , then @xmath0 is called _
separated_. for example , @xmath66 in figure [ fig1 ] is a separated , concave set of collinear circles .
clearly , if @xmath0 is a separated set of collinear circles , then @xmath0 is concave iff @xmath67 for a finite lattice @xmath68 , the set of elements with exactly one lower cover , that is the set of non - zero join - irreducible elements , is denoted by @xmath69 .
dually , @xmath70 stands for the set of elements with exactly one cover .
the _ convex dimension _ of a finite convex geometry @xmath27 is the least integer @xmath3 such that @xmath71 is the union of @xmath3 chains . in other words ,
the convex dimension is the _ width _ of the poset @xmath71 .
we are now in the position of formulating our main result , which characterizes finite convex geometries of convex dimension at most 2 as the convex geometries of fini ( separated or not necessarily separated ) concave sets of collinear circles .
[ thmmain ] * if @xmath0 is a finite , concave set of collinear circles in the plane , then @xmath14 is a convex geometry of convex dimension at most 2 .
* for each finite convex geometry @xmath27 of convex dimension at most 2 , there exists a finite , separated , concave set @xmath0 of collinear circles in the plane such that @xmath27 is isomorphic to @xmath14 .
for every @xmath72 , we have @xmath73 since the `` @xmath74 '' inclusion follows from @xmath75 while the converse inclusion comes from the obvious @xmath76 .
[ fig2 ] ] assume that @xmath72 , and let @xmath77 . since @xmath78 holds for all @xmath79 , that is @xmath80 , we obtain @xmath81 we have equality here since @xmath82 .
this implies @xmath83 , and it follows that @xmath43 is a closure operator with the property @xmath84 .
observe that a closure operator @xmath24 on @xmath20 satisfies the anti - exchange property iff @xmath85 hence , tailoring to our situation , we assum @xmath86 , @xmath87 , and @xmath88 .
we have to show @xmath89 . combining our assumption with , we obtain @xmath90
let @xmath91 and @xmath92 be the boundary of @xmath93 and that of the set given in , respectively ; see the thick closed curves in figure [ fig2 ] . in the figure , @xmath39 is depicted twice : in itself on the left and with @xmath94 on the right .
we can imagine @xmath91 as resilient rubber noose around the members of @xmath39 .
pick an @xmath95 .
observe that @xmath96 implies @xmath97 .
for a set @xmath0 of collinear circles and @xmath98 , we define the _ horizontal interval _
@xmath99 note that @xmath100 can be empty .
if @xmath101 and @xmath102 is inside the circle @xmath103 , then @xmath104 but @xmath105 .
note also that , for @xmath106 , the horizontal interval @xmath107 is the _ set of circles _ of @xmath0 that are _ inside _ @xmath102 , including @xmath102 itself .
[ fromtolemma ] if @xmath0 is a finite set of collinear circles in the plane , then the following three statements hold .
[ fromtolemmaa ] @xmath108 . if @xmath0 is concave , then even the equality @xmath109 holds .
[ fromtolemmab ] for @xmath110 , let @xmath102 and @xmath103 denote the circles in @xmath39 with least left end and greatest right end , respectively @xmath0 is concavehen @xmath111 holds in @xmath112 .
[ fromtolemmac ] if @xmath113 , then @xmath114 .
we can assume @xmath115 since otherwise the lemma is trivial .
assume @xmath116 .
clearly , as indicated in figure [ fig1 ] by the grey - colored plane shape , there exists a convex subset @xmath18 of @xmath13 such that @xmath117 but @xmath18 includes no @xmath118 as a subset .
( as in the figure , it is always possible to find an appropriate @xmath18 whose boundary is the union of an arc of @xmath102 , that of @xmath103 , and four straight line segments . note that if @xmath119 , then the only arc of @xmath102 is itself . )
this implies that @xmath120 .
thus @xmath121 .
to prove the converse inclusion under the additional assumption that @xmath0 is concave , and let @xmath122 . since @xmath0 is finite , there are a unique @xmath123 with least left end @xmath124 and a unique @xmath125 with largest right end @xmath126 . using that @xmath0 is concave
, we obtain that , for every @xmath127 , @xmath128 , that is , @xmath129 .
this implies @xmath130 . on the other hand , if @xmath131 , then the choice of @xmath102 and @xmath103 gives @xmath132 and @xmath133 , that is , @xmath127 .
hence , @xmath134 , and we obtain @xmath135 .
this proves and the first equality in .
next , we do not assume that @xmath0 is concave .
if @xmath136 , then @xmath137 holds for all @xmath138 .
that is , @xmath139 for all @xmath138 .
thus , for all @xmath131 , we have @xmath140 , which clearly implies
. finally , assume again that @xmath0 is concave . from and the first equality in ,
we obtain @xmath141 .
reverse inclusion follows from the assumption that @xmath0 is concave .
hence , holds .
[ mzrrdcgbn ] if @xmath0 is a finite set of collinear circles in the plane , then @xmath142 although the statement is intuitively more or less clear , we give an exact proof .
first , we show that , for @xmath106 , @xmath143 .
this is obvious if @xmath144 , that is , if @xmath107 is an atom in @xmath112 .
assume @xmath145 .
let @xmath103 and @xmath48 be the circles with least left end and greatest right end in @xmath146 , respectively .
( they need not be distinct . ) since the circles of @xmath0 are determined by their left ends , @xmath147 , and we obtain @xmath148 by the definition of @xmath107 . similarly , @xmath149 .
now if @xmath150 , then the choice of @xmath103 and @xmath48 gives @xmath151 . on the other hand , if @xmath151 , then @xmath152 yield @xmath150 .
thus @xmath153 is a unique lower cover of @xmath107 , because @xmath154 implies @xmath155 for all @xmath156 .
hence , @xmath143 .
this proves @xmath157 . to prove the converse inclusion ,
let @xmath158 .
we obtain from lemma [ fromtolemma ] that @xmath159 for some @xmath160
. thus @xmath161 .
[ mirlemma ] assume that @xmath0 is a finite , concave set of collinear circles in the plane , and @xmath162 .
the circle in @xmath0 with least left end and that with largest right end denoted by @xmath163 and @xmath164 , respectively @xmath165 are chains in @xmath112 , and each nonempty @xmath113 is of the form @xmath166 where @xmath167 are defined in lemma [ fromtolemma ] .
let @xmath168 .
we can assume @xmath169 since `` @xmath64 '' is a linear order .
by transitivity , @xmath170 .
therefore , the first set in the lemma is a chain . by left - right duality
, so is the second one .
next , let @xmath122 .
@xmath134 by lemma [ fromtolemma ] , and the obvious @xmath171 completes the proof .
one of the many equivalent ways , actually the first way , of defining convex geometries was to use meet - distributive lattices ; see dilworth @xcite .
now we recall some concepts from lattice theory .
a finite lattice @xmath68 is _ lower semimodular _ if whenever @xmath172 such that @xmath173 is covered by @xmath174 , in notation @xmath175 , then @xmath176 .
equivalently , if the implication @xmath177 holds for all @xmath178
. we will often use the trivial fact that this property is inherited by intervals and , more generally , by cover - preserving sublattices .
for @xmath179 in a finite lattice @xmath68 , let @xmath180 denote the meet of all lower covers of @xmath181 .
a finite lattice @xmath68 is _ meet - distributive _ if the interval @xmath182 $ ] is a distributive lattice for all @xmath183 . for other definitions ,
see adaricheva @xcite , adaricheva , gorbunov and tumanov @xcite , and caspard and monjardet @xcite ; see also czdli ( * ? ?
* proposition 2.1 and remark 2.2 ) and adaricheva and czdli @xcite for recent surveys and developments .
the study of meet - distributive lattices ( and their duals ) goes back to dilworth @xcite , 1940 .
there were a lot of discoveries and rediscoveries of these lattices and equivalent combinatorial structures ( including convex geometries ) ; see @xcite and @xcite , monjardet @xcite , and stern @xcite for surveys .
we recall the following statement ; its origin is the combination of ward @xcite ( see also dilworth @xcite , where @xcite is cited ) and avann @xcite ( see also edelman ( * ? ? ?
* theorem 1.1(e , h ) ) , where @xcite is recalled ) .
[ mdstrthnlsm ] every finite meet - distributive lattice is lower semimodular .
the _ width _ of a partially ordered set @xmath184 , denoted by @xmath185 , is the smallest @xmath186 such that @xmath184 is the union of @xmath186 appropriate chains of @xmath68 .
equivalently , see dilworth @xcite , @xmath185 is the largest @xmath186 such that there is a @xmath186-element antichain in @xmath184 . for a finite lattice @xmath68 ,
we are interested in the width of @xmath70 .
note that , clearly , @xmath187 is the smallest @xmath186 such that the union of @xmath186 maximal chains of @xmath68 includes @xmath70 .
following grtzer and knapp @xcite and , in the present form , czdli and schmidt @xcite , finite lattices @xmath68 with @xmath188 are called _ dually slim_. finite lattices @xmath68 with @xmath189 are , of course , called _ slim_. if @xmath68 is a lattice and @xmath190 , then the _ principal ideal _
@xmath191 is denoted by @xmath192 .
we have already mentioned that a finite convex geometry @xmath193 determines a lattice , the lattice @xmath194 of its closed sets .
conversely , if @xmath68 is a finite meet - distributive lattice , then we can take the combinatorial structure @xmath195 .
part of the following lemma , which asserts that finite convex geometries and finite meet - distributive lattices are essentially the same , was proved by edelman ( * ? ? ?
* theorem 3.3 ) , see also armstrong ( * ? ? ? * theorem 2.8 ) .
the rest can be extracted from adaricheva , gorbunov , and tumanov ( * ? ? ?
* proof of theorem 1.9 ) ; see also czdli ( * ? ? ?
* lemma 7.4 ) for more details .
[ geomversuslat ] if @xmath68 is a finite meet - distributive lattice and @xmath193 is a finite convex geometry , then the following three statements hold .
[ latamateqa ] @xmath196 is a finite meet - distributive lattice .
[ latamateqb ] @xmath197 is a finite convex geometry .
[ latamateqc ] @xmath198 and @xmath199 .
finite , slim , semimodular lattices are more or less understood .
therefore , so are their duals , the dually slim , lower semimodular , finite lattices . the following lemma is practically known , but we will explain how to extract it from the literature . the notation introduced before lemma [ geomversuslat ] is still in effect .
[ slimgeomlemma ] @xmath24 a closure operator on @xmath20hen @xmath193 is a finite convex geometry of convex dimension at most 2 iff@xmath200 @xmath196 is a finite , dually slim , lower semimodular lattice . in view of lemma [ geomversuslat ] ,
all we have to show is that , for a finite lattice @xmath68 , the following two conditions are equivalent : [ sdszdfa ] @xmath68 is meet - distributive and @xmath188 ; [ sdszdfb ] @xmath68 is lower semimodular and dually slim . with reference to ward @xcite and avann @xcite , we have already mentioned that meet - distributivity implies lower semimodularity .
thus implies .
conversely , assume .
@xmath68 is meet - distributive by the dual of czdli , ozsvrt , and udvari ( * ? ? ?
* corollary 2.2 ) , and @xmath188 by the definition of dual slimness . thus implies .
next , motivated by lemma [ slimgeomlemma ] , we will have a closer look at finite , dually slim , lower semimodular lattices .
a finite lattice is _ planar _ if it has a planar diagram in the obvious sense ; for more details see the next subsection .
planarity is a great help for us , because of part ( b ) of the following lemma .
a cover - preserving @xmath201 sublattice is a 5-element sublattice @xmath202 such that @xmath203 for @xmath204 .
[ slimplanar ] if @xmath68 is a finite , lower semimodular lattice , then the following two statements hold .
* @xmath68 is dually slim iff it has no cover - preserving @xmath201 sublattice . * if @xmath68 is dually slim , then it @xmath68 is planar , and each of its elements has at most two lower covers .
following libkin @xcite , a finite lattice @xmath205 is said to satisfy _
carathodory s condition _
n if for any @xmath206 such that @xmath207 , there are @xmath208 such that @xmath209 .
carathodory s classical theorem asserts that whenever @xmath210 is a point and @xmath39 is a subset of @xmath211 such that @xmath210 belongs to the convex hull @xmath212 of @xmath39 , then there exists an at most @xmath3-element subset @xmath9 of @xmath39 such that @xmath213 .
this implies that the lattices @xmath214 satisfy n since their join - irreducible elements are exactly the atoms .
we say that a _ finite convex geometry @xmath27 satisfies n _ if so does the corresponding lattice , @xmath31 .
lower semimodularity is not assumed in the following statement .
[ cpropos ] every finite , dually slim lattice satisfies 2 . also , every finite convex geometry of convex dimension at most 2 satisfies 2 .
the lattices in question are planar by , and planar lattices satisfy 2 by libkin ( * ? ? ?
* corollary 4.7 and theorem 3 ) .
the rest follows from lemma [ slimgeomlemma ] .
[ jirmaprop ] @xmath215 can be extended to an isomorphism from @xmath216 onto @xmath217 .
next , let @xmath218 denote the inverse of @xmath215 .
we define a map @xmath219 by @xmath220 .
similarly , let @xmath221 be defined by @xmath222 .
the choice in shows that @xmath215 and @xmath223 are order isomorphism .
this implies that @xmath224 and @xmath225 are extensions of @xmath215 and @xmath223 , respectively .
[ latisorem ] if @xmath216 and @xmath217 are finite lattices and @xmath226 is a bijection satisfying , then @xmath227 and @xmath215 extends to an isomorphism @xmath228 .
even if @xmath215 in proposition [ jirmaprop ] is an order - isomorphism , may fail ; its satisfaction depends mainly on the case where @xmath229 is an antichain .
this is one of the reasons why we are going to have a closer look at dually slim , lower semimodular lattices .
since dually slim lattices are planar by lemma [ slimplanar](b ) , the propositions of this subsection may look intuitively clear .
however , their exact proofs need some preparation . fortunately , the theory of planar lattices is satisfactorily developed in kelly and rival @xcite at a rigorous level , so we can often rely on results from @xcite instead of going into painful rigorousity .
whenever we deal with a planar lattice , always a fixed planar diagram is assumed .
actually , most of the concepts , like left and right , depend on the planar diagram chosen ( sometimes implicitly ) at the beginning rather than on the lattice .
this will not cause trouble since our arguments do not depend on the choice of planar diagrams .
now , we recall some necessary concepts and statements for planar lattices ; the reader may ( but need not ) look into @xcite for more exact details .
let @xmath48 be a maximal chain in a finite planar lattice @xmath68 ( with a fixed planar diagram ) .
this chain cuts @xmath68 into a _ left side _ and a _ right side _
, see kelly and rival ( * ? ? ?
* lemma 1.2 ) .
the intersection of these sides is @xmath48 .
if @xmath190 is on the left side of @xmath48 but not in @xmath48 , then @xmath46 is _ strictly on the left _ of @xmath48 .
let @xmath17 be another maximal chain of @xmath68 .
if all elements of @xmath17 are on the left of @xmath48 , then @xmath17 is _ on the left of _ @xmath48 . in this sense
, we can speak of the leftmost maximal chain of @xmath68 , called the _ left boundary chain _ , and the rightmost maximal chain , called the _ right boundary chain_. the union of these two chains is the _ boundary _ of @xmath68 .
also , if @xmath2 is a ( not necessarily maximal ) chain of @xmath68 , then the _ leftmost maximal chain through _
@xmath2 ( or extending @xmath2 ) and the rightmost one make sense .
if @xmath230 , then the leftmost maximal chain of @xmath68 through @xmath2 is the union of the left boundary chains of the intervals @xmath231}}}}$ ] , @xmath232 $ ] ,
, @xmath233 $ ] , and @xmath234 $ ] .
( the diagrams of these intervals are the respective subdiagrams of the fixed diagram of @xmath68 . )
if @xmath235 is a singleton , then chains containing @xmath236 are said to be chains through @xmath236 rather than chains through @xmath237 .
the most frequently used result of kelly and rival @xcite is probably the following one .
[ krchainlemma ] let @xmath68 be a finite planar lattice , and let @xmath238 .
if @xmath46 and @xmath239 are on different sides of a maximal chain @xmath48 in @xmath68 , then there exists an element @xmath240 such that @xmath241 .
next , let @xmath46 and @xmath239 be elements of a finite planar lattice @xmath68 , and assume that they are incomparable , in notation , @xmath242 .
if @xmath243 has lower covers @xmath244 and @xmath245 such that @xmath246 , @xmath247 , and @xmath244 is on the left of @xmath245 , then the element @xmath46 is _ on the left _ of the element @xmath239 .
if @xmath46 is on the left of @xmath239 , then we say that @xmath239 is _ on the right _ of @xmath46 .
let us emphasize that whenever left or right is used for two elements , then the elements in question are incomparable .
[ leftrightlemma ] let @xmath68 be finite planar lattice @xmath248 incomparable elementshen the following hold .
* @xmath46 is on the left of @xmath239 iff @xmath46 is on the left of some maximal chain through @xmath239 iff @xmath46 is on the left of all maximal chains through @xmath239 . *
either @xmath46 is on the left of @xmath239 , or @xmath46 is on the right of @xmath239 . *
if @xmath249 , @xmath242 , @xmath250 , @xmath46 is on the left of @xmath239 , and @xmath239 is on the left of @xmath251 , then @xmath46 is on the left of @xmath251 .
if @xmath252 is a 3-element antichain such that @xmath253 is on the left of @xmath239 and @xmath239 is on the left of @xmath254 for some ( necessarily unique ) @xmath255 , then @xmath239 is _ between _ the elements @xmath256 and @xmath244 .
[ propokoztes ] let @xmath68 be a finite lattice .
@xmath252 is a 3-element antichain in @xmath68 the following two statements hold . *
[ lemmabetweenjoina ] if @xmath68 is planar and @xmath239 is between @xmath256 and @xmath244 , then @xmath257 . * [ lemmabetweenjoinb ] if @xmath68 is a dually slim , lower semimodular lattice and @xmath257 , then @xmath239 is between @xmath256 and @xmath244 .
note that it would be unreasonable to tailor the same condition on @xmath68 in parts ( a ) and ( b ) .
the 5-element , modular , non - distributive lattice @xmath201 indicates that planarity in itself would not be sufficient in part ( b ) . on the other hand , although dual slimness ( with or without lower semimodularity ) would be sufficient in part ( a ) by lemma [ slimplanar ] , in this case the statement would be weaker and we could not use the dual of part ( a ) in the proof of part ( b ) . to prove ( a ) ,
pick a maximal chain @xmath48 through @xmath239 . without loss of generality
, we can assume that @xmath256 and @xmath258 are on the left and @xmath244 is on the right of @xmath48 .
applying lemma [ krchainlemma ] , there exists an element @xmath240 such that @xmath259 . belonging to the same chain , @xmath239 and @xmath251 are comparable .
since @xmath260 would contradict @xmath261 , we have @xmath262 , proving ( a ) . next , to prove ( b ) by contradiction , suppose that @xmath68 is a dually slim , @xmath257 , but @xmath239 is not between @xmath256 and @xmath244 .
let , say , @xmath256 be on the left of @xmath244 and @xmath244 be on the left of @xmath239 .
the interval @xmath263 $ ] contains @xmath256 , @xmath244 and @xmath239 since @xmath257 and , by the dual of part ( a ) , @xmath264 . as an interval of @xmath68
, it is lower semimodular , and it follows from lemma [ slimplanar](a ) that this interval is dually slim .
there are two cases .
first , assume @xmath265 .
let @xmath2 be a maximal chain of @xmath266 through @xmath267 .
by lemma [ leftrightlemma](a ) , @xmath256 is on the left of @xmath2 and @xmath239 is on the right of @xmath2 .
note that the left side of @xmath2 ( including @xmath2 itself ) is a cover - preserving sublattice by kelly and rival ( * ? ? ?
* proposition 1.4 ) ; this can also be derived from lemma [ krchainlemma ] easily .
hence , we can pick lower covers @xmath268 , @xmath269 of @xmath258 on the left of @xmath2 such that @xmath270 and @xmath271 .
similarly , let @xmath272 be a lower cover of @xmath258 such that @xmath273 and @xmath274 is on the right of @xmath2 .
if we had @xmath275 , then @xmath2 would contain an element @xmath251 by lemma [ krchainlemma ] , necessarily strictly above @xmath244 since @xmath276 , such that @xmath277 , which would lead to the following contradiction : @xmath278 hence , @xmath279 .
if we had @xmath280 or @xmath281 , then @xmath282 or @xmath283 , respectively , would be a contradiction .
hence @xmath268 , @xmath269 and @xmath274 are three distinct lower covers of @xmath258 , which contradicts lemma [ slimplanar](b ) .
second , assume @xmath284 .
since these elements are in @xmath266 , we have @xmath285 .
take a maximal chain @xmath12 of @xmath266 through @xmath286 , and let @xmath287 be the left side of @xmath12 . by lemma [ leftrightlemma](a ) , @xmath288 . in @xmath287 , take a maximal chain @xmath0 through @xmath256 , and let @xmath205 be the right side of @xmath0 in @xmath287 . by lemma
[ leftrightlemma](a ) again , @xmath289 .
@xmath205 is a cover - preserving sublattice of @xmath266 , again by kelly and rival ( * ? ? ?
* propositions 1.4 ) .
now @xmath0 and @xmath12 are the left and right boundary chains of @xmath205 , respectively .
like we obtained it for @xmath266 , we conclude that @xmath205 , which is a cover - preserving sublattice of @xmath266 ( and also of @xmath68 ) , is a dually slim , lower semimodular lattice . by czdli and schmidt (
* lemma 6 ) , @xmath290 . therefore , there are @xmath291 and @xmath292 such that @xmath293 .
since @xmath294 is excluded by @xmath295 and @xmath0 is a chain , @xmath296 .
similarly , using that @xmath12 is a chain and @xmath297 , we obtain @xmath298 .
thus @xmath299 , and we conclude @xmath300 . clearly , @xmath301 since otherwise @xmath302 would contradict @xmath303 .
hence @xmath304 , which leads to the contradiction @xmath305 .
we also need the following statement .
[ prszdobzf ] if @xmath68 a finite lower semimodular lattice , @xmath306 , @xmath307 , @xmath308 , and @xmath309 , then @xmath310 . to prove the statement by contradiction ,
suppose that in spite of the assumptions , @xmath311 .
if we had @xmath312 , then @xmath313 together with @xmath309 would contradict @xmath314 .
hence @xmath315 .
let @xmath316 ; we know that @xmath317
. since @xmath318 together with @xmath319 would contradict @xmath320 , we have @xmath321 . hence , @xmath322 and we can pick an element @xmath323 $ ] such that @xmath324 . denoting the unique lower cover of @xmath173 by @xmath325
, we conclude @xmath326 because of the following reasons : @xmath327 yields @xmath328 ; @xmath329 or @xmath330 would contradict @xmath315 ; and @xmath331 or @xmath332 would lead to the contradiction @xmath333 or @xmath334 . since @xmath335 , we have @xmath336 . lower semimodularity yields @xmath337 , and we obtain @xmath338 . since @xmath339 and @xmath340 , the elements @xmath325 and @xmath341 are two distinct lower covers of @xmath173 .
this contradicts @xmath314 .
the elements of @xmath342 are called _ doubly irreducible elements_. a principal filter @xmath343 of @xmath68 is a _ prime filter _ if @xmath344 is closed with respect to joins or , equivalently , if @xmath345 is a lattice ideal of @xmath68 .
[ dirronbtop ] let @xmath68 be a finite , dually slim , lower semimodular lattice @xmath346 , then the following statements hold .
[ dirronbtopa ] @xmath68 has a maximal doubly irreducible element @xmath347 , and @xmath347 belongs to the boundary of @xmath68 .
[ dirronbtopb ] if @xmath190 and @xmath348 , then @xmath349 but @xmath350 .
furthermore , @xmath343 is a chain .
[ dirronbtopc ] @xmath343 is a prime filter of @xmath68 . by lemma [ slimplanar ] ( b ) , @xmath68 is planar .
we know from kelly and rival ( * ? ? ?
* theorem 2.5 ) that each finite planar lattice @xmath68 with at least three elements has a doubly irreducible element on its boundary .
however , we only use this theorem to conclude @xmath351 .
hence , @xmath342 contains a maximal element , @xmath347 . since @xmath70 is a subset of the boundary by the dual of czdli and schmidt ( * ? ? ? * lemma 6 ) , @xmath347 belongs to the boundary .
this proves .
let , say , @xmath347 belong to the right boundary chain @xmath352 of @xmath68 .
the dual of czdli ( * ? ?
? * lemma 2.3 ) asserts @xmath353 .
we claim that @xmath354 . to prove this by contradiction ,
suppose @xmath355 .
@xmath343 is not a chain since @xmath356 is a maximal chain in @xmath343 .
hence , there are @xmath357 such that @xmath358 , and there are chains @xmath359 and @xmath360 in @xmath343 . clearly , @xmath361 .
let @xmath362 be the largest subscript such that @xmath363 , @xmath364 , and @xmath365 .
this @xmath362 exists since @xmath366 .
since @xmath358 , we have @xmath367 and @xmath368 . there are two cases .
first , if @xmath369 , then @xmath370 has at least two distinct covers .
second , if @xmath371 , then @xmath370 has at least two distinct covers again : @xmath372 and a cover belonging to @xmath352 . hence , in both cases
, @xmath370 is meet - reducible and belongs to @xmath352 , which contradicts @xmath353 .
therefore , @xmath354 and , since subsets of chains are chains , @xmath343 is a chain .
now , assume @xmath348 .
since @xmath373 and @xmath347 was a maximal doubly irreducible element , we conclude @xmath374 .
this proves . to prove by contradiction , suppose @xmath375 such that there exist elements @xmath376 with @xmath377 .
@xmath378 and @xmath379 for some @xmath380 .
it follows from proposition [ cpropos ] that there are two elements in @xmath381 whose join belongs to @xmath343
. therefore , there are incomparable elements @xmath382 such that @xmath383 .
there are two cases .
first , assume that @xmath384 is an antichain .
none of @xmath210 and @xmath385 belongs to @xmath352 , and lemma [ leftrightlemma](a ) yields that both @xmath210 and @xmath385 are on the left of @xmath347 .
this is a contradiction since proposition [ propokoztes](b ) implies that @xmath347 is between @xmath210 and @xmath385 .
second , assume that @xmath384 is not an antichain .
since @xmath347 is join - irreducible and @xmath386 , we can not have @xmath387 . hence , apart from @xmath210-@xmath385 symmetry , @xmath388 and @xmath389 .
however , proposition [ prszdobzf ] contradicts @xmath383 . before formulating the last auxiliary statement towards theorem [ thmmain ]
, remember that dual slimness implies planarity by lemma [ slimplanar](b ) .
[ lmlstm ] let @xmath68 be a finite , dually slim , semimodular lattice with a fixed planar diagram , and let @xmath0 be a finite concave set of collinear circles in the plane .
assume that we have a bijective map @xmath390 such that for any @xmath391 , [ lmlstma ] @xmath392 if and only if @xmath393 , and [ lmlstmb ] @xmath358 and @xmath181 is on the left of @xmath394 if and only if @xmath395 and @xmath396 .
@xmath397 .
lemma [ mzrrdcgbn ] allows us to define a map @xmath398 by @xmath399 .
clearly , @xmath400 therefore , by assumption , @xmath215 is an order - isomorphism .
since we want to apply propositions [ jirmaprop ] , we are going to show that @xmath215 satisfies condition . if @xmath310 or @xmath401 , then @xmath402 or @xmath403 since @xmath215 is an order - isomorphism thus , if @xmath404 is not an antichain , we have @xmath405 next , assume that @xmath404 is an antichain .
so is @xmath406 since @xmath215 is an order - isomorphism .
there are two cases : either @xmath173 is between @xmath347 and @xmath407 , or not . in the first case , we can assume that @xmath347 is on the left of @xmath173 and @xmath173 is on the left of @xmath407 .
proposition [ propokoztes ] gives @xmath309 , and yields @xmath408 since @xmath0 is concave , implies @xmath409 , which gives @xmath410 .
hence , holds in this case . in the second case , where @xmath173 is not between @xmath347 and @xmath407
, we can assume that @xmath173 is on the left of @xmath347 and @xmath347 is on the left of @xmath407 . by proposition [ propokoztes ] , we have @xmath411 , and gives @xmath412 let @xmath413 .
@xmath414 by lemma [ fromtolemma ] , and @xmath415 by the definition of horizontal intervals . since @xmath416 , we conclude @xmath417 , and holds in this case . next ,
assume @xmath418 and @xmath419 , and let @xmath325 stand for the unique lower cover of @xmath173 .
@xmath420 , and we have @xmath411 . since @xmath223 is an order - isomorphism , @xmath421 .
this gives @xmath422 , and is fulfilled again .
finally , up to @xmath347-@xmath407 symmetry , we are left with the case where @xmath315 and @xmath308 .
we can assume @xmath321 since otherwise both sides of would obviously be false and would hold . take a maximal chain @xmath48 including @xmath423 ; it does not contain @xmath347 .
let , say , @xmath347 be on the left of @xmath48 .
now , by lemma [ leftrightlemma ] , @xmath347 is on the left of @xmath407 and also on the left of @xmath173 . since @xmath424 , we conclude @xmath425 .
this and yield @xmath426 as previously , this gives @xmath427 and @xmath428 , which implies @xmath429 . since @xmath411 by proposition [ prszdobzf ] , is satisfied again .
since holds in all cases , @xmath215 satisfies .
thus propositions [ jirmaprop ] appl .
let @xmath0 be a finite , concave set of collinear circles in the plane .
proposition [ prop1 ] yields that @xmath14 is a convex geometry .
hence , by lemma [ geomversuslat ] , @xmath112 is a finite meet - distributive lattice , and it is lower semimodular by claim [ mdstrthnlsm ] .
we obtain from lemma [ mirlemma ] that this lattice is dually slim .
therefore , lemma [ slimgeomlemma ] implies that @xmath14 is a convex geometry of convex dimension at most 2 .
this proves part ( a ) . in view of lemmas
[ geomversuslat ] and [ slimgeomlemma ] , part ( b ) is equivalent to the following statement : if @xmath68 is a finite , dually slim , lower semimodular lattice , then there exists a finite , separated , concave set @xmath0 of collinear circles in the plane such that @xmath68 is isomorphic to @xmath112 .
we prove ( c ) by induction . by lemma [ lmlstm ] , it suffices to construct a pair @xmath430 such that @xmath0 is a finite , separated , concave se collinear circles in the plane and @xmath390 is a bijective map satisfying conditions and of lemma [ lmlstm ] .
since we are going to construct a separated @xmath0 , allows us to satisfy these two conditions with leftmost and rightmost endpoints rather left and right ends .
first , assume that @xmath68 is a chain or , equivalently , @xmath69 is a chain .
we can let @xmath0 and @xmath223 be a set of concentric circles and the unique map satisfying condition of lemma [ lmlstm ] , respectively ; clearly , @xmath430 is an appropriate pair . more generally , not assuming that @xmath68 is a chain
, we can prove the existence of an appropriate pair @xmath430 by induction on the size @xmath431 of @xmath68 .
since the case of chains has been settled , the induction starts at size 4 . and the corresponding separated concave set @xmath0 of collinear circles[fig3 ] ]
assume that @xmath432 and for each finite , dually slim , lower semimodular lattice of smaller size , there exists and appropriate pair .
take a maximal doubly irreducible element @xmath433 proposition [ dirronbtop ] ) . by left - right symmetry
, we can assume that @xmath407 is on the right boundary chain of @xmath68 ; see figure [ fig3 ] , on the left , for illustration .
( the figure serves only as an illustration , so the reader need not check the properties of @xmath68 .
however , we note that @xmath68 is obviously a dually slim , lower semimodular lattice by the dual of czdli and schmidt ( * ? ? ?
* theorem 12 ) . )
since @xmath434 is a prime filter , @xmath435 is ideal of @xmath68 . as an interval of @xmath68 , @xmath436 is lower semimodular .
it follows from lemma [ slimplanar](a ) that @xmath436 is dually slim .
hence , by the induction hypothesis , there exists an appropriate pair @xmath437 for @xmath436 .
we want to define @xmath430 such that @xmath438 and @xmath223 be an extension of @xmath439 . by proposition [ dirronbtop ] , @xmath440 . therefore , our purpose is to find and appropriate circle @xmath48 and to let @xmath223 be the map from @xmath69 to @xmath441 defined by @xmath442 and @xmath443 for @xmath444 . in the figure ,
@xmath445 and @xmath446 is the set of black - filled elements . on the right of the figure
, we write @xmath48 , @xmath447 , and @xmath448 instead of @xmath449 , @xmath450 , and @xmath451 , respectively ; @xmath0 is the collection of all circles , and @xmath452 . to make @xmath0
concave [ fig4 ] ] since @xmath407 is on the right boundary chain , lemma [ leftrightlemma](a ) yields that each @xmath453 is either strictly on the left of @xmath407 or comparable to @xmath407 .
this together with proposition [ dirronbtop ] imply that @xmath454 is the disjoint union of the following two sets : @xmath455 note that one of these two sets can be empty but their union , @xmath454 , is nonempty . in the figure , @xmath456 and @xmath457 .
condition of lemma [ lmlstm ] will hold iff @xmath458 similarly , condition of lemma [ lmlstm ] will hold iff @xmath459 the only stipulation that and tailor on @xmath460 is @xmath461 this can be satisfied easily . therefore , to see that we can choose @xmath462 such that and hold , it suffices to show that @xmath463 for @xmath464 and @xmath465 , there are two cases .
first , assume @xmath466 .
clearly , @xmath467 , thus @xmath468 .
since condition of lemma [ lmlstm ] holds for @xmath437 , we have @xmath469 and @xmath470 .
this yields @xmath471 since @xmath472 is separated .
second , assume @xmath242 .
let @xmath2 be a maximal chain in @xmath68 that extends @xmath473 . since @xmath239 is on the left of @xmath407
, @xmath239 is on the left of @xmath2 by lemma [ leftrightlemma](a ) .
hence , again by lemma [ leftrightlemma](a ) , @xmath239 is on the left of @xmath46 .
therefore , using condition of lemma [ lmlstm ] for the appropriate pair @xmath437 , we conclude @xmath474 .
this implies @xmath471 since @xmath0 is separated
. thus holds , and so do and . since and are strict inequalities , we can choose both @xmath462 and @xmath460 infinitely many ways
therefore , we can choose @xmath48 so that @xmath0 be separated .
finally , we have to show that @xmath0 is concave . since @xmath472 is concave and separated ,
the only case we have to consider is @xmath475 where @xmath476 ; have to show @xmath477 suppose that after choosing @xmath48 , fails for some @xmath476 , see figure [ fig4 ] .
@xmath48 is in the interior of the region between the two common tangent lines @xmath478 and @xmath479 of @xmath56 and @xmath57 .
let @xmath480 be the tangent line of @xmath48 through @xmath481 , and let @xmath482 be the intersection point of @xmath483 and @xmath480 for @xmath484 .
if @xmath460 tends to infinity while @xmath462 is unchanged , then the arc of @xmath48 between @xmath478 and @xmath479 with middle point @xmath481 approaches the line segment @xmath485 .
therefore , replacing @xmath48 by @xmath486 such that @xmath487 and @xmath488 is sufficiently large , we have @xmath489 and .
we can treat all pairs @xmath490 with @xmath491 , each after each , because @xmath460 can always be enlarged .
this proves that @xmath0 is concave for some @xmath48 .
it is not hard to see that theorem [ thmmain ] remains valid if we consider closed discs or open discs instead of circles , and modify the definitions accordingly .
the advantage of circles is that they are easier to visualize and label in figures .
open discs are particularly less pleasant than circles since they can not be singletons .
note that we can not use semicircles or half discs since , by the following example , the corresponding structure is not a convex geometry in general .
let @xmath492 rotating @xmath493 around @xmath494 by angle @xmath495 , we obtain a half circle @xmath496 . since @xmath497 @xmath498 for @xmath499 , the anti - exchange property fails ,
and we do not obtain a convex geometry from @xmath500 .
[ krkrntnk ] let @xmath0 be an @xmath3-element set of collinear circles .
assume that there is an additional circle @xmath205 such that every circle @xmath61 is internally tangent to @xmath205 ; see figure [ fig5 ] for @xmath501 , where the dotted curve is an arc of @xmath205 .
@xmath502 is the @xmath503-element boolean lattice , and the convex dimension of @xmath112 is @xmath3 .
[ cvmdfpnts ] for each natural number @xmath3 , there exists an @xmath505-element set @xmath0 of circles in the such that @xmath14 does not satisfy n. for example , we can take the inscribed circle of a regular @xmath506-gon and @xmath507 additional little circles whose centers are the vertices of the @xmath506-gon ; see figure [ fig6 ] for @xmath508 . in view of lemma [ mzrrdcgbn ]
, we have to show the following : if @xmath510 such that @xmath511 , then there exist @xmath512 such that @xmath513 .
let @xmath12 be the boundary of @xmath514 ; see the thick closed curve in figure [ fig7 ] , where @xmath515 and @xmath0 contains the solid circles and possibly some other circles not indicated .
( the dotted circle need not belong to @xmath0 . ) clearly , @xmath12 can be divided into circular arcs and straight line segments of common tangent lines of some circles belonging to @xmath516 ; these parts are separated by black - filled points in the figure . keeping its center fixed , we enlarge @xmath48 to @xmath486 such that @xmath517 and @xmath486 is internally tangent to @xmath12 at a point @xmath518 .
there are two cases . in the first case ,
we assume that @xmath519 belongs to a circular arc of @xmath12 . in the figure , @xmath519 in this case is not indicated ; it can be any point of the closed circular arc @xmath520 .
this arc is also an arc of some member , @xmath521 , of @xmath0 .
clearly , @xmath522 .
thus @xmath521 and @xmath48 are concentric circles , and @xmath523 .
hence , we can let @xmath524 and @xmath525 .
observe that circular arcs of length 0 cause no problem since then the radius of @xmath521 is zero , the center of @xmath521 belongs to @xmath12 , and @xmath526 . in the second case , we assume that @xmath519 belongs to a straight line segment @xmath527 of @xmath12 . now @xmath486 is the dotted circle in the figure , and its center is @xmath20 .
we can assume that @xmath528 since otherwise the previous case applies . clearly , @xmath184 is a point of a unique @xmath529 with center @xmath530 , and @xmath531 is on a unique @xmath532 with center @xmath533 .
since the radii @xmath534 , @xmath535 and @xmath536 are all perpendicular to the common tangent line @xmath527 , it follows that @xmath20 is between @xmath530 and @xmath533 , and @xmath537 .
therefore , @xmath538 . for a class @xmath539 of structures ,
let @xmath540 denote the class of structures that are isomorphic to some members of @xmath539 .
we consider the following classes of finite convex geometries ; by circles we mean circles in the .
results by adaricheva @xcite and bergman @xcite show that @xmath541 we obtain from examples [ krkrntnk ] and [ cvmdfpnts ] , theorem [ thmmain](a ) , and proposition [ prsozhrw ] that @xmath542 @xmath44 consist of three non - collinear points and their barycenter @xmath45 does not satisfy 2 , and we conclude from proposition [ prsozhrw ] that @xmath543 in the lattices associated with members of @xmath544 , all join - irreducible elements are atoms .
this implies that @xmath545 in spite of , , , and , we do not have a satisfactory description of the partially ordered set @xmath546 in particular , we do not know whether @xmath547
if we augment the set with convex geometries obtained from @xmath3-dimensional spheres or , say , coplanar three - dimensional spheres , then the problem becomes even more difficult .
finally , while theorem [ thmmain ] describes @xmath548 in an abstract way , we have no similar descriptions for @xmath549 and @xmath550 .
adaricheva , k. : join - semidistributive lattices of relatively convex sets . in : contributions to general algebra 14 , proceedings of the olomouc conference 2002 ( aaa 64 ) and
the potsdam conference 2003 ( aaa 65 ) , verlag johannes heyn , klagenfurt , 2004 , 114 .
see also http://arxiv.org/abs/math/0403104 [ http://arxiv.org/abs/math/0403104 ] burris , s. , sankappanavar , h. p. : a course in universal algebra , graduate texts in mathematics , * 78*. springer - verlag , new york berlin , 1981 ; the millennium edition , http://www.math.uwaterloo.ca/$\sim$snburris/htdocs/ualg.html edelman , p.h . :
abstract convexity and meet - distributive lattices . in : proc .
combinatorics and ordered sets ( arcata , calif . , 1985 ) , contemp . math .
* 57 * , 127150 , amer .
soc . , providence , ri ( 1986 ) . | let @xmath0 be a finite set of circles in the plane .
we point out that the usual convex closure restricted to @xmath0 yields a convex geometry , that is , a combinatorial structure introduced by p.h .
edelman in 1980 under the name `` anti - exchange closure system '' .
we prove that if the circles are collinear and they are arranged in a `` concave way '' , then they determine a convex geometry of convex dimension at most 2 , and each finite convex geometry of convex dimension at most 2 can be represented this way .
the proof uses some recent results from lattice theory , and some of the auxiliary statements on lattices or convex geometries could be of separate interest .
the paper is concluded with some open problems . |
in this paper , we consider the cross - view and cross - modality matching problem between street - level rgb images and a geographic information system ( gis ) . specifically , given an image taken from street - level , the goal is to query a database assembled from a gis in order to return likely locations of the street - level query image which contain similar semantic concepts in a consistent layout . relying only on visual data is important in gps - denied environments , for images where
such tags have been removed on purpose ( e.g. for applications in intelligence or forensic sciences ) , for historical images , or images from the web which are lacking any gps tags . traditionally , such matching problems are solved by establishing pairwise correspondences between interest points using local descriptors such as sift @xcite with a subsequent geometric verification stage .
unfortunately , even if top - down satellite imagery is available , such an approach based on local appearance features is not applicable to the wide - baseline cross - view matching considered in our setting , mainly because of the following two reasons .
firstly , the extremely wide baseline between top - view gis imagery and the street - level image leads to a strong perspective distortion , and secondly , there can be drastic changes in appearance , e.g. due to different weather conditions , time of day , camera response function , etc . in this paper , we present a system to handle those two challenges .
we propose to phrase the cross - view matching problem in a semantic way .
our system makes use of two cues : what objects are seen and what their geometric arrangement is .
this is very similar to the way we humans try to localize ourselves on a map .
for instance , we identify that a house can be seen on the left of a lake and that there are two streets crossing in front of this house
. then we will look for the same semantic concepts in a consistent spatial configuration in the map to find our potential locations .
inspired by this analogy , in our system , instead of matching low - level appearance - based features , we propose to extract segments from the image and label them with a semantic concept employing imperfect classifiers which are trained using images of the same viewpoint and therefore are not invalidated by the viewpoint change .
gis often already provide highly - accurate semantically annotated top - down views thereby rendering the semantic labeling superfluous for the gis satellite imagery .
hence , we assume that such a semantic map is provided by the gis .
a typical query image and an excerpt of a semantic map can be seen in .
the semantic concepts we focus on ( e.g. , buildings , lakes , roads , etc ) form large ( and quite often insignificant in number ) segments in the image , and not points . therefore , we argue that a precise point - based geometric verification , like a ransac - search@xcite with an inlier criterion based on the euclidean distance between corresponding points , is not applicable .
we address these issues by designing a descriptor to robustly capture the spatial layout of those semantic segments .
pairwise asymmetric l2 matching between these descriptors is then used to find likely locations in the gis map with a spatial layout of semantic segments which is consistent with the one in the query image .
we also develop a tree - based search method based on a hierarchical semantic tree to allow fast geo - localization in a geographically broad areas .
cross - view matching in terms of semantic segments between street - level query image and a gis map joins several previous research directions . matching across a wide baseline
has traditionally been addressed with local image descriptors for points @xcite , areas @xcite , or lines @xcite .
registration of street - level images with oblique aerial or satellite imagery is generally based only on geometric reasoning .
previous work , e.g. @xcite , has reduced the matching problem to a 2d-2d registration problem by projecting ground models along vertical directions and rectifying the ground plane . unlike our approach
, the mentioned work requires a @xmath0d point cloud at query time , either from a laser scan @xcite or from multiple views based on structure - from - motion @xcite .
more recently , @xcite considered the registration problem of a dense multi - view - stereo reconstruction from street - level images to oblique aerial views . building upon accurate 3d city models for assembling a database ,
contours of skylines in an upward pointing camera can be matched to a city model @xcite or perspective distortion can be decreased by rectifying regions of the query image according to dominant scene planes @xcite .
@xcite also relied on rectification of building facades , however , their system relied on the repetitive structure of elements in large facades , enabling a rectification without access to a 3d city model . using contours between the sky and landscape has also been shown to provide valuable geometric cues when matching to a digital elevation model @xcite .
not using any 3d information , lin et al .
@xcite proposed a method to localize a street - level image with satellite imagery and a semantic map .
their system relies on an additional , large dataset which contains gps - annotated street - level images which therefore establish an explicit link between street - level images and corresponding areas in the satellite imagery and semantic map .
similarly to the idea of information transfer in exemplarsvms @xcite , once a short - list of promising images from this additional dataset has been generated by matching appearance - based features , appropriate satellite and semantic map information can be transferred from this short - list to the query image . visual location recognition and image retrieval system emphasise the indexing aspect and can handle large image collections : bag - of - visual - words @xcite , vocabulary trees @xcite or global image descriptors such as fisher vectors @xcite have been proposed for that purpose , for example .
all those schemes do not account for any higher - level semantic information .
more recently , @xcite has therefore introduced a scheme where pooling regions for local image descriptors are defined in a semantic way : detectors assign each segment a class label and a separate descriptor ( e.g. a fisher vector ) is computed for each such segment .
those descriptors rely on local appearance features , which fail to handle significant viewpoint changes faced in the cross - view matching problem considered in our paper .
also , this approach does not encode the spatial layout _ between _ semantic segments . if the descriptors are sufficiently discriminative by themselves , encoding this spatial layout is less important . in our case
however , the information available in the query image which is shared with the gis only captures class labels and a very coarse estimate of the segment shapes .
it is therefore necessary to capture both , the presence of semantic concepts and the spatial layout between those concepts , in a joint representation .
very recently , ardeshir et al.s work @xcite considered the matching problem between street - level image and a dataset of semantic objects .
specifically , deformable - part - models ( dpms ) were trained to detect distinctive objects in urban areas from a single street - level image .
the main objective of that paper was improved object detection with a geometric verification stage using a database of objects with known locations and the gps - tag and viewing direction of the image has been assumed to be known roughly .
they also present an exhaustive search based approach to matching an image against the entire object database .
the considered dpms in @xcite are well - localized and can be reduced to the centroid of the detection bounding box for a subsequent ransac step which searches for the best 2d - affine alignment in image space .
therefore , @xcite can only handle such `` spot '' based information and is not designed to handle less accurate and potentially larger semantic segments with uncertain locations such as the ones provided by classifiers for ` road ' or ` lake ' .
a graphical illustration of our proposed system is shown in .
the building blocks proposed by our paper will be described in detail in the next section , here we provide a rough overview of the system and describe the system inputs .
the computation of this input information relies on previous work and is therefore not considered as one of our contributions .
given a street - level query image , first we split it into superpixel segments and label each segment with a semantic concept by a set of pre - trained classifiers .
we train two different types of classifiers to annotate superpixels with class labels corresponding to a subset of the labels available in the gis . for semantic concepts with large variation in appearance and shape
, we are using the work by ren et al .
@xcite . however , in street level images , it is also quite common to spot highly informative objects with small with - in class variation .
typical examples are traffic signs or lamp posts . for each of those classes ,
we therefore employ a deformable - part - model ( dpm ) , similar to the ones of @xcite .
the second piece of input information is an estimate of the vertical vanishing point or the horizon line , e.g. @xcite describe how those entities can be estimated from a single image .
the perspective distortion of the ground plane can then be undone by warping the query image with a suitable rectifying homography , which is fully determined by the horizon line , assuming a rough guess of the camera intrinsic matrix is available .
we have opted to estimate the two inputs for our system , namely the ground plane location and the superpixel segmentation , in two entirely independent steps .
we note however , that at the expense of higher computational cost this could be estimated jointly @xcite .
the semantic map from the gis and the warped and labelled query image now share the same modality and suffer from less perspective distortion . the cross - view matching problem between query and semantic gis map
is then cast as a search for consistent spatial layouts of those semantically labeled regions . in order to do so , we design a semantic segment layout ( ssl ) descriptor which captures the _ spatial _ and _ semantic _ arrangement of those segments within a local support area located at a point in the query image or the semantic map . upon extracting such descriptors from the rectified image and semantic map ,
the problem can be reduced to a well - understood matching problem .
the goal of the ssl descriptor is to capture the _ presence _ of semantic concepts and at the same time encode the _ rough geometric layout _ of those concepts .
similar to several previous descriptors @xcite , the neighborhood around the descriptor centre is captured with pooling regions arranged in an annular pattern where the size of the regions increases with increasing distance to the descriptor centre .
we instantiate separate pooling regions for each semantic concept and the overall descriptor is the concatenation of the per - concept descriptors .
similar to appearance based local descriptors , we have to choose the locations where to extract descriptors and its orientation .
we are not aware of a good way to find reliable interest points in arrangements of semantic regions .
we initially experimented with placing a separate descriptor at the center of each semantic segment .
unfortunately , this choice turns out to be very sensitive to the location of the segment projected onto the ground plane .
this projection depends on an accurate estimate of the contact region of that segment with the ground plane . based on our experiments , it is challenging to get a sufficiently accurate estimate without manual user intervention . in some preliminary experiments ,
we therefore also tried to factor in the contact region uncertainty by blurring the contributions of a neighbouring segment along the line of sight between the camera and that segment , with an increasing amount of blur the further away the segment is from the camera center .
we think this is an elegant and theoretically sound way to account for those uncertainties and we refer to for a graphical illustration of the descriptor and of the subsequent processing steps of our preliminary pipeline .
we plan to explore this pipeline in future work in more detail . in this work
however , we settled for a simpler choice : a single descriptor is placed either in the center of the rectified image ( denoted ci in the experiments ) or at the center of the camera ( cc in the experiments ) . also , despite the ssl descriptor being more general , we only choose one annular pooling region thereby putting more emphasis on capturing the direction of semantic segments rather than the direction _ and _ distance .
this choice is again motivated by the difficulty of accurately estimating contact regions .
we suggest using descriptors which are not rotation invariant and an orientation therefore needs to be assigned to each descriptor as well .
the reason for this choice is that the alternative of defining a rotation - invariant descriptor leads to a considerably less discriminative descriptor and pairwise matching score .
it is straightforward to define a canonical orientation in the semantic gis map .
for example , the first pooling region can be chosen to point to geographic north . however ,
unless compass direction is available , it is not easily possible to define a canonical orientation for the rectified query image .
hence , we choose an arbitrary direction for the descriptors extracted from the rectified query image , and cope with not knowing the rotation parameters by employing a rotation invariant distance metric at the query time ( see section ) .
there are several ways to capture the overlap between a pooling region and a semantic segment .
an intuitive approach is to compute the area of intersection between the pooling region and the segment .
this can be fairly slow for irregularly shaped segments .
moreover , the shape of the segments in the query image are quite imprecise , so an accurate computation of the intersection area might be unnecessary or even harmful . here , we propose a scheme motivated by a probabilistic point of view .
the segments can be considered as a spatial probability distribution , that a point sampled in or close to this segment takes a certain label .
similarly , the pooling regions are interpreted as probability distributions of sampling a point at a certain location .
we then have to compute a statistical measure for the similarity between the two distributions . for discrete distributions ,
the mutual information is a good candidate . in our setting
however , we have to handle continuous distributions defined over the ground plane .
the bhattacharyya distance @xcite is a good way to measure the overlap between two continuous distributions and can be efficiently computed when we deal with gaussian distributions .
hence , in practice , in order to keep the computational requirements low , we will use a two - dimensional gaussian to define a pooling region and each segment will be approximated by a gaussian , as well . in detail , let @xmath1 and @xmath2 denote the gaussian for the segment and the pooling region , respectively .
the bhattacharyya distance is then given by @xmath3 where @xmath4 .
if multiple segments that are labeled with the same semantic concept are present , they can be treated as a gaussian mixture model ( gmm ) .
the bhattacharyya distance between two gmms can be approximated @xcite by @xmath5 , where @xmath6 and @xmath7 are a gmm for the segment and the pooling region , respectively . in our case , the pooling region is always represent by a single gaussian , so @xmath8 and @xmath9 .
the bhattacharyya distance is then converted to the hellinger distance @xmath10 where @xmath11 is the bhattacharyya coefficient .
our descriptor is the concatenation of all those hellinger distances @xmath12 .
each block of this descriptor corresponding to a semantic concept is then l2-normalized independently of the other blocks .
if a concept is not present , the hellinger distance for all pooling regions of that concept are set to zero .
the descriptor extraction from the semantic map is considerably simpler than from the query ] matching -------- given a street - level query image , our goal is to generate a ` heat - map ' of likely locations where this image has been taken .
the semantic gis map is therefore split into fix sized overlapping tiles , based on parameters such as average query image field - of - view and height of camera above ground . as described previously , the rotation alignment between the descriptor used in the semantic map and the query image is unknown . in order to cope with that
, we propose to rotate one of the descriptors in discrete rotation steps and compute the l2 distance for each step .
this boils down to the computation of a circular correlation ( or circular convolution ) between blocks of the two descriptors corresponding to the same semantic concept .
this can be implemented efficiently with a circulant matrix multiplication or even with a fft using the circular convolution theorem @xcite . in our implementation
, we are currently using the l2 distance between two descriptors .
however , especially for the descriptor placed at the camera centre , several pooling regions will not be contained in the field of view .
we therefore employ an asymmetric l2 distance where the distance contribution of pooling regions in the query descriptor which are not within the field of view of the camera is set to zero .
the field of view of the camera can easily be estimated given the image resolution and focal length .
groups with the hierarchical spectral clustering , as described in the paper .
we can see how this clustering produces semantically similar clusters .
note that the white areas in the top - most layer denote tiles which are mostly empty , i.e. had no semantic entries in the gis map.,scaledwidth=45.0% ] the reference area covered by gis maps is often very broad .
this leads to a large number of reference tiles which the query descriptor should be compared against .
several techniques , such as k - means or kd - tree , have developed for fast nearest neighbor search @xcite which we can employ to speed up this process .
we describe a procedure inspired by k - means trees @xcite suitable for pre - computing a hierarchical semantic tree which arranges the tiles of the map in a semantically and spatially meaningful way .
this speeds up the matching and also enables fast semantic queries in a gis system .
our tree construction is based on hierarchical spectral clustering with @xmath13 branches on each level . the similarity matrix required for spectral clustering
is composed of the previously described asymmetric distance between pairs of tiles of the gis map .
this provides us with a @xmath14 similarity matrix , where @xmath15 denotes the number of tiles .
we use spectral clustering , rather than k - means as used in k - means trees or similar methods , since we are employing our own customized distance function ( asymmetric l2 ) while those methods often assume a standard distance .
shows the two top layers of the hierarchical tree ( @xmath16 ) obtained by applying the procedure just defined on a large gis image .
what we observe is that the area has been partitioned into three well - defined semantic concepts : water ( in blue ) , area scarcely populated ( yellow ) and densely populated area ( grey ) in the first layer . in the second layer ,
each of those three areas are decomposed into three other areas , and construction of the tree continues as such .
the tree being semantically meaningful is a byproduct of the fact that our descriptor is targeted towards capturing semantic information . at query time , we start traversing the tree at the root .
the query image is matched against a random subset of tiles contained in each cluster of the current level .
the most promising child node out of the @xmath13 children at each level , which represents the cluster , is found in this way .
we use this technique since the cluster center is not straightforward to define for our customized distance functions .
the tree is traversed all the way down to the leaf nodes to find the final match .
the tree depth is @xmath17 , which leads to a speed - up of roughly @xmath18 where @xmath19 denotes the cardinality of the random subset of tiles explored at each level .
our framework has been tested for the geolocalization of generic outdoor images taken in the entire ( @xmath20 ) of the district of columbia , us , as extensive gis databases of this area are made available to public @xcite .
the area is depicted in figure [ fig : exp1 ] and includes a variety of different regions ( water , suburban , urban ) .
we have gathered a set of @xmath21 geo - tagged images from google maps and panoramio taken at different locations , which serves as a benchmark for our system .
we have @xmath22 semantic classes , that are @xmath23\{_road , tree , building , water , lamp post , traffic signal , traffic sign _ } , and in the following experiments we are going to use different sets of classes .
the size of a tile in the gis is around @xmath24 .
we set the number of rings in our descriptors to @xmath25 , and the number of pooling regions to @xmath26 .
we assume that the focal length is approximately known .
the homography which rectifies the ground plane is then determined by the vertical vanishing point or the horizon line .
we assume that the y - axis of the image is roughly aligned with the vertical direction and the vertical vanishing point is then given by the mle including all the lines segments @xmath27 degrees w.r.t .
the y - axis , see also @xcite .
as ground plane estimation is not the topic of our paper , we manually check and correct highly inaccurate estimates of the vanishing point in our benchmark images in order not to bias our evaluation toward mistakes in the ground plane estimation ( this is done for all baselines ) .
since the semantic gis map uses metric units and the rectified image is in pixels , the size or scale of the pooling regions need to be converted between those two units .
a reasonable assumption for street - level images is that the camera is roughly at @xmath28 above ground .
the conversion factor between metric units and pixels is then given by @xmath29 = \frac{f}{d } [ \text{meters}]$ ] .
the scale of the pooling regions is chosen such that a ssl descriptor captures significant contributions from segments within roughly @xmath30 m .
reports detailed qualitative results of our proposed cross - view matching scheme for several sample queries .
several interesting observations can be made in this figure .
first , as the covered area is very large , even such generic semantic cues can narrow down the search space to often @xmath31 .
second , the semantic and geometric similarity among the top 15 matching gis tiles shows the proposed method is successfully capturing such properties and is yet forgiving of the modeled uncertainties by not being overly discriminative ; note that the ground truth location ( marked in the cdf ) is often among the top few percents .
third , the heat map correlates well with the semantic content of the image . as for quantitative results ,
we compare different nearest neighbor ( nn ) classifiers with the following feature vectors : ssl descriptor in the center of image ( per query image / gis tile ) , ssl descriptor in the center of the camera , binary indicator vector encoding the presence or absence of semantic concepts ( _ presence _ term ) , ssl descriptor in the center of the camera plus the presence term . the last method is random matching
. the superior results of the ssl plus presence matching reported in confirm the necessity of jointly using semantic and coarse geometry for a successful localization task .
we also suspect that placing the ssl descriptor in the center of camera results in better performance than placing the descriptor in the center of the image because the former approach is less sensible to tiling quantization artifacts : e.g. objects on the left side of the field of view will generally remain on the left side even when the viewpoint is moved to the closest tile location .
we used the subset @xmath32\{_building , lamp post , traffic signal , traffic sign _ } of the @xmath33 semantic classes since this subset yielded the best overall geo - localization result .
for further evaluation , depicts the results of ssl plus presence matching over four different sets of semantic classes , which shows the contribution of each semantic class in the overall geo - localization results .
interestingly , the curves also show how certain sets of semantic classes may even mislead the geo - localization process .
we believe this is due to un - informativeness ( e.g. being too common ) of some classes as well as several sources of noise whose magnitude can vary between different classes ( e.g. inaccurate entries in gis map , semantic segmentation misclassifications , etc . ) . if enough training data were available , appropriate weights could be learned for each class .
this paper proposed an approach for cross - view matching between a street - level image and a gis map .
this problem was addressed in a semantic way .
a fast semantic segment layout descriptor has been proposed which jointly captures the presence of segments with a certain semantic concept and the spatial layout of those segments .
as our experimental evaluation showed , this enabled matching a street - level image to a large reference map based on purely semantic cues of the scene and their coarse spatial layout .
the results confirm that the semantic and topological cues captured by our method significantly narrow down the search area .
this can be used as an effective pre - processing for other less efficient but more accurate localization techniques , such as street view based methods .
* acknowledgments : * this work has been supported by the max planck center for visual computing and communication .
we also acknowledge the support of nsf career grant ( n1054127 ) . | matching cross - view images is challenging because the appearance and viewpoints are significantly different .
while low - level features based on gradient orientations or filter responses can drastically vary with such changes in viewpoint , semantic information of images however shows an invariant characteristic in this respect .
consequently , semantically labeled regions can be used for performing cross - view matching . in this paper
, we therefore explore this idea and propose an automatic method for detecting and representing the semantic information of an rgb image with the goal of performing cross - view matching with a ( non - rgb ) geographic information system ( gis ) . a segmented image forms the input to our system with segments assigned to semantic concepts such as traffic signs , lakes , roads , foliage , etc .
we design a descriptor to robustly capture both , the presence of semantic concepts and the spatial layout of those segments .
pairwise distances between the descriptors extracted from the gis map and the query image are then used to generate a shortlist of the most promising locations with similar semantic concepts in a consistent spatial layout . an experimental evaluation with challenging query images and
a large urban area shows promising results . |
x - ray production by charge exchange has received attention when a plausible link was established between cometary x - ray emission and solar wind heavy ions . fully stripped and
hydroglenlike carbon , nitrogen , oxygen , and neon , which are part of the solar wind , were thought to interact with gases in the cometary coma , producing k - shell x rays via the charge exchange mechanism @xcite .
recently , high - resolution techniques became available to study the x - ray emission of highly charged ions following charge exchange @xcite .
these measurements were able to resolve most x - ray lines , even those from levels with high principal quantum number @xmath3 . because the measurements were performed on an electron beam ion trap
, these measurements also included the emission from the @xmath4 level , which could not be detected in fast - ion experiments @xcite because of the slow radiative rate associated with this transition .
as a result it is now possible to record complete charge exchange induced x - ray spectra in the laboratory and to use them to simulate the observed emission from comets .
these new capabilities made it possible to show that cometary x - ray emission can be completely described by charge exchange induced x - ray emission @xcite .
alternative x - ray production mechanisms , which ranged from lower - hybrid wave heating , scattering of solar x rays by nano - sized dust grains to electron fluorescence and conversion of the kinetic energy of dust particles to x - rays @xcite , were shown to not be needed to simulate the observed cometary x - ray spectra . in the following we present measurements of the spectra produced by o@xmath0 interacting with n@xmath1 and ch@xmath2 as well as by ne@xmath5 interacting with neutral neon recorded at our facility with a high - resolution x - ray microcalorimeter array .
our measurements are carried out at the electron beam ion trap facility at the university of california lawrence livermore national laboratory ( uc - llnl ) .
this facility has been used for almost a decade for studying of the x - ray emission of highly charged ions following charge exchange @xcite .
the early measurements involved ions with charge as high as u@xmath6 @xcite .
the charge exchange measurements were enabled by utilizing the so - called magnetic trapping mode of the electron beam ion trap @xcite . in this mode , the electron beam is turned off after the ions have been produced .
the ions are confined in the radial direction by the longitudinal field of two superconducting magnets in the helmholtz configuration , and in the axial direction by a potential on the upper and lower drift tubes .
the interaction gas is selected by puffing the desired neutral gas into the trap .
more recently we used the uc - llnl facility to study the emission of highly charged ions found in the solar wind . in a study of the k - shell emission of o@xmath7 and ne@xmath8
we showed that the shape of the k - shell emission spectrum depends on the ion - neutral collision energy below about a few kev / amu @xcite .
these measurements were made with a windowless high - purity ge detector and thus could not resolve the individual x - ray transitions . in order to perform high - resolution spectral mesaurements of the x - ray emission
, we implemented in 2000 a microcalorimeter array detector on our facility @xcite .
the x - ray microcalorimeter spectrometer ( xrs ) was originally developed for the astro - e mission @xcite .
it consists of a 32-channel , 13 mm@xmath9 detector array sensitive to x rays with energy between 200 and 12,000 ev with a resolution of 10 ev .
the latter represents more than an order of magnitude improvement over the resolution of the ge detector used earlier .
an upgrade to the higher - resolution ( 6 ev ) microcalorimeter from the astro - e2 mission was implemented in october 2003 @xcite .
the improvement in resolving power is illustated in fig .
[ f1 ] , where we show a spectrum of the ne@xmath8 k - shell emission following charge exchange of bare neon ions with atomic neon . for comparison
we show both the data previously obtained @xcite with the germanium detector and the spectrum recorded with the xrs .
the individual x - ray transitions emanating from shells with different principal quantum number @xmath3 are now clearly resolved . a slight difference between the two spectra beyond
the difference in resolution arises in the relative magnitude of the rydberg lines from levels with @xmath10 .
this is presumably due to differences in the collision energy , i.e. , the ion temperature , between the two runs .
more measurements are needed to confirm this hypothesis .
the figure clearly illustrates the resonant nature of charge exchange between a neutral with inonzation potential @xmath11 and an ion with charge @xmath12 , in which the electron is preferrentially transferred to a level with principal quantum number @xmath13 given by @xmath14 where @xmath15 is the ionization potential of hydrogen @xcite .
moreover , the strong emission from levels with principal quantum number @xmath10 is a signature of the low ion - neutral collision energy in our trap ( typically less than 20 ev / amu ) , as discussed in @xcite .
the details of the emission from high-@xmath3 levels depend on the ionization potential of the interaction gas , as shown in the above equation .
energy conservation arguments predict that electrons are captured into higher @xmath3 levels if the ionization potential of the donor electron is lower .
this has been shown experimentally in @xcite and is illustrated by the new xrs spectra in fig .
[ f2 ] . the emission from the highest-@xmath3 level shifts from @xmath16 in the interaction with ch@xmath2 , which has a 12.6 ev ionization potential , to @xmath17 in the interaction with n@xmath1 , which has a 15.6 ev ionization potential .
astro - e2 , which is scheduled to be launched in 2005 , employs a 32-channel microcalorimeter array similar to the one we use at our facility .
the satellite is expected to return cometary x - ray spectra with resolution equal or better than that of the spectra we presented .
the observations made with astro - e2 will further stimulate laboratory measurements of charge exchange produced x - ray emission from highly charged ions and will provide new challenges for atomic theory to explain the observations .
work by the university of california lawrence livermore national laboratory was performed under the auspices of the u.s.department of energy under contract no .
w-7405-eng-48 and supported by nasa s planetary atmospheres program under w-19,938 . 10 c. m. lisse , k. dennerl , j. englhauser , m. harden , f. e. marshall , m. j. mumma , r. petre , j. p. pye , m. j. ricketts , j. schmitt , j. trmper , and r. g. west , science * 274 * , 205 ( 1996 ) .
r. m. hberli , t. i. gombosi , d. l. d. zeeuw , m. r. combi , and k. g. powell , science * 276 * , 939 ( 1997 ) .
r. wegmann , h. u. schmidt , c. m. lisse , k. dennerl , and j. englhauser , planet .
space sci . * 46 * , 603 ( 1998 ) .
n. a. schwadron and t. e. cravens , astrophys .
j. * 544 * , 558 ( 2000 ) .
t. e. cravens , science * 296 * , 1042 ( 2002 ) .
p. beiersdorfer , k. r. boyce , g. v. brown , h. chen , s. m. kahn , r. l. kelley , m. may , r. e. olson , f. s. porter , c. k. stahle , and w. a. tillotson , science * 300 * , 1461 ( 2003 )
. j. b. greenwood , i. d. williams , s. j. smith , and a. chutjian , astrophys .
j. ( lett . ) * 533 * , l175 ( 2000 ) .
j. b. greenwood , i. d. williams , s. j. smith , and a. chutjian , phys .
a * 63 * , 062707 ( 2001 ) .
v. a. krasnopolsky , icarus * 128 * , 368 ( 1997 ) .
r. bingham , j. m. dawson , v. d. shapiro , d. a. mendis , and b. j. kellett , science * 275 * , 49 ( 1997 ) .
ip and v. w. chow , icarus * 130 * , 217 ( 1997 ) .
a. owens , a. n. parmar , t. oosterbroek , a. orr , l. a. antonelli , f. fiore , r. schulz , g. p. tozzi , m. c. maccarone , and l. piro , astrophys . j. ( lett . )
* 493 * , l47 ( 1998 ) .
r. bingham , b. j. kellett , j. m. dawson , v. d. shapiro , and d. a. mendis , astrophys .
j. ( suppl . ) * 127 * , 233 ( 2000 ) .
p. beiersdorfer , l. schweikhard , j. crespo lpez - urrutia , and k. widmann , rev .
* 67 * , 3818 ( 1996 ) .
l. schweikhard , p. beiersdorfer , g. v. brown , j. r. crespo lpez - urrutia , s. b. utter , and k. widmann , nucl . instrum .
methods * b 142 * , 245 ( 1998 ) . , phys .
scripta * t80 * , 121 ( 1999 ) . p. beiersdorfer , r. e. olson , g. v. brown , h. chen , c. l. harris , p. a. neill , l. schweikhard , s. b. utter , and k. widmann , phys . rev . lett . * 85 * , 5090 ( 2000 ) . p. beiersdorfer , r. e. olson , l. schweikhard , p. liebisch , g. v. brown , j. crespo lpez - urrutia , c. l. harris , p. a. neill , s. b. utter , and k. widmann , in _ the physics of electronic and atomic collisions _
, aip conference proceedings no . 500 , ed .
by y. itikawa et al .
( aip , new york , 2000 ) , p. 626
( 1990 ) .
p. beiersdorfer , b. beck , r. e. marrs , s. r. elliott , and l. schweikhard , rapid commun .
mass spectrom .
* 8 * , 141 ( 1994 ) .
p. beiersdorfer , b. beck , s. becker , and l. schweikhard , int .
j. mass spectrom .
* 157/158 * , 149 ( 1996 ) .
p. beiersdorfer , c. m. lisse , r. e. olson , g. v. brown , and h. chen , astrophys .
j. ( lett . ) * 549 * , l147 ( 2001 ) .
f. s. porter , m. d. audley , p. beiersdorfer , k. r. boyce , r. p. brekosky , g. v. brown , k. c. gendreau , j. gygax , s. kahn , r. l. kelley , c. k. stahle , and a. e. szymkowiak , proc .
spie * 4140 * , 407 ( 2000 ) .
spie * 3765 * , 114 ( 1999 ) .
f. s. porter , g. v. brown , k. r. boyce , r. l. kelley , c. a. kilbourne , p. beiersdorfer , h. chen , s. terracol , s. m. kahn , and a. e. szymkowiak , rev .
instrum . in press
( 2004 ) .
r. k. janev and h. winter , phys .
rep . * 117 * , 265 ( 1985 ) .
j. a. perez , r. e. olson , and p. beiersdorfer , j. phys .
b * 34 * , 3063 ( 2001 ) .
1 . x - ray emission of ne@xmath8 measured with high - purity ge detector ( solid trace ) and with the x - ray microcalorimeter ( dashed trace ) .
the emission is produced in the interaction of ne@xmath5 ions with atomic neon .
the collision energy is a few ev / amu .
k - shell emission of h - like o@xmath7 produced by charge exchange with ch@xmath2 ( top ) and n@xmath1 ( bottom ) .
note changes in the ratio of lyman-@xmath18 emanating from @xmath19 and lyman-@xmath20 from @xmath21 . | x - ray emission following charge exchange has been studied on the university of california lawrence livermore national laboratory electron beam ion traps ebit - i and ebit - ii using a high - resolution microcalorimeter .
the measured spectra include the k - shell emission from hydrogenlike and heliumlike c , n , o , and ne needed for simulations of cometary x - ray emission .
a comparison of the spectra produced in the interaction of o@xmath0 with n@xmath1 and ch@xmath2 is presented that illustrates the dependence of the observed spectrum on the interaction gas .
charge exchange , x - ray lines , comets , microcalorimeter 34.70.+e , 32.30.rj , 96.50.gn |
affine schubert calculus is a subject that ties combinatorics , algebraic geometry and representation theory together .
its modern development is motivated by the relation between @xmath0-schur functions and the ( co)homology of the affine grassmannian of @xmath1 .
@xmath0-schur functions were introduced by lapointe , lascoux , morse @xcite in the study of macdonald polynomial positivity , a mostly combinatorial branch of symmetric function theory . + peterson @xcite identified the equivariant homology of the affine grassmannian with a subalgebra of the affine nilhecke algebra @xmath2 , now called the peterson algebra .
lam @xcite proved that @xmath0-schur functions can be identified with the schubert classes of the homology of the affine grassmannian of @xmath1 .
the nilhecke ring acts as divided difference operators on the equivariant cohomology of kac - moody partial flag varieties . by using the correspondence
, one can investigate problems about the ( co)homology of the affine grassmannian of @xmath1 by translating them into the theory of @xmath0-schur functions and understanding the combinatorics of @xmath0-schur functions .
there are affine analogues of the classical theory of pieri rules , tableaux , and stanley symmetric functions @xcite .
+ lam , lapointe , morse and shimozono @xcite introduced the strong schur functions , indexed by elements in the affine symmetric group @xmath3 .
these strong schur functions generalize the @xmath0-schur functions combinatorially .
they conjectured a series of properties of strong schur functions such as the symmetry of the strong schur functions .
berg , saliola and serrano @xcite studied the pieri operators acting on the affine nilcoxeter ring @xmath4 to establish some of conjectures in @xcite .
+ in this paper , we prove the affine pieri rule for the cohomology of the affine flag variety conjectured in @xcite .
we introduce the cap operators acting on the affine nilcoxeter ring @xmath4 by investigating the work of kostant and kumar @xcite and show that the cap operators for pieri elements are the same as the pieri operators defined in @xcite by using the strong strips .
the affine pieri rule gives us geometric interpretation of the skew strong schur functions as an affine grassmannian part of the cap product of the schubert classes in ( co)homology of the affine flag variety .
we now describe these two operators . for @xmath5 ,
a _ marked strong cover _ @xmath6 consists of @xmath5 and an integer @xmath7 such that @xmath8 where @xmath9 and @xmath10 is the bruhat order .
we use the notation inside@xmath11 and outside@xmath12 .
+ lam , lapointe , morse and shimozono @xcite introduced the notion of _ strong strips _ to define @xmath0-schur functions and strong schur functions .
a _ strong strip _
@xmath13 of length @xmath14 from @xmath15 to @xmath16 , denoted by @xmath17 , is a path @xmath18 where @xmath19 .
we let inside@xmath20 , outside@xmath21 .
+ berg , saliola and serrano @xcite studied the _
pieri operators _ @xmath22 acting on the affine nilcoxeter ring @xmath4 defined by @xmath23 where the sum is over all strong strips from @xmath24 to @xmath15 of length @xmath14 .
they proved that the operators @xmath22 commute for all @xmath14 so that the symmetry of strong schur functions follows .
they also showed a series of properties of @xmath22 that uniquely determine the operator @xmath22 .
let us consider the affine type @xmath25 root datum and corresponding kac - moody flag variety @xmath26 ( see @xcite for details ) .
let @xmath27 denote the schubert basis for the equivariant cohomology @xmath28 of @xmath26 .
one of main problems in affine schubert calculus is to find a combinatorial formula for the structure constants @xmath29 where @xmath30 . in @xcite , kumar and kostant
identified the torus - equivariant cohomology @xmath28 of the affine flag variety and the _ dual _ of the nilhecke ring . using this connection
, one can compute the structure constants @xmath29 by calculating the coproduct structure constants of @xmath2 .
more precisely , we have @xmath31 for @xmath32 , a _ cap operator _
@xmath33 on @xmath4 is defined by @xmath34 geometrically , the cap operator is the cap product on the ordinary homology and cohomology of the affine flag variety .
more precisely , the cap operator @xmath33 can be considered as an element in @xmath35 and @xmath36 can be considered in @xmath37 so that the cap product @xmath38 can be described by @xmath39 .
note that the cap operator only keep track of ordinary cohomology since @xmath29 is constant when @xmath40 .
+ let @xmath41 be the pieri element @xmath42 in @xmath3 where indices are taken modulo @xmath43 .
we study the cap operators @xmath44 for @xmath41 and show that @xmath44 satisfy the properties of the pieri operators @xmath22 that berg , saliola and serrano proved in @xcite .
this establishes the following main theorems equivalent to the affine pieri rule conjectured in @xcite .
[ affinepieri ] for @xmath45 with @xmath46 , @xmath47 counts the number of strong strips from @xmath24 to @xmath15 . as a corollary
, one can compute @xmath48 for @xmath49 or @xmath50 for any @xmath14 by applying automorphisms on dynkin diagram of affine type @xmath25 , namely , the map @xmath51 and @xmath52 for @xmath53 .
+ the paper is structured as follows . in section 2 ,
we recall some notions about the affine symmetric groups , @xmath0-schur functions , strong schur functions and root systems . in section 3
, we define the affine nilhecke ring @xmath2 and study its properties . in section 4 , we recall some statements concerning the affine flag variety and its equivariant cohomology as well as the relationship between the equivariant cohomology of affine flag variety and the coproduct structure of the affine nilhecke ring . in section 5 , we define the cap operator .
we show that these cap operators for pieri elements agree with the pieri operators defined by berg , saliola and serrano .
we also prove the affine pieri rule for the ordinary cohomology of the affine flag variety . in section 6 , we apply the affine pieri rule to show that the structure constants @xmath54 for the cohomology of the affine flag variety can be described in terms of strong schur functions when @xmath15 is @xmath55-grassmannian .
this also gives a geometric interpretation of the skew strong schur functions .
i would like to thank thomas lam for helpful discussions and for introducing affine schubert calculus to me .
i also wish to thank kias for valuable support .
positive integers @xmath56 and @xmath57 will be fixed throughout the paper .
let @xmath3 denote the affine symmetric group with simple generators @xmath58 satisfying the relations @xmath59 where indices are taken modulo @xmath43 .
an element of the affine symmetric group may be written as a word in the generators @xmath60 .
a _ reduced word _ of the element is a word of minimal length .
the _ length _ of @xmath24 , denoted @xmath61 , is the number of generators in any reduced word of @xmath24 .
+ the _ bruhat order _ , also called _ strong order _ , on affine symmetric group elements is a partial order where @xmath62 if there is a reduced word for @xmath15 that is a subword of a reduced word for @xmath24 . if @xmath62 and @xmath63 , we write @xmath64 .
it is well - known that @xmath64 if and only if there exists a reflection @xmath65 such that @xmath66 and @xmath63 . for type @xmath25 ,
the set @xmath67 consists of transpositions @xmath68 .
see @xcite for instance .
+ the subgroup of @xmath3 generated by @xmath69 is naturally isomorphic to the symmetric group @xmath70 .
the @xmath55-grassmannian elements are minimal length coset representatives of @xmath71 .
let @xmath72 denote the ring of symmetric functions .
for a partition @xmath73 , we let @xmath74 denote the monomial , homogeneous , elementary and schur symmetric functions , respectively , indexed by @xmath73 . each of these families forms a basis of @xmath72 .
+ let @xmath75 denote the subalgebra generated by @xmath76 .
the elements @xmath77 with @xmath78 form a basis of @xmath75 .
let @xmath79 denote the quotient of @xmath72 by the ideal @xmath80 generated by @xmath81 with @xmath82 .
the image of elements @xmath81 with @xmath78 form a basis of @xmath83 . lam ( * ? ? ?
* theorem 7.1 ) showed that @xmath75(resp .
@xmath83 ) are isomorphic to homology(resp .
cohomology ) of affine grassmannian as hopf - algebras .
+ there are many conjecturally equivalent definitions of @xmath0-schur functions in @xcite .
we follow the definition of @xmath0-schur functions in @xcite using strong strips .
note that the marking of strong covers and strong strips defined in this extended abstract follows the notation of @xcite .
this differs from the definition of strong covers and strips in @xcite by reversing direction . +
a _ marked strong cover _ @xmath6 consists of @xmath5 and an integer @xmath7 such that @xmath8 where @xmath84 .
we use the notation inside@xmath11 and outside@xmath12 .
strong strip _ s of length @xmath14 from @xmath15 to @xmath16 , denoted by @xmath17 , is a path @xmath18 where @xmath19 .
we let inside@xmath20 , outside@xmath21 , and size@xmath85 .
strong tableau _ is a sequence @xmath86 of strong strips @xmath87 such that outside@xmath88 inside@xmath89 for all @xmath90 and size@xmath91 for all sufficiently large @xmath14 .
we define inside@xmath92 inside@xmath93 and outside@xmath92 outside@xmath94 for @xmath14 large .
the _ weight _
@xmath95 of @xmath96 is the sequence @xmath97 we say that @xmath96 has _
@xmath98 where @xmath99 inside@xmath100 and @xmath101 outside@xmath100 so that we have @xmath102 . for fixed @xmath103 ,
define the strong schur function @xmath104 where t runs over the strong tableaux of shape @xmath98 .
if @xmath15 is @xmath55-grassmannian and @xmath16 is the identity element , @xmath105 is a @xmath0-schur function @xmath106 where @xmath107 is a bijection between the set of 0-grassmannian elements and the set of @xmath0-bounded partitions . for details ,
see @xcite for instance .
+ we shall assume basic familiarity with weyl groups , root systems , and weights .
see @xcite for details .
+ let @xmath108 denote an affine cartan matrix , where @xmath109 , so that @xmath110 is corresponding finite cartan matrix . for type @xmath111(corresponding to @xmath3 )
, we have @xmath112 and @xmath113 let @xmath114 be the root system for @xmath70 .
let @xmath115 denote the set of positive roots , negative roots respectively .
let @xmath116 denote the simple roots and @xmath117 denote the simple coroots .
let @xmath118 denote the highest root @xmath119 of @xmath120 .
+ let @xmath121 and @xmath122 denote the affine root system and positive affine roots .
the positive simple affine roots ( resp .
coroots ) are @xmath123 ( resp .
@xmath124 ) . the null root @xmath125 is given by @xmath126 .
similarily , the canonical central element @xmath127 is given by @xmath128 . a root @xmath129 is real if it is in @xmath3-orbit of the simple affine roots , and imaginary otherwise .
the imaginary roots are exactly @xmath130 .
every real affine root is of the form @xmath131 , where @xmath132 and @xmath133 .
the root @xmath131 is positive if @xmath134 , or if @xmath135 and @xmath136 .
let @xmath137 denote the set of positive roots in @xmath121 .
+ let @xmath138 denote the root lattice and let @xmath139 denote the coroot lattice .
let @xmath140 and @xmath141 be the weight lattice and coweight lattice respectively .
we have inclusions @xmath142 and a map @xmath143 given by sending @xmath125 to @xmath55 .
let @xmath144 denote the pairing between @xmath140 and @xmath141 satisfying @xmath145 .
+ let @xmath146 be the affine weight lattice and @xmath147 be the set of affine fundamental weights .
then @xmath148 the affine weyl group @xmath3 acts on the affine weight lattice @xmath149 and the affine coweight lattice by @xmath150 this is called the _ level zero action_. an element @xmath151 can be uniquely written of the form @xmath152 where @xmath153 and @xmath154 .
then the level zero action on @xmath155 is given by @xmath156 for a real root @xmath129 , we let @xmath157 denote the corresponding reflection , defined by @xmath158 if @xmath159 . the reflection @xmath157 acts on weights by @xmath160 .
for a strong cover @xmath161 , let @xmath162 denote the unique positive root satisfying the equation @xmath163 .
let @xmath164 be the polynomial ring having a variable for each free generator of @xmath140 and let @xmath165 be the fraction field . + define the @xmath166-vector space @xmath167 with basis @xmath168 , with product given by @xmath169 for
any real root @xmath170 define the element @xmath171 by @xmath172 we write @xmath173 for @xmath174 , we have @xmath175 the @xmath176 satisfy the braid relations as the @xmath60 in @xmath168 , i.e. , @xmath177
. therefore it makes sense to define @xmath178 one can check that @xmath179 the _ nilcoxeter algebra _
@xmath4 is the subring of @xmath180 generated by @xmath176 over @xmath181 for @xmath182 .
the set @xmath183 forms a basis of @xmath4 over @xmath181 .
nilhecke algebra _
@xmath2 is the subring of @xmath180 generated by @xmath13 and @xmath4 .
we state the following result ( * ? ? ?
* proposition 4.30 ) .
let @xmath151 and @xmath184 .
then @xmath185 where @xmath129 is always taken to be a positive root in @xmath186 .
the coefficients @xmath187 are known as _
chevalley coefficients_. the affine nilhecke algebra @xmath2 defined above is slightly different from those defined in @xcite . in @xcite , kostant and kumar set @xmath188 instead of @xmath189 .
we have a projection @xmath190 and the map induces @xmath191 which is compatible with theorem [ cohomology ] .
the image of the null root @xmath192 in @xmath140 is 0 .
geometric interpretation of @xmath140 and @xmath149 will be discussed in section [ affineflagvariety ] .
a word @xmath193 with indices in @xmath194 is called _ cyclically decreasing _ if each letter occurs at most once and whenever @xmath60 and @xmath195 both occurs in the word , @xmath195 precedes @xmath60 . for @xmath196
there is the unique cyclically decreasing element @xmath197 with letters @xmath198 . for @xmath199 ,
let @xmath200 where @xmath201 and @xmath202 for @xmath203 by convention .
lam @xcite showed that the elements @xmath204 commute and freely generate a subalgebra @xmath205 of @xmath4 called _ affine fomin - stanley algebra_. it is well - known that @xmath205 is isomorphic to @xmath75 via the map sending @xmath206 to @xmath207 .
therefore , the set @xmath208 forms a basis of @xmath205 .
+ there is a coproduct structure on the nilhecke ring @xmath2 .
kostant and kumar showed that the structure constants of the coproduct in @xmath2 are the same as the structure constants of the equivariant cohomology ring of @xmath26 .
details and proofs can be found in @xcite .
+ let @xmath209 be the left @xmath166-linear map defined by @xmath210 [ delta ] the map @xmath209 induces the unique left @xmath13-module homomorphism @xmath211 such that @xmath212 for @xmath213 and any expression of @xmath214 and @xmath215 of the form @xmath216 the product of @xmath214 and @xmath215 can be computed by the naive componentwise product @xmath217 for @xmath218 , the _ equivariant schubert structure constants _
@xmath219 are defined as the coefficients in the expansion @xmath220 .
the followings are properties of @xmath221 . 1
. @xmath222 unless @xmath223 and @xmath224 .
@xmath221 is homogeneous of degree @xmath225 .
@xcite @xmath226.$ ] [ essc ] the last property in theorem [ essc ] is called _
graham positivity_. one can use theorem [ delta ] to compute @xmath221 explicitly .
[ p ] @xcite let @xmath227 and @xmath228 .
then @xmath229 where @xmath230 if @xmath231 and @xmath232 otherwise , and @xmath233 .
in this section , we define the kac - moody flag variety @xmath26 and establish the relationship between the equivariant cohomology of @xmath26 and the coproduct structure on @xmath2 .
there are two definitions of kac - moody flag variety @xmath26 in @xcite , but we only recall @xmath26 as the kac - moody flag ind - variety in @xcite .
+ let @xmath234 denote the kac - moody group of affine type associated with @xmath1 and let @xmath235 denote its borel subgroup .
the kac - moody flag ind - variety @xmath236 is paved by cells @xmath237 whose closure @xmath238 is called the schubert variety .
schubert variety defines a schubert class @xmath239_{t_{\text{af } } } \in h^*_{t_{\text{af}}}(\widetilde{fl})$ ] where @xmath240 is the maximal torus in @xmath235 . for our setting , we consider @xmath241 where @xmath96 is a maximal torus in @xmath1 .
kumar and kostant @xcite identified the equivariant cohomology ring of @xmath26 with the _ dual _ of nilhecke ring @xmath2 . by restricting the group action from @xmath240 to @xmath96 , this identification shows that the structure constant of @xmath28 is the same as @xmath29 defined in section [ coproduct ] .
[ cohomology ] @xcite the @xmath96-equivariant cohomology of @xmath26 has a basis @xmath242_t \in h^*_t(\widetilde{fl})\}$ ] over @xmath243 .
moreover , the structure constants of @xmath28 are @xmath221 , i.e. , @xmath244_t [ x^{u_2}]_t=\sum_w p^w_{u_1,u_2}[x^w]_t.\ ] ] [ ordinary ] the ordinary cohomology of @xmath26 has a basis @xmath242 \in h^*(\widetilde{fl})\}$ ] over @xmath181 and corresponding structure constants are just @xmath245 where @xmath246 is the evaluation at @xmath55 .
let @xmath247 denote the map sending a polynomial to its constant term , called the _ evaluation map _ at @xmath55 .
it extends to the map @xmath248 given by @xmath249 .
_ pieri elements _
@xmath41 are defined by @xmath250 for @xmath251 . for @xmath151 ,
let us define _ the cap operator _
@xmath252 on @xmath2 ( and @xmath4 ) by @xmath253 if @xmath254 , we call @xmath44 the _ pieri operator _ and denote by @xmath255 . note that @xmath256 when @xmath257 by theorem [ essc ] .
finding a combinatorial formula for such @xmath258 is one of the important problems in schubert calculus and it is not completely known even for finite flag varieties . in this section , we will prove that @xmath259 is the same as the pieri operator @xmath22 defined by berg , saliola , serrano in @xcite .
this identification gives a combinatorial description of the pieri rule for the ordinary cohomology of @xmath26 and therefore proves the main theorem [ affinepieri ] .
+ let us state properties of @xmath22 described in @xcite .
* theorem 4.8)[bss1 ] suppose @xmath260 and @xmath261 .
then @xmath262 [ bss2 ] ( * ? ? ?
* lemma 4.5 ) for @xmath263 and @xmath264 , @xmath265 ( * ? ? ?
* proposition 4.3 ) [ bss3 ] for @xmath266,@xmath151 and @xmath264 , we have @xmath267 for @xmath151 , there is a unique decompostion @xmath268 where @xmath269 is @xmath55-grassmannian and @xmath270 .
note that @xmath271 where @xmath107 is the bijection between @xmath55-grassmannian elements and @xmath0-bounded partitions mentioned in section 2 . in this setup ,
the set @xmath272 forms a basis of @xmath2 ( see @xcite ) .
therefore , above theorems uniquely determine @xmath22 . in this section ,
we prove the same theorems for @xmath259 .
this implies that @xmath273 and theorem [ affinepieri ] follows .
[ 0 ] suppose @xmath260 and @xmath261
. then @xmath274 _ proof .
_ let @xmath275 for @xmath227 and @xmath276 .
it is enough to show that @xmath277 for all @xmath278 by theorem [ p ] .
note that any reducd word of @xmath16 does not contain @xmath279 since @xmath280 , and @xmath41 has the unique reduced word ending with @xmath279 , namely @xmath281 .
let @xmath282 and @xmath283 .
then @xmath284 for @xmath285 .
therefore , we have @xmath286 and the theorem follows .
+ [ dh ] for @xmath263 and all @xmath264 , @xmath287 _ proof . _
this theorem follows as a corollary of ( * ? ? ?
* lemma 7.7 ) .
* lemma 7.7 ) for @xmath263 , @xmath288 since @xmath41 appears in @xmath206 , terms appearing in @xmath289 with second entry @xmath41 are exactly @xmath290 . + in order to prove analogue of theorem [ bss3 ] for @xmath255 , we need the following lemmas .
[ hp ] ( * ? ? ?
* lemma 7.2 ) let @xmath291 and @xmath292
. then @xmath293 [ translation ] let @xmath294 be the group automorphism defined by @xmath295 .
the automorphism induces the ring automorphism @xmath296 . for @xmath291
, we have @xmath297 _ proof . _
since @xmath205 is generated by @xmath206 for @xmath298 and the equality @xmath299 is obvious from the definition of @xmath206 , we are done .
+ now we are ready to prove the following property of @xmath255 .
[ dual ] for @xmath266,@xmath151 and @xmath264 , we have @xmath300 _ proof .
_ by the definition of the pieri operator and remark [ ordinary ] , we have @xmath301 note that the last equality follows from theorem [ delta ] . by ( [ naive ] ) , we can compute @xmath302 and @xmath303
seperately and multiply them together .
let @xmath304 denote the element @xmath305 for @xmath306 .
then @xmath307 .
therefore , we have @xmath308 + _ proof of theorem [ affinepieri ] .
_ theorem [ 0 ] , [ dh ] , [ dual ] uniquely determine the pieri operator @xmath255 and we have @xmath273 by theorem [ bss1 ] , [ bss2 ] , [ bss3 ] .
therefore , theorem [ affinepieri ] follows .
+ theorem [ dual ] generalizes ( * ? ? ?
* theorem 3.8 ) .
in fact , the author believes that there is a gap in the first few lines in their proof of ( * ? ? ?
* theorem 3.8 ) , and theorem [ dual ] resolves the problem as well .
let @xmath309 denote the set of @xmath55-grassmannian elements in @xmath3 . in this section , we prove that theorem [ affinepieri ] implies the following conjecture , as stated in @xcite . note that berg , saliola and serrano @xcite proved @xmath314 of theorem [ strong ] and @xmath315 when @xmath316 are in @xmath309 . to be more precise , they showed that for @xmath317 , @xmath54 is the coefficient of @xmath318 in @xmath319 , where @xmath318 is the _ affine stanley symmetric function _ labeled by @xmath24 .
lam , lapointe , morse and shimozono @xcite introduced _ weak schur functions _ defined by @xmath320 for @xmath45 where the summation is over all @xmath0-bounded partitions and @xmath321 is the inner product on @xmath4 satisfying @xmath322 .
if @xmath323 and @xmath16 is the identity element , @xmath324 is the affine stanley symmetric function @xmath318 .
+ let @xmath325 denote the affine grassmannian associated with @xmath326 . since @xmath327 for the affine kac - moody group @xmath234 and its maximal parabolic subgroup @xmath328 , we have the schubert bases @xmath329 @xmath330 lam @xcite showed that @xmath331 and @xmath332 can be represented by @xmath0-schur functions @xmath333 and affine stanley symmetric functions @xmath334 via the isomorphisms @xmath335 and @xmath336 . in fact , _ affine insertion _ @xcite offers a duality between weak and strong orders combinatorially .
+ define the _ affine cauchy kernel _ @xmath337 by @xmath338 as an element in @xmath339 where @xmath340 is the image of @xmath341 in @xmath342 .
the second equality follows from the well - known equality @xmath343 after taking the projection map @xmath344 .
+ the duality between strong and weak orders produces the following affine cauchy identity ( * ? ? ?
* corollary 4.6 ) .
now we are ready to prove theorem [ strong ] , following ( * ? ? ?
* proposition 4.19 ) .
+ _ proof of theorem [ strong ] .
_ let @xmath347 @xmath348 denote the schubert bases for the cohomology ( resp . homology ) of the affine flag variety .
in this setting , theorem [ affinepieri ] can be described as follows : theorem [ affinepieri2 ] completely determines the action of @xmath353 on @xmath35 , obtained from the inclusion @xmath354 .
+ we may consider the affine cauchy kernel @xmath355 as an element of the completion @xmath356 via the isomorphism @xmath357 .
let @xmath358 denote the inner product on @xmath35 defined by @xmath359 . by the definition of @xmath355 and theorem [ affinepieri2 ]
, we have @xmath360 where @xmath361 .
note that we have also used the definition of the strong schur functions and the equality @xmath362 ( see @xcite ) . by corollary [ affinecauchy ] , we may also write @xmath363 so that @xmath364 since @xmath365 are @xmath0-schur functions in @xmath366 , we have @xmath367 .
+ the skew strong schur function @xmath368 can be geometrically interpreted as the image of @xmath369 under the projection @xmath370 where @xmath33 is the cap product acting on @xmath37 .
indeed , by the definition of @xmath33 and remark [ ordinary ] we have @xmath371 .
since the projection @xmath370 maps @xmath372 to @xmath333 , it sends @xmath369 to @xmath373 by theorem [ strong](3 ) .
therefore , @xmath368 captures the @xmath55-grassmannian part of @xmath369 .
note that when @xmath24 is not @xmath55-grassmannian , we have @xmath374 ( see @xcite ) . note
that lam , lapointe , morse and shimozono @xcite defined the marked strong cover @xmath375 where @xmath376 for a fixed constant @xmath377 instead of @xmath378 in our setup .
the affine pieri rule conjectured in @xcite is the following .
* conjecture 4.15)[affinepieri3 ] let @xmath151 and @xmath349 .
then in @xmath35 we have @xmath379 where the sum runs over strong strips @xmath13 of size @xmath351 with respect to @xmath377 such that @xmath352
. we can obtain theorem [ affinepieri3 ] by applying @xmath380 to theorem [ affinepieri2 ] . recall that @xmath381 is the automorphism of @xmath3 sending @xmath60 to @xmath382 . by applying @xmath381 to theorem [ affinepieri2 ]
, we get the dual affine pieri rule . | we prove the affine pieri rule for the cohomology of the affine flag variety conjectured by lam , lapointe , morse and shimozono .
we study the cap operator on the affine nilhecke ring that is motivated by kostant and kumar s work on the equivariant cohomology of the affine flag variety .
we show that the cap operators for pieri elements are the same as pieri operators defined by berg , saliola and serrano .
this establishes the affine pieri rule . |
on approach to their glass transition @xcite glass forming systems display increasingly heterogeneous dynamics @xcite .
this dynamic heterogeneity is not correlated in any obvious way to structural features . in order to uncover a possible configurational origin for dynamic heterogeneity harrowell and
coworkers recently proposed the concept of dynamic propensity @xcite : a particle s propensity is defined as some measure of its mobility over a period of time , such as its mean square displacement , averaged over all dynamic trajectories which start from the same initial configuration .
this ensemble of trajectories is sometimes referred to as the iso - configurational ensemble .
dynamic propensity was found to be a good indicator of dynamic heterogeneity in simulations of atomistic models @xcite for times at least as long as the structural relaxation time .
while this suggested a configurational origin for the heterogeneity in the dynamics , no correlation was found between dynamic propensity and simple structural measures such as local free volume or local composition @xcite . in this paper
we apply the idea of dynamic propensity to a kinetically constrained model ( kcm ) of a glass former , the two - vacancy assisted triangular lattice gas @xcite , or ( 2)-tlg . we show that the heterogeneity in the dynamics can be described by an appropriately defined propensity field .
we find that propensity is not correlated to simple local configurational properties such as local density .
instead we show that a non - local structural property , the connectedness of particles to clusters of vacancies , serves as a good predictor of dynamic propensity .
we also discuss the behaviour of cross - correlations between configurations in the iso - configurational ensemble and its relation to the dynamic phase separation associated with dynamic heterogeneity .
we consider the lattice gas model introduced by jckle and krnig @xcite .
this model is itself a variant of the constrained lattice models proposed by kob and andersen @xcite .
it consists of a set of particles that move on a two - dimensional lattice of triangular geometry .
there are no static correlations between particles and at most each site can only hold one particle at a time . any particle on the lattice can only move to one of its six nearest neighbour sites if the following rules are satisfied : ( i ) the target site is unoccupied and ( ii ) both the two mutual nearest neighbours of the initial and target site are also empty , see fig .
[ tlg - schematic ] .
the model is often referred to as that of two - vacancy assisted hopping on the triangular lattice , or more simply the ( 2)-tlg .
the absence of any static correlations allows initial configurations to be constructed by randomly placing particles into the lattice until the desired density is reached .
although not intended to represent a physical system the dynamical rules can be interpreted as the steric constraint on particle motion within a dense fluid . for increasing particle density the model shows a rapid dynamical slowdown and clear dynamical heterogeneity @xcite .
the dynamics of the model are indicative of a fragile glass former @xcite . for our model
we choose to define the propensity of particle @xmath0 as its root - mean squared displacement , @xmath1 . with this choice propensity
has units of length ( rather than length squared as in @xcite ) .
the average is over all trajectories starting from the same initial configuration using lattice gas monte carlo ( mc ) dynamics ( or continuous time mc for high densities @xcite ) .
that is , for a given initial condition our propensity ensemble is that of all possible randomly attempted particle moves .
this ensures that the dynamics of the system has opportunity to proceed via a different route during each run : fig .
[ trajectories ] shows the resulting particle displacements for three different trajectories starting from the same @xmath2 initial configuration .
the propensity ensemble as defined here is the analog in the mc case to that of randomized momenta used in md simulations @xcite .
[ propensity - maps ] shows the spatial distribution of propensity at four different particle densities @xmath3 .
the times for each density is the relaxation time @xmath4 extracted from the persistence function @xcite .
the data is represented as a contour plot .
each propensity map has been averaged over @xmath5 independent trajectories with a lattice size of @xmath6 sites .
the relaxation time at @xmath7 is approximately @xmath8 times larger than of the lowest density shown , @xmath9 .
a comparison between the four panels shows the distribution of propensity becoming increasingly spatially heterogeneous as the density is increased .
high propensity regions become more localised in nature whilst domains of low propensity are observed to grow in size .
this is similar to what is observed in atomistic systems as temperature is decreased @xcite .
it is also informative to analyse the distribution of particle propensities as a function of density , fig.[propensity - distribution ] .
this is analogous to the van hove function .
the figure shows that the distribution widens significantly for increasing particle density .
interestingly , at high density the distributions appear remarkably flat over a large range of propensities .
this suggests that for these time scales , the curve is not well fit by two gaussian profiles ( or a delta and a gaussian ) , one describing the dynamics of low propensity particles , the other characteristic of high propensity behaviour . as argued in ref .
@xcite , this is a sensible approximation at times either much smaller or much larger than @xmath10 . for the times comparable to @xmath10 shown here the overall distribution appears to be that of a convolution of gaussian functions corresponding to particles which began diffusing at all times between 0 and @xmath11 @xcite
this seems to be an important feature of transport decoupling which deserves further exploration .
we next try to ascertain what aspect of the initial structure , if any , is responsible for the heterogeneity observed in the spatial distribution of propensity . as a starting point
it is useful to first consider a simple local property of the system such as the local density . in any instantaneous configuration the ability of a particle in the ( 2)-tlg to make a single step
is entirely determined by its local neighbourhood as this governs the dynamic rules of the system .
those particles which have a higher number of vacancies amongst their nearest neighbours will be more likely to move .
if local density does correlate well then one would expect the local neighbourhood of high propensity particles to be markedly different from those for which the propensity is low .
+ fig .
[ vacancy - distribution ] shows the distribution of vacancies amongst nearest neighbours for particles within the lowest 5% propensity , nearest neighbours for particles within the highest 5% propensity , nearest and next - nearest neighbours for particles within the lowest 5% propensity , nearest and next - nearest neighbours for particles within the highest 5% propensity .
it is clear that there is nothing special about the local environment of the highest propensity particles , as the distribution of vacancies appears to exactly match the low propensity curve , even when coarse - grained over several particle distances .
this is again analogous to what was found in atomistic simulations @xcite . in a system like the ( 2)-tlg , for which the dynamics are highly collective , many cooperative rearrangements are required in order for a particle to undergo a significant displacement . as such
it is perhaps unsurprising that local density fails to provide a suitable prediction of the long time propensity for motion . in order to provide a more accurate prediction of the spatial distribution one needs to consider some non - local feature of the initial structure .
interestingly , in the less cooperative ( 1)-tlg @xcite , where dynamics proceeds through diffusion of vacancy pairs , we find a similar lack of correlation between propensity and structure as that given by fig.[vacancy - distribution ] .
it is not individual vacancies but rather clusters of vacancies that help facilitate the rearrangement of particles in some collective manner @xcite . considering this aspect of the dynamics we choose to classify particles in the initial configuration according to the size of vacancy clusters to which they are immediately connected .
for example , although a particle may only have a single vacancy amongst its nearest neighbours this vacancy could itself be part of a cluster of appreciable size .
a feature of this nature would not be picked up by any localised measure . using a hoshen - kopelman algorithm @xcite we identify and label all distinct vacancy clusters within a given initial configuration . following the labeling procedure it is trivial to determine the size of each cluster .
each particle is then assigned a value equal to the sum of the vacancy cluster sizes to which it is connected .
we define this value as the cluster connectivity .
.,title="fig : " ] + fig .
[ cluster - maps ] shows two propensity maps at the highest density we studied , @xmath2 .
highlighted in each panel ( black circles ) are those particles whose cluster connectivities are above a chosen threshold ( highest 10% in this case ) .
these particles typically form closed rings around a large vacancy cluster .
it is clear that the most highly connected particles correlate extremely well with those particles showing the highest propensity .
this is not only true for the highest density configurations , for which the propensity distribution is most heterogeneous , a good agreement is also observed at lower density . in fig .
[ cluster - maps-2 ] we show the same for densities @xmath12 and @xmath9 ( the threshold for @xmath12 is 10% , and for @xmath9 is 50% ) . for the lower density of @xmath9
the high propensity regions can almost be entirely described by the particles whose cluster connectivity falls into the top 50% of the system . in fig .
[ cluster - maps-1tlg ] we show a similar propensity map for the ( 1)-tlg at a high density of particles . while less striking , the behaviour is analogous to that of the ( 2)-tlg : the propensity is heterogeneously distributed in space , and regions of high propensity correlate with highly connected clusters .
the result above is also apparent when analysing the distribution of propensity for particles with the minimum and maximum cluster connectivity , as shown by fig .
[ propensity - clusters ] for the ( 2)-tlg .
in contrast to the local density measure , the two distributions are clearly different , the particles with minimum connectivity showing a peak at low propensity whereas the peak is shifted to higher propensity for particles with high connectivity .
our results here are another indication of the difficulty in trying to connect dynamics to structure in atomistic systems . even in the simple systems we consider , like the ( 2)-tlg and ( 1)-tlg
, one has to work pretty hard to obtain any significant correlation between a structural measure and the observed dynamics , and the success has much to do with the detailed prior knowledge of the dynamics in these models . devising a similar cluster scheme which is of any use in atomistic models appears to be a very complicated task .
an important issue to understand is the time dependence of the propensity . fig .
[ propensity - evolution ] shows three propensity maps created using an initial configuration of @xmath13 sites at density @xmath2 .
the three panels show the distribution of propensity at different time scales , ( a ) @xmath14 , ( b ) @xmath15 and ( c ) @xmath16 . here
we do not use a contour plot but instead construct the distribution from circles centered about a particle s position in the initial configuration with a radius equal to its propensity .
filled circles indicate the particles whose cluster connectivity is within the top 5% of the system . comparing the figures it is evident that the highly connected particles act as seeds for mobility within the system with propensity circles seen to emanate outwards from these hot spots . in this sense ,
propensity at shorter times is a good predictor of propensity at longer times .
while clearly in our system there is no such thing as a local debye - waller factor , this spatial correlation between short and long time dynamics does resemble the one found in @xcite . as the system evolves
some interesting features can be seen in the propensity maps . in cases where two or more large vacancy clusters are in close proximity one may observe the individual clusters to coalesce .
this is often accompanied by a sudden burst of motion at the contact point as indicated by a region of newly mobile small circles .
the propensity of these particles may then rapidly grow as motion persists in the region .
this behaviour is consistent with the shape of the propensity distributions shown in fig.[propensity - distribution ] . examining the evolution of the propensity field sheds light on the nature of the ( 2)-tlg dynamics .
as suggested earlier it is the clusters of vacancies that help facilitate the cooperative rearrangement of particles within the system . from this
we find highly connected particles which themselves form clusters that move , coalescing and fragmenting into clusters of different shape and size .
it appears that these clusters underly the dynamics and are in effect the essential excitations for the model .
the initial configuration represents only a single instantaneous indication of the system s structure . by labeling particles according to their cluster connectivity
we only manage to capture the hot spots at a particular moment in time . by analysing the evolution of the most highly connected particles we should be able to track the motion of the excitations throughout the lattice .
note this is not the same as simply looking at the motion of vacancies .
the motion of one of the clusters is illustrated in fig.[cluster - motion ] .
here we show the position of the particles with the top 5% connectivity at three different times .
highlighted in the three panels is the motion of two nearby clusters , which are seen to come together and eventually coalesce . fig .
[ cluster - motion ] shows two characteristic lengthscales , the typical cluster size and the typical distance between them .
this resembles the super - defect scenario of @xcite .
one difference however , at least superficially , appears to be the dynamics : super - defects are supposed to be free to propagate ; the clusters we find have pinned end - points , they fluctuate and may combine to form larger clusters , but are not free to move away . in any case , it would be worth exploring the connection to super - defects further .
we end by considering time correlations and cross - correlations or overlaps within the propensity ensemble .
the normalized autocorrelation function is @xmath17 , where @xmath18 specifies the site occupancy , and @xmath19 indicates connected part of correlations , @xmath20 . for large enough systems
@xmath21 calculated in the propensity ensemble coincides with the standard autocorrelation function .
a second interesting function is the cross - correlation or overlap @xmath22 between two different configurations at time @xmath11 in the propensity ensemble , @xmath23 , where @xmath24 indicate different trajectories with identical initial conditions .
figure [ q](a ) shows @xmath21 and @xmath22 for the ( 2)-tlg .
as expected , @xmath25 for all @xmath11 , i.e. , the distance between initial and final configurations along a trajectory is always larger than that between two different final states .
in fact , for any reversible dynamics @xmath26 as long as the initial conditions are in equilibrium ( for a discussion of the behaviour of @xmath27 in aging see e.g. @xcite ) .
interestingly , when @xmath21 is much slower than an exponential , as in the present case where it appears to be logarithmic , the ratio @xmath28 remains almost constant at long times .
this is to be contrasted to the simple non - interacting lattice gas , where all decorrelation is through the diffusion of vacancies , and for which @xmath29 .
the above illustrates how trajectories within the propensity ensemble are correlated through their common initial conditions .
the sensitivity to the initial state can be tested by the response to perturbations in the initial conditions .
consider the case of two trajectories @xmath30 and @xmath31 whose starting configurations differ only in the location of a fraction @xmath32 of their vacancies .
we denote by @xmath33 their normalized overlap . in the case of a non - constrained lattice gas
we have @xmath34 , i.e. , the initial distance is preserved in the trajectories at all times . figs .
[ q](b ) and ( c ) show the overlaps for the ( 2)-tlg . in this case
the distance between configurations increases .
the corresponding response , @xmath35 , shown in the insets , is non - monotonic and peaks at times around the relaxation time .
as argued in @xcite , sensitivity to initial conditions should be a signature of the dynamic phase transition which underlies dynamic heterogeneity in facilitated models in general : dynamics takes place very close to phase coexistence between an active phase , where motion is plentiful , and an inactive phase , where motion is scarce .
due to this phase equilibrium , initial conditions play an important role since they act as boundary fields in space - time .
dynamic propensity appears to be a useful tool to study their relevance . and
propensity overlap @xmath22 in the ( 2)-tlg for two different densities @xmath3 ( main panel and inset ) .
the dashed line is the ratio @xmath28 .
( b ) overlaps @xmath33 for different values of @xmath32 at density @xmath12 .
the inset shows the response @xmath36 .
( c ) same as ( b ) but for density @xmath2 .
, title="fig : " ] and propensity overlap @xmath22 in the ( 2)-tlg for two different densities @xmath3 ( main panel and inset ) .
the dashed line is the ratio @xmath28 .
( b ) overlaps @xmath33 for different values of @xmath32 at density @xmath12 .
the inset shows the response @xmath36 .
( c ) same as ( b ) but for density @xmath2 .
, title="fig : " ] and propensity overlap @xmath22 in the ( 2)-tlg for two different densities @xmath3 ( main panel and inset ) .
the dashed line is the ratio @xmath28 .
( b ) overlaps @xmath33 for different values of @xmath32 at density @xmath12 .
the inset shows the response @xmath36 .
( c ) same as ( b ) but for density @xmath2 .
, title="fig : " ] we thank peter harrowell for correspondence .
this work was supported by epsrc grants gr / r83712/01 , gr / s54074/01 , and university of nottingham grant fef3024 .
99 for reviews see : h. sillescu , j. non - cryst .
solids * 243 * , 81 ( 1999 ) ; m.d .
ediger , annu .
phys . chem . *
51 * , 99 ( 2000 ) ; s.c .
glotzer , j. non - cryst .
solids , * 274 * , 342 ( 2000 ) ; r. richert , j. phys . condens .
matter * 14 * , r703 ( 2002 ) ; h. c. andersen , proc .
u. s. a. * 102 * , 6686 ( 2005 ) . for recent results see , for example , j. kurchan , l. peliti and m. sellito , europhys .
lett . * 39 * , 365 ( 1997 ) ; e. marinari and e. pitard , euro .
lett . * 69 * , 235 ( 2005 ) ; a. lawlor et al .
e * 72 * , 21401 ( 2005 ) . | we apply the concept of dynamic propensity to a simple kinetically constrained model of glass formers , the two - vacancy assisted triangular lattice gas , or ( 2)-tlg .
we find that the propensity field , defined in our case as the local root - mean square displacement averaged over the ensemble of trajectories with identical initial configurations , is a good measure of dynamical heterogeneity .
this suggests a configurational origin for spatial fluctuations of the dynamics , but just as in the case of atomistic systems , we find that propensity is not correlated to any simple structural property .
we show instead that certain extended clusters of particles connected to vacancies correlate well with propensity , indicating that these are the fundamental excitations of the ( 2)-tlg .
we also discuss time - correlations and the correlation between configurations within the propensity ensemble . |
this work was supported by the national natural science foundation of china under grant no . 11104159 , scientific research foundation of nanjing university of posts and telecommunications under grant no .
ny211008 , university natural science research foundation of jiangsu province under grant no .
11kja510002 , the open research fund of key lab of broadband wireless communication and sensor network technology ( nanjing university of posts and telecommunications ) , ministry of education , china , and a project funded by the priority academic program development of jiangsu higher education institutions .
f. g. deng , x. h. li , c. y. li , p. zhou , and h. y. zhou , phys .
a * 72 * , 044301 ( 2005 ) ; f. g. deng , x. h. li , c. y. li , p. zhou , and h. y. zhou , europ .
j. d * 39 * , 459 ( 2006 ) ; x. h. li , p. zhou , c. y. li , h. y. zhou , and f. g. deng , j. phys .
b * 39 * , 1975 ( 2006 )
. y. b. sheng , f. g. deng , and h. y. zhou , phys .
a * 77 * , 042308 ( 2008 ) ; y. b. sheng , and f. g. deng , phys .
a * 81 * , 032307 ( 2010 ) ; y. b. sheng , and f. g. deng , phys .
a * 82 * , 044305 ( 2010 ) ; y. b. sheng , f. g. deng , and h. y. zhou , europ .
j. d * 55 * , 235 ( 2009 ) ; y. b. sheng , f. g. deng , and g. l. long , phys .
lett . a * 375 * , 396 ( 2011 ) . | we present an efficient entanglement concentration protocol ( ecp ) for mobile electrons with charge detection .
this protocol is quite different from other ecps for one can obtain a maximally entangled pair from a pair of less - entangled state and a single mobile electron with a certain probability . with the help of charge detection , it can be repeated to reach a higher success probability .
it also does not need to know the coefficient of the original less - entangled states .
all these advantages may make this protocol useful in current distributed quantum information processing .
entanglement plays an important role in the current quantum communication @xcite and distributed quantum information processing field @xcite . for most of the practical quantum communication and computation protocols ,
people need to share a maximally entangled state with each other .
however , the entanglement resource is fragile , for the maximally entangled state may be degraded into a mixed state or become a less - entangled state when it interacts with the noisy environment .
people usually resort to the entanglement purification @xcite to increase the fidelity of the mixed state and the entanglement concentration @xcite which will be detailed here to recover the less - entangled state to a maximally entangled state .
currently , most of the protocols for purification and concentration are focused on optical systems @xcite , for during the transmission , the photons have weak interaction with the environment . on the other hand
, there is another candidate for the flying qubit , that is the mobile electron .
a strong interaction between different electrons makes them feasible to interact flying electron spins with other solid electron spins , since the coulomb interaction between each electrons is strongly screened . in the recent years
, the investigation about the flying electron qubits becomes an active study area @xcite . in 2004 , beenakker _
et al .
_ broke through the obstacle of the no - go theorem with the help of charge detection and constructed a controlled - not(cnot ) gate using beam splitters and spin rotations near deterministically @xcite . with the help of charge detector
, people can construct the charge qubit @xcite , perform entanglement purification @xcite , construct entangled spins @xcite , and prepare a multipartite entanglement analyzer and cluster states @xcite . especially , the flying qubit can also be used in the one dimensional system and create entanglement between two distant matter qubits .
recently , matsuzaki and jefferson proposed a protocol for distributed quantum information processing with mobile electrons @xcite . in their protocol , they used mobile electron spins as the mediators of the interaction between the static qubits at each node and finally created a high quality entanglement between each node .
unfortunately , the distributed quantum entanglement may also be degraded into the less - entangled state when it is coupled with the noisy environment .
on the other hand , with current technology , it is difficult to operate each flying quibts perfectly , which can also lead the ideal maximally entangled state to be degraded @xcite . and reflect spin down @xmath0 .
the charge detector ( c ) can distinguish the charge number 1 from 0 and 2 , but can not distinguish the number 0 and 2 .
d is the detector.,width=302 ] entanglement concentration is a powerful tool to recover a less - entangled state into a maximally entangled one with only local operation and classical communication .
the first entanglement concentration protocol ( ecp ) was proposed by bennett _
et al . _ in 1996 @xcite .
this method is called schmidt projection method , in which they need the collective measurement and need to know the exact coefficient of the initial entangled state . in 2001 , zhao _ et al . _ and yamamoto _ et al . _
proposed two similar simplified ecps with linear optical elements based on the schmidt projection method respectively .
both ecps have been realized experimentally @xcite . in 2009
, we proposed an ecp based on electrons @xcite .
almost all the current ecps need two pairs of less - entangled states , but after performing each ecp , at most one pair of maximally entangled state can be obtained with a certain success probability .
actually , using two copies of less - entangled pairs to obtain one maximally entangled pair is not the optimal way . in this paper , we show that we can perform the ecp with the same success probability by using only one pair of less - entangled state and a single electron .
compared with the previous ecps , this protocol requires less less - entangled resources . moreover , analogized with the ref .
@xcite , we adopt the charge detectors and polarization beam splitter ( pbs ) to reconstruct our protocol , which makes it have a higher success probability than those protocols with linear optics .
this protocol can also be used to concentrate the multipartite entangled system .
all these advantages may make this protocol more useful in current quantum communication and distributed quantum information processing . before we start to explain our ecp
, we first introduce the charge detector ( c ) which is a key element shown in fig .
1 . the charge detector can distinguish the occupation number 1 from the occupation number 0 and 2 .
however , it can not distinguish between the occupation number 0 and 2 @xcite , so in both two cases , we define the charge detector will show 0 for simple . in fig.1 , we suppose the source @xmath1 emits a pair of less - entangled state of the form @xmath2 . @xmath3 and @xmath0 are spin up and spin down respectively .
meanwhile , the source @xmath4 emits a single electron to bob of the form @xmath5 and @xmath6 and alice only receives one electron from @xmath7 .
bob first performs a bit - flip operation and makes @xmath8 become @xmath9 then bob lets his two electrons pass through the pbs , which fully transmits @xmath3 and reflects @xmath0 .
the @xmath10 becomes @xmath11 from above equation , one can find that the item @xmath12 means that the two electrons in bob s location are both in the spatial mode @xmath13 while the item @xmath14 means that the two electrons are both in the mode @xmath15 .
however , both the items @xmath16 and @xmath17 mean that the two electrons are in the modes @xmath15 and @xmath13 , respectively .
therefore , the charge detector will show 0 if the original state collapses to @xmath12 or @xmath14 , and it will show 1 if the original state collapses to @xmath18 the probability of obtaining the state of eq .
( [ threephoton ] ) is @xmath19 . obviously , it is the three - electron maximally entangled state .
it is easy to get a two - electron maximally entangled state from eq .
( [ threephoton ] ) .
bob needs to perform a hadamard operation on his electron in the mode @xmath13 .
it makes @xmath20 ) becomes @xmath21\nonumber\\ & = & \frac{1}{2}[(|\uparrow\rangle_{a1}|\uparrow\rangle_{c1}+|\downarrow\rangle_{a1}|\downarrow\rangle_{c1})|\uparrow\rangle_{c2}\nonumber\\ & + & ( |\uparrow\rangle_{a1}|\uparrow\rangle_{c1}-|\downarrow\rangle_{a1}|\downarrow\rangle_{c1})|\downarrow\rangle_{c2}].\end{aligned}\ ] ] then the last step for bob is to measure the spin in the basis @xmath22 . from above equation , if the measurement result is @xmath23 , the electron pair in the modes @xmath24 will become @xmath25 , the electron pair in the modes @xmath24 will become @xmath26 .
if they share the pair @xmath27 , bob will tell alice that the protocol is successful and asks alice to retain her electron . in this way
, they can share a maximally entangled state from a less - entangled state with the success probability of @xmath19 . from above description , bob chooses the case that the charge detector s result is 1 and discards the case of 0 .
it is essentially the partially parity check gate which picks up the even parity states @xmath28 and @xmath29 but discards the odd parity states @xmath30 and @xmath31 . in this case , the function of pbs is similar as it is in the optical systems @xcite . in refs .
@xcite , they need to check that both the output modes of the optical pbs contain exactly and only contain one photon .
it is so called the post - selection principle . therefore , even if they successfully perform their ecps , the maximally entangled state would be destroyed by the single photon detector .
the maximally entangled photon pair can not be remained for further application . in this protocol
, bob can judge the successful case from the charge detection results .
the charge qubit carries both the spin degree of freedom and the charge degree of freedom . as charge and spin
are commute and a measurement of charge leaves the spin qubit unaffected , the charge detection does not affect the entangled state of the electrons .
actually , the charge detector and pbs are more powerful than it has been described above , for the discarded items are essentially the lesser - entangled state which can be reconcentrated in a second step . that is to say , we can not only pick up the even parity states , but also pick up the odd parity states . in fig .
2 , we add another pbs say pbs@xmath32 to reconstruct our ecp .
we denote the whole setup p gate shown in fig .
2 . if the charge detector s result is 0 , the original state will collapse to @xmath33 it means that after passing through the pbs@xmath34 , both the two electrons are in the same spatial mode , while with the help of pbs@xmath32 , they are coupled into latexmath:[\[\begin{aligned } it means that the two electrons in bob s location are in the different spatial modes , say @xmath15 and @xmath13 , respectively . then after performing the hadamard operation and measuring the spin of the electron on the mode @xmath13 , they can get latexmath:[\[\begin{aligned } if the measurement result is @xmath3 .
they can get latexmath:[\[\begin{aligned } if the measurement result is @xmath0 . both eqs .
( [ less1 ] ) and ( [ less2 ] ) are lesser - entangled states .
they can be reconcentrated in the next step . briefly speaking ,
if they obtain @xmath38 , bob needs to choose another single electron of the form @xmath39 , it becomes @xmath40 obviously , if the charge detection c=1 , they can obtain the same three - electron state as described in eq .
( [ threephoton ] ) with a probability of @xmath41 .
otherwise , if c=0 , the remained state is a lesser - entangled state and can be reconcentrated in a third round . in this way
, they can repeat this protocol to get a higher success probability than other protocols .
particles in the multipartite ghz state from the source @xmath1 are sent to to @xmath42 parties , say , alice , bob , charlie , etc .
the source @xmath4 also emits a single electron to bob .
the p is the parity check gate shown in fig.2 .
it comprises the pbs@xmath34 , charge detector c and pbs@xmath32 . only bob needs to perform this concentration.,width=302 ]
it is straightforward to extend this protocol to multi - partite pure entangled state systems .
an @xmath42-electron less - entangled system can be described as latexmath:[\[\begin{aligned } the @xmath42 electrons are emitted from @xmath1 and sent to @xmath42 parties , say , alice , bob , charlie , etc , as shown in fig .
3 .
alice gets the electron of number 1 in the spatial mode @xmath7 .
bob gets number 2 in the spatial mode @xmath44 and charlie gets the number 3 , etc .
the source of @xmath4 also emits a single electron to bob with the same form of eq .
( [ single ] ) . after rotating it by @xmath45
, the composite system can be described as @xmath46 subsequently , the electrons in the spatial modes @xmath44 and @xmath47 in bob s location pass through the @xmath48 gate . if the charge detector s result is @xmath49 , the eq .
( [ multi ] ) will collapse to @xmath50 with the probability of @xmath19 .
finally , bob performs a hardamard operation , and measures the electron in the mode @xmath13 in the basis @xmath51 . if the measurement result is @xmath3 , they will get @xmath52 and if the measurement result is @xmath0 , they will get @xmath53 both eqs .
( [ multi1 ] ) and ( [ multi2 ] ) are the @xmath42-electron maximally entangled states . otherwise , if the charge detector s result is @xmath54 , then eq .
( [ multi ] ) becomes @xmath55 after performing the hadamard operation and measuring the electron in c2 mode in the @xmath56 basis , the above state becomes @xmath57 compared with eq .
( [ nparticle ] ) , it is also a multipartite less - entangled state which can be reconcentrated into a maximally entangled state .
+ or - is decided by the measurement result @xmath3 or @xmath0 , respectively .
so far , we have fully described our ecp .
it is interesting to compare this protocol with ref .
@xcite . in ref .
@xcite , we resort two copies of less - entangled pairs to perform the concentration .
we can get one pair of maximally entangled state with the success probability of @xmath19 . in this protocol
, we use only one pair of less - entangled state and a single electron which can reach the same success probability with ref .
@xcite .
moreover , during the whole protocol , only one - way classical communication is required .
that is alice only needs to receive the information from the bob s measurement and to judge whether the protocol is successful or fail .
if the protocol is a failure , alice needs to do nothing .
if the protocol is successful , bob will tell alice the remaining state is @xmath58 or @xmath59 according to his measurement . in previous protocols
@xcite , all of the parties have to participate the whole procedure , to measure their electrons and check their results to each other to judge the remained state if the protocol is successful .
so this protocol is much more simple , especially when it is used to concentration multi - partite system , for only one of the parties needs to perform the protocol and then report his results to others . finally , let us briefly discuss the key ingredient of this protocol here , that is charge detector @xcite .
it has been realized in a two - dimensional electron gas .
it was reported that currently achievable time resolution for charge detection is @xmath60 @xcite . in a semiconductor
it was reported that the resolution required for ballistic electrons is less than 5 ps @xcite .
compared with flying qubit , it may be more practical to use isolated electrons in an array of quantum dots , as pointed by ref . @xcite . in conclusion
, we have proposed an mobile electron ecp based on charge detection .
compared with other protocols , it has several advantages : first , it dose not require the post - selection principle , and can be repeated to reach a higher efficiency than those based on linear optical elements ; second , only one pair of less - entangled state and one - way classical communication are required , which make this protocol more economic and simple than others .
all these advantages make this ecp more useful in current quantum communication and distributed quantum information processing . |
even though `` noise '' usually represents unwanted fluctuations which blur information , noise is on the fundamental level closely connected with the dynamics of a system and it contains variety of information . in this context , there have been many recent studies on a noise spectrum of current in small devices . in ballistic systems ,
the noise power can be suppressed to zero due to pauli exclusion principle@xcite . in diffusive systems ,
the noise power is suppressed to @xmath0 of the classical value@xcite .
there have been attempts to study the effect of electron - electron interaction on the noise .
for example , noise spectrum in 1d systems was calculated from luttinger liquid approximation@xcite . in this paper , we study current fluctuations of ultra - small tunnel junctions made of two pieces of metal separated by a thin insulating barrier .
it is helpful to discuss a few limiting cases .
it is well known that conventional tunnel junctions show ohmic behavior @xmath1 when it is driven by external voltage bias and the current noise spectrum of ohmic junctions follows johnson - nyquist formula@xcite @xmath2 recent advances in fabrication technology have made it possible to construct very small devices .
it was pointed that electron - electron interaction effect becomes important for small devices .
for example , it was argued that a tunnel junction develops tunneling gap in current - voltage characteristic when the temperature is lower than the charging energy of a junction @xmath3 @xcite . the charging energy was predicted to affect noise spectrum as well .
ben - jacob _ et al_@xcite studied fluctuations of current in an open tunnel junction .
they calculated equilibrium noise of a junction and found that the low frequency noise @xmath4 is suppressed to zero in the low temperature limit @xmath5 .
following theoretical predictions on the charging effect in a small tunnel junction , many experimental attempts were made to observe the charging effect or so called coulomb blockade effect .
coulomb blockade effect was verified for multi - junction systems @xcite .
however the clear verification of the effect for a single tunnel junction has not been reported .
it was later realized that the electrical transport properties of a single tunnel junction crucially depends on the electromagnetic environment of a junction which is decided by properties of leads on junction chip@xcite .
current - voltage characteristic was calculated with explicit account of the environment by modeling the environment as linear impedance presented to the junction .
it was shown that the coulomb blockade effect is erased by the quantum fluctuations of charge when the environment has low impedance and that high impedance environment is essential ingredient for the realization of coulomb blockade effect .
the dependence on the electromagnetic environment was experimentally observed@xcite behaviors of single tunnel junctions are decided not only by the ratio of temperature and the charging energy but also by the environment .
it is a goal of this paper to extend the earlier works to include noise spectrum calculation for general environment .
more important goal of this paper is to show that noise measurement in a single tunnel junction is quite nontrivial . for illustration ,
we calculate current fluctuations at the tunneling barrier and fluctuations at the leads .
most theoretical works on noise calculated the current fluctuations at a specific point of a circuit , usually at a potential barrier right in the middle of a junction .
however as pointed out by landauer _
et al_@xcite , conventional noise measurements have no relation with the electrons crossing the potential barrier of the junction .
the noise is measured , for example , through the voltage across a series resistor in a circuit .
noise measured in experiments is not precisely the same quantity as the noise treated in various theoretical calculations .
noise measured in experiments is complicated by the intrinsic noise of noise measuring device . to make comparison with theoretical calculations
, it is usually assumed that the intrinsic noise makes linear contribution and that experiments can achieve agreement with theoretical calculations by subtracting off the intrinsic noise from the measured noise .
we will call this _ linear superposition assumption _ @xcite . in this paper ,
we assume that leads , which is connected to a junction and provides electromagnetic environment , can be used as a measuring device and calculate current fluctuations at leads .
we also calculate current fluctuations at tunneling barrier and compare them with current fluctuations at leads to test the linear superposition assumption . to briefly mention a result of the calculation , we find that the linear superposition assumption holds only when @xmath6 where @xmath7 is the impedance of the environment . in low impedance environment ,
the assumption breaks down due to finite relaxation time of environment .
in high impedance environment , noise becomes quite nontrivial and spectral density at @xmath8 can be reduced by applying small voltage bias .
we will later provide the physical origin of this novel phenomena . in sec.[sec :
formulation ] , we present a formulation of a problem which incorporates the charging energy and the environment .
the formulation allows simple perturbation calculation in the weak tunneling limit . in sec.[sec : itvsir ] , we introduce two current operators .
one corresponds to current at potential barrier of a junction , which we call _ tunneling current_. the other corresponds to current measured at the leads , which we call _ relaxed current_. expectation values of the two operators are evaluated for a few examples of environment . in sec.[sec : tunnelingnoise ] , the noise spectrum of the tunneling current is calculated .
it shows agreement with existing theories in various limiting cases . in sec.[sec : relaxednoise ] , the noise spectrum of the relaxed current is calculated .
it is compared to the noise spectrum of the tunneling current and the difference is identified .
the difference is attributed to the finite relaxation time of environment and nontrivial zero point fluctuations of environmental degree of freedom .
fig.[fig : circuit](a ) presents a schematic diagram of a system which we consider in the paper .
a junction is represented by a capacitor notation with an arrow right beside it to emphasize tunneling .
the junction is connected to the electromagnetic environment and voltage source .
it is assumed that the environment is linear so that it can be modeled as an impedance @xmath7 .
the impedance is determined by properties of leads connected to junction chip . for detailed discussion of environment and the model system in fig.[fig :
circuit](a ) , see ref.@xcite .
we use the same formulation as the one adopted in ref.@xcite . the tunnel junction connected to the environment
is described by the following hamiltonian(for more details about this hamiltonian , see @xcite ) , @xmath9 the first term @xmath10 describes quasi - particle degrees of freedom , @xmath11 where @xmath12 and @xmath13 label the quasi - particle states in the left and the right electrodes , respectively and @xmath14 represents spin degree of freedom .
the second term @xmath15 describes environmental degrees of freedom and the charging energy , @xmath16 \ , , \ ] ] where @xmath17 is the charge on the capacitor plates of the junction with the steady state contribution subtracted off @xmath18 .
@xmath19 is the phase variable across the junction , @xmath20 .
@xmath17 and @xmath19 are conjugate variables describing macroscopic state of the junction and they satisfy the quantization relation , @xmath21 = ie \ , .
\label{eqn : commutation}\ ] ] phase @xmath19 is coupled to a set of harmonic oscillators following the caldeira - leggett description@xcite of the environment .
the third term describes the tunneling between two plates of the capacitor , @xmath22 where @xmath23 is the tunneling matrix element .
it is simple to verify that @xmath24 is a charge lowering operator , @xmath25 because @xmath26 explicitly takes care of the charge increase and decrease , the quasi - particle degrees of freedom and the environment degrees of freedom can be separated so that quasi - particle operators commute with the charge and phase operators .
to calculate noise spectra , we need to construct current operators .
the tunneling current operator can be easily identified from the operator equation of motion , @xmath27 \label{eqn : tcurrent } \\ & = & -{e \over \hbar}\sum_{kq\sigma}\left ( it_{kq}c_{q\sigma}^{\dagger}c_{k\sigma}e^{-i\tilde{\varphi } } + \mbox{h.c . }
\right ) \ , .
\nonumber\end{aligned}\ ] ] this operator represents the net number of electrons tunneled from one plate of the capacitor to the other .
experimentally , @xmath28 corresponds to current measured right at the tunneling barrier(fig.[fig : circuit](b ) ) .
eq.([eqn : tcurrent ] ) is not a unique choice as a current operator and there is another way of constructing a current operator , @xmath29 \nonumber \\ & = & -{\hbar \over e}\sum_{n } { 1 \over l_n } ( \tilde{\varphi}-\varphi_n ) \ , .\end{aligned}\ ] ] for definiteness , we will call @xmath28 _ tunneling current _ and @xmath30 _ relaxed current_. these currents can be related in a simple way , @xmath31 = { d\tilde{q } \over dt } \ , , \ ] ] and from this relation , we interpret @xmath30 as a charge flow at the environment .
fig.[fig : circuit](b ) illustrates the difference between @xmath28 and @xmath30 .
the construction of the relaxed current operator is possible because the environment is explicitly taken into account . because we are interested in the weak tunneling limit , we take @xmath32 as unperturbed hamiltonian and @xmath33 as perturbation . both @xmath10 and @xmath15 are quadratic and the perturbation calculation is straightforward .
current - voltage characteristic can be obtained from the perturbation theory and one obtains @xmath34 where @xmath35 \right\rangle_0 \ , , \ ] ] and @xmath36 .
@xmath37 represents the average over unperturbed state and all operators are written in interaction picture . because the expectation values of the tunneling current and the relaxed current are the same , we will drop the subscript . by explicitly evaluating the average ,
one obtains @xmath38 where @xmath39 \tilde{\varphi}(0 ) \right\rangle_0 + iet/\hbar \right ] \ , .
\label{eqn : probability}\ ] ] eq.([eqn : iv ] ) together with eq.([eqn : probability ] ) agrees with the result of ref.@xcite , which justifies two current operators .
all information of the environment is stored in the phase - phase correlation function @xmath40 \tilde{\varphi}(0 ) \rangle_0 $ ] and from the fluctuation - dissipation theorem , @xmath41 can be related to the impedance @xmath7 , @xmath42 -i\sin\omega t \right\ } \ ,
, \label{eqn : correlation}\ ] ] where @xmath43 and @xmath44 is the resistance quantum @xmath45 .
let us present current - voltage characteristics in a few special cases .
it illustrates the effect of the environment and we later find that @xmath46 is closely related to the characteristic .
firstly , let us take @xmath47 .
then from eq.([eqn : probability ] , [ eqn : correlation ] ) , one finds @xmath48 and the junction shows ohmic behavior . the charging energy plays no role in the low impedance environment .
secondly , let us take @xmath49 and @xmath50 . then from eq.([eqn : probability],[eqn : correlation ] ) , one finds @xmath51 and the characteristic develops a gap and current does not flow for @xmath52 .
this agrees with the prediction of likharev _
et al_@xcite and one realizes that their semiclassical theory becomes valid if a junction is connected to high impedance environment .
finally , let us take ohmic environment of arbitrary resistance , @xmath53@xcite . in this case , the characteristic is rather complicated and we present only the limiting behaviors at zero temperature . with small voltage bias @xmath54 , one finds@xcite @xmath55^{2r / r_q } \
, , \label{eqn : anomaly}\ ] ] where @xmath56 is the euler constant
. note that eq.([eqn : anomaly ] ) smoothly interpolates the characteristics of the low and the high impedance environments . also note that with the finite impedance , the suppression due to coulomb blockade effect is not strong enough to develop well defined gap even at zero temperature . with large voltage bias @xmath57 , one finds@xcite
the characteristic approaches the high impedance result , @xmath58 \ , .\ ] ] the zero temperature current - voltage characteristics are presented in fig.[fig : iv ] . for the experimental measurement of the current - voltage characteristic , see ref.@xcite .
noise spectral density of current or noise spectrum is defined as fourier transform of a current - current correlation function , @xmath59 because we later find the noise spectrum of the tunneling current different from the one of the relaxed current , we will put subscript @xmath60 and @xmath61 to give distinction . calculation of @xmath46 is as straightforward as the calculation of the current - voltage characteristic and one finds @xmath62 structure of eq.([eqn : sitprelim ] ) is similar to the formula for the current - voltage characteristic eq.([eqn : iv ] ) and indeed , @xmath46 can be written in terms of @xmath63 , @xmath64 eq.([eqn : sit ] ) allows one to obtain the noise spectrum of the tunneling current from the current - voltage characteristic .
it is a simple task to verify that eq.([eqn : sit ] ) satisfies the fluctuation - dissipation theorem in equilibrium @xmath65 . to verify it ,
one introduces a coupling of the tunneling current to its conjugate force @xmath66 , @xmath67 and calculates the response function @xmath68 .
then , the fluctuation - dissipation theorem relates the response to the noise spectrum @xmath69 in the following , we consider equilibrium and excess noise for ohmic environment at zero temperature .
finite temperature noise will be discussed only when it allows a simple analytic expression .
let us first study the low impedance limit @xmath70 , where the junction becomes ohmic . in equilibrium , eq.([eqn : sit ] ) reduces to johnson - nyquist noise@xcite ( fig.[fig : sitequil ] ) @xmath71 excess noise is the difference between equilibrium and non - equilibrium noise @xmath72 .
the superscript @xmath73 is employed to denote the excess noise . at zero temperature ,
the excess noise is absent for @xmath74 and for @xmath75 , it shows piecewise linear dependence on frequency ( fig.[fig : sitnonequil](a ) ) , @xmath76 eq.([eqn : lowequilnoise ] ) and eq.([eqn : sitxlow ] ) agree with eq.([eqn : noisedahm ] ) where the charging effect was neglected .
the agreement with the previous study illustrates that the charging effect does not show up in the low impedance environment . with @xmath77
, eq.([eqn : sit ] ) reduces to @xmath78 in equilibrium where @xmath79 \ , .
\label{eqn : p}\ ] ] at high temperature @xmath80 , @xmath46 becomes @xmath81 for @xmath82 , which is the coulomb blockade correction to white noise . at low temperature @xmath5 ,
@xmath46 becomes @xmath83 for @xmath84 .
the coulomb blockade strongly suppresses the noise far below the white noise level in the low temperature ( fig.[fig : sitequil ] ) .
both high and low temperature limiting behaviors agree with earlier calculation of ben - jacob _ et al_@xcite where they calculated the equilibrium noise spectrum of an open junction which is not connected to the external circuit .
ben - jacob _ et al _ s result can be cast into the form of eq.([eqn : sitpe ] ) with @xmath85 replaced by @xmath86^{-1 } \sum_{q}\delta\left(e - e_c(2q+1)\right)\exp\left ( -{q^2 e_c \over k_b t } \right ) \ , .
\label{eqn : pbj}\ ] ] note that @xmath87 is a discrete form of @xmath85 .
@xmath85 and @xmath87 become identical in the high temperature limit @xmath80 whereas they exhibit delicate difference in the low temperature limit @xmath5 .
recalling that @xmath85 is a probability to emit energy @xmath88 to environment@xcite , one finds that eq.([eqn : pbj ] ) implies the quantization of the emitted energy whereas eq.([eqn : p ] ) contains continuous spectrum of the emitted energy .
the charge in each electrode is quantized both in the open junction and in the closed junction with high impedance environment .
hence the energy used to charge the electrodes has only discrete values . in the open junction ,
the whole emitted energy is used to charge the electrodes . in the junction closed by the environment ,
however , the emitted energy is used only in part to charge the electrodes and the rest excites the environmental degrees of freedom .
therefore @xmath85 has continuous spectrum of energy whereas @xmath87 allows only discrete energy values .
in the high impedance environment , the excess noise shows piecewise linear dependence on frequency at zero temperature(fig.[fig : sitnonequil](a , b ) ) . for small voltage bias @xmath52 ,
the excess noise has finite spectral density(fig.[fig : sitnonequil](a ) ) , @xmath89 for @xmath90 .
note that even though current does not flow when @xmath52 , the excess noise does not vanish around @xmath91 .
at higher voltage bias , finite current flows and fig.[fig : sitnonequil](b ) shows the evolution of the excess noise with the voltage .
the excess noise again shows the piecewise linear dependence .
the piecewise linear dependence in both the low and the high impedance environments originate from the narrow energy range of @xmath85 .
one can also obtain @xmath46 in arbitrary ohmic impedance environment @xmath53 . in the low frequency regime @xmath92 ,
the zero temperature equilibrium fluctuations are suppressed by the coulomb blockade effect , @xmath93^{2r / r_q } \ , .\ ] ] however the suppression is not complete even at zero temperature .
the degree of the suppression depends on the impedance and the higher the impedance , the stronger the suppression is . in the high frequency regime @xmath94
, the equilibrium noise eventually converges to the result of the high impedance environment , @xmath95 \ , , \ ] ] and the convergence is faster with the higher impedance ( fig.[fig : sitequil ] ) . with finite ohmic impedance ,
the excess noise no longer exhibits piecewise linear dependence on the frequency because @xmath85 allows broad range of energy .
since the coulomb blockade effect is blurred by the quantum fluctuations of charge in the finite impedance environment , the excess noise is finite in broad frequency range ( fig.[fig : sitnonequil](a ) ) . at zero frequency ,
@xmath96^{2r / r_q } \ , .\ ] ] before we close the discussion of the tunneling current noise spectrum , let us mention one implication of eq.([eqn : sit ] ) . in the presence of finite current , the correlation between charge tunneling events can be estimated from the noise power @xmath97 , which is defined as @xmath98 . if there is no correlation between the tunneling events , the noise power retains its full value @xmath99 .
if there is any kind of correlation which regulates tunneling , it decrease the noise power below its full value .
for example , in a single channel quantum point contact , the fermi correlation is known to reduce the noise power to @xmath100 where @xmath101 is the transmission coefficient @xcite . from eq.([eqn : sit ] ) , one finds the noise power of the tunneling current to have the full value independent of the environment , @xmath102 where the subscript @xmath60 is employed to denote that this is the noise power of the tunneling current .
the full noise power implies that fermi statistics and the charging energy do not give rise to any correlation between tunneling events .
this result is understandable since we are studying the weak tunneling limit and the time interval between two consecutive tunneling events is assumed to be large .
most theoretical works on current fluctuations calculates the noise spectrum from electron crossing of a particular plane in a circuit .
however , as landauer and martin state@xcite ` currents are measured either through the voltage developed across a small series resistor in the circuit , or through the magnetic fields generated by the displacement current .
contrary to occasional perceptions , noise measurements have no relation to the crossing , by carriers , of a particular plane along the circuit . ' to clarify the relation between conventional noise measurements and noise spectrum treated in various theoretical works @xcite , it is necessary to take an explicit account of a measuring device in the formulation .
we assume that leads connected to a junction can be used as a measuring device and calculate current fluctuations at leads .
the relaxed current @xmath30 corresponds to charge flow at leads and therefore @xmath103 corresponds to spectral density of current fluctuations measured at leads .
we will study the difference between @xmath103 and @xmath46 in this section .
no tunneling limit @xmath104 illustrates a crucial difference of @xmath103 and @xmath46 . in this limit
, the tunneling current operator itself vanishes because it is proportional to the tunneling matrix elements .
hence , one finds @xmath105 .
this result is natural since @xmath28 is the electron flow measured right at the tunneling barrier and electrons can not tunnel the barrier at all in the no tunneling limit . on the other hand ,
the relaxed current operator does not depend on the tunneling matrix elements explicitly and in the zero tunneling limit , one finds @xmath106 where @xmath107 the superscript @xmath108 denotes that this is the zeroth order contribution in the perturbation calculation .
note that @xmath109 is a generalization of johnson - nyquist noise@xcite @xmath110 .
@xmath103 measures the fluctuations of the relaxed current in the environment and @xmath103 does not vanish even in the no tunneling limit because of the zero point fluctuations in the environment . with finite tunneling , it is usually assumed that @xmath103 is a linear superposition of the intrinsic noise @xmath109 and the tunneling noise @xmath46 , which we call _ linear superposition assumption_. the assumption is frequently used to make comparison of experiments with theoretical calculations ( for example , see ref.@xcite ) .
explicit calculation of @xmath103 to the second order in the tunneling matrix produces @xmath111 where @xmath112 the superscript @xmath113 in @xmath114 denotes that this is the second order contribution in the perturbation calculation . unlike the intrinsic noise
, @xmath114 depends on the voltage . at zero frequency
, one finds @xmath115 and @xmath116 , so that that the linear superposition assumption holds at zero frequency for arbitrary environment . because @xmath109 vanishes at zero frequency
, one also finds that noise powers of the tunneling current and the relaxed current are always the same @xmath117 . for non - zero frequency , one finds that the assumption holds if @xmath118 and it does not otherwise .
eq.([eqn : valid ] ) has interesting implications .
for example , in the ohmic environment with high impedance @xmath77 where coulomb blockade effect is clear , the excess noise of the tunneling current exhibits interesting feature at frequencies around @xmath119 .
however eq.([eqn : valid ] ) predicts that at this frequency range , the excess noise of the relaxed current is noticeably different from the tunneling excess noise since eq.([eqn : valid ] ) translates to @xmath120(see sec.[subsec : high ] ) .
let us first discuss the equilibrium noise .
due to the fluctuation - dissipation theorem , the equilibrium noise spectrum can be alternatively derived by coupling the relaxed current to its conjugate force @xmath121 and calculating the response function .
this procedure produces @xmath122 it can be verified that eq.([eqn : sir2eq ] ) agrees with eq.([eqn : sir ] ) in equilibrium .
note that though both @xmath103 and @xmath46 satisfy the fluctuation - dissipation theorem , they are still different since they are related to different response functions . to make predictions from eq.([eqn : sir ] ) , one has to evaluate the real part of the retarded green s function @xmath123 .
for uniform tunneling density of states , one finds at zero temperature @xmath124 in the low impedance environment @xmath70 , one finds @xmath125 vanishes at zero temperature and the zero temperature equilibrium noise becomes @xmath126 it can be verified that eq.([eqn : lowequilr ] ) agrees , to the leading order in @xmath127 , with zero temperature equilibrium noise in the presence of a shunt resistor @xmath128 in parallel with a capacitor @xmath129 @xmath130 when the voltage bias is applied , one finds the zero temperature excess noise @xmath131 eq.([eqn : suppression ] ) can be understood if we assume that current fluctuations induced by tunneling is incoherent with equilibrium zero point current
. then the behavior of a junction is essentially classical .
the only effect of the tunneling is to introduce additional charge to plates of capacitor and from a conventional circuit analysis of a capacitor connected to impedance @xmath7@xcite , one finds the current induced by tunneling is@xcite @xmath132 where @xmath133 for @xmath134 and @xmath135 . from the incoherence assumption , one immediately finds @xmath136 which agrees with eq.([eqn : suppression ] ) .
the factor @xmath137 in eq.([eqn : excessrelaxed ] ) originates from the finite relaxation time of the environment ; the relaxation of tunneled charge to the environment is not immediate and it takes characteristic time of @xmath138 .
fig.[fig : excess](a ) shows the relaxed excess noise in comparison to the tunneling excess noise . at zero temperature , there are three relevant energy scales @xmath139 .
fig.[fig : excess](a ) assumes @xmath140 . due to the low impedance of the environment , the junction shows ohmic behavior . the tunneling excess noise(dashed line )
decays linearly with the frequency and its peak width is @xmath141 .
the relaxed excess noise(solid line ) has the same magnitude with the tunneling excess noise at zero frequency .
however , its peak width is given as min@xmath142 and for the case considered in fig.[fig : excess](a ) , it is @xmath143 , making the decay of the relaxed excess noise faster than the tunneling excess noise . in the high impedance environment ,
the junction becomes highly nonlinear and the shunt resistor model can not explain its behavior . to understand the junction behavior in the high impedance environment , the investigation of the zero point fluctuations in the environment is needed . in the no tunneling limit , the environment is decoupled from quasi - particle degrees of freedom .
the environmental hamiltonian is quadratic and therefore it can be diagonalized as a set of independent harmonic oscillators .
each harmonic oscillator shows zero point fluctuations of the magnitude @xmath144 where @xmath145 and @xmath146 are mass and frequency of the oscillator , respectively .
harmonic oscillator with frequency @xmath146 contribute to the spectral density of the relaxed current at @xmath147 and the zeroth order spectral density @xmath148 is a sum of such contributions .
the situation becomes complicated if one turn on tunneling .
the tunneling couples each harmonic oscillator to quasi - particle degrees of freedom and other harmonic oscillators .
therefore the tunneling provides each harmonic oscillator with the coupling to an effective heat reservoir which consists of other harmonic oscillators and quasi - particle degrees of freedom .
the effect of such coupling is twofold .
it provides a noise source and it also introduces dissipation .
the noise source tends to increase fluctuations in , for example , harmonic oscillator displacement variables while the dissipation has the opposite effect .
it is known ( for example , see ref.@xcite ) that the coupling to the reservoir can even reduce the zero point point fluctuations in certain situations .
the relaxed current @xmath30 is linear in the environmental variables .
therefore it can be written as a linear combination of harmonic oscillator displacements and the fluctuations of @xmath30 is directly related to the fluctuations of harmonic oscillators .
a harmonic oscillator coupled to the reservoir is a difficult problem in general .
however in the weak tunneling limit , one can perturbatively calculate the effect of the reservoir .
it is worth mentioning that the voltage bias is a variable describing quasi - particle degrees of freedom
. therefore fluctuations of harmonic oscillators and the relaxed current depend on the voltage bias .
fig.[fig : excess](b ) shows the relaxed excess noise in comparison to the tunneling excess noise at zero temperature in the high impedance ohmic environment @xmath149 .
the relaxed excess noise can be interpreted as change of harmonic oscillator fluctuations by the voltage bias .
it is assumed that @xmath150 .
note that current does not flow because the voltage bias is below the gap @xmath151 .
the tunneling excess noise shows two peaks at @xmath119 . in comparison , the relaxed excess noise(solid line )
becomes _ negative _ at @xmath152 .
the relaxed excess noise is multiplied by @xmath153 to magnify the negative dips .
though the dips are quite shallow , it is important to notice that the dips imply the decrease of the total noise upon the application of the voltage bias .
the dash - dotted line shows @xmath154 .
the dash - dotted line is also multiplied by @xmath153 . at higher voltage bias @xmath155
, the relaxed excess noise shows a peak at @xmath156 and shallow dips at @xmath152 . from fig.[fig :
excess](b ) and the relation @xmath157 , one finds that the dissipative effect is dominant at @xmath152 while the noise source effect is dominant at @xmath158 .
fig.[fig : excess](c ) shows the two excess noises at zero temperature in the ohmic environment @xmath53 .
it is assumed that @xmath159 and @xmath160 . the tunneling excess noise(dashed line )
shows two peaks , while the relaxed excess noise(solid line ) shows a single peak at @xmath156 .
the dash - dotted line shows @xmath161 for comparison .
the dash - dotted line exhibit qualitatively the same feature as @xmath103 though there are quantitative difference .
eq.([eqn : sir ] ) makes one interesting prediction in the low impedance ohmic environment . as illustrated in sec.[subsec : low ] ( fig.[fig : excess](a ) ) , the excess noise measured from the environment is different from the tunneling excess noise and follows eq.([eqn : suppression ] ) .
recently , there was an experiment which measured the excess noise of the quantum point contact in the low impedance environment@xcite .
the voltage bias @xmath162 of the order of @xmath163mv was applied and the excess noise was measured at frequency @xmath164ghz which is much lower than the voltage @xmath165 .
the frequency was chosen to avoid complications due to other noise sources at low frequency such as @xmath166 noise .
the measured excess noise was compared with theoretical calculations of the noise power(for example , ref.@xcite or @xmath167 ) based on the assumption that the non - zero frequency excess noise is essentially the same as the noise power if @xmath165 .
the experiment showed nice agreement with existing theories in ballistic regime .
however in the pinched off regime which corresponds to the weak tunneling limit , it was found that the noise was reduced to about @xmath0 of theoretical predictions .
the quantum point contact in ref.@xcite is induced electrostatistically in the plane of a 2deg embedded in gaas - algaas heterostructure . due to the very different density of charge carriers in metals and semiconductors ,
our calculation on metallic junctions can not be directly applied to analyze the experiment .
however we believe that the difference of @xmath168 and @xmath169 in the low impedance environment is a generic property of the weak tunneling limit because the difference originates from the finite relaxation time of the environment which does not depend on charge carrier density .
we argue that the noise suppression observed in ref.@xcite may have its origin in the different frequency dependence of the relaxed and tunneling excess noise as illustrated in sec.[subsec : low ] .
however , we are unable to establish quantitative agreement of the experiment with eq.([eqn : suppression ] ) because the precise values of @xmath61 and @xmath129 are not available .
more experiment on , for example , the frequency dependence of the suppression factor , is needed .
eq.([eqn : sir ] ) makes another interesting prediction in the high impedance ohmic environment . at frequency @xmath152 , the spectral density @xmath103 measured at the leads _ decreases _ below its equilibrium value when the voltage bias is applied(sec.[subsec : high ] ) , making the relaxed excess noise negative .
usually the spectral density increases when the voltage bias is applied .
therefore the observation of the negative excess noise can serve as a crucial test of the present calculation .
calculation in this paper is restricted to the weak tunneling limit @xmath170 .
it will be interesting to carry out similar study for the weak scattering limit . according to a recent experiment on quantum point contact@xcite ,
the measured excess noises show the deviation from existing theories only in the pinched off regime and as a channel opens(as a transmission coefficient @xmath101 approaches @xmath163 ) , the deviation disappears .
if the deviation is due to the difference between the relaxed excess noise and the tunneling excess noise , the experiment implies that the junction in the weak scattering limit no longer behaves as a conventional capacitor and the difference between the two excess noise vanishes .
though the experiment used junctions in semiconducting heterostructure , we expect metallic junctions show similar behavior at least in the low impedance environment .
also there was a theoretical study of a single tunnel junction using the functional integral technique@xcite .
it was claimed that the junction becomes purely ohmic and the capacitive effect vanishes in the weak scattering limit . in the weak scattering limit
, electron transport is continuous rather than being sporadic as in the weak tunneling limit and we do nt expect a conventional capacitor is a proper description of the junction .
however , further study is required to gain the quantitative understanding in the weak scattering limit .
in summary , we considered current transport through a junction connected to electromagnetic environment in the weak tunneling limit . the charging energy
is taken into account to incorporate the electron - electron interaction effect .
we calculated both the tunneling noise and the relaxed noise for arbitrary environment and showed their dependence on the charging energy and environment .
we found that contrary to conventional expectation , the current through the potential barrier of the junction and the current measured from the environment have different excess noise . in the low impedance environment
, the difference is due to the finite relaxation time of the junction . in the high impedance environment
, additional complication occurs and at frequency @xmath171 , the spectral density of the relaxed current decreases below the equilibrium level when small voltage bias is applied .
we are grateful to m. reznikov for kindly sending us their manuscript before publication .
research of l.l is partly supported by alfred sloan fellowship .
email : [email protected] .
email : [email protected] .
g. b. lesovik , pisma zh .
. fiz . * 49 * , 513 ( 1989 ) .
( jetp lett . *
49 * , 592 ( 1989 ) ) .
b. yurke and g. p. kochanski , phys .
b * 41 * , 8184 ( 1990 ) .
m. bttiker , phys .
lett . * 65 * , 2901 ( 1990 ) . c. w. j. beenakker and m. mttiker , phys .
b * 46 * , 1889 ( 1992 ) . c. de c. chamon , d. e. freed , and x. g. wen , phys .
rev . b*51 * , 2363 ( 1995 ) .
p. fendley , a. w. w. ludwig , and h. saleur , cond - mat/9505031 .
a. j. dahm , a. denenstein , d. n. langenberg , w. h. parker , d. rogovin , and d. j. scalapino , phys .
* 22 * , 1416 ( 1969 ) .
k. k. likharev and a. b. zorin , j. low .
phys . * 59 * , 347 ( 1985 ) .
d. v. averin and k. k. likharev , j. low temp .
phys . * 62 * , 345 ( 1986 ) .
u. geigenmller and g. schn , physica ( amsterdam ) *
152b * , 186 ( 1988 ) . k. k. likharev , ibm j. res . dev .
bf 32 , 144 ( 1988 ) .
d. v. averin and k. k. likharev , in : _ mesoscopic phenomena in solids _
, ed . by b. l. altshuler , p. a. lee , and r. a. webb ( elsevier , amsterdam , 1991 ) , chap . 6 .
e. ben - jacob , e. mottola , and gerd schn , phys .
. lett . * 51 * , 2064 ( 1983 ) .
j. lambe and r. c. jaklevic , phys .
* 22 * , 1371 ( 1969 ) .
t. a. fulton and g. j. dolan , phys .
lett . * 59 * , 109 ( 1987 ) .
p. j. m. van bentum , r. t. m. smokers , and h. van kempen , phys .
* 60 * , 2543 ( 1988 ) .
l. s. kuzmin , p. delsing , t. claeson , and k. k. likharev , phys .
lett . * 62 * , 2539 ( 1989 ) .
r. wilkins , e. ben - jacob , and r. c. jaklevic , phys .
lett . * 63 * , 801 ( 1989 ) .
p. delsing , k. k. likharev , l. s. kuzmin , and t. claeson , phys .
lett . * 63 * , 1180 ( 1989 ) .
l. j. geerligs , v. f. anderegg , c. a. van der jeugd , j. romijin , and j. e. mooij , europhys . lett .
* 10 * , 79 ( 1989 ) .
v. nazarov , pisma zh .
. fiz . * 49 * , 105 ( 1989 ) [ jetp lett . *
49 * , 126 ( 1989 ) ] .
m. h. devoret , d. esteve , h. grabert , g. -l .
ingold , h. pothier , and c. urbina , phys .
* 64 * , 1824 ( 1990 ) .
s. m. girvin , l. i. glazman , m. jonson , d. r. penn , m. d. stiles , phys .
lett . * 64 * , 3183 ( 1990 ) .
a. n. cleland , j. m. schmidt , and j. clarke , phys .
* 64 * , 1565 ( 1990 ) .
r. landauer and th .
martin , physica b*175 * , 167 ( 1991 ) .
f. liefrink , j. i. dijkhuis , m. j. m. de jong , l. w. molenkamp , and h. van houten , phys .
b * 49 * , 14066 ( 1994 ) . m. h. devoret and h. grabert , in : _ single charge tunneling _ , ed . by a. grabert and m. devoret ( plenum , new york , 1992 ) , chap . 1 .
ingold and yu .
v. nazarov , in : _ single charge tunneling _ , ed . by a. grabert and m. devoret ( plenum , new york , 1992 ) , chap .
2 . a. o. caldeira and a. j. leggett , phys .
lett . * 46 * , 211 ( 1981 ) ; ann .
( n.y . ) * 149 * , 374 ( 1983 ) ; * 153 * , 445(e ) ( 1984 ) . in real experiments
, @xmath7 has frequency dependence and usually it decays at high frequency regime .
however , we take ohmic environment for illustration because ohmic environment is simpler to treat and illustrates essential features of real experiments .
b. callen and t. a. welton , phys .
rev . * 83 * , 34 ( 1951 ) .
s. r. e. yang , solid state comm . * 81 * , 375 ( 1992 ) .
the shunt resistor gives only higher order effect and it can be ignored safely in the leading order calculation of the excess noise .
s. chakravarty , phys .
rev . lett . * 49 * , 681 ( 1982 ) .
m. reznikov , m. heiblum , h. shtrikman , and d. mahalu , preprint .
r. brown and e. imnek , phys .
b. * 34 * , 2957 ( 1986 ) . | we study noise spectra of currents through a tunnel junction in weak tunneling limit .
we introduce effective capacitance to take into account the interaction effect and explicitly incorporate the electromagnetic environment of the junction into the formulation .
we study the effect of charging energy and macroscopic environment on noise spectra .
we calculate current fluctuations at tunneling barrier and fluctuations measured at leads .
it is shown that two fluctuations have different noise spectra and the relation between them is nontrivial .
we provide an explanation for the origin of the difference .
experimental implications are discussed . |
the presence of circumstellar disks around ttauri stars has been suspected for a long time , but it is only very recently that these disks were directly detected , using high - angular millimetre imaging .
these images revealed extended gas structures which appeared to be in keplerian rotation around the central object . among the rare detections so far , two disks were found around binary ttauri stars : ggtau ( dutrey , guilloteau & simon 1994 ) and uyaur ( duvert et al . 1998 ) .
the separations of the binaries are 025 and 089 respectively , which correspond to projected physical separations of 35 and 125au at the distance of the taurus star - forming region ( 140pc ) .
in both cases , light scattered off the surface of the disks has been detected afterwards with adaptive optics imaging at near - infrared wavelengthes .
roddier et al .
( 1996 ) found that the ggtau ring has a clumpy appearance and that several radial spokes of material extend from the ring onto the central stars .
the ring is brighter in its northern part , but is detected in all directions .
they interpret this brightness difference as being due to the scattering geometry .
the uyaur case is very different , as close et al .
( 1998 ) only detected the disk on one side of the binary .
furthermore , they found evidences that a `` spiral arm '' splits from the main disk and gets closer to the star .
deconvolution processes were applied in both studies to retrieve the highest spatial resolution allowed by adaptive optics devices , and this may lead to some artifacts in the final images .
more recently , the first visible wavelength images of uyaur were obtained by mnard et al .
( 1999 ) at 600 and 800 nm with hst / wfpc2 .
the psf - subtracted images revealed a more complicated structure that was found by close et al .
( 1998 ) : a large `` clump '' appears to be independent from the disk itself . if true , this implies that the inclination of the system to the line - of - sight is larger than was first thought ( about 60 instead of about 40 ) . to improve our knowledge of these two circumbinary disks ,
we have performed new observations at 1 and 2 of these systems with hst / nicmos .
we used the polarimetric modes , and we obtained both intensity and polarization maps , which do not need to be deconvolved .
the ggtau polarization maps are the first ever obtained of this system , while potter et al .
( 1998 ) already presented a deconvolved j - band polarization map of uyaur which revealed a nice centrosymetric pattern .
polarization maps are powerfull tools to investigate the dust grain properties and the geometry and structure of the disks . in section2 , we summarize our observations and data processing steps , and the maps of both systems are presented and commented in section3 .
section4 describes some implications of our results on the properties of these disks .
the 1 and 2 images were obtained with camera 1 and camera 2 respectively , providing pixel scales of 0043 and 0075 .
both binaries were observed through the three polarizers at each wavelength , during three 96 seconds exposures for each filter .
the regular nicmos data reduction pipeline prooved to be unsatisfying , and we had to re - reduced all data , with specific care to the so - called `` pedestal effect '' , to obtain final images where the sky level is flat all over the detector . to allow clear detections of the disks , it is mandatory to remove the bright stellar point spread funtions ( psfs ) .
we first tried tinytim psfs , but it appeared that their match with the real ones is quite poor , so we turned to a `` natural '' star , i.e. a bright single star observed through the same filters .
the diffraction spikes subtraction , though unperfect , is quite good , and the optical ghosts induced by some polarizers are naturally removed .
some residuals in the core of the psfs , however , are still large , and nothing can be securely detected in the inner 05 at 1 . at 2 ,
some fringing can be seen at separations as large as 35 .
no deconvolution process was applied to our images , which allows an easier interpretation .
the new 1 image of the ggtau ring is presented in fig.1 .
its overall geometry is in good agreement with roddier et al .
( 1996 ) s images , though with a higher signal - to - noise ratio . however , there are some noticeable features .
first , the ring does not appear clumpy in our image .
this property was likely an artifact introduced by the deconvolution process applied to the adaptive optics images .
fitting an ellipse onto the ring , we find a semi - major axis , a position angle and an inclination in excellent agreement with the millimetre results of guilloteau et al .
it is noticeable , however , that this ellipse is not centered on the center of mass of the binary .
our image does not allow us to confirm the existence of the spokes of material discovered by roddier et al .
( 1996 ) , because of the large psf subtraction residuals inside the ring .
finally , a significant east - west asymetry in the northern part of the ring is seen in our intensity map .
the polarization vectors are strikingly well organized in a centrosymetric pattern , which is symetric about the semi - minor axis of the ring .
the brightest part of the ring , which is the closest to the observer , displays a lower polarization level than the faintest side , typically 20% as opposed to 5060% . at 2 ,
the disk is too close to the stars , and the large subtraction residuals prevent us from obtaining a clear image of the ring .
however , we calculated the polarization map at this wavelength and , though the image is strongly dominated by the unpolarized stellar fluxes , a centrosymetric pattern is found in the polarization vectors , with a typical level of 510% , indicating that the intrinsic polarization level of the light scattered by the ring is high .
the morphology of the uyaur circumbinary disk in our new 1 image is in good agreement with mnard et al .
( 1999 ) s optical image , though the former suffer from a poor signal - to - noise ratio .
as an be seen in fig.2 , the disk appears as an unresolved arc to the southwest of the binary at both wavelengthes , while a bright clump to the southeast appears to be unrelated to this structure .
noticeably , the bright arc seems to widen to the west of the binary in our image .
this can be interpreted as the arc breaking into two separate arcs : the disk itself remaining at least 25 away from the stars , and an inner arc , getting closer to the stars .
the latter would correspond to the `` spiral arm '' described by close et al .
a second feature in our map is an inner arc which is much brighter than the disk itslef .
it lies about 1 to the southwest of the secondary .
this may be a psf artifact , but it can be seen using both a natural or a tinytim - built psf . furthermore , its coincidence with a similar arc detected in the wfpc2 images is suggestive ( see fig.2 ) . at 2 ,
the only area which is clearly separated from the psf residuals is found to the southeast of the binary .
it can be traced up to 5 away from the stars .
the back side of the disk remains undetected at both wavelengthes , which provides strong constraints on the dust grain properties .
we note however , that the small arc seen to the northeast of the primary at 1 can be seen in all indivual images . the polarization pattern at 2 is well organized in its southeastern part , with all vectors well aligned , in a fashion consistent with centrosymetric .
the typical polarization level is about 40% . at 1 , however , the picture is much different : though the vectors are basically aligned in the southwestern part of the disk with a typical level of 20% , again in a more or less centrosymetrical pattern , they are quite randomly oriented in the southeastern clump . whether this is due to our reduction pipeline , to the low signal - to - noise ratio or
reflects the intrinsic pattern is unclear .
it must be noted that potter et al .
( 1998 ) found a very different behaviour .
their data reduction process , however , included a deconvolution step , which impact on the polarization is unknwon .
as already pointed out by guilloteau et al .
( 1999 ) , the shift between the apparent center of the ring and the center of mass of the binary is naturally explained by a thick ring geometry .
this is related to the fact that , in the mie theory , forwards scattering is strongly favoured .
hence , most of the light scattered towards the observer comes from the upper part of the disk s inner edge , whose projection onto the sky is not symetric about the physical center of the ring .
the quantitative model proposed by guilloteau et al . ( 1999 ) is a ring with a half - thickness of 60au at its inner radius ( 180au ) and a very sharp edge .
the observed location of the ring appears to be in excellent agreement with the prediction from this model .
roddier et al .
( 1996 ) suggested that the thickness - to - radius ratio was about one tenth at the inner edge of the ring , which seems imcompatible with our results .
the origin of the east - west asymetry is unclear .
it may be due to the presence of two illuminating stars , or be related to the slight asymetry found in the millimetre wavelength image , which itself may reveal internal structures differences .
fitting ellipses on the large southwest unresolved arc both at 1 and 600 nm yields very similar figures and , especially , an inclination to the line - of - sight of about 60 in both cases .
this is much larger than the 42 estimated by close et al .
the reason for that discrepancy is that they assumed that the southeast clump belongs to the disk , which now seems unlikely .
duvert et al .
( 1998 ) pointed out that a larger inclination ( 6070 ) is in better agreement with the millimetre observations .
if the inner arc close to the secondary star proves to be real , it will represent a challenge for theoretical studies , as most of them predict that structures at such a location should be highly unstable due to the interaction with the binary system .
the polarization levels , as well as the front - to - back side flux ratios , are tightly linked to the disk geometry and to the dust grain properties .
our next step in this study is to run monte carlo simulations to investigate these properties .
for instance , we may determine whether the dust grain size distribution in the disks is compatible with that of interstellar grains . in principle , we will also constrain the amount of `` flaring '' in both disks , as well as their geometrical height and optical depth . close ,
l. , dutrey , a. , roddier , f. , guilloteau , s. , roddier , c. , northcott , m. , mnard , f. , duvert , g. , graves , j. , potter , d. 1998 , , 499 , 833 dutrey , a. , guilloteau , s. , simon , m. 1994 , , 286 , 149 duvert , g. , dutrey , a. , guilloteau , s. , mnard , f. , schuster , k. , prato , l. , simon , m. 1998 , , 332 , 867 guilloteau , s. , dutrey , a. , simon , m. 1999 , , 348 , 570 mnard , f. , stapelfeldt , k. , krist , j. , duvert , g. , padgett , d. , burrows , c. 1999 , , 194 , 6811 potter , d. , close , l. , roddier , f. , roddier , c. , graves , j. , northcott , m. 1998 , in eso conf proceedings n56 , astronomy with adaptive optics , ed .
d. bonaccini , 353 roddier , c. , roddier , f. , northcott , m. , graves , j. , jim , k. 1996 , , 463 , 326 | we have obtained new near - infrared images of both ggtau and uyaur circumbinary disks with the polarimetric modes of nicmos aboard the hubble space telescope .
the 1 intensity map of ggtau reveals a more complete elliptical shape than in previous ground - based images .
its eastern and western sides are definitely asymmetric .
our image strongly supports the ring geometry proposed by guilloteau , dutrey & simon ( 1999 ) on the basis of their millimetre interferometry images : a geometrically thick and sharply edged ring surrounding an empty gap around the binary . around uyaur
, we identify structures that are in excellent agreement with the optical images of mnard et al .
( 1999 ) , which confirms that the inclination of the system to the line - of - sight is about 60 degrees .
we also find tentative new structures closer to the stars .
# 1 1.25 in .125 in .25 in |
diffractive events in @xmath0 collisions are characterized by the presence of a leading proton or antiproton which remains intact , and/or a rapidity gap , defined as a pseudorapidity of a particle is defined as @xmath1 , where @xmath2 is the polar angle of the particle with respect to the proton beam direction . ]
region devoid of particles .
diffractive events involving hard processes ( `` hard diffraction '' ) , such as production of high @xmath3 jets ( see fig . [
fig : diff_diagram ] ) , have been studied extensively to understand the nature of the exchanged object , the pomeron , which in qcd is a color singlet entity with vacuum quantum numbers .
one of the most interesting questions in hard diffractive processes is whether or not they obey qcd factorization , in other words , whether the pomeron has a universal , process independent , parton distribution function ( pdf ) .
results on diffractive deep inelastic scattering ( ddis ) from the @xmath4 collider hera show that qcd factorization holds in ddis
. however , single diffractive ( sd ) rates of @xmath5-boson @xcite , dijet @xcite , @xmath6-quark @xcite and @xmath7 @xcite production relative to non - diffractive ones measured at cdf are @xmath8(10 ) lower than expectations from pdfs determined at hera , indicating a severe breakdown of factorization in hard diffraction between tevatron and hera .
the suppression factor at the tevatron relative to hera is approximately equal in magnitude to that measured in soft diffraction cross sections relative to regge theory predictions based on regge factorization .
the suppression relative to predictions based on ddis pdfs is illustrated in fig .
[ fig : sd_jj ] , which shows the `` diffractive structure function '' @xmath9 measured at cdf by using diffractive dijet data with a leading antiproton detected in roman pots @xcite .
the @xmath10 ( integrated over antiproton momentum loss @xmath11 and four momentum transfer squared @xmath12 ) was obtained as a function of @xmath13 , the momentum fraction of the parton in the pomeron , @xmath14 ( @xmath15 is @xmath16-bjorken of the parton in the antiproton , see fig .
[ fig : sd_jj ] ) , by measuring the ratio of diffractive to non - diffractive dijet rates and using the known leading order pdfs of the proton .
the measured suppression of @xmath9 relative to the expectation from the h1 pdfs is approximately equal to that observed in soft diffraction .
cdf has also studied dijet events with a double pomeron exchange ( dpe ) topology ( fig .
[ fig : diff_diagram ] ) using the roman pot trigger sample at @xmath17 gev @xcite . by measuring the ratio of dpe to sd dijet rates ( @xmath18 ) and comparing it with that of sd to nd rates ( @xmath19 ) ,
a breakdown of qcd factorization was observed as a discrepancy of the double ratio @xmath20 from unity .
in run ii , being currently under way , cdf plans to study various topics on diffraction , including @xmath21 and @xmath11 dependence of @xmath9 in sd , gap width dependence of @xmath9 in dpe , production of exclusive dijet , heavy flavor and low mass states in dpe , and dijets with a large gap in - between jets .
two recently installed `` miniplug '' ( mp ) calorimeters cover the region @xmath22 , and 7 stations of scintillation counters , called beam shower counters ( bsc ) , mounted around the beam pipe , extend the coverage to the very forward region of @xmath23 . the roman pots ( rp ) used in run
i were re - installed and are being operated to trigger on leading antiprotons in the kinematic range @xmath24 and @xmath25 gev@xmath26 .
= 3.5 cm = 6.0 cm
triggering on a leading antiproton in the rp in conjunction with at least one calorimeter tower with @xmath27 gev , a study of diffractive dijet events has been performed . from a sample of 352k triggered events , about 15k sd dijet events with dijets of corrected @xmath27 gev in the range @xmath24
were obtained .
the @xmath11 ( fractional momentum loss of antiproton ) was measured by using all calorimeter information . using a non - diffractive dijet sample triggered on the same calorimeter tower requirement ,
the ratio of diffractive to non - diffractive dijet rates was measured as a function of @xmath16-bjorken of the parton in the antiproton , as shown in fig .
[ fig : r_xbj ] .
this figure shows that ( i ) the ratio observed with run ii data in approximately the same kinematic region as in run i reproduces the run i results , and ( ii ) there is no appreciable @xmath11 dependence in the ratio , as already seen in run i. measurement of the @xmath11 dependence at still lower @xmath11 values ( @xmath28 ) is one of our run ii goals and is being currently under study .
preliminary results of the @xmath21 dependence of the ratio , where @xmath21 is defined as the square of average value of the mean dijet @xmath3 , are shown in fig .
[ fig : r_xbj ] .
no significant @xmath21 dependence is observed , indicating that the pomeron evolves with @xmath21 in a similar way as the proton .
for a study of dpe dijets in run ii , a dedicated trigger has been implemented that requires a rapidity gap in the bsc in the outgoing proton direction in addition to the presence of a leading antiproton in the rp and a single calorimeter tower of @xmath27 gev .
the requirement of a bsc gap on the proton side enhances the dpe signal , as can be seen in the two - dimensional lego plot of mp versus bsc hit multiplicity of sd dijet events ( fig .
[ fig : dpe ] ) .
offline , requiring in addition a gap in the proton - side mp , we obtained about 16k dijet events ( about 100 times more data than in run i ) , which are qualitatively consistent with dpe dijets .
figure [ fig : dpe ] ( middle and right ) shows the @xmath3 , mean @xmath29 and azimuthal angle difference @xmath30 of the two leading jets for the dpe candidate events ( points ) .
as seen in run i dpe data , the @xmath3 distributions look similar to those of sd dijets ( histograms ) , while the mean @xmath29 and @xmath30 show that the dpe dijets are more central and more back - to - back .
et al . _ ,
78 * , 2698 ( 1997 ) .
et al . _ ,
79 * , 2636 ( 1997 ) .
t. affolder _
et al . _ ,
* 84 * , 232 ( 2000 ) .
t. affolder _
et al . _ ,
lett . * 87 * , 241802 ( 2001 ) .
t. affolder _
et al . _ ,
lett . * 84 * , 5043 ( 2000 ) .
t. acosta _
et al . _ ,
lett . * 88 * ,
151802 ( 2002 ) . t. affolder _
et al . _ ,
lett . * 85 * , 4215 ( 2000 ) . | we present results on hard diffraction obtained by the cdf collaboration in run ii proton - antiproton collisions at the fermilab tevatron .
run i cdf results on hard diffraction are also reviewed . |
the standard model ( sm ) of particle physics is considered to be an effective theory despite the fact that it has succeeded in describing known experimental data available up to now .
supersymmetry ( susy ) between bosons and fermions at the unification - energy scale is believed to be one of the most promising extension of the sm . among the supersymmetric theories ,
the minimal supersymmetric extension of the sm ( mssm ) is a well studied framework of susy because of its compactness . in the mssm , however , there are many unfixed parameters . for limiting the possible region of the mssm parameters ,
a promising approach is so - called natural susy " . in the framework of the natural susy ,
a light stop with a large a - term and light higgsinos give a solution of fine - tunings in the mssm@xcite .
we consider that using experimental results is the top priority for limiting the mssm parameters , and investigate the possibility of survival of a light stop .
recently , we have found the possible region of the mssm parameters@xcite within the bounds from the experimental results of ( i ) the higgs mass@xcite , ( ii ) the branching ratio , @xmath6@xcite , ( iii ) the muon @xmath1@xcite , ( iv ) the dark matter ( dm ) abundance@xcite , and ( v ) the direct searches for the lighter stop at the lhc@xcite . in the parameter searches , we use suspect2@xcite , susy - hit@xcite , and micromegas@xcite for ( i)(iv ) .
moreover , we have studied indirect searches for susy particles at the ilc by using benchmark points within the possible region of the mssm parameters@xcite .
we have calculated the 1-loop effects of susy particles in the processes , @xmath7 , @xmath8 , @xmath9@xcite , and @xmath10@xcite with the aid of grace / susy - loop@xcite .
our criterion for searching the mssm parameters is that the mssm effects are within the bounds from the following experimental results : 1 . the higgs mass , @xmath11 , 2 . the branching ratio of the rare decay mode of @xmath0-quark , @xmath6 , 3 . the muon anomalous magnetic moment , @xmath12 , 4 . the dm abundance , @xmath13 , 5 .
the lower bounds of the lighter stop mass , @xmath14 in the direct searches at the lhc .
we have performed systematic multidimensional scans in the entire range of the mssm parameter space@xcite , but we present only the results which are consistent with the five observables above .
the experimental value of the higgs mass is measured@xcite as @xmath15 the mssm contribution to the higgs mass mainly depends on the masses of stops , @xmath14 , @xmath16 , the a - term , @xmath17 , the higgsino mass parameter , @xmath18 , and @xmath19@xcite .
an alternative parameter , @xmath20 is useful due to the relation among the parameters , @xmath21 figure [ fig : xtmst1 ] shows the possible contours on the @xmath20@xmath14 plane which is consistent with for @xmath22 and @xmath23 .
the reason for the large value of @xmath18 here is explained in the next subsection . for the allowed mass of the lighter stop , @xmath14 , there are two possibilities as follows : 1
. the light stop scenario , @xmath24 with @xmath25 , 2 .
the heavy stop scenario , @xmath26 with @xmath27 .
@xmath14 plane plotted by green crosses . ] at any allowed point in in figure [ fig : xtmst1 ] , the value of @xmath17 is severely restricted by the higgs mass .
figure [ fig : ldetmu ] shows the dependence of the higgs mass , @xmath11 on @xmath17 .
thus , we obtain a solution , @xmath29 for @xmath30 .
the mssm parameters are also constrained by the experimental value of the branching ratio@xcite , @xmath31 figure [ fig : rdetmu ] shows the dependence of the branching ratio , @xmath6 on @xmath17 .
thus , we obtain a constraint , @xmath32 for @xmath29 . 0.4 and @xmath18 in the light
stop scenario.,title="fig : " ] 0.4 and @xmath18 in the light stop scenario.,title="fig : " ] the muon anomalous magnetic moment , @xmath34 has been accurately measured , thus causes the mssm contribution to be restricted@xcite .
the experimental value , @xmath35 , the sm prediction , @xmath36 , and the difference , @xmath37@xcite are @xmath38 the mssm contribution to @xmath12 depends on the slepton mass , @xmath39 , @xmath19 , @xmath18 , the bino mass , @xmath40 , and the wino mass , @xmath41@xcite . for @xmath22 , @xmath23 , and @xmath42 .
the allowed region is surrounded by the light green lines . ]
figure [ fig : gminus2 ] shows the allowed region on the @xmath39@xmath40 plane for @xmath22 , @xmath23 , and @xmath42 .
it is remarkable that the bino mass is constrained as @xmath43 , thus the lsp is binolike due to @xmath44 .
three red points numbered 1 to 3 are the benchmark points which are defined in section 2.5@xcite .
the lightest susy particle ( lsp ) is a promising candidate for the dm .
we consider the lightest neutralino , @xmath45 as the lsp .
the dm abundance can be explained by the lsp - pair annihilation@xcite , the stau
lsp co - annihilation@xcite , and the stop
lsp co - annihilation@xcite .
there are three scenarios as in table [ scenarios ] by combining the light or the heavy stop with the mass degeneracy of stau lsp or stop lsp .
if the masses of stau , stop , and lsp degenerate , both co - annihilation processes occur .
.three scenarios for the stau mass and the stop mass . [ cols="^,^,^",options="header " , ] the observed value of the dm abundance@xcite , @xmath46 restricts the mssm parameters within the narrow region .
figure [ fig : stopdm ] shows the projection of the mssm parameter scan on the @xmath47@xmath13 plane .
the data are plotted for two values of the stau mass , @xmath48 . .
] figure [ fig : stopstau ] shows the projection of the mssm parameter scan on the @xmath14@xmath49 plane which is consistent with in the light stop scenario for @xmath50 .
the data within the red ellipse are consistent with the five observables as mentioned in the opening paragraph of section 2 . in the allowed region ,
the lighter stop mass and the lighter stau mass are restricted as @xmath51 and @xmath52 .
these mass ranges depend on @xmath40 . there
, however , is a room for the heavy stop scenario if @xmath53 . in the direct searches of susy particles at the lhc
, there has been no evidence of susy particle production up to now .
we , however , still have a possible region on the mssm parameters which has not been excluded by experiments .
figure [ fig : spec ] shows the mass spectra for the benchmark points which we take from the possible region of the mssm parameters corresponding to three scenarios in table [ scenarios ] .
the mssm parameter sets named set1 and set3 are for the light stop scenario , and that named set2 is for the heavy stop scenario .
0.32 0.32 0.32 figure [ fig : lhc ] shows the mass bounds of the lighter stop , @xmath54 from the direct searches at the lhc@xcite .
the two benchmark points plotted in the figure are located outside of the excluded regions .
thus we find that two kinds of the light stop scenarios and the heavy stop scenario still survive .
0.45 , title="fig : " ] 0.45 , title="fig : " ]
for indirect searches of susy particles , susy effects are revealed through 1-loop corrections . there
, however , exist the sm contributions to 1-loop corrections , thus the following quantities are useful@xcite .
@xmath55 . ] for the indirect searches of susy particles , we consider the following processes : @xmath56 where the center - of - mass energy , @xmath57 is determined by the tree level cross sections as in figure [ fig : tree ] and by the technical design of the ilc@xcite . figure [ fig : tautau ] and figure [ fig : bb ] show the angular distributions of @xmath58 in the @xmath2-pair production process and the @xmath3-pair production one , respectively .
it is remarkable that susy effects can be revealed as excess of the sm estimation at the 1-loop level , because the value of @xmath58 is larger than the expected error bars .
for the integrated luminosity , @xmath59 , however , the parameter set of the light stop scenario , set1 can not be discriminated from that of the heavy stop scenario , set2 because of the mass relations , @xmath53 and @xmath60 in both parameter sets .
0.45 . in ( a )
, statistical errors are shown only for the results of set1 where we assumed that the integrated luminosity is @xmath61.,title="fig : " ] 0.45 . in ( a )
, statistical errors are shown only for the results of set1 where we assumed that the integrated luminosity is @xmath61.,title="fig : " ] figure [ fig : tt500 ] shows the angular distribution of @xmath58 in the @xmath4-pair production process .
there is small difference between the @xmath58 s of the two parameter sets in the forward direction of top quark . in the @xmath4-pair production process for the parameter sets , set1 and set2 at @xmath62 .
statistical errors are shown only for the results of set1 where we assumed that the integrated luminosity is @xmath63 , but are negligibly small . ] in the @xmath5 production process for the parameter sets , set1 and set2 at @xmath64 .
statistical errors are shown only for the results of set1 where we assumed that the integrated luminosity is @xmath61 . ]
figure [ fig : zh250 ] shows the angular distribution of @xmath58 in the @xmath5 production process .
there is difference between the @xmath58 s of the two parameter sets , which is larger than the range of error bars .
it is remarkable that the 1-loop corrections of this process are sensitive to the lighter stop mass , @xmath14 .
we have obtained possible mssm parameter sets which are consistent with the bounds from the experimental results of the higgs mass , the rare decay mode of @xmath0-quark , the muon @xmath1 , the dark matter abundance , and the direct searches for the lighter stop at the lhc . for the parameter sets , the virtual effects of mssm in the indirect search processes
are estimated at 1%@xmath65% in the early stage of the ilc , which are larger than the ranges of error bars .
for the purpose of discrimination of the light and heavy stop scenarios , @xmath5 is the most promising process to investigate . | we have found the possible region of parameters of the minimal supersymmetric standard model ( mssm ) within the bounds from the experimental results of the higgs mass , the rare decay mode of @xmath0-quark , the muon @xmath1 , the dark matter abundance , and the direct searches for the lighter stop ( i.e. , one of the supersymmetric partners of top quark ) at the lhc .
we present numerical results of calculations for the one loop effects of supersymmetric particles in the processes of @xmath2 , @xmath3 , @xmath4 , and @xmath5 production at the ilc by using benchmark points within the possible region of the mssm parameters . |
in the main body of the paper , we focus on the dynamics of the skyrmion density as the important quantity to describe skyrmions or antiskyrmions , respectively .
still , these objects have an internal magnetic structure given by the magnetization @xmath14 .
in particular , different kinds of dzyaloshinskii - moriya interactions ( dmi ) may stabilize different types of skyrmions . in this appendix
, we show exemplary magnetic structures for * bulk dmi given by the hamiltonian @xmath98 $ ] which stabilizes bloch - like skyrmions ( cf .
[ fig::bdmi ] and movie `` ` sk_density_vs_time_bulkdmi.avi ` '' ) , * interfacial dmi given by the hamiltonian @xmath99 $ ] which stabilizes nel - like skyrmions ( cf .
[ fig::idmi ] and movie `` ` sk_density_vs_time_interfacialdmi.avi ` '' ) , and , * no dmi with means stabilization neither of bloch - like nor of nel - like skyrmions ( cf .
[ fig::zdmi ] and movie `` ` sk_density_vs_time_zerodmi.avi ` '' ) .
all figures are snapshots taken from movies which we also provide as supplemental material online in the `` other formats '' option on the article s arxiv page . even though the explicit magnetic structures differ for the according dmi , no qualitative changes for the pair creation process , as described in the main article ,
were observed .
+ we have used the same parameter set as before .
the calculations were performed on a @xmath66 square lattice with periodic boundary conditions , external magnetic field @xmath68 , spin velocity @xmath100 m/s , gilbert damping @xmath101 and non - adiabaticity @xmath102 . to create initial fluctuations of the skyrmion density , a tiny modulation to the magnetic field pointing in the @xmath71-direction , i.e. , @xmath72 $ ] and
@xmath73 has been added to the external field .
+ 40ifxundefined [ 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 \doibase http://dx.doi.org/10.1016/0029-5582(62)90775-7 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.74.085308 [ * * , ( ) ] `` , '' in link:\doibase 10.1007/978 - 1 - 4684 - 9449 - 5_6 [ _ _ ] ( , , ) pp . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1016/j.jmaa.2007.01.013 [ * * , ( ) ] @noop * * , ( ) @noop * * ( ) @noop | magnetic skyrmions can be considered as topologically protected quasiparticles . due to their stability , their small size , and the possibility to move them by low electric currents they are promising candidates for spintronic devices . leaving the topological protection untouched ,
it is possible to create skyrmion - antiskyrmion pairs , as long as the total skyrmion number remains unchanged .
we derive a skyrmion equation of motion which reveals how spin - polarized charge currents create skyrmion - antiskyrmion pairs .
it allows to identify general prerequisites for the pair creation process .
we corroborate these general principles by numerical simulations . on a lattice , where topological protection becomes imperfect ,
the antiskyrmion partner of the pairs is annihilated and only the skyrmion survives .
this eventually changes the total skyrmion number and yields a new way of creating and controlling skyrmions .
magnetic skyrmions ( sks ) are vortex - like localized magnetization configurations @xcite which have been predicted @xcite before they were discovered experimentally @xcite in magnetic layers with a strong spin - orbit interaction @xcite . despite their size being much smaller @xcite than , for example ,
that of topologically trivial magnetic bubbles , their thermodynamic stability is amazing @xcite .
this is a consequence of the particular magnetic configuration which can be characterized by a topological sk number @xmath0 .
it can take integer values only and therefore can not be changed continuously @xcite .
this feature protects magnetic sks against typical drawbacks of solid state systems such as disorder or imperfect fabrication @xcite .
together with the property of easy repositioning by rather tiny in - plane electrical currents @xcite , this makes diluted sks attractive candidates for future racetrack memory devices @xcite .
creation of sks has been reported in the vicinity of notches @xcite or by sweeping the back ground magnetic field @xcite . controlled creation and annihilation of individual sks has been demonstrated @xcite . with in - plane electrical currents @xcite the density of sks adjusts to the respective equilibrium value when ,
for example , the strength of the externally applied zeeman magnetic field is varied in time .
this process must overcome topological stability .
the precise mechanism of sk creation remained , however , unexplored . in this work ,
we derive a sk equation of motion which reveals the details of the process how sk - numbers @xmath0 change by an applied in - plane current .
we find this to happen in three steps .
first , a skyrmion - antiskyrmion ( sk - ask ) pair is created @xcite , triggered by small spatial fluctuations of the magnetization .
pair creation does not change the total sk - number @xmath0 , since the sk and the ask have non - zero equal @xmath0 of opposite sign , respectively . then , in a second step , sk and ask get spatially separated by the dynamics we describe below .
the sk equation of motion reveals the relevant terms at work which are not captured by the common thiele approximation @xcite . finally ,
the ask , being no stable solution for a given zeeman- and dzyaloshinsky - moriya interaction strength , decays due to gilbert damping .
it is this last dissipative step which is ultimately responsible for changing @xmath0 .
all general findings are confirmed by extended numerical simulations .
the two - dimensional magnetization configuration @xmath1 of a single current - driven sk evolves in time according to the landau - lifshitz - gilbert ( llg ) equation @xcite @xmath2 where @xmath3 is a normalized vector field .
all interactions of the hamiltonian @xmath4 describing the system are contained in the effective field @xmath5 .
below , in eq .
( [ latticemodel ] ) , we specify the hamiltonian for a lattice model , but its detailed form is not relevant for the following consideration .
@xmath6 already contains the gyromagnetic ratio and we set @xmath7 .
further important parameters are the gilbert damping constant @xmath8 and the non - adiabaticity parameter @xmath9 . in this work
, we focus specifically on the impact of spin - polarized electronic currents @xmath10 @xcite flowing homogeneously in space and time in the magnetic plane with spin polarization @xmath11 and lattice constant @xmath12 , proportional to an electronic charge current density @xmath13 . with the vector field @xmath14
, we then define the sk density @xmath15\;,\ ] ] and the spatial integral which yields the quantized skyrmion number @xmath16 in fact , this homotopy invariant completely characterizes sks and is the degree of the vector field @xmath17 @xcite . for convenience , we take the direction of the spin current @xmath18 as reference direction , @xmath19 and @xmath20 .
while the topological invariant @xmath0 is conserved in time at low energies , the time evolution of @xmath21 describes the current - induced local motion of sks . in particular ,
as discussed below , it also describes the generation or annihilation of sk - ask pairs . to reveal the sk - ask pair creation mechanism ,
we decompose the effective field according to @xmath22 by combining eqs .
( [ eq::llg ] ) , ( [ eq::qxyt ] ) and ( [ eq::basis ] ) , we readily obtain the sk equation of motion @xmath23 with the sk currents [ eq::jsk ] @xmath24\mathbf v_s\label{eq::jskd1}\\ & \qquad-[{\ensuremath { ( \hat{\mathbf v } \cdot\nabla){\ensuremath{\mathbf
n}}}}]^2\mathbf v_{\perp}\big\}\ , , \label{eq::jskd2}\end{aligned}\ ] ] with contributions parallel to the current flow ( @xmath25 ) and perpendicular to it ( @xmath26 ) .
the coefficients read [ eq::jsk_coeff ] @xmath27/(1+\alpha^2)\ , , \label{eq::jt}\\ j_2 = & [ -\alpha+\beta - ( { \ensuremath{b_{\perp 1}}}-\alpha { \ensuremath{b_{\perp 2}}})]/(1+\alpha^2)\ .\label{eq::jd}\end{aligned}\ ] ] the sk equation of motion ( [ eq::eom ] ) resembles a continuity equation which connects the sk `` charge '' density @xmath28 with the sk current density .
we note , however , that conservation of @xmath0 in eq .
( [ eq::q ] ) in the present case is not a consequence of noether s theorem , albeit conserved quantities may still exist for eq .
( [ eq::eom ] ) @xcite under continuous variation of @xmath29 @xcite .
the physical meaning of @xmath30 and @xmath31 becomes apparent when we consider sk in the steady state where @xmath32 and thus @xmath33 . for not too large current densities , no major structural changes of the magnetization
occur and @xmath6 remains parallel to @xmath29 .
then , the perpendicular components @xmath34 vanish and the coefficients @xmath35 and @xmath36 in eq .
( [ eq::jsk_coeff ] ) are constant .
a case of special importance occurs when @xmath37 , which implies that @xmath38 , @xmath39 .
then , @xmath40 , such that the undistorted sk density @xmath28 moves with the velocity @xmath41 , according to eq .
( [ eq::jskt ] ) .
this motivates to call @xmath30 a sk current density .
when @xmath42 ( but still assuming @xmath43 ) , @xmath36 becomes nonzero .
then , we may rewrite eqs .
( [ eq::jskd1],[eq::jskd2 ] ) in the form @xmath44 , with the coefficients @xmath45 and @xmath46 ^ 2/q$ ] .
we can simplify @xmath47 , where @xmath48 is the angle between @xmath49 and @xmath50 .
this term @xmath51 only adds to the contribution of @xmath30 ( though with a different dependence on @xmath8 , @xmath9 , as well as on the shape of the vector field @xmath29 ) to drive sk density and ask density along @xmath52 depending on the explicit angle @xmath48 .
+ crucial for the following is the term @xmath53 .
first , it points perpendicularly to the externally applied spin current and , second , it drives sk- and ask - densities in opposite directions , as it changes sign under the inversion @xmath54 .
we therefore identify the contribution ( [ eq::jskd2 ] ) as being responsible for separating the sk from the ask density resulting in a common sk - ask pair .
the separation takes place perpendicularly to the external current .
actually , this process can be expected to be a common scenario in real materials for sufficiently strong applied charge current densities .
the only further prerequisites are @xmath55 and small spatial fluctuations of the sk density @xmath21 , which also imply finite gradients @xmath49 and @xmath50 and thus a finite @xmath31 .
a finite gradient @xmath50 is , strictly speaking , not necessary for a non - vanishing current @xmath31 [ cf .
eq .
( [ eq::jskd2 ] ) ] . nevertheless , it is important for a finite divergence @xmath56 . only in this case , the skyrmion current can not be gauged away and is physically relevant .
having such a current means that regions of negative and positive sk density @xmath21 are separated .
in addition , regions of opposite signs appear quite naturally in the topologically trivial state @xmath57 , as regions of finite @xmath28 , which we have postulated , have to cancel each other to sum up to zero . ultimately , a sk - ask pair is formed out of these fluctuations when this is energetically favorable .
we note in passing that the detailed motion of asks is typically more complicated than that of sks , since commonly , an isolated ask is not a stationary solution and thus , already for @xmath58 , @xmath6 is clearly not parallel to @xmath29 , which implies that @xmath59 . in the following , we illustrate these general principles for a concrete model realized by the hamiltonian @xcite @xmath60\nonumber\\ & -\sum_{{\ensuremath{\mathbf r}}}\mathbf b_{\mathbf r}\cdot{\ensuremath{\mathbf n_{\mathbf r}}}\nonumber\end{aligned}\ ] ] defined on lattice sites @xmath61 ( unit lattice constant ) in two dimensions .
it supports diluted sks within certain ranges of the parameters .
@xmath62 is the exchange interaction and @xmath63 the dzyaloshinsky - moriya interaction ( dmi ) strength .
we use the values @xmath64 , @xmath65 reported for mnsi @xcite . here , we only discuss a bulk dmi which stabilizes bloch sks .
yet , we have also verified our findings for systems with an interfacial dmi which stabilizes nel sks @xcite .
no qualitative modifications occur and our findings thus apply to both kinds of sks . in the numerical simulations , we use a @xmath66 square lattice with periodic boundary conditions .
depending on the magnitude of the externally controlled zeeman - field @xmath67 , either a helical phase , a sk lattice or the ferromagnetic ( field polarized ) phase is the ground state @xcite .
a field @xmath68 is in fact strong enough to align all magnetic moments , @xmath69 .
then , @xmath21 remains zero everywhere and , according to eqs .
( [ eq::eom ] ) and ( [ eq::jsk ] ) , for all times , since @xmath70 , even at non - zero applied current densities . to realize at least a small initial non - zero sk density @xmath28 , we add a tiny modulation to the magnetic field pointing in the @xmath71-direction , i.e. , @xmath72 $ ] and @xmath73 . as a matter of fact , the precise form of the initial inhomogeneous magnetization configuration is of minor importance .
the time evolution of the system is calculated by solving the llg eq .
( [ eq::llg ] ) by standard advanced numerical methods .
starting from the fully field polarized state @xmath74 , we first let the system accommodate to the additional @xmath75 field , at zero external current .
after this initial equilibration , we switch on the current at @xmath76 and calculate @xmath21 at every time step .
a movie of this evolution is available in the sm @xcite while a selection of snapshots of @xmath28 is shown in fig .
[ fig::snapshots ] .
initially , the very small amplitude @xmath77 of @xmath75 generates a tiny seed sk density of both , positive and negative sign with an overall @xmath57 .
gradually , under the influence of the external current , sk - ask pairs begin to form with growing magnitudes of @xmath28 .
consistent with our theoretical prediction , the sk and ask centers separate in the @xmath71-direction , perpendicular to the external current flow . after its full development , since it is unstable , the ask disappears on a time scale @xmath78 .
thereby its diameter shrinks relatively quickly , eventually below the lattice constant . at this moment ,
@xmath79 abruptly changes by one .
( color online ) .
black solid line : time - dependence of the total sk number @xmath79 .
black dashed line : time - dependence of the total ask number defined as @xmath80 stemming from negative sk densities only .
note that @xmath81 does not need to be integer and that the restriction to lattice points imposes some small , unimportant ambiguity on the precise determination of @xmath21 .
red dashed line : time - dependence of the energy in reference to the initial energy , @xmath82 , per lattice site . before every @xmath0-jump
, @xmath81 gradually increases , accompanied by an increase of the energy which eventually is taken from the external current . ]
this scenario is demonstrated further in fig .
[ fig::q_vs_t ] , where we show the time - dependence of @xmath79 . over large time spans , the total sk number ( black line ) takes integer values , while occasionally @xmath79 jumps to the next integer within a short ( but on the graph resolvable ) transition time .
these transitions are accompanied by sudden rises of the total ask number @xmath80 , a quantity that we define by integrating over negative sk - density only . during the times when @xmath79 stays integer , @xmath83 may decrease gradually with time .
this indicates the gradual creation of sk - ask pairs , their growth and their spatial separation , before the finally isolated , but unstable ask annihilates during a time much shorter than the duration of its creation , as described above .
this initial gradual development of the first sk - ask pair due to a weak spatial inhomogeneity of the zeeman - field is clearly seen in fig .
[ fig::q_vs_t ] . on the other hand ,
as soon as a finite number of sks exist ( after 6300 ps in fig . [ fig::q_vs_t ] )
, their intrinsic inhomogeneous magnetization suffices to facilitate further creation of sk - ask pairs in their surroundings , then even at a homogeneous zeeman - field as we have convinced ourselves independently .
since the system starts very close to the ferromagnetic ground state , the sk creation costs energy .
this energy is pumped into the system by the charge current .
figure [ fig::q_vs_t ] confirms the connection between the increase of the energy and of negative sk - density .
the evolution of sk - ask pairs is overshadowed by the relatively short life time of the ask .
the details of this process are further illustrated by an additional movie @xcite where we set the dmi to zero .
then , neither the sk nor the ask is energetically preferred and the full sk - ask pair evolves in time . the duration for sk creation can be quantified by the time @xmath84 which we define as the time span from the onset of the current flow till the creation of the first sk .
since the ask annihilation is comparably fast , it is sufficient to define @xmath84 by the condition @xmath85 .
this creation time is a combination of the time @xmath86 needed to form a sufficiently large sk - ask pair and the annihilation time @xmath87 of the ask .
since both processes happen at least partially simultaneously , the resulting @xmath84 is not a direct sum of both .
still , @xmath88 such that we can safely take @xmath89 .
since we attribute the creation of sk - ask pairs to the existence of a finite @xmath31 , we expect sks to be created faster when the magnitude of @xmath31 is larger . from eqs .
( [ eq::jskd1],[eq::jskd2],[eq::jd ] ) we find @xmath90 in the limit of vanishing @xmath91 and @xmath92 . in fig .
[ fig::tau ] , this relation between @xmath84 and @xmath31 is confirmed by the numerical results .
indeed , @xmath84 depends on @xmath93 and @xmath94 .
in particular , no sks can be created when @xmath37 which implies that the dissipative current is essential for the charge current - induced sk creation . finally , we note that even though we have chosen a particular seed magnetic field @xmath95 to create sk density fluctuations , their precise origin is not important .
in fact , only an inhomogeneous @xmath96 , besides @xmath42 and @xmath97 , is necessary for @xmath31 to become non - vanishing .
thus , a multitude of ways are eligible to create such fluctuations , for example by local fields , material modification , or by temperature . on the other hand
, a change of @xmath0 will often be undesirable in distinct set - ups .
then , @xmath31-contributions to eqs .
( [ eq::jsk ] ) should be suppressed by a proper choice of the material with a small @xmath93 , or by avoiding magnetization fluctuations , apart from simply working in the low current regime . in this work ,
we have established the skyrmion equation of motion by combining the general definition of the skyrmion density and the landau - lifshitz - gilbert equation .
we here define `` skyrmion current densities '' that conserve the total skyrmion number of a sample . in the presence of in plane
spin current densities , we identify terms that give rise to simple movement of skyrmions against the externally applied charge flow . other contributions to skyrmion current densities that we identify
explicitly drive the separation of positive skyrmion density from negative antiskyrmion density perpendicularly to the charge current flow .
these latter contributions eventually cause the creation of skyrmion - antiskyrmion pairs , already out of very small magnetic inhomogeneities .
the theoretical predictions are corroborated by numerical simulations and applied to systems with bulk and interfacial dmi .
we acknowledge support from the dfg sfb 668 ( project b16 ) . + |
the energy spectrum of mesons produced in neutrino - nucleus interactions is modified by strong interactions with the residual nucleus .
recent high - statistics measurements of charged - current @xmath3production by miniboone @xcite and minerva @xcite have shown tension with available models @xcite .
a study of @xmath0production is complementary because of differences in the nuclear interaction due to strangeness conservation .
previous measurements of neutrino - induced charged - current @xmath0production have been carried out in bubble chambers with very limited statistics @xcite .
we report the first high - statistics measurement of this process based on a sample of 1755 selected event candidates , of which 885 are estimated to be charged - current @xmath0events with @xmath4 mev . at neutrino energies below 2 gev ,
cabibbo suppressed single kaon production @xmath5 is the dominant @xmath0production mechanism . at higher energies ,
@xmath0mesons arise via associated production accompanied by strangeness @xmath6 baryons ( @xmath7 , @xmath8 ) or mesons ( @xmath9 , @xmath10 ) such that there is no net change in strangeness ( @xmath11 ) .
this can occur through an intermediate resonance state or in deep inelastic scattering ( dis ) by hadronization , the production of mesons and baryons from the struck quark .
in particular , @xmath12 pairs created in hadronization lead to pairs of strange particles in the final state .
production of @xmath0by atmospheric neutrinos is a background in experimental searches for the proton decay @xmath13 , a channel favored by grand unification theories which incorporate supersymmetry .
the simplest minimal supersymmetric models @xcite give proton lifetimes that have been excluded by experiment . however , other models @xcite allow proton lifetimes greater than @xmath14 years , consistent with the current experimental lower bound of @xmath15 years from a 260 kiloton - year exposure by super - kamiokande @xcite .
the @xmath0from proton decay is below cherenkov threshold in water , but a liquid argon time projection chamber such as dune @xcite is able to reconstruct the @xmath0momentum precisely . the @xmath0momentum spectrum in @xmath13depends on the momentum distribution of the initial - state protons inside the nucleus .
a related issue is the extent to which @xmath0mesons born inside the nucleus experience final - state interactions ( fsi ) as they emerge into the detector medium .
kaons produced by neutrinos are subject to the same interactions . measuring @xmath0production by neutrinos on carbon is a first step toward understanding the spectrum for @xmath13 in the argon of the dunefar detector .
kaon - nucleus and pion - nucleus reactions differ because of strangeness conservation .
absorption is the dominant feature in the pion - nucleus inelastic cross section at pion kinetic energies in the few 100s of mev . in @xmath9-nucleus scattering ,
the @xmath9can be absorbed , converting a bound nucleon into a hyperon .
the analogous process for @xmath0-nucleus scattering is forbidden because there are no antibaryons in the nucleus .
a @xmath0produced inside the nucleus will exit unless it charge exchanges to a @xmath16 .
in addition , @xmath0can be produced in @xmath3-nucleus reactions by strong processes such as @xmath17 . in the giessen boltzmann - uehling - uhlenbeck model @xcite
, this kind of reaction gives an enhancement to the @xmath0production cross section at low @xmath0momentum . in genie @xcite , the event generator used by minervaand many other experiments , 13% of @xmath0produced in carbon reinteract before exiting the nucleus , distorting the spectrum toward lower kaon energies .
geniedoes not include @xmath0production either by pions or charge exchange in its fsi model . this paper reports a measurement at high statistics of inclusive charged - current @xmath0production by muon neutrinos , @xmath1 ch @xmath18 .
the differential cross section in @xmath0kinetic energy is measured and compared to predictions of current neutrino event generators with and without fsi treatments .
minervais a dedicated neutrino - nucleus cross section experiment in the numi beamline @xcite at fermilab .
the detector consists of a core of strips of solid plastic scintillator `` tracker '' surrounded by calorimeters on the sides and downstream end .
the electromagnetic and hadronic calorimeters intersperse scintillator with passive planes of lead and steel , respectively .
the upstream nuclear targets region is used only to veto front - entering events for this result .
the minos near detector is located 2 m downstream of minerva .
positive muons from antineutrino - induced charged - current reactions are rejected using curvature , but the muon momentum measurement is not used in this analysis .
the scintillator strips are arranged into planes stacked perpendicular to the horizontal axis , and are rotated @xmath19 and @xmath20 with respect to the vertical axis to enable unambiguous three - dimensional tracking of charged particles .
the cross section of the strips is triangular with a base edge of 3.4 cm and a height of 1.7 cm . in the center of each strip
is a wavelength - shifting optical fiber which is mirrored at one end and read out by a 64-channel multi - anode photomultiplier tube at the other .
a hit is defined as an energy deposit in a single scintillator strip .
the uncalibrated hit time is the time of the earliest charge recorded on a single channel , with an electronics resolution of 2.2 ns . when a charge threshold is exceeded , charge is integrated for 151 ns such that subsequent energy deposits in one strip due to the same neutrino interaction accumulate onto one hit .
in particular , the timing of a delayed @xmath0decay product is lost if the decay particle overlaps spatially with prompt energy due to other particles produced in the neutrino interaction . because of this effect , the reconstruction efficiency depends on the particle multiplicity .
the timing resolution is a function of the number of observed photoelectrons ( pe ) because it is based on the decay of the fluors in the scintillator and wavelength - shifting fiber .
for many - pe hits , the timestamp will come from light that resulted from very prompt decays of the scintillator and fiber ; at smaller numbers of pe , the recorded hit times are delayed relative to the true time of the energy deposition .
the timing resolution is 10 ns for 1 - 2 pe hits , 3 ns for 6 - 12 pe hits due to minimum ionizing particles , and approaches the 2.2 ns resolution of the electronics asymptotically at very high pulse heights .
timing information is first used to separate multiple neutrino interactions within a single 10 @xmath21s beam pulse .
hits are sorted by their time , and a scan is performed to find 80-ns windows where the total energy exceeds a threshold .
the window is moved forward in time until the threshold is no longer met .
this algorithm reliably separates neutrino interactions which occur 100 ns apart , and keeps a @xmath0and its decay products together .
the design , calibration and performance of the minervadetector , including the calibration of the timing response , is described in detail in ref .
the hit timing in the data acquisition system is described in ref .
the data for this measurement were collected between march 2010 and april 2012 , corresponding to @xmath22 protons on target ( pot ) .
the horn current was configured to focus positive pions , resulting in a @xmath1-enriched beam with 10% @xmath23contamination which is largely in the high - energy tail of the flux .
the neutrino beam is simulated by a geant4-based model @xcite that is tuned to agree with hadron production measurements on carbon @xcite by the procedure described in ref .
uncertainties on the neutrino flux arise from the statistical and systematic uncertainties in these hadron production experiments , as well as uncertainties in the beamline geometry and alignment @xcite .
the integrated neutrino flux is estimated to be @xmath24{cm^{-2}/pot}$ ] .
table [ tab : nu_flux ] lists the flux as a function of energy . | production of @xmath0mesons in charged - current @xmath1interactions on plastic scintillator ( ch ) is measured using minervaexposed to the low - energy numi beam at fermilab .
timing information is used to isolate a sample of 885 charged - current events containing a stopping @xmath0which decays at rest . the differential cross section in @xmath0kinetic energy , @xmath2 ,
is observed to be relatively flat between 0 and 500 mev .
its shape is in good agreement with the prediction by the genieneutrino event generator when final - state interactions are included , however the data rate is lower than the prediction by 15% . 0=1 0=1 0=0 |
thermonuclear explosions of carbon - oxygen ( co ) white dwarfs ( wds ) , which lead to type ia supernovae ( sne ia ) , could be triggered by the merger of two wds ( the double - degenerate scenario ( dds ) @xcite ) or by the accretion of matter from a non - degenerate star through roche - lobe overflow ( the single - degenerate scenario ( sds ) @xcite ) . in the sds , the companion to the co - wd could be a main - sequence ( ms ) , red giant ( rg ) , or helium - rich ( he ) star . in the dds , the companion could be another co wd or he wd . based on current studies , both scenarios are not ruled out by observations , but are also not proven by them . if both scenarios lead to sn ia , it is still unclear which channel(s ) in the sds and dds dominate(s ) the sne ia , and by what fraction @xcite .
recent multi - dimensional hydrodynamics simulations of sn impact on the non - degenerate binary companions in the sds suggest that the companions should survive the sn impact and could be detectable . therefore , one of the direct methods to distinguish between the sds and dds is to search for the surviving companions ( scs ) in sn ia remnants ( ia snrs ) .
* hereafter p10 ) and ( * ? ? ?
* hereafter p12a ) examine the effects of a sn impact on the non - degenerate binary companions in the sds for ms , rg , and he star binary companions via multi - dimensional hydrodynamics simulations .
these simulations include the symmetry - breaking effects of orbital motion , rotation of the binary companion , and roche - lobe overflow ( rlof ) , all of which allow a better description of sn - driven shock compression , heating , and stripping of scs . examined a similar sn impact on ms and he stars with the sph approach using companion models from more sophisticated one - dimensional binary evolutions .
however liu et al .
did not study the subsequent post - impact evolution and therefore could not predict the properties of scs in historical ia snrs . on the other hand , @xcite and @xcite examined the evolution of a @xmath2 subgiant and a @xmath2 ms companion with ad hoc prescriptions for energy input and mass stripping without performing detailed hydrodynamical calculations , thereby dramatically overestimating the luminosity of scs . in (
* hereafter p12b ) and ( * ? ? ?
* hereafter p13 ) , we mapped our detailed three - dimensional hydrodynamical results into a one - dimensional stellar evolution code to simulate the post - impact evolution of ms- and he - scs and thus provided a more realistic treatment of post - impact evolution . in this paper , we calculate the time evolution of the magnitudes and colors of our models and discuss the possibility of searching for sds - scs in specific nearby ia snrs .
we present numerical results in 2 and compare them with searches for scs in galactic ia snrs in 3 , in the magellanic clouds in 4 , and in m31 and m82 in
5 . finally
, we discuss the evidence for the sds channel for sne ia and present our conclusions in the last section .
in this section , we discuss and predict some possible observables of scs in nearby ia snrs and describe differences between sc candidates and unrelated stars in ia snrs . in particular , we calculate the colors , magnitudes , effective temperatures , and photospheric radii of ms- and he - scs as functions of time .
we also predict the magnitude and effective temperature changes as functions of time .
the linear and rotational velocities of scs are also discussed .
we also suggest upper limits for the ni / fe contamination in scs . in simulations of ms- and he - scs , the bolometric luminosity , effective temperature , and photospheric radius
are directly determined using the stellar evolution code mesa ( modules for experiments in stellar astrophysics ; @xcite ) . to facilitate direct comparison with optical observations ,
the bolometric luminosity is converted to broad band magnitudes .
additional observable quantities such as the strength of absorption lines require the treatment of detailed radiation transfer effects in stellar atmosphere models , which are not considered in this paper . given the effective temperature and photospheric radius of a sc , the magnitude of the sc can be estimated under the assumption that the photosphere emits a blackbody radiation spectrum .
we have considered several different filters with their corresponding sensitivity functions in this study , including the johnson - cousins - glass ubvir system and the hst / wfc3 system .
the absolute magnitudes are calculated in the ab magnitude system . for a given extinction , @xmath3
, the extinction curve can be calculated using the fitting formula in @xcite .
the stellar parameters of ms- and he - sc models in our previous work are summarized in table [ tab_models ] .
figures [ fig_hr_all ] and [ fig_hr_bvi ] illustrate the hertzsprung - russell ( h - r ) diagram in different representations using the b and v wavebands and hst wavebands respectively .
he - scs are sdo - like stars that exhibit stronger emission in the @xmath4 and @xmath5 bands , while ms - scs exhibit greater emission in the @xmath6 and @xmath7 bands ( a - k subgiants ) .
the absolute magnitudes of ms - scs ( he - scs ) span the range @xmath8 @xmath9 .
the brightest phase of ms - scs corresponds to ia snr ages of @xmath10 yr , which are similar to the ages of most known historic nearby ia snrs ( see table [ tab1 ] ) .
therefore , if these snrs originated from normal sne ia via the sds ms or the he star channels , scs should be detectable . in the rg donor channel , almost the entire envelope of a rg
should be removed during the sn impact ( @xcite ; p12a ) .
therefore , the sc in ia snr will no longer be a giant star , but could be a helium degenerate core star with a shallow hydrogen - rich envelope .
it should be noted that in our calculations , we assume normal sne ia with chandrasekhar mass explosions , and the explosions are initially spherically symmetric .
asymmetric explosions , sub - chandrasekhar , or super - chandrasekhar mass explosions may reduce or enhance the evolution of post - impact luminosity ( p12b ) .
in addition to comparisons of the colors and magnitudes of sc candidates , the long term variation of magnitude and effective temperature changes is an alternative way to identify the sc . in figure [ fig_hr_all ]
we see that ms - scs show rapid luminosity changes but maintain similar effective temperatures in the early few hundred years after an explosion . therefore , a slight magnitude drop but without color change for a few years after the first observation is expected for ms - scs .
on the other hand , he - scs show both magnitude and color changes in the first ten years .
figure [ fig_mag ] and figure [ fig_teff ] show the magnitude and effective temperature variations of our ms- and he - scs as functions of time .
brighter sc models ( models a , b , and d ) have a maximum magnitude change rate of 0.3 mag yr@xmath11 in the v - band when their ia snrs have ages less than five hundred years . for models c , e , and f ,
the ms - scs show a smoother rate of change of magnitude during the first thousand years , but the maximum rate is less than 0.01 mag yr@xmath11 , which is very difficult to detect with current optical telescopes . model g radiates the deposited energy immediately and shows a positive magnitude gradient ( gets dimmer ) .
the rate of change of magnitude for model g is less than @xmath12 mag yr@xmath11 .
the effective temperatures of all our ms - sc models are roughly constants .
the highest rate of change is about @xmath13 k yr@xmath11 for models a , b , and d and @xmath14 k yr@xmath11 for other models .
however , the rates of change of magnitude and effective temperature for he - scs could be notable .
all our he - scs change by @xmath15 mag yr@xmath11 in the b - band during the first three years and maintain @xmath16 mag yr@xmath11 over the first decade .
furthermore , he - scs increase in temperature by @xmath17 k yr@xmath11 during the first @xmath18 years and then decrease by @xmath19 k yr@xmath11 in the first decade .
therefore , these magnitude and effective temperature changes could be detected in young and nearby ia snrs if the sc is a he star .
a sc will have an abnormal speed due to its original orbital speed plus the sn kick .
the momentum transfer between the sn ejecta with the donor star will give a kick velocity perpendicular to its orbital velocity .
the kick velocity is about @xmath20 km s@xmath11 for ms and he star donor channels , and @xmath21 km s@xmath11 for the rg donor channel ( p10 ; p12a ; ) . by adding the kick velocity to the orbital velocity , the final linear speed at the end of supernova impact is @xmath22 km s@xmath11 for ms - scs , @xmath23 km s@xmath11 for he - scs ( see table [ tab_models ] ) , and @xmath24 km s@xmath11 for rg - scs ( p12a ) .
the contribution from the kick velocity could be as high as @xmath25 in the ms donor channel .
assuming the co wd has a similar orbital speed ( but opposite direction ) at the time of explosion and the snr is moving with this orbital speed , the observational error circle has to include a radius of at least @xmath26 , where @xmath27 is the age of the snr , and @xmath28 is the inclination angle .
the theoretical estimate of the maximum @xmath29 is @xmath17 km s@xmath11 for the he star donor channel .
the surface rotational speed of a sc also could be an important signature for sc searches in nearby ia snrs . at the time of the sn
ia explosion , the companion should be synchronized by tidal locking and be rapidly rotating , resulting in a surface rotational speed close to @xmath30 km s@xmath11 for ms donors , @xmath31 km s@xmath11 for he star donors , and @xmath24 km s@xmath11 for rg donors ( p12a ; ) . however , because of the angular momentum loss accompanying mass stripping , the surface rotational speed could drop to @xmath32 km s@xmath11 for ms donors and @xmath33 km s@xmath11 for he star donors after the sn impact ( p12b ; p13 ) .
once the sn ejecta escape and the sc reaches hydrostatic equilibrium , the rotational speed could keep decreasing or increasing as a result of post - impact contraction or expansion . if we assume the specific angular momentum is conserved during the post - impact evolution , the surface rotational speed could be in the range from @xmath34 km s@xmath11 to @xmath31 km s@xmath11 , depending on the amount and depth of sn energy deposition and the age of the snr ( see figure 11 in p12b and figure 12 in p13 ) .
in p12b and p13 , we have shown that the surface rotation speed will first drop to @xmath35 km s@xmath11 after the sn impact and then keep decreasing to @xmath36 km s@xmath11 during the early expansion over a timescale of @xmath37 yr for ms - scs and @xmath38 yr for he - scs .
the stellar expansion is due to the release of energy deposited by the sn ejecta .
once the deposited energy has all been released , the envelope of the sc will contract , releasing its gravitational energy . in this phase ,
the rotational speed may increase up to @xmath38 km s@xmath11 for ms - scs and @xmath31 km s@xmath11 for he - scs .
for most of the lifespan of ia snrs with ages less than a thousand years , ms - scs are slowly rotating subgiants but he - scs are rapidly rotating sdo / b stars .
@xcite have suggested that the contamination from sn ejecta may provide observable features in iron absorption lines , which can be used to identify the scs in ia snrs . in p12a , we showed that the amount of ni / fe contamination at the surfaces of the scs is @xmath39 for ms star companions , @xmath40 for he star companions , and @xmath41 for rg companions . at the current stage
, our simplified post - impact evolution method can not predict the strength of any absorption or emission lines during the evolution , but we can provide an order - of - magnitude estimate by assuming that these contaminated material are uniformly mixed in the stellar envelope . the estimated upper limit for the nickel - to - hydrogen - plus - helium ratio is about @xmath42 for ms - scs , @xmath43 for he - scs , and @xmath44 for rg - scs ( p12a ) .
the rate of galactic sne ( including type i and type ii ) is about @xmath45 sne per century , and @xmath46 of them are sne ia @xcite .
these rates suggest that there should be more than 2,500 snrs in our galaxy , and 300 of them are ia snrs , if snrs are recognizable for ages less than @xmath47 yrs .
however , only @xmath48 galactic snrs have been identified @xcite , and only four of them are known as ( or likely to be ) ia snrs ( @xmath49 ) . therefore , a search for a sc in the central region of a snr could also be an alternative indirect method to identify ia snrs among currently - known snrs .
table [ tab1 ] shows a summary of the four possible galactic ia snrs ( and/or ia snr candidates ) : sn 1006 , sn 1572 , sn 1604 , and rcw 86 .
the young ( @xmath50 yr ) and nearby ( @xmath51 kpc , @xcite ) snr , tycho s snr , has been identified as a normal ia snr by its scattered - light echo @xcite . the non - thermal x - ray arc within tycho s snr could arise from an interaction between the sn ejecta and mass ejected from the companion star , giving support for the sds as the progenitor @xcite .
@xcite found a subgiant star , namely tycho g , characterized by a high radial velocity @xmath52 km s@xmath11 that could be related to the original orbital speed of the white dwarf companion in the sds .
* hereafter k13 ) recently updated this radial velocity to @xmath53 km s@xmath11 , which is still anomalous for the region .
figure [ fig_snrs ] shows our predictions of post - impact conditions of ms - scs compared with observations by ( * ? ? ? * hereafter gh09 ) and k13 .
gh09 reported that tycho g is a subgiant with effective temperature @xmath54 k , surface gravity @xmath55 dex , and [ fe / he ] @xmath56 dex , using keck high resolution optical spectra .
the distance of tycho g is also comparable to that of tycho s snr . using the same data but with a slightly different analysis tool
, k13 determined hotter and less luminous characteristics for tycho g ( @xmath57 k , @xmath58 dex , and [ fe / h]@xmath59 dex ; see k13 for a more detailed comparison of their results with gh09 ) . our model e ( @xmath60 k , @xmath61 dex , and @xmath62 ; see table 3 in p12b ) has a similar effective temperature to that of tycho g , but is brighter than the luminosity reported by gh09 ( @xmath63 , assuming a mass of @xmath2 ) .
we note that our model e is more massive ( @xmath64 ) than the mass ( @xmath2 ) assumed in gh09 and k13 , and the stellar radius and evolution stage are also different with what is suggested in ( * ? ? ?
* hereafter b14 ) . the evolution of scs after
the sn explosion depends not only on the amount of mass lost and energy deposition , but also on the depth of energy deposition
. therefore , a less massive companion , different evolutionary stage , asymmetric explosion , or sub - chandrasekhar mass explosion may better match tycho g. if tycho g is indeed similar to our model e , we predict a magnitude change of @xmath65 mag for every 10 years ( figure [ fig_mag ] ) .
b14 determined high - accuracy proper motions for @xmath66 sc candidates , including tycho g , in the central region of tycho s snr .
they obtained a tangential velocity for tycho g of @xmath67 km s@xmath11 . together with the observed radial velocity ( 80 km s@xmath11 ) of tycho g and the average radial velocity ( 37 km s@xmath11 ) in the direction of tycho s snr from the sun , the linear velocity with respect to the center of tycho s snr
can be determined to be @xmath68 km s@xmath11 , if tycho g were the sc ( b14 ) .
this value is half that of the linear velocities in our models in table [ tab_models ] .
an inclination angle @xmath69 , orbital separation @xmath70 , and orbital period @xmath71 days are also suggested by b14 .
these progenitor system data are consistent with the spectroscopic observation by gh09 indicating that tycho g is a g - type subgiant .
however , we note that the linear velocity does not equal the orbital speed in the binary system , since the sn kick may contribute up to one - third of its final linear speed ( p12a ; ) .
* hereafter k09 ) found an upper limit for the rotational speed of tycho g of @xmath72 km s@xmath11 , updated to @xmath73 km s@xmath11 in k13 , causing them to question tycho g as a sc candidate .
p12a , p12b , and studied the sn impact and post - impact conditions of ms - like binary companions and found that this discrepancy can be resolved due to the loss of angular momentum during the sn impact and post - impact expansion .
however , even when applying the inclination angle in b14 , our best model ( model e ) still has a @xmath74 km s@xmath11 , which lies above the upper limit observed by k09 and k13 .
similar results have been found in as well .
therefore , a less massive or less compact companion model is required to better fit with observations .
fortunately , using the inclination angle and orbital period derived in b14 , a low rotational speed of @xmath75 km s@xmath11 can be calculated by assuming tidal locking , explaining the non - detection of rotational speed in k09 and k13 . on the other hand , @xcite comments that the absence of an fe i line feature at 372 nm in tycho g argues against the sc interpretation as there is no evidence for absorption due to the sn ejecta in the stellar spectrum .
in contrast , tycho e shows a strong blueshifted fe i absorption line without redshifted lines , which implies that it is within the remnant and that its projected position is close to the center of tycho s snr , possibly qualifying it as another sc candidate .
however , k13 suggests that tycho e is far behind the snr and hence the low column density on the receding side of the remnant could explain the lack of redshifted lines .
the recent observation of tycho e reveals an effective temperature @xmath76 k and surface gravity @xmath77 dex ( k13 ) , which is also close to our model e ( @xmath60 k and @xmath61 dex )
. however , the low radial velocity ( @xmath78 km s@xmath11 ) and large distance ( @xmath79 kpc , in comparison to @xmath80 kpc ) make it less likely to be the sc , although the distance uncertainty is large .
finally , gh09 also have suggested that tycho e could be a double lined binary , which does not provide support for it as a sc .
further detailed observations of these sc candidates are necessary to establish whether any of them is connected to tycho s sn . the lack of a compact remnant star , and the amount of iron observed inside the snr @xcite indicate that sn 1006 was a sn ia .
measurements of its proper motion indicate that sn 1006 is the closest historical ia snr with a distance @xmath81 kpc @xcite and an age of @xmath82 yr . the geometric center of the remnant is also well determined in x - ray and radio @xcite .
furthermore , the foreground extinction is low due to its high galactic latitude ( @xmath83 ) , providing a good environment in which to search for sc candidates .
searches for sc candidates in sn 1006 have been done using two independent observations by ( * ? ? ? * hereafter gh12 ) and ( * ? ? ?
* hereafter k12 ) .
both teams suggest that there is no evidence for the survival of a sc in the central region of sn 1006 .
k12 observed all stars to a limit of 0.5 @xmath84 at the distance of sn 1006 within the central 2 arcminutes .
they found no stars as bright as predicted in @xcite and @xcite , and no stars show significant rotation .
similar conclusions have been drawn by gh12 using observations of the central 4 arcminutes . however , we note that the stars b90474
( @xmath85 cm s@xmath86 , @xmath87 k ) and b14707 ( @xmath88 cm s@xmath86 , @xmath89 k ) in gh12 have surface gravities similar to our model g ( @xmath90 dex , @xmath91 k ) , although with lower effective temperatures ( see figure [ fig_snrs ] ) .
furthermore , the star b90474 has a high radial velocity ( @xmath92 km s@xmath11 ) .
however , unlike the case of tycho s sn , sn 1006 lies 500 pc above the galactic plane , and therefore , the radial velocities of surrounding stars do not exhibit a simple trend .
in addition , b90474 has a larger distance uncertainty ( @xmath93 kpc ) and b14707 ( @xmath94 kpc ) lies at a much closer distance than the snr
. in both gh12 and k12 , the authors use the geometric center of the x - ray and radio observations .
the error circles included in gh12 and k12 are @xmath95 and @xmath96 respectively , corresponding to a sc moving with a speed of @xmath17 km s@xmath11 for 2,000 and 1,000 yr at a distance of 2.2 kpc . @xcite
have suggested that there is an offset between the geometric center and the center of the iron core .
furthermore , @xcite and @xcite have shown that the ejecta distribution in sn 1006 is asymmetric and concentrated in the se quadrant , suggesting a @xmath97 pc offset to the geometric center @xcite .
@xcite reported that the expansion velocity varies significantly with azimuth ( @xmath98 km s@xmath11 in the nw and @xmath99 km s@xmath11 in the se ) .
therefore , the error circles used in gh12 and k12 may not include the real explosion center .
on the other hand , the schweizer - middleditch star is a subdwarf ob ( sdob ) star which is located at the center of sn 1006 and has strong fe absorption lines @xcite .
it is relatively bright ( @xmath100 , @xmath101 ) , with low foreground extinction ( @xmath102 ; @xcite ) .
it is consistent with the he - sc models ; however , the presence of redshifted absorption lines due to supernova ejecta suggests that it is more likely a background star @xcite .
kepler s snr has been identified as a ia snr based on x - ray observations of its o / fe ratio @xcite , but its distance is still uncertain ( @xmath103 kpc ; ) .
the interaction of the supernova ejecta with circumstellar material in kepler s snr provides evidence that kepler s snr may have originated from the sds in an evolutionary channel consisting of a wd and an asymptotic giant branch ( agb ) star .
however , the circumstellar medium could be also explained by the stellar wind from a massive progenitor in core - collpase supernova .
the low iron mass of @xmath104 @xcite and a progenitor mass of @xmath105 @xcite suggest that kepler s sn may not be a sn ia .
recently , ( * ? ? ?
* hereafter k14 ) have ruled out red giants as scs due to the lack of bright stars in the central snr .
the observed radial velocity of sc candidates also shows only a small possibility of being associated with the sn explosion
. however , 24 stars with @xmath106 at the center of kepler s snr have been found by k14 , and five of them have @xmath107 , perhaps suggesting a relatively higher probability of being a sc .
additional observations are required to test these candidates . as we pointed out in 2 ,
nearly all the envelope of a rg should be removed during the sn impact in the rg donor channel , and this is likely to be the case for the agb donor channel as well .
therefore , the sc in kepler s snr would cease to be a agb star , which would explain the non - detection of a giant star in k14 .
follow - up observations and a detailed study of the evolution of the agb(rg ) channel in the sds and its scs will be necessary in order to understand the progenitor system of kepler s snr .
rcw 86 is the oldest known galactic sn ia .
it was recently identified as a ia snr by @xcite and @xcite , who estimate that the integrated fe - k emission corresponds to a total fe mass of about @xmath2 .
its possible association with sn 185 is somewhat dubious since it would imply the age of rcw 86 is 1,829 yr . the large size of its radius at an estimated distance of @xmath108 kpc @xcite ( or @xmath109 kpc ; ) suggests a very high shock speed ( @xmath110 km s@xmath11 ) or a much older age .
one explanation is that it originated in a cavity explosion , although cavity explosions are more common in core - collapse sne .
hence , the connection between rcw 86 and sn 185 has yet to be established .
in addition , the derived ambient density ( @xmath111 @xmath112 ) found by @xcite suggests that an unusually low - density cavity surrounds the snr .
this can be understood either as resulting from the existence of a strong stellar wind from the progenitor itself or as an outflow from the nearby ob association discovered by @xcite . if rcw 86 was a member of this group , the age of this group may place some constraints on the delay - time of this sn ia , although the ages of these ob stars are still unknown .
the wind - blown bubble scenario by @xcite suggests that rcw 86 originated in the sds .
therefore , a search for sc candidates in rcw 86 seems reasonable .
however , the two expansion fronts in the southwest and the northeast make it difficult to measure the center of the explosion ( k14 ) . in our calculations ,
all ms - sc models reach the brightest phase ( @xmath113 ) and highest effective temperature at around @xmath114 yr , depending on the model .
therefore the nearby distance and low extinction ( @xmath115 ; @xcite ) make it easier to distinguish sc from unrelated stars .
the magellanic clouds ( mcs ) are excellent environments to search for scs for several reasons .
ia snrs in the large magellanic cloud ( lmc ) and small magellanic cloud ( smc ) are at known distances , 50 kpc and 60 kpc , that are sufficiently close to detect scs with hst based upon the results in 2 .
the mcs are nearly face - on galaxies , minimizing confusion along the line of sight @xcite .
the foreground galactic extinction is low , and the internal extinction of the mcs is modest @xcite . finally , the star formation history ( sfh ) and metallicity of the mcs are different than those of the milky way @xcite , allowing the possibility of testing sn ia progenitors in different type of galaxies @xcite .
ten ia snrs have been reported to lie in the lmc and four in the smc ( see table [ tab1 ] ) . since only four galactic ia snrs have been identified , these fourteen ia snrs provide a larger sample of ia snrs to search for scs .
in addition , the distances of ia snrs in the lmc and smc are known more accurately than for galactic ia snrs , which reduces the uncertainties in comparing observations with the theoretical predictions .
recently , the youngest ia snr in the lmc , snr 0509 - 67.5 , has been studied by @xcite .
snr 0509 - 67.5 formed from a sn 1991t - like sn ia @xmath116 yrs ago .
the observation of the central region of snr 0509 - 67.5 within a radius of @xmath117 , corresponding to a distance of 0.36 pc from the center of snr 0509 - 67.5 , shows no star brighter than @xmath118 ( @xmath119 ) in this region .
this result rules out all standard sds channels and suggests that snr 0509 - 67.5 originated in the dds @xcite .
@xcite studied the central sources within an error circle of @xmath120 radius about the center the @xmath121 yr old ia snr 0519 - 69.0 using hst .
they found 27 ms stars brighter than @xmath122 magnitude .
this result requires the progenitor of snr 0519 - 69.0 to arise from either the dds or the sds with a supersoft source .
we point out that @xcite found a sub - giant star in snr 0519 - 69.0 ( @xmath123 mag , and @xmath124 ) .
although this star is consistent with our model g star ( @xmath125 mag , and @xmath126 ; the black line in figure [ fig_snrs ] ) , it lies close to the edge of the possible error circle , which suggests that it may not be the sc of snr 0519 - 69.0 unless the explosion was asymmetric .
besides the lmc and smc , other nearby galaxies can provide samples that probe diverse environments for scs .
in particular , the metallicities and sfhs in other nearby galaxies differ from those of the milky way and magellanic clouds , which could affect pre - sn conditions , the sn ia explosion itself , and , potentially , the post - impact evolution . in section [ sec_predictions ]
, we have shown that our predicted ms - scs ( he - scs ) span a range of absolute magnitude @xmath8 ( @xmath127 ) in f555w ( f438w ) . by using hst / wfc3 ( with u ,
b , v , and i filters ) , the limiting magnitude with s / n of 10 for point sources is @xmath128 with a one hour exposure , and @xmath129 for a 10 hour exposure , giving maximum distance moduli of @xmath130 and @xmath131 respectively .
this corresponds to a maximum detectable distance of @xmath0 mpc ( @xmath1 mpc ) if there is no extinction .
however , as extinction varies from galaxy to galaxy , the maximum detectable distance would be smaller in extreme cases .
given that he - scs reach their maximum brightness at @xmath38 yr after sn explosions and then fade within @xmath30 yr ( p13 ) , their @xmath5 magnitudes are likely less than 1 for historic ia snrs .
this yields a distance limit of @xmath132 mpc .
however , the short timescale ( @xmath38 yrs ) of their luminous phase provides the possibility to detect slow transitions of their brightness at the locations of recent nearby sne ia with @xmath133 mpc .
we note that the timescale for the decay of the light curve from the supernova may take longer than the luminous phase of the sc and , hence , that the maximum detectable distance could be smaller .
snrs in nearby galaxies , including the andromeda galaxy ( m31 or ngc 224 ; ) , ngc 300 @xcite , m82 @xcite , ngc 4214 , ngc 5204 , ngc 5585 , ngc 6946 , m81 , and m101@xcite , have been studied recently .
therefore , ia snrs in these galaxies should be targeted for sc searches .
however , it is unclear whether these snrs are ia snrs . more than 26 snrs ( including type i and type ii ) have been identified in the andromeda galaxy at 0.79 mpc by @xcite and .
m31 has a distance modulus of 24.5 , which makes it possible to detect all of our ms - scs and most he scs using hst / wfc3 .
the panchromatic hubble andromeda treasury ( phat ) multi - cycle program has observed about one third of m31 using hst @xcite .
its uvis data reached a magnitude limit of @xmath134 in the f275w and f336w bands ; acs data reach maximum depths of @xmath135 magnitudes in f475w and @xmath128 magnitudes in f814w in the uncrowded outer disk . in these same regions ,
wfc3/ir data reach maximum depths of @xmath136 and @xmath137 in f110w and f160w , respectively .
have listed 26 x - ray snrs and 20 x - ray snr candidates in m31 based on their x - ray , optical , and radio emission , which is the most recent complete list of x - ray snrs in m31 .
therefore , using their list together with phat s data , it may be possible to identify sc candidates .
sn 1885a ( s andromedae ) is a subluminous sn ia in m31 @xcite .
its young age and the small size of its remnant make it easier to search for a sc .
however , the surrounding light is dominated by the bulge of m31 , making sc searches difficult unless the sc is overluminous .
sn 2014j was discovered by stephen fossey and his students on january 21 , 2014 and reported as a sn ia on january 22 , 2014 @xcite .
sn 2014j is the closest sn ia ( @xmath138 mpc ; @xcite ) discovered in the past 42 years .
the next close one is sn 1972e ( in ngc 5253 ) with a distance of @xmath139 mpc .
sn 2011fe is also a nearby sn ia discovered in the modern astronomical era .
however , the distance of sn 2011fe ( in m101 ; @xmath140 mpc , @xcite ) is above our predicted maximum distance for ms - scs .
a search for scs in sn 2011fe will be very challenging .
furthermore , the emptiness of the surrounding medium @xcite and the lack of pre - explosion optical and x - ray source @xcite suggest a dds progenitor for sn 2011fe .
sn 2014j is almost twice as close as sn 2011fe , and the distance ( 3.5 mpc ) is within our predicted detection limit ( 4 mpc ) for ms - scs , though the distance is close to the limit .
hst archival observations do not rule out sds progenitors such as recurrent novae or some classical novae @xcite . however
, even when the snr becomes transparent , any ms - sc would need a few hundred years to become bright enough to be detected by hst .
if the sc is a he - sc , a bright luminous ob - like star is expected to be observed in the next few decades .
we have determined the colors and magnitudes of seven ms- and four he - sc models based on our stellar evolutionary calculations ( p12b , p13 ) and predicted their rates of change as functions of time .
we have also studied the linear and rotational speeds of scs and the potential ejecta contamination of scs .
comparisons of the model predictions for scs with the galactic sne 1572 and 1006 are presented , assuming all the candidate stars have the same distance as their host snrs . in particular , it is found that both tycho e ( @xmath76 k , @xmath77 dex ) and tycho g ( @xmath54 k , @xmath55 dex ) approximately fit our predicted ms - sc model e ( @xmath60 k , @xmath61 dex ) , although there are small discrepancies ( k13 ) . furthermore , two sub - giants in the error circle of sn 1006 are also found to have magnitudes consistent with our ms sc model g ( @xmath141 k , @xmath90 dex ) , but have lower effective temperatures @xcite .
however it should be noted that the uncertainties in the distances of these sc candidates are large , and that the derived effective temperatures and surface gravities are not consistent in different papers .
in addition to the galactic ia snrs , a sub - giant in the snr 0519 - 69.0 is consistent with the magnitude and color of our model g , but the projected position in the snr is too far from the center @xcite .
if such a candidate is a confirmed sc , then the original orbital speed ( plus sn kick ) would be much higher than expected unless the explosion was asymmetric .
although higher orbital speeds would be expected for he scs , there is no evidence for such a candidate .
based on the current sample of sne ia companion searches , it is more likely that most have originated from the dds or peculiar sds channels .
peculiar sds channels such as the m - dwarf channel @xcite or the spin up / down channel @xcite may explain the non - detection of sc candidates in ia snrs . to obtain a better understanding of the progenitor systems of sn ia ,
we encourage further sc searches in other galactic or nearby extragalactic ia snrs . unlike galactic ia snrs , which have large distance uncertainties , the distances of extragalactic ia snrs
are relatively well - known .
we predict that the maximum detectable distance of ms - scs ( he - scs ) is @xmath0 mpc ( @xmath1 mpc ) , if the apparent magnitude limit is 27 with no extinction , suggesting that the lmc , smc , and m31 are excellent targets in which to search for scs .
furthermore , we also predict he - scs will not only show high luminosity and effective temperature ( luminous ob - like stars ) , but also show a rate of magnitude in b of @xmath142 mag yr@xmath11 and a rate of change of effective temperature of @xmath143 k yr@xmath11 in the first decade since explosion
. future observations of he - sc candidates should look into their magnitude and color changes as well
. similar analysis can be applied to core - collapse ( cc ) sne as well .
the observed high fraction of sn ib / c can not be explained by single stellar evolution , suggesting that a high fraction of ccsne were in binary systems @xcite .
however , the ejected - mass - and - energy and companion types and separations are very different from those in sn ia . in addition
, the companion does not need to be in roche lobe overflow at the time of explosion .
therefore , the evolution of scs in ccsne could be very different and vary from case to case .
the impact of ccsne on binary companions and their subsequent evolution are important future work in this area .
we thank the anonymous referee for his / her valuable comments and suggestions .
kcp thanks you - hua chu for useful discussions about ia snrs in the lmc .
this work was supported by the computational science and engineering ( cse ) fellowship at the university of illinois at urbana - champaign , by the theoretical institute for advanced research in astrophysics ( tiara ) in the academia sinica institute of astronomy and astrophysics ( asiaa ) , and by the european research council ( erc ) grant fish .
the flash simulations presented here were carried out using the nsf xsede ranger system at the texas advanced computing center under allocation tg - ast040034n .
flash was developed largely by the doe - supported asc / alliances center for astrophysical thermonuclear flashes at the university of chicago .
9 badenes , c. , hughes , j. p. , bravo , e. , & langer , n. 2007 , , 662 , 472 bedin , l. r. , ruiz - lapuente , p. , gonzlez hernndez , j. i. , et al .
2014 , , 439 , 354 blair , w. p. , kirshner , r. p. , & chevalier , r. a. 1981 , , 247 , 879 borkowski , k. j. , reynolds , s. p. , green , d. a. , et al .
2010 , , 724 , l161 borkowski , k. j. , reynolds , s. p. , hwang , u. , et al .
2013 , , 771 , l9 burkey , m. t. , reynolds , s. p. , borkowski , k. j. , & blondin , j. m. 2013 , , 764 , 63 cao , y. , kasliwal , m. m. , mckay , a. , & bradley , a. 2014 , the astronomer s telegram , 5786 , 1 carlton , a. k. , borkowski , k. j. , reynolds , s. p. , et al .
2011 , , 737 , l22 chiotellis , a. , schure , k. m. , & vink , j. 2012 , , 537 , a139 chomiuk , l. , soderberg , a. m. , moe , m. , et al .
2012 , , 750 , 164 dalcanton , j. j. , williams , b. f. , lang , d. , et al .
2012 , , 200 , 18 della valle , m. , & melnick , j. 1992 , , 257 , l1 di stefano , r. , voss , r. , & claeys , j. s. w. 2011 , , 738 , l1 di stefano , r. , & kilic , m. 2012 , , 759 , 56 dopita , m. a. , calzetti , d. , maz apellniz , j. , et al .
2010 , , 330 , 123 edwards , z. i. , pagnotta , a. , & schaefer , b. e. 2012 , , 747 , l19 elias - rosa , n. , greggio , l. , & botticella , m. t. 2014 , the astronomer s telegram , 5849 , 1 fenech , d. , beswick , r. , muxlow , t. w. b. , pedlar , a. , & argo , m. k. 2010 , , 408 , 607 ferrand , g. , & safi - harb , s. 2012 , advances in space research , 49 , 1313 fesen , r. a. , hflich , p. a. , hamilton , a. j. s. , et al .
2007 , , 658 , 396 fossey , j. , cooke , b. , pollack , g. , wilde , m. , & wright , t. 2014 , central bureau electronic telegrams , 3792 , 1 gonzlez hernndez , j. i. , ruiz - lapuente , p. , filippenko , a. v. , et al .
2009 , , 691 , 1 gonzlez hernndez , j. i. , ruiz - lapuente , p. , tabernero , h. m. , et al . 2012 , , 489 , 533 green , d. a. , & gull , s. f. 1984 , , 312 , 527 green , d. a. , reynolds , s. p. , borkowski , k. j. , et al .
2008 , , 387 , l54 hamilton , a. j. s. , fesen , r. a. , wu , c .- c . , crenshaw , d. m. , & sarazin , c. l. 1997 , , 481 , 838 harris , j. , & zaritsky , d. 2004 , , 127 , 1531 harris , j. , & zaritsky , d. 2009 , , 138 , 1243 haschke , r. , grebel , e. k. , & duffau , s. 2011 , , 141 , 158 hatsukade , i. , tsunemi , h. , yamashita , k. , et al . 1990 , , 42 , 279 hillebrandt , w. , kromer , m. , rpke , f. k. , & ruiter , a. j. 2013 , frontiers of physics , 8 , 116 hughes , j. p. , & helfand , d. j. 1985 , , 291 , 544 iben , i. , jr . ,
& tutukov , a. v. 1984 , , 54 , 335 ihara , y. , ozaki , j. , doi , m. , et al .
2007 , , 59 , 811 justham , s. 2011 , , 730 , l34 karachentsev , i. d. , & kashibadze , o. g. 2006 , astrophysics , 49 , 3 kerzendorf , w. e. , schmidt , b. p. , asplund , m. , et al .
2009 , , 701 , 1665 kerzendorf , w. e. , schmidt , b. p. , laird , j. b. , podsiadlowski , p. , & bessell , m. s. 2012 , , 759 , 7 kerzendorf , w. e. , yong , d. , schmidt , b. p. , et al .
2013 , , 774 , 99 kerzendorf , w. e. , childress , m. , scharwchter , j. , do , t. , & schmidt , b. p. 2014
, , 782 , 27 kochanek , c. s. 2009 , , 707 , 1578 krause , o. , tanaka , m. , usuda , t. , et al .
2008 , , 456 , 617 leibowitz , e. m. , & danziger , i. j. 1983 , , 204 , 273 liu , j. , di stefano , r. , wang , t. , & moe , m. 2012 , , 749 , 141 lu , f. j. , wang , q. d. , ge , m. y. , et al .
2011 , , 732 , 11 liu , z. w. , pakmor , r. , rpke , f. k. , et al . 2012 , , 548 , a2 liu , z .- w . ,
pakmor , r. , rpke , f. k. , et al . 2013a , , 554 , a109 liu , z .- w .
, pakmor , r. , seitenzahl , i. r. , et al . 2013b , , 774 , 37 marietta , e. , burrows , a. , & fryxell , b. 2000 , , 128 , 615 maoz , d. , mannucci , f. , & nelemans , g. 2013 , arxiv:1312.0628 maoz , d. , & badenes , c. 2010 , , 407 , 1314 matonick , d. m. , & fesen , r. a. 1997 , , 112 , 49 meng , x. , & yang , w. 2011 , science in china : physics , mechanics and astronomy , 54 , 2296 millar , w. c. , white , g. l. , & filipovic , m. d. 2012 , serbian astronomical journal , 184 , 19 nomoto , k. 1982 , , 257 , 780 odonnell , j. e. 1994 , , 422 , 158 ozaki , j. , & shigeyama , t. 2006 , , 644 , 954 pakmor , r. , rpke , f. k. , weiss , a. , & hillebrandt , w. 2008 , , 489 , 943 pan , k .- c . , ricker , p. m. , & taam , r. e. 2010 , , 715 , 78 pan , k .- c . , ricker , p. m. , & taam , r. e. 2012a , , 750 , 151 pan , k .- c . , ricker , p. m. , & taam , r. e. 2012b , , 760 , 21 pan , k .- c . , ricker , p. m. , & taam , r. e. 2013 , , 773 , 49 paxton , b. , bildsten , l. , dotter , a. , et al .
2011 , , 192 , 3 paxton , b. , cantiello , m. , arras , p. , et al .
2013 , , 208 , 4 podsiadlowski , p. 2003 , arxiv : astro - ph/0303660 reynolds , s. p. , borkowski , k. j. , hwang , u. , et al .
2007 , , 668 , l135 reynolds , s. p. , borkowski , k. j. , green , d. a. , et al .
2008 , , 680 , l41 rosado , m. , ambrocio - cruz , p. , le coarer , e. , & marcelin , m. 1996 , , 315 , 243 ruiz - lapuente , p. 2004
, , 612 , 357 ruiz - lapuente , p. 2014
, arxiv:1403.4087 ruiz - lapuente , p. , comeron , f. , mndez , j. , et al .
2004 , , 431 , 1069 sasaki , m. , pietsch , w. , haberl , f. , et al .
2012 , , 544 , a144 schaefer , b. e. , & pagnotta , a. 2012 , , 481 , 164 schweizer , f. , & middleditch , j. 1980 , , 241 , 1039 shappee , b. j. , & stanek , k. z. 2011 , , 733 , 124 shappee , b. j. , kochanek , c. s. , & stanek , k. z. 2013 , , 765 , 150 smith , n. , li , w. , filippenko , a. v. , & chornock , r. 2011 , , 412 , 1522 subramanian , s. , & subramaniam , a. 2012 , , 744 , 128 tammann , g. a. , loeffler , w. , & schroeder , a. 1994 , , 92 , 487 uchida , h. , yamaguchi , h. , & koyama , k. 2013 , , 771 , 56 van den bergh , s. 2002 , , 123 , 2045 van der marel , r. p. , & cioni , m .- r . l. 2001 , , 122 , 1807 vink , j. , kaastra , j. s. , & bleeker , j. a. m. 1997 , , 328 , 628 vink , j. 2012 , , 20 , 49 wang , b. , & han , z. 2009 , , 508 , l27 wang , b. , & han , z. 2012 , , 56 , 122 wang , x. , wang , l. , filippenko , a. v. , zhang , t. , & zhao , x. 2013 , science , 340 , 170 webbink , r. f. 1984 , , 277 , 355 westerlund , b. e. 1969 , , 74 , 879 wheeler , j. c. 2012 , , 758 , 123 whelan , j. , & iben , i. , jr . 1973 , , 186 , 1007 williams , b. j. , blair , w. p. , blondin , j. m. , et al .
2011 , , 741 , 96 williams , s. c. , darnley , m. j. , bode , p. a. , et al .
2014 , the astronomer s telegram , 5824 , 1 winkler , p. f. , gupta , g. , & long , k. s. 2003 , , 585 , 324 winkler , p. f. , long , k. s. , hamilton , a. j. s. , & fesen , r. a. 2005 , , 624 , 189 winkler , p. f. , williams , b. j. , reynolds , s. p. , et al .
2014 , , 781 , 65 wu , c .- c . , leventhal , m. , sarazin , c. l. , & gull , t. r. 1983 , , 269 , l5 yamaguchi , h. , koyama , k. , & uchida , h. 2011 , , 63 , 837 zaritsky , d. , harris , j. , thompson , i. b. , & grebel , e. k. 2004 , , 128 , 1606 lccccc a ( ms ) & 1.88 & 1.25 & 1.64 & 3.87 & 179 + b ( ms ) & 1.82 & 1.50 & 1.65 & 4.76 & 179 + c ( ms ) & 1.82 & 2.63 & 1.56 & 7.61 & 136 + d ( ms ) & 1.63 & 1.19 & 1.43 & 3.42 & 188 + e ( ms ) & 1.59 & 1.42 & 1.44 & 3.91 & 191 + f ( ms ) & 1.55 & 1.97 & 1.30 & 4.09 & 143 + g ( ms ) & 1.17 & 0.79 & 0.93 & 4.45 & 271 + hewda ( he ) & 0.697 & 0.091 & 0.656 & 0.390 & 734 + hewdb ( he ) & 0.803 & 0.158 & 0.748 & 0.565 & 550 + hewdc ( he ) & 1.007 & 0.194 & 0.962 & 0.809 & 509 + hewdd ( he ) & 1.206 & 0.231 & 1.126 & 1.041 & 446 + lcrr rcw 86 ( sn 185?)@xmath144 &
@xmath145 & @xmath146 ( mw ) & @xmath147 + sn 1006 & 1,007 & @xmath148 ( mw ) & 9.3 + sn 1572 ( tycho s ) & 441 & @xmath149 ( mw ) & 3.8 + sn
1604 ( kepler s ) & 409 & @xmath150 ( mw ) & 3 + b0509 - 67.5 & @xmath151 & 50 ( lmc ) & 3.6 + n103b & @xmath152 & 50 ( lmc ) & 3.6 + b0519 - 69.0 & @xmath153 & 50 ( lmc ) & 3.9 + dem l71 & @xmath154 & 50 ( lmc ) & 8.6 + b0548 - 70.4 & @xmath155 & 50 ( lmc ) & 12.5 + dem l316a & ? & 50 ( lmc ) & 15 + b0534 - 69.9 & @xmath156 & 50 ( lmc ) & 16 + dem l238 & @xmath157 & 50 ( lmc ) & 21 + dem l249 & @xmath157 & 50 ( lmc ) & 23 + b0454 - 67.2 & @xmath158 & 50 ( lmc ) & 27 + ikt 4 & ? & 60 ( smc ) & @xmath159 + ikt 5 & ? & 60 ( smc ) & 15 + ikt 25 & ? & 60 ( smc ) & 18 + dem s128 & ? & 60 ( smc ) & 26 + | the nature of the progenitor systems of type ia supernovae is still unclear .
one way to distinguish between the single - degenerate scenario and double - degenerate scenario for their progenitors is to search for the surviving companions . using a technique that couples the results from multi - dimensional hydrodynamics simulations with calculations of the structure and evolution of main - sequence- and helium - rich surviving companions , the color and magnitude of main - sequence- and helium - rich surviving companions
are predicted as functions of time .
the surviving companion candidates in galactic type ia supernova remnants and nearby extragalactic type ia supernova remnants are discussed .
we find that the maximum detectable distance of main - sequence surviving companions ( helium - rich surviving companions ) is @xmath0 mpc ( @xmath1 mpc ) , if the apparent magnitude limit is 27 in the absence of extinction , suggesting that the large and small magellanic clouds and the andromeda galaxy are excellent environments in which to search for surviving companions .
however , only five ia snrs have been searched for surviving companions , showing little support for the standard channels in the singe - degenerate scenario . to better understand the progenitors of type ia supernovae
, we encourage the search for surviving companions in other nearby type ia supernova remnants . |
quantum pumping is a transport phenomenon originally proposed by thouless@xcite and first realized by switkes _
et al_.@xcite it proposed that directed current can be induced by time - dependent modulation of external and internal parameters without bias in a quantum phase coherent nanoscale conductor .
theoretical and experimental research of quantum pumping has become a very important and active direction in mesoscopic physics .
it is also significant in the field of quantum dynamic theory . by scale of the modulation frequency
the quantum pump can be categorized into the adiabatic and nonadiabatic ones , with the former frequency much smaller than the characteristic tunneling times and vice the latter@xcite .
adiabatic quantum pumping can be described by berry phase of the scattering matrix accumulated during the cyclic modulation in the parameter space@xcite and also by the energy quanta absorption and emission processes equivalent to the nearest sideband approximation@xcite .
nonadiabatic quantum pumping can be described by the floquet scattering scheme picturing quantities of interest in terms of sideband formation@xcite .
the non - equilibrium green s function@xcite , equation of motion@xcite , galileo transformation@xcite , and etc . also show physics of nonadiabatic quantum pumping from different views .
adiabatic and nonadiabatic quantum pumping has been investigated in various mesoscopic systems , such as nanowire@xcite , mesoscopic rings@xcite , quantum - dot structures@xcite , spin - orbit coupled conductors@xcite , magnetic tunnel junction@xcite , graphene@xcite , and superconductor junction with majorana fermions@xcite , etc .
current fluctuations are present in almost all kinds of conductors including dynamical transport systems .
the shot noise is the quantum contribution in the current fluctuation produced by the quantum coherence of charge carriers , which can give rich physical information in mesoscopic transport systems and is more significant in nanoscale quantum devices than in the traditional non - quantum devices@xcite .
although intensive work has been done on the bias driven shot noise in various mesoscopic conductors@xcite , adiabatic pumped noise is also extensively investigated@xcite , and the general scattering theory for nonadiabatic pumped shot noise@xcite is derived , the specific pumped shot noise properties in different quantum transport systems are less covered .
they represent the underlying physics of different materials and devices , some of which is beyond conductance information .
generally a transport approach covers the shot noise as well as the conductance .
the floquet scattering matrix approach was developed for nonadiabatic noise properties as detailed by moskalets _
et al_.@xcite general expressions for the pumped current , heat flow , and shot noise are derived for adiabatically and non - adiabatically driven quantum pumps .
this approach stresses the existence of sidebands of electrons passing the time - dependent scatterer and these sidebands are connected to the currents and noise directly .
recently , park and ahn@xcite derived an expression for the admittance and the current noise for a driven nanocapacitor in terms of the floquet scattering matrix and obtained a non - equilibrium fluctuation dissipation relation .
the scattering matrix renormalized by interaction has been used by devillard _
et al_.@xcite to study the effect of weak electron - electron interaction on the noise . under the geometric framework
, there have been beautiful mathematical descriptions from the current to the noise@xcite . in this work ,
we focus on the non - adiabatic quantum pump driven by a single oscillating potential well , in which fano resonance is predicted in the floquet transmission spectrum when one of the floquet levels matches the quasibound level of the static potential well@xcite . in this case , the pumped current vanishes due to time and spatial reversal symmetry and the fano resonance is unavailable in the current measurement .
however , energy and information is transfused into the pump from exterior by instantaneous transport within a driving cycle . in the pumping process
, virtual or temporary transmission within a cyclic period generates considerable noise .
supposing the resonance feature can be characterized in the correlation , we investigated its noise and heat flow properties .
we consider a one - dimensional width-@xmath0 time - dependent potential - well sketched in fig .
the time - dependent potential , which oscillates with frequency @xmath1 and is located between @xmath2 and @xmath3 , has the form @xmath4 the time - dependent hamiltonian of the electrons can be expressed as @xmath5 @xmath6 is the effective mass of electrons and our discussion is based on single electron approximation and coherent tunneling .
firstly , we consider the quasibound states within the one - dimensional static quantum well . when the electron is confined in the well with its energy @xmath7 , the wave function can be written as @xmath8 where @xmath9 and @xmath10 .
continuity equations of the wave function and its derivative at @xmath11 and @xmath12 are @xmath13 solvability of these equations gives rise to the secular equation @xmath14 roots of @xmath15 for this equation are the quasibound state energies .
it can be obtained numerically that the only quasibound state for the considered well configuration is at the energy@xcite of @xmath16 mev .
we use the floquet scattering theory to investigate the quantum pumping properties of the oscillating quantum well@xcite .
wave functions in the three scattering regimes can be written as @xmath17 in the left and right free regions , the incident and outgoing electron waves consist of infinite number of sidebands , as shown in fig
@xmath18 is the eigenenergy of the @xmath19-th order floquet state with the fermi energy @xmath20 and @xmath21 is the corresponding wave vector .
@xmath19 are integers varying from @xmath22 to @xmath23 in an ideal exactness . while @xmath24 , @xmath25 is imaginary meaning an evanescent mode , the current for this channel vanishes .
@xmath26 and @xmath27 are the probability amplitudes corresponding to those flowing out of and flowing into the left / right leads , respectively .
here we did not use flux normalization for algebra simplicity of continuity conditions and its justification is elaborated in the appendix .
the floquet scattering matrix can be constructed by relations between @xmath26 and @xmath27 .
for example , if we set @xmath28 , @xmath29 corresponds to the reflection amplitude from the left lead to the left lead in the @xmath19-order floquet channel , and so on .
@xmath30 is wavevector in the middle oscillating quantum well region .
@xmath31 is the first kind of bessel function deriving from @xmath32 , which only exists in the oscillating region .
@xmath33 and @xmath34 are the wave function amplitudes in the oscillating region and present only in the continuity equation solving processes . by solving equations of the boundary conditions at interfaces
@xmath35 and @xmath36 which must hold for all time , the floquet scattering matrix without flux normalization can be obtained by matrix algebra .
it connects different floquet modes as @xmath37 with @xmath38 and @xmath39 column vectors made up of @xmath26 and @xmath27 of all @xmath40s .
@xmath41 and @xmath42 are the transmission and reflection amplitudes incoming from the @xmath43-th floquet channel and going into the @xmath19-th floquet channel . considering the real current flux ,
the floquet scattering matrix elements are @xmath44 transmission of evanescent modes would vanish as only the real part of the outgoing wave vectors is considered . from the scattering matrix , the total transmission probability is defined as @xmath45 we are interested in the element @xmath46 of the floquet scattering matrix given in eq .
( [ fluxnormalizedscatteringmatrix ] ) .
it measures the scattering amplitude of the electron incident through lead @xmath47 with energy @xmath15 and leaving through lead @xmath48 with energy @xmath49 . by interacting with the oscillating potential the electron absorbs or loses energy quanta of @xmath50 , with its final energy @xmath51 .
for the non - adiabatic quantum pump , the floquet scattering matrix is sensitive to the spatial symmetry of the potential .
if the system is present of perturbations broken the spatial symmetry or time - reversal symmetry it can pump a dc current . with only one oscillating potential , both the spatial symmetry and time - reversal symmetry is present in the device , thus no pumped current exists with @xmath52 the heat flow is carried by the non - equilibrium particles , which occurs in the process of scattering and the direction of heat flow is defined as from the oscillating potential to the reservoirs as @xmath53 . \label{eq9}\ ] ] here @xmath54 is equilibrium fermi distribution function .
@xmath55 is the chemical potential , which is the same in all reservoirs at zero bias .
we also assume all the reservoirs have the same temperature .
the problem of current noise is closely connected with the matrix elements of @xmath46 . for a phase - coherent conductor
the noise is sensitive to the quantum - mechanical interference effects .
we can describe the correlation function of the current as@xcite @xmath56 where @xmath57 , and @xmath58 is the quantum - mechanical current operator in the lead @xmath48 , which can be expressed as@xcite @xmath59 } e^{i\left ( { e - e ' } \right)t/\hbar } , \label{eq11}\ ] ] with @xmath60 and @xmath61 annihilation operators of the incident and outgoing electrons to the driven potential and @xmath62 from eqs .
( [ eq1 ] ) to ( [ bsa ] ) the pumped shot noise and pumped heat flow noise can be expressed in terms of the floquet scattering matrix as@xcite @xmath63 ^ 2 } } } } } , \label{eq2}\ ] ] @xmath64 ^ 2 , \label{eq3}\ ] ] with @xmath65 describing the quantum - mechanical exchange during scattering of electrons with energy @xmath66 incident from leads @xmath67 and outgoing to the leads @xmath68 with energy @xmath69 , respectively .
current flux conservation secures that @xmath70 and @xmath71 .
we consider one of the four and label @xmath72 as @xmath73 , @xmath74 as @xmath75 , and @xmath76 . to magnify the resonance spectrum
, we also considered the derivatives of the noise and heat flow over the fermi energy with @xmath77
in this paper we have adopted the floquet scattering matrix approach to investigate the pumped effect of phase coherent mesoscopic systems of noninteracting electrons .
the floquet scattering matrix describes existence of sidebands of electrons entering and exiting the pump .
the nonequilibrium electrons generated by the pump carry heat from the oscillating potential to the reservoirs and transfer charge between the two reservoirs .
only the first sidebands @xmath78 are excited if the oscillating amplitude is small .
the total transmission probability @xmath79 as a function of the incident energy is shown in fig .
2 . in all of the numerical consideration @xmath80
is set to be 1 mev .
we take into account different floquet sidebands both above and below the incident energy , with @xmath81 .
@xmath82 cutoff is used , with its precision satisfactory for the small driving amplitude . in the quantum
well there exists a quasibound state , when the first order floquet sideband overlaps with the quasibound level , a
fano " resonance occurs ( also confer fig .
1 ) , which was discovered in ref . .
due to time and spatial reversal symmetry , no charge current is generated by a single oscillating quantum well .
it is known that when pumped charge current is zero , the pumped shot noise can be considerably large due to virtual transmission processes@xcite .
we suppose that the nonequilibrium transmission properties can be recorded in the shot noise spectrum and the fano " resonance can thus be observed .
we calculated the pumped current noise , heat noise and heat flow driven by the nonadiabatic oscillating quantum well using the floquet scattering scheme with eqs .
( [ eq9 ] ) , ( [ eq2 ] ) , and ( [ eq3 ] ) .
their variation as a function of the fermi energy was depicted in fig .
3 . the pumped current noise , heat noise and heat flow increases with the fermi energy when more energy channels contribute to the transport for larger fermi energies . for a small driving amplitude in our case ,
most of the transmission comes from the original incident level and the two first order floquet sidebands ( @xmath83 ) .
when these bands completely go out of the quantum well with @xmath84 , a decrease occurs in the noise spectrum .
noise is an effect of correlation , concrete or virtual . when all the active floquet bands are out of the quantum well , the transmitting electron
sees " no structure in the conductor therefore ballistic transport governs giving rise to the shot noise decrease .
an inflection could be found at the fermi energy @xmath85 mev corresponding to the fano " resonance in the total transmission .
the magnification of the inflection point is shown in the insets .
since the transport properties are an accumulating effect of all energy channels , contribution of a single energy channel of the resonance is limited .
to magnify the fano " resonance of the nonadiabatic quantum pump , we calculated the differentials of the pumped charge and heat noise to the fermi energy and dramatic resonance pattern could be found .
differentials of the pumped current noise , heat noise , and heat flow as a function of the fermi energy are shown in fig .
4 . these curves have a sharp dip followed by a peak at @xmath86 mev , demonstrating an asymmetric fano " resonance , which ensures that there exists a quasibound state ( the energy of the quasibound state is @xmath87 mev ) in the deep quantum well@xcite .
electrons in the propagating states can emit photons and drop into the quasibound state and bounce back before exiting the well , thus contributing to the transport ( see fig .
1 ) . the heat flow also shows a sharp peak at the resonance fermi energy .
the pumped noise properties can be interpreted as follows . at the resonant fermi energy , transport process and the electron - electron correlation
achieve maximal strength .
the nonequilibrium electrons created by the oscillating scatterer move in different directions carried the heat flow into the electron reservoirs of the two sides .
the differential shot noise demonstrates peaks corresponding to the
fano " resonance , as a result of the virtual motion of nonequilibrium electrons .
we considered the noise properties of a nonadiabatic quantum pump driven by an oscillating potential well .
due to time and space reversal symmetry no dc charge current can be produced .
due to virtual transmission of the electrons , heat current is not zero with its direction from the conductor into the leads at both reservoirs . to experimentally observe the
fano " resonance found in the transmission@xcite , we investigated the heat current and shot noise of the charge and heat current .
sharp fano "-
shape resonance was found .
the differential current noise , heat noise , and heat flow demonstrate peak structure from the interaction of electrons with the oscillating potential when one of the floquet sideband matches the quasibound state .
electrons in incident channel can drop into the quasibound state by emitting photons .
similarly , electron in the bound state can also absorb photons and bounce back into the floquet channels .
thus a fano " resonance occurred .
the resonance position of the fermi energy is then governed by the energy of the static quasibound state .
the authors acknowledge enlightening discussions with professor wen - ji deng . this project was supported by the national natural science foundation of china ( no .
11004063 ) and the fundamental research funds for the central universities , scut ( no .
2014zg0044 ) .
the current operator in the left lead ( far from the sample ) can be expressed in a standard way@xcite , @xmath88 } , \label{acurrent}\ ] ] where the field operator @xmath89 is defined as @xmath90}^{{1 \mathord{\left/ { \vphantom { 1 2 } } \right .
\kern-\nulldelimiterspace } 2}}}}}\left [ { { { \hat a}_l}\left ( { { e_n } } \right){e^{i{k_{n}}x } } + { { \hat b}_l}\left ( { { e_n } } \right){e^ { - i{k_{n}}x } } } \right ] } .
\label{afieldoperator}\ ] ] to avoid tediousness we consider single transverse channel and multiple energy channels , the latter of which is necessary for the nonadiabatic dynamic process . @xmath91 and @xmath92 are defined identical to the main text .
it can be seen that the field operator is naturally flux normalized with @xmath93 in the denominator .
substituting eq .
( [ afieldoperator ] ) into eq .
( [ acurrent ] ) , and using the relations , @xmath94 and @xmath95 we could have @xmath96 } \right . } } \\ \times \left [ { { e^{i\left ( { { k_{lm } } - { k_{ln } } } \right)x}}\hat a_l^\dag \left ( { { e_n } } \right){{\hat a}_l}\left ( { { e_m } } \right ) - { e^ { - i\left ( { { k_{lm } } - { k_{ln } } } \right)x}}\hat b_l^\dag \left ( { { e_n } } \right){{\hat b}_l}\left ( { { e_m } } \right ) } \right ] \\ \left .
{ - \left [ { { { \left ( { \frac{{{k_{lm}}}}{{{k_{ln } } } } } \right)}^{{1 \mathord{\left/ { \vphantom { 1 2 } } \right .
\kern-\nulldelimiterspace } 2 } } } - { { \left ( { \frac{{{k_{ln}}}}{{{k_{lm } } } } } \right)}^{{1 \mathord{\left/ { \vphantom { 1 2 } } \right .
\kern-\nulldelimiterspace } 2 } } } } \right]\left [ { { e^ { - i\left ( { { k_{lm } } + { k_{ln } } } \right)x}}\hat a_l^\dag \left ( { { e_n } } \right){{\hat b}_l}\left ( { { e_m } } \right ) - { e^{i\left ( { { k_{lm } } + { k_{ln } } } \right)x}}\hat b_l^\dag \left ( { { e_n } } \right){{\hat a}_l}\left ( { { e_m } } \right ) } \right ] } \right\ } . \\ \end{array}\ ] ] we consider the dc current driven by nonadiabatic periodic parameter variation . the time - averaged current can be calculated in an arbitrary pumping cycle as @xmath97 where @xmath98 means the quantum and statistical average of the current operator . with the integral @xmath99 we could obtain @xmath100 } } , \label{acurrentwithn}\ ] ] here it should be noted that the floquet channels occur only in the scattering process and the original and final states are energy conserved with the energy of @xmath15 .
then , we have @xmath101 } .
\label{aaveragedcurrent}\ ] ] the outgoing operators can be expressed in terms of the incoming ones by the floquet scattering matrix @xmath102 for reservoirs at equilibrium , we have @xmath103 substituting eqs .
( [ adefinesmatrix ] ) and ( [ afermidistribution ] ) into the averaged current ( [ aaveragedcurrent ] ) , we can obtain @xmath104 } , \ ] ] which reproduced the result of ref . .
unitarity of the floquet scattering matrix naturally follows from the current conservation .
the current flowing from the left lead is identical to that flowing into the right lead @xmath105 ( positive direction of the current is defined as flowing from the reservoir to the scatterer ) . from eq .
( [ acurrentwithn ] ) , we have @xmath106 } = \sum\limits_n { \left [ { \hat b_r^\dag \left ( { { e_n } } \right){{\hat b}_r}\left ( { { e_n } } \right ) - \hat a_r^\dag \left ( { { e_n } } \right){{\hat a}_r}\left ( { { e_n } } \right ) } \right ] } .\ ] ] by eq .
( [ adefinesmatrix ] ) , it follows that @xmath107 which could be written into the matrix form as @xmath108 with corresponding elements @xmath109 and @xmath110 directly from eq .
( [ asmatrixunitarity ] ) follows the unitarity relation @xmath111 all elements of the floquet scattering matrix are only well defined for incident and outgoing propagating modes . in deducing the floquet scattering matrix from continuity conditions , it is more convenient to use the wave function without flux normalization as in eq .
( [ wavefunction ] ) . by considering the real current flux ,
we do the transform of eq .
( [ fluxnormalizedscatteringmatrix ] ) to obtain the floquet scattering matrix .
we could reproduce previous derivations by defining @xmath112 and @xmath113 with @xmath114 defined in eq .
( [ eq6 ] ) and directly obtainable from continuity relations . .
there is a quasibound state in the potential well with the binding energy @xmath115 .
energy is infused into the system by ac modulation of the potential well .
fano resonance occurs when one of the floquet sideband overlaps with the quasibound state .
equilibrium well depth is @xmath116 and its variation in time has the form of @xmath117.,width=377,height=264 ] as a function of the incident energy@xcite . driving amplitude @xmath118 mev , static well depth @xmath119 mev , well width @xmath120 , and energy quanta of the driving frequency @xmath121 mev .
a resonance occurs at @xmath85 mev . , width=453,height=377 ] , ( b ) heat flow shot noise @xmath75 , and ( c ) heat flow @xmath122 , as functions of the fermi energy .
their units are obtained by substituting @xmath123 mev into the energy and absorbing additional @xmath124 into the data .
an inflection occurs at @xmath85 mev corresponding to the resonance in transmission .
insets are the zoom - in of the inflection point .
, width=415,height=377 ] , heat flow noise @xmath125 , and the heat flow @xmath126 , as functions of the fermi energy .
their units are obtained by substituting @xmath123 mev into the energy and absorbing additional @xmath124 into the data .
sharp resonance could be seen at the inflection in fig .
3 with @xmath85 mev . , width=529,height=377 ] | we use the floquet scattering theory to study the correlation properties of the nonadiabatic pumped dc current and heat flow through a time - dependent quantum well .
electrons can transit through the quasibound state to the oscillator induced floquet states leading to resonant tunneling effect .
virtual electron scattering processes can produce pumped heat flow , pumped shot noise and pumped heat flow noise , with presence of time and spatial reversal symmetry . when one of the floquet levels matches the quasibound level there strikes a fano " resonance . |
exactly solvable models of interacting fermions in one - dimension ( 1d ) have attracted theoretical interest for more than half a century . before 1950 , it was not clear how to treat the schrdinger equation for a large system of interacting fermions .
the first important breakthrough was achieved by tomonaga @xcite who showed that fermionic interactions in 1d can mediate new collective degrees of freedom that are approximately bosonic in nature . in 1963
, luttinger @xcite introduced an exactly solvable many - fermion model in 1d which consists of two types of particles , one with positive momentum and the other with negative momentum .
however , luttinger s model suffers from several flaws which include the assumption that the fermions are spinless and massless , and more importantly an improperly filled negative energy dirac sea .
mattis and lieb @xcite expanded on luttinger s work by correctly filling the negative energy states with `` holes '' . before that , lieb and liniger @xcite solved the 1d interacting bose gas with @xmath1-function interactions using bethe s hypothesis @xcite .
later mcguire solved the equivalent spin-1/2 fermion problem for the special case where all fermions have the same spin except one having the opposite spin in the repulsive @xcite and attractive @xcite regimes .
he showed that in the presence of an attractive potential a bound state is formed .
further progress by lieb and flicker @xcite followed on the two down spin problem . in 1967 , yang @xcite solved the fermion problem for the most general case where the number of spin ups and spin downs are arbitrary by making use of bethe s hypothesis .
at the same time , gaudin @xcite solved this problem for the ground state with no polarization .
sutherland @xcite then showed that the fermion model with a general @xmath0 spin symmetry is integrable and the solution is given in terms of @xmath2 nested bethe ansatz ( ba ) equations . and in 1970 ,
takahashi @xcite examined the structure of the bound states in the attractive regime with arbitrary spin and derived the ground state energy together with the distribution functions of bound states in terms of a set of coupled integral equations . using yang and yang s method @xcite for the boson case , takahashi @xcite and lai @xcite derived the so - called thermodynamic bethe ansatz ( tba ) equations for spin-1/2 fermions in both the repulsive and attractive regimes .
the spin - string hypothesis describing the excited states of spin rapidities was also introduced by both authors .
later on , schlottmann @xcite derived the tba equations for @xmath0 fermions with repulsive and attractive interactions .
see also schlottmann s epic review article on exact results for highly correlated electron systems in 1d @xcite .
the tba equations have been analyzed in several limiting cases , i.e. , @xmath3 , @xmath4 , @xmath5 and @xmath6 , where @xmath7 is the temperature and @xmath8 is the interaction strength .
the ground state properties and the elemental charge and spin excitations were also studied for some special cases .
however , the tba equations for the attractive regime @xcite are not the most convenient for the analysis of phase transitions and thermodynamics . for the attractive case
, it was shown that the ground state in the absence of symmetry breaking fields consists of spin neutral charge bound states of @xmath2 particles .
the repulsive case however consists of freely propagating charge states and spin waves with different velocities .
the phenomenon of spin - charge separation plays a ubiquitous role in the low energy physics of 1d systems @xcite .
however , the physics of these models , such as the universal thermodynamics of tomonaga - luttinger liquids , quantum criticality and the universal nature of contact interaction , are largely still hidden in the complexity of the tba equations .
it is thus important to develop new methods to extract the physics of 1d exactly solved many - body systems in order to bring them more closer to experiments .
most recently , experimental advances in trapping and cooling atoms to very low temperatures allow a test of the theoretical predictions made so far . in particular , liao _ et al . _
@xcite experimentally studied spin-1/2 fermions of ultracold @xmath9li atoms in a 2d array of 1d tubes with spin imbalance .
the phase diagram was confirmed and it was discovered that a large fraction of a fulde - ferrell - larkin - ovchinnikov ( fflo)-like phase lies in the trapping center accompanied by two wings of a fully paired phase or unpaired phase depending on the polarization .
this observation verified the theoretical predictions @xcite regarding the phase diagram and pairing signature for the ground state of strongly attractive spin-1/2 fermions in 1d .
although the fflo phase has not yet been observed directly , the experimental results pave the way to direct observation and characterization of fflo pairing @xcite . in this paper
, we derive the tba equations for a general 1d system of fermions with @xmath0 spin symmetry from sutherland s ba equations using the same approach as yang and yang for 1d bosons @xcite .
both the repulsive and attractive cases are discussed .
we also give the exact thermodynamics of the ground state of the attractive and repulsive cases in both the strong coupling and weak coupling limits .
a general relationship between the different magnetic phases and the external magnetic field is discussed for the attractive case .
how the external magnetic fields affect the different pairing phases in the attractive regime is also addressed .
this paper gives a thorough derivation of many results in a recently published paper @xcite that provides the exact low temperature thermodynamics for strongly attractive @xmath0 fermions with zeeman splitting and shows that the system behaves like a universal tomonaga - luttinger liquid in the gapless phase .
the hamiltonian for the 1d @xmath10-body problem is @xmath11 it describes @xmath10 fermions of the same mass @xmath12 confined to a 1d system of length @xmath13 interacting via a @xmath1-function potential .
the first and second terms in the hamiltonian correspond to the kinetic energy and @xmath1-interaction potential respectively .
the coupling constant @xmath14 can be expressed in terms of the interaction strength @xmath15 as @xmath16 where @xmath17 is the effective 1d scattering length . for repulsive fermions , @xmath18 and for attractive fermions , @xmath19 .
we shall assume that the system has periodic boundary conditions i.e. , @xmath20 where @xmath21 is the position of the @xmath22-th particle .
there will be @xmath2 possible hyperfine states @xmath23 is given by @xmath24 while the zeeman energy @xmath25 is determined by the magnetic moment @xmath26 and the magnetic field @xmath27 .
for brevity , we shall choose the dimensionless units @xmath28 for the rest of this paper .
the wavefunction @xmath29 for this hamiltonian satisfies the symmetry of an irreducible representation @xmath30 $ ] , where the @xmath24s are such that @xmath31 . the young tableau which corresponds to this irreducible representation is given in fig .
[ fig : young ] .
this system has @xmath0 spin symmetry and @xmath32 charge symmetry .
sutherland @xcite showed that this problem can be solved by repeatedly using the generalized bethe s hypothesis which was introduced by yang @xcite . to obtain the total momentum @xmath33 and the energy eigenspectrum @xmath34 for the system , we need to find a set of quasimomenta @xmath35 that satisfies the equations @xmath36 @xmath37 where @xmath38 , @xmath39 , @xmath40 and @xmath41 is undefined .
the rapidities @xmath42 characterize the internal spin degrees of freedom .
the quantum numbers @xmath43 are defined as @xmath44 and @xmath45 .
this set of @xmath2 coupled algebraic equations are called the ba equations . $ ] .,height=491 ]
we shall only consider the strong coupling regime where @xmath46 .
for the repulsive case , it is easily shown that the roots @xmath35 must be real @xcite .
however the rapidities @xmath42 are allowed to take on nonzero imaginary values .
it was first suggested by lai @xcite that the rapidities appear as strings in the complex plane of the form @xmath47 in the thermodynamic limit where @xmath48 while keeping the ratio @xmath49 fixed . here
@xmath50 is the length of the string , @xmath51 labels each individual string and @xmath52 denotes the real part of each string .
every string must be symmetric along the real axis , i.e. any complex solution is accompanied by its complex conjugate pair .
this solution for the rapidities hold as long as @xmath46 . in other words , it holds up to order @xmath53 where @xmath1 is a positive monotonic increasing function of the interaction strength @xmath54 .
when the system is in its ground state , the rapidities do not form strings .
the strings also obey the relation @xmath55 where @xmath56 is the number of @xmath42 strings with length @xmath50 . in the attractive regime
, it is found that complex string solutions of @xmath35 also satisfy the ba equations .
a system with @xmath2-components of fermions has @xmath0 symmetry .
the quasimomenta @xmath35 may appear as bound states of up to length @xmath2 . for convenience ,
we denote the number of @xmath35 bound states with length @xmath57 as @xmath58 . in the ground state
, none of the bound states can be broken which means that @xmath59 . a bound state in @xmath60-space of length
@xmath12 will take the form @xmath61 with real part @xmath62 .
in general , a @xmath63 bound state of length @xmath12 will be accompanied by a @xmath64 bound state of length @xmath65 , a @xmath66 bound state of length @xmath67 and so on until a @xmath68 bound state of length @xmath69 .
each accompanying bound state in @xmath70-space , @xmath71-space , , @xmath72-space will share the same real part @xmath62 .
a graphical depiction is given in fig .
[ fig : roots ] .
however , strings in @xmath73-space do not have to be accompanied by any shorter string . therefore , the difference between our definition of a bound state and a string is that a bound state is a string that originates from @xmath60-space , and is accompanied by corresponding strings in each subsequent @xmath73-space . on the other hand
a string in @xmath73-space characterizes the spin excitations that can exist independently in spin sectors .
case with @xmath74 , @xmath75 , @xmath76 and @xmath77 in the ground state ( @xmath78 ) .
bounds states are formed for charge and spin rapidities and are indicated by the dashed boundaries . accompanying bound states in each subsequent @xmath73-space share the same real part .
the maximum length of any bound state in @xmath60-space is @xmath79 . in an excited state ( @xmath80 )
, strings can also form in @xmath73-space.,height=377 ]
the tba equations which are expressed in the form of dressed energies allow us to precisely derive numerous thermodynamic quantities and to analyze the behavior of phase transitions at critical points .
explicit expressions for the free energy , grand partition function , pressure , chemical potential and so on can be directly obtained from the tba equations .
physically , the tba equations give us the energy of elementary excitations above the ground state .
there are several steps to take in order to derive these equations .
we will give an outline of each step in deriving these equations for the repulsive case , all of which follow from yang and yang s pioneering work @xcite . the string solution for @xmath42 gives us a different form of the ba equations when substituted into the original equations ( [ ba1 ] ) and ( [ ba2 ] ) .
after lengthy calculations , we obtain @xmath81 @xmath82 @xmath83 @xmath84 the functions @xmath85 and @xmath86 are @xmath87 @xmath88 where @xmath89 taking the logarithm of each equation yields @xmath90 @xmath91 @xmath92 @xmath93 where @xmath94 and @xmath95 are odd or half - odd integers depending on the quantum numbers .
the functions @xmath96 and @xmath97 are @xmath98 @xmath99 where @xmath100 let us now introduce continuous monotonic increasing functions @xmath101 and @xmath102 .
denote @xmath103 and @xmath104 as the densities of `` particles '' and `` holes '' in @xmath60-space .
similarly , denote @xmath105 and @xmath106 as the densities of `` particles '' and `` holes '' for strings with length @xmath50 in the @xmath107-th rapidity space .
we shall express equations ( [ log1 ] ) to ( [ log4 ] ) in the continuum limit as @xmath108 @xmath109 @xmath110 @xmath111 using the relations @xmath112 and @xmath113 in the thermodynamic limit , we then have @xmath114 @xmath115 @xmath116 @xmath117 where the convolution integral @xmath118 is used .
the functions @xmath119 and @xmath120 are @xmath121 @xmath122 where @xmath123 the gibbs free energy per unit length is given by @xmath124 where the first term corresponds to the ground state energy , the second term corresponds to the addition and extraction of particles from the system , the third term represents the entropy of the system and the last term is associated with the zeeman energy per unit length of the system .
@xmath125 is the chemical potential , @xmath7 is the temperature and @xmath126 is the entropy per unit length . for each infinitesimal interval @xmath127 ,
the energy is degenerate .
the total number of microstates with this energy degeneracy is given by @xmath128 using sterling s approximation @xmath129 when @xmath130 , the entropy per unit length of the system follows as @xmath131 where the entropy is defined as @xmath132 .
the zeeman energy can be expressed as @xmath133 the following identities are required for further derivation @xmath134 in the thermodynamic limit , the zeeman energy can be written as @xmath135 minimizing the gibbs free energy per unit length when the system is in equilibrium i.e. , @xmath136 finally gives the tba equations @xmath137
@xmath138 @xmath139 @xmath140 where we have defined @xmath141 and @xmath142 .
the pressure per unit length of the system is @xmath143
after substituting the complex solutions for the bound states and strings into the original ba equations ( [ ba1 ] ) and ( [ ba2 ] ) , we obtain @xmath144 @xmath145 taking the logarithm of each equation yields @xmath146 @xmath147 we introduce the continuous monotonic increasing functions @xmath148 and @xmath149 . denote @xmath150 and @xmath151 as the densities of `` particles '' and `` holes '' for the bound states with length @xmath12 . similarly , denote @xmath105 and @xmath106 as the densities of `` particles '' and `` holes '' for strings with length @xmath50 in the @xmath107-th rapidity space . in the continuum limit ,
equations ( [ log5 ] ) and ( [ log6 ] ) become @xmath152 @xmath153 from the relations @xmath154 and @xmath155 in the thermodynamic limit , we obtain @xmath156 @xmath157 the gibbs free energy for the attractive case has the same expression as the repulsive case .
however , the expressions for each term in the gibbs free energy is different from the repulsive case .
the total number of microstates with the same energy degeneracy in the attractive case is given by @xmath158 the entropy per unit length of the system is @xmath159 the ground state energy per unit length was given by takahashi @xcite as @xmath160 it can be easily derived by taking the discrete sum @xmath161 with the roots given in equation ( [ attractiveroots ] ) and then extending it to the continuum limit . therefore the zeeman energy per unit length is @xmath162 where @xmath163 because we only need @xmath164 independent parameters to describe the relative distances between the energy levels of different fermionic species due to zeeman splitting .
here we shall denote @xmath165 for brevity . minimizing the gibbs free energy with respect to deviations in the various densities yields a set of coupled integral equations .
on introducing the dressed energy terms @xmath166 and @xmath167 , we arrive at the tba equations for attractive fermions with arbitrary spin @xmath168 @xmath169 take note that from the definition @xmath170 given earlier , @xmath171 is undefined . here
@xmath172 with @xmath173 are the dressed energies for the bound states of length @xmath12 e.g. @xmath174 is for unpaired fermions , @xmath175 is for pairs , @xmath176 is for trions and so on .
the fermi level is at @xmath177 which implies that the bound states of length @xmath12 are only occupied with fermions having quasimomenta @xmath178 .
there is an equivalent description for the band fillings of the strings .
the pressure per unit length for the system is @xmath179
in this section we present the thermodynamic properties of the system in the ground state where @xmath3 and also in the antiferromagnetic case where @xmath180 for every @xmath181 . in this regime , there are no string solutions for each level of rapidity because strings only exist in excited states . hence the tba equations simplify to @xmath182 in the regime @xmath183 , we can estimate @xmath184 from equation ( [ repulsivepressure ] ) .
taking the fourier transform of equations ( [ tba - rep2 ] ) to ( [ tba - rep4 ] ) yields the difference equations @xmath185 \\
\widehat{\xi}^{(p)}(\omega ) & = & \widehat{s}(\omega)\left[\widehat{\xi}^{(p-1)}(\omega)+\widehat{\xi}^{(p+1)}(\omega)\right ] \\
\widehat{\xi}^{(\kappa-1)}(\omega ) & = & \widehat{s}(\omega)\widehat{\xi}^{(\kappa-2)}(\omega)\end{aligned}\ ] ] where @xmath186 the general solution to this set of difference equations is @xmath187 the inverse fourier transform for this function is @xmath188 the convolution integral in equation ( [ tba - rep1 ] ) can be decoupled when @xmath183 where it becomes @xmath189 using parseval s theorem , @xmath190 we obtain @xmath191 where @xmath192 is the euler - mascheroni constant and @xmath193 is the psi ( digamma ) function .
the values of @xmath194 for @xmath195 are @xmath196 , @xmath197 and @xmath198 .
therefore equation ( [ tba - rep1 ] ) becomes @xmath199 using the conditions @xmath200 , @xmath201 and @xmath202 followed by iteration to keep terms up to order 1/c gives the thermodynamic expressions @xmath203
\\ \mu & \approx & \pi^{2}n^{2}\left[1+\frac{16}{3\kappa\gamma}\left(\mathcal{c}+\psi\left(\frac{1}{\kappa}\right)\right)\right ] \\ p & \approx & \frac{2}{3}\pi^{2}n^{2}\left[1+\frac{6}{\kappa\gamma}\left(\mathcal{c}+\psi\left(\frac{1}{\kappa}\right)\right)\right ] \\ f & \approx & \frac{1}{3}\pi^{2}n^{3}\left[1+\frac{4}{\kappa\gamma}\left(\mathcal{c}+\psi\left(\frac{1}{\kappa}\right)\right)\right]\end{aligned}\ ] ] where @xmath204 .
having derived the expression for the densities of bound states in equation ( [ attractivedensity1 ] ) , we can derive the ground state energy from equation ( [ attractive - e ] ) .
as with the repulsive case , there are no string solutions in the ground state . in the strong coupling regime @xmath205 , equation ( [ attractivedensity1 ] )
simplifies to @xmath206 an expression for the fermi points @xmath207 can be derived by first evaluating the relation @xmath208 which gives @xmath209 and then rearranging to obtain @xmath210 the ground state energy per unit length ( [ attractive - e ] ) is then given by @xmath211 this expression does not include the zeeman energy , which is equal to @xmath212 .
the actual ground state energy in the presence of an external magnetic field must include this term .
one can easily derive the ground state energy up to arbitrary orders in @xmath213 by including higher order contributions to the taylor expansions of the functions we applied them to .
for better accuracy , we derived the ground state energy that includes terms up to order @xmath214 in the more compact form @xmath215 where @xmath216 and @xmath217 is the binding energy for a bound state with length @xmath12 i.e. , @xmath218 looking back at the first tba equation ( [ tba - att1 ] ) for the attractive case , we can denote @xmath219 as the effective chemical potentials for the bound states of length @xmath12 .
an expression for the effective chemical potentials can be derived as @xmath220 here we used the vector notation @xmath221 , @xmath222 and @xmath223 where @xmath224 for @xmath225 .
the function @xmath226 is the heaviside step function with properties @xmath227 when @xmath228 and @xmath229 when @xmath230 . for the special cases of @xmath231 and @xmath232 fermions , see refs .
@xcite .
zeeman splitting can be characterized by the parameters @xmath233 or @xmath234 .
@xmath233 is the zeeman energy level for the species of fermions in state @xmath235 .
@xmath234 on the other hand parameterizes the zeeman energy level for bound states with length @xmath12 .
both sets of parameters are related by the expression @xmath236 a consistent solution to this relation for all @xmath237 is @xmath238 , @xmath239 and @xmath240 for @xmath241 .
if we denote the difference between the energy levels of fermions in state @xmath242 and @xmath235 as @xmath243 , we obtain the matrix relation @xmath244 where the blank entries in the upper and lower triangular sections of the @xmath245 matrix are defined as zero .
a useful relation between @xmath234 and the effective chemical potentials is @xcite @xmath246 from the original definition of the effective chemical potential .
the external fields @xmath234 can be tuned experimentally to drive the system between different phases where bound states of different lengths exist . in the special case
called pure zeeman splitting where @xmath247 for all @xmath12 , the system has three distinct magnetic phases .
the first phase consists of only @xmath2-bound fermions when @xmath248 .
the second phase contains a mixture of @xmath2-bound fermions and unbound fermions when @xmath249 . and
the third phase is made up of only unbound fermions when @xmath250 .
the critical external fields @xmath251 and @xmath252 can be evaluated from equation ( [ externalfield ] ) by taking @xmath253 and using the expressions for the effective chemical potentials from equation ( [ effectivemu ] ) .
doing so gives @xmath254\end{aligned}\ ] ] where @xmath255 is the spin normalized magnetization per particle density i.e. , @xmath256 .
this means that while @xmath257 depicts the true magnetization of the system , @xmath255 normalizes it and only takes on values from @xmath258 to @xmath69 .
the factor @xmath259 corresponds to the species of fermion that has the highest hyperfine spin because the unbound phase is made up of these fermions only .
the critical field @xmath251 corresponds to @xmath260 , while @xmath252 corresponds to @xmath261 .
substituting these values for @xmath255 into the equation for @xmath262 yields the general results @xmath263 \\ h_{1}^{c2 } & = & n^{2}\left[\frac{\kappa^{2}-1}{12}|\gamma|^{2}+\pi^{2}\left(1-\frac{8}{3\kappa(\kappa-1)|\gamma|}\right)\right].\end{aligned}\ ] ] the system has a linear field - dependent magnetization near the critical points . for a field slightly above @xmath251 ,
the magnetization is given by @xmath264 on the other hand , for a field that is slightly below @xmath252 , the magnetization is given by @xmath265
we have presented a thorough derivation of the tba equations for a system of 1d multi - component @xmath1-function interacting fermions in the presence of external magnetic fields .
the key results , in terms of which the thermodynamic properties are obtained , are equations ( [ tba - repulsive1])([tba - repulsive4 ] ) for the repulsive case and equations ( [ tba - att1])([tba - att2 ] ) for the attractive case .
the form of our tba equations differs from those derived by schlottmann @xcite , but are nevertheless possibly interchangeable . to see how
this can be done for the @xmath266 case , the reader is referred to takahashi s book @xcite .
the nature of charge bound states describing different sizes of atomic molecules was studied in terms of ba root patterns in the attractive regime . quantum phase diagrams and quantum phase transitions
were analytically studied from the dressed energy formalism .
we also presented the ground state energies for the strongly repulsive and strongly attractive regimes .
we found that all phase transitions for 1d @xmath1-function attractive fermions are of second order with a linear field - dependent magnetization in the vicinities of critical points , rather than a square - root field - dependent magnetization @xcite .
the linear field - dependence is as found for the su(2 ) case @xcite and for the hubbard model @xcite . for the @xmath266 case ,
the tba equations provide a comprehensive understanding of fflo pairing and finite temperature thermodynamics of tomonaga - luttinger liquids @xcite .
the key features of phase diagrams and low temperature density profiles of trapped 1d spin-1/2 fermions were experimentally confirmed by matching theoretical predictions from the tba equations , see liao _ et al .
_ @xcite and references therein . for the @xmath0 case
, these tba equations were used to study the universal thermodynamics through the derivation of the equations of state @xcite .
the results presented in this paper provide the setting for further study of quantum critical behavior in 1d interacting fermi gases , where the exact ba solutions provide insight into the physical origins of quantum criticality . | using yang and yang s particle - hole description , we present a thorough derivation of the thermodynamic bethe ansatz equations for a general @xmath0 fermionic system in one - dimension for both the repulsive and attractive regimes under the presence of an external magnetic field .
these equations are derived from sutherland s bethe ansatz equations by using the spin - string hypothesis .
the bethe ansatz root patterns for the attractive case are discussed in detail .
the relationship between the various phases of the magnetic phase diagrams and the external magnetic fields is given for the attractive case .
we also give a quantitative description of the ground state energies for both strongly repulsive and strongly attractive regimes . |
macromolecules can pass through a nanoscopic hole which is much smaller in space dimension than the bulk coil size .
given the chain - like architecture , this fact itself may not be surprising , but the way how it goes threading is interesting . today
, this mode of molecular transport , called translocation , finds its great potentiality in the multidisciplinary field of nano - technology and biological sciences @xcite , e.g. , in inventing nano - pore based rapid dna sequencing devices . in typical situations , a polymer ( with @xmath0 segments ) passes through driven by the force @xmath1 acting at the pore site .
the corresponding setup has been realized in most of experiments , where the charged polymers ( dnas ) are driven by the electric field , and the voltage drop is concentrated in the pore @xcite . from the technological design perspective , one of the most important measure is the translocation time @xmath2 , i.e. , an average passage time for a polymer to go through . for this reason ,
there have been considerable experimental attempts to quantify the translocation time .
extensive numerical simulations have also revealed a rich variety of the dynamics @xcite . under certain conditions ,
i.e. , when the specific pore - polymer interactions can be neglected , these results could be described in the light of the universality inherent in the chain - like molecules @xcite .
it is not , however , clear how and to what extent this is possible . diverse apparent " scaling laws of the translocation time reported in literatures remain to be explained @xcite . .
( c ) stem - flower [ sf ] .
( d ) strong - stretching [ ss ] .
more complete processes in respective regimes are shown in figs . 5 - 7 .
the tensed moving domain " is shaded , while the translocated part is represented by a dotted curve . ] in this paper , we find three distinct nonequilibrium regimes with the different exponents through the tension - propagation mechanism proposed in @xcite . the characteristic nonequilibrium conformations ( trumpet [ tp ] , stem - flower [ sf ] , strong - stretching [ ss ] ) of the polymeric chain with the finite extensibility are shown in fig . 1 ,
through which we establish a dynamical scaling scenario covering various models / situations and broad parameter regions relevant to real experiments . and @xmath1 ( double logarithmic scale ) .
shaded region represents the case to which the near - equilibrium theory applies @xcite .
the arrows ( i)-(iii ) indicate typical sweeping paths of parameters in experiments ( see figs .
[ fig3 ] and [ fig4 ] ) .
the boundaries @xmath3 , @xmath4 , @xmath5 separate regimes characterized by different conformations ( fig . 1 ) .
we begin by clarifying the exponents relevant to equilibrium ( static ) and near - equilibrium ( dynamic ) situations .
first , the flory exponent @xmath6 characterizes the fractality of the equilibrium polymer coil whose size is given by @xmath7 ( @xmath8 : segment size ) .
equally important is the dynamical exponent @xmath9 characterizing the dynamics of critical fluctuation " ( polymer coil ) @xcite .
normally @xmath10 in dilute solutions ( non - draining ) , but the case with @xmath11 ( free draining or rouse dynamics ) is also conceptually important .
the latter indeed corresponds to models in most ( but not all ) of numerical simulations . by denoting the translocation time @xmath12 ,
our aim is to find these dynamical exponents @xmath13 , @xmath14 relevant to the driven translocation in terms of @xmath6 and @xmath9 .
there is a growing evidence that a key element lies in the dynamical property of the polymer ; it can not respond to the driving force all at once .
this becomes evident for rather weak forces @xmath15 , i.e. , the thermal energy @xmath16 divided by the equilibrium coil size @xmath17 . in what follows
, we make the length , force and time dimensionless by the units @xmath8 , @xmath18 and @xmath19 , respectively , with @xmath20 being the viscosity of the solvent . to analyze the case @xmath21 ,
let us see the sketches of a translocating polymer where the driving force acts on the segment at the pore ( @xmath22 ) as shown in fig . 1 .
note that @xmath23-axis is taken to be perpendicular to the wall with the pore .
the one chain end is first sucked at time @xmath24 , while the polymer keeps the equilibrium conformation in @xmath25 . by the time @xmath26
, @xmath27 segments have been already translocated , and the subsequent @xmath28 segments in the anterior _ cis _ side ( shaded in fig . 1 )
are set in motion under the tension , while the remaining @xmath29 segments in the rear part are still at rest .
although plausible , a mathematical formulation of the above picture is nontrivial . in fact , at any moment , each segment moves with its own velocity and the complete characterization of the transient process would require handling highly nonlinear equations in multi - dimensional space with excluded - volume / hydrodynamic interactions .
a drastic simplification is achieved , however , by introducing the two phase picture " , i.e. , a moving and a quiescent domains separated by the tension front , and applying the steady - state approximation for the former ( fig . 1 ) .
it amounts to say that the relevant properties of the moving domain can be captured by the representative velocity @xmath30 and the steady - state conformation of the @xmath28 segments dragged at its velocity . leaving its justification in the appendix i , we remark here that there remains only single gross variable , i.e. , the position of the front , or the boundary between the moving and the quiescent domains .
this coarse - graining makes the problem analytically treatable .
then , by considering the mass conservation relation , we can show that the boundary position obeys the following equation ; @xmath31 where @xmath32 is the line density of the segment and @xmath33 is the location of the boundary ( see appendix ii for this derivation ) .
note that we first consider only the tension - propagation stage ( fig .
1(b),(c),(d ) ) since this is the dominant process giving the scaling exponents in the translocation .
in addition , @xmath34-th segment is just located in the boundary during the moving domain growth and then it still retains an initial rest position .
thus , by averaging over the initial conformation at @xmath24 , we find the relation @xmath35
here we write the _ dynamical equation of state _ relating the steady - state velocity @xmath36 , the extension @xmath37 and the driving force @xmath1 in the form ; @xmath38 the scaling form @xmath39 and the segment line density at the boundary @xmath40 are expressed through the exponents : @xmath41 @xmath42 as already mentioned , @xmath3 signifies the onset of the nonequilibrium dynamics .
moreover , we find two characteristic forces @xmath4 , @xmath43 .
this leads to the identification of three distinct nonequilibrium regimes on the dynamical diagram as summarized in fig . 2 .
the derivation and the physical meaning of eqs .
( [ d_ef_state])-([sigma_r ] ) and boundaries will be clarified in the following subsections concerning each regime .
_ trumpet ( tp ) regime _ : let us first consider the situation with the driving force in the range @xmath44 . here
the blob concepts enable one to handle the space dependent length scale , whereby provide a generic description of the deformed dragged polymer irrespective of the model details @xcite ( fig .
[ fig1](b ) ) . evaluating the dragging force and the elastic restoring force in terms of the blob size @xmath45 at each point
, we have the local force balance equation as follows : @xmath46^{-1 } \leftrightarrow \xi(x)^{2-z } \simeq v ( r + x).~~ \label{semi_local_force}\end{aligned}\ ] ] substituting @xmath22 combined with the total force balance @xmath47 , we obtain the dynamical equation of state eq .
( [ d_ef_state ] ) with eq .
( [ d_ef_state_j ] ) ( top ) , which corresponds to a generalization of the trumpet model for a tethered polymer under flow @xcite . at the boundary
, there is the largest blob of size @xmath48 centered around @xmath49 .
then , eq . ( [ semi_local_force ] ) yields @xmath50 which leads to the segment line density @xmath51 eq .
( [ sigma_r ] ) ( top ) with eq .
( [ sigma_r_vq ] ) . _
strong stretching ( ss ) regime _ : at the upper threshold of the trumpet regime @xmath52 , the number of segment in the smallest blob at the pore vicinity becomes very small @xmath53 . for stronger forces , the blob description is no longer valid , and one has to take account of the finite chain extensibility of the backbone ( e.g. , fig .
[ fig1](c),(d ) ) .
the line density such highly stretched parts is estimated as @xmath54 at the scaling level , which indicates eq .
( [ sigma_r ] ) ( bottom ) with eq .
( [ sigma_r_vq ] ) .
let us consider the limiting case of the strong driving so that the moving domain of the chain is highly stretched close to the full extension @xmath55 @xcite in the entire stage of the translocation ( fig .
[ fig1](d ) ) .
this is realized when @xmath56 ) and eq .
( [ d_ef_state_j])(bottom ) , the velocity is estimated as @xmath57 when the tension arrives at another chain end in the ss regime , so that the drag force acting on the rear end segment is @xmath58 ( @xmath59 in the original unit ) at that moment .
the ss shape moreover requires @xmath60 to retain the strong stretching even at rear end segment .
eventually , by combining it with @xmath61 , we find the lower bound as @xmath62 for ss regime . ] .
neglecting the logarithmic correction due to the hydrodynamic interactions , the total drag force is given by @xmath63 , and then the scaling form of the dynamical equation of motion is specified by eq .
( [ d_ef_state_j ] ) ( bottom ) irrespective of the solvent property ( free - draining or non - draining ) .
_ stem - flower ( sf ) regime _ : in the intermediate between above two regimes @xmath64 , the moving domain takes a highly stretched conformation only in the early stage , for which the analysis in the ss regime applies .
this point will be discussed in the section vi again .
the moving domain grows with time , and the velocity @xmath36 decreases towards @xmath65 , at which the tension at the boundary becomes comparable to @xmath66 , that is a blooming point " , before which the moving domain composes of stem only . from then on ,
the conformation of the moving domain can be described as a stem - flower ( fig .
[ fig1 ] ( c ) ) . compared to the ss regime
, the chain in sf regime has smaller extension because of the coiling at the rear part .
still , scaling exponents associated with the entire domain in the sf regime is indistinguishable from that in the ss regime @xcite and satisfies @xmath67 $ ] and @xmath68 $ ] .
consequently , we find eq .
( [ d_ef_state_j ] ) ( bottom ) ; its distinction with that in the ss regime lies in a smaller numerical prefactor , while the conformation at the rear can be described by the tp model ( eq .
( [ sigma_r ] ) ( top ) ) .
therefore , the sf regime shares ( i ) the dynamical equation of state with the ss regime , ( ii ) the segment line density at the boundary with the tp regime . in the non - draining case ( double logarithmic scale ) .
( a ) and the upper / lower plots in ( b ) correspond to the parameter sweepings indicated by arrows ( i ) and ( ii)/(iii ) in fig .
[ fig2 ] . ] .predicted exponents @xmath13 and @xmath14 in the non - draining case ( @xmath10 ) .
note that @xmath69 . [
cols="^,^,^ " , ] [ table2 ] our main results are summarized in figs .
[ fig3 ] and [ fig4 ] , where we plot @xmath70 in the non - draining ( @xmath10 ) and the free - draining ( @xmath11 ) cases , respectively .
these are expected to be valid for long chains , and indeed capture a number of observations made in numerical / real experiments .
we also listed the concrete values of exponents for cases @xmath71 and @xmath72 in tables 1 and 2 .
as a primal message , we stress that the identification of the appropriate regime ( depending on @xmath0 and @xmath1 ) as well as exponents @xmath6 , @xmath9 characterizing the equilibrium polymer coil is of a crucial importance for the proper comparison with experiments .
for example , we first note that most simulations treat free - draining chains and the typical driving force is on the order of @xmath73 ( tp and/or sf regimes ) , and the reported values of @xmath13 exponent are scattered around @xmath74 in three dimension ( @xmath75 ) . in particular ,
recent high - accuracy simulation results @xmath76 ( langevin dynamics ) and @xmath77 ( atomistic molecular dynamic ) @xcite are in excellent agreement with our prediction @xmath78 .
the change of @xmath13 exponents with the increase in @xmath1 and/or with the inclusion of hydrodynamic interactions has been examined in ref .
@xcite , whose qualitative trend accords with ours . in experiments using double - stranded dna @xcite
, the non - draining approximation with @xmath71 would be appropriate @xcite , and the typical driving force range @xmath79 would correspond to sf or even ss regime for short dna chains . then , the reported values @xmath80
@xcite and @xmath81 @xcite are in good agreement with our prediction . on the other hand ,
much less attention has been paid to the @xmath14 exponents .
usually , it has been implicitly assumed to be @xmath82 , but a recent study has claimed @xmath83 for strong force @xcite in reasonable agreement with our prediction @xmath84 in sf regime ( @xmath75 , free - draining ) . is the junction point between stem and flower ) . ]
how can we go one step further beyond the leading term scaling argument ?
practically , resolving complications associated with various finite - size and/or crossover effects is an important issue , for which we point out two main sources .
the translocation time can be formally written as @xmath85 .
figures 5 , 6 and 7 show the full story of the driven translocation process in tp , sf and ss regimes , respectively . for long chains , the tension propagation stage ( fig .
5(b)@xmath86(d ) , fig .
6(c)@xmath86(e ) , fig . 7(b)@xmath86(d ) ) dominates the whole process in the nonequilibrium driven regimes ( @xmath87 ) , thus , @xmath88 represents the translocation time @xmath89 . for shorter chains ,
however , this might be severely affected by ( i ) the post - propagation stage ( @xmath90 ) and/or ( ii ) the effect associated with the initial conditions in eq .
( [ r_diff_eq ] ) ( @xmath91 and @xmath92 ) . for the point ( i )
, we can describe the process after @xmath93 by setting @xmath94 in eq .
( [ mass_free ] ) as follows : @xmath95 where @xmath96 is the location of the rear end and approaching to the hole ( fig .
5(d)@xmath86(f ) , fig .
6(e)@xmath86(h ) , fig .
7(d)@xmath86(f ) ) with @xmath97 .
the estimate for its characteristic duration time @xmath98 is obtained by setting @xmath99 , thus , @xmath100 with the exponent @xmath101 given by eq .
( [ d_ef_state_j ] ) . the point ( ii ) can be taken into account by the following estimates ; @xmath102 @xmath103 the first one ( @xmath104 ) comes from the formation of the initial tensed blob of size @xmath105 ( fig .
5(b ) ) , which becomes the microscopic time @xmath106 at the threshold @xmath107 .
this microscopic time corresponds to @xmath108 in the ss regime as in fig .
7(b ) . in the sf regime ( @xmath109 )
, we choose @xmath91 to be the blooming time , i.e. , the moment when the flower part starts to appear ( fig .
before that , there is only a stem part in the early period ( fig .
6(b ) @xmath86 ( c ) ) during which the analysis for the ss regime applies after the microscopic relaxation time of first tensed segment @xmath110 as in fig .
these lead to the time @xmath111 after which the coiled flower - conformation appears at the rear part ( @xmath112 , @xmath113 are the exponents for ss regime ) . here , by denoting the stem length as @xmath114 ( fig .
6(d ) ) , the pertinent segment number is @xmath115^{1/\nu}$ ] from eq .
( [ front_1 ] ) , where the stem length at the beginning of the late sf regime is @xmath116 ( see appendix iii )
. solving eq .
( [ r_diff_eq ] ) with these initial conditions results in the shortening of @xmath88 .
in particular , near the thresholds @xmath117 and @xmath118 , we find @xmath119 , thus , @xmath88 become vanishingly small . in the former , @xmath91 is already close to @xmath120 . in the latter case ,
the contributions from @xmath91 and @xmath90 are comparable , and @xmath90 crossovers to @xmath121 , which is a result from the equilibrium shape assumption @xcite .
these factors may result in the narrowing of the tp regime and , in general , cause the rounding - off the crossover at @xmath122 as well as @xmath123 . to summarize this section , one finds by comparing eqs .
( [ tau_p ] ) , ( [ tau_pp ] ) and ( [ tau_ini ] ) that the propagation stage dominates , i.e. , @xmath89 for large @xmath124 limit in driven translocation regime ( @xmath87 ) . for real systems with small @xmath124 , however , the finite - size correction might be apparent . in particular , @xmath90 may become dominant for short chains in two dimension , thus , an appreciable crossover scaling with respect to @xmath124 would be expected as pointed out in ref .
@xcite . on the other hand , for weaker forces ( @xmath125 )
, the equilibrium shape assumption would be valid , and @xmath126 .
we note , however , that in this weaker forces ( @xmath127 ) , the segment returning through the pore becomes apparent , and this may alter the exponent in @xmath128 @xcite , while such an effect is negligible under the strong driving force .
once again , we emphasize that the key lies in the nonequilibrium transient response property inherent in a long flexible chain . our strategy to implement
this is a two phase picture and applying a steady state approximation for the moving domain . before closing , let us make several remarks on the derived exponents .
( i ) exponents in the translocation time @xmath13 , @xmath14 reflect the exponent @xmath101 , @xmath129 characterizing the steady state conformation of the dragged chain ( eqs .
( [ d_ef_state])-([sigma_r ] ) ) .
we have adopted eqs .
( [ d_ef_state_j ] ) , ( [ sigma_r ] ) , which is correct in the scaling limit , for the purpose of the general discussion , but a care should be taken in the comparison with real experiments . in particular , the logarithmic correction due to the hydrodynamic interactions ( in non - draining case ) is neglected , and this might alter the effective exponent @xmath101 , @xmath129 , thus , @xmath13 , @xmath14 , too , in the certain range of the interest @xcite . although the results in ref .
@xcite are in good agreement with our prediction , further careful checks are awaited on it .
( ii ) as stated in appendix i , the steady state approximation is by construction marginally valid to capture the salient features of the driven translocation , but there is some room to improve the theory by incorporating a slight deviation from the approximation .
( iii ) the presence of several different regimes and the finite - size effects makes the accurate determination of the exponents very difficult ( see also the remark ( i ) , ( ii ) above ) .
we expect , however , that it would be possible to capture a global trend in the exponents by carefully analyzing numerical / experimental data obtained in widely different conditions .
( iv ) aside from exponents , the proposed physical picture is generic , and we believe that our formalism constitutes a minimal model , or a prototype in the problem . from our point of view
, it is more important to comprehend the underlying physics ( even qualitatively ) , and seek for possible improvements of theory to describe the phenomena more adequately ( see , for example , a recent attempt by rowghanian and grosberg @xcite ) .
( v ) one of the interesting predictions concerns the beta exponent ; it is not necessarily unity in contrast to claims in most other works .
we expect that such a trend ( even qualitatively ) could be detected by careful experiments ; its reliable trend should serve as a stringent test for any theory of the driven translocation .
in summary , we proposed dynamical scaling scenario applicable to the broad range of situations met in the study of polymer driven translocation .
the paper was devoted to the detailed account of tension - propagation mechanism proposed in @xcite and its extension to cover the wider parameter space through the distinct nonequilibrium conformations . also shown
are the finite size effect , various associated crossovers and their rounding off .
we believe that these elements operate as a guide for further studies towards comprehensive understanding of the driven translocation .
last of all , we expect that the emergent physical picture may be relevant to many other polymeric systems driven by local and strong stimuli .
best examples would be found in the micro - manipulation experiments as well as the biopolymer functions in cells . since @xmath34-th segment at the front
have just taken part in the moving domain and the velocity profile would be smooth even around the front , there might be substantial velocity gradient there .
if so , its magnitude can be estimated as @xmath130 , whose @xmath48 is the size of the largest blob at the front ( eq . ( [ rear_blob ] ) ) and can be regarded as the boundary layer . comparing the shear rate with the corresponding longest relaxation time @xmath131 , we find @xmath132 using eqs .
( [ rear_blob ] ) .
thus the concerned shear rate is marginal and one can apply the steady state approximation . as another possibility
, one may suppose that an uniform velocity gradient is entirely spread in the moving domain with the shear rate given by @xmath133 .
if so , the longest relaxation time in the entire tensed chain is @xmath134 , and we similarly have the relation as @xmath135 @xcite . here
@xmath136 was estimated as follows ; the dynamical equation of state ( [ d_ef_state ] ) @xmath137 implies the spring constant @xmath138 and the viscous frictional coefficient @xmath139 , which lead to the relaxation time @xmath140 . although obtaining real velocity profile in the moving domain requires a more thorough analysis , which is beyond the scope of the present study , one may expect it to be something between the above two extremes .
we can therefore expect the steady state approximation would be valid to address the salient features in the driven translocation .
* appendix ii.mass conservation : * here we discuss the derivation of eq .
( [ mass_free ] ) from the integral form of the mass conservation in the moving domain ; @xmath141 taking account of the moving boundary @xmath96 associated with the tension propagation , the time derivative of eq .
( a-1 ) leads to @xmath142}{{\rm d}t } & = & \int_{-r(t)}^{0 } \frac{\partial \sigma(t , x ) } { \partial t } { \rm d}x + \sigma ( t ,- r ) \frac{{\rm d}r}{{\rm d}t } \nonumber \\ & = & j ( t ,- r ) -j(t,0 ) + \sigma ( t ,-
r ) \frac{{\rm d}r}{{\rm d}t } , \nonumber \\ & & ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~({\rm a-2})\nonumber \label{mass_derivation}\end{aligned}\ ] ] where the continuity equation @xmath143 is used in the last equality .
the pore current in eq .
( a-2 ) would be strongly dependent on the pore specificity , but can be eliminated by its definition @xmath144
. then , eq . ( [ mass_free ] ) follows by assuming @xmath145 , where @xmath30 is the steady state velocity of the moving domain . under this assumption
, the number of the translocated segments @xmath27 should be calculated from eq .
( a-1 ) , not from time integration of @xmath146 .
the latter quantity does not satisfy the conservation of mass as @xmath30 characterizing the chain deformation slightly differs from the velocity of the @xmath27-th segment at the pore .
we note that a better estimate of the velocities @xmath147 at the boundary ( due to the slight derivation from the representative velocity @xmath30 ) would improve the result .
* appendix iii.blooming point condition in the sf regime : * at the end of the early sf regime ( composed of stem only ) , the following conditions are met for the stem length @xmath37 and the moving velocity @xmath148 these are respectively the total force balance and the force balance at the rear end with the blob of size @xmath149 . from these ,
one obtains @xmath150 , @xmath151 or in the dimensionless form @xmath152 , @xmath153 . | by analyzing the real space nonequilibrium dynamics of polymers , we elucidate the physics of driven translocation and propose its dynamical scaling scenario analogous to that in the surface growth phenomena .
we provide a detailed account of the previously proposed tension - propagation formulation and extend it to cover the broader parameter space relevant to real experiments ; in addition to a near - equilibrium regime , we identify three distinct nonequilibrium regimes reflecting the steady - state property of a dragged polymer with the finite extensibility .
finite size effects are also pointed out .
these elements are shown to be crucial for the appropriate comparison with experiments and simulations . |
carbon and oxygen are two of the most abundant elements in the universe and lines from these elements provide valuable plasma diagnostics for almost all classes of cosmic sources .
essential for many of these diagnostics are accurate electron - ion recombination rate coefficients , particularly of dielectronic recombination ( dr ) , which for most ions in electron - ionized plasmas is the dominant means of electron capture ( arnaud & rothenflug 1985 ) .
producing accurate theoretical dr rate coefficients is , however , theoretically and computationally challenging . in the past ,
semi - empirical expressions such as the burgess ( 1965 ) formula along with modified versions by burgess & tworkowski ( 1976 ) and merts et al .
( 1976 ) were developed to calculate dr rates .
more recently , a number of more sophisticated theoretical approaches have been used to calculate dr , among them single - configuration @xmath0-coupling ( bellantone & hahn 1989 ) , multiconfiguration intermediate - coupling ( pindzola , badnell , & griffin 1990 ) , and multiconfiguration fully - relativistic ( chen 1988 ) techniques , as well as undamped and damped , unified radiative recombination ( rr ) and dr calculations in @xmath0-coupling ( nahar & pradhan 1997 ; nahar 1999 )
. approximations , though , need to be made to make any of these techniques computationally tractable ( hahn 1993 ) .
currently , sophisticated dr calculations are non - existent for many ions , and in the absence of anything better , semi - empirical formulae are often still used for plasma modeling .
laboratory measurements can be used to test the different theoretical and computational techniques for calculating dr .
recently , savin et al . ( 1997 , 1999 ) developed a technique for obtaining rate coefficients from laboratory measurements of dr resonance strengths and energies .
they successfully used this technique to derive rates for @xmath2 dr of fe xviii and fe xix and to benchmark existing theoretical calculations . here
, we describe this technique in detail for the first time and apply it to recent dr measurements in c v and o viii .
kilgus et al .
( 1990 , 1993 ) and mannervik et al .
( 1997 ) have measured the resonance strengths and energies for dr of c v to c iv and o viii to o vii .
we use their results to produce dr rate coefficients to benchmark existing c v and o viii dr calculations and to provide rates for use in plasma modeling . in electron - ionized plasmas , lines from heliumlike c
v and hydrogenic o viii trace gas at @xmath3 k and @xmath4 k , respectively ( arnaud & rothenflug 1985 ; mazzotta et al .
c v and o viii lines have been observed in solar spectra ( doschek & cowan 1984 ) and o viii lines in supernova remnants ( winkler et al .
1981 ) . and with the upcoming launches of _ chandra _ and _ xmm _ and the high - resolution spectrometers aboard , c v and o viii lines are expected to be seen in may other electron - ionized , cosmic sources . using different heavy - ion storage rings ,
kilgus et al . (
1993 ) and mannervik et al .
( 1997 ) have measured dr for c v via the capture channels @xmath5 where @xmath6 for the results of kilgus et al . and @xmath7 for the results of mannervik et al .
kilgus et al .
( 1990 ) have also measured dr for o viii via the capture channels @xmath8 where @xmath9 .
the radiative stabilization of these autoionizing c v and o vii states to bound configurations results in dr .
details of the experimental techniques used are given in the references cited .
the paper is organized as follows : we describe in section [ sec : methodofcalculation ] how one produces a dr rate coefficient using measured dr resonance strengths and energies . in section [ sec : resultsanddiscussion ] we present the resulting rate coefficients and compare the derived dr rates with published theoretical rates . we also give a simple fitting formula for use in plasma modeling .
dr is a resonance process consisting , in the zero - density limit , of an infinite number of resonances .
the dr rate coefficient @xmath10 for a plasma with a maxwell - boltzmann electron distribution is given by @xmath11 where @xmath12 is the electron temperature ; @xmath13 is the energy - dependent dr cross section for resonance @xmath14 ; @xmath15 is the relative electron - ion velocity at energy @xmath16 , which is taken to be the electron energy as the ions are nearly to stationary in the center - of - mass frame ; and the sum is over all dr resonances .
the maxwell - boltzmann distribution @xmath17 is given by @xmath18 where @xmath19 is the boltzmann constant .
kilgus et al .
( 1990 , 1993 ) and mannervik et al .
( 1997 ) published measured dr resonance strengths @xmath20 and energies @xmath21 .
the dr resonance strength is defined @xmath22 where @xmath23 is the cross section for a resonance or group of resonances labeled @xmath14 and \{@xmath24,@xmath25 } is a region in energy chosen such that it contains only those resonances comprising @xmath14 .
here we are interested in calculating rate coefficients .
this involves convolving the dr resonances with the slowly varying function @xmath17 . because the energy widths of the measured resonances are smaller than the scale over which @xmath17 changes , for our purposes we can accurately approximate @xmath13 as @xmath26 where @xmath21 is the energy of resonance @xmath14 and @xmath27 is the dirac delta function .
the dr rate coefficient for maxwellian plasmas is found by substituting equation [ eq : cross1 ] into equation [ eq : rate1 ] which yields @xmath28 kilgus et al .
( 1993 ) and mannervik et al .
( 1997 ) do not report measured resonance energies for capture by c v into levels where @xmath29 . to calculate these resonance energies
@xmath30 we use the rydberg formula @xmath31 where @xmath32 is the charge of the ion before recombination , @xmath33 ev is the energy of the @xmath34 core excitation ( kelly 1987 ) , and @xmath35 is the rydberg constant .
for o viii , @xmath36 and @xmath37 ev is the energy of the @xmath38 core excitation ( kelly 1987 ) .
mannervik et al . (
1997 ) estimate that they measured dr for capture into levels @xmath39 .
kilgus et al .
( 1993 ) do not report a value of @xmath40 , so we derive an estimate here . using their ion energy , the bending radius of the dipole magnets in their experiment ( linkemann 1995 ) , and the semiclassical formula for field ionization ( brouillard 1983 ) , we estimate electrons captured into levels where @xmath41 will be field ionized and thus not detected . however , the captured electrons can radiatively decay below @xmath42 during the @xmath43 m distance they travel between the electron - ion interaction region in the experiment and the dipole magnet ( cf . , kilgus et al . 1992 ; savin et al .
1997 ) . using the ion velocity , the hydrogenic formula for radiative lifetimes of marxer & spruch ( 1991 ) , and calculations which show that dr for heliumlike ions essentially populates only angular momentum levels where @xmath44 ( chen 1986 )
, we estimate that electrons captured into levels @xmath45 will radiatively decay below @xmath42 and thus were detected by kilgus et al . for oviii , kilgus et al .
( 1990 ) estimate @xmath9 . for a given dr series , as @xmath46 increases the energy spacing between dr resonances decreases . due to the energy resolutions of the experiments , above some @xmath46 level
it is not possible to resolve the dr resonances and determine individual values of @xmath47 . for these high @xmath46 levels , kilgus et al .
( 1990 ; 1993 ) presents total resonance strengths summed from @xmath48 to @xmath40 for c v and from @xmath49 to @xmath40 for o viii .
we divide and spread out these summed resonance strengths into @xmath50 ev wide bins when using equation [ eq : rate2 ] .
the `` resonance '' energies of these bins are chosen to lie between @xmath51 and @xmath52 for c v and between @xmath53 and @xmath52 for o viii .
the bin widths are smaller than the scale over which @xmath17 changes and the small errors in the exact resonance energies of these high @xmath46 levels has an insignificant affect on the derived dr rates .
in figures [ fig : cv ] and [ fig : oviii ] we present the c v and o viii dr rate coefficients , respectively , derived using the measured resonance strengths and energies of kilgus et al .
( 1990 , 1993 ) and mannervik ( 1997 ) .
the unmeasured contribution to the dr rate due to capture into levels where @xmath54 is predicted to be insignificant ( chen 1986 ; pindzola et al .
the absolute uncertainties in the derived dr rates are estimated to be @xmath55 which corresponds to the reported absolute experimental uncertainties .
existing theoretical dr rates are also shown in figures [ fig : cv ] and [ fig : oviii ] as well as the rr rates of verner & ferland ( 1996 ) . for c
v , the single - configuration @xmath0-coupling calculations of bellantone & hahn ( 1986 ) and romanik ( 1988 ) , multiconfiguration intermediate - coupling calculations of badnell , pindzola , & griffin ( 1990 ) , and multiconfiguration fully - relativistic calculations of chen ( 1988 ) are all in good agreement with our experimentally inferred rate .
the @xmath0-coupling calculations of younger ( 1983 ) are @xmath56 larger than the experimental rate .
the recommended and commonly used dr rate of shull & van steenberg ( 1982 ) peaks at a value @xmath57 larger than ours and has a steeper low @xmath12 behavior .
pindzola et al . (
1990 ) calculated resonance strengths for o viii using intermediate - coupling , multiconfiguration hartree - fock ( mchf ) and intermediate - coupling , multiconfiguration thomas - fermi ( mctf ) techniques . using their mchf results for capture into levels @xmath58 and their mctf results for @xmath59 , as well as the experimental resonance energies , we calculate the corresponding dr rate using equation [ eq : rate2 ] .
the resulting rate is in good agreement with our derived rate .
the single - configuration @xmath0-coupling rate of bellantone & hahn ( 1986 ) is also in good agreement with the experimental rate , though with decreasing @xmath12 their rate falls off sooner than ours .
the recommended and commonly used dr rate of shull & van steenberg ( 1982 ) is @xmath60 larger than our derived rate .
not shown in figure [ fig : oviii ] , for reasons of clarity , is the rate of zhdanov ( 1978 ) whose peak rate is @xmath61 times larger than the experimental rate .
nahar & pradhan ( 1997 ) and nahar ( 1999 ) present unified rr+dr rates . to compare with their results
we add the rr rates of verner & ferland ( 1996 ) to the experimentally derived dr rates .
thus we treat rr and dr as independent processes and do not allow for the possibility of interference between the two recombination channels .
the validity of this approach is supported by recent theoretical work ( for dr on systems ranging in complexity from c ii and mg ii to u xcii ) which has shown the effect of interferences to be small ( @xcite ) .
figures [ fig : cvnah ] and [ fig : oviiinah ] show for c v and o viii , respectively , the unified rr+dr rate of nahar & pradhan and the total rr+dr rate using the derived dr results and the rr results of verner & ferland . for c
v , at peak value the undamped , unified rate of nahar & pradhan is @xmath62 larger than our resulting total rr+dr rate .
this is consistent with the estimate by pradhan & zhang ( 1997 ) that allowance for radiation damping would reduce the undamped rate by @xmath63 . for o viii , at peak value the damped , unified rate of nahar is @xmath64 larger than our resulting total rr+dr rate .
dr rate coefficients are sometimes calculated using the burgess ( 1965 ) general formula ( gf ) or versions of the gf as modified by merts et al . ( 1976 ) and by burgess & tworkowski ( 1976 ) .
we have calculated the burgess and merts et al .
rates using the formulae as given by cowan ( 1981 ) .
we use oscillator strengths and excitation energies from wiese , smith , & glennon ( 1966 ) . shown in figure
[ fig : cvburg ] for c v is the merts et al .
rate which is in fair agreement with our derived rate , though with decreasing @xmath12 it goes to zero sooner than the experimental rate . also shown in figure [
fig : cvburg ] is the burgess gf rate which is a factor of @xmath65 times larger than the experimental rate .
the burgess & tworkowski ( 1976 ) modification of the gf yields a rate @xmath66 larger than our derived rate .
this is surprisingly good considering that their modification is meant for dr forming heliumlike , not lithiumlike , ions .
figure [ fig : oviiiburg ] shows for o viii the gf rate which is @xmath67 larger than the experimental rate and , considering the expected accuracy of the burgess formula , in fair agreement .
the burgess & tworkowski rate is in good agreement .
also shown is the merts et al .
formula rate which is a factor of @xmath65 smaller than the experimental rate .
our c v and o viii results strongly suggest that for @xmath1 dr it is not possible _ a priori _ to know which formula will yield a result closer to the true rate . for use in plasma modeling
, we have fit the experimentally derived c v and o viii dr rates using the simple formula ( arnaud & raymond 1992 ) @xmath68 here @xmath69 and @xmath70 are , respectively , the strength and energy parameters for the @xmath71th fitting component .
best fit values are listed in table [ tab : fitparameters ] . for c
v , the fit is good to better 1% for @xmath72 k. below @xmath73 k , with decreasing @xmath12 the fit goes to zero faster than the derived rate . but this error is insignificant for plasma modeling as the rr rate is @xmath74 orders of magnitude larger than the dr rate at these temperatures . for o
viii , the fit is good to better 1% for @xmath75 k. below @xmath76 k , with decreasing @xmath12
the fit goes to zero faster than the derived rate .
this error is insignificant for plasma modeling as the rr rate is @xmath77 orders of magnitude larger than the dr rate at these temperatures .
we have presented a simple technique for obtaining dr rate coefficients from laboratory measurements of dr resonance strengths and energies . with this technique ,
we have derived dr rates for c v to c iv and o viii to o vii using published resonance strengths and energies .
our derived rates are in good agreement with multiconfiguration , intermediate - coupling and multiconfiguration , fully - relativistic calculations as well as with most @xmath0-coupling calculations .
our rates are not in agreement with the recommended dr rates commonly used for astrophysical plasma modeling .
we have used theoretical radiative recombination ( rr ) rates in conjunction with our derived dr rates to produce a total recombination rate for comparison with unified rr+dr calculations in @xmath0 coupling .
our results are not in agreement with undamped , unified calculations for c v but are in reasonable agreement with damped , unified calculations for o viii . also , neither the burgess general formula , the merts et al .
formula , nor the burgess & tworkowski formula consistently yield a rate which is in agreement with our derived rates .
this suggests that for @xmath78 dr it is not possible to know _ a priori _ which formula will yield a rate closer to the true dr rate .
we have also presented simple fitting formula of the experimentally derived dr rates for use in plasma modeling .
the author would like to thank n. r. badnell , m. h. chen , t. w. gorczyca , s. m. kahn , d. a. liedahl , s. n. nahar , and a. wolf for stimulating conversation and for critically reading the manuscript .
this work was supported in part by nasa high energy astrophysics x - ray astronomy research and analysis grant nag5 - 5123 .
wiese , w. l. , smith , m. w. , & glennon , b. m. 1966 , atomic transition probabilities , volume i hydrogen through neon , national standard reference data series , national bureau of standards 4 ( u. s. government printing office , washington , d.c . ) | using published measurements of dielectronic recombination ( dr ) resonance strengths and energies for c v to c iv and o viii to o vii , we have calculated the dr rate coefficient for these ions .
our derived rates are in good agreement with multiconfiguration , intermediate - coupling and multiconfiguration , fully - relativistic calculations as well as with most @xmath0 coupling calculations .
our results are not in agreement with the recommended dr rates commonly used for modeling cosmic plasmas .
we have used theoretical radiative recombination ( rr ) rates in conjunction with our derived dr rates to produce a total recombination rate for comparison with unified rr+dr calculations in @xmath0 coupling .
our results are not in agreement with undamped , unified calculations for c v but are in reasonable agreement with damped , unified calculations for o viii .
for c v , the burgess general formula ( gf ) yields a rate which is in very poor agreement with our derived rate .
the burgess & tworkowski modification of the gf yields a rate which is also in poor agreement .
the merts et al .
modification of the gf yields a rate which is in fair agreement . for
o viii the gf yields a rate which is in fair agreement with our derived rate .
the burgess & tworkowski modification of the gf yields a rate which is in good agreement . and the merts et al .
modification yields a rate which is in very poor agreement .
these results suggest that for @xmath1 dr it is not possible to know _ a priori _ which formula will yield a rate closer to the true dr rate .
we describe the technique used to obtain dr rate coefficients from laboratory measurements of dr resonance strengths and energies . for use in plasma modeling
, we also present easy - to - use fitting formulae for the experimentally derived dr rates . |
layered structures including graphite , hexagonal boron nitride , black phosphorus , transition metal dichalcogenides such as mos@xmath0 , and oxides including v@xmath0o@xmath1 are very common in nature .
the possibility to form stable two - dimensional ( 2d ) structures by mechanical exfoliation of these structures appears very attractive for a variety of applications .
the most prominent example of such 2d systems , graphitic carbon , is the structural basis not only of graphene,@xcite but also fullerenes , nanotubes , tori and schwarzites . even though the structural motif in all of these systems may be the same , their mechanical and electronic properties depend sensitively on the local morphology .
not only does the natural abundance of structural allotropes and isomers reflect their net energetic stability , but also the relative chemical reactivity of specific sites in a given structure correlates well with the local curvature and local stability .
this relationship has been well established for the reactive sites in the c@xmath2 fullerene,@xcite used to induce structural collapse leading to chemical unzipping of carbon nanotubes , and to destroy collapsed carbon nanotubes.@xcite and the gaussian curvature @xmath3 on the surface of ( a ) a sphere , ( b ) a cylinder and ( c ) in a saddle point .
( d ) determination of the local curvature at point @xmath4 using the atomic lattice and the dual lattice .
[ fig1 ] ] for very large structures , estimating the global or local stability using _ ab initio _ calculations has proven impracticable . there
, the stability has often been estimated using empirical rules or parameterized force fields including the tersoff potential and molecular mechanics , with sometimes unsatisfactory results .
application of continuum elasticity theory , which can describe stability changes due to deviation from planarity , has been successful , but limited to systems with a well - defined , constant curvature .
since strain energy is local and independent of the global morphology , it is intriguing to explore , whether the local deformation energy may be accurately determined from local morphology estimates using the atomic geometry .
if so , then the local stability in even arbitrarily shaped structures could be estimated accurately . here
we propose a fast method to determine the local curvature in 2d systems with a complex morphology using the local atomic geometry .
curvature information alone , combined with elastic constants obtained for a planar system , provides accurate stability estimates in the framework of continuum elasticity theory .
we find that relative stabilities of graphitic structures including fullerenes , nanotubes and schwarzites , as well as phosphorene nanotubes , calculated using this approach , agree closely with _
ab initio _ density functional calculations
. the continuum elasticity approach can be applied to all 2d structures and is particularly attractive in complex systems with known structure , where the quality of parameterized force fields has not been established .
the local curvature at a particular location on a surface is given by the two principal radii of curvature @xmath5 and @xmath6 , as shown in fig .
[ fig1 ] . on a spherical surface , @xmath7 . on a cylindrical surface
, @xmath5 is the cylinder radius and @xmath8 .
finally , a saddle point on a surface is characterized by opposite signs of @xmath5 and @xmath6 . knowing the principal radii of curvature everywhere , we may use continuum elasticity theory to determine the curvature energy @xmath9 with respect to a planar layer using@xcite @xmath10 here , the integral extends across the entire closed surface , @xmath11 is the flexural rigidity and @xmath12 is the poisson ratio .
simple expressions for @xmath9 can be obtained for simple morphologies such as a sphere or a cylinder , where @xmath5 and @xmath6 are constant everywhere.@xcite this is , however , not the case in general .
we find it convenient to introduce the local mean curvature @xmath13 and the local gaussian curvature @xmath14 using these quantities , we can rewrite eq .
( [ eq1 ] ) as @xmath15 \;. \label{eq4}\ ] ] in the following , we will consider the equilibrium arrangement of atoms in a planar 2d structure as the reference structure and will determine the local curvature from changes in the local morphology .
the discrete counterpart of eq .
( [ eq4 ] ) for the curvature energy @xmath9 is a sum over atomic sites @xmath16 , @xmath17 \ ; , % \label{eq5}\]]where @xmath18 is the area per atom . to use eq .
( [ eq5 ] ) for curvature energy estimates , we need to know the local curvatures @xmath19 and @xmath3 at all atomic sites .
our approach to estimate these values at a given site @xmath4 is illustrated in fig .
[ fig1](d ) . according to eq .
( [ eq2 ] ) , the local mean curvature @xmath19 should be close to the average inverse radius of curvature at that point , @xmath20since the atomic site @xmath4 and its nearest three neighbors @xmath21 , @xmath22 and @xmath23 define the surface of a sphere of radius @xmath24 , we take @xmath25 . the positions of four atoms do not allow to distinguish , whether @xmath4 is on a plane , a sphere , a cylinder , or in a saddle point .
we may obtain this additional information using the concept of angular defect . on any surface , which can be triangulated as shown in fig .
[ fig1](d ) , the angular defect at a representative vertex @xmath26 is defined by @xmath27 in radian units . the local gaussian curvature at @xmath26 is then given by@xcite @xmath28 where @xmath29 is the total area of the triangulated surface divided by the number of vertices . for trivalent molecular graphs containing 5- , 6- and 7-membered rings found in fullerenes , carbon nanotubes and schwarzites ,
a unique triangulation may be obtained by connecting the centers of adjacent polygons .
this method is referred to as the dual graph in graph theory@xcite and its use is illustrated in fig .
[ fig1](d ) .
since @xmath4 is not a vertex in the dual graph , but rather the center of the triangle @xmath30 , we must infer the local gaussian curvature at @xmath4 from the angular defects at @xmath26 , @xmath31 and @xmath32 .
if vertex @xmath33 is surrounded by @xmath34 triangles , we may assign to point @xmath4 the angular defect @xmath35 .
then , we can estimate the local gaussian curvature at @xmath4 as @xmath36 where @xmath18 is the average area per atom .
we use @xmath37 @xmath38 , the value found in the honeycomb lattice of graphene , for all graphitic structures .
the above definition of the local gaussian curvature satisfies exactly the equality @xmath39 here , @xmath40 is the euler characteristic of the surface , given by @xmath41 , where @xmath42 is the genus , meaning ` number of holes ' . of interest here
is the fact that @xmath43 for spherical objects like fullerenes and @xmath44 for cylindrical objects such as nanotubes .
equation ( [ eq11 ] ) is the discretized version of the gauss - bonnet theorem@xcite regarding the integral of the gaussian curvature over an entire closed surface , called the sum of the defect , which is usually formulated as @xmath45 .
the variation of the local gaussian curvature @xmath3 and the local curvature energy @xmath46 across the surface of carbon polymorphs , including two fullerene isomers discussed in section [ sec3].b , a nanotube and a schwarzite structure , is displayed in fig .
the local curvature energy in these @xmath47-bonded structures has been evaluated using the elastic constants of graphene@xcite @xmath48 ev and @xmath49 .
the higher stability of the c@xmath50(17 ) isomer in fig .
[ fig2](b ) is reflected in a rather uniform curvature energy and gaussian curvature distribution .
the low stability of the c@xmath50(2 ) isomer in fig .
[ fig2](a ) is reflected in a large variation of curvature energy and gaussian curvature , clearly indicating the most reactive sites .
cylindrical carbon nanotubes , such as the ( 10,10 ) nanotube displayed in fig .
[ fig2](c ) , have zero gaussian curvature and a constant curvature energy caused by the mean curvature .
schwarzites such as the c@xmath51 structure , displayed in fig .
[ fig2](d ) , have only negative gaussian curvature that may vary across the surface , causing variations in the local curvature energy .
( left panels ) and local curvature energy @xmath52 across the surface of ( a ) the least stable c@xmath50 isomer , ( b ) the most stable c@xmath50 isomer , ( c ) a ( 10,10 ) carbon nanotube , and ( d ) a schwarzite structure with 152 atoms per unit cell .
the values of @xmath3 and @xmath52 have been interpolated across the surface .
we will next test the accuracy of the continuum elasticity approach by calculating the relative stability of non - planar structures based on graphitic carbon .
an infinite number of morphologies including nanotubes , fullerenes and schwarzites may be produced by deforming a segment of a graphene layer and reconnecting its edges so that all carbon atoms are threefold coordinated . in many cases ,
the non - planar structures contain carbon pentagons and heptagons in the graphitic honeycomb arrangement of atoms as required by euler s
theorem.@xcite to validate the continuum elasticity theory results , we calculated the total energy of a graphene monolayer and selected graphitic structures using _ ab initio _ density functional theory ( dft ) as implemented in the siesta code.@xcite we used the local density approximation ( lda ) and perdew - burke - ernzerhof ( pbe)@xcite exchange - correlation functionals , norm - conserving troullier - martins pseudopotentials@xcite , and a double-@xmath53 basis including polarization orbitals .
the 1d brillouin zone of nanotubes was sampled by @xmath54 @xmath19-points and the 2d brillouin zone of graphene by @xmath55 @xmath19-points.@xcite the small brillouin zones of schwarzites with several hundred c atoms per unit cell were sampled by only @xmath56 @xmath19-point .
we used a mesh cutoff energy of @xmath57 ry to determine the self - consistent charge density , which provided us with a precision in total energy of @xmath58 mev / atom .
all geometries have been optimized using the conjugate gradient method,@xcite until none of the residual hellmann - feynman forces exceeded @xmath59 ev / .
our dft - lda results for the relative energy @xmath60 of optimized c@xmath61 fullerenes with respect to graphene are shown in fig .
[ fig3](a ) .
the various data points for one size correspond to different structural isomers , which are increasing fast in number with increasing @xmath62 .
if all fullerenes were perfect spheres , eq .
( [ eq4 ] ) would simplify to@xcite @xmath63 . using the proper elastic constants for graphene@xcite @xmath48 ev and @xmath49
, we would estimate @xmath64 ev for all fullerenes independent of size .
the numerical values for the different optimized fullerene isomers in fig .
[ fig3](a ) are all larger , indicating that variations in the local curvature and bond lengths cause a significant energy penalty . as we show in the following , considering only local curvature variations across the surface ( and ignoring precise atomic positions ) allows continuum elasticity theory to quantitatively
predict the strain energy with a precision competing with _ ab initio _ calculations . to illustrate this point , we present in fig .
[ fig3](b ) the total strain energy @xmath65 in seventeen isomers of c@xmath50 obtained using various approaches .
the strain energy @xmath60 based on dft , which is expected to represent closely the experimental results , is not only significantly lower than the predicted values @xmath66 based on the tersoff potential,@xcite but also differs from this popular bond - order potential in the prediction of relative stabilities .
next we demonstrate that accurate energy estimates may be obtained even for geometries optimized using simple potentials with only bond stretching and bond bending terms such as the keating potential @xmath67 the first term sums over nearest neighbor pairs and the second term over nearest neighbor triplets , where @xmath68 and @xmath19 share the same neighbor @xmath16 .
dft calculations for graphene yield @xmath69 as bond length , @xmath70 as bond angle , @xmath71 ev / @xmath38 and @xmath72 ev / @xmath38
. geometries of c@xmath50 fullerene isomers optimized by dft and the keating potential are presented in the supplemental material.@xcite strain energies for keating optimized fullerenes are shown in fig .
[ fig3](b ) . whereas the keating optimized geometry is close to the dft optimized geometry
, the keating strain energy @xmath73 clearly underestimates the dft values and does not correctly represent the relative stabilities of the different isomers . as an alternative
, we used the keating optimized geometry to obtain the curvature strain energy @xmath74 using the continuum approach .
we found that this approach represents the relative stabilities of isomers adequately and compares well to @xmath60 .
the curvature strain energy values are somewhat lower than the dft values , since energy penalties associated with bond stretching and bending do not appear in the continuum approach .
the small value of such corrections reflects the fact that in equilibrated structures , bond lengths and angles are near their optimum .
the largest errors are expected in frustrated structures , where not all bond lengths and angles can be optimized simultaneously .
one of the key findings of this study is that continuum elasticity theory provides not only a fast , but also a relatively robust way to determine relative stabilities that are , to some degree , insensitive to the precise geometry .
we illustrate this point in fig .
[ fig3](c ) , where we compare different ways to determine the total strain energy @xmath65 in all c@xmath50 isomers discussed in fig .
[ fig3](b ) .
@xmath60 , shown by the solid line , is the difference between the total energy in dft of dft - optimized c@xmath50 isomers and 38 carbon atoms in the graphene structure .
@xmath75 , given by the dashed line , is the curvature energy based on the dft - optimized geometry .
@xmath74 , given by the dash - dotted line , is the curvature energy based on the keating - optimized geometry .
we note that all expressions provide an accurate representation of relative stabilities .
as mentioned above , the fact that @xmath9 is about 10% lower than @xmath76 is caused by our neglecting the stretching and bending of discrete atomic bonds in the continuum approach . in carbon nanostructures with respect to the graphene reference system .
( a ) comparison between dft - based total energies @xmath60 and the curvature energy @xmath74 based on keating - optimized geometries for all fullerene isomers considered in fig .
[ fig3](a ) .
( b ) comparison between dft - based strain energies @xmath77 and curvature energies per atom @xmath78 for keating - optimized geometries of fullerenes , nanotubes and schwarzites .
dashed lines represent agreement between dft and continuum elasticity results .
[ fig4 ] ] encouraged by the level of agreement for c@xmath50 , we present in fig .
[ fig4](a ) the correlation between the curvature energy @xmath74 and @xmath60 based on dft for all fullerenes discussed in fig .
[ fig3](a ) .
the narrow spread of the data points around the @xmath79 line confirms that the continuum elasticity approach is competitive in accuracy with computationally much more involved _
ab initio _ calculations . to demonstrate the generality of our approach ,
we extend it from near - spherical fullerenes to nanotubes with cylindrical symmetry and schwarzites with local negative gaussian curvature .
since nanotubes and schwarzites are infinitely large , we compare stabilities on a per - atom basis in these structures . besides results for the fullerenes discussed in figs .
[ fig3 ] and [ fig4](a ) , fig .
[ fig4](b ) displays results for nanotubes with radii ranging between @xmath80 and for schwarzite structures with @xmath81 , @xmath82 and @xmath83 carbon atoms per unit cell .
these results again indicate an excellent agreement between curvature energies in keating - optimized structures and dft - based strain energies .
this agreement is particularly impressive , since the spread of atomic binding energies extends over more than 1 ev .
as suggested at the outset , our approach to estimate relative stabilities is particularly valuable for unexplored systems such as monolayers of blue phosphorus,@xcite where model potentials have not yet been proposed .
our dft - pbe results for a blue phosphorene monolayer indicate @xmath84 @xmath38 as the projected area per atom , @xmath85 ev and @xmath86 .
the monolayer structure , shown in the top panel of fig .
[ fig5](a ) , has an effective thickness of 1.27 .
this structure can be rolled up to phosphorene nanotubes with different radii @xmath24 using the approach used in the construction of carbon nanotubes.@xcite as seen in fig .
[ fig5](b ) , the strain energy for this geometry , obtained using continuum elasticity theory , agrees very well down to very small radii with results obtained using much more involved dft calculations.@xcite ( bottom ) .
( b ) comparison between the strain energy per atom @xmath87 based on continuum elasticity theory and @xmath77 based on dft in blue phosphorene nanotubes .
the dashed line represents agreement between dft and continuum elasticity results.[fig5 ] ]
given a set of points in space , such as atomic positions , it is possible to construct a smooth surface that contains all these points in order to characterize its shape everywhere , and to eventually determine the deformation energy using the continuum elasticity approach .
we illustrate this point by tessellating the smooth surface of a graphitic nanocapsule , consisting of a cylinder capped by hemispheres at both ends and representing c@xmath88 , in different ways . our results in fig .
[ fig6 ] show that the average curvature energy @xmath89 is rather insensitive to the tessellation density .
the horizontal dashed line at @xmath90 ev / @xmath38 , representing an extrapolation to a dense tessellation , is @xmath91% higher than the exact continuum elasticity value of @xmath92 ev / @xmath38 , obtained for an ideal capsule with cylinder and hemisphere radius @xmath93 .
the small difference arises from our approximate way to estimate the mean curvature @xmath19 on the cylinder surface and at the interface between the cylinder and the hemisphere .
the extrapolated value is also close to the @xmath94 ev / @xmath38 based on the dft - optimized c@xmath88 capsule . per area of a nanocapsule tessellated by a honeycomb lattice with different numbers of vertices .
the vertical dash - dotted red line indicates that the capsule represents the c@xmath88 structure .
the inset shows , how the capsule surface can be tessellated by a honeycomb lattice with 120 vertices or atoms , shown by the white lines , and also with 480 vertices , shown by the red lines .
the horizontal dashed black line represents an extrapolation to an infinitely dense tessellation .
[ fig6 ] ] the reverse process to determine atomic positions from the shape alone is not unique .
an informative example is the structure of a carbon nanotube .
whereas the precise atomic structure within each nanotube is defined by the chiral index , many nanotubes with different chiral indices share essentially the same diameter and the same local curvature .
thus , given only the diameter of a ( wide ) hollow cylinder representing a nanotube , it is impossible to uniquely identify the chiral index and thus the atomic position . as a matter of fact ,
identifying the precise atomic positions is not necessary , since according to continuum elasticity theory , supported by experimental evidence , the stability of nanotubes depends only on the tube diameter.@xcite from its construction , the continuum elasticity description of local and global stability is best suited for very large structures with small local curvatures .
therefore , the high level of agreement between its predictions and _ ab initio _ results in structures with large local curvatures is rather impressive . among the different allotropes , we find the continuum elasticity description to be most accurate for carbon nanotubes , where all bond lengths are at their equilibrium value . in fullerenes and schwarzites ,
the presence of non - hexagonal rings , including pentagons and heptagons , prevents a global optimization of bond lengths and bond angles , reducing the agreement with dft results .
our stability results are consistent with the pentagon adjacency rule that provides an energy penalty of @xmath95 ev for each pair of adjacent pentagons , which causes an increase of the local curvature .
while this rule is surely useful , it can not compare the stability of isomers with isolated pentagons or structures of different size .
what we consider the most significant benefit of our approach to determine local strain@xcite is to identify the least stable sites in a structure . local curvature and in - plane strain play the key role in both local stability and local electronic structure,@xcite which also controls the chemical reactivity
thus , our approach can identify the most reactive and the least stable sites , which control the stability of the entire system . +
in conclusion , we have introduced a fast method to determine the local curvature in 2d systems with arbitrary shape .
the curvature information , combined with elastic constants obtained for a planar system , provides an accurate estimate of the local stability in the framework of continuum elasticity theory .
relative stabilities of graphitic structures including fullerenes , nanotubes and schwarzites , as well as phosphorene nanotubes calculated using this approach , agree closely with _
ab initio _ density functional calculations
. the continuum elasticity approach can be applied to all 2d structures and is particularly attractive in complex systems with known structure , where the quality of parameterized force fields has not been established .
+ we acknowledge valuable discussions with zacharias fthenakis .
this work was funded by the national science foundation cooperative agreement # eec-0832785 , titled `` nsec : center for high - rate nanomanufacturing '' .
computational resources for this project were provided by the michigan state university high - performance computer center .
39ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12
& 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1073/pnas.0502848102 [ * * , ( ) ] link:\doibase 10.1126/science.267.5195.222 [ * * , ( ) ] link:\doibase 10.1038/318162a0 [ * * , ( ) ] link:\doibase 10.1038/354056a0 [ * * , ( ) ] link:\doibase 10.1021/ci800395r [ * * , ( ) ] link:\doibase 10.1038/352762a0 [ * * , ( ) ] link:\doibase 10.1088/978 - 1 - 627 - 05273 - 3 [ _ _ ] ( , ) link:\doibase 10.1021/cr030093d [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.046801 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevb.80.075427 [ * * , ( ) ] link:\doibase 10.1038/nature07919 [ * * , ( ) ] link:\doibase 10.1038/nature07872 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.61.2879 [ * * , ( ) ] link:\doibase 10.1039/a901600 g [ * * , ( ) ] link:\doibase 10.1021/ja00212a020 [ * * , ( ) ] link:\doibase 10.1038/329529a0 [ * * , ( ) ] link:\doibase 10.1103/physrevb.48.15461 [ * * , ( ) ] link:\doibase 10.1140/epjst / e2007 - 00246 - 0 [ * * , ( ) ] link:\doibase 10.1007/978 - 3 - 642 - 11392 - 5 [ _ _ ] ( , , ) \doibase http://dx.doi.org/10.1016/s0167-8396(00)00006-6 [ * * , ( ) ] link:\doibase 10.1021/ci900124z [ * * , ( ) ] , ed .
, @noop _ _ ( , , ) link:\doibase 10.1088/0953 - 8984/20/6/064208 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.23.5048 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.13.5188 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , international series of monographs on chemistry ( , ) @noop link:\doibase 10.1103/physrev.145.637 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.112.176802 [ * * , ( ) ] @noop link:\doibase 10.1063/1.463854 [ * * , ( ) ] link:\doibase 10.1021/ja00771a014 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0009-2614(88)87084-2 [ * * , ( ) ] | we propose a fast method to determine the local curvature in two - dimensional ( 2d ) systems with arbitrary shape .
the curvature information , combined with elastic constants obtained for a planar system , provides an accurate estimate of the local stability in the framework of continuum elasticity theory .
relative stabilities of graphitic structures including fullerenes , nanotubes and schwarzites , as well as phosphorene nanotubes , calculated using this approach , agree closely with _
ab initio _ density functional calculations
. the continuum elasticity approach can be applied to all 2d structures and is particularly attractive in complex systems with known structure , where the quality of parameterized force fields has not been established . |
the interest in one - dimensional ( 1d ) quantum systems with several interacting particles arguably began back in 1931 when bethe solved the famous heisenberg model of ferromagnetism @xcite , but it was only in the 1960s that people realized that the techniques invented by bethe could be used to solve a host of different many - body models @xcite .
it was subsequently realized that many 1d systems have universal low - energy behaviour and can be described by the paradigmatic tomonaga - luttinger - liquid ( tll ) theory @xcite .
this opened up the field of one - dimensional physics , which has remained a large subfield of condensed - matter physics ever since @xcite .
recently , there has been a great revival of interest in 1d systems due to the realization of 1d quantum gases in highly controllable environments using cold atomic gases @xcite .
this development implies that one may now experimentally realize 1d systems with bosons or fermions and explore the intricate nature of their quantum behaviour .
a recent frontier is the realization of multi - component systems @xcite in order to study fundamental 1d effects such as spin - charge separation @xcite .
while this effect is usually associated with spin 1/2 fermions , it turns out that it can also be explored in bose mixtures ( two - component bosonic systems ) where the phenomenon can be even richer as there can be interactions between the two components ( inter - species ) and also within each component separately ( intra - species ) @xcite .
the latter is strongly suppressed for fermions due to the pauli principle . in the case where the intra- and inter - species interactions are identical it has been shown that a ferromagnetic ground state occurs @xcite . generalizing to the case of unequal intra- and inter - species interactions may be possible , but since the proofs and techniques rely on spin algebra and representation theory , they can not be used to obtain the full spatial structure of general systems and other approaches are therefore needed . here
we consider the limit where the inter - species dominates the intra - species interactions .
this regime has been explored in recent years for small systems using various few - body techniques @xcite and behaviour different from strongly interacting fermions or single - component bosons can be found already for three particles @xcite . from the many - body side ,
the system is known to have spin excitations with quadratic dispersion , @xcite which can be shown to be a generic feature of the magnon excitations above a ferromagnetic ground state @xcite .
this goes beyond the tll theory and it has been conjectured that a new universality class ( ferromagnetic liquid ) emerges in this regime @xcite . here
we provide a particularly clean realization of a ferromagnetic system confined in a harmonic trap . using numerical and newly developed analytical techniques we obtain and analyze the exact wave function .
this allows us to explore the crossover between few- and many - body behaviour , and to demonstrate that the strongly interacting regime realizes a perfect ferromagnet in the ground state , while particular excited states will produce perfect antiferromagnetic order . in the extremely imbalanced system , with one strongly interacting impurity
, we find both numerically and analytically that the impurity will always move to the edge of the system .
this is in sharp contrast to fermionic systems where the impurity is mainly located at the center @xcite .
our work provides a rare and explicit example of perfect ferro- or antiferromagnetism using the most fundamental knowledge of a quantum system as given by the full wave function . and one @xmath0 particle as a function of interaction strength , @xmath1 , obtained by numerical calculations . in the limit @xmath2 ,
the ground state becomes doubly degenerate and has half - integer energy .
the contribution from center - of - mass motion has been removed .
for visibility , we have removed states from the attractive side that diverge to large negative energies close to @xmath2 . ]
our two - component bosonic system has @xmath3 particles split between @xmath4 and @xmath5 identical bosons of two different kinds .
all @xmath6 particles have mass @xmath7 and move in the same external harmonic trapping potential with single - particle hamiltonian @xmath8 , where @xmath9 and @xmath10 denote the momentum and position of either an @xmath11 or @xmath0 particle and @xmath12 is the common trap frequency .
the trap provides a natural set of units for length , @xmath13 , and energy , @xmath14 , which we will use throughout ( here @xmath15 is planck s constant divided by @xmath16 ) .
we assume short - range interactions between @xmath11 and @xmath0 particles that we model by a dirac delta - function parameterized by an interaction strength , @xmath1 , i.e. @xmath17 where @xmath10 and @xmath18 denote the coordinates of @xmath11 and @xmath0 particles , respectively .
the intraspecies interaction strengths are assumed to be much smaller than @xmath1 and we will therefore neglect such terms . to access the quantum mechanical properties of our system
we must solve the @xmath6-body schrdinger equation .
this will be done using novel analytical tools and using exact diagonalization . in the latter case
we have adapted an effective interaction approach that has recently been succesfully applied to fermions in harmonic traps @xcite ( see the methods section for further details ) .
the analytical and numerical methods allow us to address up to ten particles , which is larger than most previous studies not based on stochastic or monte carlo techniques .
the simplest non - trivial case is the three - body system which has two @xmath11 and one @xmath0 particle .
the energy spectrum is shown in fig . [ 2 + 1 ] as a function of @xmath1 . the most interesting feature to notice is the ground state behaviour as @xmath19 . here , an odd and an even state become
degenerate at an energy of @xmath20 .
this should be contrasted to the behaviour of single - component bosons or two - component fermions which will always have energies that are an integer times @xmath14 when @xmath2 .
furthermore , we notice how the two states that merge at @xmath21 become two excited state branches on the attractive side of the resonance but the even parity state remains the lower one .
this is opposite to the behaviour of fermions @xcite where the hierarchy of states is inverted at @xmath21 .
the ground state for large and negative @xmath1 is very different as it contains deeply bound molecules , which we will not consider further .
the fractional energies in the spectrum can be explained by a schematic three - body model and in stochastic variational calculations @xcite .
this provides a hint that larger systems could also display fractional energy states in the strongly interacting limit and begs the question as to what spatial configurations such states correspond to .
we will now show that the fractional energy states are generic for strongly interacting two - component bosons in 1d and , importantly , for the ground state they realize perfect ferromagnetic behaviour irrespective of whether the system is balanced ( @xmath22 ) or not .
the term perfect ferromagnetic behaviour implies that we have a full spatial separation of the two components in the exact ground state wave function of the system , i.e. the probability to find only @xmath11 on one side and only @xmath0 on the other side of the system is not just dominant , it is exactly _
unity_. the ground state has only a single domain wall at which an @xmath11 and a @xmath0 particle are neighbours . as a consequence ,
imagine that you detect an @xmath11 particle on the left ( right ) side of the system , then you can immediately conclude that all the @xmath0 particles reside to the right ( left ) of this @xmath11 particle .
we first consider a four - body system that has two @xmath11 and two @xmath0 particles .
the energy spectrum for @xmath23 is shown in fig .
[ 2 + 2]*a * ) .
a striking feature is the two - fold degenerate ground state for @xmath2 that has a non - integer energy similar to the three - body problem . in this strongly repulsive limit
, the system realizes a perfect spatially ferromagnetic quantum state as we will now demonstrate analytically .
first we note that the center - of - mass motion of the four - body system can be separated and thus ignored .
this leaves three jacobi coordinates to describe the system .
the details of these reductions can be found in the methods section below . in fig .
[ 2 + 2]*b * ) we show the space of the jacobi coordinates and highlight all the planes at which an @xmath24 ( solid planes ) , an @xmath25 or a @xmath26 ( checkerboard planes ) pair of particles overlap .
the main observation is that as @xmath2 , the wave function must vanish on all the solid planes in fig .
[ 2 + 2]*b * ) and we arrive at the disconnected regions shown with different colours .
the particles become effectively impenetrable and we may characterize the wave function by specifying the amplitudes of all possible spatial configurations of the four particles .
these regions correspond to specific orderings on a line of the four particles .
in particular , the large ( red ) region dominating the figure corresponds to spatial configurations @xmath27 or @xmath28 .
the green region occupies half the spatial volume of the red and corresponds to @xmath29 or @xmath30 , while the yellow region has one - fourth the volume of the red region and corresponds to @xmath31 or @xmath32 configurations .
a wave function that vanishes on all @xmath24 interfaces may now be constructed in each of these regions .
however , it is immediately clear that it will have lower energy when it can spread over a larger volume .
we thus conclude that the doubly degenerate ground state at @xmath2 has the structure @xmath33 ( taking into account the parity invariance of the hamiltonian ) . as discussed in the methods section
, one may solve a simple wave equation in the red region and obtain the ground state energy to arbitrary precision .
the triangles at the @xmath21 line in fig .
[ 2 + 2]*a * ) show the energies obtained in this manner .
we reproduce both the ground state and a set of excited states .
all of these have fractional energies and all of them are perfectly ferromagnetically ordered .
the remaning states of the spectrum can be obtained by solving in the other regions of fig .
[ 2 + 2]*b * ) .
note that states with amplitude exclusively in the yellow regions are perfectly spatial antiferromagnetic , @xmath34 , and have energies @xmath35 @xmath14 with integer @xmath36 .
they are the only parts of the spectrum which can be constructed by starting from identical fermions using girardeau s mapping techniques @xcite .
the arguments presented here are neither restricted to @xmath37 nor to a harmonic trapping potential and hold for any @xmath6 and any shape of the external confinement .
they hinge only on the fact that the @xmath27 or @xmath28 configurations occupy the largest volume .
for instance , for @xmath27 the four regions @xmath38 , @xmath39 , @xmath40 and @xmath41 are adjacent regions and are connected by bose symmetry of each component , thus they make up the connected upper red region ( @xmath42 ) of the coordinate space in fig .
[ 2 + 2]*b * ) . for @xmath29
one finds that only @xmath43 and @xmath44 are adjacent and connected .
this implies that the volume for @xmath29 is half as large . and different values of @xmath1 .
dotted ( red ) line corresponds to @xmath45 while the solid ( black ) line is for @xmath46 .
the dots show the analytical solution for @xmath21 . *
b * ) and * c * ) densities for an equal superposition ( sum ) of the ( nearly ) two - fold degenerate ground state at @xmath46 for @xmath47 , 3 , 4 , and 5 . *
b * ) shows @xmath11 particles and * c * ) shows @xmath0 particles . *
d * ) rescaled plot of the total density at @xmath46 . for @xmath48
the density has been rescaled to a total density with @xmath49 .
the dashed line corresponds to the density expected in the many - body limit @xmath50 . ]
larger systems may in principle be handled in similar fashion by solving wave equations with proper boundary conditions and obtaining the fractional energies in the limit @xmath2 .
however , the increase in dimension of the problem makes this very difficult in practice . in order to further demonstrate that balanced systems have perfect ferromagnetic ground states irrespective of particle number ,
we have numerically computed the ground state densities for systems with @xmath51 as shown in fig .
[ bdens ] .
evidence of the separation of @xmath11 and @xmath0 can be seen in the total density already in fig .
[ bdens]*a * ) for @xmath47 as @xmath1 increases ( note the perfect agreement with the analytical result in the limit @xmath2 ) .
we expect the two degenerate ground states to have structure @xmath52 . in order to prove this perfect ferromagnetic behaviour
, we consider the odd and even superposition of the two degenerate states which we expect will yield states with exclusively @xmath11 or @xmath0 particles on either side of the system ( corresponding to @xmath53 or @xmath54 ) .
the corresponding densities are shown in fig .
[ bdens]*b * ) and fig .
[ bdens]*c * ) and beautifully confirm our expectations .
as the ground state for @xmath2 is spatially separated , one may speculate that it can be understood physically as two ideal bose gases or condensates sitting on either side of the system even in this strongly interacting limit . in fig .
[ bdens]*d * ) we plot the densities in a rescaled fashion where we multiply by @xmath55 , @xmath56 , and @xmath57 on the @xmath47 , @xmath58 and @xmath59 densities respectively .
the convergence of the results toward the @xmath49 case indicates that the system does behave as two ideal bose gases as the particle number grows . in the limit @xmath50
, we would expect the overlap of the two gases to vanish as the energy cost of overlap goes to infinity .
we therefore expect that the occupied mode in this large system limit is the first excited state of the harmonic trap which vanishes at the center .
the dashed line in fig .
[ bdens]*d * ) shows this state rescaled to @xmath49 .
this analytical guess displays the same features as the numerical densities and we conclude that already for ten particles the many - body properties are emerging . or @xmath60 system with @xmath61 .
the analytical results for @xmath21 are shown as triangles . *
b * ) and * c * ) as in fig .
[ bdens]*b * ) and * c * ) but for @xmath61 with @xmath62 or @xmath60 . *
d * ) the density of the majority component ( @xmath5 ) .
all numerical results have been obtained with @xmath46 . ]
the extremely imbalanced limit , where @xmath61 and @xmath5 varies , provides a realization of a strongly interacting bose polaron in 1d , i.e.
an impurity that interacts strongly with an ideal bose gas . in fig .
[ imdens ] we plot the densities of systems with @xmath61 and @xmath62 or @xmath60 .
we see that the impurity sits at the edge of the system ( fig .
[ imdens]*a * ) ) , while the majority component tends to occupy the center ( fig .
[ imdens]*d * ) ) .
we confirm the numerical results by employing an analytical model , which shows excellent agreement .
the details can be found in the methods section . to confirm that the wave function of the strongly interacting ground state has intrinsic phase separation
, i.e. has the form @xmath63 , we plot the densities for a sum of the nearly degenerate ground states in fig .
[ imdens]*b * ) and * c * ) .
as in the balanced case above , we find a perfectly separated ground state behaviour .
for instance , if we locate the single @xmath11 particle on one side of the trap , we would thus immediately know that all the @xmath0 particles reside on the other side , and vice versa .
this behaviour is opposite to the case where the @xmath0 particles are identical fermions where the impurity resides mainly in the center of the system @xcite .
we have confirmed that this structure is also present for @xmath64 with @xmath65 , and it is therefore a generic feature that the two species are perfectly spatially separated ( ferromagnetic ) in the ground state for strong interactions . .
the filled circles show the ground state energy for @xmath46 relative to the zero - point energy as given by a single - component system of the same size .
each point is marked with the number of particles using the notation @xmath66 .
the diamonds are the results of the analytical method for the polaron case described in the methods section .
the dashed lines are quadratic interpolations for fixed number of @xmath0 particles , while the solid line is an interpolation of the energy for the balanced systems . ] a remarkable feature of the densities in fig .
[ imdens]*b * ) and * c * ) is the movement of the centroids of the peaks with particle number .
we clearly see the majority moving into the center and the impurity being pushed toward the edge .
this demonstrates how an ideal condensate is being built in the center . for large @xmath5
the energy per particle goes to @xmath67 , which implies a single - mode condensate that is becoming macroscopically occupied ( see methods section for details ) .
the relative deviation between numerical and analytical energies is below three percent for @xmath68 .
we thus have an analytic model for the crossover between the few- and many - body limit for the bosonic polaron in one dimension .
this includes the external trap that is a reality of most experiments .
we have shown that a mixture of two ideal bose systems in one dimension has unusual properties when the inter - species interactions is strong .
the systems have energies that are non - integer multiples of @xmath14 . in fig . [ enetot ]
we show the ground state energies for @xmath46 in systems with ten particles or less relative to a ground state with only @xmath0 particles . driving an @xmath11 to @xmath0 transition via radiofrequency spectroscopy would be a possible way to confirm the predicted energies in fig .
[ enetot ] . this technique has been demonstrated for fermions in recent few - body experiments in 1d @xcite . the results in fig . [ enetot ]
show that the energy per particle tends to saturate for large systems and that this happens faster the more imbalanced the system is .
we also see that the balanced case has an almost linear energy dependence . system .
the even parity ground state is shown in purple , while the lowest excited state with antiferromagnetic structure is shown in green .
for comparison the four - peak black curve represents identical fermions while the narrow peak dashed black curve shows hard - core bosons . *
b * ) ground state impurity distributions for the even ( blue ) and odd ( red ) ground states with @xmath60 ( solid ) and @xmath62 ( dashed ) .
* c * ) ground state majority distributions for @xmath60 ( solid ) and @xmath62 ( dashed ) .
the even and odd parity results coincide for the majority . *
d * ) the same as in * a * ) for @xmath69 .
all curves in the plot have been obtained for @xmath46 . ]
the ferro- and antiferromagnetic states can also be detected by measuring momentum distributions . in fig .
[ moment]*a * ) and * d * ) we show the distributions for @xmath46 with @xmath47 and @xmath69 , respectively .
the purple solid line is the even parity ground state while the solid green line is the excited state with antiferromagnetic ordering .
the striking difference of the two distributions implies that they should be easily identifiable in experiments . for comparison ,
the solid black line with a multi - peak structure shows the distribution for a system of identical fermions . comparing the solid green and black curves in fig .
[ moment]*a * ) and * d * ) clearly demonstrates that , in spite of the fact that these two states have equal energy , the correlations are very different . in fig .
[ moment]*a * ) the dashed black line corresponds to the tonks - girardeau hard - core boson state @xcite , which is also seen to be very different from the states discussed here . for imbalanced systems , we find that measuring the impurity momentum distribution , fig .
[ moment]*b * ) , yields information about the parity of the state . on the other hand , the majority distributions in fig .
[ moment]*c * ) are identical in the two opposite parity ground states . a characteristic feature seen in fig .
[ moment]*b * ) is the development of oscillatory structure as the number of majority particles increases and pushes the impurity further out in the trap , see also fig .
[ imdens]*b * ) .
the separation of components in the ground state is intrinsic to both balanced and imbalanced mixtures , as is the presence of other spatial configurations in specific excited states .
the effect is not connected to the harmonic confinement considered here and should be seen in an arbitrary confining geometry .
a simple physical picture can be given in terms of domain walls , i.e. points at which the two components interface .
the system tends to minimize the number of domain walls and this principle can be used to understand the ferromagnetic ground state and predict the ordering in energy of other configurations . in the paradigmatic two - component ( spin 1/2 ) fermi system , the ground state is never purely ferro- or antiferromagnetic for strong interactions @xcite , and bose
mixtures therefore provide a unique set of quantum ground states for exploring and exploiting magnetic behaviour .
the description of these systems clearly goes beyond the famous bose - fermi mappings @xcite and we provide not only numerical but also new analytical tools to fill this gap .
importantly , we demonstrate that the crossover from few- to many - body physics can be studied already at the level of ten particles .
this work was funded by the danish council for independent research dff natural sciences and the dff sapere aude program and the european research council under the european community s seventh framework programme - erc grant agreement no .
we solve numerically the many - body schrdinger equation by exact diagonalization with the full hamiltonian projected onto a finite basis constructed from harmonic oscillator single - particle states .
each many - body basis state is written as a product of symmetrized states of @xmath4 and @xmath5 particles .
the model space truncation is defined by an upper limit of the total energy . instead of the bare zero - range interaction in , we consider an effective two - body interaction in order to speed up the convergence of the eigenstates with respect to the size of the many - body basis .
the effective potential is constructed in a truncated two - body space , and is designed such that its solutions correspond to exact two - body solutions given by the busch formula @xcite . as explained in detail in refs .
, this is achieved using a unitary transformation that involves the lowest eigensolutions given by the busch formula . by construction
, this unitary transformation approach will reproduce exact bare hamiltonian results for the many - body system ( both energy spectrum and wave functions ) in the limit of infinite model space .
the excellent convergence property of this effective - interaction approach was demonstrated in ref .
@xcite and is key to the quality of our numerical results and to our conclusions . in the construction of the effective interactions we benefit from having access to the exact two - body solutions for short - range interactions in harmonic traps .
however , we stress that using numerical two - body solutions this approach can be generalized to study many - body systems in higher dimensions with finite - range interactions and in any trapping potential .
here we outline the calculational procedures required to obtain the exact solutions for the @xmath47 four - body system in the @xmath2 limit .
the method can in principle be extended to larger systems , but it becomes increasingly difficult . in the next subsection we provide an alternative method that works well for larger systems in the imbalanced case .
denote the @xmath11 coordinates by @xmath70 and the @xmath0 coordinates by @xmath71 , see fig .
[ 2 + 2]*b * ) .
the hamiltonian is @xmath72&\nonumber\\ & + g\delta(x_1-y_1 ) + g\delta(x_1-y_2)+g\delta(x_2-y_1)+g\delta(x_2-y_2),&\end{aligned}\ ] ] with @xmath1 the @xmath24 interacting coupling constant and we assue that the @xmath25 and @xmath26 interactions vanish .
we now perform an orthogonal coordinate transformation @xmath73= \left[\begin{matrix } \frac{1}{\sqrt{2}}&-\frac{1}{\sqrt{2}}&0&0\\ 0&0&-\frac{1}{\sqrt{2}}&\frac{1}{\sqrt{2}}\\ \frac{1}{2}&\frac{1}{2}&-\frac{1}{2}&-\frac{1}{2}\\ \frac{1}{2}&\frac{1}{2}&\frac{1}{2}&\frac{1}{2 } \end{matrix}\right ] \left[\begin{matrix}x_1 \\ x_2 \\ y_1 \\ y_2\end{matrix}\right],\end{aligned}\ ] ] where @xmath74 are as shown at the bottom of in fig .
[ 2 + 2]*b * ) while @xmath75 denotes the center - of - mass coordinate .
the quadratic kinetic and harmonic oscillator terms in @xmath76 retain their form under this transformation and one may immediately separate the center - of - mass , @xmath75 , which can be ignored from now on . for the remaining three coordinates we switch to the usual spherical coordinate system , i.e. @xmath77 , @xmath78 , @xmath79 , and @xmath80 . carrying out these transformation
we arrive at the relative hamiltonian @xmath81 where the sum in the interaction term runs over the four combinations of signs in the argument of the delta function .
the first two terms constitute a 3d harmonic oscillator with the well - known regular solution @xmath82 , where @xmath13 and @xmath83 is the generalized laguerre polynomial .
when @xmath84 , the angular functions , @xmath85 , are the usual spherical harmonic functions with @xmath86 the total and @xmath7 the projection angular momentum quantum number . however , in the limit @xmath2 we have to enforce non - trivial boundary conditions whenever @xmath11 and @xmath0 particles overlap .
let us focus on the region @xmath42 by restricting to @xmath87 ( solutions for @xmath88 may be obtained by symmetry arguments or by considering instead @xmath89 ) .
the arguments of the dirac delta - functions in eq . vanish when @xmath90 if we define @xmath91 , we have @xmath92 and @xmath93 .
the regions defined by these conditions are illustrated in fig .
[ 2 + 2]*b * ) .
the solid red planes show exactly where the arguments of the interaction dirac delta - functions have to vanish .
we now make the simple transformation @xmath94 and @xmath95 . in these new @xmath96 variables ,
the boundaries are simply @xmath97 and @xmath98 , i.e. the function must vanish on the boundary of a square .
finally , one must transform the angular part of the laplacian into the new variables which yields @xmath99 by the procedure outlined above we have transformed the problem of solving a harmonic oscillator problem in a non - trivial geometry , i.e. the red region in fig .
[ 2 + 2]*b * ) , into solving a very simple boundary value problem @xmath100 with @xmath101 for @xmath102 .
we write the eigenvalue in this way so it matches the usual 3d angular eigenvalue @xmath103 .
the equation for @xmath104 may be straighforwardly solved by using a two - dimensional fourier expansion of the wave function .
this will produce some spurious solutions as we must also impose bosonic symmetry among the two @xmath11 and two @xmath0 particles separately .
this translates to the requirement that the solution be symmetric when reflected across the two diagonals .
notice that for each eigenvalue of this problem , @xmath105 , we obtain a whole class of solutions with energies @xmath106 as we may add radial excitations . the low - lying solutions of eq .
are given in tab .
[ datatable ] .
state number 1 , 4 , 6 , 11 , 13 , and 15 have the required bosonic symmetries for the balanced system .
a number of doubly degenerate states in the spectrum may be used to construct eigenfunctions for a four - body system with @xmath107 and @xmath108 ( or vice versa ) . in this case
the wave function must vanish on the diagonal of the @xmath109 square domain which is achievable by taking proper linear combinations .
the states marked fermions in tab .
[ datatable ] are antisymmetric across the two diagonals in the @xmath109 square and provide allowed states for all four - body two - component fermi systems , i.e. 2 + 2 , 3 + 1 or four identical fermions .
the eigenenergies of the fermionic states have the exact values 7.5 , 10.5 and 11.5 .
our results differ by @xmath110 which attests to the accuracy of our method .
all states in tab .
[ datatable ] have been obtained using a modest 400 fourier basis states .
notice that even though the lowest perfectly antiferromagnetic state for @xmath47 bosons is at the same energy as the fermionic state number 5 in tab .
[ datatable ] , they are not related since states with the configurations @xmath31 or @xmath32 solve a different boundary value problem ( corresponding to the yellow regions in fig . [ 2 + 2]*b * ) )
. .low - lying spectrum of eq . .
the last column denotes the specific system for which the particular solution is an allowed eigenstate using the notation @xmath111 for bosons .
see the text for details . [
cols="^,^,^",options="header " , ] the energies obtained using this ( semi)-analytical approach for the @xmath47 system are given in fig .
[ 2 + 2]*a * ) as triangles at @xmath21 .
the two lowest triangles correspond to the angular ground state ( lowest @xmath105 value ) with @xmath112 and @xmath113 .
the two upper triangles are the first and second excited angular solutions both with @xmath112 .
all four solutions have the spatial structure @xmath33 . the blue dots in fig .
[ bdens]*a * ) show the ground state density obtained by the transformation method discussed here .
the rest of the spectrum at @xmath2 can be obtained by solving the boundary value problem in the green ( @xmath114 ) and yellow areas ( @xmath34 ) . in the latter case
a fermionized ( totally antisymmetric ) wave function is a solution .
our main interest here is to understand the ground state so we leave the remaining states and regions for future investigations .
the analytics provided here is applied for the bose polaron , @xmath61 and @xmath5 arbitrary , but can be extended to other systems .
the hamiltonian can be written @xmath115 where @xmath116 is a 1d harmonic oscillator .
the @xmath117 coordinate denotes the single @xmath11 particle , the impurity , while @xmath118 denotes the coordinates of the majority @xmath0 particles .
we introduce an adiabatic decomposition of the total wave function of the form @xmath119 where @xmath120 is a normalized eigenstate of the eigenproblem @xmath121 which depends parametrically on @xmath117 .
this expansion can be related to the born - oppenheimer approximation in which case one may consider @xmath117 the slow variable ( typically the nuclear coordinate in molecular physics ) . in the limit of interest @xmath2 , we impose the condition that the total wave function vanishes for @xmath122 , @xmath123 .
this implies that @xmath124 whenever @xmath122 .
since there are no intra - species interactions among the @xmath0 particles , we can write @xmath125 where @xmath126 denotes the symmetrization operator and @xmath127 is the @xmath128th normalized eigenstate of @xmath129 which satisfies the condition @xmath130 . the index @xmath131 on @xmath120 denotes the many different ways to distribute the @xmath5 particles among the eigenstates of @xmath129 with the appropriate boundary condition . the schrdinger equation for @xmath132 can now be written @xmath133\phi_i=\sum_j \left(q_{ij}(x_1)\phi_j+p_{ij}(x_1)\frac{\partial \phi_j}{\partial x_1}\right),\end{aligned}\ ] ] where @xmath134 the subscript @xmath18 on the brackets denote integration over all @xmath135 .
note that @xmath136 and @xmath137 @xcite . as we are interested in the ground state
, we assume that all the @xmath0 particles are in the same state , @xmath138 , that we specify below . since the nearest excited states are obtained by promoting one of the @xmath0 particles into a single - particle excited orbital
, one can show that @xmath139 and @xmath140 scales with @xmath141 while @xmath142 scales with @xmath5 .
for large @xmath5 we can thus neglect all but the @xmath142 terms .
furthermore , in the ground state we expect to find all the @xmath0 particles on one side of the impurity .
if we assume that all @xmath0 particles are to the left of the impurity , we can write the single - particle wave function , @xmath143 for @xmath144 , as @xmath145 for @xmath88 or @xmath146 and @xmath147 , while for @xmath146 and @xmath148 we write @xmath149 here @xmath150 is a normalization factor , @xmath151 and @xmath152 are the tricomi and kummer confluent hypergeometric functions , and we have used @xmath13 as the unit of length . here
@xmath153 is a function that is chosen to satisfy the requirement @xmath154 .
this is equivalent to finding the ground state solution of @xmath155 for @xmath144 with the condition that the wave function must vanish at @xmath156 .
once we have determined the functions @xmath157 and @xmath153 , we can compute the adiabatic potential for the ground state .
we have @xmath158 furthermore , @xmath159 by additivity .
the schrdinger equation for @xmath160 is then @xmath161 the energy @xmath162 provides a variational upper bound to the exact energy .
the energies computed via this method for the polaron are shown in fig .
[ enetot ] and agree with the numerical results to within a few percent for the largest particle numbers in the figure .
we expect the agreement to become better for even larger particle numbers .
furthermore , since we obtain the full wave function in an analytical form we may also compute the densities of both impurity and majority components . in fig .
[ imdens]*a * ) we show the impurity densities for @xmath62 and @xmath60 , while fig .
[ imdens]*d * ) shows the corresponding majority density .
we see a striking agreement between the numerical results and the analytically tractable model presented here . the model presented here
can be extended to excited states and also to systems with @xmath65 .
99 h. a. bethe , z. physik * 71 * , 205 - 226 ( 1931 ) .
e. h. lieb and w. liniger , phys
* 130 * , 1605 - 1616 ( 1963 ) .
j. b. mcguire , j. math .
* 6 * , 432 - 439 ( 1965 ) ; _ ibid .
_ * 7 * , 123 - 132 ( 1966 ) . c. n. yang , phys .
* 19 * , 1312 - 1315 ( 1967 ) .
e. h. lieb and f. y. wu , phys
20 * , 1445 - 1448 ( 1968 ) .
f. d. m. haldane , phys .
lett . * 47 * , 1840 ( 1981 ) .
f. d. m. haldane , j. phys .
c : solid state phys .
* 14 * , 2585 ( 1981 ) .
t. giamarchi , _ quantum physics in one dimension _ , ( oxford university press inc .
, new york , 2003 ) .
m. a. cazalilla , r. citro , t. giamarchi , e. orignac and m. rigol , rev .
mod . phys . *
83 * , 1405 ( 2011 ) .
b. paredes _ et al .
_ , nature * 429 * , 277 - 281 ( 2004 ) . t. kinoshita , t. wenger , and d. s. weiss , science * 305 * , 1125 - 1128 ( 2004 ) .
t. kinoshita , t. wenger , and d. s. weiss , nature * 440 * , 900 - 903 ( 2006 ) .
e. haller _
et al._,bscience * 325 * , 1224 - 1227 ( 2009 ) . f. serwane _
et al . _ , science * 332 * , 336 - 338 ( 2011 ) .
et al . _ ,
lett . * 108 * , 075303 ( 2012 ) .
a. n. wenz _
et al . _ ,
science * 342 * , 457 ( 2013 ) .
g. pagano _
et al . _ , nature phys . * 10 * , 198 - 201 ( 2014 ) .
a. recati , p. o. fedichev , w. zwerger , and p. zoller , phys .
lett . * 90 * , 020401 ( 2003 ) .
a. b. kuklov and b. v. svistunov , phys .
lett . * 90 * , 100401 ( 2003 ) .
duan , e. demler , and m. d. lukin , phys .
91 * , 090402 ( 2003 ) .
s. zllner , h. d. meyer , and p. schmelcher , phys .
a * 78 * , 013629 ( 2008 ) .
y. hao and s. chen , phys .
a * 80 * , 043608 ( 2009 ) .
m. a. garcia - march and th .
busch , phys .
a * 87 * , 063633 ( 2013 )
. m. a. garcia - march _ et al .
a * 88 * , 063604 ( 2013 ) .
m. a. garcia - march _ et al .
_ , arxiv:1405.7595 ( 2014 ) .
n. t. zinner _
et al . _ ,
arxiv:1309.7219 ( 2013 ) .
b. sutherland , phys .
lett . * 20 * , 98 ( 1968 ) . y .-
li , s .- j .
gu , z .- j .
ying , and u. eckern , europhys .
* 61 * , 368 ( 2003 ) .
j. n. fuchs , d. m. gangardt , t. keilmann , and g. v. shlyapnikov , phys .
lett . * 95 * , 150402 ( 2005 ) .
x. guan , m. t. batchelor , and m. takahashi , phys .
a * 76 * , 043617 ( 2007 ) .
m. b. zvonarev , v. v. cheianov , and t. giamarchi , phys .
lett . * 99 * , 240404 ( 2007 ) .
s. akhanjee and y. tserkovnyak , phys .
b * 76 * , 140408 ( 2007 ) .
a. matveev and a. furusaki , phys .
* 101 * , 170403 ( 2008 ) .
a. kamenev and l. glazman , phys .
a * 80 * , 011603(r ) ( 2009 ) .
j. caux , a. klauser , and j. van den brink , phys .
a * 80 * , 061605 ( 2009 ) . | strongly interacting one - dimensional quantum systems often behave in a manner that is distinctly different from their higher - dimensional counterparts . when a particle attempts to move in a one - dimensional environment it will unavoidably have to interact and push other particles in order to execute a pattern of motion , irrespective of whether the particles are fermions or bosons .
a present frontier in both theory and experiment are mixed systems of different species and/or particles with multiple internal degrees of freedom . here
we consider trapped two - component bosons with short - range inter - species interactions much larger than their intra - species interactions and show that they have novel energetic and magnetic properties . in the strongly interacting regime , these systems have energies that are fractions of the basic harmonic oscillator trap quantum and have spatially separated ground states with manifestly ferromagnetic wave functions . furthermore , we predict excited states that have perfect antiferromagnetic ordering .
this holds for both balanced and imbalanced systems , and we show that it is a generic feature as one crosses from few- to many - body systems . |
the nature of the bright @xmath0-ray source geminga remained elusive from the first observations using sas - b ( @xcite ) until its recognition as a pulsar with a period of 0.237 seconds in @xmath0 rays ( @xcite @xcite ) and in x - rays ( @xcite ) . based upon colour considerations an optical candidate was proposed , g " with a m@xmath3 of 25.5 ( @xcite ) .
this star had a measurable proper motion ( @xcite ) indicating a probable distance of about 100 pc and thereby making a probable association with a neutron star .
subsequent hubble space telescope observations have given a distance based upon parallax of @xmath4 pc ( @xcite ) .
optical observations in b showed geminga to be fainter than 26th magnitude ( @xcite ) - a result confirmed by hst observations ( @xcite ) . in v geminga is brighter at 25.4 .
this aspect of the spectrum has been explained by a proton cyclotron feature causing either preferential emission in v or absorption in b and i ( @xcite ) superimposed on a thermal continuum .
however , re - analysis of the euve and rosat datasets highlight an error in this earlier work , indicating that the thermal continuum would not be expected to dominate in the optical regime , based on the observed flux ( @xcite ) . such an apparent absorption feature has been previosuly observed in the crab spectrum ( @xcite ) although not confirmed by other observations ( @xcite ) .
recent spectral studies of geminga ( @xcite ) show a continuous power - law from 3700 to 8000 ( ) with no such features consequently indicating that a predominantly magnetospheric origin is preferred over a thermal one . it should be noted that these spectroscopic studies were at the limit of the observational capabilities of the keck and with a low signal - to - noise ratio . of crucial importance to
the understanding of neutron star structure is the stellar radius .
this can in principle be inferred once the distance and the black - body contribution has been measured ( @xcite ) .
however determining the black - body component of an isolated neutron star is complicated by magnetospheric and possible atmospheric effects ( @xcite ) . as geminga is very nearby it is a prime candidate for measuring the thermal component - crucial to this will be the removal of the magenetospheric component of its emission .
this is possible by determining what contribution of the optical emission is pulsed and whether it follows the hard ( magnetospheric ) or soft ( presumed thermal ) x - ray emission profile .
the faintness of the optical counterpart has precluded time - resolved observations using conventional photometers . however by using 2-d photon counting detectors , the required astrometric analysis can be carried out off - line .
consequently photon arrival times can be measured from a reduced ( seeing optimised ) aperture diaphram .
observations were made on 25th and 26th february 1995 using the 3.55 m new technology telescope ( ntt ) at la silla .
follow up observations were taken in january 1996 , using the 6 m telescope ( bta ) of the special astrophysical observatory over three nights .
two mama detectors were used ; one a b extended s-20 ( @xcite ) and the other a bialkali ( @xcite ) photocathode . by using the ucg triffid camera ( @xcite ) to record the data .
the arrival time and position of each photon was recorded to a precision of 1 @xmath5second and 25 microns .
the spatial resolution was equivalent to 0``.13 on the ntt and 0''.25 on the bta .
absolute timing was achieved using a combination of a gps receiver , which gave utc to a precision of 400nsec every 10 seconds , and an ovened 10mhz crystal which was accurate to @xmath6 1 @xmath5second per 10 second interval . on each night
the crab pulsar was observed for calibration purposes . using a crab timing ephemeris ( @xcite ) the barycentric phase of the crab pulse was determined ;
phase was maintained to within 10 @xmath5seconds over the whole period .
table 1 shows a log of the observations .
ccccccc date & utc & duration & detector & telescope & filter & seeing + & & ( s ) & & & & ( @xmath7 ) + 1995 feb 26 & 01:14:37 & 4580 & gsfc & ntt & v & 1.3 + 1995 feb 26 & 02:36:37 & 4387 & gsfc & ntt & v & 1.4 + 1995 feb 26 & 03:50:49 & 3662 & gsfc & ntt & v & 1.4 + 1995 feb 27 & 01:58:20 & 788 & eso & ntt & b & 1.3 + 1995 feb 27 & 02:42:23 & 2096 & eso & ntt & b & 1.2 + 1995 feb 27 & 03:19:28 & 3000 & eso & ntt & b & 1.7 + 1996 jan 12 & 18:07:15 & 4397 & eso & bta & b & 1.6 + 1996 jan 12 & 19:21:14 & 6409 & eso & bta & b & 1.5 + 1996 jan 12 & 21:36:04 & 884 & eso & bta & v & 1.5 + 1996 jan 12 & 21:52:02 & 413 & eso & bta & v & 1.5 + 1996 jan 12 & 22:23:56 & 2914 & eso & bta & v & 1.3 + 1996 jan 12 & 23:13:26 & 2618 & eso & bta & v & 1.4 + 1996 jan 13 & 19:28:34 & 7509 & eso & bta & b & 2.2 + 1996 jan 14 & 16:59:25 & 8182 & eso & bta & b & 1.2 + 1996 jan 14 & 19:18:44 & 2810 & eso & bta & b & 1.2 + photon positions were binned to produce an image after each exposure was made . by using the triffid image processing software ,
the images could be marginally improved by removing the effects of telescope movement ( @xcite ) .
these images were compared with hst / wfpc2 archival images to determine the position of geminga at these epochs .
after coaddition of all the b and v images from january 1996 , a faint star could be seen at the expected position of geminga .
no such object could be seen in the february 1995 data .
the reason for this was two fold : firstly the exposure time - telescope aperture product was 5 times greater in 1996 compared to 1995 and secondly the flat - fields were deeper in the later observations .
once the position of geminga was established , the photons times were extracted from a window , centred on geminga , with a diameter corresponding to the average seeing widths for each exposure . this was chosen to maximise the signal to noise ratio .
these extracted times were then translated to the solar system barycentre using the jpl de200 ephemeris .
the geminga arrival times were folded in phase using the egret ephemeris ( @xcite ) for each colour and for each observing run .
3.3truein 3.3truein figure 1 shows the light curve in b for each night of january 1996 and figure 2 shows the combined light curve .
figure 3 shows the time resolved images in b for 4 phase bands as indicated .
clearly geminga can be seen during the two on phases ; star g is to the top right .
we note that background fluctuations are consistent with the expectations of poissonian statistics .
table 2 shows the determined pulsed magnitude or 1 sigma upper limits where appropriate .
this type of light curve is best analysed using the @xmath8 statistic ( @xcite ) .
only the january 1996 b data shows significant pulsations - the @xmath9 statistic for this data set ( of 20.6 ) has a significance of 99.96% . in the january 1996 v data a weak signal , with a similar form ,
can also be observed .
ccc data set & pulsed 3 @xmath10 upper limits & flux @xmath5jy + 95v & 24.0 & 0.95 + 95b & 23.8 & 1.3 + 96v & 25.5 & 0.24 + 96b & 26.0 @xmath11 & 0.17 @xmath12 + this can be understood from the length of the v observation being about a fifth of the b , the sky brightness in v was about 1.5 magnitudes higher than in b and the detector used in 1996 had a bi - alkali photocathode with a b dqe higher than v ( @xcite ) .
no significant signal was observed in february 1995 - mainly due to the smaller telescope aperture .
given these considerations the upper limits of our data are consistent with the level of pulsations remaining constant . in order to estimate the pulsed fraction the background level had to measured .
for each data set the background was determined as the mean of the signal in a small annulus , of radius 2``.5 and width 0''.25 around the position of geminga .
the magnitude calibration was achieved using the star g ( @xcite ) which was always in the field of view .
our measured pulse fraction is consistent with 100% , but , we should stress that there is a large error in this value , due to uncertainties in measuring the sky background , as seen in figure 2
. we can however give a @xmath13 upper limit to the unpulsed component of @xmath14 .
our results agree qualitatively with the previously observed optical flux , but the form of the optical light curve resembles the @xmath0 and hard x - ray rather than the soft x - ray signature , implying a magnetospheric origin . in @xmath0-rays the spectrum is double peaked with maxima at phases 0.25 and 0.75 ( @xcite ) .
the soft x - ray is characterized by a sinusoidal modulation consistent with thermal emission with two temperatures ( @xmath15 @xmath16 and @xmath17 @xmath18k ) or from a single thermal source and a power spectrum ( @xcite , @xcite ) .
euve satellite observations , albeit with low significance , indicate that the extreme uv timing profile is similar in shape to the soft x - ray light curve .
recent reported detections of radio emission ( @xcite , @xcite ) do not show a consistent pulse profile pattern and geminga s radio emission is still a topic of some debate . however the phase agreement of the @xmath0-ray light curve with the radio profile of kuzmin and losovskii , using the egret ephemeris , would suggest that peak 2 corresponds to direct polar emission as reported in their paper . when this is combined with the doubled pulsed optical , hard x - ray and @xmath0-ray profile we can attempt to constrain models for the high energy emission .
the high energy pulses are about 0.5 phase apart which rules out a geometry similar to the crab pulsar .
it would seem that geminga has a magnetic axis at about @xmath19 to the rotation axis , with the emission sites situated close to the neutron star surface .
interestingly when we fold our data , using the malofeev and malov ephemeris , one of the optical pulse peaks arrives at phase zero - expected from most emission models where phase zero generally corresponds to the radio peak .
what is not clear is why the malofeev and malov pulse is delayed to phase 0.4 .
from figure 2 we can see that the signal shows two peaks with a phase separation of @xmath20 0.5 .
we can understand this in terms of the extrapolation of the pulsed @xmath0 emission to optical wavelengths , which is valid for the crab pulsar .
@xmath0-ray emission from geminga has been suggested to be variable ( @xcite ) in both total intensity and spectral index .
our optical observations are within the spread of variation and wide dispersion from the higher energy points ( @xcite ) .
halpern & wang s more recent analysis fitted the x - ray and hard uv data to a black - body spectrum ( t @xmath21 k ) with a power law . the low energy extrapolation of their black - body fit would produce an optical flux 2 magnitudes fainter than we observe , whilst the power law would produce a flux @xmath22 5 magnitudes brighter . as with the crab pulsar this points to a probable flattening of the spectrum in the uv region .
we should also note this extension from the x - ray to the optical does not require a proton cyclotron feature ( @xcite ) to explain the v data but it might imply that there is preferential absorption in i. this could be due to a number of processes including electron cyclotron resonance scattering or synchrotron self - absorption ( unlikely in this instance ) . such a turnoff has been observed spectroscopically in geminga ( @xcite ) albeit with low significance .
as similar i attenuation is observed in both the crab and vela pulsars spectra we are begining to develop an understanding whereby the high energy emission from these objects comes from a similar process .
our results when taken in conjunction with the spectroscopy studies favour an electron synchrotron origin for the radiation at least up to the edge of the b band .
differences can be understood in terms of the viewing geometry and path length over which the electrons are radiated .
deep , phase resolved observations in v , r , i and h will be crucial in determining whether there is a continuum infra - red turnover or whether there is an absorption line .
such observations , including a determination of the pulse fraction , will adequately separate the thermal and non - thermal components with important ramifications for models of pulsar emission .
indeed , the unpulsed components would constrain geminga s thermal continuum , which in conjunction with euve and x - ray data and the known distance would provide definitive estimates , rather than upper limits ( @xcite ) , on the neutron star s size . _
eso and goddard space flight centre is thanked for the provision of their mama detector .
peter sinclair and the eso detector workshop at la silla are thanked for their invaluable help during the eso observations .
we are also grateful to the 6-m telescope program committee of the ras for observing time allocation .
we thank the engineers of sao ras , a. maksimov for help in equipment preparation for the observations and the director of sao ras yu .
balega for arranging the observations .
this work was supported by the russian foundation of fundamental research , state programme `` astronomy '' , russian ministry of science and technical politics , and the science - educational centre `` cosmion '' . the support of forbairt , the irish research and development agency , is gratefully acknowledged .
peter o kane of university of galway is thanked for technical assistance during the observations . _
caraveo , p. a. , bignami , g. f. , mignani , r. & taff , l. g. , 1996 , ap j , 461 l91 cullum , m. , 1990 , the mama detector users manual , eso bertsch , d. l. et al . , 1992 , nature , 357 306 bignami , g. f. , caraveo , p. a. & paul , j. a. , 1988 , a & a , 202 l1 bignami , g. f. & caraveo , p. a. , 1992 , nature , 357 287 bignami , g. f. , caraveo , p. a. & mereghetti , s. , 1993 , nature , 361 704 bignami , g. f. , caraveo , p. a. , miganni , r. , edelstein j. & bowyer , s. , 1996 , ap j , 456 , l111 bignami , g. f. , 1997 , milano astrophysics preprint , n.188 buccheri , r. , & de jager , o. c. , 1989 , timing neutron stars , eds . h. b. ogelman & e. p. j. van den heuvel , kluwer , 95 fichtel , c. e. et al . , 1975 , ap j , 198 , 163 halpern , j. & holt , s. 1992 , nature 357 222 halpern j.p . and tytler , d. , 1988 , apj , ap j , 330 201 halpern , j. p. & ruderman , m. , 1993 , apj , 330 , 201 halpern , j. p. & wang , f. y , 1997 , ap .
j. , 477 , 905 halpern , j.p . , martin , c. , & marshall , h.l . , 1996 , apj , 473 , l37 komarova , v. n. , beskin , g. m. , neustroev , v. v. & plokhotnichenko , v. l. , 1996 , journal of the korean astronomical society , 29 , 217 kuzmin , a. d. & losovskii , b. y. , 1997 , astronomy letters , 23 , 283 lyne a. , pritchard . , 1996 , crab timing ephemeris , university of manchester martin , c. , halpern , j. p. & schiminovich , d. , submitted to ap j mattox , j , halpern , j. p. & caraveo , p. a. , 1996 , a&as , 120 c77 malofeev , v. m. & malov , o. i.,1997 , nature , 387 , 697 mayer - hasselwander , h. a. et al , 1994 , ap j , 421 , 276 nasuti , f. p. , mignami , r. , caraveo , p. a. & bignami , g. f. , 1996 , a & a , 314 , 849 pavlov , g. g. , stringfellow , g. s. & cordova , f. a. , 1996 , apj , 467 , l370 ramanamurthy , p. v. , 1995 , ap j , 450 , 791 redfern r. m. , shearer a. , wouts r. , okane p. , obyrne c , jordan , b. m. , 1993 , proc .
iau coll 136 , 137 shearer , a. , butler , r. , redfern , r. m. , cullum , m. & danks , a.c .
, 1996 , ap j , 473 , 115 smith , f. g. , 1986 , mnras , 219 , 729 timothy , j. g. & bybee , r. l. , 1986 , s.p.i.e .
, 687 , 1090 walter , f. m. & matthews , l. d. , 1997 , nature , 389 358 | we present optical data which shows that g " , the optical counterpart of the @xmath0-ray pulsar geminga , pulses in b with a period of 0.237 seconds .
the similarity between the optical pulse shape and the @xmath0-ray light curve indicates that a large fraction of the optical emission is non - thermal in origin - contrary to recent suggestions based upon the total optical flux .
the derived magnitude of the pulsed emission is @xmath1 . whilst it is not possible to give an accurate figure for the pulsed fraction ( due to variations in the sky background ) we can give an upper limit of @xmath2 for the unpulsed fraction . |
the discovery of the @xmath1 meson ( the lowest pseudoscalar @xmath5 state of the @xmath6 system ) has been reported in 1998 by the cdf collaboration in the 1.8 tev @xmath7 collisions at the fermilab tevatron @xcite .
the mass has been measured to be @xmath8 gev .
the fact that the mass of the quarks of quark antiquark systems built up by @xmath9 and @xmath10 quarks is much larger than the typical binding energy suggests that these systems are non - relativistic , i.e. that the heavy - quark velocity @xmath11 is small .
the typical scales of these systems are the binding energy @xmath12 and the momentum transfer @xmath13 ; moreover , because of the non - relativistic nature of the system , @xmath14 ( for the purpouses of this discussion @xmath15 and @xmath11 can be identified with the mass and the velocity of the lightest component of the bound state respectively ) .
let us call @xmath16 the scale at which non - perturbative effects become important .
if @xmath17 , then the scale @xmath18 can be integrated out order by order in @xmath19 at a scale @xmath20 .
the system is described up to order @xmath0 by a potential which is entirely accessible to perturbative qcd and at the leading order is the coulomb potential .
non - potential effects start at order @xmath21 @xcite .
this kind of system is called coulombic .
non - perturbative effects are of non - potential type . in the particular situation
@xmath22 they can be encoded into local condensates @xcite .
this condition seems to be fulfilled by the bottomonium ground state , which has been studied in this way in @xcite .
also the charmonium ground state has been analysed as a coulombic bound state by the same authors .
in both cases ( but with caveats ) the non - perturbative corrections la voloshin leutwyler , i.e. in terms of local condensates , have been claimed to be under control @xcite . for heavy quarkonium states higher than the ground state the condition @xmath17 is not fulfilled and non - perturbative terms affect the potential .
the system is no longer coulombic .
traditionally the energy of these systems has been calculated within qcd - inspired confining potential models .
a large variety of them exists in the literature and they have been on the whole quite successful ( cf .
@xcite for some recent reviews ) . however , the usual criticisms apply .
their connection with the qcd parameters is hidden , the scale at which they are defined is not clear , they can not be systematically improved and they usually contain a superposition by hand of perturbative and non - perturbative effects .
for this reason a lot of effort has been devoted , over the years , to obtaining the relevant potentials from qcd by relating them to some wilson loops expectation values @xcite .
anyway , these have to be eventually computed either via lattice simulations or in qcd vacuum models @xcite . in the specific situation
@xmath23 the scale @xmath24 can still be integrated out perturbatively , giving rise to a coulomb - type potential .
non - perturbative contributions to the potential will arise when integrating out the scale @xmath16 .
this situation has been studied in @xcite .
we will call quasi - coulombic the systems described by the situation @xmath25 , when the non - perturbative piece of the potential can be considered small with respect to the coulombic one and treated as a perturbation .
the only available theoretical predictions ( to our knowledge ) of the @xmath1 mass resort to ( confining ) potential models or to the lattice . in this work we will carry out the calculation of the perturbative @xmath1 mass up to order @xmath0 . we will call it @xmath26 ) may be seen as a definition of what we call here perturbative @xmath1 mass .
analogous definitions have to be understood for the perturbative @xmath27 and @xmath28 masses . ] .
this calculation will be relevant to a qcd determination of the @xmath6 ground state if this system is coulombic or at least quasi - coulombic .
moreover , in the way we are doing the calculation , we also assume the @xmath28 and the @xmath27 to be coulombic or at least quasi - coulombic systems .
the question if these assumptions correspond to the actual systems can not be settled at this point . on the other hand
there is no _ a priori _ reason to rule them out .
let us consider , for instance , the argument used in @xcite for the @xmath27 .
lattice data show that the static potential clearly deviates from a @xmath29 behaviour for distances larger than 1 gev@xmath30 ( see @xcite and references therein ) .
therefore the @xmath27 is coulombic or quasi - coulombic if the characteristic scale of the bound state , @xmath31 , is bigger than 1 gev .
if we assume @xmath32 gev and if we fix that scale on the bohr radius , @xmath33 , of the @xmath27 , @xmath34 , then we get @xmath35 gev . since this scale falls into the energy window between the mass scale @xmath36 and 1 gev , these figures are consistent with a coulombic or quasi - coulombic picture of the @xmath27 .
, as we will see later on in this paper , the expansion that we get shows a still convergent behaviour . ]
the main problem of the calculation of the perturbative @xmath1 mass is the well - known bad convergence of the perturbative series when using the pole mass .
this is due to a renormalon cancellation occurring between the pole mass and the static coulomb potential @xcite .
we handle the problem by expressing the @xmath10 and the @xmath9 pole mass in the perturbative expression of the @xmath1 mass as half of the perturbative mass of the @xmath27 ( @xmath37 ) and the @xmath28 ( @xmath38 ) respectively .
this corresponds to using the quark mass in the so - called @xmath2 scheme introduced in @xcite . in this way , by expressing @xmath26 in terms of quantities that are infrared safe at order @xmath39 ( the @xmath40 perturbative masses ) , the pathologies of the perturbative series , due to the renormalon ambiguities affecting the pole mass , are cured .
we will explicitly show that , in fact , we obtain a better convergence of the perturbative expansion and a stable determination of the perturbative mass of the @xmath1 , just in the energy range that the above discussion on the @xmath27 suggests to be also the relevant one for the @xmath1 .
non - perturbative terms are of potential type in the quasi - coulombic situation and of non - potential type in the coulombic situation .
they affect the identification of the perturbative masses @xmath37 , @xmath38 and @xmath26 with the corresponding physical ones . if we aim at obtaining a good estimate of the physical @xmath1 mass , it is not important for each of these contributions to be individually small , as long as the sum of them in the @xmath1 mass is small . as we will discuss at the end , a picture with non - perturbative corrections to
the @xmath1 mass of a not too large size ( @xmath41 mev ) seems to be consistent with the experimental data and with the potential models .
the paper is organized in the following way . in the next section we set up the formalism and
perform the calculation of the perturbative @xmath1 mass . in section 3
we briefly discuss the non - perturbative corrections and compare our result with other determinations of the @xmath1 mass available in the literature .
in order to calculate the @xmath1 mass in perturbation theory up to order @xmath0 , we need to consider the following contributions to the potential : the perturbative static potential at two loops , the @xmath42 relativistic corrections at one loop , the spin - independent @xmath43 relativistic corrections at tree level and the @xmath44 correction to the kinetic energy .
we will not consider here effects due to a non - zero charm quark mass on the @xmath9 ( and the @xmath45 system ) of the type discussed in @xcite .
we will follow the derivation of the heavy quarkonium mass of ref .
@xcite .
the static potential at two - loops has been calculated in @xcite .
it is useful , in order to perform an analytic calculation , to split it as @xmath46 where @xmath47 is the part that does not contain logarithms , @xmath48 \right .
\nonumber\\ & & \hspace{2 mm } \left .
+ \left({\alpha_{\rm s } \over \pi}\right)^2 \left [ \beta_0 ^ 2 \left ( { \gamma_e^2\over 4 } + { \pi^2\over 48}\right ) + \left ( { \beta_1\over 8 } + { 5\over 12 } \beta_0 ^ 2 - { 2\over 3 } c_a \beta_0 \right ) \gamma_e + { c\over 16 } \right]\right\ } , \nonumber\end{aligned}\ ] ] where @xmath49 are the @xmath50-function coefficients , @xmath51 , ... ] , @xmath52 @xmath53 @xmath54 @xmath53 @xmath55 @xmath56 @xmath57 , @xmath58 is the euler constant , @xmath59 , @xmath60 and @xmath61 is the number of flavours ( we will take @xmath62 ) and @xmath63 @xmath64 230 mev has been checked to be consistent with the central value and the errors of eq .
( [ ebc ] ) . ] ; @xmath65 is given by @xmath66\right\}.\end{aligned}\ ] ] the strong coupling constant @xmath19 is understood in the @xmath67 scheme . at the scale @xmath68
we will take the value of @xmath19 from the three - loop expression with @xmath69 @xmath64 @xmath70 mev .
the @xmath42 relativistic corrections at one loop , the @xmath43 tree - level spin - independent terms and the @xmath44 correction to the kinetic energy are given by @xcite @xmath71 @xmath72 , @xmath36 and @xmath73 are the @xmath9 , the @xmath10 and the reduced pole mass respectively .
then , the hamiltonian relevant in order to get the @xmath1 mass at @xmath74 accuracy is @xmath75 up to order @xmath74 the ground - state energy is given by ( @xmath76 means the average on the ground state ) : @xmath77 where at leading order @xmath78 with @xmath79 , the bohr radius of the system .
this corresponds to the only contribution relevant at order @xmath0 produced by the hamiltonian ( [ bc2 ] ) in second - order perturbation theory ( @xmath80 stands for the coulombic intermediate states ) and can be read off from the second reference in @xcite .
the other averages can be easily evaluated by means of the standard formulas @xmath81 after an explicit calculation we get from eq .
( [ ebc ] ) , up to order @xmath0 : @xmath82 \right .
\nonumber\\ & & \hspace{-10 mm } + \left({\alpha_{\rm s}\over \pi}\right)^2 \left [ { 3\over 4 } \beta_0 ^ 2 \ln^2\left({2 c_f \alpha_{\rm s } m_{\rm
red}\over \mu}\right ) + \left(2 c_a \beta_0 - { 9\over 4 } \beta_0 ^ 2 - { \beta_1\over 4 } \right ) \ln\left({2 c_f \alpha_{\rm s } m_{\rm red}\over \mu}\right ) \right . \nonumber\\ & & \hspace{-5 mm }
- \pi^2 c_f^2 \left ( { 1\over m_b^2 } + { 1\over m_c^2 } - { 6\over m_b m_c } \right ) m_{\rm red}^2 + { 5\over 4 } \pi^2 c_f^2 \left ( { 1\over m_b^3 } + { 1\over m_c^3}\right ) m_{\rm red}^3 \nonumber\\ & & \hspace{-5 mm } \left.\left .
+ \pi^2 c_f c_a + { 4\over 9 } c_a^2 - { 17 \over 9 } c_a \beta_0 + \left ( { 181 \over 144 } + { 1\over 2 } \zeta(3 ) + { \pi^2\over 24 } \right ) \beta_0 ^ 2 + { \beta_1 \over 4 } + { c\over 8 } \right]\right\ } , \label{ebc2}\end{aligned}\ ] ] where @xmath83 .
the main problem connected with the perturbative series ( [ ebc2 ] ) is the bad convergence in terms of the heavy - quark pole masses .
let us consider , for instance , @xmath84 gev , @xmath85 gev and @xmath86 gev .
then we get @xmath87 mev @xmath88 mev , where the second , third and fourth figures are the corrections of order @xmath89 , @xmath90 and @xmath0 respectively .
the series turns out to be very badly convergent .
this reflects also in a strong dependence on the normalization scale @xmath68 : at @xmath91 gev we would get @xmath92 mev , while at @xmath93 gev we would get @xmath94 mev . and @xmath9 pole masses , which are poorly known .
see the following discussion . ]
the origin of this behaviour can be understood in the renormalon language .
the pole mass is affected by an ir renormalon ambiguity that cancels against an ir renormalon ambiguity of order @xmath39 present in the static potential @xcite .
the non - convergence of the perturbative series ( [ ebc2 ] ) signals the fact that large @xmath95 contributions ( coming from the static potential renormalon ) are not summed up and cancelled against the pole masses . a possible solution , in order to avoid large perturbative corrections and large cancellations , or , in other words , in order to obtain a well - behaved perturbative expansion , is to resort to a different definition of the mass .
the so - called @xmath2 mass of a heavy quark @xmath96 is defined as half of the perturbative contribution of the @xmath97 @xmath98 mass @xcite . unlike the pole mass , the @xmath2 mass , containing , by construction ,
half of the total static energy @xmath99 , is free of ambiguities of order @xmath39 .
our strategy will be the following .
first , we consider the perturbative contribution ( up to order @xmath100 ) of the @xmath97 levels of charmonium and bottomonium : @xmath101 which are respectively a function of the @xmath10 and the @xmath9 pole mass and can be read off from eq .
( [ ebc2 ] ) in the equal - mass case , adding to it the spin - spin interaction energy : @xmath102 .
we invert these relations in order to obtain the pole masses as a formal perturbative expansion depending on the @xmath2 mass . finally , we insert the expressions @xmath103 and @xmath104 in eq .
( [ ebc2 ] ) . at this point
we have the perturbative mass of the @xmath1 as a function of the @xmath27 and @xmath28 perturbative masses @xmath105 if we identify the perturbative masses @xmath37 , @xmath38 with the physical ones , i.e. @xmath106 mev and @xmath107 mev @xcite , then the expansion ( [ bc ] ) depends only on the scale @xmath68 .
( 70,0)=10truecm ( 80,100)@xmath108 ( 240,5)@xmath68 0.2truecm 0.6truecm [ plotper ] in fig .
[ plotper ] we show the dependence on @xmath68 of the order @xmath109 and @xmath110 contributions to @xmath26 respectively . taking into account that the order @xmath109 contribution vanishes at @xmath111 gev , the perturbative series seems to be reliable for values of @xmath68 bigger than @xmath112 gev and lower than @xmath113 gev .
for instance , @xmath114 mev at the scale @xmath84 gev .
this is consistent with : _
i ) _ the fact that , for values of @xmath68 close to or less than 1 gev , the perturbative calculation ( and the initial assumption that @xmath1 is coulombic or quasi - coulombic ) is expected to break down ; _ ii ) _ the fact that higher values of @xmath68 do not correspond to the characteristic scale of the system ( this is signalled by the appearance of big logarithms in the perturbative expansion ) ; _ iii ) _ the estimate of the scale @xmath68 inferred in the introduction from the size of the @xmath27 . more precisely we will take in our analysis , would not change our final result ( [ ebc ] ) .
e.g. taking 1.2 gev @xmath115 2.6 gev we would get , by keeping the same central values as above , @xmath116 . ] 1.2 gev @xmath115 2.0 gev and 250 mev @xmath117 350 mev ( in terms of @xmath118 this corresponds to @xmath119 ) . in this way
we entirely cover the energy range used in @xcite in order to study the @xmath27 .
1.76 gev . ]
( 70,0)=10truecm ( 55,100)@xmath120 ( 240,5)@xmath68 0.2truecm 0.6truecm [ plot ] by varying @xmath68 from 1.2 gev to 2.0 gev and @xmath69 from 250 mev to 350 mev and by calculating the maximum variation of @xmath26 in the given range of parameters , we get as our final result @xmath121 the upper limit corresponds to the choice of parameters @xmath122 mev , @xmath123 gev , while the lower limit to @xmath124 mev and @xmath93 gev . as a consequence of the now obtained good behaviour of the perturbative series in the considered range of parameters , our result appears stable with respect to variations of @xmath68
( see fig .
[ plot ] ) .
we would like to note that the main source of error in eq .
( [ ebc ] ) comes from the border region 1.2 gev @xmath125 1.4 gev at @xmath126 350 mev , where it may become questionable to treat the @xmath1 as a coulombic or quasi coulombic system ( see fig .
[ plotper ] ) .
we have calculated the perturbative @xmath1 mass as defined by eq .
( [ ebc ] ) .
the problem of the bad behaviour of the perturbative series has been overcome by expressing the perturbative @xmath1 mass in terms of the perturbative @xmath27 and @xmath28 masses .
the series we obtain has a good convergent behaviour .
this fact is relevant since it shows that the scale hierarchy considered in the introduction ( @xmath127 ) , which led to the hamiltonian ( [ bc2 ] ) , correctly applies to the system under consideration .
system does not show any sign of convergence . ] in other words , the result we get is consistent with the assumption made that the @xmath1 system is coulombic or quasi - coulombic .
moreover , the perturbative series turns out to be weakly sensitive to variations of @xmath68 ( the renormalization scale ) and @xmath69 in the range 1.2 gev @xmath115 2.0 gev and 250 mev @xmath117 350 mev .
the result appears , therefore , reliable from a perturbative point of view .
non - perturbative contributions have not been taken into account so far .
they affect the identification of the perturbative masses @xmath26 , @xmath38 , @xmath37 , with the corresponding physical ones through eq .
( [ ebc2 ] ) .
let us call these non - perturbative contributions @xmath128 , @xmath129 and @xmath130 respectively .
as discussed in the introduction , depending on the actual kinematic situation of the system , they can be of potential or non - potential nature . in the last case
they can be encoded into non - local condensates or into local condensates .
there is no way to discriminate among these situations , since the size of what would be the energy scale @xmath131 of the system with respect to @xmath16 is unknown .
non - perturbative contributions affect the identification of eq .
( [ bc ] ) with the physical @xmath1 mass roughly by an amount @xmath132 @xmath133 @xmath134 . assuming @xmath135 mev , @xmath136 mev and @xmath137 , the identification of our result ( [ ebc ] ) with the physical @xmath1 mass may , in principle , be affected by uncertainties , due to the unknown non - perturbative contributions , as big as @xmath138 mev .
however , the different @xmath139 are correlated , so that we expect , indeed , smaller uncertainties .
if we assume , for instance , @xmath129 and @xmath130 to have the same sign , which seems to be quite reasonable , then the above uncertainty reduces to @xmath140 mev . constraining even more the form of @xmath139 , by evaluating it from the voloshin leutwyler formula ( i.e. in terms of local condensates ) , as given in ref .
@xcite , we get a negative contribution ( since the term coming from the @xmath27 is the dominating one ) of less than @xmath141 mev .
this feature , if preserved also in the other kinematic situations , would confirm , indeed , that the effect of the non - perturbative contributions is not too big and that its effect is to lower down the perturbative result given in eq .
( [ ebc ] ) .
more quantitative statements are difficult to make , since , differently from the perturbative case discussed in the previous section , they appear to be dependent on the choice of the parameters . the result we get in eq .
( [ ebc ] ) is compatible with the experimental value @xmath142
@xmath64 @xmath143 gev reported in @xcite .
we mention that opal reports in @xcite 2 candidates @xmath1 in hadronic @xmath144 decays events , with an estimated mass @xmath145 gev . also this value compares favourably with ours .
having more precise and established experimental data will make it possible to make some more definite statements .
in particular , it will be possible to give , inside a coulombic or quasi - coulombic picture , a precise estimate of the size of the non - perturbative effects in the @xmath1 mass . in the table , we also report , for comparison , some of the other determinations of the @xmath1 mass available in the literature .
the results quoted in @xcite rely on potential models ( essentially a coulomb plus a confining potential ) and are reported without errors .
the figure that appears in the table in correspondence of ref .
@xcite refers to an average of different models performed by those authors .
finally @xcite reports the result of a very recent lattice calculation .
we would like to note that , if one assumes that potential models give a @xmath1 mass close to reality , then , comparing the potential model predictions with eq .
( [ ebc ] ) , non - perturbative contributions seem to be of the order @xmath146 mev ( consistently with expectations , non - perturbative corrections become as smaller as perturbation theory better works , i.e. in correspondence of low values of @xmath69 and high values of @xmath68 ) . finally , it is interesting to notice that these figures are completely consistent with the general discussion on the uncertainties , coming from non - perturbative contributions , done above .
.mass of the @xmath147 meson .
the result labelled with eq .
( [ ebc ] ) refers to the present work .
[ cols="<,<",options="header " , ] [ tab1 ]
the authors thank andre hoang and antonio pineda for valuable comments and suggestions .
f. abe et al . , cdf collaboration , phys . rev
. lett . * 81 * , 2432 ( 1998 ) .
n. brambilla , a. pineda , j. soto and a. vairo , nucl . phys . * b 566 * , 275 ( 2000 ) .
n. brambilla , a. pineda , j. soto and a. vairo , phys . lett .
* b 470 * , 215 ( 1999 )
. m. b. voloshin , nucl . phys .
* b 154 * , 365 ( 1979 ) ; h. leutwyler , phys . lett . *
b 98 * , 447 ( 1981 ) .
s. titard and f. j. yndurain , phys . rev .
* d 49 * , 6007 ( 1994 ) ; a. pineda and f. j. yndurain , phys . rev .
* d 58 * , 094022 ( 1998 ) ; * d 61 * , 077505 ( 2000 ) .
a. pineda , nucl .
* b 494 * , 213 ( 1997 ) . n. brambilla and a. vairo , hep - ph/9904330 ; g. bali , hep - ph/0001312 . k. g. wilson , phys
d 10 * , 2445 ( 1974 ) ; e. eichten and f. l. feinberg , phys .
lett . * 43 * , 1205 ( 1979 ) ; d. gromes , z. phys . * c 26 * , 401 ( 1984 ) ; a. barchielli , e. montaldi and g. m. prosperi , nucl . phys . * b 296 * , 625 ( 1988 ) and erratum ibid . * b 303 * , 752 ( 1988 ) ; a. barchielli , n. brambilla and g. prosperi , nuovo cimento * 103 a * , 59 ( 1990 ) ; n. brambilla , a. pineda , j. soto and a. vairo , hep - ph/0002250 . n. brambilla and a. vairo , phys . rev . * d 55 * , 3974 ( 1997 ) .
m. beneke , phys .
* b 434 * , 115 ( 1998 ) ; a. hoang , m. smith , t. stelzer and s. willenbrock , phys .
d 59 * , 114014 ( 1999 ) ; a. pineda , ph .
d. thesis , ( barcelona ) , jan .
1998 ; n. uraltsev , hep - ph/9804275 .
a. h. hoang , z. ligeti and a. v. manohar , phys .
lett . * 82 * , 277 ( 1999 ) ; a. hoang , hep - ph/9909356 . a. h. hoang and a. v. manohar , phys .
* b 483 * , 94 ( 2000 ) .
y. schrder , phys . lett .
* b 447 * , 321 ( 1999 ) ; m. peter , phys .
78 * , 602 ( 1997 ) .
s. n. gupta and s. f. radford , phys .
* d 25 * , 3430 ( 1982 ) . c. caso et al .
, particle data group , eur .
j. * c 3 * , 1 ( 1998 ) ; http://pdg.lbl.gov/. e. eichten and c. quigg , phys .
* d 49 * , 5845 ( 1994 ) . s. s. gershtein , v. v. kiselev , a. k. likhoded and a. v. tkabladze , phys .
38 , 1 ( 1995 ) ; a calculation which leads to the same numerical value for the @xmath1 mass can be found in v. o. galkin , a. yu . mishurov and r. n. faustov , sov .
. phys . * 55*(8 ) , 1207 ( 1992 ) .
l. p. fulcher , phys . rev . *
d 60 * , 074006 ( 1999 ) .
e. bagan , h. g. dosch , p. gosdzinsky , s. narison and j. m. richard , z. phys . * c 64 * , 57 ( 1994 ) .
h. p. shanahan , p. boyle , c. t. h. davies and h. newton , phys .
* b 453 * , 289 ( 1999 ) .
k. ackerstaff et al .
, phys . lett .
* b 420 * , 157 ( 1998 ) . | we evaluate in perturbative qcd , up to order @xmath0 , the mass of the @xmath1 .
we use the so - called @xmath2-mass in order to improve the convergence of the perturbative series .
our result is @xmath3 .
non - perturbative effects are discussed .
a comparison with potential models seems to be consistent with non - perturbative contributions of the order @xmath4 . +
pacs numbers : 12.38.aw , 12.38.bx , 14.40.lb , 14.40.nd = 10000 = 5 pt = 24 pt 0.25 cm |
the current tally of gravitationally lensed quasars now stands at nearly 70 ( kochanek et al .
the various searches which have contributed to this total have naturally also discovered examples of the apparently rarer phenomenon of binary quasars , of which only @xmath1 systems have been documented ( mortlock , webster , & francis 1999 ; kochanek et al .
2002 ) . while the gravitationally lensed quasars have been intensively studied , binary quasars have not received nearly as much attention , even though binary quasars may provide invaluable insight into various aspects of the quasar phenomenon . of particular importance are the cases where the pair members exhibit strikingly different characteristics , perhaps making it possible to deduce what aspects of the agn environment are responsible for presently little - understood quasar behavior such as radio - loudness , broad or narrow absorption lines , or very red colors .
in addition , physically close binaries also supply limits , statistically at least , on the timescales involved because once within @xmath2 galaxy radius ( @xmath3kpc ) , dynamical considerations limit the binary lifetimes ( mortlock et al .
1999 ) . in an ongoing infrared imaging survey of quasars at keck observatory
, we have discovered that sdss j233646.2 - 010732.6 ( hereafter sdss 2336 - 0107 ) is a double with a separation of 167 .
resolved spectra show that component a is a standard quasar with a blue continuum and broad emission lines .
component b is a broad absorption line ( bal ) quasar , specifically , a bal qso with prominent absorption from and metastable , making it a member of the `` felobal '' class ( becker et al .
1997 ; 2000 ) .
the number of known felobals has increased dramatically in the last five years ( becker et al .
2000 ; menou et al . 2001 ; hall et al .
2002 ) , including a gravitationally lensed example ( lacy et al .
2002 ) and the binary discussed here , suggesting that this type of object , once thought to be rare , may in fact be fairly common and simply overlooked in most quasar surveys .
the presence of this bal quasar in a relatively small separation binary adds to the mounting evidence that the bal phenomenon is not simply due to viewing a normal quasar at a specific orientation , but rather that bals are an evolutionary phase in the life of many , if not all , quasars , and is associated with conditions found in interacting systems .
we adopt @xmath4 , @xmath5 , and @xmath6 .
sdss 2336 - 0107 was first identified by the sloan digital sky survey ( sdss ; york et al .
2000 ) using an early version of the quasar target selection algorithm ( richards et al . 2002 ) , as a quasar with z=1.285 and having modest broad absorption features ( schneider et al .
it was also earmarked as an `` extended '' quasar in the sdss images , meaning that it was slightly resolved .
quasars at @xmath7 are unresolved in sdss images , so the extended nature made it a target in our imaging survey for lensed quasars at keck observatory .
on 2001 october 30 , we obtained deep imaging of sdss 2336 - 0107 , totaling 350s in the @xmath8 band using the near infrared camera ( nirc , matthews & soifer 1994 ) on the keck i 10 m telescope in 06 seeing .
the image scale is 015 per pixel .
the reduced image reveals the double nature of sdss 2336 - 0107 ( figure 1 ) and three galaxies ( g1 , g2 , and g3 ) in the field .
photometry of the two components was measured with daophot / iraf , using a bright standard star as a psf model . the short exposure standard star psf clearly differed from the psf of the longer exposure , dithered images of the binary quasar , so we also used component b as a psf model , even though the wings of the psf can not be determined easily in this way . in both cases ,
the residuals after subtraction suggest that the shapes of two components are not quite identical and that there is extended diffuse light at @xmath8 around a , perhaps indicating that the data are just beginning to detect the host galaxies of one or both .
the separation of the two components is determined to be 1673 in @xmath8 and 1681 in @xmath9 .
the sdss images were obtained in relatively poor seeing , ranging from 25 in @xmath10 to 16 in @xmath11 . despite the seeing , the images prove adequate for photometry of the two components using daophot / iraf , a nearby bright star serving as the psf model and photometric zeropoint . in @xmath12 ,
the psf fits yield relatively small photometric errors ; the separations range from 15 to 17 , averaging @xmath13 , consistent to a fraction of an sdss pixel ( 0396 ) with the separation measured in the nirc images , lending credibility to the extracted photometry .
the redder component b is not reliably detected in @xmath10 , but an upper limited is derived from the residuals after subtracting a single psf at the location of component a. table 1 lists the optical and ir photometry , corrected for galactic reddening ( schlegel , finkbeiner , & davis 1998 ) .
we obtained resolved spectra of the two components of sdss 2336 - 0107 in 2002 january , using the echelle spectrograph and imager ( esi ; epps & miller 1998 ) on the keck ii telescope .
the seeing was again 06 .
the esi has a dispersion of 0.15 to 0.3 per pixel over a wavelength range of 4000 to 10500 , and the 1 slit used for these observations projects to 6.5 pixels .
the 900 s exposure was obtained at an airmass of 1.33 , with the slit aligned at the position angle of the components , 958 .
the spectra show that component a is a standard quasar with broad emission lines while b is found to be a bal quasar , totally lacking any emission features ( figure 2 ) .
the shape of the overall spectral energy distributions have been corrected to agree with the photometry from the sdss images .
the redshift of component a is @xmath14 from , and @xmath15 from ] ; as ] can be contaminated by , ] , and emission , we adopt the redshift .
the redshift of component b is difficult to measure accurately because of the lack of emission features and the broad nature of the absorption .
if we assume that the absorption feature at 6418.8 , which is somewhat broad and flat - bottomed , is due to the red half the the doublet at 2803.5 , the redshift is 1.2895 .
there is also a narrow absorption line at 6525.0 which , ascribed to 2853.0 , gives @xmath16 , consistent with and perhaps more accurate than the estimate . adopting this as the redshift of component b yields an apparent velocity difference between a and b of only @xmath17 .
component b also has absorption features consistent with ca h & k at a redshift of 1.2843 ( figure 2 ) .
the esi spectrum of component a reveals a @xmath18 intervening absorption system at @xmath19 , perhaps due to one of the faint galaxies which appear in the @xmath8-band image ( figure 1 ) .
there is also absorption from what may be 2515 and 2722 .
the brighter galaxy ( g1 ) is 125 to the east of a while the fainter ( g2 ) is 230 to the west , corresponding to physical impact parameters of 9.4 and 17.4 kpc , respectively .
also weakly detected in the spectrum of component a is [ ] in emission at a wavelength corresponding to @xmath20 ( figure 2 ) , within 400 of the emission features .
this feature is in a region between bright sky lines and also appears on a shorter esi exposure of sdss 2336 - 0107 taken two nights previous to the data presented here , so its reality is beyond doubt .
though noisy , the [ ] feature does appear to be broadened by a few hundred , well in excess of the instrumental resolution .
combining the redshifts with the sdss and nirc photometry , the absolute magnitudes of the two components can be computed in the quasar restframe as @xmath21 and @xmath22 . while not exceptionally bright , both components are luminous enough to qualify as bona fide quasars . the possibility that sdss 2336 - 0107 is gravitationally lensed
is immediately ruled out by the strikingly different spectra ( figure 2 ) and the short time delay between the two paths .
the analytic formula of witt , mao , & keeton ( 2000 ) yields time delay estimates of a few months to a year for any reasonable lensing geometry .
the nearly identical redshifts and tiny velocity difference , however , makes it a definite binary quasar . for our adopted cosmology
, the apparent separation of 167 translates to a physical separation of @xmath23 kpc , making sdss 2336 - 0107 the smallest separation binary ever found in ground - based imaging and the second smallest ever .
the smallest is lbqs 0103 - 2753 ( junkkarinen et al .
2001 ) , a 03 ( 2.3 kpc ) separation binary found serendipitously during an ultraviolet spectroscopy survey of bal systems using hst and stis . in lbqs 0103 - 2753 ,
the brighter component is a standard high ionization bal ( hibal ) quasar .
the two binary quasars with the smallest physical separations both contain a bal member .
although this could be simply small number statistics , it is natural to speculate that the bal features are somehow induced by the small distances between the agn .
mass transfer onto one of the black holes would naturally result in a tidally interacting pair of galaxies where the agns had reached relatively small separations of less than a galactic radius .
this is further evidence supporting the view that bal features are a short - lived evolutionary phase during the life of a quasar ( becker et al .
2000 ; gregg et al .
2000 ) , rather than the result of viewing an agn along a particular line of sight .
additional support of this argument comes from canalizo & stockton ( 2001 ) who show that all four of the known low ionization bal ( lobal ) quasars with @xmath24 are in systems which exhibit tidal tails or evidence of merging , suggesting a strong link between these phenomena and the presence of bal spectral features .
the orientation picture can not so easily account for the occurrence of bal features in the two smallest separation binary quasars or all four low redshift lobal examples , except as by chance .
if the merger / accretion origin of bals is correct , the material directly responsible for the bal spectral features is almost certainly within tens of parsecs of the agn and not within reach of ground - based imaging resolution , yet mergers and tidal debris will occur over kiloparsec scales .
it is then somewhat counter - intuitive that it is the standard quasar component ( a ) of sdss j2336 - 0107 which is sandwiched by the galaxies g1 and g2 in the nirc image ( figure 1 ) and which exhibits the narrow [ ] in its spectrum ( figure 2 ) .
the @xmath8 image hints at a light bridge between g1 and component a , perhaps a real physical connection .
the weak [ ] feature of component a is certainly consistent with this picture , indicating some level of star formation , quite expected for an intensely interacting system .
some or all of the [ ] could come from the material in the region between g1 and a ; the 1 esi slit would have entirely missed g1 , though .
the intervening system in the spectrum of a ( figure 2 ) could arise from g2 at an intervening location ; direct spectroscopy of the galaxies in this field are required to sort out the physical relationships .
the @xmath8 seeing and s / n are not good enough to explore the region directly between the two components more fully , but space - based imaging would reveal details of the expected tidal interactions which may be driving the bal behavior . if the two quasars are in the nuclei of two large galaxies separated by only @xmath23 kpc , tidal effects should be strongly pronounced . figure 3 plots the photometry of the two components , @xmath10 through @xmath8 , highlighting the contrasting color difference .
component b , the bal , is in fact brighter in the redder bands , by @xmath25 magnitude in @xmath9 and @xmath8 . for comparison , we plot photometry synthesized from the sdss quasar composite ( vanden berk , et al .
2001 ) after redshifting to z=1.2855 .
the synthetic magnitudes have been adjusted so that @xmath11 equals that of component a. the excess flux in @xmath26 in the composite may be due to host starlight contribution in the low redshift quasar spectra which contribute to the composite in this wavelength region .
the even larger excess in the bal component is probably a combination of starlight and `` back warming '' from dust extinction .
there is evidence that felobals are heavily reddened ( hall et al . 1997 ; najita , dey , & brotherton 2000 ) and
may be associated with young galaxies and star formation ( egami et al .
1996 ) .
perhaps the most intriguing finding is that component b is an felobal , a quasar with low ioniziation broad absorption from plus broad and narrow absorption from metastable states of . in figure 4 ,
the spectrum of component b is compared to two other felobals , j1556 + 3517 ( becker et al .
1997 ) and the original of the class , q0059 - 2735 ( hazard et al .
the spectrum of sdss 2336 - 0107 b exhibits strong at @xmath27 and 2750 .
the broad absorption depresses the spectrum blueward to at least 2630 , or -18,000 in the quasar restframe . while not quite as red as j1556 + 3517 , sdss 2336 - 0107 b has essentially no emission lines whatsoever .
there is now one felobal quasar ( sdss
2336 - 0107 ) known among the 15 binary qso systems and one felobal ( j1004 + 1229 ; lacy et al
. 2002 ) among the @xmath28 known gravitationally lensed quasars .
these statistics argue strongly that felobals must be much more common than their presently cataloged population suggests .
lacy et al .
( 2002 ) discuss the details of red quasar luminosity functions and lensing probabilities , but not with regard to felobals in particular . with 1 felobal in @xmath28 lensed quasars and one among 40 quasars in @xmath29 binary systems , the simplest explanation is that from 1% to a few percent of quasars are felobals , or alternatively , that the felobals phase lasts for a similar percentage of the total quasar lifetime .
the very red spectral energy distribution and weak or completely absent broad emission lines ( figure 1 ) makes these objects extremely difficult to find , especially at redshifts @xmath30 where the absorption edge moves into the @xmath9 band .
an ir - selected sample of quasars is needed to better gauge the frequency of felobal numbers , but evidence is mounting ( becker et al .
1997 ; lacy et al . 2002 ; hall et al . 2002 ) that they are much more common than suspected even just a few years ago when the rather blue example of q0059 - 2735 was the only one known ( hazard et al .
a deep imaging survey of bal quasars and a control sample of non - bal quasars , from the ground or space , is needed to document the frequency of each in systems that show tidal interactions , which would be a more comprehensive version of the work done by canalizo & stockton ( 2001 ) . if bal - ness is explained simply by viewing an ordinary quasar along a line of sight which skims the `` dusty torus '' ( weymann et al .
1991 ) , then the two populations should appear with equal frequency in merging or interacting systems . if bal features are produced by conditions created in mergers , and orientation plays less or no role , then they will be more common in chaotic systems with tidal tails and other signs of host galaxy interactions .
dynamical modeling of the tidal effects may help constrain the timescales of various quasar phases .
felobals may be a common feature of the agn landscape , especially if there are seyfert - type luminosity class analogs to the brighter objects now being unearthed .
if felobals are indeed common , then infrared sky surveys which reach deeper than 2mass , such as those being undertaken with sirtf , will turn up numerous examples along with other extreme bal phenomena in quasar and galaxy luminosity objects . the felobal - like hawaii 167 ( cowie et al .
1994 ) was discovered in just such a survey .
such surveys are needed to provide more accurate estimates of the extremely red quasar population , especially at redshifts above 2.5 where such objects are essentially invisible in the optical .
we thank mark lacy for helpful discussions .
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 .
part of the work reported here was done at the institute of geophysics and planetary physics , under the auspices of the u.s .
department of energy by lawrence livermore national laboratory under contract no . w-7405-eng-48 .
funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. llllllrrrrrrl component a & @xmath31 & @xmath32 & 0.114 & 1.2853 & 19.75 & 19.65 & 19.35 & 19.26 & 19.12 & 18.19 & 17.22 & -23.77 + component b & @xmath33 & @xmath34 & 0.114 & 1.2871 & @xmath35 & 21.37 & 19.71 & 18.94 & 18.68 & 17.24 & 16.23 & -24.72 + | in an ongoing infrared imaging survey of quasars at keck observatory , we have discovered that the @xmath0 quasar sdss j233646.2 - 010732.6 comprises two point sources with a separation of 167 .
resolved spectra show that one component is a standard quasar with a blue continuum and broad emission lines ; the other is a broad absorption line ( bal ) quasar , specifically , a bal qso with prominent absorption from and metastable , making it a member of the `` felobal '' class .
the number of known felobals has recently grown dramatically from a single example to more than a dozen , including a gravitationally lensed example and the binary member presented here , suggesting that this formerly rare object may be fairly common .
additionally , the presence of this bal quasar in a relatively small separation binary adds to the growing evidence that the bal phenomenon is not due to viewing a normal quasar at a specific orientation , but rather that it is an evolutionary phase in the life of many , if not all , quasars , and is particularly associated with conditions found in interacting systems .
# 1 # 1 |
garcia - lario , manchado , sahu , and pottasch ( 1993 , hereafter gmsp ) present the first detailed analysis of iras 06562@xmath00337 .
they argue that it is a proto - planetary nebula ( ppn ) undergoing final mass - loss episodes .
their time - series of optical spectra , obtained over a 5 year period , show the onset of forbidden line emission and the possible evolution of the central star toward hotter temperatures .
they derive a zanstra temperature of 2@xmath410@xmath5 k , with a slight increase over a two year interval .
the effective temperature of the exciting star , t@xmath6 @xmath7 3.6@xmath410@xmath5 k , also showed a slight increase in two years .
the h@xmath8 line profile changes in time , which gmsp interpret as variable high velocity winds associated with episodic mass - loss .
the appearance of [ ] emission lines in 1990 and the resulting 4363/(4959 + 5007 ) line ratio requires an ionizing region of high electron density , log(@xmath9 ) @xmath7 6.9 .
the absence of these lines in spectra obtained before and after 1990 is interpreted as collisional de - excitation due to changing densities in the ionized region effected by violent episodic mass - loss . from co observations
gmsp derive @xmath10 = 50 @xmath11 1 km sec@xmath12 , which agrees with the velocity derived from their high resolution optical spectra . adopting a model galactic rotation curve
, they estimate a distance of 4 kpc , which compares to a distance of 2.4 kpc estimated from the equivalent width of na d absorption seen in their spectra .
the iras colors fit with blackbodies show a trend of decreasing temperature with increasing wavelength which implies a gradient of dust temperatures .
gmsp integrated the optical ir spectral energy distribution of iras 06562@xmath00337 , yielding a luminosity of l = 7000 @xmath13 for their preferred distance of 4 kpc .
kerber , lercher , and roth ( 1996 , hereafter klr ) describe an additional medium resolution , high s / n spectrum of iras 06562@xmath00337 obtained in early 1996 .
[ ] emission is still absent , but a wealth of and [ ] lines are found .
these lines confirm the high electron density derived by gmsp from the [ ] lines present in 1990 .
klr argue that the spectrum also implies a considerable density gradient in the object , as [ ] lines are collisionally suppressed at densities where lines exist .
they maintain the classification of iras 06562@xmath00337 as a candidate ppn , designating it `` the iron clad nebula '' .
bachiller , gutierrez , and garcia - lario ( 1998 , hereafter bgg ) present new mm and sub - mm observations of iras 06562@xmath00337 .
they derive @xmath10 = 54.0 @xmath11 0.2 km sec@xmath12 and adopt a different model galactic rotation curve than gmsp to estimate a distance of 7 kpc .
this distance yields a luminosity of 21000 @xmath13 and a cloud mass m @xmath14 1000 @xmath15 . from the strength of the co emission and the presence of cs emission , bgg surmise that iras 06562@xmath00337 is a `` young stellar object ( or small cluster ) still associated to its parent molecular cloud . ''
bgg point out that iras 06562@xmath00337 satisfies the three criteria for a herbig ae / be star ( herbig , 1960 ) and the spectral energy distribution , which rises sharply in the far infrared ( gmsp ) , is similar to group ii herbig ae / be stars ( hillenbrand , 1992 )
. they also note the presence of blue and redshifted wings in the co emission indicating a bipolar outflow .
the co outflow may be driven by an eruptive ionized jet , which leads them to suggest the sporadic [ ] emission seen by gmsp originates in a herbig - haro object .
we present the first 2@xmath1 image of iras 06562@xmath00337 .
our image reveals a compact cluster of stars surrounding a bright , central object .
we independently confirm the result also discovered by bgg that iras 06562@xmath00337 is a young stellar object . in section 2 of this paper , we describe our near - infrared observations and stellar census of the iras 06562@xmath00337 cluster .
we also compare the co(2@xmath161 ) map of bgg with our image . in section 3
, we describe our new spectroscopic observations and summarize the resulting 1997 epoch emission line data .
we make a detailed comparison with the 1996 epoch emission line data of klr . in section 4
, we present our conclusions .
on 1997 march 26 ut , we observed iras 06562@xmath00337 with a k@xmath17 ( 1.95 to 2.35 @xmath1 ) filter in non - photometric conditions using the lick observatory 3 m telescope and the lirc ii mercury - cadmium - telluride 256@xmath4256 pixel camera ( misch , gilmore , and rank 1995 ) .
the wide field - of - view optical configuration was utilized to yield a pixel size of 0.57@xmath2 and a full image area covering [email protected]@xmath19 .
we implemented a four - point , on - source dithering pattern to obtain 5@xmath42-second exposures at each position ; this pattern was repeated 10 times .
our cumulative exposure was 400 seconds .
evening twilight sky flat and morning dark calibration frames were obtained on the same night .
all reductions were done with iraf .
the 200 object images were dark subtracted and flat corrected ( bad pixels were also masked ) , then individually sky subtracted , registered , and combined .
the field of view exposed for the complete 400 seconds was 108@xmath2 @xmath4 108@xmath2 , and the final image was trimmed to this size .
a log - scaled greyscale image of iras 06562@xmath00337 with a field of view of 70@xmath2 @xmath4 70@xmath2 is presented in figure 1 .
the image reveals a small , dense cluster of stars around a bright , central object .
the association of the variable emission - line central object with a cluster of stars lends strong support to its classification as a young stellar object and not a ppn .
also evident is a diffuse nebulosity extending approximately 30@xmath2 , whose brightness increases toward the central object .
this nebulosity may be reflected light from the central object or the unresolved light from numerous low - mass stars in the cluster .
the morphology of the faint nebulosity is somewhat affected by our choice of an on - source dithering pattern and the resulting sky subtraction . in particular , directly to the east of the central object , the apparent decrease in the brightness of the nebulosity
is an artifact of the sky subtraction .
the log - scaling emphasizes this low level data reduction artifact , which does not affect the results of this paper .
an improved observing strategy for iras 06562@xmath00337 would require chopping well away from the cluster to obtain sky images .
our final image was photometered with the stand - alone daophotii / allstarii ( stetson , 1987 ) .
an empirical psf was derived from 7 bright , fairly isolated stars with profile errors of order 3% .
we obtained satisfactory star subtraction and psf fitting photometry for all cluster stars except the central object .
the seeing in our image measured with the psf stars is 1.3@xmath2 ( fwhm ) , while the central object has a fwhm of 1.5@xmath2 .
thus we may be just resolving the central object .
as conditions were not photometric , we can only crudely calibrate our photometry to k@xmath20 .
manchado et al .
( 1989 ) give k = 9.15 @xmath11 0.02 mag for iras 06562@xmath00337 obtained with a photometer and a beam width of 15@xmath2 .
this k magnitude is on the tiede photometric system which is equivalent to k@xmath20 ( arribus and martinez , 1987 ) .
we derived an aperture correction using a 15@xmath2 beam centered on the central object and the reported allstarii magnitude .
given the slightly non - stellar profile of the central object , this procedure yields a rather large uncertainty of @xmath11 0.2 mag for the zero - point transformation from our instrumental system to k@xmath20 . with the exception of the central object ,
our relative photometry of the cluster stars is more accurate .
typical allstarii reported errors are k = 12.5 @xmath11 0.03 mag , k = 16.5 @xmath11 0.1 mag , and k = 18 @xmath11 0.5 mag ( see fig . 2 ) .
artificial star tests were performed with daophotii in 10 trials of adding 50 stars to our k@xmath17 image , randomly distributed across the image and with magnitudes from k = 14 to 20 .
we find our completeness , defined as the ratio of stars recovered to stars added , is dependent on both magnitude and position .
considering artificial stars added at all positions in the image , we are approximately 95% complete at magnitudes brighter than k = 15 , 90% complete at k = 16 , and 50% complete at k = 17 . at k = 18
we recovered almost no added stars .
we find that within a radius of @xmath7 5@xmath2 from the central object , we recovered no stars .
considering only artificial stars with k @xmath21 17 , we find our completeness to increase from approximately 85% around 10@xmath2 from the central object to nearly 100% at a radial position 60@xmath2 from the central object .
this radial dependence of our completeness is attributable to both crowding and the irregular sky background contribution of the nebulosity .
our current near - infrared data is not ideal for a stellar census of iras 06562@xmath00337 .
we primarily require a wider field of view ( e.g. a mosaic image ) to better characterize the foreground ( and background ) stars along this line of sight through the galactic plane . nevertheless , as a first accounting of the cluster members we adopt a radius of 30@xmath2 as the cluster boundary , which is roughly the extent of the nebulosity seen in our k@xmath17 image .
inside this radius we find 71 stars .
two equal area control fields , rectangular regions at the outer - most edges of our final image , contain 9 and 17 stars , predicting an average of only 13 background / foreground stars will be found inside our 30@xmath2 radius cluster boundary .
we conclude that there is a significant stellar surface density enhancement in the immediate vicinity of iras 06562@xmath00337 .
this confirms what may seem obvious from a casual inspection of figure 1 , that we have discovered a new star cluster .
in figure 2 , the top panel , we present the k luminosity function of stars found inside our 30@xmath2 radius cluster boundary .
the cluster luminosity function is shown as the unshaded solid line histogram .
the averaged luminosity function of our two control fields is shown as the shaded histogram .
these two luminosity functions sample the same area of sky . except for small differences in photometric completeness near the cluster center , they are properly comparable . a k - s test on the cluster and background luminosity functions gives a 1% probability that these are drawn from the same distribution .
corrections for completeness near the cluster center would result in even lower probabilities .
thus , in addition to the clustering in 2 dimensions on the sky , we find the k magnitude distribution of stars centered on iras 06562@xmath00337 is not likely to be a statistical fluctuation in starcounts along this line - of - sight through the galactic plane .
in the bottom panel of figure 2 , we present the cluster luminosity function corrected for foreground / background stars and completeness .
this luminosity function has 80 stars with k @xmath21 17.5 mag .
we have additionally overplotted theoretical luminosity functions corresponding to different initial mass functions ( imfs ) , parameterized by a power - law index @xmath8 .
our fiducial imf is that of salpeter ( 1955 ; @xmath8 = 2.35 ) , plotted as a solid line for k @xmath21 17.5 mag and a long - dashed line for k @xmath14 17.5 mag . a shallower and steeper imf ( @xmath8 = 1.35 and 3.35 , respectively ) are also shown , as dotted lines .
all three of the theoretical luminosity functions are normalized to have 80 stars for k @xmath21 17.5 mag .
we note that the steep imf is the best fit , although without color information and comparisons to pre - main sequence isochrones , this analysis is susceptible to systematic biases .
for instance , pre - main sequence stars will generally appear brighter at k than main sequence stars , biasing the imf slope to higher values .
we emphasize that our simple model luminosity functions demonstrate only a plausible , order - of - magnitude consistency between the observed cluster luminosity function and that expected from a salpeter imf .
nevertheless , this rough agreement supports the principal result of our paper , that iras 06562@xmath00337 is a coeval young star cluster .
the theoretical luminosity functions allow us to estimate the k flux for an unresolved population of low - mass cluster stars , perhaps seen as the diffuse nebulosity in our image .
the model luminosity functions , in this instance , can be regarded as empirical extensions of the observed luminosity function , and assigned no greater physical significance . masking the central star ( using iraf ) in an image with all other identified stars subtracted ( using allstarii )
, we find k@xmath22 @xmath7 11.3 mag inside a 30 @xmath2 beam .
the measured flux of the nebulosity is quite sensitive to the adopted procedure of masking the central star . from several trials we estimate an uncertainty of roughly @xmath11 0.5 mag .
excluding the central star ( k = 9.2 ) the integrated magnitude of the stars in the histogram of the bottom panel of figure 2 is k = 11.0 mag ; including the central star we find k = 9.0 mag . integrating the salpeter imf luminosity function beyond the last reliable bin in the completeness
corrected luminosity function , from k = 17.5 to 23 , gives k@xmath23 = 12.3 mag . integrating the steep imf from k = 17.5 to 23 gives
k@xmath23 = 11.5 mag .
we find the differences in k@xmath23 to be small when integrating to faint magnitude limits , changing only 0.2 mag when integrating to k = 21 or 23 . both of the predicted nebular fluxes are fainter than the measured flux , albeit with somewhat large uncertainties .
therefore , we might expect some contribution of reflected light from the central object to the nebulosity .
importantly , there is no gross inconsistency in our calculations , such as a prediction of too much diffuse light from faint stars .
we conclude that our theoretical fit to the k luminosity function , extended to low mass stars ( m @xmath7 0.3 @xmath15 ) , predicts an unresolved flux that is consistent with the brightness of the nebulosity observed in our k@xmath17 image .
we adopt @xmath24 for stars in the iras 06562@xmath00337 cluster .
we adopt @xmath25 from mihalas and binney ( 1981 , p. 113 ) . ] .
the zero - point is estimated as follows . from carpenter et al .
( 1997 , see their fig .
6 ) , who compared near - ir data of the young star cluster mon r2 to the pre - main sequence stellar evolution tracks of dantona and mazzitelli ( 1994 ) , we estimate an unreddened @xmath26 star would have k = 10.2 mag . assuming a relative distance modulus between iras 06562@xmath00337 and mon r2 of 5 mag , and an iras 06562@xmath00337 extinction of @xmath27 = 1 mag , yields the 19.4 mag zero - point .
our k mag / stellar mass calibration equates the bright central object , k = 9.2 mag , with an @xmath28 star .
our mass estimate for the central star is consistent with an early b or late o spectral type ( mihalas and binney , 1981 ) .
bgg estimate a b0-b2 spectral type from the luminosity , 21000 @xmath13 . for comparison ,
gmsp calculate t@xmath6 @xmath7 3.6@xmath410@xmath5 k for the central star , which implies a late o spectral type ( mihalas and binney , 1981 ) .
using the salpeter imf theoretical luminosity function to extend our observed luminosity function , we predict the following total numbers of cluster stars : 225 , 445 , 875 , 1720 , 3380 to the limiting magnitudes k = 19 , 20 , 21 , 22 , and 23 respectively . the total mass in the observed luminosity function is estimated by integrating the salpeter imf theoretical luminosity function to k = 17.5 which gives @xmath29 @xmath7 225 @xmath15 . if we integrate to k = 23 , we find @xmath29 @xmath7 1950 @xmath15 , which compares with the lower limit of molecular gas mass given by bgg of @xmath30 .
if a 50% mass - efficiency for star formation is adopted , we might expect the total mass of molecular gas in the cluster to be @xmath31 @xmath7 4000 @xmath15 .
such a comparison is necessarily crude , particularly given our assumptions deriving the k mag / stellar mass calibration and extending a salpeter imf to the extreme lower main sequence and pre - main sequence .
iras 06562@xmath00337 appears to be a large and rich young cluster .
similar young star clusters with massive ae / be central stars have been the subject of two recent surveys .
hillenbrand ( 1995 ) investigate the inner 0.3 pc of 17 clusters and find a linear relationship between central star mass and stellar surface density ; testi et al .
( 1997 ) study fields around 19 herbig ae / be stars and report a correlation between spectral type and cluster richness , with earlier spectral types resulting in significantly more cluster stars , and some evidence of a threshold effect around spectral type b5b7 . similar to hillenbrand , they find a characteristic cluster radius to be 0.2 pc . at a distance of 7 kpc ( bgg ) ,
our 30@xmath2 radius corresponds to a physical radius of 1 pc .
this is @xmath325 times the `` typical '' cluster radius , implying iras 06562@xmath00337 is a very large cluster .
the smallest distance estimate ( 2.4 kpc , gmsp ) results in an adopted cluster radius of 0.34 pc , still @xmath33 times the `` typical '' value .
barsony et al .
( 1991 ) found 100 cluster members within @xmath34 pc in the young star cluster lkh@xmath8 101 , thus iras 06562@xmath00337 is not unprecedented .
we find an average stellar surface density of @xmath3225/pc@xmath35 within 30@xmath2 , increasing towards smaller radii .
testi et al .
define a richness indicator , i@xmath36 , which measures the stellar density enhancement of a cluster , independent of cluster size , using k band starcounts .
they find early type stars have i@xmath37 , and mwc 137 ( a b0 star ) has the highest value in their sample , of i@xmath38 .
we estimate a value of i@xmath36 = 103 for iras 06562@xmath00337 .
the high apparent mass of the central star is therefore consistent with the large size and richness of the iras 06562@xmath00337 cluster .
we determined the coordinates of the central star by comparing the positions of 12 stars in our image with the digitized sky survey image .
the centroid of the central star at k@xmath17 is offset slightly to the west of the visible image by @xmath390.6@xmath2 .
taking this into account , we found the two images to be in excellent agreement , with rms scatter of less than 0.5@xmath2 .
therefore , we report the j2000 k@xmath17 ( 2@xmath1 ) position of the central object to be @xmath8 = 06:58:44.31 , @xmath40 = @xmath003:41:09.97 ( with uncertainties of @xmath41 ) .
this differs by @xmath425.7@xmath2 , @xmath431.2@xmath2 from the near infrared position given by manchado et al .
( 1989 ) .
figure 3 presents the co(2@xmath161 ) map of iras 06562@xmath00337 , graciously provided by bgg , overlaid on our k@xmath17 image .
the co data was obtained in 1993 august and 1997 may with the iram 30-m telescope with a beam size of 12@xmath2 and an efficiency of 0.45 at 230 ghz .
there is excellent agreement between the central peak of the co emission and the central object in the k@xmath17 image , even better than the agreement between the iras position and the co emission pointed out by bgg .
the peak in the co emission and the central star are offset from each other by @xmath440.6@xmath2 , @xmath453.4@xmath2 .
the central contour of the co plot is slightly elongated , extending 8.4@xmath2 e w and 11.5@xmath2 n s , which may imply a non - spherical molecular gas distribution .
the majority of the emission ( the first 12 contours ) falls within 30@xmath2 of the central star , lending support to our choice of 30@xmath2 for the cluster radius , although strong emission ( @xmath14 5 k ) extends over a region of @xmath46 radius .
the singly - peaked emission suggests that the central object in the k@xmath17 image is the only massive young star in the cluster .
the coincidence of the cluster stars and nebulosity seen in our k@xmath17 image with the strong co emission is strong evidence that the cluster is embedded in its placental molecular cloud ( see also discussion in bgg ) .
the morphology of the molecular cloud around iras 06562@xmath00337 sets it apart from other young clusters .
figure 3 clearly shows the co distribution to be roughly spherically symmetric and well centered on the massive central star .
studies of herbig ae / be clusters still associated with remnant molecular gas ( hillenbrand 1995 , barsony et al .
1991 ) do not generally find this to be the case .
in fact , although the sample size is small , in almost every case the star appears to be located near the edge of the cloud , which is often elongated or otherwise shows signs of disruption
. this may suggest interaction between the star and the cloud , such that the energy from the star is dispersing or disrupting the cloud ( hillenbrand 1995 ) .
we speculate that this implies extreme youth in the case of iras 06562@xmath00337 , where the parent molecular cloud is still in control , keeping the hot , young star snugly cocooned inside its dense layers .
on 1997 april 16 ut , we acquired 3 spectra of iras 06562@xmath00337 from the kast spectrograph ( miller & stone 1993 ) on the 3 m telescope at lick observatory .
spectrum # 1 was exposed on the blue side of kast for 900 seconds using the 600 line mm@xmath12 grism .
spectrum # 2 was exposed concurrently on the red side of kast for 900 seconds using the 830 line mm@xmath12 grating .
spectrum # 3 was exposed on the red side of kast for 600 seconds , approximately 35 minutes after the other spectra , using the 600 line mm@xmath12 grating .
all three spectra were taken through a 2 arcsec slit .
the spectra were reduced and extracted using standard iraf routines ( massey , valdes , & barnes 1992 ) .
they were wavelength - calibrated using a hehgcd arc lamp spectrum for the blue spectrum ( # 1 ) , and a henear arc lamp spectrum for the red spectra ( # s 2 and 3 ) .
fine corrections to the zero points of the wavelength scales were made via inspection of the night sky lines .
the resultant wavelength coverage and resolution ( mean fwhm of lines in the corresponding arc spectrum ) for each spectrum are 30955206 ( resolution 3.2 ) for spectrum # 1 , 60288061 ( resolution 4.2 ) for spectrum # 2 , and 42747059 ( resolution 5.5 ) for spectrum # 3 .
the instrumental response function was removed from the spectra and they were flux - calibrated using observations of spectrophotometric standard stars obtained on the same night with the same grating settings ; feige 34 ( massey et al . 1988 ) was used for spectrum # 1 and hd 84937 ( oke & gunn 1983 ) for spectra # s 2 and 3 . a composite spectrum of iras 06562@xmath00337 is shown in figure 4 .
it combines the data from 31005200 in spectrum # 1 , 52007050 in spectrum # 3 , and 70508050 in spectrum # 2 .
the bottom panel of figure 4 shows the full wavelength and flux scales of the composite spectrum . the most prominent feature is a very strong emission line of h@xmath8 that dwarfs any other spectral features .
the continuum level increases to the red .
a power law fit to the continuum , @xmath47 , gives an index of @xmath48 . in the middle panel of figure 4 , we have reduced the range of the flux axis so that the weaker features are more apparent .
the first three balmer lines and several terrestrial atmospheric features are indicated .
we were able to identify balmer lines all the way to h11 ( @xmath49 ) ; blueward of h@xmath40 , the balmer lines are in absorption rather than emission .
although the spectrum becomes somewhat noisy at its blue end , a small central emission component ( peaking well below the local continuum level ) is apparent in the bottom of each of the h@xmath50h11 absorption lines .
this suggests that the balmer profiles in iras 06562@xmath00337 are made up of at least two components : an absorption line and a superposed emission line .
the absorption depth does not change greatly from h@xmath50 to h11 , but the intensity of the emission component increases up the balmer series such that the lines switch from mostly absorption at h@xmath50 to mostly emission at h@xmath40 .
this is qualitatively similar to the balmer emission observed in other be stars ( burbidge and burbidge , 1953 ) . in the top panel of figure 4
the wavelength and flux scales of the composite spectrum are expanded , with several typical features labeled .
in addition to the balmer lines and diffuse interstellar bands ( dibs ; herbig 1975 ) , we find several absorption lines of and many ( permitted and forbidden ) emission lines of metals .
most of the latter are lines of and [ ] ; the remainder are and [ ] .
there are a few unidentified and/or extremely weak features in the composite spectrum ; hence , we can not rule out the possible presence of lines of other elements or additional lines of the identified elements . tables 1a , 1b , and 1c list the equivalent widths ( ews ) and integrated fluxes for the identifiable lines and dibs in each of the spectra .
the ews and fluxes were measured via direct integration of pixel values between manually - selected endpoints in the flux - calibrated spectra .
several trials of each measurement were accomplished to estimate the uncertainty ; in general , the measurements are accurate to @xmath51 or better ( values with larger probable uncertainties are denoted with a `` : '' ) .
a negative flux value in the tables indicates an absorption line .
the 1996 ( klr ) and 1997 ( this paper ) epoch spectra have similar resolution and high s / n , which allows the first investigation of the variability of many fainter lines . in table 2 we have assembled select emission line strengths from our tables 1a , b , c and klr ( 1996 , their table 1 ) to make the comparison .
lines are included in table 2 if they are relatively unblended in our 1997 epoch spectra ( typical uncertainties of 10% in the line fluxes ) and were also identified in the klr spectrum . in some instances , particularly useful lines , such as the [ ] lines ,
were included despite their larger flux uncertainties . in these cases , we mantain the use of the `` : '' as in tables 1a , b , c to designate the larger uncertainty . in other cases , the flux of two blended lines of the same ionization species are listed for 1997 , and the combined flux of the two resolved lines in the klr spectrum is listed for 1996 .
klr do not report absolute fluxes , so we have normalized the fluxes to h@xmath52 = 100 . in the bottom part of table 2
, we have listed several line ratios ; typical uncertainties are 14% unless otherwise discussed below .
klr suggest that the variable emission seen by gmsp may be understood as a patchy circumstellar envelope around a single star . using the balmer decrements
, we investigate the hypothesis that variable extinction , possibly due to orbiting dust clouds , could mimic variable emission between 1996 and 1997 .
we estimate the uncertainties in our flux measurements for the h@xmath8 , h@xmath52 and h@xmath53 balmer lines is @xmath7 5% , or 7% in the ratios .
klr do not report uncertainties ; however , we will assume they are 10% , or 14% in the ratios . to derive the extinction
, we adopt the intrinsic decrements h@xmath8/h@xmath52 = 2.80 and h@xmath53/h@xmath52 = 0.47 from osterbrock s ( 1974 ) compilation of case b recombination line ratios , appropriate for a temperature near 2@xmath410@xmath5 k ( gmsp ) , and the reddening law of cardelli , clayton , and mathis ( 1987 ) . for 1997
we find @xmath54 = 8.15 @xmath11 0.25 and @xmath54 = 8.6 @xmath11 0.4 from h@xmath8/h@xmath52 and h@xmath53/h@xmath52 respectively .
similarly , for 1996 we find @xmath54 = 8.35 @xmath11 0.4 and @xmath54 = 8.2 @xmath11 0.9 .
the weighted average of all four measurements gives @xmath54 = 8.3 @xmath11 0.35 mag , or a color excess of @xmath55 = 2.68 @xmath11 0.12 mag .
note that our estimate of the uncertainty does not include a contribution from the reddening law or the adopted intrinsic line ratios .
the color excesses derived in 1996 and 1997 agree well , which implies a negligible change in extinction over the one year interval .
our color excess can be compared with that derived by gmsp of @xmath55 = 1.75 @xmath11 0.25 mag .
the discrepancy is significant , and may be due to different assumptions in the reddening law or intrinsic line ratios or a real change in the extinction between 1990 and 1996/1997 .
gmsp do not give observed line fluxes , only dereddened fluxes , and do not describe their procedure for calculating the reddening in detail .
for these reasons , we do not make any additional comparisons to the gmsp spectral line data .
lastly , the cardelli , clayton , and mathis reddening law gives @xmath56 = 0.114@xmath4@xmath54 , or @xmath56 = 0.95 mag for iras 06562@xmath00337 , which is what we assumed in section 2 .
keenan et al . ( 1995 ) present an [ ] ( 6300 + 6364/5577 ) temperature and density diagnostic diagram , as do bautista and pradhan ( 1995 ) . in our data ,
the @xmath57 and @xmath58 lines were both blended and have large uncertainties , estimated to be approximately 30% .
the [ ] ( 6300 + 6364/5577 ) line ratio ( see table 2 ) is consistent with no change over the 1996 to 1997 interval , although this result is fairly uncertain .
it is noteworthy that the unblended and strong [ ] @xmath59 line intensity decreased by a factor of two over this year ( a high significance ) , yet the [ ] line ratio remained constant .
we derive log(@xmath9 ) @xmath7 5.8 assuming an electron temperature , @xmath60 = 2@xmath410@xmath5 k ( gmsp ) .
we additionally derive an upper limit to the electron density , log(@xmath9 ) @xmath61 15.45 , from the inglis - teller formula ( 1939 ) and our observation of the h11 balmer line .
our use of the inglis - teller formula is limited by the noise at the blue end of our spectra . for comparison ,
viotto ( 1976 ) tabulated electron densities and temperatures for a number of be stars .
they are typically @xmath62 @xmath63 and @xmath64 1 - 2@xmath65 k , with earlier spectral types having the higher electron temperatures .
the ( low ) electron density derived from the [ ] line ratio implies an origin for these lines at the `` outskirts '' of the ionized region .
this may reflect a partially ionized zone at the edge of a wind - swept ionized cavity ( e.g. barsony et al .
1990 , becker and white , 1988 ) or the `` outer wind '' analagous to that seen in p cygni ( stahl et al . 1991 ) .
gmsp argued that their time - series of spectral data imply an increasing temperature , consistent with a rapidly evolving central star in a ppn .
we consider whether the 19961997 spectral line variability can be understood as simple changes in electron temperature or density .
the permitted lines of and show a general trend of decreasing intensity over the 1996 to 1997 interval . of five lines selected for comparison , one line ( @xmath66 ) shows an @xmath7 1@xmath67 increase , while four other lines show 1 - 2@xmath67 decreases .
the average decrease in line intensity is 80% . if the permitted lines of and behave similarly to the hydrogen and helium recombination lines , then their intensities depend very weakly on electron density ( e.g. collisional effects , osterbrock 1974 ) , and with a temperature dependence , @xmath68 ( oliva et al . , 1989 ) .
this implies an increase in the electron temperature of approximately 20% in one year . assuming an initial electron temperature of @xmath69 k
, the increase would be @xmath70@xmath71 k. the most dramatically varying line in table 2 is [ ] @xmath59 , which decreased by a factor of two ( approximately 5@xmath67 ) between 1996 and 1997 .
the electron temperature and density dependence of this forbidden line is given by : @xmath72 where @xmath73 = c2/@xmath74 , c2 is the second radiation constant ( 1.438769 cm k ) and @xmath74 is in units of cm ; for [ ] @xmath59 , t@xmath75 = 22837 .
@xmath76(t ) is the collision strength which we will assume has a weak temperature dependence ( osterbrock 1974 ) , @xmath9 is the electron density , and @xmath77 is the density of the ion species , in this case , neutral oxygen .
a change in the electron density of 0.2 dex is reasonably allowed by the uncertainty of our [ ] line ratio .
therefore , we can not rule out a factor of two decrease in the electron density as the cause of the factor of two decrease in the [ ] @xmath59 line intensity . if the flux variability is entirely due to a change in temperature , we find @xmath40@xmath78 k , with an estimated uncertainty of about 5000 k , consistent with the temperature change implied by the permitted lines .
however , an apparent temperature increase of @xmath325000 k in one year seems unlikely to be related to stellar evolution ( the suggestion of gmsp , klr ) , and rather implies more complex physical processes .
the intensity of all lines of [ ] listed in table 2 increased between 1996 and 1997 and appear anti - correlated with the intensities of the [ ] lines , which decreased in the same interval .
typically , a positive correlation is found between [ ] and [ ] emission strength in a wide range of objects including supernova remnants , seyfert galaxies , and regions such as orion ( bautista and pradhan , 1997 , 1995 ; bautista , pradhan and osterbrock 1994 ; mouri et al .
1990 ) . in these objects ,
[ ] emission is believed to originate in partially ionized zones ( pizs ) , lying at the edges of fully ionized zones .
the [ ] / [ ] correlation is attributed to the filling factor of the pizs and the coincidence of the [ ] and [ ] species in the pizs . in iras
06562@xmath00337 , a simple change in the electron temperature or density will not explain the behavior of both the [ ] and [ ] lines .
it is conceivable that dust grains are being blown away and iron depleted onto the grains is being returned to a gas state , enough to offset the effect of an increasing temperature .
an alternative , and preferred , explanation is that the [ ] and [ ] emission are not originating in the same zone , which argues against a presumably static shell - like piz at the edge of a wind - swept cavity and favors a more dynamic `` outer wind '' origin for these forbidden lines .
lastly , we note that the suggestion of klr that the [ ] emission is due to flourescence is unlikely at densities of log(@xmath9 ) @xmath7 5.8 , as shown by bautista , peng , and pradhan ( 1996 ) .
the bright central object in iras 06562@xmath00337 is probably a single herbig ae / be class star , with @xmath3 .
there is some evidence that it is seen at least partially in reflection .
for example , the central star is poorly fit by the stellar point spread function derived from other stars in our k@xmath17 image .
in addition , the centroids in the visible and at k@xmath17 are offset by @xmath390.6@xmath2 .
a comparison of our 1997 epoch optical spectra to the 1996 epoch spectrum presented by klr has provided new evidence for spectral line variability in iras 06562@xmath00337 that is not easily interpreted as simple changes in temperature or density .
this argues against previous suggestions that the already - published spectral data reflect real - time evolutionary changes of the central star .
understanding the immediate environs of the central star is an essential first step to better understanding the emission line variability that has been observed .
we note that our analysis of the 19961997 spectral variability of iras 06562@xmath00337 using simple zones and changes in one parameter ( such as the electron temperature or electron density ) is only intended as an interpretive tool and should be treated with some caution .
the inconsistencies of this analysis suggest the complexity of the actual emission line region .
if the central star is analagous to a herbig be star , then its `` fully ionized zone '' should have a very high electron density .
thus , the low density derived with the [ ] emission implies a stratification of densities , likely in the extended atmosphere of the central star in iras 06562@xmath00337 .
the anti - correlation of [ ] and [ ] emission is not easily understood if they arise from a common partially ionized zone ( perhaps located at the edge of a compact region around the central star ) , but is likely consistent with an origin in a dynamic stratified atmosphere .
spectral variability is also consistent with a stellar wind origin .
stellar winds from herbig be type stars are highly variable , extremely complex , and poorly understood ( e.g. , they are possibly modulated by magnetic fields or non - radial pulsations ; catala et al .
wind line strengths and profiles can show nightly variations , some of which are correlated with temperature changes of the star ( scuderi et al .
the appearance of [ ] emission in iras 06562@xmath00337 during 1990 ( gmsp ) can be understood as shocks in stellar jets from a variable stellar wind ( hartigan & raymond 1993 ) .
additionally , we emphasize that the size of the ionized region in iras 06562@xmath00337 is small .
six cm radio observations yielded a surprising non - detection given the strong h@xmath52 flux , and allow an upper limit angular diameter of @xmath79 arcsec to be set for the ionized region ( gmsp ) .
if we assume the 7 kpc distance estimate of bgg , then the ionized region in iras 06562@xmath00337 has an upper limit diameter of 23 au .
we conclude that stellar winds in the extended , stratified atmosphere of a very young , @xmath3 star is the most plausible explanation for the puzzling spectral variability of iras 06562@xmath00337 , the `` iron clad nebula . ''
a valuable time series of spectra of the central object in iras 06562@xmath00337 , covering a 10 year interval , are now available in the literature .
a useful complementary study would obtain more densely sampled spectroscopic observations , to investigate the short timescale behavior of the emission .
in particular , correlated line variability and characteristic timescales could be used to constrain the dimensions of the ionized region , which we have suggested is likely to be an extended stellar atmosphere .
we emphasize the need for high resolution and high s / n spectroscopic observations in the near - uv , and of the entire balmer series , as these data can be used to study the extended atmospheres of be stars ( burbidge & burbidge 1953 ) .
it is important to build a consistent physical picture of the environs around the central star in order to answer questions such as `` is the central star seen in reflection ? ''
this problem in particular could be addressed with high resolution imaging and polarization data .
lastly , as a young star cluster , iras 06562@xmath00337 is interesting and notable for its richness . multi - color near - infrared photometric observations would allow an accurate measurement of the imf , and an investigation into problems such as mass segregation and star formation efficiencies ( e.g. , barsony et al .
alves research at a doe facility is supported by an associated western universities laboratory graduate fellowship .
work at llnl performed under the auspices of the usdoe contract no .
hoard s research supported by nsf grant no .
ast 9217911 .
alves and hoard thank our respective advisors , kem cook and paula szkody .
rodgers is supported by the nasa graduate student research program , and thanks her advisors bruce balick and diane wooden .
alves acknowledges thor vandehei for his exuberant assistance while observing at the lick observatory .
we also thank bob becker , hien tran , carlos de breuck , and bill latter for their useful comments and discussions that significantly improved this paper .
lccl 3774 & 1.8 & @xmath01.6 & @xmath49 ( h11 ) 3797 & 1.3 & @xmath01.2 & @xmath80 ( h10 ) 3812 & 1.1 & 1.0 & @xmath81 3837 & 2.0 & @xmath02.0 & @xmath82 ( h9 ) 3856 & 1.4 & 1.4 & @xmath83 3892 & 2.1 & @xmath02.1 & @xmath84 , @xmath84 ( h8)3916 & 0.9 : & 0.9 : & @xmath85 3973 & 3.0 & @xmath03.0 & @xmath86 ( h@xmath50)4100 & 1.0 & 1.1 & @xmath87 ( h@xmath40 ) 4170 & 0.8 : & 1.0 : & @xmath88 4245 & 0.9 : & 1.1 : & [ ] @xmath89 4340 & 2.8 & 3.7 & @xmath90 ( h@xmath53 ) 4390 & 0.7 : & @xmath01.0 : & @xmath91 4415 & 0.9 & 1.2 & [ ] @xmath92 , @xmath93 4430 & 2.1 & @xmath02.9 & dib 44284472 & 0.4 : & @xmath00.6 : & @xmath94 4494 & 0.4 : & 0.6 : & @xmath95 4503 & 0.2 : & @xmath00.3 : & dib 45024584 & 0.5 & 0.7 & @xmath96 4716 & 0.2 : & @xmath00.4 : & @xmath97 4731 & 0.3 : & 0.4 : & @xmath98 4817 & 0.6 & 1.0 & @xmath99 , [ ] 4863 & 18.1 & 31.7 & @xmath100 ( h@xmath52 ) 4883 & 0.6 & @xmath01.1 & dib 48824891 & 0.4 : & 0.7 : & @xmath101 4924 & 0.6 & 1.0 & @xmath102 5020 & 1.0 & 2.0 & @xmath103 5044 & 0.3 & 0.7 & @xmath104 5058 & 0.5 : & 1.1 : & @xmath105 5151 & 0.3 : & 0.6 : & @xmath106 5160 & 1.6 : & 3.6 : & [ ] @xmath1075169 & 1.6 : & 3.6 : & @xmath108 lccl 4436 & 3.4 : & @xmath04.7 : & dib 44284475 & 1.4 : & @xmath01.9 : & @xmath94 4820 & 1.3 : & 2.1 : & @xmath99 , [ ] 4868 & 19.0 & 32.0 & @xmath100 ( h@xmath52 ) 5026 & 1.0 & 1.9 & @xmath66 5164 & 1.6 : & 3.5 : & [ ] @xmath1075174 & 1.6 : & 3.5 : & @xmath108 5204 & 0.4 : & 0.9 : & @xmath109 5267 & 1.7 : & 4.0 : & [ ] @xmath110 5277 & 1.7 : & 4.0 : & [ ] @xmath111 5337 & 0.8 & 2.0 & [ ] 5380 & 0.5 & 1.3 & [ ] 5579 & 0.2 : & 0.6 : & [ ] @xmath58 5706 & 0.5 & @xmath02.0 & dib 57055750 & 0.6 : & 1.7 : & [ ] @xmath112 5754 & 0.6 : & 1.7 : & [ ] @xmath113 5780 & 1.3 & @xmath04.0 & dib 5778,57805797 & 0.3 : & @xmath00.8 : & dib 57975870 & 0.8 & 2.5 & @xmath114 5891 & 2.6 : & @xmath08.4 : & @xmath115 ( d - lines ) 5897 & 2.6 : & @xmath08.4 : & @xmath116 ( d - lines ) 5951 & 0.6 : & 2.1 : & @xmath117 5959 & 0.6 : & 2.1 : & @xmath118 5980 & 0.6 & 1.9 & @xmath119 6048 & 0.4 : & 1.4 : & [ ] 6126 & 0.4 : & 1.5 : & @xmath120 6284 & 3.2 & @xmath011.7 & dib 62846302 & 1.1 & 4.1 & [ ] @xmath59 6320 & 0.6 & 2.4 & 6348 & 0.6 & 2.2 & @xmath121 6365 & 0.9 : & 3.5 : & [ ] @xmath57 6374 & 0.9 : & 3.5 : & @xmath122 6386 & 0.7 & 2.8 & @xmath123 6443 & 0.7 : & 2.9 : & @xmath124 6564 & 302&1265 & @xmath125 ( h@xmath8 ) 6615 & 0.4 : & @xmath01.8 : & dib 66146659 & 1.1 : & @xmath04.9 : & dib 66616681 & 0.4 : & @xmath01.8 : & @xmath126 lccl 6250 & 0.2 : & 0.5 : & @xmath127 6286 & 3.0 & @xmath011.8 & dib 67846303 & 1.3 & 4.9 & [ ] @xmath59 6320 & 0.6 : & 2.2 : & 6348 & 0.7 & 2.8 & @xmath121 6366 & 0.8 : & 3.1 : & [ ] @xmath57 6371 & 0.8 : & 3.1 : & @xmath122 6388 & 0.8 & 3.3 & @xmath123 6443 & 0.2 : & 0.9 : & @xmath124 6565 & 283 & 1143 & @xmath125 ( h@xmath8 ) lcc @xmath90 ( h@xmath53 ) & 11.5 & 12.2 @xmath100 ( h@xmath52 ) & 100 & 100 @xmath66 & 5.9 & 5.1 @xmath128 & 12.5 & 11.9 @xmath129 & 6.3 & 3.2 @xmath130 & 4.1 & 3.1 @xmath58 & 1.9 : & 4.5 @xmath131 & 5.3 & 2.2 @xmath119 & 5.9 & 7.1 @xmath59 & 12.8 & 23.5 @xmath132 & 7.5 & 12.7 @xmath121 & 6.8 & 8.6 @xmath123 & 8.8 & 12.3 @xmath125 ( h@xmath8 ) & 3953 & 4221 @xmath57 & 5.5 : & 6.3 & & h@xmath8/h@xmath52 & 39.5 & 42.2 h@xmath53/h@xmath52 & 0.115 & 0.122 ( 6300 + 6364/5577 ) & 8.0 & 6.6 @xmath59/ [ ] @xmath129 & 2.0 & 7.3 | iras 06562@xmath00337 has been the recent subject of a classic debate : proto - planetary nebula or young stellar object ?
we present the first 2@xmath1 image of iras 06562@xmath00337 , which reveals an extended diffuse nebula containing approximately 70 stars inside a 30@xmath2 radius around a bright , possibly resolved , central object .
the derived stellar luminosity function is consistent with that expected from a single coeval population , and the brightness of the nebulosity is consistent with the predicted flux of unresolved low - mass stars .
the stars and nebulosity are spatially coincident with strong co line emission .
we therefore identify iras 06562@xmath00337 as a new young star cluster embedded in its placental molecular cloud .
the central object is likely a herbig be star , @xmath3 , which may be seen in reflection .
we present medium resolution , high s / n , 1997 epoch optical spectra of the central object .
comparison with previously published spectra shows new evidence for time variable permitted and forbidden line emission , including , , [ ] , and [ ] .
we suggest the origin is a dynamic stellar wind in the extended , stratified atmosphere of the massive central star in iras 06562@xmath00337 . |
we say that a vector field on a riemannian manifold is geodesic if its integral curves are reparametrized geodesics .
the present paper discusses@xmath0 we observe ( remark [ dppdm ] ) that for @xmath1 and @xmath2 , where @xmath3 and @xmath4 are the maximum and minimum level sets of @xmath5 , one then has @xmath6 and every @xmath7 satisfying ( [ ddm ] ) is realized by some @xmath8 with ( [ trp ] ) .
one of our three main results , theorem [ cpbdl ] , classifies the triples ( [ trp ] ) such that @xmath9 here @xmath10 , while @xmath11 sends each @xmath12 to the unique point nearest @xmath13 in @xmath14 .
( in case ( [ trp ] ) @xmath15 always are disk - bundle projections , and their vertical distributions @xmath16 span a vector subbundle of @xmath17 , cf
. section [ cc ] ; however , ( [ trp ] ) does not imply ( [ spn ] ) see remark [ nintg ] . ) as a consequence of theorem [ cpbdl ] , in every triple with ( [ trp ] ) and ( [ spn ] ) , @xmath18 the remaining two main results of the paper , theorems [ tgimm ] and [ dicho ] , deal with the general case of ( [ trp ] ) , that is , do not assume ( [ spn ] ) . according to theorem [ tgimm ] , whenever @xmath19 is a leaf of either ( obviously integrable ) vertical distribution @xmath20 , the other projection @xmath21 maps @xmath22 onto the image @xmath23 of some totally geodesic holomorphic immersion @xmath24 inducing on @xmath25 a multiple of the fubini - study metric , with @xmath26 given by @xmath27 .
both @xmath28 are themselves ( connected ) totally geodesic compact complex submanifolds of @xmath29 , cf .
remark [ ascdt](iii ) .
the third main result reveals a dichotomy involving the assignment @xmath30 @xmath31 being the complex grassmannian , with @xmath32 defined as before .
specifically , theorem [ dicho ] states that one of the following two cases has to occur .
first , ( [ asg ] ) may be constant on every leaf of @xmath33 in @xmath34 , that is , on every fibre @xmath19 of the projection @xmath11 restricted to @xmath34 , with either sign @xmath35 .
otherwise , @xmath36 and @xmath32 are positive for both signs @xmath35 , while ( [ asg ] ) restricted to any such leaf @xmath19 must be a composite mapping @xmath37 formed by a holomorphic bundle projection @xmath38 , having the fibre @xmath39 , and a nonconstant holomorphic embedding @xmath40 .
the first case of theorem [ dicho ] is equivalent to condition ( [ spn ] ) , and the immersions @xmath41 , mentioned in the above summary of theorem [ tgimm ] , are then embeddings , for both signs @xmath35 , while their images constitute foliations of @xmath14 , both with the same leaf space @xmath42 appearing in ( [ bdl ] ) .
see remark [ infty ] . in the second case ( cf .
remark [ nncst ] ) the images of these immersions , rather than being pairwise disjoint , are totally geodesic , holomorphically immersed complex projective spaces , an uncountable family of which passes through each point of @xmath14 .
three special classes of the objects ( [ trp ] ) have been studied before .
one is provided by the gradient khricci solitons discovered by koiso @xcite and , independently , cao @xcite , where @xmath43 is the soliton function ; two more by special khricci potentials @xmath43 on compact khler manifolds @xcite , and by triples with ( [ trp ] ) such that @xmath44 is a ( compact ) complex surface @xcite .
each of these three classes satisfies ( [ spn ] ) .
the papers @xcite provide complete explicit descriptions of the classes discussed in them .
our theorem [ cpbdl ] generalizes their classification results , namely , ( * ? ? ?
* theorem 16.3 ) and ( * ? ? ? * theorem 6.1 ) . for more details on the preceding two paragraphs , see remark [ clone ]
. functions with geodesic gradients on arbitrary riemannian manifolds , usually called _ transnormal _ , have been studied extensively as well @xcite .
both authors research was supported in part by a fapesp
osu 2015 regular research award ( fapesp grant : 2015/50265 - 6 ) .
the authors wish to thank fangyang zheng for helpful comments .
manifolds , mappings and tensor fields , including riemannian metrics and functions , are by definition of class @xmath45 . a ( sub)manifold is always assumed connected .
our sign convention about the curvature tensor @xmath46 of a connection @xmath47 in a vector bundle @xmath48 over a manifold @xmath44 is @xmath49}\w\hh\xi\ ] ] for any section @xmath50 of @xmath48 and vector fields @xmath51 tangent to @xmath29 .
one may treat @xmath52 , the covariant derivative @xmath53 , and any function @xmath54 on @xmath44 as bundle morphisms @xmath55 sending @xmath50 or @xmath56 as above to @xmath57 , @xmath58 or , respectively , @xmath59 .
notation of ( [ nwt ] ) is used in the next three displayed relations . in the case of a riemannian manifold @xmath60 ,
the symbol @xmath47 will always stand for the levi - civita connection of @xmath61 as well as the @xmath62-gradient . given a function @xmath43 and vector fields @xmath63 on @xmath60 ,
one has the lie - derivative relation @xmath64(\hn w , w\hh'\hh)=2\hh g(\dv\nh w , w\hh'\hh)\hh,\hskip8pt\mathrm{where } \hskip5ptv=\navp\hskip5pt\mathrm{and}\hskip5pt\dv = \nabla\nh v : t\nnh\mf\to t\nnh\mf\hh,\ ] ] due to the local - coordinate equalities @xmath65_{jk}\w = v_{j , k}\w+v_{k , j}\w=2v_{j , k}\w$ ]
. for vector fields @xmath66 on a manifold @xmath44 and a bundle morphism @xmath67 , the leibniz rule gives @xmath68u=[v,\bl u]-\bl[v , u ] = [ \nabla\hskip-3pt_u\w\bl]u+\bl\nabla\hskip-3pt_u\w v -\nabla\hskip-3pt_{\bl u}\w v$ ] , and so @xmath69\hh.\ ] ] next , let @xmath70 be a killing vector field on a riemannian manifold @xmath60 .
the ricci and bianchi identities imply , as in ( * ? ? ?
* bottom of p. 572 ) , the well - known relation @xmath71 since the flow of a killing field preserves the levi - civita connection , ( [ scd ] ) also follows from the classical lie - derivative equality@xmath72\hn _ v\w w = [ \nabla\hskip-3pt_v\w\du]\hs w - r\hh(\hn u , v)\hh w$ ] , with @xmath73 , cf .
* formula ( 1.8 ) on p. 337 ) , valid for any connection @xmath47 in @xmath74 .
whenever @xmath75 is a function on a riemannian manifold @xmath60 , we have @xmath76 as one sees noting that , in local coordinates , @xmath77 .
also , obviously @xmath78 [ dvwwp]relation ( [ lvg ] ) becomes @xmath79=2\hh g(\dv\nh w , w\hh'\hh)\,$ ] if , in addition , @xmath80 commutes with @xmath81 and @xmath82 .
namely , @xmath83 on functions , so that we may evaluate @xmath79\,$ ] using the leibniz rule for the lie derivative with @xmath84 .
[ gpgrd]whenever the @xmath62-gradient @xmath85 of a function @xmath43 on a riemannian manifold @xmath86 is tangent to a submanifold @xmath87 with the submanifold metric @xmath88 , the restriction of @xmath56 to @xmath87 obviously equals the @xmath88-gradient of @xmath89 .
[ cifot]given a manifold @xmath44 and @xmath90 , we call @xmath91 _ a @xmath92 function of _ @xmath43 if @xmath43 is nonconstant ( so that its range @xmath93 is an interval ) and @xmath94 for some @xmath92 function @xmath95 .
note that @xmath96 is then uniquely determined by @xmath97 and @xmath5 .
we will denote by @xmath91 both the original function @xmath98 and the function @xmath99 of the variable @xmath100 .
let @xmath101 be a fixed variation of curves in a manifold @xmath29 , that is , a @xmath92 mapping in which the real variables @xmath102 range independently over intervals .
the partial derivative @xmath103 ( or , @xmath104 ) then assigns to each @xmath105 the velocity vector at @xmath106 ( or , @xmath107 ) of the curve @xmath108 or , respectively , @xmath109 .
( thus , @xmath103 and @xmath104 are sections of a specific pullback bundle . )
a connection @xmath47 on @xmath44 allows us to define the mixed second - order partial derivatives @xmath110 and @xmath111 of the variation , so that , for instance , the value of @xmath110 at @xmath105 is the @xmath112-covariant derivative , at the parameter @xmath107 , of the vector field @xmath113 along the curve @xmath109 , and analogously for @xmath111 . obviously , @xmath114 when @xmath47 is torsion - free , cf .
@xcite .
[ jcobi]for a torsion - free connection @xmath47 on a manifold @xmath44 and a smooth variation @xmath115 of @xmath112-geodesics , with @xmath116 near @xmath117 in @xmath118 and a vector field @xmath119 along a curve @xmath120 , let @xmath121 be the jacobi vector field along the geodesic @xmath122 defined by @xmath123 ( notation of the last paragraph )
. then @xmath124(t_0\w ) = [ \nabla\hskip-3pt_{\dot y}\w\hs\xi](s_0\w)$ ] , which is nothing else than @xmath114 ( see above ) at @xmath125 .
also , clearly , @xmath126 .
note that @xmath127 may be nonzero even if @xmath128 is a constant curve , as it then equals the ordinary derivative of @xmath129 with respect to @xmath130 .
[ tlspc]let @xmath131 be a vector bundle over a manifold @xmath132 .
we use the same symbol @xmath133 for its total space , which we identify , as a set , with @xmath134 . given a connection @xmath135 in @xmath133 and vector fields @xmath51 tangent to @xmath132 , @xmath136 ( notation of ( [ nwt ] ) ) .
namely , at any @xmath137 , the vertical ( or , horizontal ) component of the lie bracket of the horizontal lifts of @xmath56 and @xmath81 equals @xmath138 ( an easy exercise ) or , respectively , the horizontal lift of @xmath139_y\w$ ] , cf .
@xcite .
[ nexpm]let @xmath135 be the normal connection in the normal bundle @xmath140 of a totally geodesic submanifold @xmath141 in a riemannian manifold @xmath60 .
we denote by @xmath142 the normal exponential mapping of @xmath132 , the domain of which is an open submanifold @xmath143 of the total space @xmath140 such that , for every normal space @xmath144 , where @xmath145 , the intersection @xmath146 is nonempty and star - shaped ( in the sense of being a union of line segments emanating from @xmath147 ) .
remark [ jcobi ] leads to the following well - known description of the differential @xmath148 of @xmath149 at any @xmath150 , cf .
remark [ tlspc ] .
specifically , we may assume that @xmath151 since , clearly , @xmath152 when @xmath153 , under the obvious isomorphic identification @xmath154 = t\hskip-3pt_y\w\sd\oplus\nr\hskip-2.4pt_y\w\sd = t\hskip-3pt_y\w\mf\nh$ ] .
the point @xmath155 and the normal vector @xmath156 thus have the property that the nontrivial geodesic @xmath157 is defined for all @xmath158 $ ] . if @xmath159 , a vector tangent to @xmath140 at @xmath160 can be uniquely written as @xmath161 , where @xmath162\,$ ] is vertical and @xmath163 denotes the @xmath164-horizontal lift of some @xmath165 .
then , for the jacobi field @xmath166 along our geodesic @xmath167 such that @xmath168 and @xmath124(0)=\eta$ ] , @xmath169\hh.\ ] ] in fact , linearity of both sides in @xmath170 allows us to consider two separate cases , @xmath171 and @xmath172 .
for @xmath173 close to @xmath174 in @xmath175 and @xmath158 $ ] , let us set @xmath176 , where in the former case @xmath177 , and in the latter @xmath178 is the @xmath164-parallel normal vector field with @xmath179 along a fixed curve @xmath180 such that @xmath181 and @xmath182 .
thus , in both cases , the curve @xmath183 in @xmath184 has , at @xmath185 , the velocity @xmath186 .
the velocity at @xmath185 of its @xmath187-image curve @xmath188 therefore equals the left - hand side of ( [ dxp ] ) . at the same time
this last velocity is @xmath189 for @xmath190 defined as in remark [ jcobi ] with the variable @xmath191 and @xmath105 replaced by @xmath192 and @xmath193 .
now ( [ dxp ] ) follows since the two definitions of @xmath190 agree : according to remark [ jcobi ] , both jacobi fields denoted by @xmath190 satisfy the same initial conditions at @xmath185 .
[ kiljc]every killing vector field @xmath70 on a riemannian manifold is a jacobi field along any geodesic @xmath194 . in fact , the local flow of @xmath195 , applied to the geodesic , yields a variation of geodesics .
( equivalently , one may note that ( [ scd ] ) with @xmath196 , evaluated on @xmath197 , is precisely the jacobi equation . )
[ ttgim]let @xmath198 be a totally geodesic immersion of a manifold @xmath199 in a riemannian manifold @xmath200 . if @xmath201 and @xmath202 for submanifolds @xmath203 of @xmath199 and @xmath141 of @xmath29 , such that @xmath141 is totally geodesic in @xmath200 , then , for @xmath132 endowed with the submanifold metric , @xmath204 is a totally geodesic immersion .
in fact , every point of @xmath203 has a neighborhood @xmath205 in @xmath87 on which @xmath206 is an embedding with a totally geodesic image @xmath207 .
our claim now follows since the submanifold @xmath208 of @xmath132 , being the intersection of the totally geodesic submanifolds @xmath207 and @xmath132 , must itself be totally geodesic .
[ crvcm]let @xmath209 and @xmath210 be the curvature tensors of connections @xmath211 in vector bundles @xmath212 over a fixed base manifold and , respectively , of the connection @xmath213 induced by them in the vector bundle @xmath214 .
then @xmath210 is given by the commutator - type formula @xmath215\vartheta\hs-\hs\vartheta[r(\hn v , w)]$ ] , cf .
( [ nwt ] ) , for any section @xmath216 of @xmath217 ( that is , any vector - bundle morphism @xmath218 ) and vector fields @xmath51 tangent to the base .
this trivially follows from ( [ cur ] ) and the fact that @xmath219\hh\xi = \nabla'\hskip-6pt_v\w(\vartheta\hh\xi)-\vartheta\nabla\hskip-3pt_v\w\xi\,$ ] whenever @xmath50 is a section of @xmath220 .
as usual , whenever @xmath221 is a mapping between manifolds , we say that a vector field @xmath81 ( or , a distribution @xmath222 ) on @xmath44 is @xmath223-projectable if @xmath224 for some vector field @xmath70 ( or , some distribution @xmath225 ) on @xmath42 and all @xmath226 .
[ liebr]let @xmath221 be a bundle projection .
a vector field @xmath81 on @xmath44 is @xmath223-projectable if and only if , for every section @xmath56 of the vertical distribution @xmath227 , the lie bracket @xmath139\,$ ] is also a section of @xmath228 .
this is easily verified in local coordinates for @xmath44 that make @xmath229 appear as a cartesian - product projection .
[ cmmut]for @xmath230 as in remark [ liebr ] , a @xmath223-projectable vector field @xmath81 on @xmath29 , and @xmath231 such that @xmath232 , every prescribed value @xmath233 is realized by a local section @xmath70 of @xmath234 commuting with @xmath235 .
namely , we may first prescribe such @xmath70 along a fixed codimension - one submanifold containing @xmath236 , which is transverse to @xmath81 at @xmath236 , and then use the local flow of @xmath81 to spread @xmath70 over a neghborhood of @xmath236 .
[ prjct]given a vector field @xmath56 and a distribution @xmath237 on a manifold , the local flow @xmath238 of @xmath56 preserves @xmath237 if and only if , whenever @xmath81 is a local section of @xmath239 , so is @xmath139 $ ] .
namely , @xmath139=\lie\hskip-.5pt_v\w w$ ] , while , denoting by @xmath240 the push - forward action of @xmath241 on tensor fields @xmath216 of any type , we have @xmath242\hh / dt\ , = \,-\hs(de^{tv})\lie\hskip-.5pt_v\w\vartheta\hh.\ ] ] in fact , when @xmath243 , ( [ ddt ] ) is just the definition of @xmath244 , while , for arbitrary @xmath245 it follows from the group - homomorphic property of @xmath246 .
[ pralg]we say that a vector field @xmath81 ( or , a distribution @xmath225 ) on a manifold @xmath247 is projectable along an integrable distribution @xmath234 on @xmath29 , or @xmath228-projectable , if it is @xmath223-projectable , as in ( [ prj ] ) , when restricted to any open submanifold of @xmath44 on which @xmath248 forms the vertical distribution @xmath249 of a bundle projection @xmath223 . for @xmath81
this amounts to invariance of @xmath234 under the local flow of @xmath235 , cf .
remarks [ liebr ] and [ prjct ] . [ tglvs]for an integrable distribution @xmath250 on a riemannian manifold @xmath60 , the following two conditions are equivalent .
1 . @xmath251=0\,$ ] for all local sections @xmath56 of @xmath250 and @xmath81 of @xmath252 such that @xmath81 is nonzero , @xmath253-projectable , and @xmath139=0 $ ] .
2 . every leaf ( maximal integral manifold ) of @xmath250 is totally geodesic in @xmath60 .
in fact , let @xmath254 be the second fundamental form of the leaves of @xmath253 , with @xmath255 equal to the @xmath252 component of @xmath256 .
if @xmath56 and @xmath81 commute , @xmath257=2\hh g(\nabla\hskip-3pt_{w}\w v , v ) = 2\hh g(\nabla\hskip-3pt_{v}\w w , v)=-\nnh2\hh g(\nabla\hskip-3pt_{v}\w v , w ) = -\nnh2\hh g(b(\hn v , v),w)$ ] , while @xmath254 is symmetric and @xmath51 as in ( i ) realize , at any @xmath226 , any given elements of @xmath258 and @xmath259 ( see remark [ cmmut ] )
. [ frthr]clearly , ( ii ) in remark [ tglvs ] also follows when @xmath257=0\,$ ] for all @xmath51 satisfying specific further conditions besides ( i ) , as long as the last line of remark [ tglvs ] still applies .
[ intgr ] for two integrable distributions @xmath260 on a manifold @xmath44 such that the span @xmath237 of @xmath261 and @xmath262 has constant dimension , the following conditions are all equivalent . 1
. @xmath263 is integrable .
@xmath264 is projectable along @xmath265 .
@xmath266 is projectable along @xmath261 .
if ( a ) ( c ) hold , the distributions that @xmath260 locally project onto are integrable as well .
we may assume that @xmath267 is the vertical distribution of a bundle projection @xmath221 with connected fibres .
first , let @xmath268 be integrable . since @xmath237 contains @xmath269 , its leaves are unions of fibres and so their @xmath223-images form a foliation of @xmath270 , tangent to a distribution @xmath271 satisfying ( [ prj ] ) , which proves projectability of @xmath239 , and hence of @xmath262 , along @xmath272 .
in other words , ( a ) implies ( c ) .
conversely , assuming ( c ) , we obtain @xmath273 for all @xmath231 and some distribution @xmath225 on @xmath270 , which is necessarily integrable : its leaves are @xmath223-images of the leaves of @xmath262 .
integrability of @xmath237 now follows , as its leaves are the @xmath223-preimages of those of @xmath225 .
finally , as ( a ) involves @xmath267 and @xmath274 symmetrically , it is also equivalent to ( b ) .
for khler manifolds we use symbols such as @xmath60 , where @xmath44 stands for the underlying complex manifold . generally , in complex manifolds , @xmath275 let @xmath56 be a vector field on a khler manifold @xmath60 .
since @xmath276 , one has @xmath277 @xmath278 being viewed as bundle morphisms @xmath279 , cf .
( [ nwt ] ) . for the curvature tensor @xmath280 of a khler manifold @xmath86 and any vector fields @xmath281 on @xmath29 , @xmath282 in fact
, the condition @xmath283 turns @xmath47 into a connection in @xmath284 treated as a complex vector bundle , @xmath61 being the real part of a @xmath285-parallel hermitian fibre metric , that is , @xmath286 for all vector fields @xmath63 on @xmath29 , and so the curvature operators @xmath287 are all complex - linear and skew - hermitian .
the former property now amounts to commutation in ( [ rcm ] ) , the latter to the equality @xmath288 , with any vector fields @xmath289 .
real - holomorphic vector fields @xmath56 on khler manifolds will always be briefly referred to as holomorphic .
since they are characterized by @xmath290 , formula ( [ lie ] ) for @xmath291 implies that , given a vector field @xmath56 on a khler manifold @xmath60 , @xmath292 where @xmath293 as in ( [ nwt ] ) . for any holomorphic vector field @xmath80,@xmath294 in fact , for @xmath295 and @xmath73 ,
( [ ajs ] ) ( [ hol ] ) give @xmath296 , and so @xmath297 , while the local - gradient property of @xmath56 amounts to @xmath298 , and the killing condition for @xmath70 reads @xmath299 .
[ kilxp]as shown by kobayashi @xcite , if @xmath70 is a killing vector field on a riemannian manifold @xmath60 , the connected components of the zero set of @xmath70 are mutually isolated totally geodesic submanifolds of even codimensions .
[ nontr ] if a complex manifold @xmath44 admits a khler metric @xmath62 , with the khler form @xmath300 , and @xmath301 is a nonconstant holomorphic mapping , then @xmath302 represents a nonzero de rham cohomology class in @xmath303 .
whether a holomorphic mapping @xmath301 is constant , or not , the same is the case for all holomorphic mappings @xmath304 sufficiently close to @xmath305 in the @xmath306 topology .
clearly , @xmath305 remains nonconstant ( and holomorphic ) when restricted to a suitable projective line @xmath307 .
in addition to being positive semidefinite everywhere , the restriction @xmath308 of @xmath309 to @xmath310 must also be positive definite somewhere ( or else @xmath311 , being hermitian , would vanish identically , making @xmath305 constant on @xmath310 ) .
the integral of @xmath312 over @xmath310 is thus positive , proving our first claim .
the second one follows since nearby continuous mappings are , obviously , homotopic to @xmath313 .
[ fibra]we need the following well - known fact , valid both in the @xmath92 and complex ( holomorphic ) categories : any integrable distribution with compact simply connected leaves constitutes the vertical distribution of a bundle projection .
the required local trivializations are provided by the necessarily trivial
holonomy of the underlying foliation ; see , for instance , @xcite .
[ known]we need two more well - known facts ; cf .
* example 1 of sect .
2.2 ) . 1 .
a continuous function @xmath314 on an open set @xmath315 , holomorphic on @xmath316 , where @xmath317 is discrete , is necessarily holomorphic everywhere in @xmath318 .
2 . the only injective holomorphic mappings @xmath319 are biholomorphisms .
[ hlext]if @xmath198 is a continuous mapping between complex manifolds , and a codimension - one complex submanifold @xmath203 of @xmath320 , closed as a subset of @xmath320 , has the property that the restrictions of @xmath206 to @xmath87 and to the complement @xmath321 are both holomorphic , then @xmath206 is holomorphic on @xmath320 . in fact
, let @xmath322 . when @xmath323 , our claim is obvious from remark [ known](a )
generally , in local holomorphic coordinates @xmath324 for @xmath87 such that @xmath325 on the intersection of @xmath203 with the coordinate domain , the complex partial derivatives of the components of @xmath206 ( relative to any local holomorphic coordinates in @xmath247 ) all clearly exist : for @xmath326 this follows from the case @xmath323 .
[ posit]as usual , we call a differential @xmath327-form @xmath328 on a complex manifold positive if it equals the khler form @xmath329 of some khler metric @xmath62
. this amounts to requiring closedness of @xmath328 along with symmetry and positive definiteness of the twice - covariant tensor field @xmath330 .
in any complex manifold , @xmath331 and @xmath332 symmetric whenever @xmath333 or , equivalently , @xmath334\,$ ] for a real - valued function @xmath335 , with the @xmath336-form @xmath337 , also denoted by @xmath338 , which sends any tangent vector field @xmath56 to @xmath339 .
clearly , @xmath340 if @xmath54 is a @xmath92 function of a function @xmath96 on the same manifold ( cf . remark [ cifot ] ) .
the exterior - derivative and exterior - product conventions used here , for any @xmath336-forms @xmath341 and vector fields @xmath342 , are @xmath343-d_v[\kappa(u)]-\kappa([u , v])\,$ ] and @xmath344 .
when , in addition , @xmath56 is real - holomorphic , one has @xmath345\hskip10pt\mathrm{for}\hskip7pt\omega = i\hs\partial\hskip1.7pt\cro\hskip-1.5ptf.\ ] ] see ( * ? ? ?
* lemma 2 ) ; the khler metric used in @xcite always exists locally .
[ ddnsq]for the real part @xmath346 of a hermitian inner product in a fidimensional complex vector space @xmath347 , let @xmath348 and @xmath234 be the norm function and complex radial distribution on @xmath349 , so that @xmath350 and @xmath351 . 1 .
@xmath352 is obviously given by @xmath353 .
2 . @xmath354 coincides with twice the khler form @xmath355 of the constant metric @xmath356 .
3 . @xmath357 , on @xmath349 , equals @xmath358 times the restriction of @xmath355 to @xmath228 .
in fact , ( b ) ( c ) are immediate from ( a ) and remark [ posit ] .
[ cptfb]let @xmath221 be a surjective submersion between manifolds . 1 . if the preimages @xmath359 , @xmath360 , are all compact , then @xmath229 can be factored as @xmath361 , with a bundle projection @xmath362 having compact ( connected ) fibres , and a finite covering projection @xmath363 .
2 . in the case where @xmath364 and @xmath44 is compact , @xmath229 must necessarily be a ( finite ) covering projection .
namely , ( a ) is a well - known fact , easily verified using parallel transports corresponding to a fixed vector subbundle @xmath271 of @xmath284 for which @xmath365 cf .
* remark 1.1 ) , or derived as in remark [ fibra ] , since the foliation with the leaves @xmath366 has trivial holonomy .
part ( b ) in which the dimension equality means that @xmath229 is locally - diffeomorphic easily follows from ( a ) .
[ ftcvr]for a khler manifold @xmath367 with @xmath368 , any holomorphic mapping @xmath369 such that @xmath370 is a positive constant multiple of the fubini - study metric on @xmath371 ( cf .
remark [ fbstm ] ) must be a biholomorphism .
in fact , @xmath372 is then a covering projection ( remark [ cptfb](b ) ) and our claim follows since , due to a result of kobayashi @xcite , @xmath87 has to be simply connected .
given a manifold @xmath44 endowed with a fixed connection @xmath285 , we refer to a vector field @xmath56 on @xmath44 as geodesic if the integral curves of @xmath56 are reparametrized @xmath112-geodesics .
equivalently , for some function @xmath373 on the open set @xmath374 on which @xmath375 , @xmath376 a function @xmath43 on a riemannian manifold @xmath86 is said to have a geodesic gradient if its gradient @xmath56 is a geodesic vector field relative to the levi - civita connection @xmath285 .
functions with geodesic gradients are also called _ transnormal _ @xcite .
[ ggqft ] for a function @xmath43 on a riemannian manifold @xmath60 , the gradient of @xmath43 is a geodesic vector field if and only if @xmath377 is , locally in @xmath378 , a function of @xmath5 . by ( [ tnd ] ) , condition ( [ nvv ] )
is equivalent to @xmath379 .
[ ggktr]a geodesic - gradient khler triple @xmath8 consists of any khler manifold @xmath86 and a nonconstant real - valued function @xmath43 on @xmath44 such that the @xmath62-gradient @xmath85 is both geodesic and real - holomorphic .
speaking of compactness of @xmath380 , or its dimension , we always mean those of the underlying complex manifold @xmath29 , and we call two such triples @xmath381 isomorphic if @xmath382 and @xmath383 for some biholomorphism @xmath384
. for @xmath8 as above , whenever the extrema of @xmath43 exist , we will also consider @xmath385 [ trivl]a geodesgradient khler triple @xmath8 can be trivially modified to yield @xmath386 , with any real constants @xmath387 and @xmath388 .
( clearly , @xmath28 in ( [ tmm ] ) then become switched if @xmath389 . )
any such @xmath8 and any complex submanifold @xmath390 of @xmath29 , tangent to @xmath85 ( that is , forming a union of integral curves of @xmath80 ) , and not contained in a single level set of @xmath5 , give rise ( cf .
remark [ gpgrd ] ) to the new geodesgradient khler triple @xmath391 , where @xmath392 are the restrictions of @xmath61 and @xmath43 to @xmath320 .
as shown next , geodesgradient khler triples naturally arise from suitable cohomogeneity - one isometry groups .
[ chone ] let a connected lie group @xmath393 acting by holomorphic isometries on a khler manifold @xmath60 , and having some orbits of real codimension @xmath336 , preserve a nontrivial holomorphic killing field @xmath70 with zeros . if @xmath394 , then @xmath8 is a geodesic - gradient khler triple and @xmath395 for some @xmath396-invariant function @xmath43 on @xmath29 . since @xmath394 , ( [ kil ] ) implies both the existence of a function @xmath5 with @xmath397 , and the fact that its gradient @xmath398 is holomorphic
thus , elements of @xmath399 preserve @xmath43 up to additive constants .
let @xmath400 now be a fixed connected component of the zero set of @xmath195 , so that @xmath396 , being connected , leaves @xmath400 invariant , while @xmath43 is constant on @xmath400 ( cf .
remark [ kilxp ] ) .
the additive constants just mentioned are therefore equal to @xmath147 . due to their @xmath396-invariance , the functions @xmath43 and @xmath401 are constant along codimension - one orbits of @xmath399 and , consequently , functionally dependent .
( note that the union of such orbits is dense in @xmath29 . )
consequently , by lemma [ ggqft ] , the gradient @xmath85 is a geodesic vector field .
[ reddt]the assumptions about triviality of @xmath402 and holomorphicity of @xmath70 in lemma [ chone ] are well - known to be redundant when @xmath44 is compact ( * ? ? ?
* , corollary 4.5 ) ; see also ( * ? ? ? * formula ( a.2c ) and theorem a.1 ) .
the following fact will be used in the proof of theorem [ jacob ] .
[ jcbfl ] if a vector field @xmath81 on a riemannian manifold @xmath86 is orthogonal to a geodesic gradient @xmath56 and commutes with @xmath80 , then @xmath81 is a jacobi field along every integral curve of @xmath403 in the set @xmath404 where @xmath375 .
fix @xmath75 with @xmath405 . for @xmath406 , ( [ tnd ] ) and ( [ nvv ] ) give @xmath379 , so that @xmath407 is , locally in @xmath378 , a @xmath92 function of @xmath43 ( cf .
remark [ cifot ] ) . as @xmath408 due to the orthogonality assumption , and @xmath409 , we now have @xmath410 on @xmath378 .
the local flow of @xmath235 , applied to any integral curve of @xmath411 , thus yields a variation of unit - speed geodesics , and our claim follows . [ dmone]a compact geodesgradient khler triple of complex dimension @xmath412 is essentially , up to isomorphisms , nothing else than the sphere @xmath413 with a rotationally invariant metric .
in fact , necessity of rotational invariance is due to ( [ kil ] ) , while its sufficiency follows from lemma [ chone ] with @xmath414 , for the sphere treated as @xmath310 with a khler metric .
( once the @xmath415 action is chosen , the required function @xmath43 becomes unique up to trivial modifications , cf .
remark [ trivl ] . )
in this section vector spaces are complex ( except for remark [ dfrtl ] ) and fidimensional . by @xmath416_-planes _ in a vector space @xmath417 we mean dimensional vector subspaces of @xmath418 . when @xmath419 , they will also be called lines in @xmath418 .
given a vector space @xmath417 and @xmath420 , the grassmannian @xmath421 is the set of all @xmath416-planes in @xmath418 .
each @xmath421 naturally forms a compact complex manifold ( see remark [ grass ] ) , and @xmath422 is the projective space of @xmath418 , provided that @xmath423 .
we will use the standard identification @xmath424 of @xmath425 with the disjoint union of an open subset biholomorphic to @xmath417 and a complex submanifold biholomorphic to @xmath426 via the biholomorphism sending @xmath427 , or the line @xmath428 spanned by @xmath429 , to the line @xmath430 or , respectively , @xmath431 .
the projectivization of a holomorphic vector bundle @xmath131 over a complex manifold @xmath132 is , as usual , the holomorphic bundle @xmath432 of complex projective spaces over @xmath141 with @xmath433_y\w\,=\,\hs\mathrm{p}\vs\hskip7pt\mathrm{for}\hskip7pt\vs\ , = \,\nr\hskip-2.4pt_y\w\hh,\hskip7pt\mathrm{whenever}\hskip7pty\in\sd\hh.\ ] ] for a subspace @xmath434 of a vector space @xmath417 such that @xmath435 , let @xmath399 be the group of all complex - linear automorphisms of @xmath417 preserving both @xmath434 and a fixed hermitian inner product in @xmath418 .
we now define a compact complex manifold @xmath44 by @xmath436 then the hypotheses , and hence conclusions , of lemma [ chone ] are satisfied by these @xmath437 , any @xmath396-invariant khler metric @xmath61 on @xmath29 , and some @xmath195 . specifically , @xmath70 is a vector field arising from the central circle subgroup @xmath415 of @xmath399 formed by all unimodular elements of @xmath399 acting in both @xmath438 as multiples of @xmath439 .
see the remarks below .
the triples @xmath8 arising via lemma [ chone ] in cases ( [ dta].i ) and ( [ dta].ii ) will from now on be called grassmannian triples and , respectively , cp triples . since @xmath399 as above contains all unit complex multiples of @xmath439 , its action on @xmath44 is not effective
. lemma [ chone ] does not require effectiveness of the action .
[ grcpt]the cohomogeneity - one assumption of lemma [ chone ] follows here from the fact that the orbits of @xmath399 coincide with the levels of a nonconstant real - analytic function @xmath440 .
specifically , in case ( [ dta].i ) , @xmath441 , where @xmath442 and @xmath443 is some@xmath444any unit vector spanning @xmath445 , which yields @xmath396-invariance of @xmath446 .
conversely , if @xmath447 and @xmath448 , an element of @xmath399 sending @xmath449 to @xmath450 is provided by any linear isometry mapping the quadruple @xmath451 onto @xmath452 .
( such an isometry will preserve @xmath453 . ) in case ( [ dta].ii ) we may use @xmath54 given by @xmath454 , with @xmath455 standing for some@xmath444any unit vector that spans the line @xmath456 .
[ zeros]for a grassmannian or cp triple @xmath380 , critical points of @xmath5 , that is , the zeros of @xmath70 or , equivalently , the fixed points of the central circle subgroup @xmath415 mentioned above , form the disjoint union of two ( connected ) compact complex submanifolds , which since @xmath43 is clearly constant on either of them must be the same as @xmath28 in ( [ tmm ] ) . with @xmath457 denoting biholomorphic equivalence , these @xmath28 are @xmath458\hh,\hskip9pt & \mathrm{b)}&\{\ws\in\mf:\ws\subseteq\ls\hskip-2.5pt^\perp\}\nh \approx\nh\mathrm{gr}\nh_k\w\ls\hskip-2.5pt^\perp\nnh,\\ \mathrm{c)}&\{\ws\in\mf:\ws\subseteq\ls\}\nh \approx\nh\mathrm{p}\ls\hh,&\mathrm{d)}&\{\ws\in\mf:\ws\subseteq\ls\hskip-2.5pt^\perp\}\nh \approx\nh\mathrm{p}\ls\hskip-2.5pt^\perp\nnh , \end{array}\ ] ] where ( [ spm].a ) ( [ spm].b ) correspond to ( [ dta].i ) , and ( [ spm].c )
( [ spm].d ) to ( [ dta].ii ) .
in fact , each of the four sets clearly consists of fixed points of @xmath459 .
conversely , suppose that @xmath460 does not lie in the union of the sets ( [ spm].a )
( [ spm].b ) ( or , ( [ spm].c ) ( [ spm].d ) ) , and @xmath461
. then @xmath462 .
namely , if @xmath463 preserved @xmath456 , the equalities @xmath464 and @xmath465 , along with @xmath466 and @xmath467 , would imply analogous equalities for the lines spanned by @xmath468 ( case ( [ dta].i ) ) , or by @xmath469 ( case ( [ dta].ii ) ) , cf .
( [ prx ] ) , @xmath443 being any unit vector in the line @xmath445 ( or , respectively , in the line @xmath470 ) . in either case , the plane @xmath471 spanned by the two @xmath472-invariant lines would contain a third such line , in the form of @xmath434 ( or , respectively , @xmath470 ) .
thus , @xmath463 restricted to @xmath471 would be a multiple of @xmath439 , leading to a contradiction : in both cases , @xmath471 contains nonzero vectors from @xmath434 and from @xmath473 , which are eigenvectors of @xmath463 for two distinct eigenvalues .
[ cpone]all compact geodesgradient khler triples of complex dimension @xmath336 are obviously isomorphic to cp triples constructed from the data ( [ dta].ii ) with @xmath474 and @xmath475 .
see remark [ dmone ] .
[ fbstm]given the real part @xmath346 of a hermitian inner product in a vector space @xmath418 , the fubini - study metric on @xmath426 associated with @xmath346 is , as usual , uniquely characterized by requiring that the restriction of the projection @xmath476 to the unit sphere of @xmath346 be a riemannian submersion .
another such real part @xmath477 yields the same fubini - study metric as @xmath346 only if @xmath477 is a constant multiple of @xmath356 .
in fact , a @xmath478-linear automorphism of @xmath417 taking @xmath346 to @xmath477 descends to an isometry @xmath479 and hence equals a @xmath356-unitary automorphism of @xmath417 followed by @xmath480 times @xmath481 for some @xmath482 , which gives @xmath483 .
[ grass]given a vector space @xmath417 and @xmath484 , we denote by @xmath485 and @xmath486 the stiefel manifold of all linearly independent ordered @xmath416-tuples of vectors in @xmath417 ( forming an open submanifolds of the @xmath487th cartesian power of @xmath488 ) and , respectively , the projection mapping sending each @xmath489 to @xmath490
. then @xmath421 has a unique structure of a compact complex manifold of complex dimension @xmath491 , where @xmath492 , such that @xmath229 is a holomorphic submersion .
[ tgvec]we will use the canonical isomorphic identification @xmath493\,\ , = \,\,\mathrm{hom}\hs(\ws,\vs\hskip-1.8pt/\hh\ws)\hskip12pt\mathrm{whenever } \hskip7ptk\in\{0,1,\dots,\dimc\nnh\vs\}\hh,\ ] ] for the tangent space of the complex manifold @xmath421 at any @xmath416-plane @xmath456 , where @xmath494 means ` the space of linear operators ' and @xmath417 is any vector space .
namely , let @xmath495 and @xmath486 be as in remark [ grass ] . under the identification ( [ twg ] )
, @xmath496 corresponds to @xmath497\,$ ] for any linear lift @xmath498 of @xmath499 and any basis @xmath500 of @xmath449 which , clearly , does not depend on how such @xmath501 and @xmath500 were chosen .
[ dfrtl]given a real ( or , complex ) manifold @xmath184 and real ( or , complex ) vector spaces @xmath502 , let @xmath503 be a @xmath92 ( or , holomorphic ) mapping giving rise to a constant function @xmath504 or , equivalently , leading to the same value of @xmath505 for all @xmath506 .
then the mapping @xmath507 is of class @xmath92 ( or , holomorphic ) and its differential @xmath508 at any @xmath506 , with @xmath509 , cf .
( [ twg ] ) , sends @xmath510 to the unique @xmath511 having a linear lift @xmath512 such that @xmath513 here @xmath514 , so that @xmath515 and @xmath516 are linear operators @xmath517 .
in fact , let ` regular ' mean @xmath92 or holomorphic .
given @xmath506 , we may select a subspace @xmath417 of @xmath518 so that @xmath519 , where @xmath520 .
for all @xmath521 near @xmath50 in @xmath522 , the restriction of @xmath523 to @xmath417 is clearly an isomorphism onto the image of @xmath524 . denoting by @xmath525 its inverse isomorphism
, we see that @xmath526 and @xmath527 coincide with the direct - sum projections of @xmath528 onto @xmath417 and @xmath529 .
a fixed basis @xmath530 of @xmath449 thus gives rise to the basis @xmath531 of any such @xmath529 , depending regularly on @xmath532 , which constitutes a regular local lift of ( [ ker ] ) valued in the stiefel manifold @xmath533 ( see remark [ grass ] ) , proving regularity of ( [ ker ] ) .
replacing our @xmath50 with @xmath534 for a curve @xmath535 , and letting @xmath536 be a fixed basis of @xmath537 , we set @xmath538\,$ ] and @xmath539 . suppressing from the notation the dependence on @xmath540 ,
one thus gets @xmath541\hh e_j\w\nh=0\,$ ] and , by differentiation , @xmath542\hh e_j\w\nh + [ f(\xi)]\hh\dot e_j\w\nh=0 $ ] . the operators @xmath543 defined by @xmath544 , @xmath545 , yield ( [ fxh ] ) at @xmath546 for @xmath547 instead of @xmath548 . on
the other hand , @xmath549 with @xmath550 is a regular lift of @xmath551 to @xmath552 , showing that @xmath553 is the image of @xmath554 under the differential of ( [ ker ] ) at @xmath555 .
this equality , at @xmath243 , uniquely determines @xmath556 for which our @xmath557 is a linear lift realizing the image just mentioned as in the final paragraph of remark [ tgvec ] ) , so that @xmath558 , and our claim follows .
we will repeatedly consider quadruples @xmath559 formed by @xmath560\hs\mathrm{,\ a\ constant\ } \,\,a\in(0,\infty)\mathrm{,}\\ \mathrm{and\ a\ } \,c^\infty\mathrm{\ function\ } \,q\,\mathrm{\ of\ the\ variable\ } \,\vt\in[\hh\tm,\tp]\mathrm{,\nh\ positive}\\ \mathrm{on\ } \,(\tm,\tp)\mathrm{,\nh\ such\ that\ } \nnh\,q=0\,\mathrm{\ and\ } \hn\,dq\hh/\nh d\vt=\mp2a\hn\,\mathrm{\ at\ } \hn\,\vt=\tpm\hh , \end{array}\ ] ] @xmath561 being the opposite sign of @xmath562 .
as explained below , we may then also choose @xmath563 a function @xmath564 , unique up to an additive constant , with @xmath565 and the unique increasing diffeomorphism @xmath566 such that @xmath567 the inverse of @xmath568 in ( [ sgn ] ) being used to treat @xmath569 as functions of @xmath570 , for @xmath571 in fact , one easily verifies that @xmath568 and @xmath572 with the stated properties exist , while @xmath573 is finite . see ( * ? ? ?
* remark 5.1 ) , ( * ? ? ?
* theorem 10.2(iii ) ) , @xcite .
[ xtens ] any @xmath54 satisfying ( [ dfr ] ) has a @xmath92 extension to @xmath574 , which is also a @xmath92 function of @xmath575 in a sense analogous to remark [ cifot ] .
the first and second derivatives of @xmath54 with respect to @xmath576 obviously are @xmath577 and @xmath578 . since @xmath579 , it clearly suffices to establish @xmath45-extensibility of @xmath580 to @xmath574 . to this end , note that , according to ( * ? ? ?
* top line on p. 83 and remark 4.3(ii ) ) , @xmath581 ( or , @xmath582 ) may be extended to a @xmath92 function of @xmath583\smallsetminus\{\tmp\}\,$ ] ( or , of @xmath584 ) with a nonzero value at @xmath585 ( or , respectively , at @xmath147 ) . however , by ( [ sgn ] ) , @xmath586 , so that the variables @xmath43 and @xmath576 just mentioned depend diffeomorphically on each other .
[ uniqe]it is the limit condition in ( [ dsr ] ) that makes @xmath587 unique ; by contrast , @xmath570 with ( [ sgn ] ) is only unique up to a positive constant factor .
[ tends]in ( [ sgn ] ) , the increasing function @xmath588 of the variable @xmath43 clearly tends to @xmath174 as @xmath589 , and to @xmath590 as @xmath591 .
[ compo]the composite @xmath592 of the inverses of the above two diffeomorphisms is the unique solution of the autonomous equation @xmath593 , with the sign @xmath594 fixed as in ( [ sgn ] ) , such that @xmath589 as @xmath595 .
( we say ` autonomous ' since ( [ pbd ] ) makes @xmath596 a function of @xmath5 . ) in fact , any two solutions of the equivalent equation @xmath597 differ by a constant .
[ iddbf]with ( [ pbd ] ) ( [ dfr ] ) fixed as above , suppose that @xmath598 simultaneously denotes a positive function on a complex manifold @xmath34 , which also turns @xmath54 into a function @xmath599 .
now ( [ idd ] ) for @xmath600 and the formulae in lemma [ xtens ] give @xmath601 [ smdiv]if a @xmath92 function @xmath602 on an interval @xmath603 having an endpoint @xmath604 vanishes at @xmath604 , then @xmath605 for a @xmath92 function @xmath96 on @xmath603 equal , at @xmath604 , to the derivative @xmath606 of @xmath607 .
this is the taylor formula , with @xmath608 .
[ adtet]given a @xmath92 function @xmath609 on an interval @xmath603 having the endpoint @xmath174 such that @xmath610 , where @xmath611 , and @xmath612 on @xmath613 , there exists a @xmath92 function @xmath614 , unique up to multiplications by positive constants , for which @xmath615 and @xmath616 everywhere in @xmath617 , while @xmath618 .
namely , remark [ smdiv ] implies that some @xmath92 functions @xmath619 ( without zeros ) and @xmath620 on @xmath603 satisfy the conditions @xmath621 and @xmath622 .
thus , @xmath623 has the antiderivative @xmath624 on @xmath613 , and we may set @xmath625 .
let @xmath346 be the real part of a hermitian fibre metric in a holomorphic complex vector bundle @xmath131 over a complex manifold @xmath132 .
the chern connection of @xmath356 , also called its hermitian connection , is the unique connection @xmath135 in @xmath131 which makes @xmath356 parallel and satisfies the condition @xmath626 , meaning that , for any section @xmath50 of @xmath627 , the complex - antilinear part of the real vector - bundle morphism @xmath628 equals @xmath629 , the image of @xmath630 under the cauchy - riemann operator .
the following five properties of the chern connection @xmath135 are well known ( e ) is obvious ; for ( a ) ( d ) see @xcite , ( * ? ? ?
* propositions 1.3.5 , 1.7.19 and 1.4.18 ) . 1 .
@xmath631 depends on @xmath131 and @xmath346 functorially with respect to all natural operations , including @xmath494 , direct sums , and pullbacks under holomorphic mappings .
2 . @xmath632 , with the notation of ( [ nwt ] ) and ( [ jcs ] ) , where @xmath633 are any vector fields on @xmath141 and , respectively , the curvature tensor of @xmath634 .
@xmath631 is the levi - civita connection of @xmath346 if @xmath635 and @xmath346 is a khler metric .
@xmath631 coincides with the normal connection in the normal bundle @xmath140 for any totally geodesic complex submanifold @xmath141 a khler manifold @xmath86 and the riemannian fibre metric @xmath346 in @xmath131 induced by @xmath62 .
( in addition , it follows then that @xmath131 must be a holomorphic subbundle of @xmath74 . )
@xmath636-parallel sections of @xmath131 are holomorphic .
any given local holomorphic coordinates @xmath637 in @xmath141 and local holomorphic trivializing sections @xmath638 for @xmath627 , on the same domain , associate with the hermitian fibre metric @xmath639 having the real part @xmath356 , sections @xmath50 of @xmath627 , and any connection @xmath634 , their component functions @xmath640 and @xmath641 , the latter characterized by @xmath642 , as well as @xmath643 and @xmath644 . here
@xmath645 and @xmath646 denote the @xmath164-covariant differentiations in the direction od the complexified coordinate vector fields @xmath647 and @xmath648 , repeated indices are summed over and , with @xmath649 and @xmath650 standing for the complex conjugates of @xmath651 and @xmath652 , hermitian symmetry of @xmath639 amounts to @xmath653 , while @xmath654 whenever @xmath655 are sections of @xmath627 .
the real coordinate vector fields corresponding to the real coordinates @xmath656 then are @xmath657 , and so , given a complexified vector field @xmath658 , @xmath659 for a connection @xmath135 to make @xmath660 parallel it is clearly necessary and sufficient that @xmath661 and @xmath662 , where @xmath663 and @xmath664 are defined to be the complex conjugates of @xmath665 and @xmath666 .
on the other hand , the condition @xmath667 is obviously equivalent to @xmath668 , that is , @xmath669 .
existence and uniqueness of the chern connection @xmath634 now follow , its component functions being @xmath670 consequently , the chern connection @xmath135 has the curvature components @xmath671 ( which implies ( b ) above ) .
here @xmath672 , and analogously for the other three pairs @xmath673 of indices .
we obtain ( [ ccd ] ) from ( [ cur ] ) via differentiation by parts , noting that @xmath674 = [ \hh\partial_\lambda\w,\partial_{\bar\mu}\w]=0\,$ ] while , by ( [ chc ] ) , @xmath675 and @xmath676 .
for instance , @xmath677 .
similarly , @xmath678 equals @xmath679 , which is symmetric in @xmath680 ( and so @xmath681 ) .
let @xmath682 . with the notational conventions of remark [ posit ] , @xmath683\ , dz^\lambda\nh\wedge\hh d\bar z^{\bar\mu}\hh,\ ] ] as @xmath684\,dz^\lambda\nh + [ \hh\partial_{\bar\mu}\w f\hh]\,d\bar z^{\bar\mu}$ ] and
, in the case where @xmath685 ( or , @xmath686 ) , the complex - valued @xmath336-form @xmath687 equals @xmath688 or , respectively , @xmath689 .
equivalently , @xmath690 thus , by ( [ ddb ] ) and ( [ rvf ] ) , for @xmath691 and any real vector fields @xmath66 on @xmath132 , @xmath692 [ ddrsq ] with @xmath693 and @xmath694 denoting the bundle projection and the norm function of @xmath356 , for @xmath695 as above , the chern connection @xmath135 of @xmath346 and the @xmath327-form @xmath696 satisfy the following conditions . 1 . the horizontal distribution of @xmath135 constitutes a complex vector subbundle of @xmath697 , and is @xmath698-orthogonal , in an obvious sense , to the vertical distribution @xmath699 .
2 . part ( b ) of remark [ ddnsq ] describes @xmath700 restricted to any fibre @xmath701 of @xmath702 , where @xmath703 .
3 . whenever @xmath137 , cf .
remark [ tlspc ] , the restriction of @xmath704 to the horizontal space of @xmath135 at @xmath13 equals the @xmath705-pullback of the @xmath327-form @xmath706 at @xmath703 .
the chern connection @xmath707 of @xmath708 , for any function @xmath709 , is related to @xmath135 by @xmath710 , so that @xmath711 and @xmath712
. also , @xmath713 , where @xmath714 and @xmath715 denote the @xmath164-horizontal and @xmath716-horizontal lifts of any @xmath717 to @xmath718 . in terms of complexified coordinate vector fields
@xmath719 in @xmath141 and their analogs @xmath720 corresponding to the local holomorphic coordinates @xmath721 in @xmath722 , the @xmath164-horizontal lifts @xmath723 of the former are given by @xmath724 since @xmath725 and @xmath726 , ( [ hlo ] ) implies complex - linearity of the @xmath164-horizontal lift operation relative to the complex - structure tensors @xmath727 , proving the first part of ( i ) .
assertion ( ii ) is in turn obvious from naturality of the operator @xmath728 .
next , applying ( [ lmz ] ) to @xmath729 and the coordinates @xmath721 rather than just @xmath730 , we obtain @xmath731 , @xmath732 , @xmath733 , @xmath734 , and @xmath735 .
thus , @xmath736 by ( [ chc ] ) , which amounts to the remaining claim in ( i ) ; similarly , ( [ hlo ] ) and ( [ ccd ] ) yield @xmath737 . now ( iii ) follows : for the @xmath164-horizontal lifts @xmath738 of any real vectors @xmath739 , ( [ lmz ] ) ( [ ovu ] ) and the last equality give @xmath740 , while at the same time @xmath741 is imaginary , and so @xmath742 which , analogously , equals @xmath743 . finally , ( [ chc ] ) and ( [ hlo ] ) imply ( iv ) .
the geodesgradient khler triples constructed in this section are all noncompact .
what makes them relevant is the fact that some of them serve as universal building blocks for compact geodesgradient khler triples .
( see theorem [ first ] . )
we begin with data @xmath744 and @xmath745 consisting of 1 .
the real part @xmath746 of a hermitian fibre metric in a holomorphic complex vector bundle @xmath133 of positive fibre dimension over a khler manifold @xmath747 , 2 . some objects @xmath748 and @xmath745 satisfying ( [ pbd ] ) ( [ dfr ] ) .
letting @xmath749 stand for the bundle projection , @xmath135 for the chern connection of @xmath346 ( see section [ ck ] ) , and @xmath598 both for the variable in ( ii ) and for the norm function @xmath750 , we use the inverse mapping of @xmath751 , cf .
( [ sgn ] ) , to @xmath752 denoting by @xmath753 ( rather than @xmath727 ) the complex - structure tensor of @xmath754 , we define a khler metric @xmath755 on @xmath131 by requiring the khler forms @xmath756 and @xmath757 to be related by @xmath758 , which amounts to @xmath759 as @xmath760 should be positive ( remark [ posit ] ) , it is necessary to assume here that @xmath761 the above construction uses the objects ( i )
( ii ) with ( [ thd ] ) , and leads to what is shown below ( theorem [ ehggk ] ) to be a geodesgradient khler triple @xmath762 .
it is convenient , however , to provide the following equivalent , though less concise , description of @xmath763 and @xmath753 restricted to the complement @xmath764 of the zero section in @xmath627 .
it uses the complex direct - sum decomposition @xmath765 in which @xmath766 is the horizontal distribution of @xmath135 and @xmath767 equals the vertical distribution , with the summands @xmath768 and @xmath769 forming , on each punctured fibre @xmath770 , the complex radial distribution ( remark [ ddnsq ] ) and , respectively , its @xmath356-orthogonal complement in @xmath770 .
( the word ` complex ' preceding ( [ tnp ] ) is justified by lemma [ ddrsq](i ) . ) to describe @xmath763 and @xmath771 , we declare that the three summands of ( [ tnp ] ) are @xmath771-invariant and mutually @xmath772-orthogonal , that @xmath753 restricted to @xmath768 agrees , along each punctured fibre @xmath770 , with its standard complex - structure tensor of the complex vector space @xmath773 , that the differential of @xmath229 at every @xmath774 , cf . remark [ tlspc ] , maps @xmath775 complex - linearly onto @xmath776 and , with the constant @xmath777 and function @xmath778 appearing in ( i ) and ( [ trt ] ) , @xmath779 at any @xmath780 , where @xmath63 are any two vectors in @xmath781 , and @xmath782 denote their @xmath164-horizontal lifts to @xmath236 .
the vertical vector fields @xmath783 with @xmath784 allow us to characterize the restrictions of @xmath763 and @xmath753 to @xmath785 by @xmath786 note that the symmetry of @xmath787 in @xmath782 reflects ( b ) in section [ ck ] .
lemma [ ddrsq ] easily implies that the definition ( [ hgv ] ) of @xmath763 is actually equivalent to ( [ hoh ] ) , while condition ( [ thd ] ) is nothing else than positivity of the right - hand side in ( [ hgv].c ) whenever @xmath788 .
[ ehggk ] for any data ( i )
( ii ) with ( [ thd ] ) , let us define @xmath789 by ( [ trt ] ) ( [ hoh ] ) . 1 . @xmath790 is a geodesic - gradient khler triple .
2 . the fibres @xmath791 , @xmath792 , are totally geodesic complex 3 .
the zero section @xmath793 coincides with @xmath794 , the @xmath795 level set of @xmath5 .
4 . the @xmath772-gradient @xmath796 and @xmath797 satisfy ( [ lvf ] ) ( [ ana ] ) and the equality @xmath798 the assumptions being the same as in ( [ hgv].c ) .
the @xmath799-directional derivatives of the norm squared @xmath576 and of @xmath598 are , obviously , @xmath800 and @xmath801 .
as @xmath802 in ( [ sgn ] ) , we see using ( [ trt ] ) and ( [ ana ] ) that 1 .
@xmath803 equals the @xmath799-directional derivative @xmath804 of @xmath805 .
furthermore , @xmath164-parallel transports preserve the real fibre metric @xmath356 .
therefore , 1 .
@xmath806 and @xmath807 are constant along every @xmath164-horizontal curve in @xmath627 , due to ( ii ) and ( [ trt ] ) .
the equality @xmath796 now follows : @xmath808 , and @xmath778 is a function of the norm @xmath570 , so that @xmath809 is @xmath772-orthogonal to @xmath810 and @xmath811 by ( f ) , ( [ hgv].a ) , ( [ ana ] ) , and ( e ) .
also , ( [ lvf ] ) clearly gives holomorphicity of @xmath799 , while closedness and positivity of the form @xmath812 imply that @xmath763 is a khler metric , and @xmath778 has a geodesic @xmath772-gradient @xmath799 , its @xmath772-norm squared @xmath807 being a function of @xmath805 , cf .
( [ trt ] ) and lemma [ ggqft ] .
we have thus proved ( a ) .
next , remark [ tends ] yields ( c ) .
the @xmath223-projectable local sections of @xmath766 are precisely the same as the @xmath164-horizontal lifts of local vector fields tangent to @xmath132 , and their local flows act as @xmath164-parallel transports between the fibres . as the the submanifold metrics of the fibres are defined by ( [ hgv].a ) ( [ hgv].b )
, this last action consists by ( f ) of isometries which , being linear , also preserve the vertical vector field @xmath811 with ( [ lvf ] ) . hence 1 .
@xmath813 commutes with all local @xmath164-horizontal lifts @xmath235 , and , at the same time , applying remark [ tglvs ] to any such @xmath814 we obtain ( b ) . finally , by ( g ) and remark [ dvwwp ] , the left - hand side of ( [ gsw ] ) equals the @xmath799-directional derivative of the right - hand side in ( [ hgv].c ) .
to evaluate the latter , note that only the factor @xmath815 in the second term needs to be differentiated , as the first term and the remaining factor of the second one are constant along @xmath811 ( due to constancy along @xmath811 of @xmath816 , obvious from ( [ lvf ] ) ) .
now ( e ) implies ( [ gsw ] ) , completing the proof . a special khler - ricci potential @xcite on a khler manifold @xmath86 is any nonconstant function @xmath75 such that @xmath85 is real - holomorphic , while , at points where @xmath375 , all nonzero vectors orthogonal to @xmath56 and @xmath817 are eigenvectors of both @xmath818 and the ricci tensor , with @xmath819 as in ( [ nwt ] ) .
we then call @xmath8 an skrp triple .
all skrp triples @xmath8 are geodesgradient khler triples , due to their easily - verified property ( * ? ? ?
* remark 7.1 ) that @xmath405 , wherever nonzero , is an eigenvector of @xmath818 .
( [ nvv ] ) .
compact skrp triples @xmath8 have been classified in ( * ? ? ?
* theorem 16.3 ) .
they are divided into class 1 , in which @xmath44 is the total space of a holomorphic @xmath310 bundle , and class 2 , with @xmath44 biholomorphic to @xmath820 for @xmath821 .
[ cltwo ] up to isomorphisms , in the sense of definition [ ggktr ] , compact skrp triples of class 2 are the same as cp triples constructed using ( [ dta].ii ) with @xmath822 .
see ( * ? ? ?
* remark 6.2 ) .
( note that the case @xmath823 in ( [ dta].ii ) obviously leads to the same isomorphism type . ) in ( i ) above , @xmath824 , which allows the possibility of a one - point base manifold @xmath825 , so that , as a complex manifold , @xmath131 is a complex vector space , namely , the fibre @xmath773 . according to @xcite , under the standard identification ( [ inc ] ) for @xmath826 ,
both @xmath763 and @xmath778 then can be extended to the projective space @xmath827 , giving rise to a class 2 skrp triple @xmath828 , where @xmath829 .
[ xplic ] the skrp triples @xmath830 just mentioned , with @xmath829 , represent all isomorphism types of compact skrp triples of class 2 .
such types include all compact geodesgradient khler triples of complex dimension @xmath336 .
for the first part , see ( * ? ? ?
* remark 6.2 ) .
the final clause is in turn immediate from remark [ cpone ] and lemma [ cltwo ] .
[ fbstu]as a consequence of the second part of remark [ trivl ] , for @xmath831 as in theorem [ ehggk ] , every fibre @xmath773 is the underlying complex manifold of a geodesic - gradient khler triple , realizing a special case of theorem [ ehggk ] : that of a one - point base manifold @xmath832 .
its projective compactification @xmath833 constitutes , for reasons mentioned above , the underlying complex manifold of an skrp triple of class 2 .
the resulting submanifold metric on the complement of @xmath773 in @xmath833 ( that is , on the projective hyperplane at infinity , identified via ( [ inc ] ) with @xmath834 ) equals @xmath835 times the fubini - study metric associated as in remark [ fbstm ] with @xmath356 .
namely , let @xmath836 have @xmath837 and @xmath838 .
the curve @xmath839 of vectors @xmath840 tangent to @xmath773 at the points @xmath841 , satisfies , in view of ( [ hgv].b ) and remark [ tends ] , the limit relation @xmath842 as @xmath843 . at the same time , @xmath844 ( or , the tangent vector @xmath845 ) tends , as @xmath843 , to the point @xmath846 of @xmath847 , identified with @xmath848 or , respectively , to the vector tangent to @xmath834 at @xmath849 which is the image of @xmath521 under 1 . the differential of the projection @xmath476 restricted to the unit sphere of @xmath356 .
the claim about the tangent vectors , which clearly implies our assertion , can be justified as follows .
the vector @xmath840 equals @xmath850 ( notation preceding remark [ jcobi ] ) with @xmath851 , so that @xmath852 and , setting @xmath853^{-\nnh1\nh/2}(1,x(t , s ) ) \in\bbc\times\nnh\nr\hskip-2.4pt_y\w$ ] , we get @xmath854 for the direct - sum euclidean norm . identifying @xmath855 with @xmath856 , we see that @xmath845 , treated as tangent to @xmath833 at @xmath857 , is the image , under the analog of ( @xmath858 ) for @xmath859 , of the vector @xmath860 tangent to the unit sphere of @xmath859 at the point @xmath861 and having the required limit @xmath862 as @xmath843 , which we identify with @xmath532 .
[ unnec]the construction summarized in theorem [ ehggk ] has an obvious generalization , arising when , in ( [ pbd ] ) , @xmath863 is assumed to be only defined on the half - open interval @xmath864\smallsetminus\{\tmp\}$ ] , and @xmath865 at @xmath866 just for one fixed sign @xmath35 .
our discussion focuses on a narrower case since this is the case relevant to the study of compact geodesic - gradient khler triples .
throughout this section @xmath8 is a fixed geodesgradient khler triple ( definition [ ggktr ] ) .
we use the symbols @xmath867 for the complex - structure tensor @xmath868 of the underlying complex manifold @xmath29 , the gradient @xmath405 , its @xmath727-image @xmath869 , the open set @xmath404 where @xmath375 , the function @xmath373 on @xmath404 with ( [ nvv ] ) , the function @xmath870 on @xmath29 , the distribution @xmath871 on @xmath378 , its orthogonal complement , as well as the endomorphisms @xmath295 and @xmath872 of @xmath74 , cf .
( [ nwt ] ) . under the above hypotheses ,
@xmath873\,\mathrm{\ for\ any\ vector\ field\ } \,w\hh,\\ \mathrm{d)}&\dv\,\mathrm{\ is\ self}\hyp\mathrm{ad\-joint\ and\ } \,j,\du\,\mathrm{\ are\ skew}\hyp\mathrm{ad\-joint\ at\ every\ point\ of\ } \,\mf\hh,\\ \mathrm{e)}&g([w , w\hh'],u)=-\nh2\hh g(\du w , w\hh'\hh)\,\mathrm{\ for\ any\ local\ sections\ } \,w , w\hh'\mathrm{\ of\ } \,\mathcal{v}^\perp\hh,\\ \mathrm{f)}&\nabla\hskip-3pt_{v}\w v=\psi\hskip.4ptv = -\nabla\hskip-3pt_u\w u\,\mathrm{\ and\ } \,\nabla\hskip-3pt_u\w v=\nabla\hskip-3pt_{v}\w u = \psi\hskip.4ptu\,\mathrm{\ everywhere\ in\ } \,\mf',\\ \mathrm{g)}&q\,\mathrm{\ is,\ locally\ in\ } \,\mf'\nh,\mathrm{\ a\ function\ of\ } \,\vt\hh,\mathrm{\ and\ } \,2\psi=\hs d\hh q / d\vt\hh,\\ \mathrm{h)}&j,\hs\dv,\hs\du\,\mathrm{\
and\ the\ local\ flows\ of\ } \,u\,\mathrm{\ and\ } \,v\,\mathrm{\ leave\ } \,\mathcal{v}\,\mathrm{\ and\ } \,\mathcal{v}^\perp\mathrm{\ invariant . } \end{array}\ ] ] in ( [ loc].c ) , @xmath280 denotes the curvature tensor of @xmath62 , and the notation of ( [ nwt ] ) is used .
in fact , holomorphicity of @xmath56 ( cf . definition [ ggktr ] ) combined with ( [ ajs ] ) ( [ hol ] ) gives ( [ loc].a ) , @xmath70 being holomorphic due to ( [ hol ] ) , as @xmath874 commutes with @xmath727 .
next , ( [ loc].b ) follows from ( [ kil ] ) and the lie - bracket equality @xmath875=\nabla\hskip-3pt_u\w v-\nabla\hskip-3pt_{v}\w u=\dv\nh u-\du v=\dv\nh u-\dv\nh j\nh v=0 $ ] , obvious in view of ( [ loc].a ) , while ( [ loc].c ) ( or , ( [ loc].d ) ) is a direct consequence of ( [ scd ] ) and ( [ loc].a ) or , respectively , of of ( [ loc].b ) combined with the fact that @xmath56 is a gradient .
we now obtain ( [ loc].e ) from ( [ loc].d ) , noting that @xmath876 . on the other hand , ( [ loc].b ) , ( [ loc].a ) and ( [ nvv ] ) yield @xmath877 and so @xmath878 , establishing ( [ loc].f ) , while lemma [ ggqft ] , ( [ tnd ] ) and ( [ loc].f ) imply ( [ loc].g ) . that @xmath879 all leave @xmath871 invariant is clear as @xmath880 and @xmath881 while , by ( [ loc].f ) , @xmath882 are sections of @xmath883 . the same conclusion for @xmath884 is now immediate from ( [ loc].d ) . by ( [ loc].b ) ,
the local flows of @xmath56 and @xmath70 preserve @xmath66 and @xmath885 .
the @xmath195-invariance of @xmath884 now follows from ( [ loc].b ) .
finally , let @xmath81 be a section of @xmath886 .
writing @xmath346 for @xmath62 , we get @xmath887,v\ra=\la\nabla\hskip-3pt_{v}\w w-\nabla\hskip-3pt_{w}\w v , v\ra = -\hh\la w,\nabla\hskip-3pt_{v}\w v\ra-\la\dv\nh w , v\ra=-\la\dv\nh w , v\ra = -\la w,\dv\nh v\ra=0 $ ] , cf .
( [ loc].d ) and ( [ loc].f ) . similarly ,
@xmath887,u\ra=\la\nabla\hskip-3pt_{v}\w w-\nabla\hskip-3pt_{w}\w v , u\ra = -\hh\la w,\nabla\hskip-3pt_{v}\w u\ra-\la\dv\nh w , u\ra=-\la\dv\nh w , u\ra = -\la w,\dv\nh u\ra=0 $ ] .
thus , @xmath139\,$ ] is a section of @xmath884 as well . in view of remark [ prjct ] , this completes the proof of ( [ loc].h ) . for easy reference , note that , by ( [ loc].a ) ( [ loc].b ) , @xmath888 [ dvgww ] under the assumptions preceding ( [ loc ] ) , on @xmath34 , 1 . the distribution @xmath871 is integrable and has totally geodesic leaves , 2 .
a local section of @xmath889 is projectable along @xmath890 if and only if it commutes with @xmath891 and @xmath892 , 3 . if local sections @xmath81 and @xmath893 of @xmath894 commute with @xmath70 and @xmath80 , then @xmath895\,=\,2\hh g(\dv\nh w , w\hh'\hh)\hh , & \mathrm{ii)}&d_v\w[\hs g(\dv\nh w , w\hh'\hh)]\,=\,2\psi\hh g(\dv\nh w , w\hh'\hh)\hh,\\ \mathrm{iii)}&d_v\w[q^{-\nnh1}g(\dv\nh w , w\hh'\hh)]\,=\,0\hh , & \end{array}\ ] ] 4 .
@xmath896 and @xmath897\nh=\nh d_u\w[\hs g(\dv\nh w , w\hh'\hh)]\nh = \nh d_u\w q\nnh=0\,$ ] for any @xmath63 as in ( c ) , 5 .
@xmath898\hh w\,=\,2(\psi\hs-\hs\dv)\dv\hh w\hs\,$ ] whenever @xmath81 is a local section of @xmath899 .
assertions ( a ) ( b ) are obvious from ( [ loc].b ) and , respectively , remark [ liebr ] combined with ( [ loc].h ) .
next , let @xmath900 . since @xmath901 and @xmath902 act on functions as @xmath903 and @xmath904 , ( [ lvg ] ) implies ( [ dvg].i ) , and @xmath905=0\,$ ] as @xmath906 by ( [ loc].b ) . for similar reasons , @xmath907 = \lie\hskip-.5pt_u\w[\hs g(\dv\nh w , w\hh'\hh)]=0 $ ] .
( namely , ( [ loc].c ) gives @xmath908 , so that ( [ loc].a ) and ( [ lie ] ) , with @xmath909 rather than @xmath910 , yield @xmath911 . ) on the other hand , by ( [ gvv ] ) , @xmath912 . now ( [ tnd ] ) , ( [ loc].f ) and ( [ loc].b ) imply that @xmath913 and @xmath914 , establishing ( d ) . using ( [ loc].a )
we get @xmath915 which , by ( [ loc].e ) , is nothing else than @xmath916,u)/2 $ ] .
hence @xmath917=2\lie\hskip-.5pt_v\w[\hs g(\dv\nh w , w\hh'\hh ) ] = -\lie\hskip-.5pt_v\w[\hs g(\hn u,[w , j\nh w\hh ' ] ) ) ] = -\hh[\nh\lie\hskip-.5pt_v\w g](\hn u,[w , j\nh w\hh']))$ ] .
( our assumption that @xmath84 gives @xmath918 , as @xmath56 is holomorphic , which in turn yields @xmath919=0 $ ] , while @xmath920 , cf .
( [ loc].b ) . ) from ( [ lvg ] ) , ( [ loc].f ) and ( [ loc].a ) we now obtain @xmath921=-\hh[\nh\lie\hskip-.5pt_v\w g](\hn u,[w , j\nh w\hh']))/2 = -\hh g(\dv\nh u,[w , j\nh w\hh'])=-\nh2g(\psi u,[w , j\nh w\hh ' ] ) = 2\psi\hh g(\du w,\nh j\nh w\hh'\hh)\nh=\nh-\nnh2\psi\hh g(j\nnh\du w , w\hh'\hh)\nh = 2\psi\hh g(\dv\nh w , w\hh'\hh)$ ] , that is , ( [ dvg].ii ) , which , since @xmath922 by ( d ) , also proves ( [ dvg].iii ) .
finally , ( [ loc].h ) and the equality @xmath923 $ ] , cf .
( [ loc].c ) , combined with ( a ) , imply that @xmath924 leaves @xmath884 invariant . to obtain ( e )
, it now suffices to show that @xmath925\hh w-(2\psi\hs-\hs\dv)\dv\hh w\,$ ] is orthogonal to @xmath893 for any local sections @xmath63 of @xmath899 .
we are free to assume here that @xmath926 ( due to self - adjointness of @xmath927 ) and that @xmath81 commutes with @xmath70 and @xmath56 ( see ( b ) ) .
differentiation by parts gives , by ( [ dvg].iii ) and ( [ loc].d ) , @xmath928\hh w , w)=d_v\w[\hs g(\dv\nh w , w ) ] -g(\dv\nabla\hskip-3pt_{v}\w w , w)-g(\dv\nh w,\nabla\hskip-3pt_{v}\w w ) = 2\psi\hh g(\dv\nh w , w)-2g(\dv\nh w,\dv\nh w)$ ] , as required , with @xmath929 since @xmath139=0 $ ] .
in addition to using the assumptions and notations of section [ lp ] , we now let @xmath930 stand for the underlying dimensional manifold of a fixed maximal integral curve of @xmath56 in @xmath34 . we restrict the objects in ( [ sym ] ) to @xmath931 without changing the notation , and select a unit - speed parametrization @xmath932 of the geodesic @xmath930 such that @xmath933 as an obvious consequence of ( [ dxe ] ) , ( [ dvt ] ) and lemma [ dvgww](d ) , @xmath934 any constant @xmath935\cup\{\infty\}$ ] , where @xmath936 is the range of @xmath43 on @xmath930 , gives rise to the function @xmath937 defined by @xmath938\hh,\ ] ] the convention being that @xmath939 is identically zero when @xmath940 .
we denote by @xmath941 the set of all @xmath899-valued vector fields @xmath942 along @xmath931 satisfying the equation @xmath943 of particular interest to us are @xmath944 such that @xmath945\cup\{\infty\}\,\,\mathrm{\ and\ } \,\,\mathcal{w}[\zx]\ne\{0\}\hh,\mathrm{\ \
where}\\ \mathrm{b ) } & \mathcal{w}[\zx]\,=\,\{w\in\mathcal{w}:\dv\nh w=\lambda_\zx\w\nh w\}\hh . \end{array}\ ] ] about projectability along @xmath234 in ( i ) below , see remark [ pralg ] and lemma [ dvgww](a ) .
[ jacob ] under the above hypotheses , the following conclusions hold . 1 .
@xmath946-valued solutions @xmath81 to ( [ nxw ] ) are nothing else than restrictions to @xmath947 of those local sections of @xmath889 with domains containing @xmath947 which are projectable along @xmath228 .
all @xmath81 as in ( i ) , that is , all elements of @xmath948 , are jacobi fields along @xmath949 .
3 . every vector in @xmath950 equals @xmath951 for some unique @xmath952 .
4 . @xmath953 is a complex vector space of complex dimension @xmath954 , and the direct sum of all @xmath955\,$ ] for @xmath956 in ( [ wze].a ) , with @xmath957 serving as the multiplication by @xmath958 .
a function @xmath959 on the parameter interval of @xmath932 satisfies the equation @xmath960 , with @xmath961 evaluated at @xmath962 , if and only if @xmath963 , cf .
( [ lgm ] ) , for some @xmath935\cup\{\infty\}\,$ ] and all @xmath540 . 6 . at any @xmath964 , the eigenvalues of @xmath965 , cf .
( [ loc].h ) , are precisely the values @xmath966 for all @xmath944 in ( [ wze].a ) .
the eigenspace of @xmath965 corresponding to @xmath966 is @xmath967\}$ ] .
@xmath968 on @xmath969 for sections @xmath970 of @xmath886
. 8 . if @xmath971 is bounded , then @xmath972 on @xmath886 .
any @xmath81 as in the second line of ( i ) , restricted to @xmath949 , becomes both a jacobi field ( by lemmas [ jcbfl ] and [ dvgww](b ) ) and a @xmath899-valued solution to ( [ nxw ] ) ( since @xmath927 , so that ( [ dxe ] ) and lemma [ dvgww](b ) give @xmath973 ) . with @xmath931
replaced by suitable shorter subgeodesics covering all points of @xmath949 , the inclusion just established between the two vector spaces appearing in ( i ) is actually an equality : in either class , the vector field in question is uniquely determined by its initial value at any given point @xmath974 .
this proves ( i )
( ii ) as well as ( iii ) ( iv ) , the latter in view of the fact that @xmath975 , cf .
( [ loc].a ) . for a @xmath976 function @xmath977 defined on the parameter interval of @xmath194
, one has @xmath978 if and only if either @xmath979 identically , or @xmath980 everywhere and the function @xmath981 characterized by @xmath982 is constant .
( in fact , the either - or claim about vanishing of @xmath977 is due to uniqueness of solutions of initial - value problems , while ( [ dtq ] ) yields @xmath983 $ ] . ) now ( v ) easily follows , all nonzero initial conditions for ( [ dlt ] ) at fixed @xmath191 being realized by suitably chosen constants @xmath984 ( and @xmath979 satisfying ( v ) with @xmath940 ) . on the other hand , from ( [ dxe ] ) and lemma [ dvgww](e ) , @xmath985\hh w\ , = \,2q^{-\nnh1\nh/2}(\psi\hs-\hs\dv)\dv\hh w\hh , \hskip9pt\mathrm{if\ } \,w\,\mathrm{\ is\ a\ } \,\mathcal{v}^\perp\nh\hyp\mathrm{val\-ued\ vector\ field\ along\ } \,\ic.\ ] ] next , we fix @xmath986 and express any prescribed eigenvalue - eigenvector pair for @xmath965 as @xmath966 and @xmath987 , with some unique @xmath935\cup\{\infty\}$ ] and @xmath952 . by ( v ) , @xmath988 satisfies ( [ dlt ] ) , so that , in view of ( [ nxw ] ) and ( [ nds ] ) , the vector field @xmath989 is a solution of the linear homogeneous differential equation @xmath990 . since @xmath190 vanishes at @xmath991 , it must vanish identically , which establishes ( vi ) .
now let @xmath952 .
as @xmath992 ( see the lines following ( [ dlt ] ) ) , the jacobi equation and ( [ nxw ] ) give , by ( ii ) and ( [ nds ] ) , @xmath993 = q^{-\nnh1}(\psi\hs-\hs\dv)\dv\hh w$ ] , that is , @xmath994 , the second equality in ( vii ) . also , lemma [ dvgww](e ) , ( [ loc].c ) and ( [ rcm ] ) yield @xmath995\hh w = -\nh j[r(\hn u , v)\hh w]=-\hh r(\hn u , v)\hh j\nh w = r(\hn v , u)\hh j\nh w = r(\hn v , j\nh v)\hh j\nh w$ ] , the last equality in ( vii ) . combining the two relations , and repeatedly using ( [ rcm ] ) ,
we get @xmath996 , that is , @xmath997 . thus , from the bianchi identity , @xmath998 , which proves ( vii ) .
finally , ( viii ) is an easy consequence of ( vi ) and ( [ loc].d ) .
let @xmath8 be a fixed geodesgradient khler triple ( definition [ ggktr ] ) .
we use the notation of ( [ sym ] ) , ( [ tmm ] ) and in ( i ) below the terminology of remark [ cifot ] .
[ ascdt]according to ( * ? ? ?
* lemmas 11.1 , 11.2 and remark 2.1 ) , the following holds for any @xmath8 as above with compact @xmath29 , the objects ( [ tmm ] ) , and @xmath405 . 1 . @xmath999 is a @xmath92 function of @xmath5 , leading to data @xmath559 with ( [ pbd ] ) .
the flow of the killing vector field @xmath1000 is periodic .
3 . @xmath1001 are ( connected ) totally geodesic compact complex submanifolds of @xmath29 .
@xmath1002 is the zero set of @xmath80 , that is , the set of critical points of @xmath5 .
( conclusion ( iv ) is a special case of a result due to wang ( * ? ? ?
* lemma 3 ) . ) furthermore , restricting @xmath863 in ( i ) to the open interval @xmath1003 we have @xmath1004 @xmath1005 being the function with ( [ nvv ] ) on the open set @xmath404 on which @xmath1006 ( so that @xmath373 is also a @xmath92 function of @xmath5 ) .
this is clear as ( [ tnd ] ) and ( [ nvv ] ) give @xmath1007 on @xmath1008 . finally , by (
* lemma 1 ) ( see also ( * ? ? ?
* example 8.1 and lemma 8.4(iv ) ) ) , @xmath1009 [ prcrm]under the assumptions of remark [ ascdt ] , for @xmath1010 as in remark [ ascdt](i ) , @xmath1011 is the unique nonzero eigenvalue of the hessian of @xmath43 ( that is , of @xmath927 ) at any critical point @xmath145 .
the @xmath1012-eigenspace of @xmath1013 is the normal space @xmath1014 , and @xmath1015 ( which thus constitutes the @xmath147-eigenspace of @xmath1013 unless @xmath1016 ) .
in fact , as @xmath43 is a morse - bott function ( * ? ? ?
* example 8.1 ) , applying ( * ? ? ?
* lemma 8.4(i ) ) we see that @xmath1017 is the eigenspace of @xmath1013 for its unique nonzero eigenvalue , and so @xmath1015 in view of self - adjointness of @xmath1013 . that the nonzero eigenvalue equals @xmath1011 is obvious from ( [ dqt ] ) , since ( [ nvv ] ) amounts to @xmath1018 .
* theorem 1.3 ) . still assuming compactness of a geodesgradient khler triple @xmath380 ,
let @xmath1019 be the bundle of radius @xmath573 normal open disks around the zero section in the normal bundle @xmath1020 , with @xmath573 characterized by ( [ int ] ) . according to ( * ? ? ? * lemma 10.3 ) , @xmath573 is then the distance between @xmath3 and @xmath1021 , while , with @xmath149 as in remark [ nexpm ] , @xmath1022 cf .
@xcite , ( * ? ? ?
* lemma 2 ) , ( * ? ? ?
* theorem 1.1 ) .
its inverse @xmath1023 , composed with the projection @xmath1024 , yields a new disk - bundle projection @xmath1025 [ ptnrs]clearly , @xmath1026 is the normal - bundle projection @xmath1027 . also , according to the lines preceding ( [ dbp ] ) , @xmath1028 which implies ( * ? ? ? * remark 4.6 , example 8.1 and theorem 10.2(iii)(vi ) ) that @xmath15 sends every @xmath12 to the unique point nearest @xmath13 in @xmath14 . in the next lemma , by a leaf we mean as usual a maximal integral manifold .
[ vsbkr ] under the above hypotheses , @xmath1029 for the integrable distribution @xmath871 on @xmath10 , cf .
lemma [ dvgww](a ) and remark [ ascdt](iv ) . if @xmath1030 is a unit vector normal to @xmath28 at a point @xmath1031 , then , with @xmath573 as in ( [ nex ] ) , 1 .
the punctured radius @xmath573 disk @xmath1032 in @xmath1017 is mapped by @xmath1033 diffeomorphically onto a leaf @xmath1034 of @xmath228 .
furthermore , every leaf @xmath1035 of @xmath234 has the following properties . 1 .
the closure of @xmath1034 in @xmath44 is a totally geodesic complex submanifold , biholomorphic to @xmath310 and equal to @xmath1036 , where @xmath1037 are such that @xmath1038 .
2 . the leaf @xmath1034 arises from ( a ) for some unit normal vector @xmath50 at the point @xmath1039 corresponding to @xmath1040 as in ( b ) , and then @xmath1041 @xmath1042 being the normal exponential mapping @xmath1043 .
let us fix @xmath1044 and @xmath1045 as in ( [ img ] ) .
due to remark [ ascdt](iv ) , the killing field @xmath1000 vanishes along @xmath14 , so that its infinitesimal flow at @xmath1046 preserves both @xmath1047 and @xmath1014 .
the images of @xmath1045 under the flow transformations of @xmath70 thus are geodesic segments normal to @xmath28 emanating from @xmath1046 and , as a consequence of ( [ img ] ) , @xmath15 maps them all onto @xmath832 .
in other words , the union of such segments , with the point @xmath1031 removed , is simultaneously a subset of the @xmath1048-preimage of @xmath1046 as well as according to ( [ vtg ] ) , ( [ gvv ] ) and parts ( ii ) , ( iv ) of remark [ ascdt ] a surface embedded in @xmath378 .
this surface is , due to its very definition and ( [ vtg ] ) , tangent to both @xmath70 and @xmath56 which , in view of ( [ nex ] ) , yields ( a ) ; note that , by ( [ loc].a ) and remark [ prcrm ] , the orbit of @xmath50 under the flow of @xmath872 at @xmath1046 consists of all unit complex multiples of @xmath1030 . what we just observed about the orbit of @xmath50 clearly ensures smoothness of the closure of the leaf at @xmath1031 .
by ( [ nex ] ) and ( [ vtg ] ) , the union of @xmath1045 and its analog for the same point @xmath13 and the other projection @xmath21 is a length @xmath573 geodesic segment joining @xmath1049 to its other endpoint @xmath1050 .
the above discussion of the images of such a segment under the flow of @xmath70 applies equally well to @xmath1051 , so that ( b )
( c ) follow from lemma [ dvgww](a ) and the fact that @xmath1052 was arbitrary .
[ ppmgc]let @xmath8 be a grassmannian or cp triple , constructed as in section [ eg ] from some data ( [ dta].i ) or ( [ dta].ii ) .
we use the notation of ( [ sym ] ) and ( [ dbp ] ) . 1 .
we already know that the critical manifolds @xmath28 of @xmath43 are given by ( [ spm ] ) .
2 . in the case of ( [ spm].c ) ( or , ( [ spm].b ) and ( [ spm].d ) ) , @xmath15 acts on @xmath449 as the orthogonal projection into @xmath434 ( or , respectively , into @xmath1053 ) 3 .
when @xmath28 has the form ( [ spm].a ) , @xmath15 sends @xmath449 to @xmath1054 .
the leaf of @xmath234 through any @xmath1055 consists 1 . for ( [ dta].i ) of all @xmath1056 , where @xmath1057 and @xmath1058 is any line in the plane @xmath1059 other than the lines @xmath434 and @xmath1060 themselves , 2 . for ( [ dta].ii ) of all lines other than @xmath1061 and @xmath1062 in the plane @xmath1063 , where @xmath1061 and @xmath1062 denote the orthogonal projections of @xmath449 into @xmath434 and @xmath473
namely , in both cases , let @xmath1064 be the complex lie group of all complex - linear automorphisms of @xmath417 preserving both @xmath434 and @xmath473 .
the obvious action of @xmath1064 on the stiefel manifold @xmath533 ( see remark [ grass ] ) descends to a holomorphic action on @xmath1065 , which becomes one on @xmath422 when @xmath419 .
the elements of the center of @xmath1066 , restricted to both subspaces @xmath434 and @xmath473 , are complex multiples of @xmath439 , and the action of the center on @xmath421 includes the circle subgroup @xmath415 of @xmath399 generated by the killing field @xmath195 , mentioned in the lines following ( [ dta ] ) .
holomorphicity of the action implies that the flow of the gradient @xmath405 , related to @xmath70 via @xmath869 , also consists of transformations of @xmath421 arising from the action of the center , and for dimensional reasons
the orbits of the center coincide with the leaves of @xmath885 .
this easily gives ( d ) .
now ( b ) ( c ) follow : by lemma [ vsbkr](b ) , the two @xmath1067-images of any leaf of @xmath234 are the two points that , added to the leaf , yield its closure .
as @xmath1029 ( lemma [ vsbkr ] ) , we may define vector subbundles @xmath1068 of @xmath1069 by @xmath1070 [ dcomp ] given a geodesic - gradient khler triple @xmath8 with compact @xmath29 , @xmath1071 be defined as in ( [ sym ] ) , ( [ tmm ] ) and ( [ hpm ] ) .
then the bundle endomorphism @xmath1072 of @xmath74 , restricted to @xmath899 , has constant rank on @xmath378 , while @xmath1073\hh,\ ] ] and some subbundle @xmath271 of @xmath1069 yields an @xmath1074-invariant complex orthogonal decomposition @xmath1075 furthermore , for any @xmath1076 chosen as at the beginning of section [ hj],the closure of @xmath930 in @xmath44 admits a unit - speed @xmath45 parametrization @xmath1077\ni t\mapsto x(t)\,$ ] which , restricted to @xmath1078 , is a parametrization of @xmath947 satisfying ( [ dxe ] ) along with the following conditions . 1 .
the endpoint @xmath1079 lies in @xmath14 , and @xmath1080 is normal to @xmath28 at @xmath1081 .
every solution @xmath1082 of ( [ nxw ] ) along @xmath947 has a @xmath45 extension to @xmath1077\,$ ] such that @xmath1083=w(t_\pm\w)\,$ ] whenever @xmath1084 .
the bundle projection @xmath11 is holomorphic .
4 . if @xmath1085 $ ] , cf .
( [ wze ] ) , then , in ( b ) , @xmath1086 , and @xmath1087(t_\pm\w)\,$ ] is normal to @xmath28 at @xmath1079 as well as orthogonal to @xmath1080 and @xmath1088 .
if @xmath81 lies in the direct sum of spaces @xmath1089\ne\{0\}\,$ ] with @xmath1090 , for a fixed sign @xmath35 , then @xmath1091 is tangent to @xmath28 at @xmath1092 , and @xmath1087(t_\pm\w)=0 $ ] . 6 .
whenever @xmath1093 and @xmath991 , the assignment @xmath1094(t_\pm\w))$ ] , with @xmath81 as in ( b ) , is a @xmath478-linear isomorphism @xmath1095 , where @xmath1039 and @xmath1096 denotes the orthogonal complement of @xmath1097 in @xmath1014 . at the same time , @xmath951 then equals the image , under the differential of the normal exponential mapping @xmath1098 at @xmath1099 given by @xmath1100 and @xmath1101 , of the vector tangent to @xmath1102 at @xmath1103 which equals the sum of the vertical vector @xmath1104(t_\pm\w)\,$ ] and the @xmath1105-horizontal lift of @xmath1091 to @xmath1106 , for the normal connection @xmath1107 in @xmath1108 .
similarly , @xmath1109 , for @xmath869 , is the image , under the differential of @xmath149 at @xmath1106 , of the vertical vector @xmath1110 .
7 . for any @xmath1111 ,
the function @xmath1112 is constant on @xmath947 and the restriction of @xmath1113 to @xmath947 is an affine function of @xmath1114 with the derivative @xmath1115/\nh d\vt=2\hh q^{-\nnh1}g(\dv\nh w , w\hh'\hh)$ ] .
explicitly , in ( g ) , with @xmath1010 as in remark [ ascdt](i ) , either sign @xmath35 , and @xmath1116 , 1 .
@xmath1117 if @xmath1118\,$ ] and @xmath1119 , 2 .
@xmath1120 if @xmath63 both satisfy the assumption of ( e ) , 3 .
@xmath1121_\pm\w , [ \nabla\hskip-3pt_{\dot x}\w w\hh'\hh]_\pm\w)\,$ ] if @xmath1122 $ ] , where the subscript @xmath594 next to @xmath1123 or @xmath1124 represents their evaluation at @xmath1125 .
[ bthex]since @xmath1126 and @xmath1127 , applying @xmath1128 to the right - hand side in ( h1 ) , or ( h2 ) , or ( h3 ) , we get the three values @xmath1129_\pm\w , \nh[\nabla\hskip-3pt_{\dot x}\w w\hh'\hh]_\pm\w).\ ] ] as a consequence of parts ( g ) ( h ) of theorem [ dcomp ] , this triple provides the three expressions for @xmath1130 in the cases ( h1 ) , ( h2 ) and ( h3 ) , respectively .
note that the three different formulae for @xmath1113 in ( h1 ) , ( h2 ) and with the reversed sign in ( h3 ) , are all simultaneously valid when @xmath1131 $ ] .
[ inequ]under the assumptions of theorem [ dcomp ] , 1 . the relation @xmath1132 in ( f ) clearly gives @xmath1133 in ( h2 ) , 2 . by ( d )
( f ) , the images under the differential of @xmath149 of vertical ( or , horizontal ) vectors tangent to @xmath1102 at the point @xmath1103 appearing in ( f ) have the form 1 .
@xmath1134 for @xmath81 satisfying the hypothesis of ( d ) ( or , respectively , of ( e ) ) , 3 .
the differential of @xmath1135 at any @xmath1052 maps the summands @xmath1136 and @xmath1137 in ( [ tme ] ) isomorphically onto the images @xmath1138 and @xmath1139 , orthogonal to each other in @xmath1047 for @xmath1140 , 4 .
one has @xmath1141 whenever @xmath1142 and @xmath1143 at any @xmath1144 , while @xmath1140 , 5 .
( [ tmm ] ) and ( a ) imply the inequality of theorem [ jacob](viii ) everywhere in @xmath34 .
only ( iii ) and ( iv ) require further explanations . for ( iii ) , @xmath1145 is injective on the space @xmath1146 , orthogonal , by ( [ hpm ] ) and ( [ tme ] ) , to its kernel @xmath1147 .
orthogonality in ( [ hpm ] ) also shows , via ( [ hvk ] ) , ( [ wze].b ) and ( [ loc].d ) , that vectors in @xmath1136 ( or , in @xmath1148 ) have the form ( @xmath858 ) with @xmath991 , cf .
( f ) , and the former remain orhogonal to the latter as @xmath191 varies , leading to ( iii ) as a consequence of the final clause of ( b ) . assertion ( iv ) is nothing else than ( h1 ) for @xmath1149 at @xmath991 , cf .
( f ) , where @xmath1150 and @xmath1151 by ( [ img ] ) and ( b ) . [ holom]as another immediate consequence of theorem [ dcomp ] , the assignment @xmath1152 defines a holomorphic section of the bundle over @xmath404 arising via the pullback under @xmath15 from @xmath1153 , for a suitable integer @xmath32 . here
@xmath1154 is the grassmannian bundle with the fibres @xmath1155 , @xmath1049 ( cf .
section [ eg ] ) , holomorphicity and the equality @xmath1156 are clear from theorem [ dcomp](c ) ( which also implies , due to ( [ hpm ] ) , that @xmath1157 is a holomorphic subbundle of @xmath1069 ) and ( [ tme ] ) ( which , combined with ( [ hpm ] ) , ensures constancy of the dimension @xmath32 of the spaces @xmath1138 ) .
we begin by establishing ( a ) - ( f ) under the stated assumptions about @xmath949 .
let @xmath1158 be a parametrization of @xmath947 with ( [ dxe ] ) .
as @xmath43 then is clearly an increasing function of @xmath540 , it has some limits @xmath1159 as @xmath1160 , finite due to boundedness of @xmath5 .
the length of @xmath947 obviously equals the integral of @xmath1161 over @xmath1162 , and so it is finite in view of ( [ int ] ) .
this implies the existence of limits @xmath1163 of @xmath1164 as @xmath1160 .
furthermore , each @xmath1163 lies in @xmath28 since , if one @xmath1163 did not , remark [ ascdt](iv ) would yield @xmath1006 at @xmath1165 , contradicting maximality of @xmath949 .
thus , @xmath1077\mapsto x(t)\,$ ] parametrizes the closure of @xmath947 .
next , @xmath1166 is , by ( [ nex ] ) and ( [ vtg ] ) , a disjoint union of maximal integral curves of @xmath80 , each of which has two limit points , one in @xmath4 and one in @xmath1167 , and the corresponding limit directions of the curve are normal to @xmath4 and @xmath1167 . since @xmath947 is one of these curves , ( a ) follows . in ( b ) , a @xmath92 extension to @xmath1077\,$ ] must exist as @xmath81 is a jacobi field ; see theorem [ jacob](ii ) . to obtain ( d ) ( e ) , we fix @xmath1168 $ ] , so that , from ( [ lgm ] ) ( [ wze ] ) , @xmath1169\hh,\hskip22pt \mathrm{ii)}\hskip6pt\nabla\hskip-3pt_{\dot x}\w w\,=\,q^{\hs1\nh/2}\hs w/[2(\vt-\zx ) ] \hh.\ ] ] let @xmath1100 and @xmath1170 . by remark [ ascdt](iv ) ,
@xmath1171 on @xmath28 while , in view of ( a ) and ( [ dqt ] ) , @xmath1172\,$ ] evaluated at @xmath1164 tends to @xmath1173 as @xmath1160 .
if @xmath1174 , ( [ swe].ii ) multiplied by @xmath1175 thus yields @xmath1176 , and the relation @xmath1177\,$ ] for @xmath1178 , obvious from ( [ swe ] ) , implies that @xmath1087(t_\pm\w)$ ] lies in the @xmath1012-eigenspace of @xmath1013 . when @xmath1179 , ( [ swe].i ) and ( [ swe].ii ) give , respectively , @xmath1180 and @xmath1087(t_\pm\w)=0 $ ] .
due to remark [ prcrm ] , this proves ( d ) and ( e ) : orthogonality in
( d ) follows since @xmath81 and @xmath1181 take values in @xmath899 , for @xmath885 ( so that @xmath1182 ) , while @xmath1183 by ( [ dxe ] ) , and @xmath869 . furthermore , the assignment in ( f ) is well - defined , injective , complex - linear and @xmath1184-valued due to parts ( iii ) , ( ii ) , ( iv ) of theorem [ jacob ] and , respectively , ( d ) ( e )
. the first claim of ( f ) thus follows since both spaces have the same dimension .
the second ( or , third ) one is in turn immediate from ( [ dxp ] ) applied , at @xmath1185 , to any @xmath952 , cf .
theorem [ jacob](ii ) ( or , to @xmath1186 ) , with @xmath1187 as in ( f ) , and @xmath190 defined by @xmath1188 .
( that @xmath166 then is a jacobi field along the geodesic @xmath1189 follows from theorem [ jacob](ii ) or , respectively , ( [ loc].b ) and remark [ kiljc ] , while , in the latter case , due to ( [ loc].a ) along with remarks [ ascdt](iv ) and [ prcrm ] , @xmath1190 realizes the initial conditions @xmath1191 at @xmath1192 . ) the remaining equality @xmath1083=w(t_\pm\w)\,$ ] in ( b ) now becomes an obvious consequence of the second part of ( f ) combined with the first line of remark [ ptnrs ] .
this proves ( b ) and , combined with theorem [ jacob](iv ) , implies ( c ) .
next , for @xmath932 as in ( a )
( f ) , any @xmath1084 , a fixed sign @xmath562 , and @xmath991 , theorem [ jacob](iii ) , ( [ hpm ] ) and ( b ) give @xmath1193 . writing any @xmath1194 as @xmath1195 , where @xmath1196\,$ ] and @xmath1197 lies in the direct sum of the spaces @xmath1089\ne\{0\}\,$ ] with @xmath1090 , cf . theorem [ jacob](iv ) , we see that , by ( d ) ( e ) , the isomorphism in ( f ) sends @xmath1198 and @xmath1199 , respectively , to pairs of the form @xmath1200 and @xmath1201 .
thus , @xmath1202 if and only if @xmath1203 , that is , @xmath1085 $ ] . combining theorem [ jacob](vi ) with ( [ lgm ] ) and ( [ wze].b ) , one now obtains ( [ hvk ] ) , so that ( [ hpm ] ) implies the constant - rank assertion preceding ( [ hvk ] ) .
on the other hand , @xmath1204 and @xmath1205 are mutually orthogonal at every @xmath1144 , being , by ( [ hvk ] ) , contained in eigenspaces corresponding to different eigenvalues of the self - adjoint operator @xmath1206 , cf .
( [ loc].d ) , so that ( [ tme ] ) follows .
let @xmath1111 .
constancy of @xmath1112 along @xmath947 trivially follows from ( [ dvg].iii ) and ( [ vtg ] ) , cf . lemma [ dvgww](b ) and parts ( i ) ( ii ) of theorem [ jacob ] . the operators @xmath1128 and @xmath903 acting on functions @xmath1207 are in turn related by @xmath1208 , since ( [ dxe ] ) gives @xmath1209 , while @xmath1210 due to ( [ dtq ] ) .
now ( g ) is immediate from ( [ dvg].ii ) . in ( h ) , all three right - hand sides are affine functions of @xmath43 with the correct values at @xmath1211 ( that is , limits at the endpoint @xmath1092 ) . proving ( h ) is thus reduced by ( g ) to showing that , in each case , @xmath1212 coincides with the derivative of the right - hand side provided by remark [ bthex ] , which even though @xmath96 is constant on @xmath949 , cf . (
g ) will be achieved via evaluating the limit of @xmath96 at @xmath1213 or , equivalently , at @xmath1214 $ ] .
when @xmath1118 $ ] , ( [ wze].b ) and ( [ lgm ] ) imply that @xmath1215 and , consequently , @xmath1216 has the value ( and limit ) @xmath1217 at @xmath1039 , as required in ( h1 ) .
let @xmath63 now satisfy the hypotheses of ( e ) .
consequently , along @xmath1218 , @xmath1219 in fact , @xmath1220 by ( a ) .
next , @xmath1221 is bounded near the endpoint @xmath1046 of @xmath1222 ( and similarly for @xmath1223 ) ; to see this , we may assume that @xmath1168\,$ ] with @xmath1179 , cf .
( e ) , and then ( [ wze].b ) and ( [ lgm ] ) give @xmath1224 , which is bounded as @xmath589 since , due to ( b ) , @xmath81 has a limit at @xmath1211 .
now ( [ ttz ] ) follows . in view of ( [ ttz ] ) and
( b ) , we may now evaluate the limit of @xmath1225 as @xmath1160 using lhpital s rule : it coincides with the limit of @xmath1226/\dot q$ ] . by ( [ loc].c ) for @xmath1124 rather than of @xmath235 , ( [ nxw ] ) , ( [ loc].d ) and ( [ dtq ] ) , this last expression is the sum of two terms , @xmath1227 and @xmath1228 . according to ( [ ttz ] ) and ( [ dqt ] )
, only the first term contributes to the limit and , as it equals @xmath1229 , cf .
( [ gvv ] ) and ( [ vtg ] ) , relation ( [ dqt ] ) yields ( h2 ) .
finally , suppose that @xmath1122 $ ] .
it follows that @xmath1230_\pm\w\hskip13pt \mathrm{at}\hskip6pty\hh,\hskip6pt\mathrm{where}\hskip6pty = y_\pm\w\hh,\ ] ] and analogously for @xmath82 .
namely , @xmath596 and @xmath81 vanish at @xmath1046 ( see ( a ) , ( d ) ) , while @xmath1231 by ( [ dtq ] ) , and so @xmath1087/(q^{\hs1\nh/2})\nnh\dot{\phantom o}\nnh = \psi^{-\nnh1}\nabla\hskip-3pt_{\dot x}\w w$ ] .
lhpital s rule and ( [ dqt ] ) now imply ( [ qoh ] ) .
since @xmath1232_\pm\w\nh = \mp\hh a\hh[\nabla\hskip-3pt_{\dot x}\w w]_\pm\w$ ] by ( d ) and remark [ prcrm ] , assertion ( h3 ) is obvious from ( [ qoh ] ) , completing the proof of theorem [ dcomp ] .
[ dppdm]with the same notations and assumptions as in theorem [ dcomp ] , denoting by @xmath1233 and @xmath1234 the complex fibre dimensions of the subbundles @xmath1235 and @xmath271 of @xmath17 , we have , for @xmath1236 and @xmath2 , @xmath1237 as one sees adding the equalities @xmath1238 and @xmath1239 ( the former due to ( [ dbp ] ) and ( [ hpm ] ) , the latter to ( [ tme ] ) ) .
consequently , @xmath1240 with equality if and only if the distribution @xmath271 in ( [ tme ] ) is dimensional , that is , if @xmath1241 the explicit descriptions of @xmath28 in ( [ spm].c ) ( [ spm].d ) clearly show that @xmath1242
[ realz]for any two functions @xmath863 and @xmath1243 having the properties listed in ( [ pbd ] ) , with the same @xmath585 and @xmath1244 , there must exist an increasing @xmath45 diffeomorphism @xmath864\ni\vt\mapsto\hat\vt\in[\hh\tm,\tp]\,$ ] which realizes @xmath1245\,\mathrm{\ expressed}\\ \mathrm{in\ terms\ of\
the\ two\ dif\-feo\-mor\-phic\-al\-ly}\hyp\mathrm{re\-lat\-ed\ coordinates\ } \,\hat\vt\,\mathrm{\ and\ } \,\vt .
\end{array}\ ] ] such a diffeomorphism is unique up to compositions from the left ( or , right ) with transformations forming the flow of the first ( or , second ) vector field in ( [ qdt ] ) . to see this ,
apply remark [ adtet ] to @xmath1246 and @xmath1247 ( or , @xmath1248 and @xmath1249 ) , obtaining a function @xmath1250 ( or , @xmath1251 ) , unique up to a positive constant factor and vanishing at @xmath1252 ( or , @xmath1253 ) , with @xmath1254 ( or , @xmath1255 ) , the derivative being positive everywhere in @xmath864 $ ] . adjusting the constant factor
, we may require @xmath1256 and @xmath1257 to be increasing diffeomorphisms of @xmath864 $ ] onto the same interval having the endpoint @xmath147 , and then define @xmath1258 by declaring @xmath1259 `` equal to '' @xmath1260 , that is , letting @xmath1258 be @xmath1256 followed by the inverse of @xmath1260 . consequently ,
@xmath1261 on @xmath1262 , which amounts to ( [ qdt ] ) .
the uniqueness clause is obvious : the only self - diffeomorphisms @xmath1263 of @xmath1262 preserving a given vector field without zeros are its flow transformations , since @xmath607 acts on an integral curve as a shift of the parameter .
[ ntrmd ] for the data @xmath585 and @xmath863 related via remark [ ascdt](i ) to a given compact geodesic - gradient khler triple @xmath380 , and any increasing @xmath45 diffeomorphism @xmath864\ni\vt\mapsto\hat\vt\in[\hh\tm,\tp]$ ] , there exists a @xmath45 function @xmath864\ni\vt\mapsto\phi\in\bbr$ ] , unique up to additive constants , such that @xmath1264 . with
@xmath1265 treated , due to their dependence on @xmath5 , as functions on the complex manifold @xmath29 , the formula @xmath1266 then defines another khler metric on @xmath29 , and 1 .
@xmath1267 is a new geodesic - gradient khler triple .
in addition , denoting by @xmath1243 the analog of @xmath1268 arising when remark [ ascdt](i ) is applied to @xmath828 , and by @xmath1269 the @xmath772-gradient of @xmath805 , one has ( [ qdt ] ) and @xmath1270 . as @xmath1271 , where @xmath1272 , our assumption about @xmath1258 gives @xmath1273 and @xmath1274 , leading to the inequalities @xmath1275 note that @xmath1276 exists since , by remark [ smdiv ] and
( [ pbd ] ) , @xmath1277 and @xmath596 are smoothly divisible by @xmath1278 , their quotients being equal at @xmath585 to the value of @xmath1279 and @xmath1280 , respectively , and so , as @xmath1273 , @xmath1281 for the self - adjoint bundle endomorphism @xmath1282 of @xmath284 with @xmath1283 one has @xmath1284\hh,\ ] ] where @xmath1285 are , as usual , given by @xmath405 , @xmath869 , and @xmath927 .
we proceed to prove positivity of @xmath1282 at all points of @xmath29 , considering two separate cases : @xmath1049 and @xmath1286 , cf .
remark [ ascdt](iv ) .
if @xmath1046 lies in either critical manifold @xmath14 , the relations @xmath1287 and @xmath1288 imply positivity of @xmath1289 as a consequence of ( [ fgt ] ) since , by remark [ prcrm ] , any eigenvalue of @xmath1013 must be equal to @xmath174 or @xmath1012 .
on @xmath34 , we use the @xmath1074-invariant decomposition @xmath1290 , cf .
( [ tme ] ) . in view of ( [ loc].f ) ( [ loc].g ) ,
the restriction of @xmath1291 to @xmath871 equals @xmath1292 times @xmath439 . using ( [ cei ] ) and ( [ gvv ] )
we now see that @xmath1282 acts in @xmath234 via multiplication by the function @xmath1293 , which is positive according to ( [ neq ] ) . theorem [ jacob](vi ) states in turn that the eigenvalues of @xmath965 , for @xmath1144 , have the form @xmath966 with ( [ lgm ] ) and ( [ wze].a ) .
writing @xmath1294 instead of their values at @xmath236 , we conclude from ( [ cei ] ) that the corresponding eigenvalues of @xmath1295 are @xmath1296 and so , due to the ( strict ) first inequality of ( [ neq ] ) , they all lie in the interval @xmath1297 .
positivity of @xmath1282 on @xmath234 thus easily follows both when @xmath1298 and when @xmath1299 .
consequently , @xmath763 is a khler metric on @xmath29 , with the khler form @xmath1300 related to @xmath1301 by @xmath1302 . applying ( [ ojv ] ) to @xmath85 and @xmath1276 rather than @xmath446
, we obtain @xmath1303 .
( note that @xmath880 and ( [ gvv ] ) gives @xmath1304 , since @xmath1276 is a function of @xmath5 . ) as @xmath1305 , cf .
( [ dvt ] ) , @xmath56 is thus the @xmath772-gradient of @xmath1306 . on the other hand , again from ( [ dvt ] ) , @xmath1307 equals @xmath1308 , which is a function of @xmath5 , and of @xmath805 , proving both ( a ) ( see lemma [ ggqft ] ) and ( b ) .
[ alltq]let @xmath399 be the group of all automorphisms ( definition [ ggktr ] ) of a given compact geodesic - gradient khler triple @xmath380 .
then every quadruple @xmath1309 , @xmath1310 satisfying the analog of ( [ pbd ] ) arises when remark [ ascdt](i ) is applied to a suitably chosen @xmath396-invariant geodesic - gradient khler triple @xmath1311 with the same underlying complex manifold @xmath29 . in fact
, a trivial modification ( see remark [ trivl ] ) followed by rescaling of the metric allows us to assume that @xmath585 and @xmath1010 are the same as those for @xmath380 .
our claim is now obvious from remark [ realz ] and theorem [ ntrmd ] .
[ spcas]as a special case of remark [ alltq ] , for the first triple using the fubini - study metric @xmath61 and @xmath399 as in the lines preceding ( [ dta ] ) , all quadruples @xmath1312 with ( [ pbd ] ) are realized , via remark [ ascdt](i ) , by cp triples @xmath8 having arbitrarily fixed values of @xmath1236 and @xmath1313 that satisfy ( [ dem ] ) .
[ cnvrs]conversely , we can apply remark [ realz ] and theorem [ ntrmd ] to canonically modify any given cp triple , obtaining one with the fubini - study metric and the same group @xmath396 .
we will not use the easily - verified fact that , for such a fubini - study cp triple , @xmath1314 and , in ( [ dta].ii ) , the value of @xmath43 at @xmath1315 , where @xmath1316 and @xmath1317 , equals @xmath1318 for some sign @xmath35 .
in this section @xmath8 is a fixed compact geodesgradient khler triple ( definition [ ggktr ] ) .
we use the notation of ( [ sym ] ) , denote by @xmath559 the data ( [ pbd ] ) associated with @xmath8 ( see remark [ ascdt](i ) ) , and choose for them the further data ( [ sgn ] ) ( [ int ] ) , so that a sign @xmath594 is fixed as well .
we also let @xmath1319 and @xmath135 stand for @xmath794 , the normal bundle @xmath1108 , the submanifold metric of @xmath132 , the riemannian fibre metric in @xmath131 induced by @xmath62 , and the chern connection of @xmath346 in @xmath627 , cf .
( d ) in section [ ck ] .
we write @xmath1320 when @xmath155 and @xmath1321 , as in remark [ tlspc ] . using the normal exponential diffeomorphism @xmath1322 in ( [ nex ] ) , we define @xmath1323 , depending on the sign @xmath35 , to be the composite @xmath1324 where @xmath1325 is given by @xmath1326 if @xmath1327 and , otherwise , @xmath1328 note that @xmath1329 is a homeomorphism and , restricted to the complement @xmath764 of the zero section , it becomes a diffeomorphism @xmath1330 .
in fact , @xmath1331 with @xmath1332 determines @xmath50 ( smoothly if @xmath1333 ) , since @xmath1334 and @xmath91 determines @xmath570 according to the line preceding ( [ dsr ] ) .
consequently , @xmath1335 is a homeomorphism , and the restriction @xmath1336 a diffeomorphism .
in addition , @xmath1337 due to ( [ phe ] ) , the fibre - preserving property of @xmath1338 , and the first line of remark [ ptnrs ] .
[ sends]suppose that a vector field @xmath81 on @xmath1339 is 1 . the @xmath164-horizontal lift of a vector field on @xmath132 , or 2 . a vertical vector field of the form @xmath1340 for some complex - linear vector - bundle morphism @xmath1341 , skew - adjoint relative to @xmath346 at every point .
then @xmath1338 , restricted to @xmath1342 , sends @xmath81 onto its restriction to @xmath1343 in fact , let @xmath1344 be an integral curve of @xmath235 . then the function @xmath1345 is constant , and so , by ( [ dyx ] ) , @xmath1346 with some real constant @xmath604 .
this proves our claim since , in case ( b ) , @xmath81 restricted to every fibre @xmath773 , being a linear vector field on @xmath773 , is invariant under multiplications by scalars . [ first ] for either critical manifold @xmath1347 of @xmath43 in any compact geodesic - gradient khler triple @xmath380 , the triple @xmath1348 is isomorphic to one constructed in section [ ev ] .8ptfrom some data ( [ pbd ] ) ( [ sgn ] ) and @xmath1349 .
the data consist of ( [ pbd ] ) associated with @xmath8 as in remark [ ascdt](i ) , any choice of @xmath568 with ( [ sgn ] ) for ( [ pbd ] ) and our fixed sign @xmath35 , the submanifold metric @xmath308 and normal bundle @xmath1350 of @xmath1351 , and the fibre metric @xmath346 in @xmath133 induced by @xmath62 .
furthermore , 1 .
the required isomorphism @xmath1352 is provided by the mapping @xmath1353 with ( [ phe ] ) , which , in particular , must be biholomorphic , 2 .
@xmath1354 sends the horizontal distribution of the chern connection @xmath135 of @xmath346 in @xmath627 , cf .
( d ) of section [ ck ] , onto the summand @xmath1157 in ( [ tme ] ) , 3 .
the leaves of @xmath1355 are precisely the same as the @xmath1356-images of all punctured complex lines through @xmath174 in the normal spaces of @xmath132 . in the special case
where @xmath1357 , that is , the summand distribution @xmath225 is dimensional , formula ( [ hgv].c ) used in the construction of section [ ev ] may also be replaced by the following equality , using the simplified notation of ( [ hgv].c ) : @xmath1358 it suffices to prove that the restriction of @xmath1359 to @xmath1360 is an isomorphism between the geodesic - gradient khler triples @xmath1361 and @xmath1362 , since the analogous conclusion about @xmath1359 itself then follows from ( * ? ? ?
* lemma 16.1 ) .
we start by establishing the equality @xmath1363 namely , @xmath1364 for any @xmath1365 and any @xmath1366 with @xmath1367 , so that @xmath1368 , where @xmath1369 and @xmath587 depends on @xmath598 as in ( [ dsr ] ) .
since @xmath1370 is a unit - speed geodesic , ( [ vtg ] ) and ( [ dtq ] ) give @xmath1371/d\sigma=\mp\hs q^{\hs1\nh/2}\nh$ ] , the sign factor being due to the relation @xmath1372 ( immediate from ( [ tmm ] ) with @xmath405 ) . here
@xmath870 depends on @xmath1373 as in remark [ ascdt](i ) .
however , according to remark [ compo ] and the text preceding ( [ hgv].a ) ( [ hgv].b ) , the same autonomous equation @xmath1374/d\sigma=\mp\hh q^{\hs1\nh/2}$ ] holds when @xmath1373 is replaced by @xmath1375 , with the same dependence of @xmath596 on the unknown function . the uniqueness clause of remark [ compo ] thus gives @xmath1376 , as required .
one has two complex direct - sum decompositions , @xmath1377 and @xmath1378 , orthogonal relative to @xmath61 and , respectively , @xmath772 .
the former arises from ( [ tme ] ) if one sets @xmath1379 . in the latter @xmath1380 and @xmath766 are the distributions introduced in the lines following ( [ tnp ] ) .
first , for @xmath1381 as in ( [ lvf ] ) and our @xmath869 , where @xmath405 , we show that @xmath1382 more precisely , @xmath1329 ( or , @xmath149 ) appearing in ( [ dyx ] ) ( or , ( [ nex ] ) ) , restricted to @xmath1339 ( or , @xmath1383 ) , sends @xmath1384 onto their restrictions to @xmath1383 ( or , respectively , onto @xmath1385 ) .
the claims about @xmath768 in ( [ prs].i ) ( [ prs].ii ) follow as @xmath1329 clearly preserves each leaf of @xmath1386 , that is , each punctured complex line through @xmath174 in the normal space @xmath1387 at any point @xmath703 , while , by lemma [ vsbkr](a ) , @xmath149 maps the leaves of @xmath768 intersected with @xmath1388 onto leaves of @xmath228 .
this also proves ( ii ) .
next , the class of vertical vector fields of remark [ sends](b ) obviously includes @xmath1381 and , locally , some of them span @xmath1389 .
remark [ sends ] thus yields the remainder of ( [ prs].i ) , while ( [ prs].iii ) for @xmath1329 follows from complex - linearity of the @xmath164-horizontal lift operation ( due to lemma [ ddrsq](i ) ) , and the fact that @xmath1329 acts on the vertical vector fields in remark [ sends](b ) as the identity operator . on the other hand , ( [ prs].ii ) in the case of @xmath769 and @xmath766 ( or , of @xmath1390 ) is an immediate consequence of the second ( or , third ) claim in theorem [ dcomp](f ) .
( to be specific , for @xmath769 and @xmath766 this is clear from remark [ inequ](ii ) combined with ( [ hpm ] ) ( [ tme ] ) . )
finally , the complex - linearity assertion of theorem [ dcomp](f ) implies ( [ prs].iii ) . by ( [ prs ] ) , the diffeomorphism @xmath1391 maps @xmath1380 and @xmath766 onto @xmath1392 and @xmath1393 .
proving the theorem is thus reduced to showing that @xmath1394 to begin with , for @xmath807 as in section [ ev ] , @xmath811 given by ( [ lvf ] ) , and our @xmath405 , @xmath1395 in the case of @xmath807 this amounts to @xmath1396 , which is a trivial consequence of ( [ vcf ] ) and the fact that @xmath807 was defined in section [ ev ] to be the same function of @xmath778 as @xmath596 is of @xmath5 . for @xmath1390 , ( [ qav ] ) follows from ( [ prs ] ) and ( [ phe ] ) .
next , any integral curve of @xmath811 in @xmath1397 has , up to a shift of the parameter , the form @xmath1398 with a unit vector @xmath1321 , so that @xmath1399 , where in addition to the curve parameter @xmath1400 , two more real variables are used : @xmath1401 , and @xmath91 related to @xmath598 via ( [ dsr ] ) . the chain rule
thus yields @xmath1402 , while @xmath1403 for @xmath1404 .
since @xmath1405 is a unit - speed geodesic , ( [ vtg ] ) and ( [ dxe ] ) give @xmath1406/d\sigma=\mp\hs q^{\hs1\nh/2}\nh$ ] , with @xmath596 evaluated at @xmath1407 , and the sign factor arising from ( [ tmm ] ) , as @xmath405 . applying the chain rule again , we obtain @xmath1406/dr = v\nh_{x(\nh\sigma\nh)}\w$ ] and , consequently , ( [ qav ] ) . the claim made in ( [ mps ] ) about @xmath1408 and @xmath871 is now obvious from ( [ qav ] ) and ( [ gvv ] ) along with ( [ ana ] ) . for the remaining two pairs of summands , ( [ mps ] ) in the case of @xmath1409 ( or , @xmath1410 ) is a direct consequence of ( [ prs ] ) and ( [ phe ] ) ( or , respectively , of ( i ) ( ii ) in remark [ inequ ] along with parts ( h2 ) ( h3 ) of theorem [ dcomp ] , ( [ vcf ] ) and ( [ hgv ] ) ) .
note that , by ( [ dyx ] ) , @xmath1329 leaves @xmath1411 unchanged , while @xmath1412 in ( h2 ) .
finally , if @xmath1413 in ( [ tme ] ) , remark [ inequ](ii ) allows us to use ( h1 ) in theorem [ dcomp ] , instead of ( h2 ) , obtaining ( [ alt ] ) .
[ totgd ] suppose that @xmath8 is a compact geodesic - gradient khler triple .
then , for @xmath234 and @xmath1235 appearing in ( [ tme ] ) , with either sign @xmath35 , the distribution @xmath1414 is integrable and its leaves are totally geodesic in @xmath1415 .
use theorem [ first ] and theorem [ ehggk](b ) ( or for integrability
( [ hpm ] ) ) .
in the next result the inclusions @xmath1416 and @xmath1417 come from the standard identification ( [ inc ] ) for @xmath826 , where @xmath145 .
let us also note that , by ( [ hpm ] ) and corollary [ totgd ] , the restriction to the normal space @xmath1418 of the biholomorphism @xmath1419 ( see theorem [ first ] ) constitutes @xmath1420 [ tgimm ] given a compact geodesic - gradient khler triple @xmath8 and a fixed sign @xmath35 , let @xmath1046 be a point of the critical manifold @xmath14 . then the following conclusions hold . 1 .
the embedding @xmath1421 with ( [ tgh ] ) has an extension to a totally geodesic holomorphic immersion @xmath1422 .
the mapping @xmath206 in ( a ) restricted to the projective hyperplane @xmath1423 at infinity is a totally geodesic holomorphic immersion @xmath1424 , and the metric that it induces on @xmath834 equals @xmath1425 times the fubini - study metric , cf .
remark [ fbstm ] , arising from the inner product @xmath1426 in @xmath1427 , for @xmath1428 as in remark [ ascdt](i ) , 3 .
the images of the immersion @xmath1429 in ( b ) and of its differential at any point @xmath1430 , where @xmath1099 and @xmath1333 , coincide with the @xmath1067-image of the leaf of @xmath1431 in @xmath34 passing through @xmath1432 and , respectively , with the subspace @xmath1433 of @xmath1434 . as a consequence of theorems [ first ] and [ dcomp](c ) , the composite @xmath1435 maps @xmath1436 ( the complement of the zero section in @xmath1102 ) holomorphically into @xmath1437 .
the restriction of @xmath1435 to @xmath1438 , being , by ( [ pex ] ) and ( [ phe ] ) , constant on each punctured complex line through @xmath147 , thus descends to @xmath1439 where the immersion property of @xmath372 is an immediate consequence of the fact , established below , that both @xmath1440 and @xmath1441 have constant ( complex ) rank , equal to @xmath1442 .
as @xmath1359 is a biholomorphism , it suffices to verify this last claim for the former mapping ; we do it noting that @xmath1443 coincides with the @xmath1048-preimage of @xmath1046 ( due to ( [ pcf ] ) and remark [ ptnrs ] ) , and hence forms a leaf of @xmath1444 restricted to @xmath34 , cf .
( [ hpm ] ) . that @xmath1445 satisfies
the required rank condition is now clear : the kernel of its differential at any point @xmath13 coincides , by ( [ hpm ] ) and ( [ tme ] ) , with @xmath1446 , while @xmath885 .
the mapping @xmath1422 , equal to @xmath1359 on @xmath773 and to @xmath372 on @xmath834 , is continuous .
namely , if it were not , we could pick a sequence @xmath1447 , @xmath1448 , such that @xmath1449 and @xmath1450 as @xmath1451 for some unit vector @xmath1321 , while no subsequence of the image sequence @xmath1452 tends to @xmath1453 .
the resulting limit relation @xmath1454 , where @xmath1455 corresponds to @xmath1456 as in the line preceding ( [ dsr ] ) , combined with ( [ phe ] ) , now gives @xmath1457 which due to continuity of @xmath149 and ( [ pex ] ) converges to @xmath1458 , for a specific point @xmath1051 . however , ( a ) ( b ) in lemma [ vsbkr ] and the definition of @xmath372 also give @xmath1459 , which contradicts our choice of @xmath1460 , proving continuity of @xmath1461 .
holomorphicity of @xmath206 is now obvious from remark [ hlext ] applied to @xmath1462 and its codimension - one complex submanifold @xmath1463 .
furthermore , @xmath1464 to see this , first note that @xmath206 has two restrictions , @xmath372 to @xmath834 and @xmath1359 to the dense open submanifold @xmath773 , already known to be immersions , the former into @xmath1465 , cf .
( [ tgh ] ) ( [ dsc ] ) .
next , for any unit vector @xmath1321 , if @xmath1466 denotes the projective line in @xmath1467 joining @xmath1468 to the point @xmath848 ( identified via ( [ inc ] ) with @xmath1469 ) , then the restriction of @xmath206 to @xmath1466 is an embedding with the image @xmath1470 forming a complex submanifold of @xmath29 , biholomorphic to @xmath1471 , and intersecting each of @xmath3 and @xmath4 orthogonally at a single point .
in fact , lemma [ vsbkr ] yields all the claims just made except the ` embedding ' property ; we obtain the latter from remark [ known](b ) , which we use to conclude that the resulting holomorphic mapping @xmath1472 , being injective ( since so is @xmath1473 ) , must be a biholomorphism . now ( [ imm ] ) follows . for obvious reasons of continuity , ( [ tgh ] )
implies that the holomorphic immersion @xmath1474 is totally geodesic , which establishes ( a ) .
finally , remarks [ fbstu ] , [ ascdt](iii ) , [ ttgim ] and theorem [ first ] give rise to ( b ) , completing the proof .
[ codim]for @xmath1475 as in remark [ dppdm ] , the codimension @xmath1476 of the immersion @xmath372 in theorem [ tgimm](b ) equals @xmath388 .
in fact , @xmath1477 , and so , by ( [ dmq ] ) , @xmath1478 .
[ spmcp]suppose that the distribution @xmath271 in ( [ tme ] ) is dimensional or , in other words , @xmath1357 .
then , for either sign @xmath35 , the critical manifold @xmath14 , with its submanifold metric , must be biholomorphically isometric to a complex projective space carrying the fubini - study metric multiplied by @xmath1479 . in fact , the isometric immersion @xmath372 of theorem [ tgimm](b ) , having codimension zero ( cf .
remark [ codim ] ) , is necessarily a biholomorphism ( remark [ ftcvr ] ) .
the results stated and proved below use definition [ ggktr ] , the notations of ( [ sym ] ) , ( [ dbp ] ) , ( [ hpm ] ) , and the notion of projectability introduced in section [ pd ] . [ intpr ] for a compact geodesic - gradient khler triple @xmath380 , the following three conditions are mutually equivalent . 1
. the distribution @xmath1480 on @xmath1481 is integrable .
@xmath1482 is @xmath1483-projectable .
@xmath1484 is @xmath1485-projectable . in ( ii )
( iii ) one may also replace @xmath1486 by @xmath1235 or @xmath253 .
if ( i ) ( iii ) hold , then : 1 .
the immersions of theorem [ tgimm](c ) are all embeddings .
the @xmath1487-images @xmath1488 of the integrable distribution @xmath1489 on @xmath404 are integrable holomorphic distributions on @xmath28 and have totally geodesic leaves biholomorphically isometric to complex projective spaces carrying @xmath1425 times the fubini - study metric , cf .
theorem [ tgimm](b ) .
these leaves coincide with the images of the embeddings in ( iv ) , and form the fibres of holomorphic bundle projections @xmath1490 for some compact complex base manifolds @xmath1491 .
the summand @xmath271 in ( [ tme ] ) is @xmath1487-projectable and its @xmath1487-image coincides with the orthogonal complement of @xmath1492 in @xmath1493 .
the leaf space @xmath1494 admits a unique structure of a compact complex manifold such that the quotient projection @xmath1495 constitutes a holomorphic fibration while , for either sign @xmath594 and @xmath1490 as in ( iv ) , the mapping @xmath1496 , sending each leaf of @xmath250 to its image under @xmath1497 , is a biholomorphism .
there exists a unique holomorphic bundle projection @xmath221 with @xmath1498 on @xmath404 such that , for both signs @xmath35 , the restriction of @xmath229 to @xmath404 equals @xmath1499 , where @xmath1500 is the inverse of the biholomorphism @xmath1501 in ( vii ) .
@xmath1502 , with the notation of ( [ nwt ] ) , for the submanifold metric @xmath308 of @xmath14 , the normal connection @xmath135 in its normal bundle @xmath1350 , any vector field @xmath893 on @xmath14 , and any section @xmath81 of @xmath1503 , cf .
( v ) . since @xmath1157 are both integrable by ( [ hpm ] ) , the mutual equivalence of ( i ) , ( ii ) , ( iii ) and the integrability claim in ( v ) are all immediate from lemma [ intgr ] applied to @xmath1504 , along with ( [ hpm ] ) and ( [ tme ] ) .
the immersions mentioned in theorem [ tgimm](c ) thus have nonsingular images , namely , the leaves @xmath87 of the distribution @xmath1492 in ( v ) , so that ( iv ) follows from remark [ ftcvr ] applied to @xmath834 standing for @xmath1505 , with @xmath36 defined in remark [ dppdm ] , and such a leaf @xmath320 .
the remaining part of ( v ) is a direct consequence of theorem [ tgimm](b ) and remark [ fibra ] . at any @xmath145
, the image @xmath1138 is now independent of the choice of @xmath1052 with @xmath1506 , and hence so is its orthogonal complement @xmath1139 in @xmath1047 ( see remark [ inequ](iii ) ) , proving assertion ( vi ) .
the mappings @xmath1501 in ( vii ) are obviously bijective , and lead to an identification @xmath1507 which is a biholomorphism , as one sees restricting @xmath15 to `` local '' complex submanifolds of @xmath404 which the composite bundle projections @xmath1508 ( with fibres provided by the leaves of @xmath253 ) send biholomorphically onto open submanifolds of @xmath1491 .
this yields ( vii ) . for ( viii ) , it suffices to note that the two composite bundle projections @xmath1509 agree , by ( vii ) , on the intersection @xmath404 of their domains , cf
. remark [ ascdt](iv ) , while the union of their domains is @xmath29 . for ( ix )
, theorem [ first ] allows us to identify @xmath1510 with @xmath131 so that ( [ hgv].c ) and ( [ gsw ] ) hold under the assumptions following ( [ hgv ] ) . since @xmath81 lies in the @xmath1487-image @xmath1488 of @xmath1068 , cf .
( ii ) , ( iii ) , ( v ) , formula ( [ hvk ] ) gives @xmath1511 for its @xmath164-horizontal lift , also denoted by @xmath235 .
replacing @xmath1512 in ( [ gsw ] ) with @xmath1513 and multiplying the result by @xmath1514 , we get an expression for @xmath1113 which , equated to ( [ hgv].c ) , yields @xmath1515 , since @xmath1516 while , obviously , @xmath1517 . applying the last equality to @xmath1518 instead of @xmath235 , and using ( b ) in section [ ck ] along with hermitian symmetry of @xmath1519 in @xmath1520 , we obtain the required relation in ( ix ) . note that the above proof of ( ix ) in lemma [ intpr ] actually uses the assumptions ( i ) ( iii ) : without them , the formula @xmath1515 , rather than being valid for any given @xmath1521 , @xmath145 , and all vectors @xmath50 normal to @xmath28 at @xmath1031 , would hold only when @xmath81 lies in some subspace of @xmath1047 depending on @xmath1030 .
let us now fix a khler manifold @xmath1522 , and consider pairs @xmath1523 formed by a holomorphic complex vector bundle @xmath131 over @xmath1524 and the real part @xmath346 of a hermitian fibre metric in @xmath627 , the chern connection of which see section [ ck ] satisfies the curvature condition @xmath1525 for any vector fields @xmath63 tangent to @xmath1526 , where the notation of ( [ nwt ] ) is used .
[ rdehj ] whenever @xmath1524 is simply connected and such @xmath1523 exist , they are essentially unique , in the sense that , given another pair @xmath1527 with the same property , some holomorphic vector - bundle isomorphism @xmath1528 takes @xmath346 to @xmath1529 .
remark [ crvcm ] implies that the chern connections @xmath135 and @xmath1530 induce a flat metric connection in the bundle @xmath1531 .
the required isomorphism is now provided by a global parallel section of @xmath1532 chosen so as to transform @xmath356 into @xmath477 at one point , and its holomorphicity follows from ( e ) in section [ ck ] .
[ cptrp ] for a compact geodesic - gradient khler triple @xmath380 , the following two conditions are equivalent . 1 .
@xmath1533 is isomorphic to a cp triple , defined as in section [ eg ] .
@xmath1534 , where @xmath1236 and @xmath2 .
in other words , cf .
remark [ dppdm ] , @xmath1535 , that is , @xmath271 in ( [ tme ] ) is dimensional . in this case , the assertion of theorem [ first ] , including ( [ alt ] ) , is satisfied by @xmath1536 , with either fixed sign @xmath594 and @xmath1537 biholomorphically isometric to a complex projective space carrying @xmath1425 times the fubini - study metric , @xmath131 and @xmath346 being , up to a holomorphic vector - bundle isomorphism , the normal bundle of the latter treated as a linear variety in @xmath820 and its hermitian fibre metric induced by the fubini - study metric of @xmath1538 .
furthermore , the isomorphism types of cp triples @xmath8 having any given values of @xmath1539 and @xmath1540 in ( ii ) are in a natural bijective correspondence , obtained by applying remark [ ascdt](i ) , with quadruples @xmath1541 that satisfy ( [ pbd ] ) .
first , ( i ) implies ( ii ) according to ( [ dem ] ) .
assuming now ( ii ) , let us use remark [ spcas ] to select a cp triple @xmath1542 realizing the same data @xmath1543 and @xmath1268 , in ( ii ) above and remark [ ascdt](i ) , as our @xmath8 ( which also establishes the surjectivity part of the final clause ) . with either fixed sign @xmath35 , denoting @xmath1544 by @xmath1545 , and their analogs for @xmath1546 by @xmath1547 , we choose the isomorphisms @xmath1548 and @xmath1549 by applying theorem [ first](i ) to both triples . as ( i ) has already been shown to yield ( ii ) , we may now also apply remark [ spmcp ] to both of them , identifying the critical manifolds @xmath1550 ( and their submanifold metrics ) with a complex projective space @xmath1524 ( and , respectively , with the fubini - study metric @xmath1551 multiplied by @xmath835 ) .
next , ( ix ) in lemma [ intpr ] holds for both triples , so that the pairs @xmath1523 and @xmath1527 associated with them via theorem [ first ] satisfy , along with @xmath1552 and @xmath1553 , the assumptions as well as the conclusion of lemma [ rdehj ]
. thus , some holomorphic vector - bundle isomorphism @xmath1528 takes @xmath346 to @xmath477 and , since the metrics @xmath1554 on @xmath131 and @xmath1339 constructed in section [ ev ] depend only on @xmath1555 ( aside from the data fixed above and shared by both triples ) , this isomorphism is a holomorphic isometry of @xmath1556 onto @xmath1557 , sending @xmath43 to its analog on @xmath1342 . in view of (
* lemma 16.1 ) , it can be extended to an isomorphism between the triples @xmath8 and @xmath1542 .
we thus obtain injectivity in the final clause and the fact that ( ii ) yields ( i ) .
once again , we use the notation of ( [ sym ] ) , ( [ tmm ] ) and ( [ nex ] ) , assuming @xmath8 to be a compact geodesic - gradient khler triple ( definition [ ggktr ] ) . [ ttgds ]
suppose that conditions ( i )
( iii ) along with the other assumptions of lemma [ intpr ] hold for a triple @xmath380 , and @xmath1558 are as in lemma [ intpr](viii ) . 1 . given a @xmath223-projectable nonzero local section @xmath81 of the distribution @xmath271 in ( [ tme ] ) , 1 .
@xmath1559 commutes with the vector fields @xmath85 and @xmath869 , 2 .
@xmath1559 is @xmath1487-projectable for both signs @xmath35 , 3 .
the local flow of @xmath81 in @xmath404 preserves the distributions @xmath1560 and @xmath1561 .
the leaves of the integrable distribution @xmath1480 on @xmath1481 are totally geodesic complex submanifolds of @xmath1481 and all the local flows mentioned in ( a3 ) act between them via local isometries .
any @xmath81 in ( a ) is normal to the totally geodesic leaves of the integrable distributions @xmath1157 ( see corollary [ totgd ] ) , while @xmath66 are both tangent to them , as @xmath885 .
therefore , @xmath1562 and @xmath1563 , being , as a result , also normal to those leaves for both signs @xmath35 , are by ( [ tme ] ) sections of @xmath225 .
the same is true of @xmath1564 ( and hence of @xmath139,[u , w]$ ] ) due to @xmath1074-invariance in ( [ tme ] ) , with @xmath927 and @xmath1565 , cf .
( [ loc].a ) . at the same time , @xmath223-projectability of @xmath235 implies , via remark [ liebr ] and lemma [ intpr](viii ) , that @xmath139\,$ ] and @xmath1566\,$ ] are sections of @xmath1567 .
we thus obtain ( a1 ) along with ( a3 ) for @xmath228 .
next , ( a2 ) follows : due to @xmath223-projectability of @xmath235 , with @xmath1049 fixed , @xmath1568 is independent of the choice of @xmath1052 such that @xmath1506 , and hence so must be @xmath1569 , as the differential at @xmath1046 of the bundle projection @xmath1570 ( see ( vii ) ( viii ) in lemma [ intpr ] ) sends @xmath1569 to @xmath1568 , which determines @xmath1569 uniquely due to its being orthogonal , by lemma [ intpr](v ) , to @xmath1571 , for the vertical distribution @xmath1492 of @xmath1572 .
we obtain the remainder of ( a3 ) by noting that , for either fixed sign @xmath35 , @xmath1573 namely , since the @xmath1487-image @xmath190 of @xmath81 is obviously @xmath1574-projectable , we may prescribe the @xmath1487-image @xmath1575 of @xmath1576 to be a local section of @xmath1492 commuting with @xmath190 ( see ( v )
( vi ) in lemma [ intpr ] and remark [ cmmut ] ) , and then lift @xmath1575 to @xmath1577 , using remark [ inequ](iii ) . for the resulting lift @xmath1578 , ( [ cmt ] ) and parts ( ix ) , ( vi ) of lemma [ intpr ] give @xmath1579=0
we now derive ( b ) from remark [ tglvs ] . according to remark [ frthr ]
, it suffices to establish ( i ) in remark [ tglvs ] for local sections of @xmath250 having the form @xmath1580 with @xmath1576 satisfying ( [ lft ] ) and @xmath1581 equal to a constant - coefficient combination of @xmath80 and @xmath195 .
orthogonality in ( [ tme ] ) combined with ( [ gvv ] ) shows that @xmath1582 equals a constant multiple of @xmath596 plus the sum of the terms @xmath1583 .
verifying part ( i ) of remark [ tglvs ] thus amounts to showing that @xmath1584=0 $ ] . in terms of the @xmath1487-image @xmath1575 of @xmath1578 ,
remark [ inequ](iv ) gives @xmath1585 , where @xmath308 is the submanifold metric of @xmath14 .
since @xmath585 are constants , our claim is thus reduced to two separate parts , @xmath1586 and @xmath1587=0 $ ] .
the former part is immediate : @xmath596 is a function of @xmath5 , cf .
remark [ ascdt](i ) , while @xmath81 and @xmath85 are sections of the mutually orthogonal summands @xmath271 and @xmath885 in ( [ tme ] ) . for the latter part , ( a2 ) allows us to replace @xmath81 by its @xmath1487-image @xmath1588 , noting that @xmath1589 is the @xmath1487-pullback of a function defined ( locally ) in @xmath14 . now @xmath1587=0 $ ] due to the fact that ( ii ) implies ( i ) in remark [ tglvs ] , and @xmath1588 , or @xmath1590 , is normal or , respectively , tangent to the totally geodesic leaves of the integrable distribution @xmath1503 , while @xmath1588 , besides being as noted above
projectable along @xmath1503 , also commutes with @xmath1575 ( see ( v ) ( vi ) in lemma [ intpr ] and ( [ lft ] ) ) .
we say that a ( locally - trivial ) holomorphic fibre bundle carries a specific local - type fibre geometry if such a geometric structure is selected in each of its fibres and suitable local @xmath92 trivializations make the structures appear the same in all nearby fibres .
for instance , holomorphic complex vectors bundle endowed with hermitian fibre metrics may be referred to as 1 .
holomorphic bundles of hermitian vector spaces .
the fact that ( i ) leads to the presence of the distinguished chern connection ( section [ ck ] ) has obvious generalizations to two situations ( ii ) ( iii ) discussed below . by a horizontal distribution for a holomorphic bundle projection @xmath1591 between complex manifolds , also called a connection in the holomorphic bundle @xmath1592 we mean any @xmath92 real vector subbundle @xmath271 of @xmath74 , complementary to the vertical distribution @xmath699 , so that @xmath284 is the direct sum of @xmath249 and @xmath225 .
horizontal lifts of vectors tangent to @xmath270 , and of piecewise @xmath976 curves in @xmath270 , as well as parallel transports along such curves , are then defined in the usual fashion , although the maximal domain of a lift of a curve ( or , of a parallel transport ) may in general be a proper subinterval of the original domain interval . this last possibility does not , however , occur in bundles with compact fibres , or in vector bundles with linear connections , where horizontal lifts of curves and parallel transports are all global .
we proceed to describe the chern connection @xmath271 in the cases of 1 .
holomorphic bundles of fubini - study complex projective spaces , and 2 .
holomorphic bundles of cp triples , over any complex manifold @xmath270 .
their fibre geometries consist of fubini - study metrics ( remark [ fbstm ] ) and , respectively , the structures of a cp triple ( section [ eg ] ) . for ( ii ) , @xmath271 arises since local @xmath92 trivializations mentioned earlier may be chosen so as to share their domains with local holomorphic trivializations ; the former make the fibre geometry appear constant , and the latter turn the bundle , locally , into the projectivization ( [ prz ] ) of a holomorphic vector bundle @xmath48 endowed with a hermitian fibre metric @xmath660 that induces the fubini - study metrics of the original fibres . since @xmath639 is unique up to multiplications by positive functions ( remark [ fbstm ] ) , lemma [ ddrsq](iv ) easily implies that its choice does not affect the resulting parallel transports between the projectivized fibres , thus giving rise to @xmath225 .
the chern connection @xmath271 now also arises in case ( iii ) since , according to remark [ cnvrs ] , ( iii ) is a subcase of ( ii ) .
the situation is , however , more special : the critical manifolds analogs of ( [ tmm ] ) in the
fibres now constitute two holomorphic bundles @xmath28 of fubini - study complex projective spaces over @xmath42 ( with fibre dimensions that need not be both positive ; see remark [ spmcp ] ) , contained as subbundles in the original bundle , and invariant under all @xmath225-parallel transports . also , the fibre - geometry gradients and their @xmath727-images ( analogous to what we normally denote by @xmath85 and @xmath195 ) together form two holomorphic vertical vector fields @xmath56 and @xmath869 on the total space .
this is immediate from the preceding paragraph , with the two subbundles @xmath28 corresponding to a @xmath639-orthogonal holomorphic decomposition @xmath1593 of the locally - defined vector bundle @xmath1594 , cf .
( [ dta].ii ) and ( [ spm].c ) ( [ spm].d ) , while the flow of @xmath195 , described in the lines following ( [ dta ] ) , acts in both @xmath1595 via multiplications by two ( unrelated ) constant unit complex scalars . in case ( ii ) , or ( iii ) ,
@xmath1596 which holds for ( ii ) since it does for ( i ) , cf .
section [ ck ] and , consequently , also extends to the case of ( iii ) via the canonical modifications in remarks [ alltq ] and [ cnvrs ] .
the following assumptions and notations will now be used to construct compact geodesic - gradient khler triples , each of which we call a horizontal extension of the cp triple provided by any fibre @xmath1597 . 1
. @xmath1598 and @xmath271 are the bundle projection and the chern connection of a holomorphic bundle of cp triples with a compact base @xmath42 and the triple fibres @xmath1597 , @xmath1599 , while @xmath28 stand for the above subbundles of fubini - study complex projective spaces , invariant under @xmath225-parallel transports .
2 . we let @xmath1600 be the data associated with some@xmath444any fibre @xmath1597 as in remark [ ascdt](i ) , and @xmath75 ( or , @xmath11 ) be the @xmath92 function ( or , holomorphic bundle projection ) which , restricted to each @xmath1601 , equals @xmath1602 or , respectively , the version of ( [ dbp ] ) corresponding to @xmath1597 .
we also set @xmath10 .
one is given two khler metrics @xmath1603 on the total spaces @xmath28 of our holomorphic bundles of fubini - study complex projective spaces such that either @xmath1603 makes the fibres @xmath1604 , @xmath1599 , orthogonal to @xmath271 along @xmath28 and , restricted to each fibre , @xmath1603 equals @xmath1425 times the fubini - study metric of @xmath1604 .
we define a riemannian metric @xmath61 on @xmath404 by requiring that @xmath271 be @xmath62-orthogonal to the vertical distribution @xmath699 , that @xmath61 agree on the fibres @xmath1605 with the metrics @xmath1606 , and that @xmath1607 on @xmath225 , the symbols @xmath1603 being also used for the @xmath1487-pullbacks of @xmath1608 , cf .
( b ) ( c ) . 5 .
our final assumption is that the riemannian metric @xmath61 on the dense open submanifold @xmath404 has an extension to a khler metric on @xmath247 ( still denoted by @xmath62 ) .
[ adggk]under the hypotheses ( a )
( e ) , the resulting horizontal extension @xmath8 is actually a geodesic - gradient khler triple .
namely , being a part of the geometry of the fibres @xmath1597 , the functions @xmath1602 are preserved by @xmath225-parallel parallel transports , that is , @xmath43 is constant along @xmath225 , and so its ( vertical ) @xmath62-gradient must , by remark [ gpgrd ] .
coincide with the holomorphic vertical vector field @xmath56 described in the lines preceding ( [ prt ] ) . on the other hand ,
the function @xmath1609 , equal - consequently - to its fibre version , is a specific function of @xmath5 .
thus , by lemma [ ggqft ] , @xmath43 has a holomorphic geodesic @xmath62-gradient .
a compact geodesic - gradient khler triple @xmath8 is a horizontal extension arising as in remark [ adggk ] , the distribution @xmath1610 on @xmath404 coming from the decomposition ( [ tme ] ) for @xmath8 coincides , on @xmath34 , with the vertical distribution @xmath249 of the bundle projection @xmath1611 ( see ( a ) above ) and , consequently , @xmath250 is integrable . in fact
, applying remark [ tglvs ] to @xmath225-horizontal lifts @xmath81 of local vector fields on @xmath270 , we see that , by ( [ prt ] ) , @xmath249 has totally geodesic leaves . using ( [ img ] ) for both @xmath8 and the fibres @xmath1597 ,
we now conclude that the projections @xmath11 defined in ( b ) are the same as those in ( [ dbp ] ) .
( note that , due to the orthogonality requirement in ( c ) , the minimizing geodesic segment in @xmath1601 , @xmath1599 , joining a point @xmath1612 to @xmath1613 a normal to @xmath1604 , serves as the segment with the same properties for @xmath44 rather than @xmath1601 . )
now ( [ hpm ] ) implies that the distribution @xmath1614 is contained in @xmath249 and , restricted to every fibre , equals the analog of @xmath1614 for the fibre , that is , its tangent bundle ( see theorem [ cptrp](ii ) ) .
thus , @xmath250 coincides with the full vertical distribution @xmath699 .
[ cpbdl ] a geodesic - gradient khler triple @xmath380 , with compact @xmath29 , satisfies one@xmath1615all of the mutually - equivalent conditions ( i )
( iii ) of lemma [ intpr ] , if and only if it is isomorphic to a horizontal extension of a cp triple , defined as above using ( a )
remark [ hhext ] clearly yields the ` if ' part of our claim .
conversely , let @xmath8 satisfy ( i ) ( iii ) in lemma [ intpr ] .
lemma [ intpr](viii ) states that @xmath1610 coincides , on @xmath34 , with the vertical distribution @xmath699 of the holomorphic bundle projection @xmath1591 . also , in view of remark [ trivl ] , the leaves of @xmath250 form geodesic - gradient khler triples , due to their being complex submanifolds of @xmath44 tangent to @xmath85 ( since @xmath885 ) and , as they are also totally geodesic ( see lemma [ ttgds](b ) ) , ( [ hvk ] ) and the @xmath1074-invariance in ( [ tme ] ) , with @xmath927 , imply via theorem [ cptrp ] that they are all isomorphic to cp triples .
the local isometries of lemma [ ttgds](b ) can obviously be made global due to compactness ( see the lines preceding ( ii ) above ) which , consequently , turns @xmath44 into a holomorphic bundle of cp triples over @xmath270 , in the sense of ( iii ) .
on the other hand , the @xmath62-orthogonal complement of @xmath1616 is equal , on @xmath34 , to the summand @xmath271 in ( [ tme ] ) .
thus , @xmath271 constitutes a connection in the bundle @xmath44 over @xmath270 , as defined in the lines following ( i ) , and being the intersection of the horizontal distribution of the chern connections @xmath1157 in the normal bundles @xmath1350 , cf . theorem [ first](ii )
@xmath271 itself is , according to ( a ) in section [ ck ] , the chern connection of the holomorphic bundle @xmath44 of cp triples over @xmath270 .
this provides parts ( a ) ( b ) of the data ( a )
( e ) required above , with @xmath28 and @xmath1600 given by ( [ tmm ] ) and , respectively , remark [ ascdt](i ) .
the submanifold metrics @xmath1603 of @xmath28 have , by ( v ) ( vi ) in lemma [ intpr ] and the final clause of theorem [ dcomp](b ) , all the properties needed for ( c ) . to show that @xmath61 satisfies ( d ) , consider two @xmath223-projectable nonzero local sections @xmath63 of the distribution @xmath1617 , cf .
( [ tme ] ) . according to lemma [ ttgds](a ) and the last line of remark [ pralg ] , @xmath81 and @xmath893 are projectable along @xmath234 and @xmath1577 , as well as @xmath1487-projectable , for either sign @xmath35 .
their restrictions to any fixed normal geodesic segment @xmath947 emanating from @xmath28 thus lie in the space @xmath941 ( cf .
( [ vtg ] ) and ( i ) ( ii ) in theorem [ jacob ] ) and , by theorem [ dcomp](g ) , @xmath1113 restricted to @xmath947 is a ( possibly nonhomogeneous ) linear function of @xmath5 .
the same linearity condition obviously holds for @xmath1113 when @xmath61 is defined as in ( d ) , rather than being the metric of our triple @xmath380 .
the two definitions of @xmath1113 must now agree , as the two linear functions have in view of remark [ inequ](iii ) and the final clause of theorem [ dcomp](b ) the same values @xmath1618 at either endpoint @xmath585 of the interval @xmath864 $ ] .
[ clone]all compact skrp triples of class 1 ( cf .
section [ ev ] ) must be @xmath1619 while those of class 2 are themselves cp triples of a special type .
the former claim is easily verified using ( * ? ? ?
* theorem 16.3 ) ; for the latter , see lemma [ cltwo ] .
the classification result of ( * ? ? ?
* theorem 6.1 ) may be rephrased as the conclusion ( [ cdo ] ) about all compact geodesic - gradient khler triples @xmath8 with @xmath1620 other than class 2 skrp triples are .
similarly , ( [ cdo ] ) is the case by their very construction for the gradient khricci solitons of koiso @xcite and cao @xcite , mentioned in the introduction .
in this section all vector spaces are fidimensional and complex .
bilinear mappings of the type discussed here arise in any compact geodesic - gradient khler triple ( see theorem [ ctrkm ] ) , which leads to the dichotomy conclusion of theorem [ dicho ] .
a constant - rank multiplication is any bilinear mapping @xmath1621 , where @xmath1622 are vector spaces , such that the function @xmath1623 is constant or , equivalently , @xmath1624 is the same for all nonzero @xmath1625 . when @xmath1626 for all @xmath1627 , we also say that @xmath1628 has the constant rank @xmath1629 . with the notations of section [ eg ] ,
such @xmath1630 leads to a mapping @xmath1631 [ dfphi ] for @xmath1630 and @xmath305 as above , @xmath1632 and @xmath313 are both holomorphic . in terms of the identification ( [ twg ] ) , the differential of the former mapping at any @xmath1633 sends @xmath1634 to the unique @xmath1635 with @xmath1636 for all @xmath1637 , where @xmath1638 is any linear lift of @xmath1639 .
this is obvious if one sets @xmath1640 in remark [ dfrtl ] .
[ cstrk]any given constant - rank multiplication @xmath1628 leads to further such multiplications , @xmath1641 and @xmath1642 , obtained by setting @xmath1643\,$ ] and @xmath1644^*\nh$ ] . here
@xmath1645 are vector spaces , @xmath1646 ( or , @xmath1647 ) is surjective ( or , injective ) and linear , while @xmath1648^*$ ] stands for the dual of a vector space or a linear operator .
[ holem ] if @xmath1628 has the constant rank @xmath1649 and @xmath305 with ( [ phi ] ) is nonconstant , then @xmath305 is a holomorphic embedding . whether @xmath305 is constant , or not , the same is the case for all multiplications @xmath1650 of the constant rank @xmath1651 , sufficiently close to @xmath1652 .
let @xmath1653 . the subset of @xmath1654 consisting of @xmath174 and all with @xmath1655 is a vector subspace .
in fact , if @xmath1656 and @xmath1657 , then @xmath1658 for any @xmath1659 and , unless @xmath1660 , this inclusion is actually an equality due to the constant - rank property of @xmath1652 .
therefore , @xmath1661-preimages of points of @xmath1662 are linear subvarieties in @xmath1663 .
if @xmath313 is nonconstant , all these subvarieties are zedimensional , that is , @xmath305 has to be injective .
namely , by lemma [ nontr ] , for the khler form @xmath1664 of any khler metric on @xmath1665 , the integral of @xmath312 over any projective line @xmath1666 in @xmath1667 is nonzero , and so @xmath1668 can not lie in the @xmath1661-preimage of a point .
also , lemma [ dfphi ] guarantees holomorphicity of @xmath313 .
let @xmath1661 now be nonconstant
. then @xmath305 must be an embedding , that is , @xmath1669 is injective at any @xmath1670 or , equivalently , the differential of @xmath1671 at any @xmath1633 has the kernel @xmath1430 .
namely , in lemma [ dfphi ] we may set @xmath1672 when @xmath1673 , and so @xmath521 lies in the kernel if and only if the inclusion @xmath1674 holds for @xmath1675 .
unless @xmath172 , this inclusion is , as before , an equality , and injectivity of @xmath305 then yields @xmath1676 , which completes the proof , the final clause being an immediate consequence of that in lemma [ nontr ] .
given a compact geodesic - gradient khler triple @xmath380 , we use the notation of ( [ sym ] ) and ( [ tmm ] ) to set , for @xmath1677 and @xmath1678 , with either fixed sign @xmath35 , @xmath1679 thus , @xmath1680 , as @xmath655 are tangent , and @xmath81 normal , to the totally geodesic leaf through @xmath1046 of the @xmath727-invariant integrable distribution @xmath1681 , cf .
( [ hpm ] ) , theorem [ dcomp](c ) , corollary [ totgd ] , and the first line of remark [ ptnrs ] . also , denoting by @xmath1682 the endomorphism @xmath1683 of @xmath1684
, one has @xmath1685=[z_y^\pm\nh(\xi,\eta)]j\hskip-2pt_y\w\hh,\ ] ] as an obvious consequence of ( [ rcm ] ) and ( [ gjw ] ) .
next , we define a complex - bilinear mapping @xmath1686 by @xmath1687 by @xmath1688 we mean here ` the space of antilinear operators ' and @xmath1689 is treated as a complex vector space in which the multiplication by @xmath1690 acts via composition with @xmath1691 from the left . ( the product thus equals the given operator @xmath1692 followed by @xmath1691 . ) antilinearity of @xmath1693 and complex - bilinearity of @xmath1694 are both obvious from ( [ zex ] ) .
[ ctrkm ] for a compact geodesic - gradient khler triple @xmath380 , a fixed sign @xmath35 , and any point @xmath145 , the mapping @xmath1694 with ( [ mxw ] ) is a constant - rank multiplication , cf .
section [ cr ] .
furthermore , if @xmath1695 corresponds to @xmath1696 as in ( [ phi ] ) and @xmath50 is any nonzero vector normal to @xmath28 at @xmath1031 , then @xmath1697 whenever @xmath1698 and @xmath1699 , we have @xmath1700 in fact , let @xmath1701 as in theorem [ dcomp ] , with some fixed @xmath1084 . according to ( [ hvk ] ) and parts ( iii ) , ( iv ) , ( vi ) of theorem [ jacob ] , the vectors forming @xmath1136 are precisely the values @xmath951 for all @xmath81 as in theorem [ dcomp](e ) which also have the property that @xmath1702 whenever @xmath893 satisfies the hypotheses of theorem [ dcomp](e ) .
since the values @xmath1703 in theorem [ dcomp](h2 ) fill @xmath1047 ( cf .
assertions ( d ) ( f ) of theorem [ dcomp ] ) , replacing @xmath1113 and @xmath1704 in the last equality with the expressions provided by theorem [ dcomp](h2 ) and remark [ bthex ] , we easily verify , using ( [ rcm ] ) and remark [ inequ](i ) , that @xmath1705 if and only if @xmath1706 . now the final clause of theorem [ dcomp](b ) ( or , remark [ holom ] ) yields the first to simplify notations , let us write @xmath1707 rather than @xmath1708 . since @xmath1698 in ( [ krz ] ) and @xmath1359 is holomorphic ( theorem [ first ] ) , ( [ krz ] ) and remark [ holom ] clearly imply that , for a suitable integer @xmath32 , the resulting mapping @xmath1709 the @xmath92 version of the assumptions listed in remark [ dfrtl ] is thus satisfied if one chooses @xmath1710 to be @xmath1711 and sets @xmath1712 . by ( [ fxh ] )
, the differential of ( [ mpg ] ) at any nonzero @xmath1713 sends any @xmath1714 to the unique @xmath1715 , where @xmath1716 , with a linear lift @xmath1717 such that @xmath1718 equals the restriction of @xmath1719 to @xmath456 .
( we have @xmath1720 since @xmath1721 is real - bilinear and symmetric in @xmath1520 , cf .
( [ zex ] ) . )
consequently , @xmath1722 complex - linearity of the differential , due to ( [ mpg ] ) , means that ( [ cpl ] ) will still hold if we replace @xmath521 with @xmath1723 and @xmath1724 with @xmath1725 .
then , from ( [ zex ] ) and ( [ cpl ] ) , @xmath1726 = -\nnh2j[z(\xi,\eta)\hh w]=-\nnh2z(\xi,\eta)\hh j\nh w$ ] .
in other words , @xmath1727 whenever @xmath1728 and @xmath1729 .
thus , by ( [ mxw ] ) , @xmath1730 , while the opposite inclusion is obvious since ( [ zex ] ) gives @xmath1731 , and so the expression @xmath1727 for @xmath1732 equals @xmath1733 . the equality @xmath1734 and ( [ krz ] ) ( [ mpg ] ) complete the proof .
the description of @xmath1735 in the lines preceding ( [ mpg ] ) also gives @xmath1736 which one sees taking the limit of the equality in theorem [ dcomp](h2 ) with @xmath1737 as @xmath1093 approaches the other endpoint @xmath1738 ( and so @xmath591 ) .
this section uses the notations listed at the beginning of section [ cc ] and the symbols @xmath1233 of remark [ dppdm ] .
any @xmath1049 leads to the assignment @xmath1739 @xmath1740 being defined by ( [ phe ] ) .
( due to ( [ hpm ] ) and ( [ tme ] ) , @xmath1145 is injective on @xmath1136 . )
[ dicho ] given any compact geodesic - gradient khler triple @xmath380 , one and only one of the following two cases occurs . 1 .
either the mappings ( [ xtd ] ) are all constant , for both signs @xmath35 , or 2 .
each of ( [ xtd ] ) , for both signs @xmath35 , descends to a nonconstant holomorphic embedding @xmath1741 , where @xmath834 is the projective space of @xmath1742 .
condition ( a ) holds if and only if @xmath8 satisfies ( i )
( iii ) in lemma [ intpr ] . in view of theorem [ ctrkm ]
, we may use lemma [ holem ] for @xmath1695 corresponding to @xmath1696 as in ( [ phi ] ) , concluding ( from an obvious continuity argument ) that , with either fixed sign @xmath35 , all the mappings ( [ xtd ] ) descend to holomorphic embeddings of @xmath834 unless they are all constant . their constancy for one sign implies , however , the same for the other , since it amounts to ( ii ) or ( iii ) in lemma [ intpr ] , while ( ii ) and ( iii ) are equivalent .
this completes the proof .
[ infty]case ( a ) of theorem [ dicho ] is equivalent to ( [ spn ] ) , as one sees combining lemma [ intpr](i ) with ( [ hpm ] ) .
according to ( iv ) ( vi ) in lemma [ intpr ] ) , the immersions of theorem [ tgimm](c ) are then embeddings and their images form the leaves of foliations on @xmath1437 , both of which have the same leaf space @xmath270 .
[ nncst]when ( b ) holds in theorem [ dicho ] , images of the totally geodesic holomorphic immersions of theorem [ tgimm](c ) pass through every point @xmath145 , realizing an uncountable family of tangent spaces : the image of the embedding ( [ xtd ] ) .
we continue using the asumptions and notation of section [ he ] . the leaf space @xmath1743 of the integrable distribution @xmath871 on @xmath10 , cf .
lemma [ dvgww](a ) , carries a natural structure of a compact complex manifold of complex dimension @xmath1744 , with @xmath821 , such that the quotient - space projection @xmath1745 forms a holomorphic fibration and , for either sign @xmath35 , the projectivization @xmath1746 of the normal bundle @xmath1350 , defined as in ( [ prz ] ) , is biholomorphic to @xmath1747 via the biholomorphisms sending each complex line @xmath1666 through @xmath174 in the normal space of @xmath28 at any point to the @xmath1748-image of the punctured radius @xmath573 disk in @xmath1749 , the latter image being a leaf of @xmath234 according to lemma [ vsbkr](a ) .
the mappings ( [ dbp ] ) , restricted to @xmath34 , descend to holomorphic bundle projections @xmath1750 also denoted by @xmath1048 , which , under the biholomorphic identifications @xmath1751 of the preceding paragraph , coincide with the bundle projections @xmath1752 .
the restrictions @xmath1753 given by ( [ phe ] ) with the two possible signs @xmath594 are biholomorphisms ( theorem [ first ] ) , and hence so is the composite of one of them followed by the inverse of the other . at the same time , by theorem [ first](iii ) , either of them descends to a bijection @xmath1754 , and the composite just mentioned yields a biholomorphism between @xmath1755 and @xmath1756 .
this turns @xmath1747 into a compact complex manifold in a manner independent of the bijection used .
our assertion is now immediate from ( [ pcf ] ) .
[ trint]the direct sum of the two vertical distributions @xmath16 of the projections ( [ hbp ] ) is a distribution on @xmath1757 , since , at every point @xmath1758 , they intersect trivially : @xmath1759 .
in fact , as a consequence of ( [ tme ] ) , the original vertical distributions on @xmath34 , given by ( [ hpm ] ) , intersect along @xmath228 . for a grassmannian triple @xmath8 obtained as in section [ eg ] from some data ( [ dta].i ) ,
the descriptions of @xmath28 provided by ( [ spm].a ) , and @xmath1760 the equality meaning a natural biholomorphic identification .
if @xmath8 is in turn a cp triple , arising from ( [ dta].ii ) , @xmath28 must be as in ( [ spm].b ) , and ( [ mpv ] ) is replaced by @xmath1761 , while @xmath15 in ( [ hbp ] ) then become the factor projections .
[ ttlds ] for a fidimensional complex vector space @xmath418 , any @xmath1762 , and @xmath1747 given by ( [ mpv ] ) , let @xmath1763 .
then there exist an integer @xmath1764 and @xmath1765 , @xmath1766 , with @xmath1767 and whenever @xmath1768 , the notation @xmath1769 meaning that @xmath1770 or @xmath1771 . if @xmath1772 , our claim is obvious as @xmath1773 .
otherwise we may first choose @xmath1774 and @xmath1775 such that @xmath1776 , and then select @xmath1777 along with @xmath1778 spanned by @xmath1775 and a vector in @xmath1779 .
now @xmath1780 and @xmath1781 .
this step may be repeated for @xmath1782 instead of @xmath1783 , as long as @xmath1784 .
[ strbg ] let @xmath8 be any grassmannian triple , arising from some data ( [ dta].i ) as in section [ eg ] .
then the direct sum @xmath1785 appearing in lemma [ intpr](i ) is a strongly bracket - generating distribution on @xmath34 , in the sense that any two points of @xmath404 can be joined by a piecewise @xmath92 curve tangent to @xmath1786 . according to ( [ mpv ] ) , whenever @xmath1787 in lemma [ ttlds ] , both @xmath1788 and @xmath1789 must lie in the same fibre of one of the bundle projections ( [ hbp ] ) .
as the fibres of either projection ( [ hbp ] ) , being complex projective spaces ( see the last line in lemma [ lfspc ] ) , are connected , the strong bracket - generating property thus follows for the direct - sum distribution of remark [ trint ] .
our claim is now immediate since @xmath1785 projects onto that latter distribution under the quotient - space projection @xmath1790 , which also has connected fibres ( biholomorphic to twice - punctured complex projective lines , cf .
lemma [ vsbkr](b ) ) .
[ nintg]a compact geodesic - gradient khler triple need not , in general , satisfy conditions ( i ) ( iii ) of lemma [ intpr ] , that is , ( [ spn ] ) .
examples are provided by all grassmannian triples @xmath8 arising via lemma [ chone ] from data ( [ dta].i ) such that @xmath1791 , where @xmath492 .
namely , in ( [ dmq ] ) , @xmath1792 as @xmath1793 ( see remark [ grass ] ) and , similarly , @xmath1794 from ( [ spm].a ) ( [ spm].b ) , where @xmath1795 by ( [ dta].i ) .
thus , @xmath1796 and @xmath1785 in ( [ tme ] ) is a proper subbundle of @xmath17 .
consequently , due to corollary [ strbg ] , it can not be integrable .
[ twweb]for any compact geodesic - gradient khler triple @xmath380 , the leaf space @xmath1747 carries what might be called a holomorphic @xmath327-web of complex projective spaces , formed by the two holomorphic fibrations ( [ hbp ] ) with fibres biholomorphic to ( posidimensional ) complex projective spaces , having the trivial - intersection property of remark [ trint ] .
there is also a natural holomorphic complex line bundle over @xmath1757 , the restriction of which to every fibre of @xmath1797 ( or , @xmath1798 ) , with ( [ hbp ] ) , is biholomorphically isomorphic to the tautological ( or , respectively , dual tautological ) bundle of the fibre . specifically , the complex line attached to a leaf @xmath1035 of @xmath234 is @xmath1799 , cf .
theorem [ first](iii ) ; that changing the sign @xmath594 to @xmath1800 leads to its dual complex line follows from ( * ? ? ?
* remark 4.1 ) and ( [ hgv].a ) ( [ hgv].b ) . , transformation groups in differential geometry , reprint of the 1972 edition , classics in mathematics , springer - verlag , berlin , 1995 . , foundations of differential geometry , vol.1.9pti , wiley , new york , 1963 .
, on rotationally symmetric hamilton s equation for kheinstein metrics , in : `` recent topics in differential and analytic geometry '' ( t. ochiai , ed . ) , adv .
stud . pure math . * 18*i , academic press , boston , ma , 1990 , 327337 . | a vector field on a riemannian manifold is called geodesic if its integral curves are reparametrized geodesics .
we classify compact khler manifolds admitting nontrivial real - holomorphic geodesic gradient vector fields that satisfy an additional integrability condition .
they are all biholomorphic to bundles of complex projective spaces .
0.5 un champ de vecteurs sur une varit riemannienne est dit godsique si ses courbes intgrales sont godsiques non paramtrs . on classifie des
varits khlriennes compactes qui admettent des gradients godsiques rels -.64ptholomorphes -.64ptnon -.64pttriviaux -.64ptsatisfaisant -.64pt -.64ptune -.64ptcondition -.64ptadditionnelle -.64ptdintgrabilit .
elles sont toutes biholomorphes fibrs en espaces projectifs complexes
. [ section ] [ theorem]lemma [ theorem]corollary [ theorem]proposition [ theorem]claim [ theorem]definition [ theorem]remark [ theorem]example [ theorem]notation |
solitons play an important role in many areas of physics . as classical solutions of non - linear field theories
, they are localized structures with finite energy , which are globally regular . in general
, one can distinguish between topological and non - topological solitons .
while topological solitons @xcite possess a conserved quantity , the topological charge , that stems ( in most cases ) from the spontaneous symmetry breaking of the theory , non - topological solitons @xcite have a conserved noether charge that results from a symmetry of the lagrangian .
the standard example of non - topological solitons are @xmath0-balls @xcite , which are solutions of theories with self - interacting complex scalar fields .
these objects are stationary with an explicitly time - dependent phase . the conserved noether charge @xmath0 is then related to the global phase invariance of the theory and is directly proportional to the frequency .
@xmath0 can e.g. be interpreted as particle number @xcite .
while in standard scalar field theories , it was shown that a non - normalizable @xmath1-potential is necessary @xcite , supersymmetric extensions of the standard model ( sm ) also possess @xmath0-ball solutions @xcite . in the latter case ,
several scalar fields interact via complicated potentials .
it was shown that cubic interaction terms that result from yukawa couplings in the superpotential and supersymmetry ( susy ) breaking terms lead to the existence of @xmath0-balls with non - vanishing baryon or lepton number or electric charge .
these supersymmetric @xmath0-balls have been considered as possible candidates for baryonic dark matter @xcite and their astrophysical implications have been discussed @xcite . in @xcite , these objects have been constructed numerically using the exact form of a scalar potential that results from gauge - mediated susy breaking . however , this potential is non - differentiable at the susy breaking scale . in @xcite a differentiable approximation of this potential
was suggested and the properties of the corresponding @xmath0-balls have been investigated .
@xmath0-ball solutions in @xmath2 dimensions have been studied in detail in @xcite using a @xmath1-potential .
it was realized that next to non - spinning @xmath0-balls , which are spherically symmetric , spinning solutions exist .
these are axially symmetric with energy density of toroidal shape and angular momentum @xmath3 , where @xmath0 is the noether charge of the solution and @xmath4 corresponds to the winding around the @xmath5-axis . approximated solutions of the non - linear partial differential equations were constructed in @xcite by means of a truncated series in the spherical harmonics to describe the angular part of the solutions .
the full partial differential equation was solved numerically in @xcite .
it was also realized in @xcite that in each @xmath6-sector , parity - even ( @xmath7 ) and parity - odd ( @xmath8 ) solutions exist .
parity - even and parity - odd refers to the fact that the solution is symmetric and anti - symmetric , respectively with respect to a reflection through the @xmath9-@xmath10-plane , i.e. under @xmath11 .
complex scalar field models coupled to gravity possess so - called `` boson star '' solutions @xcite . in @xcite boson stars have been considered that have flat space - time limits in the form of @xmath0-balls .
these boson stars are hence self - gravitating @xmath0-balls .
the gravity gauge theory duality @xcite has attracted a lot of attention in the past years .
the most famous example is the ads / cft correspondence @xcite which states that a gravity theory in a @xmath12-dimensional anti - de sitter ( ads ) space time is equivalent to a conformal field theory ( cft ) on the @xmath13-dimensional boundary of ads . recently
, this theory has been used to describe so - called holographic conductor / superconductor phase transitions with the help of black holes in higher dimensional space
time @xcite .
close to the horizon of the black hole the effective mass of the scalar field can become negative with masses below the breitenlohner
freedman bound @xcite such that the scalar field becomes unstable and possesses a non vanishing value on and close to the horizon of the black hole .
interestingly , the model used to describe conductor / superconductor phase transitions by using black holes in higher - dimensional ads can be modified to describe insulator / superconductor phase transitions by considering soliton solutions in higher - dimensional ads .
this was done in @xcite and @xcite in ( 4 + 1)-dimensional ads in the probe limit and taking backreaction into account , respectively as well as in ( 3 + 1)-dimensional ads in @xcite .
the ads soliton is related to the black hole by a double wick rotation .
moreover , one of the coordinates is compactified to a circle .
the ads soliton has been used before in the context of the description of a confining vacuum in the dual gauge theory @xcite since it possesses a mass gap .
while for spherically symmetric black holes in ads there is the hawking - page phase transition from the ads black hole to global ads space - time when lowering the temperature @xcite , this is different for ads black holes with ricci - flat horizons used in the description of holographic superconductors . in that case
, there is a phase transition between the ads black hole and the ads soliton @xcite which was interpreted as a confining / deconfining phase transition in the dual gauge theory . now taking the viewpoint of condensed matter this phase transition describes a 1st order insulator / conductor phase transition . to complete the picture it was realized in @xcite that one can add a chemical potential to the ads soliton .
increasing the chemical potential @xmath14 the ads soliton becomes unstable to the formation of scalar hair above some critical value @xmath15 .
this was interpreted as a 2nd order insulator / superconductor phase transition that is possible even at zero temperature .
the study done in the probe limit in @xcite was extended to take backreaction into account @xcite .
phase diagrams were constructed and it was shown that there is a new type of phase transition from a superconductor to an insulator for sufficiently strong backreaction .
qualitatively similar results exist when describing insulator / superfluid phase transitions with non - vanishing superfluid velocity @xcite .
charged solitons in global ads have also recently been studied in 4 dimensions @xcite as well as in 5 dimensions @xcite .
the limit of vanishing gauge coupling has also been studied in @xcite . in that case
, the soliton solutions correspond to planar boson stars in ads space - time .
since the scalar field is uncharged the interpretation in terms of insulators / superconductors is difficult in this case .
however , since the ads / cft correspondence connects strongly coupled cfts to weakly coupled gravity theories the prototype example of a strongly coupled field theory comes to mind - quantum chromodynamics ( qcd ) .
as such the planar boson stars in ads have been interpreted as bose - einstein condensates of glueballs .
glueballs are color - neutral bound states of gluons predicted by qcd and the scalar glueball ( which is also the lightest possible glueball ) is predicted to have a mass of 1 - 2 gev ( see e.g. @xcite for an overview on experimental results ) .
since these glueballs appear due to non - linear interactions and as such can not be described by a perturbative approach it is very difficult to make predictions within the framework of quantum field theory .
however , holographic methods have been applied to make predictions about glueball spectra ( see e.g. @xcite and reference therein ) .
the spherically symmetric counterparts of boson stars in ads space - time have been studied before using a massive scalar field without self - interaction @xcite . in this paper
, we will extend the results of @xcite by including a self - interaction potential of exponential form motivated from supersymmetric extensions of the standard model @xcite .
in contrast to the study in @xcite our solutions exist also in the limit of vanishing gravitational coupling and correspond to @xmath0-balls in an ads background .
moreover , in @xcite the frequency appearing in the periodic time - dependence of the solutions has been scaled to unity . in this paper , we will explicitly keep this parameter .
our paper is organised as follows : in section 2 , we give the model , the ansatz and the equations of motion . in section 3 and 4 ,
respectively , we discuss @xmath0-balls in an ads background and boson stars in ads space - time . in section 5 ,
we give the holographic interpretation of our solutions , while we conclude in section 6 .
the existence of @xmath0-balls in supersymmetric extensions of the standard model has been suggested some time ago . in @xcite these @xmath0-balls
have been constructed using the following potential @xmath16 where @xmath17 corresponds to the scale below which supersymmetry is broken and is roughly @xmath18 tev , while @xmath6 denotes the scalar boson mass .
this potential is not differentiable at @xmath19 .
this is why the following approximation of the above potential has been suggested @xcite @xmath20 in the following we will study non - spinning @xmath0-balls and boson stars in a model containing this potential in asymptotically anti - de sitter ( ads ) space - time .
the action @xmath21 reads : @xmath22 where @xmath23 is the ricci scalar , @xmath24 denotes newton s constant , @xmath25 is the negative cosmological constant with @xmath26 the anti - de sitter radius and @xmath27 is the matter lagrangian given by @xmath28 where @xmath29 denotes a complex scalar field and we choose as signature of the metric @xmath30 .
@xmath31 is the potential given in ( [ potential ] ) .
the matter lagrangian @xmath27 ( [ lag ] ) is invariant under the global u(1 ) transformation @xmath32 as such the locally conserved noether current @xmath33 , @xmath34 , associated to this symmetry is given by @xmath35 the globally conserved noether charge @xmath0 of the system then reads @xmath36 finally , the energy - momentum tensor is given by @xmath37 + \partial_{\mu } \phi^ * \partial_{\nu } \phi + \partial_{\nu}\phi^ * \partial_{\mu } \phi\ .\end{aligned}\ ] ] for the metric we use the following ansatz in spherical schwarzschild - like coordinates @xmath38 where @xmath39 the non - vanishing components of the einstein tensor in this space - time read @xmath40 @xmath41 @xmath42 and @xmath43 . for the complex scalar field
, we use a stationary ansatz that contains a periodic dependence of the time - coordinate @xmath44 : @xmath45 where @xmath46 is a constant and denotes the frequency .
the coupled system of ordinary differential equations is then given by the einstein equations @xmath47 with @xmath48 given by ( [ em ] ) and the klein - gordon equation @xmath49 in order to be able to use dimensionless quantities we introduce the following rescalings @xmath50 and find that the equations depend only on the dimensionless coupling constant @xmath51 where @xmath52 is the planck mass .
note that @xmath53 denotes the ratio between the susy breaking scale and the planck mass which we expect to be small @xmath54 if the susy breaking scale is on the order of @xmath18 tev . in this paper
, we will however also study larger values of @xmath53 in order to understand the qualitative behaviour of the solutions in curved space - time .
note that with these rescalings the scalar boson mass @xmath55 becomes equal to unity .
the coupled ordinary differential equations then read @xmath56 @xmath57 @xmath58 where the prime now and in the following denotes the derivative with respect to @xmath59 .
note that ( [ eq1 ] ) corresponds to the @xmath60-component of the einstein equation , while ( [ eq2 ] ) is a combination of the @xmath60 and @xmath61 component : @xmath62 .
these equations have to be solved numerically subject to appropriate boundary conditions .
we want to construct globally regular solutions with finite energy . at the origin
we hence require @xmath63 while we choose @xmath64 ( any other choice would just result in a rescaling of the time coordinate ) .
moreover , while the scalar field function falls of exponentially for @xmath65 with @xmath66 it falls of power - law for @xmath67 with @xmath68 where @xmath69 is a constant that has to be determined numerically and which can be interpreted via ads / cft as the value of the condensate of glueballs in the dual theory living on the boundary of global ads .
the noether charge @xmath0 of the solutions is given by the following expression @xmath70 which is well finite also in asymptotically ads space - time .
for @xmath71 , we can read of the mass @xmath72 from the behaviour of the metric function @xmath73 at infinity .
this reads @xcite @xmath74 where @xmath75 is a constant that depends on @xmath76 .
for @xmath77 we have @xmath78 , @xmath79 and the mass of the solution is given by the volume integral of the energy - density , which is given by @xmath80 : @xmath81 in our case this gives @xmath82 r^2 dr \ , \ ] ] which is again a finite quantity in both asymptotically flat as well as asymptotically ads space - time .
for @xmath77 we find that @xmath79 and @xmath83 , hence @xmath84 . in this case
, we are studying a @xmath0-ball in the background of a global ads space - time .
hence , the space - time does not backreact .
the equation for the scalar field then reads @xmath85 we can rewrite the last term in ( [ eqnew ] ) as @xmath86 . and integrate over @xmath87 .
we then find @xmath88 the last term is a friction term , @xmath89 is an integration constant . in the classical mechanics picture
adapted e.g. in @xcite , where @xmath87 is considered to describe the position of the particle and @xmath59 the `` time - coordinate '' , this describes a particle with total energy @xmath89 subject to friction moving in a `` time''-dependent effective potential given by @xmath90 note that this potential becomes `` time''-independent for @xmath91 and @xmath65 , i.e. for @xmath0-balls in a fixed minkowski space - time since @xmath92 . in that case
, bounds for the frequency @xmath46 can be found .
these read @xcite @xmath93 such that @xmath94 and hence @xmath95 corresponds to a maximum of the potential and @xmath96 this latter condition on the frequency comes from the requirement that the potential @xmath97 is positive for some @xmath98 , i.e. @xmath99 for some @xmath98 .
in contrast to the @xmath100-potential used in @xcite we can always find a @xmath101 for any choice of @xmath102 $ ] that fulfills this condition .
+ as soon as either @xmath25 or @xmath24 are non - vanishing the metric function @xmath103 depends on @xmath59 . in fig.[potentialv ] we show the potential @xmath97 for @xmath91 , different values of @xmath25 , @xmath46 and of the `` time''-coordinate @xmath59 .
fig.[pot1 ] and fig.[pot2 ] show the potential for fixed @xmath104 . in fig.[pot1 ]
the cosmological constant is fixed @xmath105 and @xmath46 varied .
we observe that the potential becomes more narrow when increasing @xmath46 . in fig.[pot2 ] we show the potential for @xmath106 and different values of @xmath25 . here , we observe that the potential flattens when decreasing @xmath25 .
finally , in fig.[pot3 ] we show the potential for @xmath106 , @xmath105 and different values of the `` time''-coordinate @xmath59 .
as expected , the potential flattens for increasing @xmath59 such that in the limit @xmath107 the potential tends to zero .
it is important to note that in contrast to the @xmath100-potentials used in previous studies of @xmath0-balls in minkowski space - time @xcite our potential does not possess local maxima at some @xmath98 .
we have to construct the solutions to ( [ eqnew ] ) numerically .
we have done this applying a newton - raphson method that uses an adaptive grid scheme @xcite .
we have first studied the dependence of the mass @xmath72 and charge @xmath0 on @xmath46 for different values of @xmath25 .
this is shown in fig.[qballs ] .
we observe that the mass @xmath72 and charge @xmath0 of the @xmath0-balls diverge at @xmath108 , respectively @xmath109 in asymptotically flat space - time , i.e. for @xmath65 .
this is very similar to the qualitative behaviour in the case of a @xmath100-potential @xcite . decreasing the cosmological constant and hence considering @xmath0-balls in an ads background space - time that does not backreact , we find that @xmath72 and @xmath0 still diverge at @xmath108 .
moreover , we find that @xmath108 does not seem to depend on @xmath25 . however , at @xmath109 we find that now both @xmath0 and @xmath72 tend to zero - very similar to what happens for self - gravitating @xmath0-balls , so - called boson stars ( see e.g. @xcite ) .
moreover , @xmath109 increases with decreasing @xmath25 , e.g. we find @xmath110 for @xmath65 , @xmath111 for @xmath112 and @xmath113 for @xmath114 . at @xmath109 the value of @xmath95 and hence @xmath115 .
this is shown in fig .
[ qballs2 ] and is , of course , related to the fact that for @xmath116 the potential @xmath97 no longer possesses a maximum at @xmath117 , but rather a minimum . in flat space - time
( @xmath65 ) the profile of @xmath87 would hence become more and more spread out over all space and just before becoming zero everywhere would be nearly constant ( and non - vanishing ) on a larger domain of space - time .
this is what makes the integrals for @xmath0 and @xmath72 diverge . on the other hand , as soon as @xmath118 the ads background acts as a confining box and the integral never diverges . at @xmath108
we find that @xmath119 tends to very large values .
this is see in fig.[qballs2 ] .
the profile of @xmath120 becomes hence very sharply peaked on a small interval of @xmath59 which makes the integrals for the charge and mass diverge , even if @xmath118 .
this changed behaviour of the charge and mass at @xmath109 for @xmath118 now leads to qualitatively different plots for the charge @xmath0 over the mass @xmath72 .
this is shown in fig.[qballs3 ] , where we show @xmath0 as function of @xmath72 and compare @xmath72 with the mass of @xmath0 free boson @xmath121 ( in our rescaled units ) . while for @xmath65 there exist two solutions with different @xmath72 for a given charge , one of which has @xmath122 ( stable branch ) and the other @xmath123 ( unstable branch ) , there exists only one branch of solutions as soon as @xmath118 .
the @xmath0-balls in ads space - time exist for arbitrarily small values of the mass and the charge .
we observe that the mass drops below @xmath124 for sufficiently large values of the charge .
we would hence expect that the @xmath0-balls are unstable to a decay into @xmath0 free bosons with mass @xmath125 for small charges , but stable against this decay for large charges @xmath0 .
we also observe that the critical value of @xmath0 , @xmath126 at which the mass of the @xmath0-ball is @xmath127 depends on @xmath25 .
the smaller @xmath25 the larger @xmath126 such that for @xmath128 the @xmath0-balls are unstable , while for @xmath129 the @xmath0-balls are stable .
a smaller @xmath25 corresponds to a larger absolute value of @xmath25 and hence to a smaller ads radius .
we show the mass @xmath72 as function of the charge @xmath0 for @xmath0-balls in an anti - de sitter background for different values of the cosmological constant @xmath25 . for comparison
we also show the values for @xmath0-balls in a flat space - time background ( @xmath65 ) .
note that the line @xmath130 corresponds to the mass of @xmath0 free boson @xmath131 with mass @xmath132 ( in our rescaled variables).,width=264 ] next to the fundamental solutions presented above , it has been observed previously @xcite that @xmath0-ball solutions with a finite number @xmath133 , @xmath134 , of nodes of the scalar field function @xmath120 exist .
these solutions are typically interpreted as radially excited solutions .
we also find these solutions for our scalar potential and when including a negative cosmological constant . in fig.[nodes_profiles ] we show the profiles of the solutions with @xmath135 nodes for @xmath136 , where the @xmath137 solution corresponds to the fundamental @xmath0-ball solution discussed above .
we observe that the smallest value of @xmath59 at which @xmath120 becomes zero decreases with increasing @xmath133 .
moreover , the zeros appear at small values of @xmath59 , i.e. close to the core of the @xmath0-ball .
we show the profile of the scalar field function @xmath120 for @xmath0-balls with @xmath135 nodes , respectively , for @xmath136 .
, width=264 ] we show the mass @xmath72 as function of the charge @xmath0 for @xmath0-balls in an anti - de sitter background with @xmath136 for the @xmath135 solutions .
the dashed line is @xmath130 and corresponds to the mass of @xmath0 free bosons with mass @xmath121 ( in our rescaled variables ) .
, width=264 ] + now it is interesting to see how the mass @xmath72 and charge @xmath0 of the @xmath0-ball solutions with nodes depend on @xmath46 .
this is shown in fig.[qballsex ] for @xmath136 and @xmath135 , respectively .
our results suggest that the maximal value of @xmath46 up to which the solutions exist depends on the number of nodes .
we find that @xmath138 increases with increasing node number , i.e. radially excited @xmath0-ball solutions exist for larger values of the frequency @xmath46 .
moreover , both the charge ( see fig.[charge_qballex ] ) as well as the mass ( see fig.[mass_qballex ] ) increase with @xmath133 for a fixed value of @xmath46 .
this is not surprising since these are radially excited solutions . in fig.[qballsex2 ] we show the mass as function of the charge for @xmath0-balls in an ads background with @xmath136 and for @xmath135 .
clearly , for a given charge @xmath0 the mass @xmath72 of the solutions increases with increasing @xmath133 .
moreover , the mass @xmath72 of the @xmath139 solutions is always above that of @xmath0 free bosons with mass @xmath121 , while for @xmath137 it is above that curve for small @xmath0 and below for large @xmath0 . of course
, we would expect the excited solutions to be unstable and our numerical results here confirm exactly this expectation .
here , we choose @xmath71 . in this case , we have to solve the full set of coupled and non - linear equations ( [ eq1])-([eq3 ] ) numerically .
again , we have done this with a newton - raphson method @xcite . in fig.[bs1 ]
we show the charge @xmath0 and @xmath72 of the boson stars in dependence on the frequency @xmath46 for different values of @xmath25 and two different values of @xmath53 .
we observe that the local maximum of the charge and mass , respectively , that exists close to @xmath140 for @xmath65 and @xmath71 disappears as soon as @xmath118 . moreover , as for @xmath77 the value of @xmath140 at which the charge and mass become zero does not depend on @xmath53 .
this has been observed before in asymptotically flat space - times ( see e.g. @xcite ) and seems to be true also for @xmath118 .
for @xmath71 the solutions exist down to a minimal finite value of @xmath141 .
this @xmath142 depends on @xmath53 as well as on @xmath25 .
it increases with increasing @xmath53 and increases with decreasing @xmath25 , i.e. the larger the curvature of the space - time and the larger the interaction between the curvature and the matter content , the larger @xmath142 . at @xmath142
a second branch of solutions exists that has smaller mass and charge than the first ( main ) branch of solutions .
we have only constructed the second branch of solutions here , but believe that in analogy to boson stars in asymptotically flat space - time several further branches exist that form a spiral .
+ in fig.[bs2 ] we show the charge @xmath0 as a function of the mass @xmath72 of boson stars for different values of @xmath53 and for the asymptotically flat case ( see fig.[charge_bs4 ] ) as well as for two asymptotically ads cases with @xmath143 ( see fig.[mass_bs5 ] ) and @xmath136 ( see fig.[charge_bs6 ] ) , respectively .
we observe that the features of these plots change qualitatively when either choosing @xmath53 or @xmath25 non - vanishing . for @xmath144 ,
i.e. in the case of a fixed minkowski space - time we find that the solutions exist only down to a minimal , non - vanishing value of the charge and the mass and that from there a second branch of solutions exists .
the first branch of solutions has lower mass than the mass of @xmath0 free scalar boson with mass @xmath125 , while the second branch has ( within our numerical accuracy ) mass equal to the mass of @xmath0 scalar bosons .
we would hence expect that the lower branch of @xmath0-balls is stable , while the second branch possesses an instability .
choosing @xmath71 for @xmath65 , we find that the mass and charge of the boson stars in asymptotically flat space - time are bounded from above , i.e. boson stars for the exponential scalar potential employed here exist only up to a maximal value of the charge and mass .
this is similar to the observation for other potentials ( see e.g. @xcite ) .
moreover , all solutions have mass lower than that of @xmath0 free scalar bosons .
this is natural since gravity acts to from a bound state out of the scalar bosons .
our results further suggest that the larger @xmath53 , the smaller is the maximal value of the mass and charge , i.e. the closer the susy breaking scale is to the planck mass , the smaller the maximal mass and charge of the corresponding boson stars will be . for @xmath118 and @xmath77 the solutions exist for arbitrary values of the charge and the mass and
there is only one branch of solutions . for small @xmath0 this branch of solutions
is below the @xmath130 curve of @xmath0 scalar bosons , while for large @xmath0 it is below this curve . on the other hand , for @xmath118 and @xmath71 ,
the solutions exist up to a maximal value of the charge and mass and from these values on , a second branch of solutions exists .
again , the maximal value of the mass and charge depend strongly on @xmath53 , while the dependence on @xmath25 seems to be very small ( at least for the values of @xmath25 that we have chosen here ) .
interestingly , the solutions on both branches have mass below @xmath130 .
+ + + +
in the following , we want to use the ads / cft correspondence to interpret our solutions in ads space - time in terms of a bose - einstein condensate of scalar glueballs .
this has been suggested in @xcite for planar boson star solutions .
here , we are studying global ads and hence our dual and strongly coupled cft lives on the surface of a 2-sphere .
we have first studied the case @xmath77 , i.e. neglecting the backreaction of the space - time on the condensate of glueballs . in fig.[holo1 ] we show the value of the condensate @xmath145 in dependence on @xmath119 ( a ) , in dependence on the charge @xmath0 ( b ) and in dependence on the mass @xmath72 ( c ) , respectively .
we find that the condensate is an increasing function of @xmath119 , @xmath0 and @xmath72 , respectively .
moreover , decreasing the cosmological constant increases the value of the condensate for a fixed value of @xmath119 , @xmath0 and @xmath72 .
this is related to the fact that the increase of the absolute value of the negative cosmological constant is , of course , related to a decrease of the radius of the ads space - time .
this in turn allows to built up larger condensates .
we find that at some intermediate values of @xmath119 , @xmath0 and @xmath72 the condensate has a maximal value .
our numerical results suggest that the larger the absolute value of @xmath25 the smaller in the value of @xmath119 at which the condensate reaches its maximum . however , the corresponding values of the mass and charge at which the maximal condensate appears are significantly larger than the values of the mass and charge necessary to form a condensate in the first place . we have also studied the value of the condensate for the radially excited @xmath0-ball solutions mentioned above .
our results are shown in fig.[holo_ex1 ] , where we present @xmath145 in dependence on @xmath119 ( a ) , in dependence on the charge @xmath0 ( b ) and in dependence on the mass @xmath72 ( c ) , respectively , in an ads background with @xmath146 and for the fundamental solution without nodes ( @xmath137 ) and for the radially excited solutions with @xmath139 nodes .
it is apparent that the value of the condensate increases with increasing node number for fixed @xmath119 , @xmath72 or @xmath0 .
hence , larger condensates can built up in the presence of radially excited @xmath0-ball solutions . however , since we expect these radially excited solutions to be unstable to decay to the fundamental solution , these condensates are very likely unstable .
it would have to be computed explicitely how long - lived these excited solutions are - and hence of how much physical relevance these condensates are .
this corresponds to @xmath71 .
our results for fixed @xmath146 and different values of the backreaction are shown in fig.[holo2 ] .
plotting the condensate as function of @xmath119 ( see fig.[holo2_1 ] ) we find that a value of @xmath147 exists at which the condensate has a maximal value .
this was also observed for planar boson stars in @xcite .
we find that the larger the backreaction , i.e. the larger @xmath53 the smaller is @xmath148 .
moreover , at a critical value of @xmath149 , the condensate tends to zero again .
hence , glueball condensates are only possible in a finite range of the parameter @xmath119 .
the interval of @xmath119 on which glueball condensates exist decreases when increasing the backreaction .
when plotting the condensate as function of the charge @xmath0 and the mass @xmath72 ( see fig.[holo2_2 ] and fig.[holo2_3 ] , respectively ) we find that glueball condensates exist only up to a maximal finite mass of the charge and mass .
moreover , the value of the condensate increases for fixed @xmath0 or @xmath72 when increasing the backreaction . this is a standard observation in applications of the ads / cft correspondence applied to strongly coupled qfts ( see e.g. @xcite and references therein ) .
interestingly , we observe that in a small interval close to the maximal charge @xmath0 ( respectively maximal mass @xmath72 ) two different values of the condensate are possible for a fixed @xmath72 and @xmath0 .
one of the branches corresponds to a stable boson star , while the other would probably be unstable . for the unstable boson star ,
the condensate can be larger .
however , it would likely be much shorter lived than that building up for a stable boson star . in order to decide
how short - lived this is , one would probably have to simulate the dynamical decay of the unstable boson star to the stable one .
this is out of the scope of the present paper .
we have also studied the condensate building up on radially excited boson stars in ads space - time .
our results for @xmath150 and @xmath105 are given in fig.[holo2ex ] .
we show the condensate for the fundamental solution ( @xmath137 ) as well as for the two radially excited solutions with @xmath139 .
it is apparent that the values of the condensate can be much higher in the space - time of radially excited boson stars .
this is similar to what we have observed for the @xmath0-balls in an ads background ( see above ) .
we find that the difference of the condensate on the first and second branch of solutions increases with increasing node number .
while the difference between the two branches for @xmath137 is quite small , it is already quite big for @xmath151 .
again , we expect these solutions to be unstable and hence the condensates to decay .
in this paper , we have studied ads @xmath0-balls and boson stars that arise in a complex scalar field model with exponential potential .
this potential appears in gauge - mediated susy breaking in the mssm .
we find that the qualitative features in the case of an asymptotically flat space - time are similar to those observed for other potentials .
when choosing a non - vanishing cosmological constant the qualitative features change .
a study on boson stars in ads already exists for a massive scalar field without self - interaction @xcite , however , in this latter paper the frequency has been scaled to one . in the present paper , we have kept the frequency as a parameter and have studied the dependence of the mass and charge of the solutions on this parameter .
we observe that @xmath0-balls in an ads background exist for arbitrary values of the charge and mass and that only one branch of solutions exists .
moreover , we find that radially excited solutions exist both for a minkowski as well as for an ads background . when including backreaction of the space - time , we find that boson stars exist only up to a maximal value of the mass and charge and that at the minimal value of the frequency a spiraling behaviour appears .
plotting the value of the scalar field on the ads boundary and interpreting our solutions as holographic duals of bose - einstein condensates of scalar glueballs , we find that a maximal value of the condensate exists in the probe limit as well as in the backreacted case .
moreover , for fixed values of the charge and mass of the solutions , we find that more than one value of the condensate is possible .
however , the condensate with the larger value is ( likely ) associated to an unstable solution and would hence decay . in this paper
, we have only studied non - spinning @xmath0-balls and boson stars .
it would be interesting to extend our results to the case of rotation .
these solutions have already been studied for a @xmath100-potential in asymptotically flat space - time @xcite and we would expect that they also exist in asymptotically ads and with an exponential potential .
the question is then whether the boundary theory would also possess rotation and if one could hence describe rotating bose - einstein condensates of glueballs with our model .
this is currently under investigation .
+ + * acknowledgement * we gratefully acknowledge the deutsche forschungsgemeinschaft ( dfg ) for financial support within the framework of the dfg research training group 1620 _ models of gravity_. 99 n.s . manton and p.m. sutcliffe , _ topological solitons _ , cambridge university press , 2004 .
r. friedberg , t. d. lee and a. sirlin , phys .
d * 13 * ( 1976 ) 2739 .
t. d. lee and y. pang , phys . rep .
* 221 * ( 1992 ) , 251 .
s. r. coleman , nucl .
b * 262 * ( 1985 ) , 263 . m.s .
volkov and e. whnert , phys .
d * 66 * ( 2002 ) , 085003 .
a. kusenko , phys .
b * 404 * ( 1997 ) , 285 ; phys .
b * 405 * ( 1997 ) , 108 .
a. kusenko , hep - ph/0009089 .
k. enqvist and j. mcdonald , phys .
b * 425 * ( 1998 ) , 309 ; s. kasuya and m. kawasaki , phys . rev .
d * 61 * ( 2000 ) , 041301 ; a. kusenko and p. j. steinhardt , phys .
* 87 * ( 2001 ) , 141301 ; t. multamaki and i. vilja , phys .
b * 535 * ( 2002 ) , 170 ; m. fujii and k. hamaguchi , phys . lett .
b * 525 * ( 2002 ) , 143 ; m. postma , phys . rev .
d * 65 * ( 2002 ) , 085035 ; k. enqvist , _ et al .
_ , phys .
b * 526 * ( 2002 ) , 9 ; m. kawasaki , f. takahashi and m. yamaguchi , phys . rev .
d * 66 * ( 2002 ) , 043516 ; a. kusenko , l. loveridge and m. shaposhnikov , phys . rev .
d * 72 * ( 2005 ) , 025015 ; y. takenaga _ et al . _ [ super - kamiokande collaboration ] , phys .
b * 647 * ( 2007 ) , 18 ; s. kasuya and f. takahashi , jcap * 11 * ( 2007 ) , 019 . l. campanelli and m. ruggieri , phys .
d * 77 * ( 2008 ) , 043504 ; l. campanelli and m. ruggieri , phys . rev .
d * 80 * ( 2009 ) 036006 .
e. copeland and m. tsumagari , phys.rev .
d * 80 * 025016 ( 2009 ) .
b. kleihaus , j. kunz and m. list , phys .
d * 72 * ( 2005 ) , 064002 .
b. kleihaus , j. kunz , m. list and i. schaffer , phys .
d * 77 * ( 2008 ) , 064025 .
y. brihaye and b. hartmann , nonlinearity * 21 * ( 2008 ) , 1937 .
d. j. kaup , phys . rev . * 172 * ( 1968 ) , 1331 .
e. mielke and f. e. schunck , proc .
8th marcel grossmann meeting , jerusalem , israel , 22 - 27 jun 1997 , world scientific ( 1999 ) , 1607 .
r. friedberg , t. d. lee and y. pang , phys .
d * 35 * ( 1987 ) , 3658 .
p. jetzer , phys . rept . * 220 * ( 1992 ) , 163 .
f. e. schunck and e. mielke , class .
* 20 * ( 2003 ) r31 .
f. e. schunck and e. mielke , phys .
a * 249 * ( 1998 ) , 389 .
y. brihaye and b. hartmann , phys .
d * 79 * ( 2009 ) , 064013 .
o. aharony , s. s. gubser , j. m. maldacena , h. ooguri and y. oz , phys . rept .
* 323 * ( 2000 ) 183 ; e. dhoker and d. z. freedman , _ supersymmetric gauge theories and the ads / cft correspondence _ ,
arxiv : hep - th/0201253 ; m. benna and i. klebanov , _ gauge - string duality and some applications _ , arxiv : 0803.1315 [ hep - th ] . j. maldacena , adv .
* 2 * ( 1998 ) 231 ; int .
* 38 * ( 1999 )
1113 . s. s. gubser , phys .
d * 78 * ( 2008 ) 065034 .
s. a. hartnoll , c. p. herzog and g. t. horowitz , phys .
* 101 * ( 2008 ) 031601 ; jhep * 0812 * ( 2008 ) 015 ; g. t. horowitz and m. m. roberts , phys .
d * 78 * ( 2008 ) 126008 .
c. p. herzog , j. phys .
a * 42 * ( 2009 ) 343001 ; s. a. hartnoll , class .
* 26 * ( 2009 ) 224002 ; g. horowitz , _ introduction to holographic superconductors _ , arxiv:1002.1722 .
p. breitenlohner and d. z. freedman , annals phys .
* 144 * ( 1982 ) 249 .
t. nishioka , s. ryu and t. takayanagi , jhep * 03 * ( 2010 ) 131 .
g. t. horowitz and b. way , jhep * 1011 * ( 2010 ) 011 [ arxiv:1007.3714 [ hep - th ] ] .
y. brihaye and b. hartmann , phys .
d * 83 * , 126008 ( 2011 ) .
e. witten , adv .
math . phys .
* 2 * ( 1998 ) 505 .
g. horowitz and r. myers , phys .
d * 59 * ( 1998 ) 026005 .
s. hawkinga and d. page , commun .
* 87 * ( 1983 ) 577 .
s. surya , k. schleich and d. witt , phys .
* 86 * ( 2001 ) 5231 .
s. a. gentle , m. rangamani and b. withers , _ a soliton menagerie in ads _ , arxiv:1112.3979 [ hep - th ] .
o. j. c. dias , p.
figueras , s. minwalla , p. mitra , r. monteiro and j. e. santos , _ hairy black holes and solitons in global ads 5 _ , arxiv:1112.4447 [ hep - th ] . y. brihaye and b. hartmann , _ hairy charged gauss - bonnet solitons and black holes _ , arxiv:1203.3109 [ gr - qc ] . v. crede and c. a. meyer , prog .
* 63 * ( 2009 ) 74 . v. mathieu , n. kochelev and v. vento , int . j. mod
. phys .
e * 18 * ( 2009 ) 1 .
d. astefanesei and e. radu , nucl .
b * 665 * ( 2003 ) 594 [ gr - qc/0309131 ] .
u. ascher , j. christiansen and r. d. russell , math . comput . * 33 * ( 1979 ) , 659 ; acm trans .
* 7 * ( 1981 ) , 209 . | we study non - spinning @xmath0-balls and boson stars in 4-dimensional anti - de sitter ( ads ) space - time .
we use an exponential scalar field potential that appears in gauge - mediated supersymmetry ( susy ) breaking in the minimal supersymmetric extension of the standard model ( mssm ) .
we investigate the dependence of the charge and mass of these non - topological solitons on the negative cosmological constant , the frequency that appears in the periodic time - dependence as well as on the ratio between the susy breaking scale and the planck mass .
next to fundamental solutions without nodes in the scalar field function we also construct radially excited solutions . in the second part of the paper we put the emphasis on the holographic interpretation of these solutions in terms of bose - einstein condensates of scalar glueballs that are described by a strongly coupled quantum field theory ( qft ) on the boundary of global ads .
pacs numbers : 04.40.-b , 11.25.tq |
the muon spectrometer of the atlas experiment @xcite will be operated in the toroidal magnetic field of a superconducting air - core magnet system with 36 tm bending power.it is designed to provide stand - alone muon momentum resolution of @xmath4 for transverse momenta between 6 gev and 1 tev over a pseudo - rapidity range of @xmath5 .
this requires very accurate track sagitta measurement with three layers of muon chambers and high - precision optical alignment monitoring systems .
precision drift chambers , the monitored drift tube ( mdt ) chambers , have been developed to provide a track position resolution of @xmath0 m over an active area of 5500 m@xmath6 . ] ] ] ] ] ] lcc + parameter & x - ray fit & design value + @xmath7-pitch @xmath8m@xmath9 $ ] & @xmath10 & @xmath11 + @xmath12-pitch @xmath8m@xmath9 $ ] & @xmath13 & @xmath14 + @xmath15 @xmath16mm@xmath9 $ ] & @xmath17 & @xmath18 + @xmath19 @xmath8m@xmath9 $ ] & @xmath20 & @xmath21 + + parameter & x - ray fit & design value + @xmath7-pitch @xmath8m@xmath9 $ ] & @xmath22 & @xmath11 + @xmath12-pitch @xmath8m@xmath9 $ ] & @xmath13 & @xmath14 + @xmath15 @xmath16mm@xmath9 $ ] & @xmath23 & @xmath24 + @xmath19 @xmath8m@xmath9 $ ] & @xmath25 & @xmath21 + & & + [ -2 mm ] + parameter & x - ray fit & design value + @xmath7-pitch @xmath8m@xmath9 $ ] & @xmath26 & @xmath27 + @xmath12-pitch @xmath8m@xmath9 $ ] & @xmath28 & @xmath14 + @xmath15 @xmath16mm@xmath9 $ ] & @xmath29 & @xmath30 + @xmath19 @xmath8m@xmath9 $ ] & @xmath31 & @xmath21 + + parameter & x - ray fit & design value + @xmath7-pitch @xmath8m@xmath9 $ ] & @xmath32 & @xmath27 + @xmath12-pitch @xmath8m@xmath9 $ ] & @xmath28 & @xmath14 + @xmath15 @xmath16mm@xmath9 $ ] & @xmath33 & @xmath30 + @xmath19 @xmath8m@xmath9 $ ] & @xmath34 & @xmath21 + lccc + & high voltage end & readout end & center + @xmath7-coord . @xmath8m@xmath9 $ ] & @xmath35 & @xmath36 & 15.0 + @xmath12-coord . @xmath8m@xmath9 $ ] & @xmath37 & @xmath38 & 16.1 + combined @xmath8m@xmath9 $ ] & & 15.6 + & & + [ -2 mm ] + & high voltage end & readout end & center + @xmath7-coord .
@xmath8m@xmath9 $ ] & 14.3 & 15.3 & 13.2 + @xmath12-coord . @xmath8m@xmath9 $ ] & 10.5 & 14.1 & 7.7 + combined @xmath8m@xmath9 $ ] & & 10.5 + ] ] ] ] of the measured @xmath7 coordinates of the wires in the 6 tube layers ( distances between layers compressed ) with respect to the expected wire grid at the readout end of the bos 98 prototype chamber .
the uncompensated cross plate deformations during assembly of each layer are indicated .
[ res1 ] ] ] ( 93:7 ) gas mixture at 3 bar.[resol ] ] ] the mdt chambers ( see fig .
[ mdt ] ) consist of 3 or 4 layers of precise aluminum drift tubes with @xmath39 mm outer diameter and @xmath40 m wall thickness on either side of a space frame carrying an optical monitoring system to correct for chamber deformations .
the drift tubes are operated at a gas pressure of 3 bar to provide a single - tube position resolution of at least @xmath1 m ( rms ) with an ar : co@xmath41 ( 93:7 ) gas mixture and at the low gas gain of @xmath42 required to prevent ageing of the drift tubes at the high background rates at the lhc .
the sense wires of the drift tubes have to be positioned in individual tubes with an accuracy of @xmath43 m ( rms ) and in the whole chamber with an accuracy of @xmath2 m ( rms ) in order to obtain a chamber position resolution of @xmath0 m ( rms ) . in total 1200 mdt chambers containing 400000 drift tubes of 16 m length have to be constructed for the atlas muon spectrometer at 13 production sites over a period of 4 years . in munich , the production of 88 of the largest mdt chambers with 432 drift tubes of 3.8 m length in 6 layers and with a width of 2.16 m has started .
the drift tubes for these chambers are fabricated in a joint facility at the joint institute for nuclear research ( jinr ) in dubna , russia .
the first mdt chamber for the atlas detector ( named ` bos-0 ' ) has been completed in august 2000 . in spring 1998 , the full - scale prototype of a mdt chamber of this type ( named ` bos 98 ' ) has been built with the methods developed for large - scale production @xcite . with the prototype chamber , it has first been demonstrated that the required high mechanical accuracy can be achieved . over the last two years
, operation experience with mdt chambers was gained with the prototype in the muon test beam at cern .
for the prototype chamber , the sense wires were positioned and fixed at the tube ends using external references for tubes and wire and fast - curing glue ( see fig . [ glue ] ) .
the effect of glue shrinkage on the wire position was measured and taken into account .
the endplug of the drift tubes designed for this method ( see fig .
[ endplug1 ] ) does not require high precision in the fabrication .
after assembly , the wire positioning accuracy at the ends of the drift tubes with respect to the outer tube diameter was measured to be @xmath43 m ( rms ) in both coordinates including the non - roundness of the tubes ( see fig .
[ wp1_2d ] ) using a stereo x - ray technique with a resolution of @xmath44 m . for the large - scale production , a precisely machined endplug variant ( see fig .
[ endplug2 ] ) has been adopted where the wire is located in a spiral hole concentric with an aluminum reference ring on which the drift tube ends are positioned during chamber assembly ( see below ) .
the injection moulding process of the insulating plastic ( noryl ) body of the endplugs with the metal inserts has been carefully optimised in order to prevent stresses and the development of cracks which can make the drift tubes leak .
the x - ray measurements show a wire positioning accuracy of @xmath45 m ( rms ) ( see fig .
[ wp2_1d ] ) .
reliable ground contact of the aluminum tubes ( the cathodes of the drift tubes ) is provided by spot welds to the aluminum ring on the endplug using a specially developed laser welding technique employing filler wire .
the contact resistance at a current of 10 ma stays below 1 m@xmath46 even after an accelerated corrosion test with exposure to salt spray for 48 hours .
the drift tubes for serial production are assembled semi - automatically in a clean room of class 10000 with temperature and relative humidity controlled to be @xmath47c and @xmath48 , respectively .
they have to fulfil stringent quality criteria which include wire position measurement with x - rays within @xmath49 ( @xmath50 m ) in both coordinates , gas leak rate at 3 bar below @xmath51 bar@xmath52l / s , wire tension within @xmath53 of the nominal value of 350 g , and high voltage leakage current for ar : co@xmath41 ( 93:7 ) at 3 bar and 3.4 kv below 8 na . at present , the drift tube rejection rate is about @xmath54 . in the drift tubes of the bos 98 and the bos-0 chamber ,
the wire tension , determined from the measurement of the oscillation frequency @xmath55 of the @xmath56 m diameter gold - plated tungsten - rhenium ( 97:3 ) wire , is uniform within @xmath57 ( rms ) .
during the assembly of a mdt chamber @xcite , the ends of the drift tubes for each layer are positioned on precision aluminum combs with an accuracy of @xmath58 m ( rms ) in horizontal ( @xmath12 ) and vertical ( @xmath7 ) direction .
the combs have been produced industrially by spark erosion and are installed on a flat granite table in a clean room of class 100000 with temperature and relative humidity controlled to be @xmath47c and @xmath48 , respectively . over their whole length the tubes are held straight in 9 parallel rows of combs with vacuum suction .
the tube layers inserted in the combs are glued subsequently to the aluminum space frame ( three cross plates connected by two long beams ; see fig . [ table1 ] ) which for this purpose is positioned with respect to the combs with an accuracy of @xmath59 m in @xmath7 and @xmath12 on precision towers at the ends of the three cross plates ( see fig .
[ table1 ] ) .
the positioning of the space frame is monitored with laser beams and transparent optical position sensors @xcite-@xcite .
while the space frame is supported on the reference towers during glueing of a tube layer , the 2.16 m wide cross plates of the large chambers bend between the support points under the weight of the chamber by up to @xmath1 m at the ends and @xmath60 m in the middle .
the gravitational sag of the cross plates during assembly is measured with optical sensors installed on the cross plates ( see fig .
[ sag2 ] ) . in order to prevent deformations of the tube layers after glueing them to the space frame ,
the cross plates are also supported via the long beams applying forces with computer controlled pneumatic actuators until the sag is compensated without lifting the chamber from the reference towers .
after positioning the chamber on the reference towers , the pneumatic actuators instantaneously apply the required forces at the ends of 4 bars inserted through holes in the long beams close to the cross plates as shown in fig .
[ table1 ] .
the wire positions in the completed chambers have been measured at cern @xcite with scans with stereo x - ray sources perpendicular to the wires .
the reproducibility of the x - ray measurements of the wire coordinates was within @xmath61 m ( rms ) during the scans of the prototype chamber in 1998 and now is within @xmath58 m ( rms ) .
a fit of an ideal wire grid to the measured wire coordinates @xmath7 ( perpendicular to the tube layers ) and @xmath12 ( parallel to the tube layers ) allows a determination of the geometrical parameters of the chamber and to evaluate the wire positioning accuracy . in table
[ fit ] the fitted parameters , the horizontal ( @xmath12 ) and vertical ( @xmath7 ) wire pitch ( at @xmath62c ) and the @xmath7- and @xmath12-separations @xmath15 and @xmath19 between the two triple - layers , from the scans at the chamber ends are compared to the design values for the prototype chamber ( bos 98 ) and for the first production chamber ( bos-0 ) . for the bos 98 chamber , the x - ray results are the average of several scans at the same position along the tubes .
the @xmath15 values at both ends of this chamber differ because of a known common @xmath7-offset of the wire positions in the tubes at the readout end . during the assembly of the bos-0 chamber , an unexpectedly large shrinkage of the glue between tube layers and frame
was observed which is taken into account in the design values . with the different type of glue used for the bos 98 chamber , no significant glue shrinkage was observed . for both chambers ,
the fitted parameters agree very well with the design values .
distributions of the residuals of the measured wire coordinates @xmath7 and @xmath12 with respect to the fitted grid are shown in figs .
[ xray1 ] and [ xray2 ] for the two chambers .
the widths of the distributions for x - ray measurements at both ends of the chambers as well as near the center are summarised in table [ width ] .
the wire locations at intermediate positions between the tube ends are determined by the wire locations at the ends ( weighted averages ) and by the gravitational sag @xmath63 of the wires .
the wire sag is known from the measurement of the wire oscillation frequency @xmath55 via the relation @xmath64 where @xmath65 is the gravitational acceleration .
variations in the maximum wire sag of @xmath57 ( rms ) of the nominal @xmath66 m due to the variations of the wire tension are negligible . the residuals of the @xmath7-coordinates are shown in figs .
[ res1 ] and [ res2 ] .
the gravitational sag of the cross plates which has to be compensated for the assembly of each tube layer ( see fig . [ sag2 ] ) is indicated .
the bos-0 chamber , for cost reasons , has less stiff cross plates than the prototype chamber and therefore larger cross plate sag .
the statistical fluctuations of the wire locations are larger in the prototype chamber because the aluminum tube walls , instead of the precise endplugs , were used as references for the positioning of the drift tubes the wire positioning accuracy in the prototype chamber is better than the required @xmath2 m ( rms ) . with the first production chamber ,
a wire positioning accuracy of better than @xmath3 m ( rms ) has been achieved for one of the largest chamber types in the atlas muon spectrometer .
the main improvement with respect to wire positioning accuracy compared to the prototype chamber is the introduction of the aluminum reference ring on the endplugs which allows more precise positioning of the drift tube ends on the assembly combs than the aluminum tube walls used before .
based on the present experience , a wire positioning accuracy of about @xmath43 m ( rms ) is reachable for such large chambers .
the prototype chamber has been tested in a 300 gev muon beam at cern at perpendicular incidence to the tube layers .
ar : co@xmath67 @xmath68 at 3 bar was used as drift gas . using a silicon strip detector telescope as external reference , the space to drift - time relationship and the position resolution as a function of the drift distance r have been determined .
( see fig .
[ resol ] ) .
the average single - tube resolution at the low interaction rates is @xmath69 m ( rms ) .
the r - t - relationship measured locally was applied to the other drift tubes in the beam allowing only the maximum drift time to vary within @xmath70 ns ( rms ) because of varying operating conditions . requiring the track residual distributions as function of the drift distance to be symmetric left and right of the wires provides information about displacements of the wires from their nominal positions in @xmath12-direction with respect to a reference wire .
comparison with the x - ray measurements of the @xmath12-coordinates of the wires shows a good correlation ( see fig .
[ wirepos ] ) .
both measurements agree within @xmath43 m ( rms ) .
we wish to thank our collegues at the jinr , dubna for the fabrication of such excellent drift tubes for the bos-0 chamber and are indebted to the x - ray tomograph group at cern for the prompt and accurate measurements of our chambers .
a. airapetian , v. grabsky , h. hakopian , a. vartapetian , f. fares , g.f .
moorhead , et al .
, `` atlas muon spectrometer technical design report , '' cern , geneva , switzerland , cern internal report , cern / lhcc/97 - 22 , may 1997 .
w. blum , h. kroha , and p. widmann , `` a novel laser - alignment system for particle tracking detectors using transparent silicon strip sensors , '' _ ieee trans .
_ , vol .
43 , no . 3 , pp . 1194 - 1199 , june 1996 .
j. berbiers , d. drakoulakos , c.w .
fabjan , s. grau , e. gschwendtner , j .-
maugain , et al .
, `` high - precision x - ray tomograph for quality control of the atlas muon monitored drift chambers , '' _ nucl .
_ , vol . 419 , pp .
a342-a350 , 1998 . | the monitored drift tube ( mdt ) chambers for the muon spectrometer of the atlas detector at the large hadron collider ( lhc ) consist of 34 layers of pressurised drift tubes on either side of a space frame carrying an optical deformation monitoring system .
the chambers have to provide a track position resolution of @xmath0 m with a single - tube resolution of at least @xmath1 m and a sense wire positioning accuracy of @xmath2 m ( rms ) .
the feasibility was demonstrated with the full - scale prototype of one of the largest mdt chambers with 432 drift tubes of 3.8 m length .
for the atlas muon spectrometer , 88 chambers of this type have to be built .
the first chamber has been completed with a wire positioning accuracy of @xmath3 m ( rms ) .
drift tubes , drift chambers , muon spectrometer , atlas detector |
effects of simple inorganic salts on the molecular properties of water are at the heart of a vast number of interesting and complex processes such as the stability of proteins and nucleic acids @xcite and environmentally relevant processes @xcite . understanding the effect of dissolved ions on the structural and dynamical properties of water is essential in this regard . the hofmeister effect , which includes highly ion - specific effects on aggregation dynamics of proteins @xcite and other biologically relevant processes @xcite has been of significant interest lately .
spectroscopic techniques have been instrumental in probing the cooperative ion hydration mechanism and consequent long - range structural and dynamical effects of certain salts , or ion combinations , on water @xcite .
earlier experiments suggested that the effect of ions on dynamical properties of water is largely restricted to their first hydration shell @xcite .
however , recent experiments , using a combination of femtosecond time resolved infrared ( fs - ir ) and dielectric relaxation spectroscopy , have shown the existence of a fraction of reorientationally slow water molecules @xcite well beyond the first hydration shells of dissolved mgso@xmath1 ions . in these experiments ,
two sub - populations of water molecules were identified for various salts : one with reorientation timescales comparable to pure water ( @xmath22.6 ps ) and the other showing characteristically slower reorientations ( @xmath210 ps ) .
the fraction of total water molecules contributing to the sub - population of slow - water molecules was seen to increase with increase in salt concentration for all salts and the magnitude of increase was observed to be highly dependent on ion combinations : being largest for combinations of strongly hydrated ion species ( mg@xmath3 , so@xmath4 ) . despite the recent experimental results @xcite , the existence of long - range temporal effects of salts , beyond first hydration shell , is controversial owing to results from other experimental @xcite as well as simulation @xcite studies which suggest the contrary .
while an intense cooperative slowdown of water reorientation has been observed in presence of ions with high charge densities , the range of presence of slow water molecules has been found to be confined to the first hydration shell of the ions @xcite .
the focus of present work is on the extended hydrogen bond network in bulk water and the domain formation and existence of defect water molecules at boundaries of such domains , when salt is added . instead of classifying water molecules based on radially
varying spherical hydration shells around ions , sub - populations depending on whether they are bulk - like or defect water molecules are considered . using molecular dynamics ( md ) simulations ,
we present evidence for reorientational slowdown of water molecules well beyond the first hydration shell for solutions of mgcl@xmath0 and effect of salt concentration on the same .
however for cscl - water solutions , no significant effect on reorientational dynamics of water ( irrespective of cscl concentration ) was found , which is in good agreement with the hofmeister series .
classical atomistic molecular dynamics ( md ) simulations were performed with tip4p - ew @xcite and tip3p @xcite water models and using simulation package namd 2.9 @xcite . for simulations involving tip4p - ew water model , recently developed parameters for divalent ion mg@xmath3 and optimized for correct coordination number @xcite were used .
monovalent ( cl@xmath5 , cs@xmath6 ) ion parameters for the tip4p - ew simulations were taken from extensively used halide and alkali ion parameters @xcite .
standard charmm parameters were used for all ions @xcite while simulating with tip3p water .
all systems were first equilibrated under constant pressure and temperature ( npt ) and further under constant volume and energy ( nve ) conditions .
production runs for all systems were performed under nve conditions .
for the npt simulations , pressure was maintained at 1 atm using langevin piston @xcite and temperature coupling to external heat bath was used to maintain temperature at 298 k for simulations with tip4p - ew model and at 305 k for tip3p .
lennard - jones interactions were smoothly truncated with cutoff @xmath7 using switching function between @xmath8 and @xmath7 .
long - range electrostatic interactions were computed using particle mesh ewald ( pme ) method .
timesteps of 1 and 2 fs were used for simulations involving tip4p - ew and tip3p models respectively . + _
tip4p - ew water systems : _ initially a @xmath9 box of water containing 5251 water molecules was equilibrated for 5 ns under npt conditions and further for 4 ns under nve .
_ solvate _ plugin of vmd @xcite was used to produce mgcl@xmath0 ( 2 m , 3 m , 4 m ) and cscl ( 3 m , 4 m ) solutions from this equilibrated configuration and each of the salt - water systems was equilibrated for 5 ns under npt conditions and further 4 ns under nve resulting in total equilibration time of 9 ns for each of six systems involving tip4p - ew water . + _
tip3p water systems : _ initially a @xmath10 box of water containing 4972 water molecules was equilibrated for 10 ns under npt conditions .
_ solvate _ plugin of vmd was used to produce salt solutions of the same concentrations as for tip4p - ew water molecules from this equilibrated configuration and each of these salt - water systems was equilibrated for 15 ns under npt conditions .
the pure water system was further simulated for 5 ns , resulting in 15 ns npt equilibration simulations for each of the six systems considered .
all above systems were equilibrated for further 9 ns under nve conditions resulting in total equilibration time of 24 ns for systems involving tip3p water .
+ further details of system setup are included in the table si of supplementary information ( _ supp .
_ ) @xcite .
relatively high salt concentrations of salt were chosen in this study based on recently studied experimental concentrations @xcite as well as to account for good statistics .
for computation of reorientational autocorrelation functions and other dynamic observables , production runs of 1 ns each were carried out for all systems under nve conditions , with system configurations saved every 0.1 ps . all the dynamical analyses were performed over the last 0.5ns of the production run data .
$ ] for all systems with tip4p - ew water model considered in the study and for slab thickness of 4@xmath11 .
the data are averaged over 5000 frames and 0.5 ns of simulation time . ] to identify hydrogen bond ( o - o distance @xmath12 3.8@xmath11 and h - o@xmath13h angle @xmath12 60@xmath14 definition of hydrogen bond was used in the present study ) network domains in all the twelve systems considered in this study , each system was divided into multiple overlapping slabs along each of the three orthonormal directions . over each such slab , hydrogen bonded domains of water molecules were identified .
the criterion for inclusion of water molecules to a single domain is the existence of a bidirected path between every pair , via a network of hydrogen bonds formed by water molecules in the same slab . using this definition of water domain
, it was observed that for pure water all values of slab thickness @xmath152.7@xmath11 consistently resulted in a single spanning domain along the slab .
addition of salt to the water systems resulted in formation of multiple water domains , even when higher values of slab thickness ( 4@xmath11 ) were considered .
the size distribution of different domains of tip4p - ew water systems for a slab thickness of ( 4@xmath11 ) and various concentrations of added salts , in comparison with the case of pure water is shown in fig .
1 . from the figure
, it can be seen that for the case of pure water , a single spanning domain exists . with the addition of mgcl@xmath0 salt ,
the size distribution of water domains depends on the concentration . for 2 m salt concentration ,
the domain size distribution is similar to the pure water case , with additional appearance of smaller sized water domains .
when the concentration of mgcl@xmath0 is increased to 3 m and above , the domain size distribution differs significantly from that of pure water .
no single spanning domain remains in the system and the system primarily consists of many domains of water molecules of varying size distribution.similar behavior is aslo observed when tip3p water model is used and the results are given in _
si ) @xcite .
ion specific local structural effects on water has been widely studied in the literature @xcite , generally classifying ions into structure - makers and structure - breakers @xcite based on their effect on the hydrogen bond network of water .
high charge density ions have been known to exert strong patterning effects on first solvation shell water molecules leading to a reduction in water - water hydrogen bonding @xcite .
the domain size distribution in cscl - water system , in contrast to mgcl@xmath0-water systems , is very similar to the pure water system even at 4 m concentration consistent with a recent ab initio md study of other low - charge density salts such as nacl and csi @xcite .
following the identification of domains , water molecules residing at the domain boundaries were identified and recorded for all the three orthonormal directions . from this data , the water molecules which appear in all the three lists , corresponding to three orthonormal directions ,
are identified and labeled defect water molecules .
this set of defect water molecules however may contain water molecules which are within the first hydration shell of any ions . from this super set of defect water molecules , a subset of water molecules which are not within the first hydration shell of any ion
were identified , which are referred to henceforth as _ waterd _ ( pure water is referred to as _ waterp _ ) .
a further sub - population of defect water molecules were also identified : water molecules which are at domain boundaries in any two orthonormal directions ( instead of three ) and not in the first hydration shell of ions .
these will be referred as _
waterd2 _ and constitute a more relaxed definition of defect water molecules and their sub - population size would be larger than that of _ waterd _ at moderate salt concentrations .
the probability distribution of local tetrahedral order parameter ( @xmath16 ) was computed for _ waterp _ and sub - populations _ waterd _ for systems involving tip4p - ew model of water .
@xmath16 for a water molecule is defined as @xcite @xmath17 where @xmath18 is the angle formed by the lines joining the oxygen atom of a given molecule and those of its nearest neighbours j and k @xmath19 .
@xmath16 takes values 1 and 0 for ideal tetrahedral structures and ideal gas respectively .
( t ) for _ waterd _ in the presence of mgcl@xmath0 ( colored curves ) at the concentrations studied and pure water ( black ) using tip4p - ew model .
the error bars in all plots have been magnified 5 times ] the tetrahedrality results , in conjunction with hydrogen bonds per water molecules , suggest that the sub - population of _ waterd _ molecules differ significantly from pure water molecules .
the typical bimodal distributions @xcite of tetrahedrality of water molecules , measured as in eq(1 ) , exhibit deviations in the case of defect water molecules ( results are included in the _ supp .
sv ) @xcite .
the number of hydrogen bonds per defect water molecule ( _ waterd _ ) were found to be less than 2 .
reorientation autocorrelation functions p@xmath20(t ) and p@xmath21(t ) defined as first and second legendre polynomials of water dipole vector ( @xmath22 ) , a unit bisector of the h - o - h angle , were computed for various sub - populations of water molecules over 20 ps : @xmath23 where the angular brackets denote average over number of water molecules in each sub - population and time .
the errors for p@xmath20(t ) and p@xmath21(t ) have been obtained by computing standard deviations using block averages . + the computed correlation functions p@xmath20(t ) for pure water and mgcl@xmath0 solutions using tip4p - ew water model are plotted in fig .
2 ( plots of p@xmath21(t ) for tip4p - ew and both p@xmath20(t ) and p@xmath21(t ) for tip3p are given in fig .
sii and fig .
siii respectively in _ supp . info .
_ ) @xcite . a significant slowing down of reorientational times for sub - population _ waterd _ as a function of salt concentration
can be seen from the figure .
these results suggest that slow water molecules exist beyond the first hydration shell of both cations and anions .
earlier experiments and simulations show that the propensity of formation of ion - water clusters is higher at higher concentration of salts in water @xcite .
water molecules trapped in such clusters can experience very slow reorientational times .
it is to be noted that though the salt concentrations considered in this study are high , the sampling of sub - population _
waterd _ will not include such trapped water molecules , by definition .
the _ waterd _
molecules considered can , at best , be part of three solvent separated ion pairs .
extensive studies of the reorientation of water molecules @xcite have shown that autocorrelation functions of a body - set vector , including dipole vector , of water molecules involve distinct time scales . a fast , sub - picosecond decay due to librational motion followed by a slower component that can be attributed to structural changes such as hydrogen bond exchange and reorientations of hydrogen bonded water molecules .
multiexponential functions thus serve as good fit functions for such correlations and they have been fit to either bi- ( @xmath24 ) or tri - exponential functions of the form p@xmath25 ( t ) for _ waterd _ comparing 3 m and 4 m cscl with 2 m mgcl@xmath0 . the black curve is for pure water using tip4p - ew model .
the error bars in all plots have been magnified 5 times ] depending on the system .
however , the attribution of specific physical processes to these time constants is not attempted in this work .
a bi - exponential function ( @xmath24 ) was the best fit for reorientational time curves for the pure water case . in the presence of mgcl@xmath0
, a tri - exponential function was found to be more appropriate , with dynamics of defect water molecules introducing a new time scale into the problem .
the decay times ( @xmath26 ) are much larger than the slowest component of orientational relaxation for bulk water and show an increase with the concentration of mgcl@xmath0 salt .
these results are shown in table 1 .
simulation results show that for the largest concentration of mgcl@xmath0 studied ( 4 m ) , the longest mode in decay time for defect water molecules , which are beyond first hydration shell of any ions , is as high as 21 ps .
the p@xmath20(t ) curves for cscl salt solutions are plotted in fig . 3 and deviation from the case of pure water is much less significant compared to effects of mgcl@xmath0 and independent of salt concentration .
the p@xmath20(t ) curves for cscl are also fit with a tri - exponential function and the values are given in table 1 .
the corresponding reorientational time constants obtained for p@xmath21(t ) for sub - population _ waterd _ for tip4p - ew and p@xmath27(t ) for tip3p water models are given in _
sii ) @xcite .
the anion used in both the salts , cl@xmath5 , is known to be weakly hydrated ion and such anions are expected to affect the dynamics of oh vector preferentially over dipole vectors of water molecule @xcite .
thus the observed difference in reorientational dynamics between the two cations studied in this work , mg@xmath3 and cs@xmath6 reflect the difference between a strongly hydrated vs a weakly hydrated cation , while both have the same counterion .
similar differences between mg@xmath3 and cs@xmath6 are observed for the sub - population of water molecules _
waterd2 _ as well and the results are included in _ supp .
siv ) @xcite for both water models used .
these results seem to be consistent with the ordering of cations in well - known hofmeister series @xcite . [ cols="^,^,^,^,^,^,^,^",options="header " , ]
to summarize , md simulation results in this study support the concentration dependent effects of strongly hydrated ion species ( mg@xmath3 ) on the reorientational dynamics of water molecules beyond the first hydration shell . a likely mechanism for
the same has been suggested in terms of salt induced defects in the underlying hydrogen bond network of water , which is in agreement with concepts of ion induced patterning of water at long distances . while the actual number of water molecules beyond first hydration shell of ions that exhibit slow reorientational times is a small fraction of the total water molecules , especially at low concentrations ,
they have been found at large spatial separations from the ions .
the fractional number has been observed to increase monotonically with increase in the concentration of mgcl@xmath0 .
it is to be noted that the long - range effect of mg@xmath3 ions in the presence of a weakly hydrated counterion cl@xmath5 is smaller than experimentally observed effects in the presence of so@xmath4 @xcite , but not insignificant .
comparison with weakly hydrated cationic species ( cs@xmath6 ) has been seen to be in agreement with the hofmeister series .
the results in the present study are obtained with two non - polarizable models of water and these results are expected to be enhanced when polarizable models of water are used and are a part of future studies .
all the simulations in this work have been carried out on 1024-cpu annapurna cluster at the institute of mathematical sciences , chennai , india . | effects of presence of ions , at moderate to high concentrations , on dynamical properties of water molecules are investigated through classical molecular dynamics simulations using two well known non - polarizable water models .
simulations reveal that the presence of magnesium chloride ( mgcl@xmath0 ) induces perturbations in the hydrogen bond network of water leading to the formation of bulk - like domains with defect sites on boundaries of such domains : water molecules at such defect sites have less number of hydrogen bonds than those in bulk water .
reorientational autocorrelation functions for dipole vectors of such defect water molecules are computed at different concentrations of ions and compared with system of pure water .
earlier experimental and simulation studies indicate significant differences in reorientational dynamics for water molecules in the first hydration shell of many dissolved ions .
results of this study suggest that defect water molecules , which are beyond the first hydration shells of ions , also experience significant slowing down of reorientation times as a function of concentration in the case of mgcl@xmath0 . however , addition of cesium chloride(cscl ) to water does not perturb the hydrogen bond network of water significantly even at higher concentrations .
this difference in behavior between mgcl@xmath0 and cscl is consistent with the well - known hofmeister series . |
the real- or imaginary - time dynamics of systems described by a finite hamiltonian matrix , representing bosonic or fermionic degrees of freedom , admits an exact probabilistic representation in terms of a proper collection of independent poisson processes @xcite . for a lattice system
, the poisson processes are associated to the links of the lattice and the probabilistic representation leads to an optimal algorithm @xcite which coincides with the green function quantum monte carlo method in the limit when the latter becomes exact @xcite . in the recent ref .
we have exploited the above probabilistic representation to derive analytical expressions for the matrix elements of the evolution operator in the long time limit . in this way ,
the ground - state energy as well as the expectation of a generic operator in the ground state of a lattice system without sign problem are obtained as the solution of a simple scalar equation .
the result is based on the application of a central limit theorem to the rescaled multiplicities of the values assumed by the potential and hopping energies in the configurations dynamically visited by the system . as a consequence
, the probabilistic expectations can be calculated by using a gaussian - like probability density . in this paper
, we briefly review the approach developed in ref . and discuss in detail the choice of the asymptotic probability density used in the calculation .
we illustrate our approach in the case of imaginary - time dynamics for a system of hard - core bosons described by the hamiltonian @xmath0 where @xmath1 is a finite @xmath2-dimensional lattice with @xmath3 sites and @xmath4 the commuting destruction operators at site @xmath5 and spin index @xmath6 with the property @xmath7 .
the potential operator @xmath8 is arbitrary , _
e.g. _ for the hubbard model @xmath9 . for simplicity , we assume @xmath10 if @xmath5 and @xmath11 are first neighbors and @xmath12 otherwise . in order to study the ground - state properties of the hamiltonian @xmath13 it is sufficient to evaluate the long time behavior of @xmath14 , where @xmath15 are the lattice occupation numbers taking the values 0 or 1 . in fact , the ground - state energy is given by @xmath16 while the quantum expectation of a generic operator @xmath17 in the ground state of @xmath13 can be obtained via the hellman - feynman theorem @xcite by evaluating the ground - state energy @xmath18 of the modified hamiltonian @xmath19 . at any finite time @xmath20 ,
the matrix elements of the evolution operator considered above admit the exact probabilistic representation @xmath21 where @xmath22 is a stochastic functional defined in terms of independent poisson processes associated to the links of the lattice , see ref . for a detailed description .
at each jump of a poisson process relating sites @xmath5 and @xmath11 with spin @xmath6 and taking place at a given configuration @xmath23 , a particle of spin @xmath6 moves from site @xmath5 to site @xmath11 or vice versa if the mod 2 sum of the occupations of these two sites is @xmath24 , while the lattice configuration @xmath23 remains unchanged if @xmath25 .
hereafter , links with @xmath26 will be called active . by ordering the jumps according to the times @xmath27 , @xmath28 , at which they take place in the interval @xmath29
, we define a trajectory as the markov chain @xmath30 generated from the initial configuration @xmath31 .
the number of jumps @xmath32 is , of course , a random integer associated to each trajectory .
we associate to each trajectory also two sequences , @xmath33 and @xmath34 , representing the number of active links and the potential energy of the visited configurations @xmath35 here , @xmath36 is the set of system links , _
i.e. _ the pairs @xmath37 with @xmath38 and @xmath39 such that @xmath40 . the stochastic functional @xmath22 which appears in eq .
( [ eq1 ] ) actually depends on the jump times @xmath41 and on the corresponding sequences @xmath33 and @xmath34 .
evaluating the expectation @xmath42 over the detailed sequences above specified can be done numerically by a monte carlo method @xcite . in ref . we have demonstrated that an analytical expression of @xmath42 can be obtained in the limit of long times . this result
is reached in four steps described in the next subsections .
the crucial point is that , if one integrates over all the possible jumps times , what matter are not the detailed sequences @xmath33 and @xmath43 but the multiplicities @xmath44 and @xmath45 of the possibles values which the variables @xmath46 and @xmath47 may assume .
we call @xmath48 and @xmath49 the sets of these values and @xmath50 and @xmath51 their cardinalities .
it is clear that the nature of these sets depends only on the structure of the system hamiltonian , not on the values of the hamiltonian parameters .
the expectation @xmath42 is reduced to an average over @xmath44 and @xmath45 . for @xmath52
, this average can be evaluated analytically by using saddle - point techniques and a central limit theorem . referring to ref . for the details ,
we decompose the expectation as a series of conditional expectations with a fixed number of jumps ( canonical averages ) @xmath53 integrating over the @xmath54 jumps times , each term of the series ( [ expansion ] ) can be written as @xmath55 where @xmath56 means average over the trajectories with @xmath54 jumps generated by extracting with uniform probability one of the active links available at the configurations @xmath57 , and @xmath58 , named weight , is defined as @xmath59 according to their definition , the weights satisfy a recursive differential equation which is easily solved in terms of the laplace transform @xmath60 @xcite , _ i.e. _ @xmath61 while this expression shows that @xmath58 depends on the multiplicities @xmath45 for any value of @xmath54 , the explicit inversion of the laplace transform can be done analytically only for @xmath54 large .
however , this is the limit we are interested in since the weights @xmath62 have a maximum at some @xmath54 which increases by increasing @xmath20 . by using a complex saddle - point method which is asymptotically exact for @xmath63
, we get @xcite @xmath64 } } { \sqrt { 2\pi \sum_{v\in\mathscr{v}}\frac{\epsilon^2 n_{v}}{(x_{0}+v)^{2 } } } } , \end{aligned}\ ] ] where @xmath65 is the solution of the equation @xmath66 to evaluate the canonical averages it is useful to introduce the frequencies , @xmath67 and @xmath68 , which for @xmath54 large become continuously distributed in the range @xmath69 $ ] with the constraints @xmath70 note that for @xmath54 large we will not distinguish the different normalizations , @xmath71 and @xmath54 , of @xmath45 and @xmath44 , respectively . equation ( [ average ] ) can be then rewritten as @xmath72 where @xmath73 and @xmath74 are vectors with @xmath75 components defined as @xmath76 and @xmath77 \ldots;\ldots \log a \ldots)$ ] , respectively . for later use , we also define @xmath78 and @xmath79 .
note that @xmath74 , @xmath80 and @xmath81 depend on @xmath73 through @xmath82 and @xmath83 , @xmath84 .
the probability density @xmath85 is given by the fraction of trajectories branching from the initial configuration @xmath31 and having after @xmath54 jumps multiplicities @xmath86 and @xmath87 .
for @xmath54 large , it can be approximated in the following way .
we rewrite the multiplicities as @xmath88 and @xmath89 , where @xmath90 if @xmath91 and @xmath92 otherwise , and similarly for @xmath93 .
since the configurations @xmath94 form a markov chain with finite state space , a central limit theorem applies to each rescaled sum @xmath95 or @xmath96 @xcite .
therefore , these rescaled variables are completely described in terms of the mean values @xmath97 and of the covariance matrix @xmath98 , which are easily measured by sampling over trajectories with a large number of jumps .
due to the constraints ( [ constraints ] ) , it is easy to see that @xmath99 @xmath100 in the next two subsections we will describe two different choices for the density @xmath85 .
the two densities differ only for the fact that the first one satisfies the constraints ( [ constraints ] ) in mean , while the other one identically .
we will show that , at least in the limit of large @xmath54 , the results obtained by assuming the two densities coincide . in this case , we assume a purely gaussian probability density @xmath101 due to eq .
( [ blocks ] ) , we have that @xmath102 , _ i.e. _ @xmath98 is singular and the density ( [ gaussian ] ) ill defined . nevertheless , this singularity is eliminable and the calculation with this density has the advantage to be rather simple .
we rewrite eq .
( [ average1 ] ) as @xmath103 where @xmath104 and @xmath105 is a smooth function .
the parametric dependence of @xmath106 and @xmath107 on @xmath20 and @xmath54 is omitted for brevity .
we will perform the integral ( [ average2 ] ) with a saddle - point method by solving the @xmath108-dimensional stationary problem for @xmath109 .
the derivatives of @xmath106 with respect to the @xmath47 and @xmath46 components of @xmath73 read @xmath110 on the other hand , by using the definition of @xmath111 , eq .
( [ x ] ) , it is easy to see that @xmath112 , so that @xmath109 is stationary for @xmath113 solution of the equation @xmath114 note that , due to eqs .
( [ means ] ) and ( [ blocks ] ) , we have @xmath115 and only @xmath116 equations in ( [ saddle ] ) are independent . the result of the saddle - point integration then is @xmath117 where @xmath118_{\bm{{\nu}}=\bm{{\nu}}^\mathrm{sp}}\end{aligned}\ ] ] and @xmath119 , @xmath120 being the @xmath121 matrix with elements @xmath122 by writing @xmath123 and observing that @xmath124 cancels out , eq .
( [ average3 ] ) becomes @xmath125 } } { \sqrt{2 \pi n \epsilon^2 ( \bm{\nu},\bm{w } ) } } \right|_{\bm{\nu}=\bm{\nu}^\mathrm{sp}}.\end{aligned}\ ] ] note that for @xmath126 the matrix @xmath120 is uniform and @xmath127 .
in general , @xmath128 so that @xmath129 up to terms of second order in @xmath120 . in the previous subsection
we have evaluated the canonical averages by exploiting a central limit theorem for the rescaled variables @xmath95 and @xmath96 . however , due to the constraints ( [ constraints ] ) , the joint probability for these @xmath108 rescaled sums is not purely gaussian .
given an arbitrary set of @xmath130 @xmath47-like components and @xmath131 @xmath46-like components , we can assume a joint probability density as the product of a gaussian density for this set of @xmath116 variables and two delta functions which take into account the constraints ( [ constraints ] ) , _ i.e. _ @xmath132 where @xmath133 is the normal density defined in terms of the vector @xmath134 having the @xmath116 chosen components of @xmath73 @xmath135 in the following , with the symbol @xmath136 over a vector or a matrix we will indicate the projection onto the chosen @xmath116 dimensional index space . by using the density ( [ density ] ) , we rewrite eq .
( [ average1 ] ) as @xmath137 where @xmath138 is a smooth function of @xmath139 and , if @xmath140 and @xmath141 are the two components chosen to reduce to @xmath116 the dimension of the index space , @xmath142 note that @xmath143 is the solution of eq .
( [ x ] ) where the constraints ( [ constraints ] ) have been explicited .
the function @xmath144 can be rewritten as @xmath145 where @xmath146 and @xmath147 . by evaluating the derivatives of @xmath144 with respect to the components of @xmath139
, we get @xmath148 where @xmath149 . observing that @xmath150
, we can rewrite these derivatives as @xmath151 therefore , the saddle - point equation is @xmath152 finally , we calculate @xmath153 . by using eqs .
( [ phi1 ] ) and ( [ speq ] ) and the identities @xmath154 and @xmath155 , we get @xmath156_{\hat{\bm{\nu}}=\hat{\bm{\nu}}^\mathrm{sp } } \nonumber \\ & = & \left [ \hat{x}_0\frac{t}{n}+ \left(\overline{\bm{\nu}},\bm{u}(\hat{x}_0 ) \right ) + \frac12 \left ( \bm{\sigma } \bm{u}(\hat{x}_0),\bm{u}(\hat{x}_0)\right ) \right]_{\hat{\bm{\nu}}=\hat{\bm{\nu}}^\mathrm{sp}}.\end{aligned}\ ] ] due to the identity @xmath157 , we have @xmath158 , where @xmath159 is the value obtained with the purely gaussian density . therefore , eq . ( [ phisp ] ) shows that , at least at the saddle - point level , the chosen probability density provides the same results obtained in the previous subsection . in order to evaluate the expectation
@xmath160 we need to resum the series ( [ expansion ] ) . for @xmath161
, the sum can be substituted with the integral @xmath162 where @xmath163.\end{aligned}\ ] ] the integrand in ( [ psiint ] ) is exponentially peaked at @xmath164 , where @xmath165 satisfies @xmath166 . for a generic @xmath54
we have @xmath167 \right\ } \partial_n x_{0 } + ( \overline{\bm{\nu}}{},{\bm{u } } ) + \frac{1}{2 } \left ( \bm{\sigma } { \bm{u } } , { \bm{u } } \right ) \right]_{\bm{\nu}=\bm{\nu}^\mathrm{sp } } \nonumber \\ & = & \left [ ( \overline{\bm{\nu}}{},{\bm{u } } ) + \frac{1}{2 } \left ( \bm{\sigma } { \bm{u } } , { \bm{u } } \right ) \right]_{\bm{\nu}=\bm{\nu}^\mathrm{sp}},\end{aligned}\ ] ] where , due to eqs .
( [ x ] ) and ( [ saddle ] ) , we noticed that the term @xmath168 \right\}$ ] vanishes for @xmath113 .
the stationarity condition for @xmath169 is , therefore , @xmath170_{\bm{\nu}=\bm{\nu}^\mathrm{sp},{n = n^\mathrm{sp } } } = 0.\end{aligned}\ ] ] equation ( [ e1 ] ) is a time independent equation which determines @xmath171 as a function of @xmath172 and @xmath98 .
according to eq .
( [ x ] ) , this means that the quantity @xmath165 increases linearly with @xmath20 so that @xmath171 becomes independent of time in the limit @xmath173 . by evaluating the second derivative of @xmath169 @xmath174 where @xmath175 and approximating @xmath176 in the exponent and @xmath177 elsewhere , the integral ( [ psiint ] ) reduces to a gaussian one which gives @xmath178 note that , since the peak of the gaussian at @xmath164 moves to infinity linearly with @xmath20 while its width increases only as @xmath179 , the result ( [ emt ] ) is asymptotically exact for @xmath161 . according to eq .
( [ e0 ] ) , the result ( [ emt ] ) shows that the ground - state energy of the hard - core boson system is @xmath180 equation ( [ e1 ] ) is , therefore , the equation for the ground - state energy .
it defines @xmath181 in terms of @xmath172 and @xmath98 and explicitly reads @xmath182 in the case @xmath183 , the ground - state energy @xmath184 can be solved analytically @xmath185 .
\end{aligned}\ ] ]
by using saddle - point techniques and a central limit theorem , we have exploited an exact probabilistic representation of the quantum dynamics in a lattice to derive analytical approximations for the matrix elements of the evolution operator of a system of hard - core bosons in the limit of long times .
the approach yields a simple scalar equation for the ground - state energy .
this equation depends on the values of the generalized potentials @xmath47 and of the kinetic quantities @xmath46 , and on the statistical moments @xmath172 and @xmath98 of their asymptotic multiplicities @xmath45 and @xmath44 . in turn , these moments depend only on the structure of the system hamiltonian , not on the values of the hamiltonian parameters .
this implies that the statistical moments must be measured _
una tantum _ for a given hamiltonian structure and , once @xmath172 and @xmath98 are known , our approach provides the ground - state energy analytically as a function of the hamiltonian parameters .
the ground state energies obtained with the present formula reveal a small systematic error when compared with exact results @xcite .
in fact , a more accurate analysis which takes into account the large deviations neglected by the central limit theorem shows that the present approach corresponds to a second order truncation of an exact cumulant expansion for the ground - state energy @xcite . | we review a novel approach to evaluate the ground - state properties of many - body lattice systems based on an exact probabilistic representation of the dynamics and its long time approximation via a central limit theorem .
the choice of the asymptotic density probability used in the calculation is discussed in detail . |
the pion nucleon form factor @xmath2 plays a very important , but not less controversial role in the framework of @xmath5 and @xmath6 dynamics . in general ,
meson baryon form factors are used in one boson exchange potentials ( obep ) for the @xmath6 force in order to account for the microscopic , i.e. quark and gluon structure of the mesons and baryons at a given vertex , and therefore provide a natural cutoff description for the interaction potential at short distances .
most of the realistic obep fit the data with a monopole form for the @xmath1 form factor @xmath7 with a large monopole mass @xmath8 @xcite .
on the other hand almost all other hadronic theories advocate a much softer form factor with @xmath9 : chiral soliton and quark models for the nucleon @xcite , the goldberger - treiman discrepancy between @xmath10 and @xmath11 @xcite , general considerations on the structure of the @xmath1 vertex @xcite , charge exchange reactions @xcite , threshold @xmath12 production @xcite , deep inelastic lepton nucleon scattering ( sullivan process ) @xcite .
more recent studies within the obep , which either include the @xmath13 resonance @xcite , use a different form of the scalar exchange potential @xcite or consider a correlated @xmath14exchange @xcite also indicate a softer cutoff mass ( @xmath15 ) . for these reasons it seems highly desirable to perform a calculation of @xmath2 , which does not suffer from the ambiguities of the model and parameterization schemes mentioned above and
is connected to qcd as closely as possible . in this context
a quenched lattice qcd calculation has recently been carried out rendering a monopole cutoff mass of @xmath16 @xcite .
it is the aim of our work to study this phenomenon within the framework of qcd sum rules @xcite , which have turned out to be a very successful method for calculating hadronic properties at intermediate energies ( for reviews c.f.@xcite ) without employing the computer time consuming lattice gauge calculations . in order to calculate the pion nucleon coupling constant @xmath17 itself ,
i.e. for on shell @xmath12 , one can consider either : ( i ) the vacuum three point correlator of two nucleon and one pseudoscalar meson interpolating fields , which are saturated with resonances in the nucleon @xmath18 and pion @xmath12 channels on the phenomenological side @xcite or ( ii ) the 2 point function of 2 nucleon interpolating fields sandwiched between the vacuum and one @xmath12 state and saturating only with @xmath18 resonances @xcite . with both methods it was possible to obtain rather reasonable results for @xmath19 ,
although it should be stated that the uncertainties in the borel analysis are relatively high , especially in method ( ii ) , even if higher order power and @xmath20 corrections are taken into account @xcite .
if one is interested in the momentum dependence of @xmath2 at intermediate @xmath0 and therefore with off shell @xmath12 , only method ( i ) , which has been successfully used for the calculation of meson formfactors @xcite , can be applied .
our paper is organized as follows : in section [ sec2 ] we introduce the 3 point function for the @xmath21 vertex and saturate with nucleon intermediate states . in section [ sec3 ] the operator product expansion ( ope )
is carried out . in section [ sec4 ]
we perform the borel analysis .
the results are discussed and summarized in section [ sec5 ] .
our starting point is the 3 point function ( fig.[fig1 ] ) @xmath22 of a pseudoscalar , charge neutral current @xmath23 and two ioffe nucleon interpolating fields @xcite @xmath24 \label{eq4}\ ] ] hereby @xmath25 denote the spinors for the quarks with mass @xmath26 , @xmath27 the color indices and @xmath28 the charge conjugation matrix .
the three momenta at the vertex are related by @xmath29 .
due to restrictions from lorentz , parity and charge conjugation invariance the 3 point function @xmath30 has the general form @xmath31 where @xmath32 , @xmath33 .
the functions @xmath34,@xmath35,@xmath36 are symmetric and @xmath37 is antisymmetric under the interchange @xmath38 .
the matrix element of the pseudoscalar current @xmath39 between on shell nucleon states defines the pseudoscalar nucleon formfactor @xmath40 where @xmath41 denotes a free nucleon spinor in momentum space .
the pion nucleon coupling constant @xmath42 is defined by the @xmath5 interaction : @xmath43 with the form factor @xmath2 for on shell nucleons it is equivalent to use the coupling of eq.([eq5a ] ) or the chirally invariant vector coupling @xmath44 by means of the chiral ward identity @xmath45 for the axial current @xmath46 as well as the pcac relation @xmath47 one can relate the pseudoscalar nucleon form factor @xmath48 ( eq.[eq5 ] ) to the pion nucleon form factor @xmath2 : @xmath49 therefore saturating the eq.([eq2 ] ) with nucleon states @xmath50 renders , after continuing to euclidean momenta ( @xmath51 , @xmath52 , @xmath53 ) @xmath54_n = i \lambda_n ^2 \left ( { { m_\pi ^2 f_\pi}\over { m_0 } } \right ) \ : { { g_{\pi nn } ( q^2 ) } \over { q^2 + m_\pi ^2 } } \ : { { \left [ ( \not{\ ! p_1 } + m_n ) \gamma_5 ( \not{\
} + m_n ) \right ] } \over { ( p_1 ^2 + m_n ^2 ) ( p_2 ^2 + m_n ^2 ) } } \label{eq9}\ ] ] where @xmath55 is the overlap between the ioffe current @xmath56 and a nucleon state @xmath57 using @xmath58 the @xmath59 $ ] term in eq.([eq9 ] ) can be cast into the form @xmath60 the contribution of the first higher resonance @xmath61 with mass @xmath62 to eq.([eq2 ] ) has the form @xmath54_{n^ * } = i \lambda_n \lambda_{n^ * } \left ( { { m_\pi ^2 f_\pi}\over { m_0 } } \right ) \ : { { g_{\pi n { n^ * } } ( q^2 ) } \over { q^2 + m_\pi ^2 } } \ : { { \left [ ( \not{\ ! p_1 } + m_n ) \gamma_5 ( \not{\ ! p_2
} + m_n ^ * ) \right ] } \over { ( p_1 ^2 + m_n ^2 ) ( p_2 ^2 + { m_n ^*}^2 ) } } \quad + \quad
n \leftrightarrow n^ * \label{eq12}\ ] ]
as already stated in eq.([eqf ] ) there are essentially 4 lorentz structures contained in @xmath30 . in the following we work with massless quarks , i.e @xmath63 . from a simple dimensional analysis of both sides of the sum rule it is easy to see @xcite that in the sum rule for the functions @xmath34 and @xmath36 , which both contain an even number of external momenta , only operators of even dimension ( @xmath64 , @xmath65 , @xmath66 , etc . ) contribute , whereas the odd dimension operators enter with a mass factor @xmath67 and therefore vanish for @xmath68 . on the other hand for the functions @xmath35 and @xmath37 containing an odd number of external momenta , only operators of odd dimension ( @xmath69 , @xmath70 )
contribute . by multiplying both sides of the sum rule with @xmath71 and taking the trace over the dirac matrices
, we can single out the odd dimensional structure , i.e. the sum rule for @xmath35 and @xmath37 . up to order 4 we only have to account for the diagrams in fig.[fig2 ] containing the quark condensate @xmath72 .
the contribution from the diagrams in fig.[fig2]a renders : @xmath73 _ { 2a } = 16 \ , \gamma_5 \
, \langle \bar{q } q \rangle \ , { { \not{\ ! q } } \over { q^2 } } \ , { 1\over 2 } \ , \left [ p_1 ^2 i ( p_1 ^2 ) + p_2 ^2 i(p_2 ^2 ) \right ] \label{eq13}\ ] ] whereas for the diagrams of fig.[fig2]b one obtains after a rather lengthy calculation : @xmath74 _ { 2b } \ , = \ , ( & - & 16 ) \ , \gamma_5 \ , \langle \bar{q } q \rangle \ , { { \not{\ ! q } } \over 4 } \ , \left [ ( p_1 ^2 + p_2 ^2 ) j(p_1 ^2 , p_2 ^2 , q^2 ) + ( p_1 ^2 - p_2 ^2 ) k_+ ( p_1 ^2 , p_2 ^2 , q^2 ) \right ] \ , + \nonumber\\ ( & - & 16 ) \ , \gamma_5 \
, \langle \bar{q } q \rangle \ , { { \not{\ ! p } } \over 2 } \ , \left [ ( p_2 ^2 - p_1 ^2 ) j(p_1 ^2 , p_2 ^2 , q^2 ) + ( p_1 ^2 - p_2 ^2 ) k_- ( p_1 ^2 , p_2 ^2 , q^2 ) \right ] \ , + \nonumber\\ ( & - & 16 ) \ , \gamma_5 \ , \langle \bar{q } q \rangle \ , { { \not{\ ! q } } \over 4 } \ , \left [ { 1\over 2 } i(p_1 ^2 ) + { 1\over 2 } i(p_2 ^2 ) + 2 i ( q^2 ) \right ] \ ,
+ \nonumber\\ ( & - & 16 ) \ , \gamma_5 \
, \langle \bar{q } q \rangle \ , { { \not{\ ! p } } \over 2 } \ , \left [ { 1\over 2 } i(p_1 ^2 ) - { 1\over 2 } i(p_2 ^2 ) \right ] \label{eq14}\end{aligned}\ ] ] the expressions @xmath75,@xmath76,@xmath77 and @xmath78 arise from loop integrations and can be most conveniently represented as double parameter integrals @xmath79 @xmath80 and @xmath81 \label{eq17}\ ] ] with @xmath82 in the uv divergent integral of eq.([eq15 ] ) we have applied the standard dimensional renormalization at the renormalization point @xmath83 .
for the ope to be valid @xmath0 has to large , i.e. @xmath84 . in this case
we can neglect the pion mass in the pole term in eq.([eq13 ] ) , which is consistent with putting @xmath63 . in refs.@xcite
a sum rule for the pion nucleon coupling @xmath85 has been obtained and analyzed by identifying the residua of the @xmath86 pole in eq.([eq9 ] ) and in the ope contribution from fig.[fig2]a ( [ eq13 ] ) . by introducing @xmath87 and analogous expressions for the higher resonance contributions , we therefore can write down the sum rule : @xmath88_{2b } \label{eq20}\ ] ] following
refs.@xcite we apply the double borel transform with respect to the nucleon momenta @xmath89 and @xmath90 @xmath91 @xmath92 and put @xmath93 . the momentum @xmath0 of the pion is kept fixed at a spacelike value @xmath84 , so that both sides of the sum rule depend on @xmath94 and @xmath0 now . after applying the operator @xmath95 on the ope side the terms which depend only on either @xmath89 or @xmath90 vanish .
the borel transform of the other expressions can be performed by using the relations @xmath96 taking the first resonance with the quantum numbers of the nucleon @xmath61 explicitly into account in addition to the ground state @xmath18 we obtain the borel sum rule @xmath97 where the function @xmath98 is defined as : @xmath99 e^ { -x \frac{\rho}{1-\rho } } \label{eq24}\ ] ] in order to eliminate the parameters of the higher resonance contribution as far as possible we can take the derivative @xmath100 on both sides of eq.([eq24 ] ) and substitute the so obtained sum rule back into eq.([eq24 ] ) .
after employing the gell - mann oakes renner relation @xmath101 we obtain finally @xmath102 with @xmath103 \ ,
\left ( \frac{m}{m_n } \right ) ^6 \ , e^ { 2 \frac{m_n ^2}{m^2 } } \label{eq27}\ ] ] before performing the borel analysis , let us comment about the region of @xmath0 , where our method can be assumed to work .
if @xmath0 is too low ( @xmath104 ) the ope breaks down due to higher order power corrections in @xmath105 . on the other hand
we have used the pcac relation ( [ eq7 ] ) in order to relate @xmath2 to the pseudoscalar nucleon form factor @xmath48 ( eq.([eq8 ] ) ) .
this is equivalent to assume pion pole dominance for @xmath48 , i.e. saturating @xmath106 with a pion state and using the coupling @xmath2 .
this assumption likely breaks down if @xmath0 is of the same magnitude as the first pionic resonance @xmath13 , i.e. @xmath107 , because then @xmath13 contributes in the dispersion relation to a similar extent than @xmath12 does . keeping this in mind we use for the the borel analysis @xmath0 values in the interval @xmath108 , where our approach can be assumed to be reliable .
hereby we follow the spirit of ref.@xcite and treat the resonance mass @xmath62 as an effective mass , which value is adjusted in order to obtain maximal borel stability . from figs.([fig3],[fig4],[fig5 ] ) we can deduce that for the whole @xmath0 interval mentioned above an effective resonance mass of @xmath109 ( @xmath110 ) gives the largest borel plateau of @xmath111 .
from eqs.([eq26],[eq27 ] ) we obtain the final expression for the form factor @xmath112 defined in eq.([eq1 ] ) : @xmath113 due to the discussion in the last section , we use @xmath114 and @xmath115 .
for @xmath116 we take the experimental value @xmath117 . in order to avoid uncertainties from the sum rule for @xmath42
we will also use the experimental value @xmath118 instead of the value from the sum rule of refs .
the parameter @xmath55 is not experimentally known but has been determined from various analyses using the nucleon sum rule , the most recent and concerning uncertainties obviously most reliable one coming from ref .
@xcite : @xmath119 as we have discussed in the last section formula ( [ eq29 ] ) can not be applied for low @xmath0 .
this means that we have to interpolate from @xmath120 to @xmath10 , where @xmath121 .
it is interesting to not that the values for @xmath112 which one obtains by this procedure at low @xmath0 are practically identical to the ones one would get if one applied eq.([eq29 ] ) literally to the low @xmath0 region
. this might be purely accidental , but a possible reason could be that higher power corrections in@xmath105 blowing up the ope at @xmath122 are canceled or suppressed after applying the double borel transform @xmath95 ( [ eq21 ] ) . in order to answer this question
one has to examine higher order power corrections to the ope diagrams in fig.[fig2 ] , which is rather cumbersome for a 3 point function consisting of four quark lines and will be postponed to a separate analysis .
anyhow , from fig.[fig6 ] we can see that the interpolation of @xmath112 from intermediate @xmath0 , where we can apply eq.([eq29 ] ) , to small @xmath0 is very smooth .
furthermore in the whole interval @xmath123 the @xmath112 of our calculation can be fitted very accurately by a monopole form ( [ eq1 ] ) with a cutoff mass of @xmath124 .
this result is very close to the one recently obtained from quenched lattice qcd @xcite , which indicates that both methods give a similar description of the @xmath1 vertex at intermediate @xmath0 .
this work has been supported by the nsf grant # phys-9310124 .
i would like to thank e.henley ( university of washington ) for various valuable discussions and comments and the institute for nuclear theory at the university of washington for its hospitality during the programs int-95 - 1 and int-95 - 2 .
r.machleidt , k.holinde and c.elster , phys.rep . * 149 * , 1 ( 1987 ) .
t.d.cohen , phys.rev .
d * 34 * , 2187 ( 1986 ) .
n.kaiser , u.g.meissner and w.weise , phys.lett .
b * 198 * , 319 ( 1987 ) . s.theberge et.al .
, phys.rev .
d * 22 * , 2838 ( 1980 ) .
k.bruer et.al . , z.phys .
a * 320 * , 609 ( 1985 ) .
p.alberto et.al .
, phys.lett .
b * 208 * , 75 ( 1988 ) ; z.phys . a * 336 * , 449 ( 1990 ) .
t.meissner and k.goeke , nucl.phys .
a * 524 * , 719 ( 1991 ) .
j.gasser , m.sainio and a.svarc , nucl.phys .
b * 307 * , 779 ( 1988 ) . d.schtte and a.tillemans , phys.lett .
b * 206 * , 1 ( 1988 ) . h.esbensen and
t.s.h.lee , phys.rev .
c * 32 * 1966 ( 1985 ) .
t.s.h.lee , nucl - th-9502005 , ( 1995 ) .
a.w.thomas , phys.lett .
b * 126 * , 97 ( 1983 ) . l.frankfurt , l.mankiewicz and m.i.strikman , z.phys .
a * 334 * , 343 ( 1989 ) . w-y.p.hwang ,
j.speth and g.e.brown , z.phys .
a * 339 * , 383 ( 1991 ) . k.holinde and a.w.thomas , phys.rev .
c * 42 * , ( 1990 ) , r1195 .
s.deister et.al .
, few body syst . * 10 * , 36 ( 1991 ) .
g.janssen et.al .
, phys.rev.lett . * 71 * , 1978 ( 1993 ) .
k.f.liu et.al .
, phys.rev.lett . * 74 * , 2121 ( 1995 ) .
m.a.shifman , a.i.vainshtein and v.i.zakharov , nucl.phys .
b * 147 * , 385 ; * 147 * , 448 ( 1979 ) .
l.j.reinders , h.rubinstein and s.yazaki , phys.rep .
* 127 * , 1 ( 1985 ) .
s.narison , _ qcd spectral sum rules _ ( world scientific , singapore , 1989 ) .
l.j.reinders , h.rubinstein and s.yazaki , nucl.phys .
b * 213 * , 109 ( 1983 ) .
l.j.reinders , act.phys.pol . *
15 * , 329 ( 1984 ) .
s.narison and n.paver , phys.lett .
b * 135 * , 159 ( 1984 ) . h.shiomi and t.hatsuda , hep - ph-9504353 , ( 1995 ) .
b.l.ioffe and a.v.smilga , phys.lett .
b * 114 * , 353 ( 1982 ) .
v.a.nesterenko and a.v.radyushkin , phys.lett .
b * 115 * 410 ( 1982 ) .
b.l.ioffe and a.v.smilga , nuc.phys .
b * 216 * , 373 ( 1983 ) .
b.l.ioffe , nucl.phys .
b * 188 * , 317 ( 1981 ) .
t.hatsuda et.al .
, phys.rev .
c * 49 * , 452 ( 1994 ) .
e.m.henley and j.pasupathy , nucl.phys .
a * 556 * , 467 ( 1993 ) . | qcd sum rules are used to calculate the @xmath0 dependence of the @xmath1 coupling @xmath2 in the spacelike region @xmath3 .
we study the borel sum rule for the three point function of one pseudoscalar and two nucleon currents up to order four in the operator product expansion .
the borel transform is performed with respect to the nucleon momenta , whereas the momentum @xmath0 of the pseudoscalar vertex is kept fixed at spacelike values .
the results can be well fitted using a monopole form with a cutoff mass of about @xmath4 . |
the work of c.s.k . was supported by grant no . r02 - 2003 - 000 - 10050 - 0 from brp of the kosef .
y.d . is supported in part by nfsc under contract no.10305003 , henan provincial foundation for prominent young scientists under contract no.0312001700 and in part by the project sponsored by srf for rocs , sem .
this work is also supported by grant no .
f01 - 2004 - 000 - 10292 - 0 of kosef - nsfc international collaborative research grant . | the pure penguin process @xmath0 is one of the most important probes of physics beyond the standard model .
recently babar and belle have measured the unexpectedly large transverse polarization in the decays @xmath0 , which may single out new physics effects beyond the standard model .
we study the possibility that the phenomenon could serve as an important probe of anomalous tensor interactions . we find that a spin flipped tensor interaction with a small strength and a phase could give a possible solution to the polarization puzzle . + * pacs numbers 13.25.hw , 12.60.jv * ( 0,0 ) looking for signals of physics beyond the standard model is one of the most important missions of high energy physics .
it is well known that flavor - changing neutral currents induced in @xmath1 decays are one of the best probes of new physics beyond the standard model because they arise only through loop effects in the standard model ( sm ) . to this end , the decays @xmath0 are of particular interests , since they are pure penguin processes and have interesting polarization phenomena as well as relatively clear experiment signature . within the sm
, it is expected that both @xmath2 and @xmath3 are mainly longitudinally polarized , while its transverse polarization is suppressed by the power of @xmath4 . however , last year both babar and belle had observed rather small longitudinal polarizations in the decays @xmath5 due to @xmath6 , both groups have measured unexpectedly large transverse polarizations in the @xmath7 decays . this summer babar collaboration has again reported their full angular analysis of the the decay @xmath8
@xcite @xmath9 which has confirmed their previous measurements and called urgent theoretical explanations . the final states @xmath2 and @xmath10 are fast moving in the @xmath1 meson frame and any spin flip of fast flying quark will be suppressed by power of @xmath11 .
the charge interaction currents structure of the sm is left - handed , therefore , will result in the dominance of longitudinal polarization .
such situation has been known to us for many years @xcite .
so that , the recent measurements of large transverse polarizations in @xmath0 are referred as a puzzle within the high energy physics community @xcite . the analysis of the decays within the sm can be performed in terms of an effective low - energy theory with the hamiltonian @xcite @xmath12 the amplitude for the decay within the sm can be written as @xmath13~.\end{aligned}\ ] ] since @xmath1 meson is a pseudoscalar , the final two vector mesons must have the same helicity . in the helicity basis
, the amplitude can be decomposed into three helicity amplitudes , which are @xmath14~,\nonumber \\ h_{\pm\pm}&=&i\frac{g_f}{\sqrt{2}}v_{tb}v_{ts}^ { * } a(\phi k^ * ) m_{\phi}f_{\phi } \left [ ( m_{b}+m_{k^*})a_1 \mp \frac{2m_b p_c } { m_b + m_{k^*}}v \right]~.\end{aligned}\ ] ] in naive factorization @xcite , @xmath15 then the branching ratio is thus read as @xmath16 and the longitudinal and the transverse polarization rates are @xmath17 using the wilson coefficients @xmath18 evaluated at scale of @xmath19 @xcite , and the decay constants @xmath20 gev , @xmath21 gev and the form factors of light - cone qcd sum - rules @xcite , one can get @xmath22 it must be reminded that a theoretical estimation of the branching ratios depend very strongly on the form factors from different hadronic models and the theoretical frameworks of @xmath1 meson nonletponic decays , even though most frameworks and form factors predict dominance of the longitudinal polarizations . for example , recent calculation of @xmath23 by cheng and yang by using qcd factorization @xcite gives @xmath24 for lcsr and @xmath25 for bsw form - factors @xcite , respectively , while pqcd @xcite calculation gives @xmath26 where form - factors are not inputs . however , both studies present the dominance of longitudinal polarization .
after babar and belle measurements of the abnormal large transverse polarization , there have been some theoretical explanations ; namely , through the final state interactions ( fsi ) contributions @xcite , large annihilation contributions and new physics from right - hand current interactions @xcite and transverse @xmath2 from the emitted gluon of @xmath27 which might be enhanced by new physics @xcite . in this letter , we investigate the possibility that the abnormally large transverse polarization may arise from a new tensor interaction beyond the sm ( bsm ) , @xmath28 where @xmath29 is the relative interaction strength normalized to that of @xmath30 in the sm and @xmath31 is the new physics phase . in principle , such a tensor operator could be produced even in mssm @xcite .
interestingly , the recent study of radiative pion decay @xmath32 at pibeta @xcite has found deviations from the sm in the high-@xmath33 kinematic region , which may indicate the existence of a tensor quark - lepton interaction @xcite .
we also note that kagan mentioned the case of tensor operator for resolving the puzzle @xcite .
our starting point arises from the observation that the tensor interaction only contributes to transverse polarization but not to longitudinal one .
the matrix element reads @xcite @xmath34 which is scaled as @xmath35 since @xmath36 for fast flying @xmath2 .
however , if the @xmath2 meson is produced instead from a vector interaction vertex , we will have @xmath37 , and it is easy to understand that the longitudinal polarization dominate over the transverse one by a large factor @xmath38 because of @xmath39 . using the form factors defined in ref .
@xcite , we can write down the amplitude of the tensor operator in eq .
( [ tensor ] ) in naive factorization approximation , @xmath40 from this equation , we can get the new physics contributions , @xmath41~.\end{aligned}\ ] ] compared with eq.9 , the tensor interaction contributions to @xmath42 are enhanced by a factor of @xmath43 . [ cols="^,^ " , ] numerical results are presented in fig .
1 . from fig .
1(a ) , we can find that the transverse polarization in @xmath44 is very sensitive to the presence of new tensor interactions .
for @xmath45 , we can easily find solutions to the polarization puzzle depending on the phase of the tensor interaction .
for example , to account for @xmath46 within @xmath47 , we get intervals @xmath48 , @xmath49 , @xmath50 for @xmath51 , respectively . of course , the branching ratio measurements could also give constraints on such a tensor interaction operator , which are presented in fig .
1(b ) . here
we can see the windows are very narrow because the longitudinal contribution estimated within the sm already saturate the experimental branching ratio .
however , it is well known that theoretical calculations of the branching ratios of hadronic @xmath1 decays suffer from large uncertainties .
it is believed that polarization fractions could be predicted more accurately than the branching ratios , because some of hadronic uncertainties could be cancelled in the former ones . in the future , if theoretical frameworks for hadronic @xmath1 decays could achieve @xmath52 accuracy and their predictions of longitudinal branching ratio still saturate the experimental measurement , the tensor interaction scenario could be ruled out . in such a case , we need not only new physics contributions to transverse part but also new contributions destructive to longitudinal part
. however , it would be very hard to account for the large branching ratio of @xmath53 because the similarity between the amplitudes of @xmath53 and the longitudinal amplitude of @xmath54 in the heavy @xmath55 limit . in conclusion
, we have studied the large transverse polarization puzzle in @xmath54 decays , which is taken as an important probe of an anomalous tensor interactions .
we find that a relatively weak tensor interaction could resolve the puzzle .
if we take the coupling @xmath56 , such a solution might be a signal of new physics with tensor interaction at tev scale . with the running of @xmath1 factories babar and belle ,
we have witnessed many challenging phenomena .
theoretically , we need more accurate and complete framework to clarify whether the sm could explain those abnormal phenomena or not . + _ note added : when we finished our work , we note the paper@xcite where the same tensor operator is also studied . _ |
transitional disks have properties that can be interpreted as intermediate between the protoplanetary and debris disk phases , with an optically thick outer disk and typically a wide ( tens of au ) gap within ( e.g. , * ? ? ?
* ; * ? ? ?
one possible origin for such gaps is dynamical clearing by forming planets .
the transitional disk host lkca 15 has emerged as a benchmark target for this scenario .
lkca 15 is a k5-type star at a distance of @xmath0140pc located in the taurus - auriga star - forming region @xcite .
resolved imaging ranging from the visible to millimeter wavelengths ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) confirmed the morphological structure implied from spectral energy distribution ( sed ) fitting @xcite , all showing an outer disk truncated inward of @xmath050 au .
recent scattered light imaging also revealed an inner disk component within the gap structure @xcite .
lkca 15 stands out in that directly imaged planet candidates have been reported in the system .
the first candidate was published by @xcite , and subsequently two more were proposed by @xcite .
we follow the three - object nomenclature of @xcite , in which ` b ' refers to a candidate at @xmath014.7 au which displays h@xmath3 emission , with ` c ' and ` d ' following counterclockwise .
the h@xmath3 emission of ` b ' makes a particularly compelling case for a protoplanet , as it implies active gas accretion . here
we present new deep , high - contrast images of lkca 15 with sphere ( spectro - polarimetric high - contrast exoplanet research ; @xcite ) , offering the first full map of the gap s outer edge and revealing the inner disk structures hinted at by earlier studies .
these images also serve to complement the observations in which the planet candidates were discovered , where the precise morphology of the emission can not be uniquely determined ( with the exception of the h@xmath3 imaging ) , since they are based on sparse - aperture masking ( sam ) and thus do not fully sample the pupil plane of the telescope .
on 2015 - 12 - 19 , we obtained two sets of imaging polarimetry data with the irdis branch @xcite of sphere at the european southern observatory s very large telescope .
the extreme adaptive optics ( saxo ; @xcite ) yielded a point - spread function ( psf ) with a fwhm of @xmath04 pixels ( @xmath4mas ) in the broadband @xmath1 filter ( ` bb_j ` , 1.26@xmath5 m ) under good conditions ( seeing 0608 in @xmath6-band ) . the first observation , hereafter named deep , aimed at maximizing sensitivity for the faint disk structures and therefore employed the ` n_alc_yj_s ` coronagraph ( apodized lyot , inner working angle 80mas ) with long exposures ( 32s ) .
the observation comprised 15 polarimetric cycles of 2 exposures at each of the 4 standard half - wave plate positions for a total of 3840s of exposure .
the second observation , fast , was executed immediately afterwards without coronagraph and with minimal exposure times ( 0.85s ) to leave the star unsaturated and grant access to the innermost regions .
three polarimetric cycles with 40 exposures per position yielded 102s of exposure .
frame registration was performed on fast by fitting two - dimensional gaussians to the unsaturated target star in each frame . since those measured star locations appeared stable ( jitter @xmath02mas ) with no visible drift
, we adopted their median star location for all deep frames . apart
from removing the coronagraph , no changes were made to the instrument setup between deep and fast runs .
the data were reduced using custom ` idl ` routines implementing the double - ratio polarimetry and empirical correction of instrumental and stellar polarization from @xcite .
we present the results as polar - coordinate stokes components ( @xmath7 , @xmath8 ) , in which positive @xmath7 signals correspond to azimuthally oriented linear polarization @xcite . in face - on or
mildly inclined circumstellar disks , this convention maps almost all scattered light into positive @xmath7 signal .
however , due to lkca 15 s inclination of 50@xmath9 @xcite , multiple - scattering events also produce faint @xmath8 signals @xcite .
@xmath8 therefore provides an upper limit for the noise level .
furthermore , on 2015 - 11 - 29 , we obtained non - polarimetric imaging , hereafter named full , with sphere irdis in dual - band imaging mode ( dbi ; @xcite ) with the ` k1k2 ` filter ( @xmath10 @xmath11 m , @xmath12 @xmath11 m ) as part of the sphere infrared survey for exoplanets ( shine ; chauvin et al .
2016 in prep ) .
pupil tracking was used to allow for angular differential imaging ( adi , @xcite ) .
we obtained a sequence of 4288s totalling a field rotation of 24.5@xmath9 , using the ` n_alc_yjh_s ` coronagraph ( apodized lyot , inner working angle 93mas ) .
non - coronagraphic frames were obtained before and after the coronagraphic sequence for photometric calibration .
all data were reduced with the sphere pipeline @xcite and additional analysis tools implemented at the sphere data centre , including adi with ` klip ` ( 5 subtracted modes , 24.5@xmath9 field rotation ; * ? ? ?
* ; * ? ? ?
however , since these observations are less sensitive to planets than previous studies , we only use them to complement our disk analysis ( see section [ s : outer ] ) .
our @xmath1-band polarimetry of lkca 15 is shown in figure [ f : images ] .
the deep @xmath7 images ( fig . [ f : images]a ) reveal the inner and outer disk components and the gap between them at much higher s / n than previously achieved @xcite . to quantify the detection , we calculate s / n maps ( fig .
[ f : images]b ) by dividing the @xmath7 and @xmath8 images by a radial noise profile obtained by calculating the standard deviation of pixel values in concentric annuli in the @xmath8 image .
both disk components are detected at high significance in @xmath7 , with local s / n exceeding 5@xmath13 over an area of 0.23arcsec@xmath14 and peaking at 11@xmath13 . finally , we present the @xmath7 and @xmath8 images after scaling with an @xmath15 map , where @xmath16 represents the estimated physical distance of each pixel s scattering material from the star , assuming it lies on the disk s inclined midplane ( 50@xmath9 ; fig . [
f : images]c ) .
this counteracts the @xmath17 dependency of the scattered - light intensity and thus reduces the dynamic range of the image , rendering disk components at different separations directly comparable .
this reveals the inner edge of the outer disk as a sharp intensity step ( @xmath060% over 1fwhm , cf .
section [ s : profiles ] ) in @xmath7 , allowing us to trace the edge at all position angles for the first time .
the residuals in the @xmath8 image appear as a faint quadrupole pattern diagonal to the principal axes of the projected disk , as expected from higher - order scattering @xcite .
the brightest structures in @xmath15-scaled @xmath8 are @xmath04 times fainter than those in @xmath15-scaled @xmath7 . figures [ f : images]d f show the same images for fast . despite the lower sensitivity ,
the same structures as in deep can be discerned .
the calculated s / n drops sharply within a radius of @xmath080mas , though this could be due to astrophysical signal from higher - order scattering dominating the @xmath8 image at those separations , causing the noise to be overestimated . for
now , we assume that fast does not detect the disk at smaller separations than deep
. however , the disk structures appear notably sharper in fast .
this may be due to the fact that those were registered on a frame - to - frame basis , whereas the long exposure times and global registration of deep may have introduced some smearing . whereas the imagery in @xcite and @xcite sufficed to discover and confirm the presence of scattered light from the inner disk , our new data now allow us to investigate its morphology .
figure [ f : inner ] shows a comparison of the inner regions of the @xmath15-scaled @xmath7 images from deep ( panel a ) and fast ( b , c ) as well as our previously published @xmath2-band data ( d ; @xcite ) .
all three images show the inner disk in strikingly similar morphology . the polarized flux is contained in a roughly elliptical area comparable in shape and orientation to the outer disk , but about half as large in each dimension .
we note a number of morphological features that appear persistent between the three images , as marked in figure [ f : inner]c : a bright crescent on the near side of the disk , reminiscent of the full - intensity scattered - light appearance of the outer disk @xcite ; a slight asymmetry between the two crescent arms , with the western arm ( ` x ' ) seemingly trending outward while the eastern arm ( ` y ' ) curls inward ; and a local brightening along the far side of the minor axis ( ` z ' ) .
furthermore , the area within this crescent appears darkened , giving the impression of another gap in the inner disk .
however , we caution that polarimetric efficiency drops in the innermost resolution elements , where the positive / negative pattern in @xmath18 and @xmath19 is too small to be cleanly resolved @xcite .
this could mimic a gap .
figure [ f : inner]c also indicates the positions of the three proposed planet candidates ( b , c , d ; cf . * ? ? ?
* lbt joint fit results ) , which coincide with the bright , forward - scattering side of the inner disk .
the analysis and interpretation of sam data is challenging and potentially even misleading in the presence of scattered light from a disk ( e.g. , @xcite , @xcite , @xcite , caceres et al . in prep . ) ; thus , the inner disk might account for some or even all of the sam signal attributed to protoplanets .
on the other hand , h@xmath3 imaging makes a strong case for a protoplanet at ` b ' , which could be confirmed with a re - detection in h@xmath3 showing orbital motion .
such a protoplanet could be responsible for the substructure in the inner disk .
in the @xmath15-scaled @xmath7 images , the outer disk appears as a diffuse region of roughly flat brightness bounded on the inside by a sharp elliptical edge .
we employ the maximum merit method @xcite to fit a parametric ellipse onto the deep @xmath7 image that maximizes the brightness contrast between the annular regions immediately inside and outside of the ellipse , as detailed in @xcite .
we fix the position angle of the major axis to 60@xmath9 ( cf.@xcite ) . the resulting ellipse is shown in solid blue in figure [ f : ellipses]a
. the fit ellipse matches the perceived gap edge accurately . to separate the inner and outer disk for further analysis
, we also fit a second ellipse along which the @xmath15-scaled @xmath7 intensity is minimized .
this ellipse is shown in dotted blue in figure [ f : ellipses]a .
the numerical parameter of both ellipses are given in table [ t : num ] .
for errors , we `` fit '' similar ellipses to @xmath8 and measure the standard deviation of the resulting merits .
the error intervals represent the family of ellipses whose @xmath7 merits are less than one standard deviation below the best fit .
figure [ f : ellipses]b compares the gap edge fit ellipses in @xmath1-band polarimetry ( this work ; solid blue ) to those in @xmath2-band polarimetry ( @xcite ; long - dashed red ) and sub - millimeter interferometry ( @xcite ; short - dashed green ) with respect to the star s position .
our new data confirm the eccentric scattered - light gap contrasting with the symmetric gap in sub - millimeter thermal emission . due to the tapered gap wall
, inclination can not explain this eccentricity @xcite , whereas shadowing from the inner disk could @xcite .
the eccentricity appears slightly lower than in @xmath2-band .
this could be due to the poorly constrained far - side edge in the @xmath2-band data , color - dependent polarized reflectivity , chromatic filtering by the inner disk halo , or a temporal evolution of the shadowing postulated as the origin of the eccentric scattered - light gap .
figure [ f : ellipses]c shows the @xmath1-band gap ellipse superimposed on the @xmath20-band full - intensity klip image of the full data .
the ellipse coincides exactly with the sharp inner edge of the bright forward - scattering crescent , confirming the long - held assumption that such crescents in adi images accurately trace the near - side gap edge even though the fainter parts of the disk suffer from oversubtraction effects .
the deep data are sensitive enough to search for substructure in the outer disk .
while we find no sign of spirals or structural asymmetries , we note significant variations in the azimuthal @xmath7 distribution .
in particular , the disk appears locally dimmed in radial lanes around azimuths 50@xmath9 , 135@xmath9 , 200@xmath9 , and 325@xmath9 ( see figure [ f : ellipses]d , e ) .
furthermore , near - side gap edge shows similar substructure in the full and deep images .
-scaled deep @xmath7 image . *
( b ) * comparison of the best - fit gap edge in @xmath1-band ( this work ; blue solid line ) with those in @xmath2-band ( @xcite ; red long - dashed line ) and sub - millimeter interferometry ( @xcite ; green short - dashed line ) . * ( c ) * full - intensity klip image ( 5 subtracted modes ) of the full data in the ` k1k2 ` filter for comparison .
the gap edge derived from the @xmath7 image coincides very well with the edge of the bright crescent in the klip image . *
( d ) * the image in panel ( a ) at a harder stretch , emphasizing the surface brightness variations in the outer disk .
four position angles with reduced brightness are marked , possibly indicating transient shadowing from the inner disk . *
( e ) * azimuthal profile of the outer disk in @xmath15-scaled deep @xmath7 ( solid ) and @xmath8 ( dashed ) , evaluated in a 7 pixels ( 86mas ) wide annulus outside the best - fit gap edge at 5@xmath9 resolution .
the four position angles from panel ( d ) are indicated with vertical dashes .
the @xmath8 profile is dominated by the multiple - scattering quadrupole . ]
one possible scenario is that these dark lanes indicate shadows cast by inner disk regions @xcite or magnetospheric accretion columns @xcite .
the inner and outer disks do have similar azimuthal structure .
an alternative explanation is that a polarized scattering phase function with peaks in both the forward and backward directions ( e.g. , * ? ? ?
* ) could brighten the projected disk along its minor axis ( stolker et al . in prep . ) .
a case study for this scenario is hd100546 @xcite , where a bright wedge surrounded by two dark wedges was detected along the near - side minor axis .
another analogy between these two datasets is the presence of a tenuous counterpart to the bright wedge in @xmath8 , which is roughly 45@xmath9 displaced from it .
this detection may suggest small departures from centrosymmetric scattering as expected from multiple scattering in inclined disks @xcite . to quantify the amount of polarized light detected ,
we calculate surface brightness profiles in the @xmath7 image without @xmath15 scaling .
we extract a 86mas wide strip along the major axis of the projected disk ( position angle 60@xmath9 ; @xcite ) and a 257mas wide strip along the minor axis , and average the pixel values across the width to obtain one - dimensional profiles .
we then divide each profile by the sum of all pixel values in the unsaturated stellar full - intensity psf ( out to 123 ) and by the solid angle on sky corresponding to a pixel ( 00123@xmath14 ) .
the values of the resulting profiles therefore measure the local polarized flux density of the disk in units of stellar flux per square arcsecond .
the resulting values do not explicitly depend on pixel size or wavelength and therefore allow for direct comparison between optical and infrared measurements .
( this differs from the normalization used in @xcite . )
figures [ f : profiles]a and b show the profiles for the major and minor axis , respectively .
the solid black curve corresponds to deep . to visualize the uncertainty level ,
we extract a number of equivalent profiles from the deep @xmath8 image and add them to the measured @xmath7 profiles , yielding the family of grey curves .
the thin orange curves are the @xmath7 profiles for fast , which match the deep results well .
we also show the corresponding @xmath7 profiles for the @xmath2-band data from @xcite as thin red curves .
while the morphology of the @xmath2- and @xmath1-band profiles is similar , the latter is significantly brighter overall .
we investigate this effect by summing the @xmath7 image over the detectable parts of the inner and outer disks in each filter .
we sample the inner disk from a radius of 70mas out to the flux - minimum fit ellipse ( dotted blue in fig .
[ f : ellipses]a ) , and the outer disk from that ellipse out to a radius of 1225 .
we find that the polarized disk flux to stellar flux ratio is 6.4 times larger in @xmath1 than in @xmath2 for the inner disk , and 6.6 times for the outer disk .
the polarized flux ratio between the inner and outer disk area appears stable with wavelength ( 0.70 in @xmath1 , 0.73 in @xmath2 ) .
judging from the @xmath8 images , these figures are accurate on the order of 10% .
the numerical results are summarized in table [ t : num ] . on sky .
* ( b ) * profiles measured in a 257mas wide strip along the minor axis , centered on the star .
positive separations correspond to a position angle of 330@xmath9 .
solid black lines show the deep @xmath1-band @xmath7 surface brightness profiles .
the family of thin grey lines represents the results of adding strips from the @xmath8 image taken from different position angles to the @xmath7 strip to illustrate the estimated noise level .
thin orange lines show the corresponding profiles from fast , which match deep very well .
thin red lines represent the @xmath2-band profiles from @xcite for comparison , demonstrating that the disk is significantly more reflective in polarization in @xmath1-band than in @xmath2-band . dashed blue vertical lines demarcate the bounds of the @xmath1-band gradient fit ellipse from fig .
[ f : ellipses ] .
dotted red vertical lines show the corresponding bounds for the @xmath2-band fit ellipse from @xcite . ]
l|r@ r@ r@r@ r@ r@ & & + + semimajor axis @xmath21 ( mas ) & 367 & @xmath22 & @xmath23 & 338 & @xmath24 & @xmath25 + ( au ) & 51.5 & @xmath26 & @xmath27 & 47.4 & @xmath28 & @xmath29 + semiminor axis @xmath30 ( mas ) & 235 & @xmath31 & @xmath32 & 245 & @xmath33 & @xmath34 + ( au ) & 32.9 & @xmath35 & @xmath36 & 34.3 & @xmath37 & @xmath38 + major - axis offset @xmath39 ( mas ) & @xmath40 & @xmath41 & @xmath32 & @xmath42 & @xmath24 & @xmath43 + ( au ) & @xmath44 & @xmath45 & @xmath36 & @xmath46 & @xmath28 & @xmath47 + minor - axis offset @xmath48 ( mas ) & @xmath49 & @xmath22 & @xmath50 & @xmath51 & @xmath24 & @xmath25 + ( au ) & @xmath52 & @xmath26 & @xmath53 & @xmath54 & @xmath28 & @xmath29 + + semimajor axis @xmath21 ( mas ) & 276 & @xmath55 & @xmath56 + ( au ) & 38.6 & @xmath57 & @xmath58 + semiminor axis @xmath30 ( mas ) & 196 & @xmath59 & @xmath60 + ( au ) & 27.4 & @xmath59 & @xmath61 + major - axis offset @xmath39 ( mas ) & @xmath62 & @xmath63 & @xmath64 + ( au ) & @xmath62 & @xmath65 & @xmath66 + minor - axis offset @xmath48 ( mas ) & @xmath49 & @xmath67 & @xmath68 + ( au ) & @xmath52 & @xmath69 & @xmath70 + + inner disk & & + outer disk & & + inner / outer ratio & & + + notes .
@xmath1-band results are based on this work ; @xmath2-band results are taken from @xcite .
the gap fitting parameters are explained in section [ s : outer ] , the contrast values in section [ s : profiles ] .
the accuracy of the contrast values is estimated at @xmath010% .
+ the photometry is likely affected by psf - dependent loss of polarimetric sensitivity at small separations @xcite .
however , the fact that we measure consistent color ratios for the inner and outer disks indicates that the differential effect between the two bands is small , most likely because both data sets were taken with the same telescope and ao system .
a higher polarized disk flux at longer wavelengths was also observed for the transition disk hd 135344 b by @xcite , who measured a factor @xmath02 between @xmath1 and @xmath2 bands .
this color factor offers an additional diagnostic for the composition of scattering dust .
our ongoing sphere observations are measuring polarized colors for many protoplanetary disks , which will allow for a comparative study in the near future .
this work has been carried out within the frame of the nccr planets supported by the snsf , and is supported by anr-14-ce33 - 0018 .
eth members acknowledge support from snsf ; inaf members from miur s `` progetti premiali '' ; j.o . from the millennium nucleus rc130007 ( chilean ministry of economy ) and alma / conicyt project 31130027 .
we thank the referee for constructive input . | lkca 15 hosts a pre - transitional disk as well as at least one accreting protoplanet orbiting in its gap .
previous disk observations have focused mainly on the outer disk , which is cleared inward of @xmath050au .
the planet candidates , on the other hand , reside at orbital radii around 15au , where disk observations have been unreliable until recently . here we present new @xmath1-band imaging polarimetry of lkca 15 with sphere irdis , yielding the most accurate and detailed scattered - light images of the disk to date down to the planet - hosting inner regions .
we find what appear to be persistent asymmetric structures in the scattering material at the location of the planet candidates , which could be responsible at least for parts of the signals measured with sparse - aperture masking .
these images further allow us to trace the gap edge in scattered light at all position angles and search the inner and outer disks for morphological substructure .
the outer disk appears smooth with slight azimuthal variations in polarized surface brightness , which may be due to shadowing from the inner disk or a two - peaked polarized phase function .
we find that the near - side gap edge revealed by polarimetry matches the sharp crescent seen in previous adi imaging very well . finally , the ratio of polarized disk to stellar flux is more than six times larger in @xmath1-band than in the @xmath2 bands .
= 1 |
supernova explosions following the gravitational collapse of a massive star ( @xmath1 ) are among the most fascinating events in the universe as they radiate as much energy as the sun is expected to emit over its whole life span @xcite .
nuclear physics is an essential ingredient in the numerical simulations which aim to describe these events , since realistic astrophysical descriptions of the collapse and post - bounce evolution rely on the accuracy of the implementation of weak processes and equation(s ) of state ( eos ) @xcite .
determining the eos constitutes a particularly difficult task since phenomenology ranges from a quasi - ideal un - homogeneous gas to strongly interacting uniform matter and , potentially , deconfined quark - matter .
the situation is even more difficult if phase transitions are experienced , since mean - field models fail in such situations @xcite . the coulomb - quenched liquid - gas ( lg ) phase transition taking place at densities smaller than the nuclear saturation density ( @xmath2 @xmath3 ) is , probably , the most notorious and best understood case @xcite . at highest densities
, a quark - gluon plasma is expected , but predictions on the exact location of the transition are strongly model dependent @xcite . in the intermediate density domain
simple energetic considerations show that additional degrees of freedom may be available , such as hyperons , nuclear resonances , mesons or muons @xcite .
the possibility that the onset of hyperons could pass via a first order phase transition in neutron stars has been evoked in ref .
@xcite , using a relativistic mean field model ( rmf ) , and in ref .
@xcite , a phase transition between phases with different hyperonic species has been observed for cold matter .
the possibility of a first order phase transition to hyperonic matter in effective rmf models has been discussed in refs .
@xcite , too . within the latter models , the phase transition region
is located at sub - saturation densities , and is thus not relevant for star matter . using a simple two - component @xmath4 model ,
we have recently studied the complete phase diagram of strange baryonic matter showing that it exhibits a complex structure with first and second order phase transitions @xcite . however , the exploratory calculation of ref .
@xcite neglects the fact that in addition to baryon number @xmath5 and strangeness @xmath6 , the charge @xmath7 and lepton @xmath8 quantum numbers are also populated .
the thermodynamics of the complete system should thus be studied in the four - dimensional space of the associated charges @xmath9 .
the strict electroneutrality constraint @xmath10 , necessary to obtain a thermodynamic limit @xcite , makes the physical space three - dimensional .
as it is known from the eos studies at sub - saturation density @xcite , the introduction of the charge degree of freedom can have a very strong influence on the phase diagram and can not be neglected .
in this work we therefore introduce a four - component model constituted of neutrons , protons , electrons and @xmath0-hyperons .
electrons are treated as an ideal gas .
we present , in sec .
[ sec : thermo ] of this paper , the thermodynamics and phase transition of the n , p , e and @xmath0 system , and discuss the influence of the coulomb interaction .
the consequence of the phase transition on the cooling of proto - neutron stars , through the neutrino mean free path , is qualitatively discussed in sec .
[ sec : neutrinos ] . finally , we present our conclusions in sec . [
sec : conclusions ] .
in the widely used mean - field approximation @xcite the total baryonic energy density is given by the sum of the mass , kinetic and potential energy density functionals which represents a surface in the three - dimensional space defined by the baryon , strange and charge density given , in our case , by @xmath11 , @xmath12 and @xmath13 .
in the non - relativistic formalism valid in the considered domains of density and temperature it reads @xmath14 the single - particle densities are given by the fermi integrals @xmath15 where @xmath16 is the fermi - dirac integral , @xmath17 is the inverse temperature , @xmath18 is the effective @xmath19-particle mass and @xmath20 is the effective chemical potential of the @xmath19-species self - defined by the single - particle density . a full thermodynamics characterization of the system is provided by the pressure @xmath21 together with the entropy density @xmath22 in mean - field , @xmath23~. \label{eq : entropy}\ ] ] the thermodynamical definition @xmath24 allows to infer the relation among the chemical potentials @xmath25 and the effective parameters @xmath20 as @xmath26 , with @xmath27 . within the numerical applications we shall use the potential energy density proposed by balberg and gal @xcite , @xmath28 accounting for nucleon - nucleon , nucleon-@xmath0 and @xmath0-@xmath0 interactions . @xmath29 denotes the third isospin component of particle @xmath19 . in the non - strange sector
the form of the interaction is the same as in the widely used lattimer - swesty @xcite eos .
let us mention that the observation of a neutron star ( psr j 1614 - 2230 ) with a mass of almost two solar masses @xcite imposes stringent constraints on the hyperonic interaction in dense neutron star matter .
the maximum mass for a @xmath30 system as studied in the present manuscript is @xmath31 with the parameter set bg i for the coupling constants ( see table [ table : bg1 ] ) in agreement with the mass of psr j 1614 - 2230 .
including all hyperonic degrees of freedom , the maximum neutron star mass obtained with parametrisation bg i decreases and becomes slightly too low .
however , the qualitative results discussed here about the thermodynamics of the system and the consequences on the neutrino mean free path are independent of the parametrisation used . in particular , the same qualitative results are obtained with the parametrisations from ref .
@xcite , in agreement with the mass of psr j 1614 - 2230 even upon including all the different hyperons .
quantitative differences are very small , such that we have chosen here to use for numerical applications one parametrisation from the original paper by balberg and gal @xcite , bg i. [ cols="^,^,^,^,^,^,^,^,^,^ " , ] matter stability with respect to phase separation can be checked in any point of the extensive variable space by analyzing the eigen - values of the curvature matrix @xcite , @xmath32 , where @xmath33 and @xmath34 is the total free - energy .
the occurrence of , at least , one negative eigen - value in a certain domain of @xmath35 means that the system is unstable versus phase separation . the associated @xmath36-dimensional gibbs construction can be reduced to a simpler 1-dimensional maxwell construction @xcite by performing a legendre transformation with respect to two out of the three chemical potentials @xmath37 , @xmath38 and @xmath39 .
we have chosen to work in the hybrid ensemble @xmath40 defined by : @xmath41 .
upper ( middle ) : @xmath42-mixture in @xmath43 ( @xmath44 ) coordinates .
lower : @xmath45-mixture in @xmath46 coordinates .
red : liquid - gas phase transition of non - strange dilute nuclear matter ; blue : non - strange to strange phase transition .
the arrows mark the directions of phase separation . ] if the associated equation of state @xmath47 as a function of @xmath48 presents a slope inversion , the relation @xmath49 is three - valued within a given interval of @xmath50
. then a maxwell equal - area construction on this function allows defining two values @xmath51 , @xmath52 , which are characterized by the complete gibbs equilibrium conditions @xmath53 for a multi - component system .
the hyper - surface connecting the two points @xmath54,@xmath55 is the usual gibbs construction .
this procedure is independent of the choice of the densities ( here : @xmath56 ) to be legendre - transformed , provided the order parameter has a non - vanishing component along the remaining density ( here : @xmath48 ) .
if this was not the case , that is if there was no jump in @xmath48 at the phase transition , @xmath57 , the information on the phase transition could not be extracted from the hybrid ensemble eq.([hybrid ] ) .
we have verified that this is never the case , and the phase transition we will identify always separates a more diluted ( lower @xmath48 ) from a denser ( higher @xmath48 ) phase .
for a generic physical multi - component system , this is not always the case and in the general case the convexity properties of the free energy have to be examined with care in order to identify phase transitions in such systems . in particular for our specific physics application of strangeness phase transition
, we will explicitly show that the charge density is almost unaffected by the phase transition .
this means that the concavity of the free energy @xmath58 as a function of @xmath59 is extremely small .
this means that working in that statistical ensemble would have rendered the observation of the phase transition very difficult .
we stress that the one - dimensional maxwell construction in the hybrid ensemble eq .
( [ hybrid ] ) , as long as a @xmath48 jump occurs through the phase transition as it does here , is strictly equivalent to the complete gibbs construction . in particular ,
the pressure , as function of one density with the other densities kept constant , has not a constant value in the mixed phase @xcite .
for instance , @xmath60 at fixed @xmath61 is not constant in the mixed phase region . on the contrary
, a maxwell construction on @xmath60 or @xmath49 at constant values of @xmath56 is never theoretically justified .
the upper part of fig . [ fig1 ] illustrates the projection of the @xmath62 phase diagram in the @xmath43 plane for @xmath63 .
the arrows mark the direction of phase separation which , in case of phase coexistence , coincides with the order parameter .
two phase - coexistence domains may be identified .
the one lying along @xmath64 at sub - saturation density corresponds to the well known lg like phase transition taking place in dilute nuclear matter @xcite .
the second domain lies at supra - saturation densities ( @xmath65 ) , and the direction of phase separation is dominated by the strange density . in the density range shown by the figure ,
this domain is not upper limited in @xmath66 . we observe a bending at very high density meaning that the domain is finite as for the sub - saturation lg transition , but since we do not consider these densities as being realistically described within the present model , we refrain from showing the whole domain here .
this transition is consistent with our previous findings within a simpler 2d model @xcite , though it obviously depends on the assumed strengths of the @xmath67 and @xmath68-interactions .
it is at first sight surprising to observe that the coexistence border of the latter transition is given by simple straight lines . in principle
the coexistence borders of a first order phase transition with three conserved charges are given by two surfaces in the three dimensional space .
for their projection on the @xmath43 plane to be given by a one - dimensional curve , these surfaces have to be perpendicular to that plane , that is independent of @xmath59 . the observed independence on the electric charge shows that the strange charge is the dominant order parameter for this transition .
however , we can expect that some dependence on the electric charge would arise if charged strange particles were included , because of the correlation which would then exist between @xmath59 and @xmath69 . the middle part of fig .
[ fig1 ] illustrates the projection of the phase coexistence domains in the @xmath70 plane for @xmath63 , corresponding to the strangeness - equilibrium condition which is relevant for star matter . as discussed above , the coexistence domain being three - dimensional this representation depends on the value of the third variable given by @xmath71 ( or @xmath69 ) .
the well - known isospin dependence of the lg phase transition occurring at @xmath64 @xcite is apparent .
we have just noticed that the order parameter of the strangeness phase transition is given by a combination of the strange and baryonic density . not surprisingly , the direction of phase separation of this transition in the @xmath70 plane is thus dominated by the baryonic density . the order parameter component in the direction of the electric charge
can be understood as due to the correlation between the different densities .
we are facing a transition between a relatively diluted , non - strange phase to a relatively dense , more strange one .
since @xmath0 s are neutral , the positively charged component of the baryonic density is relatively less important in the dense phase , which explains the slope of the separation direction .
-mixture at t=0 and @xmath72 mev in @xmath73 coordinates .
red : liquid - gas phase transition of non - strange dilute nuclear matter ; blue : phase transition from non - strange to strange compressed baryonic matter .
[ fig : phd_npl_mu ] offers the complementary image on how the two phase coexistence domains look like when plotted with respect to @xmath74 . in this case the condition @xmath63 is released and alternative arbitrary constraints on @xmath75 mev are imposed .
the absence of @xmath0-hyperons at sub - saturation densities makes the conjugated chemical potential undefined .
mathematically , this means that any @xmath76 is possible .
this makes @xmath71 span a semi - infinite domain lower limited by @xmath77 . reminding that - in @xmath78 coordinates - the nuclear matter lg phase coexistence is figured by a curve whose extremities are the critical points
, it is easy to understand that , by fixing @xmath79 , one fixes both @xmath80 and @xmath81 .
now , one can straightforwardly identify the semi - infinite horizontal coexisting lines as the ones corresponding to lg .
the strangeness - driven phase coexistence at fixed @xmath79 appears in @xmath74 as two merged semi - infinite linear segments .
the merging point corresponds to the state where the equilibrium counter - part of the dense phase jumps from vacuum ( low @xmath71 ) to a dilute mixture ( high @xmath71 ) .
we now turn to investigate the influence of coulomb effects on the phase diagram . for simplicity , we will consider only electrons and neglect other charged leptons or mesons . , and @xmath48 at the corresponding critical temperature for @xmath63 .
] electrons are coupled to charged baryons through the electromagnetic interaction , which can modify the baryonic phase diagram .
however , the charge neutrality condition @xmath10 makes the associated chemical potential @xmath79 ill - defined and keeps the problem three - dimensional @xcite . since in homogeneous matter with the condition @xmath82 the coulomb interaction exactly vanishes @xcite ,
the total mean field pressure can be written as a sum over independent terms @xmath83 .
we shall still concentrate on the @xmath84 contribution , as the other terms do not affect the convexity properties of the thermodynamical potential on which phase transitions rely .
we have constructed the full phase diagram of the @xmath85 system in the @xmath86 space using the hybrid ensemble @xmath87 in practice , the charge neutrality condition gives @xmath88 , which allows to infer the electron chemical potential , @xmath89 , via @xmath90 , \label{eq : mue_ne}\ ] ] and to obtain @xmath91 .
note that @xmath92 stands here for the net electron density .
if neutrinos were in equilibrium , then @xmath93 would correspond to the electron neutrino chemical potential .
in particular , in a cold neutron star in @xmath94-equilibrium we would have @xmath95 fixing @xmath96 . in core collapse events , on the other hand ,
neutrinos can not be considered in equilibrium in the major part of the system . here
we want to study the entire phase diagram , not restricting to @xmath94-equilibrium , and we therefore leave @xmath93 free .
neutrinos , even in equilibrium , would not change the phase properties , and are thus neglected for the sake of simplicity in the present discussion .
the lowest panel of fig .
[ fig1 ] depicts the phase coexistence regions of the @xmath85 system at @xmath62 and @xmath63 . in agreement with the results of ref
@xcite , a strong coulomb quenching of the lg - phase transition is obtained .
however , the coexistence domain of the strangeness - driven phase transition is practically un - modified .
this can be easily understood from the fact that the effect of the neutrality condition @xmath82 on the two phase transitions is very different .
the phase transition at sub - saturation density has the total baryonic density as order parameter . at such densities
, @xmath48 is strongly correlated to @xmath97 because of the nuclear symmetry energy which favorizes symmetric @xmath98 matter .
the phase transition thus implies a discontinuity in @xmath88 , which is strongly disfavored by the huge electron incompressibility @xcite . at supersaturation densities
the order parameter is given by @xmath69 which is very loosely correlated to @xmath59 .
the phase transition thus does not imply any strong change in the electron distribution and the presence of electrons thus does not influence much the phase diagram .
the temperature dependence of the phase diagram along @xmath63 is presented in fig .
we can observe that the direction of phase separation is almost independent of @xmath99 .
more interesting , starting from a finite value of @xmath99 , a critical point appears and survives up to very high temperature . on fig .
[ fig2bis ] the critical temperature and the electron fraction @xmath100 are shown as a function of baryon density .
these values are typically reached within the cooling proto - neutron star , meaning that effects of criticality should be experienced .
the cooling of proto - neutron stars is mostly driven by neutrino diffusion during the first seconds . to explore the consequence of criticality for the cooling of proto - neutron star
, we therefore turn to calculate the mean free path for the neutrino scattering off n , p , and @xmath0 particles including the long - range correlations , essential for the study of criticality , in the linear response approximation . in the non - relativistic limit for the baryonic components , the mean free path at temperature @xmath99 of a neutrino with initial energy @xmath101
is given by @xmath102 @xcite where the contribution of the vector channel is defined as , @xmath103 and that of the axial channel , @xmath104 in eqs .
( [ eq : lambdav])-([eq : lambdaa ] ) , @xmath105 is the fermi constant , @xmath106 is the angle between the initial and final neutrino momentum ( = @xmath107 ) , @xmath108 is the transferred energy - momentum , @xmath109 , and @xmath110 is the fermi - dirac distribution of the outgoing neutrino .
@xmath111 ( @xmath112 ) are the dynamical response function in the vector ( axial ) channel .
since this study is focused on the impact of the density fluctuations close to the critical point , only the vector channel is considered . for densities close to the critical point
, spin - density fluctuations are however expected to be small @xcite .
the dynamical response function in the vector channel is defined as @xmath113 where @xmath114 is the vector - polarization matrix for the three species n , p , and @xmath0 , given by the lindhard functions in the case of the mean - field approximation and by the solution of the bethe - salpeter equations in the case of the mean - field+rpa approximation @xcite .
the vector coupling constants are set to be : -1 ( n ) , 0.08 ( p ) , -1 ( @xmath0 ) @xcite .
the residual p - h interaction is derived from the potential energy ( [ eq : epspot ] ) and is closely related to the curvature matrix without electrons @xcite .
the neutrino mean free path along an arbitrary @xmath115 trajectory in the phase diagram which passes by the critical point ( see dot - dashed line in fig . [
fig2](a ) ) is shown in fig .
[ fig : mfp ] .
as expected , the rpa correlations strongly reduce the neutrino mean free path close to the critical point , similar to the critical opalescence effect observed for the photon scattering off matter in critical water .
the ratio of the neutrino mean free path in mean - field+rpa approximation over that at the mean - field level is shown in panel ( b ) exploring different neutrino energies around the neutrino chemical potential defined at beta equilibrium .
the effect of the rpa correlations around the critical point is almost independent of the neutrino energy in agreement with the interpretation as critical opalescence .
to conclude , we have studied the phase diagram of a mixture constituted of interacting neutrons , protons and @xmath0-hyperons under the condition of strangeness - equilibrium , relevant for supernovae and neutron star physics . at supra - saturation densities ,
a strangeness - driven phase transition can take place , depending on the assumed strengths of nucleon-@xmath0 and @xmath0-@xmath0 interactions @xcite .
this second transition survives the screening effect of electrons and persists over a large domain of temperatures such that it may have an impact on star phenomenology . for a first study of this equation of state ( eos ) within core - collapse supernovae ,
see @xcite .
in addition to the eos , linear response theory shows that the neutrino mean - free path dramatically decreases close to the critical point of this phase transition , which occurs in a thermodynamic domain accessible to newly - born proto - neutron stars .
these results present a first step , and quantitative results might be somewhat modified in the presence of other strange- and non - strange baryonic , leptonic or mesonic degrees of freedom .
this work is in progress and it will make the subject of a forthcoming publication .
this work has been partially funded by the sn2ns project anr-10-blan-0503 and it has been supported by compstar , a research networking program of the european science foundation .
r. r acknowledges partial support from the romanian national authority for scientific research under grants pn - ii - id - pce-2011 - 3 - 0092 and pn 09 37 01 05 and kind hospitality from lpc - caen .
99 h. a. bethe , supernova mechanisms
. reviews of modern physics , 62 , 801 ( 1990 ) .
janka , k. langanke , a. marek , g. martinez - pinedo and b. mller , phys . rept .
* 442 * ( 2007 ) 38 .
j. m. lattimer and m. prakash , phys .
442 , 109 ( 2007 ) .
n. k. glendenning , phys .
342 , 394 ( 2001 ) . c. ducoin , ph .
chomaz and f. gulminelli , nucl .
phys . a 771 , 68 ( 2006 ) .
m. barranco and j. r. buchler , phys .
c 24 , 1191 ( 1981 ) .
s. shlomo and v. m. kolomietz , rep .
68 , 1 ( 2005 ) .
f. douchin , p. haensel and j. meyer , nucl .
phys . a 665 , 419 ( 2000 ) .
a. rios , nucl .
phys . a 845 , 58 ( 2010 ) .
i. sagert _
et al . _ ,
* 102 * , 081101 ( 2009 ) .
n. glendenning , phys .
b114 , 392 ( 1982 ) .
j. schaffner - bielich , m. hanauske , h. stcker and w. greiner , phys .
* 89 * ( 2002 ) 171101 .
j. schaffner - bielich and a. gal , phys .
c * 62 * ( 2000 ) 034311 .
p. wang , d. b. leinweber , a. w. thomas , a. g. williams and , phys .
c * 70 * ( 2004 ) 055204 .
l. yang , w. l. qian , r. -k .
su , h. q. song and , phys .
c * 70 * ( 2004 ) 045207 .
l. yang , s. y. yin , w. l. qian , r. -k .
su and , phys .
c * 73 * ( 2006 ) 025203 .
f. gulminelli , ad .
r. raduta and m. oertel , phys .
c * 86 * , 025805 ( 2012 ) . c. ducoin
, k. h. o. hasnaoui , p. napolitani , ph .
chomaz and f. gulminelli , phys .
c * 75 * 065805 ( 2007 ) .
c. ducoin , ph . chomaz and f. gulminelli , nucl .
a789 * , 403 ( 2007 ) .
j. r. stone , p. a. m. guichon , and a. w. thomas , arxiv e - prints ( 2010 ) , 1012.2919 .
m. oertel , a. f. fantina and j. novak , phys .
c * 85 * , 055806 ( 2012 ) .
s. weissenborn , d. chatterjee and j. schaffner - bielich , nucl . phys . *
a881 * , 62 ( 2012 ) .
i. bednarek , p.haensel , j. l. zdunik , m. bejger and r. manka , astron .
& astrophys . 543
( 2012 ) 157 .
f. hofmann , c. keil and h. lenske , phys .
c * 64 * , 025804 ( 2001 ) . l. bonanno and a. sedrakian , astron . & astrophys .
* 539 * , a16 ( 2012 ) .
r. lastowiecki , d. blaschke , h. grigorian and s. typel , acta phys . pol .
suppl . 5 , 535 ( 2012 ) .
s. balberg and a. gal , nucl
a625 * , 435 ( 1997 ) . j. m. lattimer and f. d. swesty , nucl . phys . *
a535 * , 331 ( 1991 ) .
p. demorest _
_ , nature * 467 * ( 2010 ) 1081 . j. margueron and ph .
chomaz , phys . rev .
* c 67 * , 041602 ( 2003 ) . c. ducoin , j. margueron and ph .
chomaz , nucl . phys . * a 809 * , 30 ( 2008 ) . c. providencia , l. brito , s. s. avancini , d. p. menezes and ph .
chomaz , phys .
c * 73 * , 025805 ( 2006 ) .
n. iwamoto and c.j .
pethick , phys . rev .
* d 25 * , 313 ( 1982 ) .
j. navarro , e.s .
hernandez and d. vautherin , phys .
rev . * c 60 * , 045801 ( 1999 ) .
a. polls , a. ramos and i. vidaa , phys . rev .
* c 65 * , 035804 ( 2002 ) .
j. margueron , i. vidaa and i. bombaci , phys . rev .
* c 68 * , 055806 ( 2003 ) .
s. reddy , m. prakash and j.m .
lattimer , phys . rev . *
d 58 * , 013009 ( 1998 ) .
b. peres , m. oertel and j. novak , phys . rev . *
d 87 * , 043006 ( 2013 ) . | the phase diagram of a system constituted of neutrons , protons , @xmath0-hyperons and electrons is evaluated in the mean - field approximation in the complete three - dimensional space given by the baryon , lepton and strange charge .
it is shown that the phase diagram at sub - saturation densities is strongly affected by the electromagnetic interaction , while it is almost independent of the electric charge at supra - saturation density . as a consequence ,
stellar matter under the condition of strangeness equilibrium is expected to experience a first as well as a second - order strangeness - driven phase transition at high density , while the liquid - gas phase transition is expected to be quenched .
an rpa calculation indicates that the presence of this critical point might have sizable implications for the neutrino propagation in core - collapse supernovae . |
the shape of the stellar initial mass function ( imf ) and whether it is universal or not are key issues in astrophysics . for clusters within 2 kpc , there is no compelling evidence for variations in the stellar imf ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or the brown dwarf imf ( e.g. * ? ? ?
* ) . however , these clusters only span a limited range in total cluster mass ( @xmath4 m@xmath2 ) and all have a metallicity similar to the solar value .
thus , we are forced to observe more extreme regions of star formation in search of variations in the imf as a function of environment .
it has been suggested that the shape of the imf and in particular the characteristic mass where the imf flattens from a salpeter power
law could depend on the metallicity in the molecular cloud out of which the stars are formed .
@xcite , @xcite , and @xcite suggest that a lower metallicity results in higher temperatures in the molecular cloud which would increase the jeans mass . this would in turn result in a top heavy imf relative to the solar metallicity imf . the closest place with massive metal
poor young star clusters is the large magellanic cloud ( lmc ) .
the metallicity is only @xmath5 the solar value @xcite and star clusters can be studied in some detail despite a distance of @xmath150 kpc @xcite .
of particular interest is the 30 dor cluster which is powering the most luminous hii region in the local group @xcite .
the cluster has a mass of at least 2.2@xmath6 m@xmath2 within a radius of 4.7 pc @xcite and is a relatively low - mass analog to the more distant starburst clusters .
r136 lies at the center of the 30 dor cluster and has long commanded significant attention : once thought to be a single @xmath11000 m@xmath2 star @xcite , the region is now known to host numerous o stars @xcite .
the whole 30 dor region , with a size of 200 pc , appears to have an age spread of @xmath120 myr @xcite with stars still forming @xcite .
r136 appears to have a much smaller age spread of at most a few myr @xcite .
an age of 2 myr or less is inferred from spectroscopy of the o stars in the very cluster center @xcite , whereas the intermediate mass population is thought to be @xmath134 myr old @xcite .
@xcite obtained hst spectroscopy of the 65 bluest and most luminous sources within 17 of the cluster center .
they derived the imf over the mass range 15120 m@xmath2 and found it to be well approximated by a power
law @xmath7 with a slope of @xmath8 , consistent with a salpeter slope imf @xcite .
@xcite obtained ( ) and ( ) band optical photometry utilizing hst / wfpc2 in order to resolve the cluster s intermediate mass stellar population .
the imf derived for different annuli out to a radius of 4.7 pc was found to be in the range @xmath9 for the mass range 2.815 m@xmath2 , again consistent with a salpeter slope imf .
@xcite combined their results for the high mass imf with the results from @xcite in order to constrain the imf from 2.8 m@xmath2 up to 120 m@xmath2 . comparing the number of high mass stars predicted by the intermediate
mass imf from @xcite , they found the number of massive stars was consistent with a single power
law imf with a salpeter slope , i.e. @xmath10 . combining the two data sets used in @xcite , @xcite derived the imf between 1.35 m@xmath2 and 6.5 m@xmath2 , extending the imf determination into the mass range where the stars are still in their pre
main sequence phase .
the imf was derived in a box with the dimensions @xmath1304@xmath0268 ( [email protected] ) , but excluding the inner most [email protected] ( [email protected] ) .
again , a salpeter slope was found down to 2 m@xmath2 , but the imf was found to be flatter than salpeter , @xmath11 , between 1.35 m@xmath2 and 2 m@xmath2 , suggesting the characteristic mass is higher in this massive , metal poor cluster than @xmath1 0.5 m@xmath2 as found in the galactic field @xcite .
the foreground ( a@xmath12 mag ) and differential extinction ( a@xmath13 mag ) within the cluster @xcite makes it desirable to observe the cluster in the infrared , for example the band where the extinction is less than 20% that of the band .
in addition , pre
main sequence stars are often associated with circumstellar disks and outflows which will introduce additional extinction for the clusters low
mass content .
we have observed r136 with hst / nicmos camera 2 through the band , which is similar to a ground based filter .
the observations were aimed at being sensitive to objects below 1 m@xmath2 for a stellar population with an age of 3 myr .
preliminary results have previously been presented in @xcite , and @xcite .
the paper is structured as follows .
the data and their reduction is described in section 2 .
section 3 shows the results for the band imaging .
the imf is derived in section 4 and compared with the imf derived by @xcite .
we point out several plausible reasons for the different results in the optical and near infrared , including mass segregation , and differential extinction .
finally , our conclusions are presented in section 5 .
we have obtained hst / nicmos camera 2 images through the band of the central 56@xmath057 region around r136 in the 30 dor cluster ( hst program i d 7370 ) .
the observations were centered on the cluster ( ra , dec)=(05:38:43.3,@xmath1469:06:08 ) and on two adjacent control fields centered on ( 05:38:42.4,@xmath1468:52:00 ) , and ( 05:38:56.9,@xmath1468:52:00 ) .
the observing dates were oct 14 and 16 , 1997 .
the field - of - view of the 256@xmath0256 pixel nicmos camera 2 is 19 @xmath019 with a pixel scale of 0075 , resulting in nyquist sampling of diffraction limited band data . each position in a 3@xmath03 mosaic centered on r136 was observed four times with small dithers of @xmath116 pixels .
the data were obtained in non destructive multiaccum mode such that the photometry of the bright stars can be retrieved due to the first short integration in each exposure . the integration time for each dither position was 896 seconds , resulting in a total integration time of 3584 seconds for each position in the mosaic .
the two control fields were observed in a similar manner .
the location of the mosaic is shown in fig .
[ overviewfig ] and the nicmos mosaic is shown in fig .
[ 30dormos ] .
the faintest stars visible with the stretch used here have an magnitude of @xmath121.5 mag , corresponding to a mass of 0.8 m@xmath2 , based on the pre - main sequence models of @xcite , adopting an age of 3 myr @xcite , half solar metallicity , and an extinction of a@xmath15 mag ( see section 3.2 ) .
for comparison , the similar detection limit in an uncrowded environment without nebulosity would be @xmath123.5 mag according to the nicmos exposure time calculator .
each individual image was processed through the calnica and calnicb pipelines as well as the biaseq and pedsky procedures within the iraf environment .
the tasks are described in detail in the nicmos data handbook .
we used synthetic dark frames and flat fields created for the appropriate instrument temperature at each exposure . the biaseq task corrects differences in bias levels for each chip between different sub - exposures .
the pedsky task corrects differences in the bias level for each quadrant of the chip when the array is reset before the exposure .
the data for each position in the mosaic were combined using the drizzle task . the reduced pixel size ( 00375 )
was chosen as half the detector pixel size .
bad pixels , bad columns , and the coronagraphic hole were flagged as bad pixels before the images were combined .
source detection was done using daofind and photometry was performed via point spread function ( psf ) photometry utilizing allstar within the iraf environment .
it was difficult to obtain a good psf model from the data due to the high degree of crowding and the spatial variability of the psf .
instead , the tinytim software @xcite was used to create a synthetic psf .
tinytim allows to create a psf that varies as a function of the location on the array .
the source detection and photometry was performed on each individual position in the mosaic due to the linearly varying point spread function ( psf ) .
a hot template star ( o5v ) was used for the spectral energy distribution in order to achieve the best fit for the brightest stars to limit their residuals .
the tinytim psf was created for five different positions on the nicmos camera 2 array and the psfs were placed in an empty frame with the same number of pixels as the nicmos camera 2 array .
four frames were created with offsets between each psf identical to the offsets used for the science data in order to replicate the data as closely as possible .
the four psf frames were then combined using drizzle together in the same manner as the science data and a linearly varying psf was created from the drizzled frame .
source detection is complicated due to the diffraction features present in nicmos data .
adoption of a low threshold for source detection led to numerous diffraction spots from bright stars being erroneously identified as fainter stars .
instead the source detection was done in the following way in order to limit false detections .
we first detected the brightest stars ( brighter than 1000@xmath16 ) in each frame and used allstar to remove these with the synthetic psf .
a search for fainter stars ( brighter than 500@xmath16 ) was then performed in the frame with the bright stars removed .
since the removal of the brightest stars also removed the diffraction pattern associated with them , we did not detect the diffraction spots as stars .
the two star lists ( the brightest stars and the fainter stars ) were joined into one and these stars were removed from the original frame , again using allstar .
fainter stars are then found from the frame with the already detected stars removed .
this process was iterated until stars at 10 @xmath16 peak pixel intensity over the background were detected and removed .
the frame with the stars removed was then ring
median filtered to remove stellar residuals but to retain the large - scale nebulosity in each frame .
the median filtered image was then removed from the original frame and the star detection process was repeated in this frame but now continued to a detection threshold of 5@xmath16 .
a 5@xmath16 instead of e.g. a 3@xmath16 threshold was selected to limit the risks of false detections due to noise spikes .
we finally made sure by visual inspection that every detection indeed was a point source and that it was not a spurious detection due to the diffraction spikes and spots from bright stars .
the main interest here is in the low mass ( faint ) stellar content in r136 and one concern is the detection of residuals from the bright stars as false stellar objects .
some false sources were detected by daofind but are rejected during the psf fitting routine .
a few remained from the brightest stars .
they typically produced at most a few false detections in the diffraction spikes that were @xmath167 mag fainter than the bright source .
we removed these detections together with other false detection through the visual inspection of all sources .
we have further utilized the artificial star experiments described below to examine how many detections are false due to the residuals from bright stars .
we had only false positives associated with the brightest artificial stars ( @xmath17 12 mag ) .
for artificial stars fainter than @xmath18 mag , no false detections were present .
the false detections for the bright stars were located at the diffraction spikes and would have been identified in the manual inspection of the source list .
we found a total of 10108 uniquely detected sources with a formal error smaller than 0.1 mag and brighter than = 22.5 mag in the 9 frames .
below this magnitude limit the incompleteness is substantial , as discussed below .
table [ sources ] presents the list of detected stars .
the effects of crowding were examined by placing artificial stars in the individual frames using the psf created from the synthetic tinytim psf .
the artificial stars followed a luminosity function with a similar slope to that of the observed stars ( see section 3 ) but with a surface density 10% that of the detected number of stars to avoid affecting the crowding characteristics of the real stars .
we performed 100 artificial star experiments for each frame , for a total of 10 times more artificial stars than real stars .
[ 30dor_corrections ] shows the resulting recovery fractions as a function of the input magnitudes for several annuli around the cluster center .
the difference of the size of the error bars as a function of distance from the cluster center is due to a lower number of artificial stars placed in the central parts of the cluster .
this is a consequence of adding 10% artificial stars relative to observed stars in each artificial star experiment and the relative number of stars in each annulus .
the imf is not determined in regions with this low completeness .
we are mainly interested in the low - mass stellar content of the cluster , which is below the 50% completeness in the central parts . the uncertainty in the completeness corrections for the inner parts of the cluster will therefore not affect the conclusions drawn for the stellar populations further out .
the completeness is a strong function of the radial distance from the center . for the outer regions of the cluster , 50% or more of the stars brighter than @xmath19 mag
are detected , whereas only the very brightest stars are detected in the innermost region . in an annulus at 0.6 - 1 pc radius from the center ,
we detect 50% or more of the stars brighter than @xmath20 mag . adopting the pms models of @xcite and the main sequence models of @xcite ,
@xmath19 mag corresponds to a 0.8 m@xmath2 , half solar metallicity , 3 myr old object , whereas @xmath21 mag corresponds to a 7.5 m@xmath2 star , assuming an extinction of a@xmath15 mag in both cases .
we have investigated the accuracy of the derived photometry by using the stars detected in the overlap regions of several fields .
[ brandl_comp ] shows the difference in derived magnitude for stars detected in the overlap regions of the mosaic .
dots denote stars outside a 2 pc radius and plus signs denote stars between 1.252 pc radius , respectively . the band photometry has been compared with the ground based band photometry obtained using adaptive optics observations by @xcite .
[ brandl_comp ] shows the magnitude difference between the adaptive optics photometry and this study based on 829 stars common to both datasets .
stars were considered detected in both datasets if the spatial position coincided within 2.5 drizzled pixels , corresponding to 0094 .
some scatter is present between the two datasets , especially for the fainter stars .
however , the median difference between the magnitudes derived for the two datasets is less than 6% for objects @xmath22 mag .
we have in the following treated the observations as a standard cousins band .
conversely , there appears to be a tendency for the fainter stars to be brighter in the data than in the band data of @xcite .
the tendency for the fainter stars to be skewed towards fainter band magnitudes is an effect also seen in other comparisons between hst / nicmos and ao data ( e.g. * ? ? ?
* ) who suggest it is due to the extended halos present in ao observations around bright stars .
the immediate results from the band hst photometry are presented . after discussing the luminosity function for different annuli ,
the luminosity profile for the cluster is derived .
the band data are combined with the optical hst data by @xcite and the color
magnitude diagrams are presented .
utilizing the two color
magnitude diagrams we show that the spread observed for the higher mass stars is consistent with that expected due to reddening .
we estimate the average age for the stellar population and discuss the possible presence of an age spread .
the star counts in the central 0.6 pc radius region is heavily affected by low number statistics , crowding even for the brightest stars , and relatively uncertain incompleteness corrections .
we therefore focus on the sources outside 0.6 pc in this paper .
[ lfs ] shows the band luminosity functions for the 0.67 pc radius region of the 30 dor cluster divided into several radial bins to show the difference in photometric depth due to crowding .
overplotted are the completeness
corrected lfs , where each bin has been divided by the corresponding recovery fraction from the artificial star experiments . the completeness corrected luminosity functions are relatively smooth and have been fitted with power - laws down to the 50% completeness limit
. the derived slopes with their 1@xmath16 uncertainties and the 50% completeness limits are presented in table [ hlf_slopes ] .
although the slope in the inner annulus is found to be more shallow , the derived slopes are consistent with each other within 2@xmath16 with an average slope of 0.31 and the shallow slope is not significant .
the completeness corrected combined histogram for the stars detected in the two off cluster control fields is shown in the lower right panel in fig .
[ lfs ] . from the histogram
, it can be seen that the field star contamination found from the star counts is @xmath231015% for the faintest stars in the 57 pc annulus and less closer to the center as well as for brighter stars .
although the contamination of field stars is found to be relatively small it is not negligible and they are therefore statistically subtracted from the cluster population in the following analysis ( section 4 ) .
next , the band photometry is combined with the optical data presented by @xcite .
a star was considered detected in both surveys if the spatial position agreed within 2.5 drizzled nicmos camera 2 pixels ( 0094 ) . in the cases where two optical stars were located within the search radius of the star detected in the nicmos camera 2 observations ,
the brightest star was chosen as the match .
we find in total 2680 in common with the @xcite survey that detected 3623 stars the inner 35 of the cluster .
1848 of those sources have a combined formal photometric error in the color of less than 0.1 mag . within the area covered by @xcite
we detect a total of 5095 sources .
most of the stars detected by the nicmos survey but not the wfpc2 observations are fainter than @xmath24 20 mag . assuming an object age of 3 myr and an average extinction of a@xmath15 mag ( see below ) , the similar object would have a magnitude in the band of @xmath122 mag . for objects with more extinction , they will be even harder to detect in the band .
@xcite essentially do nt detect any stars within a 1 pc radius at this magnitude or fainter .
only 1 in 4 stars in the magnitude interval = 2122 mag was detected outside 1 pc .
it is thus not surprising that a significant population of faint stars are detected in the nicmos survey relative to the wfpc2 survey . nevertheless , the lower spatial resolution of the nicmos observations results in a low recovery fraction at these magnitudes in the central few pc of the cluster . the majority of the sources not detected in the nicmos survey but at optical wavelengths are located within a radius of 1 pc .
the lack of detection is due to the lower spatial resolution in this study relative to the optical hst data .
the resolution is almost a factor of two better in the band than in the band .
sources not detected in the nicmos data outside 1 pc are mainly due to crowding as well .
indeed , visual inspection of the location of the stars detected in the optical but not near infrared shows they are often located either very close to the core or on the first airy ring of a bright source .
the - versus color magnitude diagram is shown in fig .
overplotted are a 3 myr isochrone for the high mass stars adopted from the @xcite models and 2 , 3 , and 4 myr isochrones adopted from @xcite for stars below 7 m@xmath2 .
the stars above 7 m@xmath2 and up to the maximum mass we fit the imf in section 4.1 ( 20 m@xmath2 ) are all expected to be on the main sequence .
both isochrones were calculated adopting a metallicity of half the solar value , typical for the lmc @xcite .
the two isochrones have a small offset in both the ( 0.06 mag ) and ( 0.07 mag ) band .
we have forced the @xcite isochrone to match the @xcite isochrone at 7 m@xmath2 .
it is evident there is a significant scatter in the color magnitude diagram .
the scatter is likely due to a combination of binary systems ( both physical and chance alignments ) , differential extinction , photometric errors and a possible age spread .
the median extinction is found for the main sequence part of the isochrone . for objects in the range 720 m@xmath2
, we find a median extinction of a@xmath15 mag which is slightly higher than the reddening found by @xcite in the inner part of the 30 dor region . at masses below 7 m@xmath2 ,
the spread in the color magnitude diagram is larger but almost exclusively extends to the red part of the diagram .
this indicates the lower
mass objects on average have an excess amount of extinction relative to the higher mass objects .
@xcite observed stars more massive than @xmath25 m@xmath2 and would not detect the additional reddening for the lower mass sources .
the possible sources for the additional reddening is described in sec .
we have estimated an average age for the cluster by utilising the fact the isochrone is almost horizontal in the color range -=1.52.5 mag and around @xmath119 mag .
the median magnitude is 19.0 mag in this region of the color magnitude diagram . adopting an average extinction of a@xmath15 mag
, this corresponds to the magnitudes of the 3 myr isochrone in the same color range .
we have thus adopted 3 myr as the mean age of the low mass cluster population and a 3 myr isochrone is adopted to create a mass luminosity relation in order to turn the luminosities into masses for objects below 7 m@xmath2 and the 3 myr @xcite isochrone above .
we will in sect .
4 the effects on the derived imf adopting an age spread of 2 myr . the right hand panel in fig .
[ cmd ] shows the versus color magnitude diagram .
it is evident that the clustering around the isochrone is tighter than for the versus color magnitude diagram .
this is expected if a large part of the scatter is due to differential extinction .
we can calculate the scatter around the main sequence in both color magnitude diagrams and compare with the difference predicted from extinction .
if the spread in the color magnitude diagrams is due to extinction we expect the ratio of spread in the versus diagram to be ratio of the extinction in each color , i.e. @xmath26 times larger than in the versus color magnitude diagram .
since the isochrone is almost vertical in both diagrams , we have calculated the standard deviation around the reddened isochrone in both color magnitude diagrams .
we have used the stars with good photometry , better than 5% in each filter , and in the magnitude range @xmath27mag .
the standard deviation found for the and color magnitude diagrams are 0.60 mag , and 0.36 mag and the ratio is 1.7 .
if the measurement errors are taken into account this ratio increases .
the typical errors for the culled sample are 0.04 , 0.02 , and 0.03 mag for the v , i , and bands , respectively . after taking the measurement errors into account ,
the ratio is found to be 1.9 , assuming the measurement errors in two filters are independent . due to blending , this is not necessarily the case .
thus , 1.9 is an upper limit and we thus find the ratio to be between 1.7 and 1.9 , in agreement with the scatter being due to differential extinction . since the amount of differential extinction does not affect the band photometry significantly , the single band photometry presented here is competitive with the 2band optical photometry .
there is a unique translation from the band magnitude to the object mass for the majority of the mass range . for the optical photometry
, the color information is used to determine the extinction and the mass function is thus effectively determined by the de reddened band magnitude .
the main advantage of the optical relative to the near infrared hst photometry is the improved resolution due to the smaller diffraction limit .
the stellar content can therefore be resolved to lower masses closer to the cluster core than is possible with the near infrared observations .
however , phenomena associated with the star formation process can introduce additional reddening that can complicate the derivation of the low
mass imf from optical data .
mass objects may still be associated with a circumstellar disk .
there is evidence from e.g. the orion nebula cluster that circumstellar disks can survive the uv radiation from massive stars @xcite .
even if the disks are being evaporated by the ration field from the early type stars , the evaporated material will be a further source of reddening .
patchy extinction associated with the 30 dor complex and located in the foreground of r 136 will be an additional source of differential reddening .
there are signs in the optical images presented in fig . 1 of @xcite of patches of extinction , e.g. to the east north east of the cluster center .
if variable extinction is present or if a significant fraction of the stars are associated with disks or outflows , an extinction limited sample has to be created in order to avoid a biases against detection of the low mass stars .
the near infrared photometry is effected by differential extinction as well but the effect is less than 20% of that measured in the band . thus , whereas the imf derived from optical observations where an extinction limited sample is not defined might be severely affected for the low
mass objects , the effect on near infrared observations is modest
. therefore , in the outer parts of the cluster where crowding is a smaller issue than closer to the center , the near infrared observations are more suitable to detect and characterise the low mass stellar population in the cluster .
on the other hand , single band photometry has the disadvantage that there is no information on the age of individual objects .
we investigate in the next section how this might affect the derived imf .
we note that if differential extinction is present , the situation is no better for the optical photometry .
even though the cluster was observed through two filters in the optical , there is still a degeneracy between age and extinction .
@xcite converted the
photometry into an effective temperature and used that effective temperature to obtain a bolometric correction . without de reddening the sources , the age of a cluster member can be in error and hence the mass estimates will be uncertain .
we construct a mass luminosity relation by combining the main sequence models by @xcite , and the pre main sequence models of @xcite in order to infer the stellar mass from the band magnitude .
we then derive the mass functions for r136 outside 0.6 pc where the 50% completeness limit corresponds to a stellar mass below 10 m@xmath2 . deriving the imf this way
is a well established procedure @xcite .
we further discuss the potential effect of extinction on the derived imf . finally , we search for evidence for mass segregation in the outer parts of the cluster using the cumulative luminosity functions .
a mass luminosity relation is needed to convert the derived band magnitude for each star to a mass .
we use the @xcite isochrones for stars below 7 m@xmath2 and the @xcite 3 myr isochrone for the more massive stars as discussed in section 3.2 .
the age of the cluster is first assumed to be 3 myr and is later varied to examine the effects on the derived imf for different cluster ages .
stars below @xmath13 m@xmath2 are on the pre
main sequence isochrone whereas the more massive stars up to our upper mass limit of 20 m@xmath2 ( see below ) are on the main sequence .
the adopted mass luminosity relation is shown in fig .
[ mlrel ] .
we have limited knowledge of the extinction for the majority of our objects .
instead , we have adopted an average extinction of a@xmath15 mag , as determined from the versus color magnitude diagram in fig .
[ cmd ] . since the amount of extinction ranges between a@xmath28 mag @xcite the extinction for an individual object might be wrong by up to a@xmath29 mag , a maximum error @xmath17 0.2 mag in the band .
this corresponds to an error of @xmath110% when the luminosity is transformed into a mass .
[ massfuncs ] shows the derived mass functions outside 0.6 pc for a 3 myr isochrone after field stars have been subtracted statistically in each annulus .
the mass functions are in general smooth and well fit by power laws . however , there appears to be some structure in the derived imfs at intermediate masses , 24 m@xmath2 , which is the region where the pre
main sequence track joins the main sequence .
the mass luminosity relation is plagued by a non monotonous feature at this mass range ( see fig . [ mlrel ] ) , which marks the radiative - convective gap @xcite and the transition region from pre - main sequence to main sequence ( see also * ? ? ?
a similar structure in the derived imf is seen in the results from e.g. ngc 3603 but at a slightly higher mass since the cluster is younger ( e.g. * ? ? ?
the turn on mass is higher for a younger cluster .
thus we would expect the kink in the mass luminosity relation to move to higher masses for a younger cluster . since this is what is seen comparing ngc 3603 and r 136
, it indicates indeed a feature of the isochrones and not a feature intrinsic to the cluster .
the number of stars in each mass bin is provided in table [ numbers ] .
power laws have been fitted to each of the histograms in order to derive the slopes of the mass function in each annulus .
the fit was done over the mass range from 20 m@xmath2 down to the 50% completeness limit for each annulus .
the mass for stars above @xmath120 m@xmath2 is very poorly constrained from near infrared observations due to uncertainties in the bolometric corrections ( e.g. * ? ? ?
the derived slopes @xmath30 , where @xmath31 , are indicated in fig .
[ massfuncs ] and are also presented in table [ slopes ]
. the derived slopes for annuli outside 1 pc are consistent with each other within 2@xmath16 error bars . for the 35 pc and 57 pc annuli where the data are complete to below 2 m@xmath2
, the slopes are found to be @xmath32 and @xmath33 , respectively , slightly shallower than the slope of @xmath34-@xmath35 derived by @xcite above 2 m@xmath2 , except that in our case the imf continues as a power
law down to 0.8 m@xmath2 . has the fact that we used the whole mass range for our power - law fit washed out a possible flattening at the low mass end ? to test this possibility , we have additionally fitted a separate power
law to the low mass part of the imf . only the part of the mass function that is not influenced by the kink in the mass luminosity relation is used .
this region is limited to masses below 1.7 m@xmath2 for the 3 myr isochrone .
it is therefore only for the 57 pc annulus that a reasonable mass range is covered to fit the imf .
we find the slope to be @xmath36 , which is more shallow but consistent at the 2@xmath16 level with a salpeter imf and is consistent with the slope derived for the full mass range .
we have derived the imf in the same boxes as done by @xcite .
the completeness correction was calculated independently for each box before the imfs were combined to the average imf for direct comparison with the imf presented by @xcite .
the 50% completeness limit for the nicmos data varies from 2.8 to 1.4 m@xmath2 for the four boxes .
following sirianni et al .
, we have derived an average completeness limit for the three regions of 2.2 m@xmath2 . as evident ,
the agreement is good for the common mass range .
we appear to underestimate the stars at @xmath1 6m@xmath2 compared to @xcite .
however those appear to be recovered at 8 m@xmath2 .
the color
magnitude diagrams show a large spread in the main sequence to pre main sequence transition at = 1819 mag .
although shown in section 3.2 that this scatter can be explained by differential extinction , it can not be ruled out that there is an age spread present as well as suggested in previous studies @xcite .
it is therefore reasonable to take a star formation history different than a single burst at 3 myr into account .
we show in fig .
[ imf_spread ] the imf in the outer two annuli assuming a cluster age of 2 and 4 myr , respectively .
we also show an average imf found as the average of the imf s derived for the age range 24 myr in 0.5 myr increments .
the lower mass limit in the average imf was determined from the 4 myr isochrone which provides the most restrictive mass limit .
we find that both in the case of a 2 and 4 myr isochrone the imf is well fit by power
the derived slopes are steeper assuming an older isochrone relative to the younger ones .
there is no indication for a flattening below 2 m@xmath2 in either case .
the average imf is also found to be represented by a power law with a slope consistent with a salpeter slope .
the slopes of the derived power
laws are given in table [ slopes ] .
for the average imf , the number of stars averaged over the different ages in each mass bin is derived .
error bars for the average imf have been determined as the standard deviation around the mean number of objects in each mass bin .
as was the case for the 3 myr isochrone , the slopes of the imf for different assumed ages have also been calculated and are provided in table [ slopes ] .
the slopes are found to be shallower than a salpeter slope , but at the @xmath37 level consistent with a salpeter slope .
the slopes are also consistent with those derived for all masses up to 20 m@xmath2 , as was the case assuming the 3 myr isochrone .
the lack of a flattening in the imf below 2 m@xmath2 is in contrast to the results presented by @xcite , who derived the imf closer to the cluster center .
there can be several possible reasons for the difference in the derived imf slope in the two surveys .
first , due to the different spatial resolution in the two studies , the nicmos imf is derived further away from the center of the cluster than the wfpc2 imf by @xcite .
the imf was derived in the areas shown in fig .
[ 30dormos ] as regions b , c , and d. thus , all of their surveyed area is outside a radius of 1 pc and the majority of their surveyed area is between 2 and 5 pc where crowding precludes nicmos from detecting stars less massive than 2.2 m@xmath2 for a 3 myr isochrone .
one possibility for the difference in the derived slopes for the nicmos and wfpc2 data can therefore be a variation of the imf as a function of radius .
another possible reason can be differential extinction as suggested by @xcite .
both possibilities are discussed in subsections 4.2 and 4.3 as was suggested by @xcite , the presence of differential extinction can potentially alter the low mass end of the imf if an extinction - limited sample is not used . in order to estimate the possible effect on the imf if differential extinction is not taken into account , we have constructed a simple model of the cluster which includes differential extinction and the depth of the dataset from @xcite . a salpeter slope imf and a cluster age of 3 myr were assumed .
each object within the artificial cluster was assigned a band magnitude based on its mass from the 3 myr isochrone computed for a half solar metallicity by @xcite .
the objects were then reddened by a foreground extinction chosen randomly from a normal distribution with a standard deviation of 0.7 and shifted to peak at a@xmath38=1.85 mag . if the extinction was found to be less than a@xmath12 mag , a new extinction was calculated .
stars were then considered detected if their reddened magnitude is within the 50% completeness limit presented by @xcite .
the model is obviously an oversimplification of the real situation .
nevertheless it is expected to illustrate how the derived imf might differ from the underlying imf .
figure [ checksirianni ] shows the input salpeter imf ( solid line ) , the derived imf ( dashed line ) together with the measurements by @xcite .
the model mimics a flattening in the observed imf similar to that deduced by @xcite .
the ratio of the number of stars below and above 2 m@xmath2 respectively has been calculated both for the model cluster and the data from @xcite . for the model cluster
it is found to be 0.87 , which is in reasonable agreement with the ratio of 0.76 derived from the observations .
another explanation for the difference between the results obtained here and the results by @xcite can be mass segregation .
we have searched for evidence for mass segregation in the two outer annuli in our survey .
we used the luminosity functions instead of the mass functions to avoid additional uncertainties due to the mass luminosity relation .
the results obtained for the mass functions are very similar to those from the luminosity functions .
the cumulative luminosity distributions are shown in fig .
[ cumu_lf ] for the outer two radial bins .
these are the only bins where the 50% completeness limit is below 2 m@xmath2 .
the two cumulative distributions are very similar .
we have performed a kolmogorov
smirnov test to quantify the similarity of the cumulative luminosity distributions .
the maximum difference between the two distributions is 0.039 and the probability for the two distributions to be drawn from the same parent distribution is 10% .
thus , there is no strong evidence ( less than 2@xmath16 ) for mass segregation in the outer parts of the cluster .
the fact that there is little evidence for mass segregation outside 3 pc does not exclude the possibility that the cluster is mass segregated out to a radius of several pc .
both @xcite and @xcite found evidence for mass segregation of the massive stars in the center of the cluster .
@xcite showed the half
mass relaxation time to be 7.8@xmath39 yr , much longer than the cluster age .
they also point out that the massive stars will experience mass segregation on a much shorter time scale than the lower mass stars ; the time scale depends inversely on the stellar mass .
it is thus not surprising , from a dynamical point of view , that there is no evidence for mass segregation outside the half
mass radius of 1.7 pc @xcite .
on the other hand , this does not rule out the possibility that the cluster might be mass segregated at birth closer to the cluster center .
evidence for mass segregation has been found in e.g. the orion nebula cluster ( onc ) @xcite .
@xcite showed evidence for mass segregation down to stellar masses of 12 m@xmath2 . due to the youth of the onc , they concluded the mass segregation had to be at least partly primordial .
it is thus possible that r136 is also affected by primordial mass segregation close to the cluster center and that mass segregation is the reason for the difference between the nicmos and wfpc2 imfs .
we can obtain a rough estimate of the cluster mass from the near infrared observations .
the main limitation in our mass estimate is the amount of confusion due to crowding in the cluster centre : our data mainly samples the imf down to and below 1.4 m@xmath2 outside 3 pc . nevertheless , we can utilise the mass estimates within 2 pc from @xcite to complement our mass estimate down to 2.1 m@xmath2 . inside
2 pc and into 0.15 pc , the results are extrapolated from the local completeness limit mass down to 2.1 m@xmath2 assuming an underlying salpeter imf .
no stars have been detected less massive than 20 m@xmath2 within the central 0.15 pc radius due to crowding .
the mass in the very center has been estimated from the surface density profile down to 2.8 m@xmath2 in @xcite to be @xmath40 m@xmath2pc@xmath41 , resulting in a mass of 3700 m@xmath2 down to a lower mass limit of 2.1 m@xmath2 .
we find the cluster total mass down to 2.1 m@xmath2 to be @xmath42 m@xmath2 .
the directly determined mass down to 2.8 m@xmath2 within 4.7 pc is found to be @xmath43 m@xmath2 , almost the same as found by @xcite . if the imf follows a salpeter slope down to 0.5 m@xmath2 as observed in the galactic field and nearby lower
mass clusters @xcite , the total mass in the central region would be roughly double the amount given above , and the total cluster mass would be close to @xmath44m@xmath2 . the velocity dispersion , and hence the dynamical mass , of the whole ngc 2070 region , including r 136 has been determined by @xcite .
the dynamical mass was determined to be 4.5@xmath45m@xmath2 , almost 5 times higher than expected for r 136 alone , but consistent with the photometric mass for the same area @xcite .
if we take into account that the half mass radius of r 136 is 1.7 pc @xcite , compared to 14 pc for the whole ngc 2070 region and assuming the velocity dispersion is the same in the inner parts of the cluster , we would expect a dynamical mass of 4.5@xmath46m@xmath2=5.5@xmath47m@xmath2 which is lower than the mass expected if the imf is consistent with a galactic imf down to 0.5 m@xmath2 .
thus , at face value , the velocity dispersion would be low enough that the cluster can stay bound .
however , a measurement of the velocity dispersion for the inner regions is necessary to directly compare the photometric mass with the dynamical mass .
we can directly derive the surface brightness profile of the region around r136 in the 30 dor cluster since the data does not suffer from saturated stars .
although bright stars will saturate through the one hour exposure , the non - destructive readout mode ensures that only the first reads are used to derive the magnitude of the brightest stars .
the surface brightness profile is shown in fig .
[ lightprof ] . between @xmath10.2 and 2 pc ,
the light profile is well fit by a power
law , whereas inside 0.2 pc the light profile appears to be flattening .
we have therefore fitted the light profile with a power law modified by a core radius , similar to the approach in @xcite . constraining the fit to inside 2 pc
, we find a slope of @xmath48 , slightly more shallow than @xmath49 derived outside 0.1pc by @xcite using planetary camera onboard hst observations .
the core radius is found to be @xmath50pc , which is less than the resolution of the observations and is thus likely and upper limit .
previous hst optical studies determined a small core radius , @xmath51 pc @xcite , consistent with our findings here .
however , since the derived core radius is smaller than the resolution of the observations , it s evidence is weak .
one or two bright stars off center by only a small amount could mimic a cluster core .
how does the low mass end of the imf in 30 dor compare with that determined for other massive and dense stellar clusters ?
a top heavy imf in massive dense clusters has been suggested on theoretical grounds ( e.g. * ? ? ?
the most convincing example of a young cluster with a present - day mass function departing significantly from a salpeter imf above 1 m@xmath2 is the arches cluster @xcite .
@xcite found an average slope of @xmath52 for the central parsec of the arches cluster , flatter than a salpeter slope of -1.35 .
deeper observations found that the present day mass function in arches to be well approximated by a power law with a slope of @xmath53 down to 1.3 m@xmath2 @xcite .
however , recent work taking differential extinction into account suggest the slope of the power law is only slightly more shallow than a salpeter slope , @xmath54 @xcite .
@xcite note that even if the observed imf is slightly flatter thann a salpeter imf , this can be explaned by mass segregation .
the mass segregation would be accelerated in the cluster due to the strong gravitational field from the galactic center . by adopting realistic parameters for a model cluster and an appropriate distance from the galactic center
, they found that an input salpeter slope imf would be transformed to the observed present day mass function via strong dynamical evolution .
@xcite showed that the imf of the cluster powering the ngc 3603 hii region was well fitted by a power law but with a slope flatter than salpeter , @xmath55 .
they further showed evidence for mass segregation for the more massive stars , m@xmath56 m@xmath2 .
the data indicated a slight flattening of the low
mass content ( m@xmath17 3 m@xmath2 ) .
ngc 3603 is younger than the arches cluster and not affected by a strong tidal gravitational field .
thus it is expected to be less influenced by dynamical mass segregation .
the even more massive starburst clusters appear to be the primary sites ( unit cells ) of star formation in starburst galaxies , including interacting / colliding galaxies such as the antennae or the cartwheel . if starburst clusters are the basic building blocks of certain star
forming galaxies , their stellar content ( imf ) will affect much of the observed chemical and photometric evolution of galaxies , both at the present epoch and perhaps even more so in the high - redshift past @xcite .
several observational claims have been made that the imf in unresolved starburst clusters is top
heavy @xcite , although observations of the antennae gave a mixed result @xcite .
however , it has been suggested that the high mass to light ratios found in some young starburst clusters are artificially high related to their not being in virial equilibrium due to gas expulsion from the clusters @xcite . during the first 50 myr of the cluster , the velocity dispersion and
hence the cluster mass might be overestimated if the cluster is assumed to be virialized .
@xcite suggest that the top heavy imfs inferred in young unresolved extragalactic star clusters might be spurious due to their non - virialized dynamical state . with the present dataset it is clear that the imf in the outer parts of r136 continues as a power
law down to 1 m@xmath2 , similar to what is found in other star clusters and the slope is similar to what is found in the field . whether this is true for the cluster as a whole depends on the cause for the flattening observed closer to the cluster center
. it would be interesting to know the imf if the observations could be extended closer to the characteristic mass where the galactic field star imf flattens ( 0.5 m@xmath2 * ? ? ?
* ) , a mass that can be reached in massive young clusters ( @xmath57 4 myr ) in the lmc with ao systems .
it has long been suggested r136 might be a proto globular cluster @xcite .
the question has been whether r136 would remain bound over a hubble time .
one consequence of a top heavy imf is that the cluster would dissolve soon after gas expulsion and mass loss due to evolution of the high mass stars .
however , the detection of stars in r136 less massive than 1 m@xmath2 gives the first _
direct _ evidence that low
mass stars are formed in a starburst cluster .
the fact that the imf in the outer parts of r136 appears to be a salpeter imf down to at least 1 m@xmath2 gives support to the notion the cluster might be a proto globular cluster , albeit a light one .
early gas expulsion and subsequent mass loss through stellar evolution will disrupt star clusters deficient in low
mass stars during the first 5 gyr of the clusters life @xcite however , a determination of the velocity dispersion in the inner parts of the cluster is necessary to determine its final fate .
thus , the presence of low mass stars is a necessary , but not sufficient condition for the possibility of the cluster to evolve into a globular cluster .
the median mass of galactic globular clusters is 8.1@xmath47 m@xmath2 @xcite , comparable to the mass of r136 .
even if r 136 will remain bound it will lose some mass and might end up as a low mass globular cluster .
we have analyzed hst / nicmos band data covering the central [email protected] around r136 in the ngc 2070 cluster in the lmc .
we have reached the following conclusions : * from the color magnitude diagram obtained by combining our photometry with previously published hst / wfpc 2 data we constrain the age of the lower mass stellar content in the cluster to be 24 myr , consistent with previous estimates .
we derive individual masses for the objects detected adopting a 3 myr isochrone .
* we have detected stars in the cluster down to 0.5 m@xmath2 at @xmath58 pc , assuming an age of 3 myr . *
the derived imf is consistent with a salpeter slope imf with no evidence for a flattening at low masses down to the 50% completeness limit corresponding to a mass of 1.1 m@xmath2 outside a radius of 5 pc for a 3 myr population and 1.4 m@xmath2 if the oldest stars are 4 myr . *
the result is in disagreement with the flattening of the imf below 2 m@xmath2 observed by @xcite using optical data covering a region closer to the cluster center .
we suggest two possible reasons for the discrepancy : differential extinction and mass segregation .
* we find no evidence for mass segregation outside 3 pc , but with the current data , we can not rule out that closer to the center the low mass stars are segregated . * from the radial surface brightness profile we have derived a core radius for the cluster of 0.025 pc ( 01 ) , consistent with previous estimates by @xcite . *
the mass of the cluster within 7 pc between 25 m@xmath2 and down to 2.1 m@xmath2 is estimated to be 5@xmath47 m@xmath2 .
if the imf continues with a salpeter slope down to 0.5 m@xmath2 the total mass estimate will double . *
the total mass of the cluster combined with the large number of low mass stars suggests that the 30 dor cluster may survive to become a proto globular cluster depending on the cluster velocity dispersion .
we thank richard larson for discussions in the early phases of the project , eddie bergeron for assistance with the drizzle software , and matthew kenworthy for commenting on an early version of the manuscript .
m. a. and h. z. acknowledges support from the dlr grant 50or9912 : `` data analysis of nicmos / hst images of the 30 dor cluster '' and partial funding through the dlr grant 50or0401 .
m.a thanks the astrophysikalisches institut potsdam for providing a stimulating and supportive environment for carrying out this thesis work .
additional support was funded through the european commission fifth framework programme research training network `` the formation and evolution of young stellar clusters '' ( hprn - ct-2000 - 00155 ) .
the astronomische gesellschaft is acknowledged for providing funding for travel .
support for this work was provided by nasa through grant number go-07370.01 - 96a from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
andersen , m. , meyer , m. r. , greissl , j. , oppenheimer , b. d. , kenworthy , m. a. , mccarthy , d. w. , & zinnecker , h. 2005 , iau symposium , 227 , 285 , cambridge university press , eds .
r. cesaroni , e. churchwell , m. felli , & c.m .
walmsley .
zinnecker , h. , brandl , b. , brandner , w. , moneti , a. , & hunter , d. 1999 , iau symp . 190 : new views of the magellanic clouds , 190 , the astronomical society of the pacific , edited by y .- h . chu , n. suntzeff , j. hesser , & d. bohlender zinnecker , h. , andersen , m. , brandl , b. , brandner , b. , hunter , d. , larson , r. , mccaughrean , m. j. , meylan , g. & moneti , a. 2002 , iau symp . 207 : extragalctic star clusters , 531 , san francisco : astronomical society of the pacific , edited by d. geisler , e.k .
grebel , and d. minniti . | we present deep hubble space telescope ( hst ) nicmos 2 band observations of the central 56@xmath057 ( [email protected] ) region around r136 in the starburst cluster 30 dor ( ngc 2070 ) located in the large magellanic cloud .
our aim is to derive the stellar initial mass function ( imf ) down to @xmath11 m@xmath2 in order to test whether the imf in a massive metal
poor cluster is similar to that observed in nearby young clusters and the field in our galaxy .
we estimate the mean age of the cluster to be 3 myr by combining our photometry with previously obtained hst wfpc2 optical and band photometry and comparing the stellar locus in the color magnitude diagram with main sequence and pre main sequence isochrones . the color
magnitude diagrams show the presence of differential extinction and possibly an age spread of a few myr .
we convert the magnitudes into masses adopting both a single mean age of 3 myr isochrone and a constant star formation history from 2 to 4 myr .
we derive the imf after correcting for incompleteness due to crowding .
the faintest stars detected have a mass of 0.5 m@xmath2 and the data are more than 50% complete outside a radius of 5 pc down to a mass limit of 1.1 m@xmath2 for 3 myr old objects .
we find an imf of @xmath3 over the mass range 1.120 m@xmath2 only slightly shallower than a salpeter imf .
in particular , we find no strong evidence for a flattening of the imf down to 1.1 m@xmath2 at a distance of 5 pc from the center , in contrast to a flattening at 2 m@xmath2 at a radius of 2 pc , reported in a previous optical hst study .
flattening at 2 m@xmath2 at a radius of 2 pc previously found .
we examine several possible reasons for the different results including the possible presence of mass segregation and the effects of differential extinction , particularly for the pre
main sequence sources .
if the imf determined here applies to the whole cluster , the cluster would be massive enough to remain bound and evolve into a relatively low mass globular cluster . |
ever since it was introduced in @xcite , the notion of a nash equilibrium and its refinements have remained among the most prominent solution concepts of noncooperative game theory . in its turn , not only has noncooperative game theory found applications in such diverse topics as economics , biology and network design , but it has also become the standard language to actually _ describe _ complex agent interactions in these fields . still , the issue of why and how players may arrive to equilibrial strategies in the first place remains an actively debated question .
after all , the complexity of most games increases exponentially with the number of players and , hence , identifying a game s equilibria quickly becomes prohibitively difficult .
accordingly , as was first pointed out by aumann in @xcite , a player has no incentive to play his component of a nash equilibrium unless he is convinced that all other players will play theirs . and
if the game in question has multiple nash equilibria , this argument gains additional momentum : in that case , even players with unbounded deductive capabilities will be hard - pressed to choose a strategy . from this point of view
, rational individuals would appear to be more in tune with aumann s notion of a correlated equilibrium where subjective beliefs are also taken into account @xcite .
nevertheless , the seminal work of maynard smith on animal conflicts @xcite has cast nash equilibria in a different light because it unearthed a profound connection between evolution and rationality : roughly speaking , one leads to the other .
so , when different species contend for the limited resources of their habitat , evolution and natural selection steer the ensuing conflict to an equilibrial state which leaves no room for irrational behavior . as a consequence ,
instinctive `` fight or flight '' responses that are deeply ingrained in a species can be seen as a form of rational behavior , acquired over the species evolutionary course .
of course , this evolutionary approach concerns large populations of different species which are rarely encountered outside the realm of population biology .
however , the situation is not much different in the case of a finite number of players who try to learn the game by playing again and again and who strive to do better with the help of some learning algorithm .
therein , evolution does not occur as part of a birth / death process ; rather , it is a byproduct of the players acquired experience in playing the game see @xcite for a most comprehensive account .
it is also worth keeping in the back of our mind that in some applications of game theory , `` rationality '' requirements precede evolution .
for example , recent applications to network design start from a set of performance aspirations ( such as robustness and efficiency ) that the players ( network devices ) seek to attain in the network s equilibrial state .
thus , to meet these requirements , one has to literally reverse - engineer the process by finding the appropriate game whose equilibria will satisfy the players the parallel with mechanism design being obvious . in all these approaches ,
a fundamental selection mechanism is that of the _ replicator dynamics _ put forth in @xcite and @xcite which reinforces a strategy proportionately to the difference of its payoff from the mean ( taken over the species or the player s strategies , depending on the approach ) . as was shown in the multi - population setting of samuelson and zhang @xcite ( which is closer to learning than the self - interacting single - population scenaria of @xcite and @xcite ) , these dynamics are particularly conducive to rationality .
strategies that are suboptimal when paired against any choice of one s adversaries rapidly become extinct , and in the long run , only rationally admissible strategies can survive .
even more to the point , the only attracting states of the dynamics turn out to be precisely the ( strict ) nash equilibria of the game see @xcite for a masterful survey .
we thus see that nash equilibria arise over time as natural attractors for rational individuals , a fact which further justifies their prominence among noncooperative solution concepts . yet , this behavior is also conditional on the underlying game remaining stationary throughout the time horizon that it takes players to adapt to it and unfortunately , this stationarity assumption is rarely met in practical applications . in biological models , for example , the reproductive fitness of an individual may be affected by the ever - changing weather conditions ; in networks , communication channels carry time - dependent noise and interference as well as signals ; and when players try to sample their strategies , they might have to deal with erroneous or imprecise readings .
it is thus logical to ask : _ does rational behavior still emerge in the presence of stochastic perturbations that interfere with the underlying game _ ? in evolutionary games ,
these perturbations traditionally take the form of `` aggregate shocks '' that are applied directly to the population of each phenotype .
this approach by fudenberg and harris @xcite has spurred quite a bit of interest and there is a number of features that differentiate it from the deterministic one .
for example , cabrales showed in @xcite that dominated strategies indeed become extinct , but only if the variance of the shocks is low enough .
more recently , the work of imhof and hofbauer @xcite revealed that even equilibrial play arises over time but again , conditionally on the variance of the shocks .
be that as it may , if one looks at games with a finite number of players , it is hardly relevant to consider shocks of this type because there are no longer any populations to apply them to .
instead , the stochastic fluctuations should be reflected directly on the stimuli that incite players to change their strategies : their payoffs .
this leads to a picture which is very different from the evolutionary one and is precisely the approach that we will be taking . in this paper
, we analyze the evolution of players in stochastically perturbed games of this sort .
the particular stimulus - response model that we consider is simple enough : players keep cumulative scores of their strategies performance and employ exponentially more often the one that scores better .
after a few preliminaries in section [ sec : preliminaries ] , this approach is made precise in section [ sec : replicator ] where we derive the stochastic replicator equation that governs the behavior of players when their learning curves are subject to random perturbations .
the replicator equation that we get is different from the `` aggregate shocks '' approach of @xcite and , as a result , it exhibits markedly different rationality properties as well . in stark contrast to the results of @xcite , we show in section [ sec : dominated ] that dominated strategies become extinct irrespective of the noise level ( proposition [ prop : dominated ] ) and provide an exponential bound for the rate of decay of these strategies ( proposition [ prop : timetolive ] ) .
in fact , by induction on the rounds of elimination of dominated strategies , we show that this is true even for _ iteratively _ dominated strategies : despite the noise , only rationally admissible strategies can survive in the long run ( theorem [ thm : rational ] ) .
then , as an easy corollary of the above , we infer that players will converge to a strict equilibrium ( corollary [ cor : dominance ] ) whenever the underlying game is dominance - solvable .
we continue with the issue of equilibrial play in section [ sec : congestion ] by making a suggestive detour in the land of congestion games . if the noise is relatively mild with respect to the rate with which players learn , we find that the game s potential is a lyapunov function which ensures that strict equilibria are stochastically attracting ; and if the game is dyadic ( i.e. , players only have two choices ) , this tameness assumption can be dropped altogether . encouraged by the results of section [ sec : congestion ] , we attack the general case in section [ sec : equilibrium ] .
as it turns out , strict equilibria are _ always _ asymptotically stochastically stable in the perturbed replicator dynamics that stem from exponential learning ( theorem [ thm : stability ] ) .
this begs to be compared to the results of @xcite where it is the equilibria of a suitably modified game that are stable , and not necessarily those of the actual game being played .
fortunately , exponential learning seems to give players a clearer picture of the original game and there is no need for similar modifications in our case . given a finite set @xmath0 , we will routinely identify the set @xmath1 of probability measures on @xmath2 with the standard @xmath3-dimensional simplex of @xmath4 and @xmath5 . under this identification
, we will also make no distinction between @xmath6 and the vertex @xmath7 of @xmath1 ; in fact , to avoid an overcluttering of indices , we will frequently use @xmath8 to refer to either @xmath9 or @xmath7 , writing , for example , `` @xmath10 '' or `` @xmath11 '' instead of `` @xmath6 '' or `` @xmath12 , '' respectively . to streamline our presentation
, we will consistently employ latin indices for players ( @xmath13 ) and greek for their strategies ( @xmath14 ) , separating the two by a comma when it would have been sthetically unpleasant not to . in like manner , when we have to discriminate between strategies
, we will assume that indices from the first half of the greek alphabet start at @xmath15 ( @xmath16 ) while those taken from the second half start at @xmath17 ( @xmath18 ) .
finally , if @xmath19 is some stochastic process in @xmath20 starting at @xmath21 , its law will be denoted by @xmath22 or simply by @xmath23 if there is no danger of confusion ; and if the context leaves no doubt as to which process we are referring to , we will employ the term `` almost surely '' in place of the somewhat unwieldy `` @xmath23-almost surely . ''
as is customary , our starting point will be a ( finite ) set of @xmath24 _ players _ , indexed by @xmath25 .
the players possible actions are drawn from their _ strategy sets _ @xmath26 and they can combine them by choosing their @xmath27th ( pure ) strategy with probability @xmath28 . in that case , the players _ mixed strategies _ will be described by the points @xmath29 or , more succinctly , by the _ strategy profile _ @xmath30 . in particular , if @xmath31 denotes the @xmath8th vertex of the @xmath32th component simplex , the ( pure ) profile @xmath33 simply corresponds to player @xmath32 playing @xmath34 . on the other hand ,
if we wish to focus on the strategy of a particular player @xmath35 against that of his _ opponents _ @xmath36 , we will employ the shorthand notation @xmath37 to denote the profile where @xmath32 plays @xmath38 against his opponents strategy @xmath39 .
so , once players have made their strategic choices , let @xmath40 be the reward of player @xmath32 in the profile @xmath41 , that is , the payoff that strategy @xmath34 yields to player @xmath32 against the strategy @xmath42 of @xmath32 s opponents . then
, if players mix their strategies , their expected reward will be given by the ( multilinear ) _ payoff functions _
@xmath43 : @xmath44 under this light , the payoff that a player receives when playing a pure strategy @xmath45 deserves special mention and will be denoted by @xmath46 this collection of _ players
_ @xmath35 , their _ strategies _ @xmath34 and their _ payoffs _ @xmath47 will be our working definition for a _ game in normal form _ , usually denoted by @xmath48or @xmath49 if we need to keep track of more data .
needless to say , rational players who seek to maximize their individual payoffs will avoid strategies that always lead to diminished payoffs against any play of their opponents .
we will thus say that the strategy @xmath38 is ( _ strictly _ ) _ dominated _ by @xmath50 and we will write @xmath51 when @xmath52 for all strategies @xmath53 of @xmath32 s opponents @xmath54 . with this in mind , dominated strategies can be effectively removed from the analysis of a game because rational players will have no incentive to ever use them .
however , by deleting such a strategy , another strategy ( perhaps of another player ) might become dominated and further deletions of _ iteratively dominated _ strategies might be in order ( see section [ sec : dominated ] for more details ) .
proceeding ad infinitum , we will say that a strategy is _ rationally admissible _ if it survives every round of elimination of dominated strategies .
if the set of rationally admissible strategies is a singleton ( e.g. , as in the prisoner s dilemma ) , the game will be called _ dominance - solvable _ and the sole surviving strategy will be the game s _ rational solution_. then again , not all games can be solved in this way and it is natural to look for strategies which are stable at least under unilateral deviations .
hence , we will say that a strategy profile @xmath55 is a _
nash equilibrium _ of the game @xmath48 when @xmath56 if the equilibrium profile @xmath57 only contains pure strategies @xmath58 , we will refer to it as a _ pure equilibrium _ ; and if the inequality ( [ eq : nash ] ) is strict for all @xmath59 , the equilibrium @xmath57 will carry instead the characterization _
strict_. clearly , if two pure strategies @xmath60 are present with positive probability in an equilibrial strategy @xmath61 , then we must have @xmath62 as a result of @xmath47 being linear in @xmath63 .
consequently , only pure profiles can satisfy the strict version of ( [ eq : nash ] ) so that strict equilibria must also be pure .
the converse implication is false but only barely so : a pure equilibrium fails to be strict only if a player has more than one pure strategies that return the same rewards . since this is almost always true ( in the sense that the degenerate case can be resolved by an arbitrarily small perturbation of the payoff functions )
, we will relax our terminology somewhat and use the two terms interchangeably . to recover the connection of equilibrial play with strategic dominance , note that if a game is solvable by iterated elimination of dominated strategies , the single rationally admissible strategy that survives will be the game s unique strict equilibrium .
but the significance of strict equilibria is not exhausted here : strict equilibria are exactly the evolutionarily stable strategies of multi - population evolutionary games proposition 5.1 in @xcite . moreover , as we shall see a bit later , they are the only asymptotically stable states of the multi - population replicator dynamics
again , see chapter 5 , pages 216 and 217 of @xcite . unfortunately , strict equilibria do not always exist , rock - paper - scissors being the typical counterexample .
nevertheless , pure equilibria do exist in many large and interesting classes of games , even when we leave out dominance - solvable ones .
perhaps the most noteworthy such class is that of _ congestion games_.
[ def : congestion ] a game @xmath64 will be called a _ congestion game _ when : 1 .
all players @xmath35 share a common set of _ facilities _ @xmath65 as their strategy set : @xmath66 for all @xmath35 ; 2 . the payoffs are functions of the number of players sharing a particular facility : @xmath67 where @xmath68 is the number of players choosing the same facility as @xmath32 . amazingly enough
, monderer and shapley made the remarkable discovery in @xcite that these games are actually equivalent to the class of _ potential games_. [ def : potential ] a game @xmath64 will be called a _ potential game _ if there exists a function @xmath69 such that @xmath70 for all players @xmath35 and all strategies @xmath71 , @xmath72 .
this equivalence reveals that both classes of games possess equilibria in pure strategies : it suffices to look at the vertices of the face of @xmath73 where the ( necessarily multilinear ) potential function @xmath74 is minimized . as one would expect , locating the nash equilibria of a game is a rather complicated problem that requires a great deal of global calculations , even in the case of potential games ( where it reduces to minimizing a multilinear function over a convex polytope ) .
consequently , it is of interest to see whether there are simple and distributed learning schemes that allow players to arrive at a reasonably stable solution .
one such scheme is based on an exponential learning behavior where players play the game repeatedly and keep records of their strategies performance . in more detail , at each instance of the game all players @xmath35 update the cumulative scores @xmath75 of their strategies @xmath45 as specified by the recursive formula @xmath76 where @xmath77 is the players strategy profile at the @xmath78th iteration of the game and , in the absence of initial bias , we assume that @xmath79 for all @xmath80 .
these scores reinforce the perceived success of each strategy as measured by the average payoff it yields and hence , it stands to reason that players will lean towards the strategy with the highest score .
the precise way in which they do that is by playing according to the namesake exponential law : @xmath81 for simplicity , we will only consider the case where players update their scores in continuous time , that is , according to the coupled equations @xmath82 then , if we differentiate ( [ eq : detlogit ] ) to decouple it from ( [ eq : detscore ] ) , we obtain the _ standard _ ( _ multi - population _ ) _ replicator dynamics _
@xmath83 alternatively ,
if players learn at different speeds as a result of varied stimulus - response characteristics , their updating will take the form @xmath84 where @xmath85 represents the _ learning rate _ of player @xmath32 , that is , the `` weight '' which he assigns to his perceived scores @xmath75 . in this way , the replicator equation evolves at a different time scale for each player , leading to the _ rate - adjusted _ dynamics @xmath86 naturally , the uniform dynamics ( [ eq : rd ] ) are recovered when all players learn at the `` standard '' rate @xmath87 .
if we view the exponential learning model ( [ eq : discretelogit ] ) from a stimulus - response angle , we see that that the payoff of a strategy simply represents an ( exponential ) propensity of employing said strategy .
it is thus closely related to the algorithm of _ logistic fictitious play _
@xcite where the strategy @xmath88 of ( [ eq : ratelogit ] ) can be seen as the ( unique ) best reply to the profile @xmath89 in some suitably modified payoffs @xmath90 .
interestingly enough , @xmath91 turns out to be none other than the _ entropy _ of @xmath88 : @xmath92 that being so , we deduce that the learning rates @xmath85 act the part of ( player - specific ) inverse temperatures : in high temperatures ( small @xmath85 ) , the players learning curves are `` soft '' and the payoff differences between strategies are toned down ; on the contrary , if @xmath93 the scheme `` freezes '' to a myopic best - reply process .
the replicator dynamics were first derived in @xcite in the context of population biology , first for different phenotypes within a single species ( single - population models ) , and then for different species altogether ( multi - population models ; @xcite and @xcite provide excellent surveys ) . in both these cases , one begins with large populations of individuals that are programmed to a particular behavior ( e.g. , for `` hawks '' or for `` doves '' ) and matches them randomly in a game whose payoffs directly affect the reproductive fitness of the individual players .
more precisely , let @xmath94 be the population size of the phenotype ( strategy ) @xmath45 of species ( player ) @xmath95 in some multi - population model where individuals are matched to play a game @xmath48 with payoff functions @xmath47 .
then , the relative frequency ( share ) of @xmath8 will be specified by the _ population state _
@xmath96 where @xmath97 .
so , if @xmath24 individuals are drawn randomly from the @xmath24 species , their expected payoffs will be given by @xmath98 , @xmath35 , and if these payoffs represent a proportionate increase in the phenotype s fitness ( measured as the number of offsprings in the unit of time ) , we will have @xmath99 as a result , the population state @xmath100 will evolve according to @xmath101 which is exactly ( [ eq : rd ] ) viewed from an evolutionary perspective .
on the other hand , we should note here that in _ single - population _ models the resulting equation is cubic and not quadratic because strategies are matched against themselves . to wit , assume that individuals are randomly drawn from a large population and are matched against one another in a ( symmetric ) 2-player game @xmath48 with strategy space @xmath102 and payoff matrix @xmath103 . then
, if @xmath104 denotes the population share of individuals that are programmed to the strategy @xmath105 , their expected payoff in a random match will be given by @xmath106 ; similarly , the population average payoff will be @xmath107 .
hence , by following the same procedure as above , we end up with the single - population replicator dynamics @xmath108 which behave quite differently than their multi - population counterpart ( [ eq : erd ] ) .
as far as rational behavior is concerned , the replicator dynamics have some far - reaching ramifications .
if we focus on multi - population models , samuelson and zhang showed in @xcite that the share @xmath109 of a strategy @xmath45 which is strictly dominated ( even iteratively ) converges to zero along any interior solution path of ( [ eq : rd ] ) ; in other words , _ dominated strategies become extinct in the long run_. additionally , there is a remarkable equivalence between the game s nash equilibria and the stationary points of the replicator dynamics : _ the asymptotically stable states of coincide precisely with the strict nash equilibria of the underlying game _
@xcite .
a large part of our work will be focused on examining whether the rationality properties of exponential learning ( elimination of dominated strategies and asymptotic stability of strict equilibria ) remain true in a stochastic setting . however , since asymptotic stability is ( usually ) too stringent an expectation for stochastic dynamical systems , we must instead consider its stochastic analogue .
that being the case , let @xmath110 be a standard wiener process in @xmath20 and consider the stochastic differential equation ( sde ) @xmath111 following @xcite , the notion of asymptotic stability in this sde is expressed by the following .
[ def : stability ] we will say that @xmath112 is _ stochastically asymptotically stable _ when , for every neighborhood @xmath113 of @xmath114 and every @xmath115 , there exists a neighborhood @xmath74 of @xmath114 such that @xmath116 for all initial conditions @xmath117 of the sde ( [ eq : sde ] ) .
much the same as in the deterministic case , stochastic asymptotic stability is often established by means of a lyapunov function . in our context , this notion hinges on the second order differential operator that is associated to ( [ eq : sde ] ) , namely the _ generator _
@xmath118 of @xmath19 : @xmath119 the importance of this operator can be easily surmised from it s lemma ; indeed , if @xmath120 is sufficiently smooth , the generator @xmath118 simply captures the drift of the process @xmath121 : @xmath122 in this way , @xmath118 can be seen as the stochastic version of the time derivative @xmath123 ; this analogy then leads to the following .
[ def : lyapunov ] let @xmath112 and let @xmath113 be an open neighborhood of @xmath114 .
we will say that @xmath124 is a ( local ) _ stochastic lyapunov function _ for the sde ( [ eq : sde ] ) if : 1 .
@xmath125 for all @xmath126 , with equality iff @xmath127 ; 2 .
there exists a constant @xmath128 such that @xmath129 for all @xmath126 .
whenever such a lyapunov function exists , it is known that the point @xmath112 where @xmath124 attains its minimum will be stochastically asymptotically stable for example , see theorem 4 in pages 314 and 315 of @xcite .
a final point that should be mentioned here is that our analysis will be constrained on the compact polytope @xmath130 instead of all of @xmath131 .
accordingly , the `` neighborhoods '' of definitions [ def : stability ] and [ def : lyapunov ] should be taken to mean `` neighborhoods in @xmath73 , '' that is , neighborhoods in the subspace topology of @xmath132 .
this minor point should always be clear from the context and will only be raised in cases of ambiguity .
of course , it could be argued that the rationality properties of the exponential learning scheme are a direct consequence of the players receiving accurate information about the game when they update their scores .
however , this is a requirement that can not always be met : the interference of nature in the game or imperfect readings of one s utility invariably introduce fluctuations in ( [ eq : detscore ] ) , and in their turn , these lead to a perturbed version of the replicator dynamics ( [ eq : rd ] ) . to account for these random perturbations
, we will assume that the players scores are now governed instead by the _
differential equation @xmath133 where , as before , the strategy profile @xmath134 is given by the logistic law @xmath135 in this last equation , @xmath136 is a standard wiener process living in @xmath131 and the coefficients @xmath137 measure the impact of the noise on the players scoring systems . of course
, these coefficients need not be constant : after all , the effect of the noise on the payoffs might depend on the state of the game in some typically continuous way .
for this reason , we will assume that the functions @xmath138 are continuous on @xmath73 , and we will only note en passant that our results still hold for essentially bounded coefficients @xmath137 ( we will only need to replace @xmath139 and @xmath140 with @xmath141 and @xmath142 , respectively , in all expressions involving @xmath137 ) . a very important instance of this dependence can be seen if @xmath143 for all @xmath144 , in which case equation ( [ eq : score ] ) becomes a convincing model for the case of insufficient information .
it states that when a player actually uses a strategy , his payoff observations are accurate enough ; but with regards to strategies he rarely employs , his readings could be arbitrarily off the mark .
now , to decouple ( [ eq : score ] ) and ( [ eq : logit ] ) , we may simply apply it s lemma to the process @xmath19 . to that end ,
recall that @xmath136 has independent components across players and strategies , so that @xmath145 ( the kronecker symbols @xmath146 being @xmath15 for @xmath147 and @xmath17 , otherwise )
. then , it s formula gives @xmath148\\[-8pt ] & = & \sum_{\beta } \biggl(u_{i\beta}(x)\,\frac{\partial x_{i\alpha}}{\partial u_{i\beta } } + \frac{1}{2}\eta_{i\beta}^{2}(x)\,\frac{\partial^{2}x_{i\alpha } } { \partial u_{i\beta}^{2 } } \biggr ) \,dt\nonumber\\ & & { } + \sum_{\beta } \eta_{i\beta}(x)\frac{\partial x_{i\alpha } } { \partial u_{i\beta } } \,dw_{i\beta}.\nonumber\end{aligned}\ ] ] on the other hand , a simple differentiation of ( [ eq : logit ] ) yields @xmath149 and by plugging these expressions back into ( [ eq : dxprelim ] ) , we get @xmath150 \,dt\nonumber\\ & & { } + x_{i\alpha } \biggl[\frac{1}{2}\eta_{i\alpha}^{2}(x)(1 - 2 x_{i\alpha } ) - \frac{1}{2}\sum_{\beta}\eta_{i\beta}^{2}(x ) x_{i\beta } ( 1 - 2x_{i\beta } ) \biggr ] \,dt\\ & & { } + x_{i\alpha } \biggl[\eta_{i\alpha}(x ) \,dw_{i\alpha } - \sum_{\beta } \eta _ { i\beta}(x ) x_{i\beta } \,dw_{i\beta } \biggr]\nonumber.\end{aligned}\ ] ] alternatively , if players update their strategies with different learning rates @xmath85 , we should instead apply it s formula to ( [ eq : ratelogit ] ) .
in so doing , we obtain @xmath151 \,dt\nonumber\\ & & { } + \frac{\lambda_{i}^{2}}{2}x_{i\alpha } \biggl[\eta_{i\alpha}^{2}(x)(1 - 2 x_{i\alpha } ) -\sum_{\beta}\eta_{i\beta}^{2}(x ) x_{i\beta } ( 1 - 2x_{i\beta } ) \biggr ] \,dt\nonumber\\[-8pt]\\[-8pt ] & & { } + \lambda_{i } x_{i\alpha } \bigl[\eta_{i\alpha}(x)\ , d w_{i\alpha } - \sum\eta_{i\beta}(x ) x_{i\beta } \,dw_{i\beta } \bigr]\nonumber\\ & = & b_{i\alpha}(x ) \,dt + \sum_{\beta } \sigma_{i,\alpha\beta } ( x ) \,d w_{i\beta},\nonumber\end{aligned}\ ] ] where , in obvious notation , @xmath152 and @xmath153 are , respectively , the drift and diffusion coefficients of the diffusion @xmath19 .
obviously , when @xmath87 , we recover the uniform dynamics ( [ eq : srd ] ) ; equivalently ( and this is an interpretation that is well worth keeping in mind ) , the rates @xmath154 can simply be regarded as a commensurate inflation of the payoffs and noise coefficients of player @xmath35 in the uniform logistic model ( [ eq : logit ] ) .
equation ( [ eq : srd ] ) and its rate - adjusted sibling ( [ eq : slrd ] ) will constitute our stochastic version of the replicator dynamics and thus merit some discussion in and by themselves . first , note that these dynamics admit a ( unique ) strong solution for any initial state @xmath155 , even though they do not satisfy the linear growth condition @xmath156 that is required for the existence and uniqueness theorem for sdes ( e.g. , theorem 5.2.1 in @xcite ) .
instead , an addition over @xmath45 reveals that every simplex @xmath157 remains invariant under ( [ eq : srd ] ) : if @xmath158 , then @xmath159 and hence , @xmath160 will stay in @xmath161 for all @xmath162actually , it is not harder to see that every face of @xmath73 is a trap for @xmath19 .
so , if @xmath163 is a smooth bump function that is equal to @xmath17 on some open neighborhood of @xmath164 and which vanishes outside some compact set @xmath165 , the sde @xmath166 _ will _ have bounded diffusion and drift coefficients and will thus admit a unique strong solution .
but since this last equation agrees with ( [ eq : srd ] ) on @xmath73 and any solution of ( [ eq : srd ] ) always stays in @xmath73 , we can easily conclude that our perturbed replicator dynamics admit a unique strong solution for any initial @xmath167 .
it is also important to compare the dynamics ( [ eq : srd ] ) , ( [ eq : slrd ] ) to the `` aggregate shocks '' approach of fudenberg and harris @xcite that has become the principal incarnation of the replicator dynamics in a stochastic environment .
so , let us first recall how aggregate shocks enter the replicator dynamics in the first place .
the main idea is that the reproductive fitness of an individual is not only affected by deterministic factors but is also subject to stochastic shocks due to the `` weather '' and the interference of nature with the game .
more precisely , if @xmath168 denotes the population size of phenotype @xmath45 of the species @xmath95 in some multi - population evolutionary game @xmath48 , its growth will be determined by @xmath169 where , as in ( [ eq : offspring ] ) , @xmath134 denotes the population shares @xmath170 . in this way , it s lemma yields the _ replicator dynamics with aggregate shocks _ : @xmath171 \,dt\nonumber\\[-8pt]\\[-8pt ] & & { } + x_{i\alpha } \bigl[\eta_{i\alpha } \,dw_{i\alpha } - \sum\eta _ { i\beta } x_{i\beta } \,dw_{i\beta } \bigr].\nonumber\end{aligned}\ ] ] we thus see that the effects of noise propagate differently in the case of exponential learning and in the case of evolution . indeed , if we compare equations ( [ eq : srd ] ) and ( [ eq : asrd ] ) term by term , we see that the drifts are not quite the same : even though the payoff adjustment @xmath172 ties both equations back together in the deterministic setting ( @xmath173 ) , the two expressions differ by @xmath174 \,dt.\ ] ] innocuous as this term might seem , it is actually crucial for the rationality properties of exponential learning in games with randomly perturbed payoffs .
as we shall see in the next sections , it leads to some miraculous cancellations that allow rationality to emerge in all noise levels .
this difference further suggests that we can pass from ( [ eq : srd ] ) to ( [ eq : asrd ] ) simply by modifying the game s payoffs to @xmath175 .
of course , this presumes that the noise coefficients @xmath137 be constant
the general case would require us to allow for games whose payoffs may not be multilinear . this apparent lack of generality does not really change things but we prefer to keep things simple and for the time being , it suffices to point out that this modified game was precisely the one that came up in the analysis of @xcite . as a result
, this modification appears to play a pivotal role in setting apart learning and evolution in a stochastic setting : whereas the modified game is deeply ingrained in the process of natural selection , exponential learning seems to give players a clearer picture of the actual underlying game .
thereby armed with the stochastic replicator equations ( [ eq : srd ] ) , ( [ eq : slrd ] ) to model exponential learning in noisy environments , the logical next step is to see if the rationality properties of the deterministic dynamics carry over to this stochastic setting . in this direction , we will first show that dominated strategies always become extinct in the long run and that only the rationally admissible ones survive . as in @xcite ( implicitly ) and @xcite ( explicitly ) , the key ingredient of our approach will be the _ cross entropy _ between two mixed strategies @xmath176 of player @xmath35 : @xmath177 where @xmath178 is the _ entropy _ of @xmath179 and @xmath180 is the intimately related _ kullback leibler divergence _ ( or _ relative entropy _ )
: @xmath181 this divergence function is central in the stability analysis of the ( deterministic ) replicator dynamics because it serves as a distance measure in probability space @xcite .
as it stands however , @xmath180 is not a distance function per se : neither is it symmetric , nor does it satisfy the triangle inequality .
still , it has the very useful property that @xmath182 iff @xmath88 employs with positive probability all pure strategies @xmath45 that are present in @xmath179 [ i.e. , iff @xmath183 or iff @xmath179 is absolutely continuous w.r.t .
@xmath88 ] . therefore ,
if @xmath185 for all dominated strategies @xmath179 of player @xmath32 , it immediately follows that @xmath88 can not be dominated itself . in this vein , we have the following . [ prop : dominated ] let @xmath19 be a solution of the stochastic replicator dynamics ( [ eq : srd ] ) for some interior initial condition @xmath186 .
then , if @xmath38 is ( strictly ) dominated , @xmath187 in particular , if @xmath188 is pure , we will have @xmath189 ( a.s . ) : strictly dominated strategies do not survive in the long run .
note first that @xmath190 and hence , @xmath160 will almost surely stay in @xmath191 for all @xmath192 ; this is a simple consequence of the uniqueness of strong solutions and the invariance of the faces of @xmath193 under the dynamics ( [ eq : srd ] ) .
let us now consider the cross entropy @xmath194 between @xmath179 and @xmath160 : @xmath195 as a result of @xmath160 being an interior path , @xmath194 will remain finite for all @xmath162 ( a.s . ) .
so , by applying it s lemma we get @xmath196\\[-8pt ] & = & -\sum_{\beta } \frac{q_{i\beta}}{x_{i\beta } } \,dx_{i\beta } + \frac{1}{2 } \sum_{\beta } \frac{q_{i\beta}}{x_{i\beta}^{2 } } ( d x_{i\beta } ) ^{2}\nonumber\end{aligned}\ ] ] and , after substituting @xmath197 from the dynamics ( [ eq : srd ] ) , this last equation becomes @xmath198 \,dt\nonumber\\[-8pt]\\[-8pt ] & & { } + \sum_{\beta } q_{i\beta}\sum_{\gamma } ( x_{i\gamma } - \delta _ { \beta \gamma } ) \eta_{i\gamma } ( x)\,dw_{i\gamma}.\nonumber\end{aligned}\ ] ] accordingly , if @xmath199 is another mixed strategy of player @xmath32 , we readily obtain @xmath200\\[-8pt ] & & { } + \sum_{\beta } ( q_{i\beta } ' - q_{i\beta } ) \eta_{i\beta}(x ) \,dw_{i\beta}\nonumber\end{aligned}\ ] ] and , after integrating , @xmath201\\[-8pt ] & & { } + \sum_{\beta}(q_{i\beta}'-q_{i\beta})\int_{0}^{t}\eta_{i\beta } ( x(s ) ) \,dw_{i\beta}(s).\nonumber\end{aligned}\ ] ] suppose then that @xmath202 and let @xmath203 . with @xmath204 compact
, it easily follows that @xmath205 and the first term of ( [ eq : gint ] ) will be bounded from below by @xmath206 .
however , since monotonicity fails for it integrals , the second term must be handled with more care . to that end , let @xmath207 and note that the cauchy schwarz inequality gives @xmath208\\[-8pt ] & \leq & s_{i}\eta_{i}^{2 } \sum_{\beta } ( q'_{i\beta } -q_{i\beta})^{2}\leq 2 s_{i } \eta_{i}^{2},\nonumber\end{aligned}\ ] ] where @xmath209 is the number of pure strategies available to player @xmath32 and @xmath210 ; recall also that @xmath211 for the last step .
therefore , if @xmath212 denotes the martingale part of ( [ eq : gcomp ] ) and @xmath213 is its quadratic variation , the previous inequality yields @xmath214(t ) = \int_{0}^{t}\xi _ { i}^{2}(s)\,ds \leq2 s_{i } \eta_{i}^{2 } t.\ ] ] now , if @xmath215 , it follows from the time - change theorem for martingales ( e.g. , theorem 3.4.6 in @xcite ) that there exists a wiener process @xmath216 such that @xmath217 . hence , by the law of the iterated logarithm we get @xmath218 on the other hand , if @xmath219 , it is trivial to obtain @xmath220 by letting @xmath221 in ( [ eq : gint ] ) .
therefore , with @xmath222 , we readily get @xmath223 ( a.s . ) ; and since @xmath224 for all pure strategies @xmath45 , our proof is complete . as in @xcite , we can now obtain the following estimate for the lifespan of pure dominated strategies .
[ prop : timetolive ] let @xmath19 be a solution path of ( [ eq : srd ] ) with initial condition @xmath225 and let @xmath23 denote its law . assume further that the strategy @xmath45 is dominated ; then , for any @xmath226 and for @xmath78 large enough , we have @xmath227 where @xmath228 is the number of strategies available to player @xmath32 , @xmath229 and the constants @xmath205 and @xmath230 do not depend on @xmath78 .
the proof is pretty straightforward and for the most part follows @xcite .
surely enough , if @xmath231 and we use the same notation as in the proof of proposition [ prop : dominated ] , we have @xmath232 where @xmath233 and @xmath234
. then @xmath235\\[-8pt ] & = & \frac{1}{2}\operatorname{erfc}\biggl(\frac{m - h_{i}(x_{i } ) - v_{i}t}{\sqrt{2\rho _ { i}(t ) } } \biggr)\nonumber\end{aligned}\ ] ] and , since the quadratic variation @xmath213 is bounded above by @xmath236 ( [ eq : quvar ] ) , the estimate ( [ eq : timetolive ] ) holds for all sufficiently large @xmath78 [ i.e. , such that @xmath237 .
some remarks are now in order : first and foremost , our results should be contrasted to those of cabrales @xcite and imhof @xcite where dominated strategies die out only if the noise coefficients ( shocks ) @xmath137 satisfy certain tameness conditions . the origin of this notable difference is the form of the replicator equation ( [ eq : srd ] ) and , in particular , the extra terms that are propagated there by exponential learning and which are absent from the aggregate shocks dynamics ( [ eq : asrd ] ) . as can be seen from the derivations in proposition [ prop : dominated ] , these terms are precisely the ones that allow players to pick up on the true payoffs @xmath238 instead of the modified ones @xmath239 that come up in @xcite ( and , indirectly , in @xcite as well ) .
secondly , it turns out that the way that the noise coefficients @xmath240 depend on the profile @xmath241 is not really crucial : as long as @xmath242 is continuous ( or essentially bounded ) , our arguments are not affected . the only way in which a specific dependence influences the extinction of dominated strategies is seen in proposition [ prop : timetolive ] : a sharper estimate of the quadratic variation of @xmath243 could conceivably yield a more accurate estimate for the cumulative distribution function of ( [ eq : timetolive ] ) .
finally , it is only natural to ask if proposition [ prop : dominated ] can be extended to strategies that are only _ iteratively _ dominated . as it turns out , this is indeed the case .
[ thm : rational ] let @xmath19 be a solution path of ( [ eq : srd ] ) starting at @xmath244 . then ,
if @xmath38 is iteratively dominated , @xmath245 that is , _ only rationally admissible strategies survive in the long run . _ as in the deterministic case @xcite , the main idea is that the solution path @xmath19 gets progressively closer to the faces of @xmath246 that are spanned by the pure strategies which have not yet been eliminated .
following @xcite , we will prove this by induction on the rounds of elimination of dominated strategies ; proposition [ prop : dominated ] is simply the case @xmath247 . to wit ,
let @xmath248 , @xmath249 and denote by @xmath250 the set of strategies @xmath251 that are admissible ( i.e. , not dominated ) with respect to any strategy @xmath252 .
so , if we start with @xmath253 and @xmath254 , we may define inductively the set of strategies that remain admissible after @xmath3 elimination rounds by @xmath255 where @xmath256 ; similarly , the pure strategies that have survived after @xmath3 such rounds will be denoted by @xmath257 .
clearly , this sequence forms a descending chain @xmath258 and the set @xmath259 will consist precisely of the strategies of player @xmath32 that are rationally admissible .
assume then that the cross entropy @xmath260 diverges as @xmath261 for all strategies @xmath262 that die out within the first @xmath263 rounds ; in particular , if @xmath264 this implies that @xmath265 as @xmath266
. we will show that the same is true if @xmath179 survives for @xmath263 rounds but is eliminated in the subsequent one .
indeed , if @xmath267 but @xmath268 , there will exist some @xmath269 such that @xmath270 now , note that any @xmath271 can be decomposed as @xmath272 where @xmath273 is the `` admissible '' part of @xmath89 , that is , the projection of @xmath89 on the subspace spanned by the surviving vertices @xmath274 . hence , if @xmath275 , we will have @xmath205 and , by linearity , @xmath276 moreover , by the induction hypothesis , we also have @xmath277 as @xmath266 .
thus , there exists some @xmath278 such that @xmath279 for all @xmath280 [ recall that @xmath281 is spanned by already eliminated strategies ] .
therefore , as in the proof of proposition [ prop : dominated ] , we obtain for @xmath280 @xmath282 where @xmath283 is a constant depending only on @xmath278 . in this way , the same reasoning as before gives @xmath284 and the theorem follows . as a result ,
if there exists only one rationally admissible strategy , we get the following .
[ cor : dominance ]
let @xmath19 be an interior solution path of the replicator equation ( [ eq : srd ] ) for some dominance - solvable game @xmath48 and let @xmath285 be the ( unique ) strict equilibrium of @xmath48 . then @xmath286 that is , _ players converge to the game s strict equilibrium _
_ ) . in concluding this section
, it is important to note that all our results on the extinction of dominated strategies remain true in the adjusted dynamics ( [ eq : slrd ] ) as well : this is just a matter of rescaling . the only difference in using different learning rates @xmath85 comes about in proposition [ prop : timetolive ] where the estimate ( [ eq : timetolive ] ) becomes @xmath287 as it stands , this is not a significant difference in itself because the two estimates are asymptotically equal for large times . nonetheless , it is this very lack of contrast that clashes with the deterministic setting where faster learning rates accelerate the emergence of rationality .
the reason for this gap is that an increased learning rate @xmath85 also carries a commensurate increase in the noise coefficients @xmath288 , and thus deflates the benefits of accentuating payoff differences .
in fact , as we shall see in the next sections , the learning rates do not really allow players to learn any faster as much as they help diminish their shortsightedness : by effectively being lazy , it turns out that players are better able to average out the noise .
having established that irrational choices die out in the long run , we turn now to the question of whether equilibrial play is stable in the stochastic replicator dynamics of exponential learning .
however , before tackling this issue in complete generality , it will be quite illustrative to pay a visit to the class of congestion games where the presence of a potential simplifies things considerably . in this way
, the results we obtain here should be considered as a motivating precursor to the general case analyzed in section [ sec : equilibrium ] . to begin with
, it is easy to see that the potential @xmath74 of definition [ def : potential ] is a lyapunov function for the deterministic replicator dynamics .
indeed , assume that player @xmath35 is learning at a rate @xmath289 and let @xmath100 be a solution path of the rate - adjusted dynamics ( [ eq : lrd ] ) .
then , a simple differentiation of @xmath290 gives @xmath291\\[-8pt ] & = & -\sum_{i } \lambda_{i } \biggl(\sum_{\alpha } x_{i\alpha } u^{2}_{i\alpha } ( x ) - u_{i}^{2}(x ) \biggr)\leq0,\nonumber\end{aligned}\ ] ] the last step following from jensen s inequality recall that @xmath292 on account of ( [ eq : potential ] ) and also that @xmath293 .
in particular , this implies that the trajectories @xmath100 are attracted to the local minima of @xmath74 , and since these minima coincide with the strict equilibria of the game , we painlessly infer that strict equilibrial play is asymptotically stable in ( [ eq : lrd])as mentioned before , we plead guilty to a slight abuse of terminology in assuming that all equilibria in pure strategies are also strict .
it is therefore reasonable to ask whether similar conclusions can be drawn in the noisy setting of ( [ eq : slrd ] ) . mirroring the deterministic case , a promising way to go about
this question is to consider again the potential function @xmath74 of the game and try to show that it is stochastically lyapunov in the sense of definition [ def : lyapunov ] . indeed ,
if @xmath294 is a local minimum of @xmath74 ( and hence , a strict equilibrium of the underlying game ) , we may assume without loss of generality that @xmath295 so that @xmath296 in a neighborhood of @xmath297 .
we are thus left to examine the negativity condition of definition [ def : lyapunov ] , that is , whether there exists some @xmath128 such that @xmath298 for all @xmath299 sufficiently close to @xmath297 . to that end ,
recall that @xmath300 and that @xmath301 .
then , the generator @xmath118 of the rate - adjusted dynamics ( [ eq : slrd ] ) applied to @xmath74 produces @xmath302\\[-8pt ] & & { } -\sum_{i,\alpha } \frac{\lambda_{i}^{2}}{2 } x_{i\alpha } u_{i\alpha}(x ) \biggl(\eta_{i\alpha}^{2}(1 - 2x_{i\alpha } ) - \sum_{\beta}\eta_{i\beta}^{2 } x_{i\beta}(1 - 2x_{i\beta } ) \biggr),\nonumber\end{aligned}\ ] ] where , for simplicity , we have assumed that the noise coefficients @xmath138 are constant
. we will study ( [ eq : genv ] ) term by term by considering the perturbed strategies @xmath303 where @xmath304 belongs to the face of @xmath193 that lies opposite to @xmath305 ( i.e. , @xmath306 , @xmath307 and @xmath308 ) and @xmath309 measures the distance of player @xmath32 from @xmath305 . in this way
, we get @xmath310\\ & = & u_{i,0}(x ) -\varepsilon_{i } \sum_{\mu } y_{i\mu } \delta u_{i\mu } + \mathcal o(\varepsilon_{i}^{2 } ) , \nonumber\end{aligned}\ ] ] where @xmath311 .
then , by going back to ( [ eq : genv ] ) , we obtain @xmath312\nonumber\\ & & \qquad= ( 1-\varepsilon_{i } ) u_{i,0}(x ) [ u_{i,0}(x ) - u_{i}(x ) ] \nonumber\\ & & \qquad\quad { } + \varepsilon_{i } \sum_{\mu } y_{i\mu } u_{i\mu } ( x ) [ u_{i\mu}(x ) - u_{i}(x ) ] \nonumber\\ & & \qquad= ( 1-\varepsilon_{i } ) u_{i,0}(x ) \cdot\varepsilon_{i } \sum_{\mu } y_{i\mu } \delta u_{i\mu}\\ & & \qquad\quad{}- \varepsilon_{i } \sum_{\mu } y_{i\mu } u_{i\mu}(q_{0 } ) \delta u_{i\mu } + \mathcal o(\varepsilon_{i}^{2 } ) \nonumber\\ & & \qquad= \varepsilon_{i } \sum_{\mu } y_{i\mu } u_{i,0}(q_{0 } ) \delta u_{i\mu } - \varepsilon_{i } \sum_{\mu } y_{i\mu } u_{i\mu}(q_{0 } ) \delta u_{i\mu } + \mathcal o(\varepsilon_{i}^{2 } ) \nonumber\\ & & \qquad= \varepsilon_{i } \sum_{\mu } y_{i\mu } ( \delta u_{i\mu } ) ^{2 } + \mathcal o(\varepsilon_{i}^{2 } ) .\nonumber\end{aligned}\ ] ] as for the second term of ( [ eq : genv ] ) , some easy algebra reveals that @xmath313\\[-8pt ] & & \qquad\quad { } + 2(1-\varepsilon_{i})^{2 } \eta_{i,0}^{2 } + 2\varepsilon_{i}^{2}\sum_{\mu } \eta_{i\mu}^{2 } y_{i\mu } ^{2}\nonumber\\ & & \qquad= -\varepsilon_{i } \biggl(\eta_{i,0}^{2 } + \sum_{\mu } y_{i\mu } \eta _ { i\mu } ^{2 } \biggr ) + \mathcal o(\varepsilon_{i}^{2 } ) \nonumber\end{aligned}\ ] ] and , after a ( somewhat painful ) series of calculations , we get @xmath314\\[-8pt ] & & \qquad= -\varepsilon_{i } u_{i,0}(q_{0 } ) \biggl ( \eta_{i,0}^{2 } + \sum_{\mu}y_{i\mu}\eta_{i\mu}^{2 } \biggr)\nonumber\\ & & \qquad\quad{}+ \varepsilon_{i}\sum_{\mu } y_{i\mu } u_{i\mu}(q_{0 } ) ( \eta_{i\mu } ^{2 } + \eta_{i,0}^{2 } ) + \mathcal o(\varepsilon_{i}^{2 } ) \nonumber\\ & & \qquad= -\varepsilon_{i}\sum_{\mu } y_{i\mu } \delta u_{i\mu } ( \eta _ { i\mu } ^{2 } + \eta_{i,0}^{2 } ) + \mathcal o(\varepsilon_{i}^{2 } ) .\nonumber\end{aligned}\ ] ] finally , if we assume without loss of generality that @xmath295 and set @xmath315 ( i.e. , @xmath316 and @xmath317 for all @xmath35 , @xmath318 ) , we readily get @xmath319\\[-8pt ] & = & -\sum_{i,\alpha } u_{i\alpha}(q_{0 } ) \xi_{i\alpha } + \mathcal o(\varepsilon^{2 } ) \nonumber\\ & = & \sum_{i}\varepsilon_{i}\sum_{\mu } y_{i\mu}\delta u_{i\mu } + \mathcal o(\varepsilon^{2 } ) , \nonumber\end{aligned}\ ] ] where @xmath320 . therefore , by combining ( [ eq : paystep ] ) , ( [ eq : noisestep ] ) and ( [ eq : potstep ] ) , the negativity condition @xmath321 becomes @xmath322\nonumber\\[-8pt]\\[-8pt ] & & \qquad\geq k\sum_{i}\varepsilon_{i } \sum_{\mu } y_{i\mu}\delta u_{i\mu } + \mathcal o ( \varepsilon^{2}).\nonumber\end{aligned}\ ] ] hence , if @xmath323 for all @xmath324 , this last inequality will be satisfied for some @xmath128 whenever @xmath325 is small enough .
essentially , this proves the following .
[ prop : potential ] let @xmath326 be a strict equilibrium of a congestion game @xmath48 with potential function @xmath74 and assume that @xmath327 . assume further that the learning rates @xmath85 are sufficiently small so that , for all @xmath328 and all @xmath35 , @xmath329 then @xmath114 is stochastically asymptotically stable in the rate - adjusted dynamics ( [ eq : slrd ] ) .
we thus see that no matter how loud the noise @xmath288 might be , stochastic stability is always guaranteed if the players choose a learning rate that is slow enough as to allow them to average out the noise ( i.e. , @xmath330 ) . of course , it can be argued here that it is highly unrealistic to expect players to be able to estimate the amount of nature s interference and choose a suitably small rate @xmath85 . on top of that , the very form of the condition ( [ eq : payvsnoise ] ) is strongly reminiscent of the `` modified '' game of , a similarity which seems to contradict our statement that exponential learning favors rational reactions in the _ original _ game .
the catch here is that condition ( [ eq : payvsnoise ] ) is only _ sufficient _ and proposition [ prop : potential ] merely highlights the role of a potential function in a stochastic environment . as we shall see in section [ sec : equilibrium ]
, nothing stands in the way of choosing a different lyapunov candidate and dropping requirement ( [ eq : payvsnoise ] ) altogether . to gain some further intuition into why the condition ( [ eq : payvsnoise ] ) is redundant , it will be particularly helpful to examine the case where players compete for the resources of only two facilities ( i.e. , @xmath331 for all @xmath332 ) and try to learn the game with the help of the uniform replicator equation ( [ eq : srd ] ) .
this is the natural setting for the el farol bar problem @xcite and the ensuing minority game @xcite where players choose to `` buy '' or `` sell '' and are rewarded when they are in the minority buyers in a sellers market or sellers in an abundance of buyers .
as has been shown in @xcite , such games always possess strict equilibria , even when players have distinct payoff functions .
so , by relabeling indices if necessary , let us assume that @xmath333 is such a strict equilibrium and set .
then , the generator of the replicator equation ( [ eq : srd ] ) takes the form @xmath334\ , \frac{\partial}{\partial x_{i}}\nonumber\\[-8pt]\\[-8pt ] & & { } + \frac{1}{2}\sum_{i}x_{i}^{2}(1-x_{i})^{2 } \eta_{i}^{2}(x)\ , \frac { \partial^{2}}{\partial x_{i}^{2}},\nonumber\end{aligned}\ ] ] where now @xmath335 and @xmath336 .
it thus appears particularly appealing to introduce a new set of variables @xmath304 such that @xmath337 ; this is just the `` logit '' transformation : @xmath338 . in these new variables , ( [ eq:2dgen ] ) assumes the astoundingly suggestive guise @xmath339 which reveals that the noise coefficients can be effectively decoupled from the payoffs .
we can then take advantage of this by letting @xmath118 act on the function @xmath340 ( @xmath341 ) : @xmath342 hence , if @xmath343 is chosen small enough so that @xmath344 for all sufficiently large @xmath304 [ recall that @xmath345 since @xmath297 is a strict equilibrium ] , we get @xmath346 where @xmath347 . and since @xmath124 is strictly positive for @xmath348 and only vanishes as @xmath349 ( i.e. , at the equilbrium @xmath297 ) , a trivial modification of the stochastic lyapunov method ( see , e.g. , pages 314 and 315 of @xcite ) yields the following . [ prop : minority ] the strict equilibria of minority games are stochastically asymptotically stable in the uniform replicator equation ( [ eq : srd ] ) .
it is trivial to see that strict equilibria of minority games will also be stable in the rate - adjusted dynamics ( [ eq : slrd ] ) : in that case we simply need to choose @xmath343 such that @xmath350 .
a closer inspection of the calculations leading to proposition [ prop : minority ] reveals that nothing hinges on the minority mechanism per se : it is ( [ eq:2dygen ] ) that is crucial to our analysis and @xmath118 takes this form whenever the underlying game is a _ dyadic _ one ( i.e. , @xmath351 for all @xmath35 ) . in other words ,
proposition [ prop : minority ] also holds for all games with 2 strategies and should thus be seen as a significant extension of proposition [ prop : potential ] .
[ prop : dyadic ] the strict equilibria of dyadic games are stochastically asym ptotically stable in the replicator dynamics ( [ eq : srd ] ) , ( [ eq : slrd ] ) of exponential learning .
in deterministic environments , the `` folk theorem '' of evolutionary game theory provides some pretty strong ties between equilibrial play and stability : strict equilibria are asymptotically stable in the multi - population replicator dynamics ( [ eq : rd ] ) @xcite . in our stochastic setting
, we have already seen that this is always true in two important classes of games : those that can be solved by iterated elimination of dominated strategies ( corollary [ cor : dominance ] ) and dyadic ones ( proposition [ prop : dyadic ] ) .
although interesting in themselves , these results clearly fall short of adding up to a decent analogue of the folk theorem for stochastically perturbed games
. nevertheless , they are quite strong omens in that direction and such expectations are vindicated in the following .
[ thm : stability ] the strict equilibria of a game @xmath48 are stochastically asymptotically stable in the replicator dynamics ( [ eq : srd ] ) , ( [ eq : slrd ] ) of exponential learning . before proving theorem
[ thm : stability ] , we should first take a slight detour in order to properly highlight some of the issues at hand . on that account ,
assume again that the profile @xmath352 is a strict equilibrium of @xmath48 .
then , if @xmath297 is to be stochastically stable , say in the uniform dynamics ( [ eq : srd ] ) , one would expect the strategy scores @xmath353 of player @xmath32 to grow much faster than the scores @xmath354 of his other strategies .
this is captured remarkably well by the `` adjusted '' scores [ eq : ratescore ] @xmath355 where @xmath289 is a sensitivity parameter akin ( but not identical ) to the learning rates of ( [ eq : slrd ] ) ( the choice of common notation is fairly premeditated though ) .
clearly , whenever @xmath356 is large , @xmath353 will be much greater than any other score @xmath357 and hence , the strategy @xmath358 will be employed by player @xmath32 far more often . to see this in more detail ,
it is convenient to introduce the variables [ eq : ydef ] @xmath359 where @xmath360 is a measure of how close @xmath361 is to @xmath362 and @xmath363 is a direction indicator ; the two sets of coordinates are then related by the transformation @xmath364 , @xmath45 , @xmath365 . consequently , to show that the strict equilibrium @xmath366 is stochastically asymptotically stable in the replicator equation ( [ eq : srd ] ) , it will suffice to show that @xmath360 diverges to infinity as @xmath266 with arbitrarily high probability .
our first step in this direction will be to derive an sde for the evolution of the @xmath367 processes . to that end ,
it s lemma gives @xmath368\\[-8pt ] & = & \sum_{\beta } \biggl ( u_{i\beta}\frac{\partial y_{i\alpha}}{\partial u_{i\beta } } + \frac{1}{2 } \eta_{i\beta}^{2 } \,\frac{\partial^{2 } y_{i\alpha } } { \partial u_{i\beta}^{2 } } \biggr ) \,dt + \sum_{\beta } \eta_{i\beta } \,\frac{\partial y_{i\alpha}}{\partial u_{i\beta } } \,dw_{i\beta},\nonumber\end{aligned}\ ] ] where , after a simple differentiation of ( [ eq : y0 ] ) , we have @xmath369\end{aligned}\ ] ] @xmath370 and , similarly , from ( [ eq : ym ] ) @xmath371\end{aligned}\ ] ] @xmath372\\[-8pt ] \frac{\partial^{2 } y_{i\mu}}{\partial u_{i\nu}^{2 } } & = & \lambda_{i}^{2 } y_{i\mu}(\delta_{\mu\nu } - y_{i\nu})(1 - 2y_{i\nu}).\nonumber\end{aligned}\ ] ] in this way , by plugging everything back into ( [ eq : itoy ] ) we finally obtain [ eq : dy ] @xmath373 \,dt\nonumber\\[-8pt]\\[-8pt ] & & { } + \lambda_{i } y_{i,0 } \biggl [ \eta_{i,0 } \,dw_{i,0 } -\sum_{\mu } \eta_{i\mu } y_{i\mu } \,dw_{i\mu } \biggr],\nonumber\\ \label{eq : dym } dy_{i\mu } & = & \lambda_{i}y_{i\mu } [ u_{i\mu } - \sum_{\nu } u_{i\nu } y_{i\nu } ] \,dt\nonumber\\ & & { } + \frac{\lambda_{i}^{2}}{2}y_{i\mu } \biggl[\eta_{i\mu}^{2}(1 - 2 y_{i\mu } ) -\sum_{\nu}\eta_{i\nu}^{2 } y_{i\nu } ( 1 - 2y_{i\nu } ) \biggr ] \,dt\\ & & { } + \lambda_{i } y_{i\mu } \biggl[\eta_{i\mu } \,dw_{i\mu } - \sum_{\nu } \eta _ { i\nu } y_{i\nu } \,dw_{i\nu } \biggr],\nonumber\end{aligned}\ ] ] where we have suppressed the arguments of @xmath47 and @xmath288 in order to reduce notational clutter . this last sde is particularly revealing : roughly speaking , we see that if @xmath85 is chosen small enough , the deterministic term @xmath374 will dominate the rest ( cf . with the `` soft '' learning rates of proposition [ prop : potential ] ) . and
, since we know that strict equilibria are asymptotically stable in the deterministic case , it is plausible to expect the sde ( [ eq : dy ] ) to behave in a similar fashion .
proof of theorem [ thm : stability ] tying in with our previous discussion , we will establish stochastic asymptotic stability of strict equilibria in the dynamics ( [ eq : srd ] ) by looking at the processes @xmath375 of ( [ eq : ydef ] ) . in these coordinates
, we just need to show that for every @xmath376 and any @xmath115 , there exist @xmath377 such that if @xmath378 , then , with probability greater than @xmath379 , @xmath380 and @xmath381 for all @xmath162 . in the spirit of the previous section , we will accomplish this with the help of the stochastic lyapunov method
. our first task will be to calculate the generator of the diffusion @xmath382 , that is , the second order differential operator @xmath383 where @xmath384 and @xmath385 are the drift and diffusion coefficients of the sde ( [ eq : dy ] ) , respectively .
in particular , if we restrict our attention to sufficiently smooth functions of the form @xmath386 , the application of @xmath118 yields @xmath387\ , \frac{\partial f_{i}}{\partial y_{i,0}}\\ & & { } + \frac{1}{2 } \sum_{i\in\mathcal{n } } \lambda_{i}^{2 } y_{i,0}^{2 } \biggl [ \eta_{i,0}^{2 } + \sum_{\mu } \eta_{i\mu}^{2 } y_{i\mu}^{2 } \biggr]\ , \frac{\partial^{2 } f_{i}}{\partial^{2}y_{i,0}}.\nonumber\end{aligned}\ ] ] therefore , let us consider the function @xmath388 for @xmath389
. with @xmath390 and @xmath391 , ( [ eq : generator ] ) becomes @xmath392\\[-8pt ] & & \hspace*{54.1pt } { } - \frac{\lambda_{i}}{2}\sum_{\mu } y_{i\mu}(1-y_{i\mu})\eta _ { i\mu } ^{2 } \biggr].\nonumber\end{aligned}\ ] ] however , since @xmath366 has been assumed to be a strict nash equilibrium of @xmath48 , we will have @xmath393 for all @xmath394 .
then , by continuity , there exists some positive constant @xmath205 with @xmath395 whenever @xmath396 is large enough ( recall that @xmath397 ) .
so , if we set @xmath398 and pick positive @xmath85 with @xmath399 , we get @xmath400 for all sufficiently large @xmath396 .
moreover , @xmath124 is strictly positive for @xmath348 and vanishes only as @xmath401 .
hence , as in the proof of proposition [ prop : minority ] , our claim follows on account of @xmath124 being a ( local ) stochastic lyapunov function . finally , in the case of the rate - adjusted replicator dynamics ( [ eq : slrd ] ) , the proof is similar and only entails a rescaling of the parameters @xmath85 .
if we trace our steps back to the coordinates @xmath402 , our lyapunov candidate takes the form @xmath403 .
it thus begs to be compared to the lyapunov function @xmath404 employed by imhof and hofbauer in @xcite to derive a conditional version of theorem [ thm : stability ] in the evolutionary setting . as it turns out , the obvious extension @xmath405 works in our case as well , but the calculations are much more cumbersome and they are also shorn of their ties to the adjusted scores ( [ eq : ratescore ] ) .
we should not neglect to highlight the dual role that the learning rates @xmath85 play in our analysis . in the logistic learning model ( [ eq : ratelogit ] ) , they measure the players convictions and how strongly they react to a given stimulus ( the scores @xmath75 ) ; in this role , they are fixed at the outset of the game and form an intrinsic part of the replicator dynamics ( [ eq : slrd ] ) . on the other hand
, they also make a virtual appearance as free temperature parameters in the adjusted scores ( [ eq : ratescore ] ) , to be softened until we get the desired result . for this reason ,
even though theorem [ thm : stability ] remains true for any choice of learning rates , the function @xmath406 is lyapunov only if the sensitivity parameters @xmath154 are small enough .
it might thus seem unfortunate that we chose the same notation in both cases , but we feel that our decision is justified by the intimate relation of the two parameters .
our aim in this last section will be to discuss a number of important issues that we have not been able to address thoroughly in the rest of the paper ; truth be told , a good part of this discussion can be seen as a roadmap for future research . in single - population evolutionary models ,
an evolutionarily stable strategy ( ess ) is a strategy which is robust against invasion by mutant phenotypes @xcite .
strategies of this kind can be considered as a stepping stone between mixed and strict equilibria and they are of such significance that it makes one wonder why they have not been included in our analysis .
the reason for this omission is pretty simple : even the weakest evolutionary criteria in multi - population models tend to reject all strategies which are not strict nash equilibria @xcite .
therefore , since our learning model ( [ eq : rd ] ) corresponds exactly to the multi - population environment ( [ eq : erd ] ) , we lose nothing by concentrating our analysis only on the strict equilibria of the game .
if anything , this equivalence between ess and strict equilibria in multi - population settings further highlights the importance of the latter .
however , this also brings out the gulf between the single - population setting and our own , even when we restrict ourselves to 2-player games ( which are the norm in single - population models ) .
indeed , the single - population version of the dynamics ( [ eq : asrd ] ) is : @xmath407 \,dt\nonumber\\[-8pt]\\[-8pt ] & & { } + x_{\alpha } \bigl[\eta_{\alpha } \,dw_{\alpha } - \sum\eta_{\beta } x_{\beta } \,dw_{\beta } \bigr].\nonumber\end{aligned}\ ] ] as it turns out , if a game possesses an interior ess and the shocks are mild enough , the solution paths @xmath19 of the ( single - population ) replicator dynamics will be recurrent ( theorem 2.1 in @xcite ) . theorem
[ thm : stability ] rules out such behavior in the case of strict equilibria ( the multi - population analogue of ess ) , but does not answer the following question : if the underlying game only has mixed equilibria , will the solution @xmath19 of the dynamics ( [ eq : srd ] ) be recurrent ? this question is equivalent to showing that a profile @xmath299 is stochastically asymptotically stable in the replicator equations ( [ eq : srd ] ) , ( [ eq : slrd ] ) only if it is a strict equilibrium . since theorem [ thm : stability ] provides the converse `` if '' part , an answer in the positive would yield a strong equivalence between stochastically stable states and strict equilibria ; we leave this direction to be explored in future papers . for comparison purposes ( but also for simplicity ) ,
let us momentarily assume that the noise coefficients @xmath137 do not depend on the state @xmath19 of the game . in that case , it is interesting ( and very instructive ) to note that the sde ( [ eq : score ] ) remains unchanged if we use stratonovich integrals instead of it ones : @xmath408 then , after a few calculations , the corresponding replicator equation reads @xmath409 the form of this last equation is remarkably suggestive .
first , it highlights the role of the modified game @xmath239 even more crisply than ( [ eq : srd ] ) : the payoff terms are completely decoupled from the noise , in contrast to what one obtains by introducing stratonovich perturbations in the evolutionary setting @xcite .
secondly , one can seemingly use this simpler equation to get a much more transparent proof of proposition [ prop : dominated ] : the estimates for the cross entropy terms @xmath410 are recovered almost immediately from the stratonovich dynamics .
however , since ( [ eq : stratonovich ] ) takes this form only for constant coefficients @xmath138 ( the general case is quite a bit uglier ) , we chose the route of consistency and employed it integrals throughout our paper . before closing , it is worth pointing out the applicability of the above approach to networks where the presence of noise or uncertainty has two general sources .
the first of these has to do with the time variability of the connections which may be due to the fluctuations of the link quality because of mobility in the wireless case or because of external factors ( e.g. , load conditions ) in wireline networks .
this variability is usually dependent on the state of the network and was our original motivation in considering noise coefficients @xmath137 that are functions of the players strategy profile ; incidentally , it was also our original motivation for considering randomly fluctuating payoffs in the first place : travel times and delays in traffic models are not determined solely by the players choices , but also by the fickle interference of nature .
the second source stems from errors in the measurement of the payoffs themselves ( e.g. , the throughput obtained in a particular link ) and also from the lack of information on the payoff of strategies that were not employed .
the variability of the noise coefficients @xmath411 again allows for a reasonable approximation to this problem .
indeed , if @xmath412 is continuous and satisfies @xmath413 for all @xmath80 , this means that there are only errors in estimating the payoffs of strategies that were not employed ( or small errors for pure strategies that are employed with high probability ) . of course
, this does not yet give the full picture [ one should consider the discrete - time dynamical system ( [ eq : discretescore ] ) instead where the players _ actual _ choices are considered ] , but we conjecture that our results will remain essentially unaltered .
we would like to extend our gratitude to the anonymous referee for his insightful comments and to david leslie from the university of bristol for the fruitful discussions on the discrete version of the exponential learning model . | we study repeated games where players use an exponential learning scheme in order to adapt to an ever - changing environment .
if the game s payoffs are subject to random perturbations , this scheme leads to a new stochastic version of the replicator dynamics that is quite different from the `` aggregate shocks '' approach of evolutionary game theory .
irrespective of the perturbations magnitude , we find that strategies which are dominated ( even iteratively ) eventually become extinct and that the game s strict nash equilibria are stochastically asymptotically stable .
we complement our analysis by illustrating these results in the case of congestion games . and . |
with the introduction of microarrays biologist have been witnessing entire labs shrinking to matchbox size .
this paper invites quality researchers to join scientists on their _ fantastic journey _ into the world of microscopic high - throughput measurement technologies . building a biological organism as laid out by the genetic code
is a multi - step process with room for variation at each step .
the first steps , as described by the _ dogma of molecular biology , _ are genes ( and dna sequence in general ) , their transcripts and proteins .
substantial factors contributing to their variation in both structure and abundance include cell type , developmental stage , genetic background and environmental conditions .
connecting molecular observations to the state of an organism is a central interest in molecular biology .
this includes the study of the gene and protein functions and interactions , and their alteration in response to changes in environmental and developmental conditions .
traditional methods in molecular biology generally work on a `` one gene ( or protein ) in one experiment '' basis . with the invention of _ microarrays _
huge numbers of such macromolecules can now be monitored in one experiment .
the most common kinds are _ gene expression microarrays , _ which measure the mrna transcript abundance for tens of thousands of genes simultaneously . for biologists , this high - throughput approach has opened up entirely new avenues of research . rather than experimentally confirming the hypothesized role of a certain candidate gene in a certain cellular process , they can use genome - wide comparisons to screen for all genes which might be involved in that process .
one of the first examples of such an exploratory approach is the expression profiling study of mitotic yeast cells by @xcite which determined a set of a few hundred genes involved in the cell cycle and triggered a cascade of articles re - analyzing the data or replicating the experiment .
microarrays have become a central tool in cancer research initiated by the discovery and re - definition of tumor subtypes based on molecular signatures ( see e.g. @xcite , @xcite , @xcite , @xcite ) . in section [ b ]
we will explain different kinds of microarray technologies in more detail and describe their current applications in life sciences research .
a dna microarray consists of a glass surface with a large number of distinct fragments of dna called probes attached to it at fixed positions .
a fluorescently labelled sample containing a mixture of unknown quantities of dna molecules called the target is applied to the microarray . under the right chemical conditions ,
single - stranded fragments of target dna will base pair with the probes which are their complements , with great specificity .
this reaction is called hybridization , and is the reason dna microarrays work .
the fixed probes are either fragments of dna called complementary dna ( cdna ) obtained from messenger rna ( mrna ) , or short fragments known to be complementary to part of a gene , spotted onto the glass surface , or synthesized in situ .
the point of the experiment is to quantify the abundance in the target of dna complementary to each particular probe , and the hybridization reaction followed by scanning allows this to be done on a very large scale .
the raw data produced in a microarray experiment consists of scanned images , where the image intensity in the region of a probe is proportional to the amount of labelled target dna that base pairs with that probe . in this way we can measure the abundance of thousands of dna fragments in a target sample .
microarrays based on cdna or long oligonucleotide probes typically use just one or a few probes per gene .
the same probe sequence spotted in different locations , or probe sequences complementary to different parts of the same gene can be used to give within array replication .
short oligonucleotide microarrays typically use a larger number per gene , e.g. 11 for the hu133 affymetrix array per gene .
such a set of 11 is called probeset for that gene , and the probes in a probe set are arranged randomly over the array . in the biological literature
, microarrays are also referred to as ( gene ) chips or slides . when the first microarray platforms were introduced in the early 90s ,
the most intriguing fact about them was the sheer number of genes that could be assayed simultaneously .
assays that used to be done one gene at a time , could suddenly be produced for thousands of genes at once .
a decade later , high - density microarrays would even fit entire genomes of higher organisms .
after the initial euphoria , the research community became aware that findings based solely on microarray measurements were not always as reproducible as they would have liked and that studies with inconclusive results were quite common . with this high - throughput measurement technology becoming established in many branches of life sciences research , scientists in both academic and corporate environments raised their expectations concerning the validity of the measurements .
data quality issues are now frequently addressed at meetings of the microarray gene expression database group ( mged ) .
the _ microarray quality control project , _ a community - wide effort , under the auspices of the u.s . food and drug administration ( fda ) , is aiming at establishing _ operational metrics _ to objectively assess the performance of seven microarray platform and develop minimal quality standards .
their assessment is based on the performance of a set of standardized external rna controls .
the first formal results of this project have been published in a series of articles in the september 2006 issue of _ nature biotechnology .
_ assessing the _ quality of microarray data _ has emerged as a new research topic for statisticians . in this paper , we conceptualize microarray data quality issues from a perspective which includes the technology itself as well as their practical use by the research community .
we characterize the nature of microarray data from a quality assessment perspective , and we explain the different levels of microarray data quality assessment
. then we focus on short oligonucleotide microarrays to develop a set of specific statistical data quality assessment methods including both numerical measures and spatial diagnostics .
assumptions and hopes about the quality of the measurements have become a major issue in microarray purchasing . despite their substantially higher costs , affymetrix short oligonucleotide microarrays
have become a widespread alternative to cdna chips .
informally , they are considered the industrial standard among all microarray platforms . more recently , agilent s
non - contact printed high - density cdna microarrays and illumina s bead arrays have fueled the competition for high quality chips .
scientist feel the need for systematic quality assessment methods allowing them to compare different laboratories , different chip generation , or different platforms .
they even lack good methods for selecting chips of good enough quality to be included in statistical data analysis beyond preprocessing .
we have observed several questionable practices in the recent past : * skipping hybridization qa / qc all together * discarding entire batches of chips following the detection of a few poor quality chips * basing hybridization qa / qc on raw data rather than data that has already been had large - scale technical biases removed * delaying any qa / qc until all hybridizations are completed , thereby losing the opportunity to remove specific causes of poor quality at an early stage * focussing on validation by another measurement technolgy ( e.g , quantitative pcr ) in publication requirements rather than addressing the quality of the microarray data in the first place * merging of data of variable quality into one database with the inherent risk of swamping it with poor quality data ( as this produced at a faster rate due to few replicates , less quality checks , less re - doing of failed hybridizations etc . ) the community of microarray users has not yet agreed on a framework to measure accurary or precision in microarray experiments . without universally accepted methods for quality assessment , and guidelines for acceptance , statisticians judgements about data quality may be perceived as arbitrary by experimentalists .
users expectations as to the level of gene expression data quality vary substantially .
they can depend on time frame and financial constraints , as well as on the purpose of their data collection .
@xcite , p.120/21 , explained the standpoint of the applied scientist : + _ he knows that if he were to act upon the meagre evidence sometimes available to the pure scientist , he would make the same mistakes as the pure scientist makes in estimates of accuracy and precisions .
he also knows that through his mistakes someone may lose a lot of money or suffer physical injury or both .
[ ... ] he does not consider his job simply that of doing the best he can with the available data ; it is his job to get enough data before making this estimate . _ + following this philosophy , microarray data used for medical diagnostics should meet high quality standards .
in contrast , microarray data collected for a study of the etiology of a complex genetic disease in a heterogeneous population , one may decide to tolerate lower standards at the level of individual microarrays and invest the resources in a larger sample size .
scientists need informative quality assessment tools to allow them to choose the most appropriate technology and optimal experimental design for their precision needs , within their time and budget constraints .
the explicit goals of quality assessment for microarrays are manifold .
which goals can be envisioned depends on the resources and time horizon and on the kind of user single small user , big user , core faculity , multi - center study , or `` researcher into quality '' .
the findings can be used to simply exclude chips from further study or recommend to have samples reprocessed .
they can be imbedded in a larger data quality management and improvement plan .
typical quality phenomena to look for include : * outlier chips * trends or patterns over time * effects of particular hybridization conditions and sample characteristics * changes in quality between batches of chips , cohorts of samples , lab sites etc .
* systematic quality differences between subgroups of a study some aspects of quality assessment and control for cdna have been discussed in the literature . among these ,
@xcite and @xcite emphasize the need for quality control and replication .
@xcite define a quality score for each spot based on intensity characteristics and spatial information , while @xcite approach this with baysian networks . @xcite and @xcite
suggest explicit statistical quality measures based on individual spot observations using the image analysis software spot from @xcite . @xcite
apply multivariate statistical process control to detect single outlier chips .
the preprocessing and data management software package arraymagic of @xcite includes quality diagnostics .
the book by @xcite is a comprehensive collection of quality assessment and control issues concerning the various stages of cdna microarray experiments including sample preparation , all from an experimentalist s perspective .
@xcite suggest spot quality scores based on the variance of the ratio estimates of replicates ( on the same chip or on different chips ) .
spatial biases have also been addressed .
in examining the relationship between signal intensity and print - order , @xcite reveals a plate - effect .
the normalization methodology by @xcite incorporates spatial information such as print - tip group or plate , to remove spatial biases created by the technological processes . @xcite and @xcite found pairwise correlations between genes due to their relative positioning of the spots on the slide and suggest a localized mean normalization method to adjust for this .
@xcite proposed a method of identifying poor quality spots , and of addressing this by assigning quality weights .
@xcite developed an approach for the visualization and quantitation of regional bias applicable to both cdna and affymetrix microarrays .
for affymetrix arrays , the commercial software @xcite includes a _ quality report _ with a dozen scores for each microarray ( see subsection [ d_a ] ) .
none of them makes use of the gene expression summaries directly , and there are no universally recognized guidelines as to which range should be considered good quality for each of the gcos quality scores .
users of short oligonucleotide chips have found that the quality picture delivered by the gcos quality report is incomplete or not sensitive enough , and that it is rarely helpful in assigning causes to poor quality .
the literature on quality assessment and control for short oligonucleotide arrays is still sparse , though the importance of the topic has been stressed in numerous places , and some authors have addressed have looked at specific issues .
an algorithm for probeset quality assessment has been suggested by @xcite .
@xcite transfer the weight of a measurement to a subset of probes with optimal linear response at a given concentration .
@xcite investigate the effect of updating the mapping of probes to genes on the estimated expression values .
@xcite define four types of degenerate probe behaviour based on free energy computations and pattern recognition .
@xcite evaluated the affymetrix quality reports of over 5,000 chips collected by st .
jude children s research hospital over a period of three years , and linked some quality trends to experimental conditions .
@xcite extend traditional effect size models to combine data from different microarray experiments , incorporating a quality measure for each gene in each study .
the detection of specific quality issues such as the extraction , handling and amount of rna , has been studied by several authors ( e.g.@xcite , @xcite , @xcite , @xcite ) . before deriving new methods for assessing microarray data quality
, we will relate the issue to established research into data quality from other academic disciplines , emphasizing the particular characteristics of microarray data ( section [ d_c ] ) . a conceptual approach to the statistical assessment of microarray data quality
is suggested in subsection [ d_a ] , and is followed by a summary of the existing quality measures for affymetrix chips .
the theoretical basis of this paper is section [ m ] , where we introduce new numerical and spatial quality assessment methods for short oligonucleotide arrays . two important aspects of our approach are : * the quality measures are based on _ all the data _ from the array . *
the quality measures are computed after hybridization and data preprocessing .
more specifically , we make use of probe level and probeset level quantities obtained as by - products of the robust multichip analysis ( rma / fitplm ) preprocessing algorithm presented in @xcite , @xcite and @xcite . our _ quality landscapes _
serve as tools for visual quality inspection of the arrays after hybridization .
these are two dimensional pseudo - images of the chips based on probe level quantities , namely the weights and residuals computed by rma / fitplm .
these quality landscapes allow us to immediately relate quality to an actual location on the chip , a crucial step in detecting special causes for poor chip quality .
our numerical quality assessment is based on two distributions computed at the probeset level , the
_ normalized unscaled standard error ( nuse ) _ and the _ relative log expression ( rle ) . _ given a fairly general biological assumption is fulfilled , these distributions can be interpreted for chip quality assessment .
we further suggest ways of conveniently visualizing and summarizing these distributions for larger chip sets and of relating this quality assessment with other factors in the experiment , to permit the detection of special causes for poor quality to reveal biases .
quality of gene expression data can be assessed on a number of levels , including that of probeset , chip and batch of chips . another aspect of quality assessment concerns batches of chips .
we introduce the _ residual scale factor ( rsf ) , _ a measure of chip batch quality .
this allows us to compare quality across batches of chips within an experiment , or across experiments .
all our measures can be computed for all available types of short oligonuceotide chips given the raw data ( cel file ) for each chip and the matching cdf file .
software packages are described in @xcite and available at www.bioconductor.org . in section [ r ]
we extensively illustrate and evaluate our quality assessment methods on the experimental microarray datasets described in section [ s ] .
to reflect the fact that quality assessment is a necessary and fruitful step in studies of any kind , we use a variety of datasets , involving tissues ranging from fruit fly embryos to human brains , and from academic , clinical , and corporate labs .
we show how quality trends and patterns can be associated with sample characteristics and/or experimental conditions , and we compare our measures with the affymetrix gcos quality report .
after the hunt for new genes has dominated genetics in the 80s and 90s of the last century , there has be a remarkable shift in molecular biology research goals towards a comprehensive understanding of the function of macromolecules on different levels in a biological organism .
how and to what extend do genes control the construction and maintenance of the organism ?
what is the role of intermediate gene products such as _ rna transcripts _ ? how do the macromolecules interact with others ?
the latter may refer to _ horizontal _ interaction , such as genes with genes , or proteins with proteins .
it may also refer to _ vertical _ interaction , such as between genes and proteins .
_ genomics _ and _ proteomics _ in professional slang summarized as _
omics sciences _ have started to put an emphasis on _ functions .
_ as the same time , these research areas have become more quantitative , and they have broadened the perspective in the sense of observing huge numbers of macromolecules simultaneously .
these trends have been driven by recent biotechnological inventions , the most prominent ones being _ microarrays . _ with these _ high - throughput _ molecular measurement instruments , the relative concentration of huge numbers of macromolecules can be obtained simultaneously in one experiment
. this section will give an overview of the biological background and the applications of microarrays in biomedical research .
for an extended introduction to _
omics sciences _ and to microarray - based research we refer to the excellent collections of articles in the three nature genetics supplements _ the chipping forecast i , ii , iii _
( 1999 , 2002 , 2005 ) and to the recent review paper by @xcite .
though the popular belief about genes is still very deterministic once they are put into place , they function in a preprogrammed straight forward way for biologists the effect of a gene is variable .
most cells in an organism contain essentially the same set of genes .
however , cells will look and act differently depending on which organ they belong to , the state of the organ ( e.g.healthy vs.diseased ) , the developmental stage of the cell , or the phase of the cell cycle .
this is predominantly the result of differences in the abundance , distribution , and state of the cells proteins . according to the _ central dogma of molecular biology _ the production of proteins
is controlled by dna ( for simplicity , the exceptions to this rule are omitted here ) .
proteins are polymers built up from 20 different kinds of amino acids .
genes are _ transcribed _ into dna - like macromolecules called _ messenger rna ( mrna ) _ , which goes from the chromosomes to the _ ribosomes .
_ there , _ translation _ takes place , converting mrna into the amino acid chains which fold into proteins .
the term _ gene expression _ is defined as the relative concentration of mrna and protein produced by that gene .
depending on the context , however , it is often used to refer to only one of the two .
the _ gene expression profile _ of a type of cell usually refers to the relative abundance of each of the mrna species in the total cellular mrna population . from a practical point of view , in particular by many areas of medical research , protein abundance is seen as generally more interesting than mrna abundance .
the measurement of protein abundances , however , is still much more difficult to measure on a large scale than mrna abundance .
there is one property which is peculiar to nucleic acids : their complementary structure . dna is reliably replicated by separating the two strands , and complementing each of the single strands to give a copy of the original dna .
the same mechanism can be used to detect a particular dna or rna sequence in a mixed sample .
the first tool to measure gene expression in a sample of cells of a was introduced in 1975 .
southern blot _ named for its inventor is a multi - stage laboratory procedure which produces a pattern of bands representing the activity of a small set of pre - selected genes . during the 1980s spotted arrays on nylon holding bacterial colonies carrying different genomic inserts were introduced . in the early 1990s , the latter
would be exchanged for preidentified cdnas .
the introduction of _ gene expression microarrays _ on glass slides in the mid 1990s brought a substantial increase in feature density . with the new technology ,
gene expression measurements could be taken in parallel for thousands of genes .
modern microarray platforms even assess the expression levels of tens of thousands of genes simultaneously .
a gene expression microarray is a small piece of glass onto which _ a priori _ known dna fragments called _ probes _ are attached at fixed positions . in a chemical process called _ hybridization , _
the microarray is brought into contact with material from a sample of cells .
each probe binds to its complementary counterpart , an mrna molecule ( or a complementary dna copy ) from the sample , which we refer to as the _
target_. the hybridization reaction product is made visible using fluorescent dyes or other ( e.g.radioactive ) markers , which are applied to the sample prior to hybridization .
the readout of the microarray experiment is a scanned image of the labelled dna .
microarrays are specially designed to interrogate the genomes of particular organisms , and so there are yeast , fruit fly , worm and human arrays , to name just a few .
there are three major platforms for microarray - based gene expression measurement : _ spotted two - color cdna arrays , _ _ long oligonucleotide arrays _ and _ short oligonucleotide arrays .
_ in the platform specific parts of this paper we will focus on the latter . on a short oligonucleotide microarray , each gene is represented on the array by a _ probe set _ that uniquely identifies the gene .
the individual probes in the set are chosen to have relatively uniform hybridization characteristics . in the affymetrix hu133 arrays , for example ,
each probe set consists of 11 to 20 probe sequence pairs .
each pairs consists of a _ perfect match ( pm ) _ probe , a 25 bases long oligonucleotide that matches a part of the gene s sequence , and a corresponding _ mismatch ( mm ) _ probe , that has the same sequence as the pm except for the center base being flipped to its complementary letter .
the mm probes are intended to give an estimate of the random hybridization and cross hybridization signals , see @xcite and @xcite for more details .
other affymetrix gene expression arrays may differ from the hu133 in the number of probes per probe set .
exon arrays do not have mm probes .
most of the arrays produced by nimblegen are composed from 60mer probes , but some are using 25mer probes .
the number of probes per probeset is adapted to the total number of probesets on the array to make optimal use of the space . besides being more efficient than the classical gene - by - gene approach , microarrays open up entirely new avenues for research .
they offer a comprehensive and cohesive approach to measuring the activity of the genome .
in particular , this fosters the study of interactions .
a typical goal of a microarray based research project is the search for genes that behave differently between different cell populations .
some of the most common examples for comparisons are diseased vs. healthy cells , injured vs. healthy tissue , young vs. old organism , treated vs. untreated cells . more explicitly
, life sciences researchers try to find answers to questions such as the following .
which genes are affected by environmental changes or in response to a drug ? how do the gene expression levels differ across various mutants ?
what is the gene expression signature of a particular disease ?
which genes are involved in each stage of a cellular process ? which genes play a role in the development of an organism ? or , more generally , which genes vary their activity with time ?
the principle of microarray measurement technology has been used to assess molecules other than mrna .
a number of platforms are currently at various stages of development ( see review by @xcite ) .
snp chips detect single nucleotide polymorphisms .
they are an example for a well developed microarray - based genotyping platform .
cgh arrays are based on comparative genome hybridization .
this method permits the analysis of changes in gene copy number for huge numbers of probes simultaneously .
a recent modification , representational oligonucleotide microarray analysis ( roma ) , offers substantially better resolution .
both snp chips and cgh arrays are genome - based methods , which , in contrast to the gene expression - based methods , can exploit the stability of dna .
the most common application of these technologies is the localization of disease genes based on association with phenotypic traits .
antibody protein chips are used to determine the level of proteins in a sample by binding them to antibody probes immobilized on the microarray .
this technology is still considered semi - quantitative , as the different specificities and sensitivities of the antibodies can lead to an inhomogeneity between measurements that , so far , can not be corrected for .
the applications of protein chips are similar to the ones of gene expression microarrays , except that the measurements are taken one step further downstream .
more recent platforms address multiple levels at the same time .
chip - on - chip , also known as genome - wide location analysis , is a technique for isolation and identification of the dna sequences occupied by specific dna binding proteins in cells .
the still growing list of statistical challenges stimulated by microarray data is a _
tour dhorizon _ in applied statistics ; see e.g. @xcite , @xcite and @xcite for broad introductions . from a statistical point of view a microarray experiment
has three main challenges : ( i ) measurement process as multi - step biochemical and technological procedure ( array manufacturing , tissue acquisition , sample preparation , hybridization , scanning ) with each step contributing to the variation in the data ; ( ii ) huge numbers measurements of different ( correlated ) molecular species being take in parallel ; ( iii ) unavailability of gold - standards covering a representative part of these species .
statistical methodology has primarily been developed for gene expression microarrays , but most of the conceptual work applies directly to many kinds of microarrays and many of the actual methods can be transferred to other microarray platforms fitting the characteristics listed above .
the first steps of the data analysis , often referred to as _ preprocessing _ or _ low level analysis _ , are the most platform - dependent tasks . for two - color cdna arrays this includes image analysis ( see e.g. @xcite ) and normalization ( see e.g. @xcite ) . for short oligonucleotide chip data this includes normalization ( see e.g. @xcite ) and the estimation of gene expression values ( see e.g. @xcite and @xcite as well as subsequent papers by these groups ) .
questions around the design of micorarray experiments are mostly relevant for two - color platforms ( see e.g. ch . 2 in @xcite , @xcite and further references there ) .
analysis beyond the preprocessing steps is often referred to as _ downstream analysis .
_ the main goal is to identify genes which act differently in different types of samples .
exploratory methods such as classification and cluster analysis have quickly gained popularity for microarray data analysis . for reviews on such methods from a statistical point of view
see e.g. ch . 2 and ch . 3 in @xcite and ch
. 3 - 7 in @xcite . on the other side of the spectrum ,
hypothesis - driven inferential statistical methods are now well established and used .
this approach typically takes a single - gene perspective in the sense that it searches for _ individual _ genes that are expressed differentially across changing conditions ; see e.g. @xcite .
the main challenge is the imprecision of the gene - specific variance estimate , a problem that has been tackled by strategies incorporating a gene - unspecific component into the estimate ; see e.g. @xcite , @xcite , @xcite and references therein , and @xcite for the case of microarray time course data .
testing thousands of potentially highly correlated genes at the same time with only a few replicates raises a substantial multiple testing problem that has been systematically addressed by various authors incorporating benyamini s and hochberg s _ false discovery rate ( fdr ) _ ; see e.g. @xcite and the review @xcite .
the joint analysis of pre - defined groups of genes based on _ a priori _ knowledge has become an established alternative to the genome - wide exploratory approaches and the gene - by - gene analysis ; see e.g. @xcite and @xcite .
while methodology for microarray data analysis has become a fast growing research area , the epistemological foundation of this research area shows gaps . among other issues ,
@xcite addresses the problem of simultaneous validation of research results and research methods .
@xcite offer a review of the main approaches to microarray data analysis developed so far and attempt to unify them .
many software packages for microarray data analysis have been made publicly available by academic researchers .
in particular , there is the bioconductor project , a community - wide effort to maintain a collection of r - packages for genomics applications at www.bioconductor.org .
many of the main packages are described in @xcite .
data quality is a well established aspect of many quantitative research fields .
the most striking difference between assessing the quality of a measurement as opposed to assessing the quality of a manufactured item is the additional layer of uncertainty .
concerns around the accuracy of measurements have a long tradition in physics and astronomy ; the entire third chapter of the classical book @xcite is devoted to this field .
biometrics , psychometrics , and econometrics developed around similar needs , and many academic fields have grown a strong quantitative branch .
all of them facing data quality questions .
clinical trials is a field that is increasingly aware of the quality of large data collections ( see @xcite and other papers in this special issue ) . with its recent massive move into the quantitative field , functional genomics gave birth to what some statisticians call _ genometrics .
_ we now touch on the major points that characterize gene expression microarray data from the point of view of qa / qc . these points apply to other high - dimensional molecular measurements as well . *
unknown kind of data : * being a new technology in the still unknown terrain of functional genomics , microarrays produce datasets with few known statistical properties , including shape of the distribution , magnitude and variance of the gene expression values , and the kind of correlation between the expression levels of different genes .
this limits access to existing statistical methods .
* simultaneous measurements : * each microarray produces measurements for thousands of genes simultaneously .
if we measured just one gene at a time , some version of shewhart control charts could no doubt monitor quality .
if we measured a small number of genes , multivariate extensions of control charts might be adequate . in a way , the use of control genes is one attempt by biologists to scale down the task to a size that can be managed by these classical approaches .
control genes , however , can not be regarded as typical representatives of the set of all the genes on the arrays .
gene expression measures are correlated because of both the biological interaction of genes , and dependencies caused by the common measurement process .
biologically meaningful correlations between genes can potentially `` contaminate '' hybridization quality assessment .
* multidisciplinary teams : * microarray experiments are typically planned , conducted and evaluated by a team which may include scientists , statisticians , technicians and physicians . in the interdisciplinarity of the data production and handling , they are similar to large datasets in other research areas . for survey data , @xcite names
the risk associated with such a _
`` mlange of workers''_. among other things , he mentions : radically different purposes , lack of communication , disagreements on the priorities among the components of quality and concentration on the `` error of choice '' in their respective discipline .
the encouragement of a close cooperation between scientists and statisticians in the care for measurement quality goes all the way back to @xcite , p.70/71 : _ `` where does the statistician s work begin ? [ ... ] before one turns over any sample of data to the statistician for the purpose of setting tolerances he should first ask the scientist ( or engineer ) to cooperate with the statistician in examining the available evidence of statistical sontrol .
the statistician s work solely as a statistician begins after the scientist has satisfied himself through the application of control criteria that the sample has arisen under statistically controlled conditions . '' _
* systematic errors : * as pointed out by @xcite , and , in the context of clinical trials , by @xcite , systematic errors in large datasets are much more relevant than random errors .
microarrays are typically used in studies involving different experimental or observational groups .
quality differences between the groups are a potential source of confounding . * heterogenous quality in data collections :
* often microarray data from different sources are merged into one data collection .
this includes different batches of chips within the same experiment , data from different laboratories participating in a single collaborative study , or data from different research teams sharing their measurements with the wider community . depending on the circumstances
, the combination of data typically takes place on one of the following levels : raw data , preprocessed data , gene expression summaries , lists of selected genes .
typically , no quality measures are attached to the data .
even if data are exchanged at the level of cel files , heterogeneity can cause problems .
some laboratories filter out chips or reprocess the samples that were hybridized to chips that did not pass screening tests , others do not .
these are decision processes that ideally should take place according to the same criteria .
the nature of this problem is well known in data bank quality or _ data warehousing _ ( see e.g. @xcite , @xcite , @xcite ) . *
re - using of shared data : * gene expression data are usually generated and used to answer a particular set of biological questions .
data are now often being placed on the web to enable the general community to verify the analysis and try alternative approaches to the original biological question .
data may also find a secondary use in answering modified questions .
the shifted focus potentially requires a new round of qa / qc , as precision needs might have changed and artifacts and biases that did not interfere with the original goals of the experiment may do so now . * across - platform comparison : * @xcite , p. 112
, already values the consistency between different measurement methods higher than consistency in repetition . for microarrays ,
consistency between the measurements of two or more platforms ( two - color cdna , long oligonucleotide , short oligonucleotide ( affymetrix ) , commercial cdna ( agilent ) , and real - time pcr ) on rna from the same sample has been addressed in a number of publications .
some of the earlier studies show little or no agreement ( e.g. @xcite , @xcite , @xcite , @xcite ) , while others report mixed results ( e.g.@xcite , @xcite , @xcite ) .
more recent studies improved the agreement between platforms by controlling for other factors . @xcite and @xcite
restrict comparisons to subsets of genes above the noise level .
@xcite use sequence - based matching of probes instead of gene identifier - based matching .
@xcite , @xcite and @xcite use superior preprocessing methods and systematically distinguish the lab effect from the platform effect ; see @xcite and @xcite for detailed reviews and further references .
for affymetrix arrays , @xcite , @xcite and @xcite found inter - laboratory differences to be managable .
however , merging data from different generations of affymetrix arrays is not as straightforward as one might expect ( e.g. @xcite , @xcite , @xcite , @xcite , @xcite ) .
quality assessment for microarray data can be studied on at least seven levels : * the raw chip ( pre - hybridization ) * the sample * the experimental design * the multi - step measurement process * the raw data ( post - hybridization ) * the statistically preprocessed microarray data * the microarray data as entries in a databank the last two items are the main focus of this paper .
the quality of the data after statistical processing ( which includes background adjustment , normalization and probeset summarization ) is greatly affected , but not entirely determined by the quality of the preceeding five aspects .
the raw microarray data ( 5 ) are the result of a multi - step procedure . in the case of the expression
microarrays this includes converting mrna in the sample to cdna , labelling the target mrna via an in vitro transcription step , fragmenting and then hybridizing the resulting crna to the chip , washing and staining , and finally scanning the resulting array .
temperature during storage and hybridization , the amount of sample and mixing during hubridization all have a substantial impact on the quality of the outcome .
seen as a multi - step process ( 4 ) the quality management for microarray experiments has a lot in common with chemical engineering , where numerous interwoven quality indicators have to be integrated ( see e.g. @xcite ) .
the designer of the experiment ( 3 ) aims to minimize the impact of additional experimental conditions ( e.g. hybridization date ) and to maximize accuracy and precision for the quantities having the hightest priority , given the primary objectives of the study .
sample quality ( 2 ) is a topic in its own right , strongly tied to the organism and the institutional setting of the study .
the question how sample quality is related to the microarray data has been investigated in @xcite based on a variety of rna quality measures and chip quality measures including both affymetrix scores and and ours .
the chip before hybridization ( 1 ) is a manufactured item .
the classical theory of quality control for industrial mass production founded by @xcite provides the appropriate framework for the assessment of the chip quality before hybridization .
the affymetrix software gcos presents some chip - wide quality measures in the expression report ( rtp file ) .
they can also be computed by the bioconductor r package simpleaffy described in @xcite .
the document `` qc and affymetrix data '' contained in this package discusses how these metrics can be applied .
the quantities listed below are the most commonly used ones from the affymetrix report ( descriptions and guidelines from @xcite and @xcite ) . while some ranges for the values are suggested , the manuals mainly emphasize the importance of _ consistency _ of the measures within a set of jointly analyzed chips using similar samples and experimental conditions .
the users are also encouraged to look at the scores in conjuction with others scores . * * average background : * average of the lowest 2@xmath1 cell intensities on the chip .
affymetrix does not issue official guidelines , but mentions that values typically range from 20 to 100 for arrays scanned with the genechip scanner 3000 .
a high background indicates the presence of nonspecific binding of salts and cell debris to the array . *
* raw q ( noise ) : * measure of the pixel - to - pixel variation of probe cells on the chip .
the main factors contributing to noise values are electrical noise of the scanner and sample quality .
older recommendations give a range of 1.5 to 3 .
newer sources , however , do not issue official guidelines because of the strong scanner dependence .
they recommend that data acquired from the same scanner be checked for comparability of noise values .
* * percent present : * the percentage of probesets called _ present _ by the affymetrix detection algorithm .
this value depends on multiple factors including cell / tissue type , biological or environmental stimuli , probe array type , and overall quality of rna .
replicate samples should have similar percent present values .
extremely low percent present values indicate poor sample quality .
a general rule of thumb is human and mouse chips typically have 30 - 40 percent present , and yeast and e. coli have 70 - 90 percent present . *
* scale factor : * multiplicative factor applied to the signal values to make the 2@xmath1 trimmed mean of signal values for selected probe sets equal to a constant . for the hu133 chips , the default constant is 500 .
no general recommendation for an acceptable range is given , as the scale factors depend on the constant chosen for the scaling normalization ( depending on user and chip type ) .
* * gapdh 3 to 5 ratio ( gapdh 3/5 ) : * ratio of the intensity of the 3 probe set to the 5 probe set for the gene gapdh .
it is expected to be an indicator of rna quality .
the value should not exceed 3 ( for the 1-cycle assay ) .
* perfect match ( pm ) : * the distribution of the ( raw ) pm values . while we do not think of this as a full quality assessment measure
, it can indicate particular phenomena such as brightness or dimness of the image , or saturation . using this tool in combination with other quality measures , can help in detecting and excluding technological reasons for poor quality .
a convenient way to look at the pm distributions for a number of chips is to use boxplots .
alternatively , the data can be summarized on the chip level by two single values : the median of the pm of all probes on the chip , abbreviated _
med(pm ) , _ and the interquartile range of the pm of all probes on the chip , denoted by _
_ our other assessment tools use probe level and probeset level quantities obtained as a by - product of the robust multichip analysis ( rma ) algorithm developed in @xcite , @xcite and @xcite .
we now recall the basics about rma and refer the reader to above papers for details .
consider a fixed probeset .
let @xmath2 denote the intensity of probe @xmath3 from this probeset on chip @xmath4 usually already background corrected and normalized .
rma is based on the model @xmath5 with @xmath6 a _ probe affinity effect _ , @xmath7 representing the log scale expression level for chip @xmath4 and @xmath8 an i.i.d .
centered error with standard deviation @xmath9 .
for identifiability of the model , we impose a zero - sum constraint on the @xmath10s .
the number of probes in the probeset depends on the kind of chip ( e.g.11 for the hu133 chip ) . for a fixed probeset
, rma robustly fits the model using iteratively weighted least squares and delivers a probeset expression index @xmath11 for each chip .
the analysis produces residuals @xmath12 and weights @xmath13 attached to probe @xmath3 on chip @xmath14 the weights are used in the irls algorithm to achieve robustness .
probe intensities which are discordant with the rest of the probes in the set are deemed less reliable and downweighted .
the collective behaviour of all the weights ( or all the residuals ) on a chip is our starting point in developing post - hybridization chip quality measures .
we begin with a `` geographic '' approach images of the chips that highlight potential poorly performing probes and then continue with the discussion of numerical quality assessment methods .
* quality landscapes : * an image of a hybridized chip can be constructed by shading the positions in a rectangular grid according to the magnitude of the perfect match in the corresponding position on the actual chip .
in the same way , the positions can be colored according to probe - level quantities other than the simple intensities .
a typical color code is to use shades of red for positive residuals and shades of blue for negative ones , with darker shades corresponding to higher absolute values .
shades of green are used for the weights , with darker shades indicating lower weights .
as the weights are in a sense the reciprocals of the absolute residuals , the overall information gained from these two types of quality landscapes is the same . in some particular cases ,
the sign of the residuals can help to detect patterns that otherwise would have been overlooked ( see both fruit fly datasets in sections [ r ] for examples ) .
if no colors are available , gray level images are used .
this has no further implications for the weight landscapes . for the residual landscapes , note that red and blue shades are translated into similar gray levels , so the sign of the residuals is lost .
positive and negative residuals can plotted on two separate images to avoid this problem . *
normalized unscaled standard error ( nuse ) : * fix a probeset .
let @xmath15 be the estimated residual standard deviation in model ( [ rmamodel ] ) and @xmath16 the _ total probe weight _ ( of the fixed probeset ) in chip @xmath14 the expression value estimate for the fixed probeset on chip @xmath4 and its standard error are given by @xmath17 the residual standard deviations vary across the probesets within a chip .
they provide an assessment of overall goodness of fit of the model to probeset data for all chips used to fit the model , but provide no information on the relative precision of estimated expressions across chips .
the latter , however , is our main interest when we look into the quality of a chips compared to other chips in the same experiment .
replacing the @xmath18 by @xmath19 gives what we call the _ unscaled standard error ( use ) _ of the expression estimate .
another source of heterogeneity is the number of `` effective '' probes in the sense of being given substantial weight by the rma fitting procedure . that this number varies across probeset is obvious when different numbers of probes per probeset are used on the same chip
another reason is dysfunctional probes , that is , probes with high variabiliy , low affinity , or a tendency to crosshybridize . to compensate for this kind of heterogeneity
, we divide the use by its median over all chips and call this _ normalized unscaled standard error ( nuse ) .
_ @xmath20 an alternative interpretation for the nuse of a fixed probeset becomes apparent after some arithmetic manipulations .
for any odd number of positive observations @xmath21 we have @xmath22 since the function @xmath23 is monotone . for an even number @xmath24
this identity is it still approximatively true .
( the reason for the slight inaccuracy is that , for an even number , the median is the average between the two data points in the center positions . )
now we can rewrite @xmath25 the total probe weight can also be thought of as an _ effective number of observations _ contributing to the probeset summary for this chip .
its square root serves as the divisor in the standard error of the expression summaries , similarly to the role of @xmath26 in the classical case of the average of @xmath27 independent observations .
this analogy supposes , for heuristic purposes , that the probes are independent ; in fact this is not true due to normalization , probe overlap and other reasons .
the median of the total probe weight over all chips serves as normalization constant . in the form
, we can think of the nuse as the reciprocal of the normalized square root of total probe weight .
the nuse values fluctuate around 1 .
chip quality statements can be made based on the distribution of all the nuse values of one chip .
as with the pm distributions , we can conveniently look at nuse distributions as boxplots , or we can summarize the information on the chip level by two single values : the median of the nuse over all probesets in a particular chip , _
med(nuse ) , _ and the interquartile range of the nuse over all probesets in the chip , _
iqr(nuse ) . _
* relative log expression ( rle ) : * we first need a reference chip .
this is typically the _ median chip _ which is constructed probeset by probeset as the median expression value over all chips in the experiment .
( a computationally constructed reference chips such as this one is sometimes called `` virtual chip '' . ) to compute the rle for a fixed probeset , take the difference of its log expression on the chip to its log expression on the reference chip .
note that the rle is not tied to rma , but can be computed from any expression value summary .
the rle measures how much the measurement of the expression of a particular probeset in a chip deviates from measurements of the same probeset in other chips of the experiment .
again , we can conveniently look at the distributions as boxplots , or we can summarize the information on the chip level by two single values : the median of the rle over all probesets in a particular chip , _
med(rle ) , _ and the interquartile range of the rle over all probesets in the chip , _
iqr(rle ) . _
the latter is a measure of deviation of the chip from the median chip .
a priori this includes both biological and technical variability . in experiments where it can be assumed that @xmath28 iqr(rle ) is a measure of technical variability in that chip
even if biological variability is present for most genes , iqr(rle ) is still a sensitive detector of sources of technical variability that are larger than biological variability .
med(rle ) is a measure of bias . in many experiments
there are reasons to believe that @xmath29 in that case , any deviation of med(rle ) from @xmath30 is an indicator of a bias caused by the technology .
the interpretation of the rle depends on the assumptions ( and ) on the biological variability in the dataset , but it provides a measure that is constructed independently of the quality landscapes and the nuse . for quality assessment , we summarize and visualize the nuse , rle , and pm distributions .
we found series of boxplots to be very a convenient way to glance over sets up to 100 chips .
outlier chips as well as trends over time or pattern related to time can easily be spotted .
for the detection of systematic quality differences related to circumstances of the experiment , or to properties of the sample it is helpful to color the boxes accordingly .
typical coloring would be according to groups of the experiment , sample cohort , lab site , hybridization date , time of the day , a property of the sample ( e.g.time in freezer ) . to quickly review the quality of larger sets of chips , shorter summaries such as the above mentioned median or the interquartile range of pm , nuse and rle .
these single - value summaries at the chip level are also useful for comparing our quality measures to other chip quality scores in scatter plots , or for plotting our quality measures against continuous parameters related to the experiment or the sample .
again , additional use of colors can draw attention to any systematic quality changes due to technical conditions . while the rle is a form of absolute measure of quality ,
the nuse is not .
the nuse has no units .
it is designed to detect differences _ between chips within a batch .
_ however , the magnitudes of these differences have no interpretation beyond the batch of chips analyzed together .
we now describe a way to attach a quality assessment to a set of chips as a whole .
it is based on a common residual factor for a batch of jointly analyzed chips , rma estimates a common residual scale factor .
it enables us to compare quality between different experiments , or between subgroups of chips in one experiment .
it has no meaning for _ single _ chips .
* residual scale factor ( rsf ) : * this is a quality measure for batches of chips .
it does not apply to individual chips , but assesses the quality of batches of chips .
the batches can be a series of experiments or subgroups of one experiment ( defined , e.g.by cohort , experimental conditions , sample properties , or diagnostic groups ) . to compute the rsf , assume the data are background corrected .
as the background correction works on a chip by chip basis it does not matter if the computations were done simultaneously for all batches of chips or individually . for the normalization , however , we need to find one target distribution to which we normalize all the chips in all the batches .
this is important , since the target distribution determines the scale of intensity measures being analyzed .
we then fit the rma model to each batch separately .
the algorithm delivers , for each batch , a vector of the estimated _ residual scales _ for all the probesets .
we can now boxplot them to compare quality between batches of chips .
the median of each is called _ residual scale factor ( rsf ) . _
a vector of residual scales is a heterogeneous set . to remove the heterogeneity ,
we can divide it , probeset by probeset , by the median over the estimated scales from all the batches .
this leads to alternative definitions of the quantities above , which we call _ normalized residual scales _ and _ normalized residual scale factor ( nrsf ) . _
the normalization leads to more discrimination between the batches , but has the drawback of having no units .
software for the computation and visualization of the quality measures and the interpretation of the statistical plots is discussed in @xcite .
the code is publicly available from www.bioconductor.org in the r package affyplm .
note that the implementation of the nuse in affyplm differs slightly from the above formula .
it is based on the `` true '' standard error as it is comes from m - estimation theory instead of the total weights expression in [ m : nuse ] .
however , the difference is small enough not to matter for any of the applications the nuse has in chip quality assessment .
* affymetrix hu95 spike - in experiments : * here 14 human crna fragments corresponding to transcripts known to be absent from rna extracted from pancreas tissue were spiked into aliquots of the hybridization mix at different concentrations , which we call chip - patterns .
the patterns of concentrations from the spike - in crna fragments across the chips form a latin square .
the chip - patterns are denoted by a , b, ... ,s and t , with a, ... ,l occurring just once , and m and q being repeated 4 times each .
chip patterns n , o and p are the same as that of m , while patterns r , s , and t are the same as q. each chip - pattern was hybridized to 3 chips selected from 3 different lots referred to as the l1521 , the l1532 , and the l2353 series . see www.affymetrix.com/support/technical/sample.data/datasets.affx for further details and data download . for this paper
, we are using the data from the 24 chips generated by chip patterns m , n , o , p , q , r , s , t with 3 replicates each . * st .
jude children s research hospital leukemia data collection : * the study by @xcite was conducted to determine whether gene expression profiling could enhance risk assignment for pediatric acute lymphoblastic leukemia ( all ) .
the risk of relapse plays a central role in tailoring therapy intensity .
a total of 389 samples were analyzed for the study , from which high quality gene expression data were obtained on 360 samples .
distinct expression profiles identified each of the prognostically important leukemia subtypes , including t - all , e2a - pbx1 , bcr - abl , tel - aml1 , mll rearrangement , and hyperdiploid@xmath3150 chromosomes . in addition , another all subgroup was identified based on its unique expression profile .
@xcite re - analized 132 cases of pediatric all from the original 327 diagnostic bone marrow aspirates using the higher density u133a and b arrays .
the selection of cases was based on having sufficient numbers of each subtype to build accurate class predictions , rather than reflecting the actual frequency of these groups in the pediatric population .
the follow - up study identified additional marker genes for subtype discrimination , and improved the diagnostic accuracy .
the data of these studies are publicly available as supplementary data .
* fruit fly mutant pilot study : * gene expression of nine fruit fly mutants were screened using affymetrix drosgenome1 arrays .
the mutants are characterized by various forms of dysfunctionality in their synapses .
rna was extracted from fly embryos , pooled and labelled .
three to four replicates per mutant were done .
hybridization took place on six different days . in most cases ,
technical replicates were hybridized on the same day .
the data were collected by tiago magalhes in the goodman lab at the university of california , berkeley , to gain experience with the new microarray technology .
* fruit fly time series : * a large population of wild type ( canton - s ) fruit flies was split into twelve cages and allowed to lay eggs which were transferred into an incubator and aged for 30 minutes . from that time onwards , at the end of each hour for the next 12 hours , embryos from one plate were washed on the plate , dechorionated and frozen in liquid nitrogen .
three independent replicates were done for each time point . as each embryo sample contained a distribution of different ages , we examined the distribution of morphological stage - specific markers in each sample to correlate the time - course windows with the nonlinear scale of embryonic stages .
rna was extracted , pooled , labeled and hybridized to affymetrix drosgenome1 arrays .
hybridization took place on two different days .
this dataset was collected by pavel tomanck in the rubin lab at the university of california , berkeley , as a part of their comprehensive study on spatial and temporal patterns of gene expression in fruit fly development @xcite .
the raw microarray data ( .cel files ) are publically available at the project
s website www.fruitfly.org/cgi-bin/ex/insitu.pl .
* pritzker data collection : * the pritzker neuropsychiatric research consortium uses brains obtained at autopsy from the orange country coroner s office through the brain donor program at the university of california , irvine , department of psychiatry .
rna samples are taken from the left sides of the brains .
labeling of total rna , chip hybridization , and scanning of oligonucleotide microarrays are carried out at independent sites ( university of california , irvine ; university of california , davis ; university of michigan , ann arbor ) .
hybridizations are done on hu95 and later generations of affymetrix chips . in this paper
, we are looking at the quality of data used in two studies by the pritzker consortium . the _ gender study _ by @xcite is motivated by gender difference in prevalence for some neuropsychiatric disorders .
the raw dataset has hu95 chip data on 13 subjects in three regions ( anterior cingulate cortex , dorsolateral prefrontal cortex , and cortex of the cerebellar hemisphere ) .
the _ mood disorder study _ described in @xcite is based on a growing collection of gene expression measurements in , ultimately , 25 regions .
each sample was prepared and then split so that it could be hybridized to the chips in both michigan and either irvine or davis .
we start by illustrating our quality assessment methods on the well known affymetrix spike - in experiments .
the quality of these chips is well above what can be expected from an average lab experiment .
we then proceed with data collected in scientific studies from variety of tissue types and experimental designs .
different aspects of quality analysis methods will be highlighted throughout this section .
our quality analysis results will be compared with the affymetrix quality report for several sections of the large publicly available st .
jude children s research hospital gene expression data collection .
\(a ) * outlier in the affymetrix spike - in experiments : * 24 hu 95a chips from the affymetrix spike - in dataset .
all but the spike - in probesets are expected to be non - differentially expressed across the arrays
. as there are only 14 spike - ins out of about twenty thousand probesets , they are , from the quality assessment point of view , essentially 24 identical hybridizations .
a glance at the weight ( or residual ) landscapes gives a picture of homogenous hybridizations with almost no local defects on any chip but # 20 ( fig .
the nuse indicates that chip # 20 is an outlier .
its median is well above 1.10 , while all others are smaller than 1.05 , and its iqr is three and more times bigger than it is for any other chip ( fig .
the series of boxplots of the rle distributions confirms these findings .
the median is well below @xmath32 and the iqr is two and more times bigger than it is for any other chip .
chip # 20 has both a technologically caused bias and a higher noise level .
the affymetrix quality report ( fig .
a3 ) , however , does not clearly classify # 20 as an outlier .
its gapdh 3/5 of about 2.8 is the largest within this chip set , but the value 2.8 is considered to be acceptable . according to all other affymetrix quality measures percent present , noise , background average , scale factor chip # 20 is within a group of lower quality chips , but does not stand out .
\(b ) * outlier in st .
jude s data not detected by the affymetrix quality report : * the collection of mll hu133b chips consists of 20 chips one of which turns out to be an outlier .
the nuse boxplots ( fig . b1 , bottom line ) show a median over 1.2 for chip # 15 while all others are below 1.025 .
the iqr is much larger for chip # 15 than it is for any other chip .
the rle boxplots ( fig .
b1 , top line ) as well distinguish chip # 15 as an obvious outlier .
the median is about @xmath33 for the outlier chip , while it is very close to @xmath30 for all other chips .
the iqr is about twice as big as the largest of the iqr of the other chips .
b1 displays the weight landscapes of chip#15 along with those of two of the typical chips .
a region on the left side of chip # 15 , covering almost a third of the total area , is strongly down weighted , and the chip has elevated weights overall .
affymetrix quality report ( fig .
b3 ) paints a very different picture
chip @xmath3415 is an outlier on the med(nuse ) scale , but does not stand out on any of common affymetrix quality assessment measures : percent present , noise , scale factor , and gapdh 3/5. \(c ) * overall comparison of our measures and the affymetrix quality report for a large number of st .
jude s chips : * fig .
d1 pairs the med(nuse ) with the four most common gcos scores on a set of 129 hu133a chips from the st .
jude dataset .
there is noticable linear association between med(nuse ) and percent present , as well as between med(nuse ) and scale factor .
gapdh 3/5 does not show a linear association with any the other scores .
\(d ) * disagreement between our quality measures and the affymetrix quality report for hyperdip@xmath3150 subgroup in st .
jude s data : * the affymetrix quality report detects problems with many chips in this dataset . for chip
a , raw q ( noise ) is out of the recommended range for the majority of the chips : @xmath3412 , @xmath3414 , c1 , c13 , c15 , c16 , c18 , c21 , c22 , c23 , c8 and r4 .
background detects chip @xmath3412 as an outlier .
scale factor does not show any clear outliers .
percent present is within the typical range for all chips .
gapdh 3/5 is below 3 for all chips . for chip
b , raw q ( noise ) is out of the recommended range for the @xmath3412 , @xmath348 , @xmath3418 and r4 .
background detects chip @xmath3412 and @xmath348 as outliers .
scale factor does not show any clear outliers .
percent present never exceeds 23@xmath1 in this chip set , and it is below the typical minimum of 20@xmath1 for chips @xmath348 , c15 , c16 , c18 , c21 and c4 .
gapdh 3/5 is satisfactory for all chips .
our measures suggest that , with one exception , the chips are of good quality ( fig .
the heterogeneity of the perfect match distributions does not persist after the preprocessing . for chip
a , @xmath3412 has the largest iqr(rle ) and is a clear outlier among the nuse distributions .
two other chips have elevated iqr(rle ) , but do not stand out according to nuse . for chip
b , the rle distributions are very similar with @xmath3412 again having the largest iqr(rle ) .
the nuse distributions are consistently showing good quality with the exception of chip @xmath3412 .
\(e ) * varying quality between diagnostic subgroups in the st .
jude s data : * each boxplot in fig .
e1 sketches the residual scale factors ( rsf ) of the chips of all diagnostic subgroups .
they show substantial quality differences .
the e2a@xmath35pbx1 subgroup has a much higher med(rsf ) than the other subgroups .
the t@xmath35all subgroup has a slightly elevated med(rsf ) and a higher iqr(rsf ) than the other subgroups .
\(f ) * hybridization date effects on quality of fruit fly chips :
* the fruit fly mutant with dysfunctional synapses is an experiment of the earlier stages of working with affymetrix chips in this lab .
it shows a wide range of quality . in the boxplot series of rle and nuse ( fig .
f1 ) a dependency of the hybridization date is striking . the chips of the two mutants hybridized on the day colored yellow show substantially lower quality than any of the other chips .
f2 shows a weight landscape revealing smooth mountains and valleys .
while the pattern is particularly strong in the chip chosen for this picture , it is quite typical for the chips in this dataset .
we are not sure about the specific technical reason for this , but assume it is related to insufficient mixing during the hybridization .
\(g ) * temporal trends or biological variation in fruit fly time series : * the series consists of 12 developmental stages of fruit fly embryos hybridized in 3 technical replicates each . while the @xmath36(pm ) distributions are very similar in all chips , we can spot two kinds of systematic patterns in the rle and nuse boxplots ( fig .
one pattern is connected to the developmental stage . within each single one of the three repeat time series ,
the hybridizations in the middle stages look `` better '' than the ones in the early stages and the chips in the late stages .
this may , at least to some extent , be due to biological rather than technological variation . in embryo development , especially in the beginning and at the end
, huge numbers of genes are expected to be affected , which is a potential violation of assumption .
insufficient staging in the first very short developmental stages may further increase the variability .
also , in the early and late stages of development , there is substantial doubt about the symmetry assumption .
another systematic trend in this dataset is connected to the repeat series .
the second dozen chips are of poorer quality than the others .
in fact , we learned that they were hybridized on a different day from the rest .
the pairplot in fig .
g2 looks at the relationship between our chip quality measures .
there is no linear association between the raw intensities summarized as med(pm ) and any of the quality measures .
a weak linear association can be noted between med(rle ) and iqr(rle ) .
it is worth to note that is becomes much stronger when focusing on just the chips hybridized on the day colored in black .
iqr(rle ) and med(nuse ) again have a weak linear association which becomes stronger when looking only at one of the subgroups , except this time it is the chips colored in gray . for the pairing med(rle ) and med(nuse ) , however , there is no linear relationship . finally ( not shown ) , as in the dysfunctional synapses mutant fruit fly dataset , a double - wave gradient , as seen in fig .
f2 for the other fruit fly dataset , can be observed in the quality landscapes of many of the chips .
although these experiments were conducted by a different team of researchers , they used the same equipment as that used in generating the other fruit fly dataset .
\(h ) * lab differences in pritzker s gender study : * we looked at hu95 chip data from 13 individuals in two brain regions , the cortex of the cerebellar hemisphere ( short : cerebellum ) and the dorsolateral prefrontal cortex .
with some exceptions , each sample is hybridized in both lab m and lab i. the nuse and rle boxplots ( fig .
h1 ) for the cerebellum dataset display an eye - catching pattern : they show systematically much better quality in lab m then in lab i. this might be caused by overexposure or saturation effects in lab i. the medians of the raw intensities ( pm ) values in lab i are , on a @xmath36-scale between about 9 and 10.5 , while they are very consistently about 2 two 3 points lower in lab m. the dorsolateral prefrontal cortex hybridizations show , for the most part , a lab effect similar to the one we saw in the cerebellum chips ( plots not shown here ) .
\(i ) * lab differences in pritzker s mood disorder study : * after the experiences with lab differences in the gender study , the consortium went through extended efforts to minimize these problems .
in particular , the machines were calibrated by affymetrix specialists .
i1 summarizes the quality assessments of three of the pritzker mood disorder datasets .
we are looking at hu95 chips from two sample cohorts ( a total of about 40 subjects ) in each of the brain regions anterior cingulate cortex , cerebellum , and dorsolateral prefrontal cortex . in terms of med(pm ) , for each of the three brain regions , the two replicates came closer to each other : the difference between the two labs in the mood disorder study is a third or less of the difference between the two labs in the gender study ( see first two boxes in each of the three parts of fig .
i1 , and compare with fig .
this is due to lab i dropping in intensity ( toward lab m ) and the new lab d also operating at that level .
the consequence of the intensity adjustments for chip quality do not form a coherent story . while for cerebellum the quality in lab m is still better than in the replicate in one of the other labs , for the other two brain regions the ranking is reversed .
effects of a slight underexposure in lab m may now have become more visible .
generally , in all brain regions , the quality differences between the two labs are still there , but they are much smaller than they in the gender study data .
\(j ) * assigning special causes of poor quality for st .
jude s data : * eight quality landscapes from the early st .
jude s data , a collection of 335 hu133av2 chips .
the examples were picked for being particularly strong cases of certain kinds of shortcomings that repeatedly occur in this chip collection .
they do not represent the general quality level in the early st .
jude s chips , and even less so the quality of later st . jude s chips .
the figures in this paper are in gray levels . if the positive residual landscape is shown , the negative residual landscape is typically some sort of complementary image , and vice versa .
colored quality landscapes for all st . jude s chips can be downloaded from bolstad s _ chip gallery _ at www.plmimagegallery.bmbolstad.com . fig .
j1 `` bubbles '' is the positive residual landscape of chip hyperdip-50 - 02 .
there are small dots in the left upper part of the slide , and two bigger ones in the middle of the slide .
we attribute the dots to dust attached to the slide or air bubbles stuck in this place during the hybridization .
further , there is an accumulation of positive residuals in the bottom right corner .
areas of elevated residuals near the corners and edges of the slide are very common , often much larger than in this chip .
mostly they are positive .
the most likely explanation are air bubbles that , due to insufficient mixing during the hybridization , got stuck close to the edges where they had gotten when this edges was in a higher position to start with or brought up there by the rotation .
note that there typically is some air in the solution injected into the chip ( through a little hole near one of the edges ) , but that the air is moved around by the rotation during the hybridization to minimize the effects on the probe measurements .
j2 `` circle and stick '' is the residual landscape of hyperdip47 - 50-c17 .
this demonstrates two kinds of spatial patterns that are probably caused by independent technical shortcomings .
first , a circle with equally spaced darker spots ( approximately ) .
the symmetry of the shape suggests it was caused by a foreign object scratching trajectories of the rotation during the hybridization into the slide .
second , there are little dots that almost seem to be aligned along a straight line connecting the circle to the upper right corner .
the dots might be air bubbles stuck to some invisible thin straight object or scratch . fig .
j3 `` sunset '' is the negative residual landscape of hyperdip-50-c6 .
this chip illustrates two independent technical deficiencies .
first , there is a dark disk in the center of the slide .
it might be caused by insufficient mixing , but the sharpness with which the disk is separated from the rest of the image asks for additional explanations .
second , the image obviously splits into an upper and a lower rectangular part with different residual , separated by a straight border . as a most likely explanation ,
we attribute this to scanner problems .
j4 `` pond '' is the negative residual landscape of tel - aml1 - 2m03 .
the nearly centered disc covers almost the entire slide .
it might be caused by the same mechanisms that were responsible for the smaller disc in the previous figure .
however , in this data collection , we have only seen two sizes of discs the small disk as in the previous figure and the large one as in this figure .
this raises the question why the mechanism that causes them does not produce medium size discs .
j5 `` letter s '' is the positive residual landscape of hypodip-2m03 .
the striking pattern the letter s with the `` cloud '' on top is a particularly curious example of a common technical shortcoming .
we attribute the spatially heterogeneous distribution of the residuals to insufficient mixing of the solution during the hybridization .
j6 `` compartments '' is the positive residual landscape of hyperdip-50 - 2m02 .
this is a unique chip .
one explanation would be that the vertical curves separating the three compartments of this image are long thin foreign objects ( e.g. hair ) that got onto the chip and blocked or inhabited the liquid from being spread equally over the entire chip .
j7 `` triangle '' is the positive residual landscape of tel - aml1 - 06 .
the triangle might be caused by a long foreign object stuck to the center of the slide on one end and free , and hence manipulated by the rotation , on the other end .
j8 `` fingerprint '' is the positive residual landscape of hyperdip-50-c10 .
what looks like a fingerprint on the picture might actually be one . with the slide measuring 1 cm by 1 cm , the pattern has about the size of a human fingerprint or the middle part of it .
* quality landscapes : * the pair of the positive and negative residual landscapes contains the maximum information . often , one of the two residual landscapes can already characterize most of the spatial quality issues . in the weight pictures ,
the magnitude of the derivation is preserved , but the sign is lost . therefore , unrelated local defects can appear indistinguishable in weight landscapes .
the landscapes allow a first glance at the overall quality of the array : a square filled with low - level noise typically comes from a good quality chip , one filled with high - level noise comes from a chip with uniformly bad probes . if the landscape is reveals any spatial patterns , the quality may or may not be compromised depending on the size of the problematic area .
even a couple of strong local defects may not lower the chip quality , as indicated by our measures .
the reason lies in both the chip design and the rma model .
the probes belonging to one probeset are scattered around the chip assuring that a bubble or little scratch would only affect a small number of the probes in a probeset ; even a larger under- or overexposed area of the chip may affect only a minority of probes in each probeset . as the rma model is fitted robustly , its expression summaries are shielded against this kind of disturbance .
we found the quality landscape most useful in assigning special causes of poor chip quality .
a quality landscape composed of smooth mountains and valleys is most likely caused by insufficient mixing during the hybridization .
smaller and sharper cut - out areas of elevated residuals are typically related to foreign objects ( dust , hair , etc . ) or
air bubbles .
symmetries can indicate that scratching was caused by particles being rotated during hybridization .
patterns involving horizontal lines may be caused by scanner miscalibration .
it has to be noted , that the above assignment of causes are educated guesses rather than facts .
they are the result of extensive discussions with experimentalist , but there remains a speculative component to them .
even more hypothetical are some ideas we have regarding how the sign of the residual could reveal more about the special cause .
all we can say at this point is that the background corrected and normalized probe intensity deviate from what the fitted rma model would expect them to be .
the focus in this paper is a global one : chip quality .
several authors have worked on spatial chip images from a different perspective , that of automatically detecting and describing local defects ( see @xcite , or the r - package harshlight by @xcite ) .
it remains an open question how to use this kind of assessment beyond the detection and classification of quality problems . in our approach , if we do not see any indication of quality landscpape features in another quality indicator such as nuse or rle , we suppose that it has been rendered harmless by our robust analysis . this may not be true .
* rle : * despite its simplicity the rle distribution turns out to be a powerful quality tool . for a small number of chips , boxplots of the rle distributions of each chip allow the detection of outliers and temporal trends .
the use of colors or gray levels for different experimental conditions or sample properties facilitates the detection of more complex patterns . for a large number of chips , the iqr(rle )
is a convenient and informative univariate summary .
med(rle ) should be monitored as well to detect bias . as seen in the drosophila embryo data , these assumptions are crucial to ensuring that what the rle suggests really are technical artifacts rather than biological differences .
note that the rle is not tied to the rma model , but could as well be computed based on expression values derived from other algorithms .
the results may differ , but our experience is that the quality message turns out to be similar .
* nuse : * as in the case of the rle , boxplots of nuse distributions can be used for small chip sets , and plots of their median and interquartile ranges serve as a less space consuming alternative for larger chip sets . for the nuse ,
however , we often observe a very high correlation between median and interquartile range , so keeping track of just the median will typically suffice .
again , colors or gray levels can be used to indicate experimental conditions facilitating the detection their potential input on quality .
the nuse is the most sensitive of our quality tools , and it does not have a scale
. observed quality differences , even systematic ones , have therefore to carefully assessed .
even large relative differences do not necessarily compromise an experiment , or render useless batches of chips within an experiment .
they should always alert the user to substantial heterogeneity , whose cause needs to be investigated . on this matter
we repeat an obvious but important principle we apply .
when there is uncertainty about whether or not to include a chip or set of chips in an analysis , we can do both analyses and compare the results .
if no great differences result , then sticking with the larger set seems justifiable . * raw intensities and quality measures : * raw intensities are not useful for quality prediction by itself , but they can provide some explanation for the poor performance according to the other quality measures .
all the pritzker datasets , for example , suffer from systematic differences in the location of the pm intensity distribution ( indicated by med(pm ) ) . sometimes the lower location was worse too close to underexposure
and sometimes the higher was worse too close to overexposure or saturation .
we have seen examples of chips for which the raw data give misleading quality assessment .
some kinds of technological shortcomings can be removed without trace by the statististical processing , while others remain .
* comparison with affymetrix quality scores : * we found good overall agreement between our quality assessment and two of the affymetrix scores : percent present and scale factor .
provided the quality in a chips set covers a wide enough range , we typically see at least a weak linear association between our quality measures and these two , and sometimes other affymetrix quality scores .
however , our quality assessment does not always agree with the affymetrix quality report . in the st .
jude data collection we saw that the sensitivity of the affymetrix quality report could be insufficient . while our quality assessment based on rle and nuse clearly detected the outlier chip in the mll chip b dataset , none of the measures in the affymetrix quality did .
the reverse situation occured in the hyperdip chip a dataset .
while most of the chips passed according to our quality measures , most of the chips got a poor affymetrix quality scores .
* rsf : * the residual scale factor can detect quality differences between parts of a data collection , such as the diagnostic subgroups in the st .
jude s data . in the same way
, it can be employed to investigate quality differences between other dataset divisions defined by sample properties , lab site , scanner , hybridization day , or any other experimental condition .
more experience as to what magnitudes of differences are acceptable is still needed .
in this paper , we have laid out a conceptual framework for a statistical approach for the assessment and control of microarray data quality . in particular , we have introduced a quality assessment toolkit for short oligonucleotide arrays .
the tools highlight different aspects in the wide spectrum of potential quality problems .
our numerical quality measures , the nuse and the rle , are an efficient way to detect chips of unusually poor quality . furthermore , they permit the detection of temporal trends and patterns , batch effects , and quality biases related to sample properties or to experimental conditions .
our spatial quality methods , the weight and residual landscapes , add to the understanding of specific causes of poor quality by marking those regions on the chip where defects occur .
furthermore , they illustrate the robustness of the rma algorithm to small local defects .
the rsf quantifies quality differences between batches of chips .
it provides a broader framework for the quality scores of the individual chips in an experiment .
all the quality measures proposed in this paper can be computed based on the raw data using publicly available software .
deriving the quality assessment directly from the statistical model used to compute the expression values is more powerful than basing it on the performance of a particular set of of control probes , because the control probes may not behave in a way that is representative for the whole set of probes on the array .
the model - based approach is also preferable to metrics less directly related to the bulk of the expression values .
some of the affymetrix metrics , for example , are derived from the raw probe values and interpret any artifacts as quality problems , even if they are removed by routine preprocessing steps .
a lesson from the practical examples in this paper is the importance of a well designed experiment .
one of the most typical sources for bias , for example , is an unfortunate systematic connection between hybridization date and groups of the study a link that could have been avoided by better planning .
more research needs to be done on the attribution of specific causes to poor quality measurements . while our quality measures , and , most of all , our quality landscapes , are a rich source for finding specific causes of poor quality , a speculative component remains . to increase the credibility of the diagnoses , systematic quality experiments need to be conducted .
a big step forward is bolstad s _ chip gallery _ at www.plmimagegallery.bmbolstad.com , which collects quality landscapes from affymetrix chip collections , provides details about the experiment and sometimes offers explanations for the technical causes of poor quality .
started as a collection of chip curiosities this website is now growing into a visual encyclopedia for quality assessment .
contributions to the collection , in particular those with known causes of defects , are invited ( anonymous if preferred ) .
further methodological research is needed to explore the use of the spatial quality for statistically `` repairing '' local defects , or making partial use of locally damaged chips .
we are well aware that the range of acceptable values for each quality scores is the burning question for experimentalists .
our quality analysis results with microarray datasets from a variety of scientific studies in section [ r ] show that the question about the right threshold for good chip quality does not have a simple answer yet , at least not as the present level of generality .
thresholds computed for gene expression measurements in fruit fly mutant screenings can not necessarily be transferred to brain disease research or to leukemia diagnosis .
thresholds need to be calibrated to the tissue type , the design , and the precision needs of the field of application .
we offer two strategies to deal with this on different levels : 1 . *
individual researchers : * we encouraged researchers to look for outliers and artificial patterns in the series of quality measures of the batch of jointly analyzed chips . furthermore , any other form of unusual observations
e.g.a systematic disagreement between nuse and rle , or inconsistencies in the association between raw intensities and quality measures potentially hints at a quality problem in the experiment .
* community of microarray users : * we recommend the development of quality guidelines .
they should be rooted in extended collections of datasets from scientific experiments .
complete raw datasets are ideal , where no prior quality screening has been employed .
careful documentation of the experimental conditions and properties help to link unusual patterns in the quality measures to specific causes .
the sharing of unfiltered raw chip data from scientific experiments on the web and the inclusion of chip quality scores in gene expression databank entries can lead the way towards community - wide quality standards . besides , it contributes to a better understanding how quality measures relate to special causes of poor quality .
in addition , we encourage the conduction of _ designed microarray quality experiments . _
such experiments aim at an understanding of the effects of rna amount , experimental conditions , sample properties and sample handling on the quality measures as well as on the downstream analysis .
they give an idea of the range of chip quality to be expected under given certain experimental , and , again , they help to characterize specific causes of poor quality .
benchmarking of microarray data quality will happen one way or another ; if it is not established by community - wide agreements the individual experimentalist will resort to judging on the basis of anecdotal evidence .
we recommend that benchmarks be actively developed by the community of microarray researchers , experimentalists and data analysts .
the statistical concepts and methods proposed here may serve as a foundation for the quality benchmarking process .
we thank st .
jude children s research hospital , the pritzker consortium , tiago magalhes , pavel tomanck and affymetrix for sharing their data for quality assessment purposes . 99 natexlab#1#1 affymetrix ( 2001 ) , _ guidelines for assessing data quality _ ,
affymetrix inc , santa clara , ca .
alizadeh , a. a. , eisen , m. b. , davis , r. e. , ma , c. , lossos , i. s. , rosenwald , a. , boldrick , j. c. , sabet , h. , tran , t. , yu , x. , powell , j. i. , yang , l. , marti , g. e. , moore , t. , hudson , j. , lu , l. , lewis , d. b. , tibshirani , r. , sherlock , g. , chan , w. c. , greiner , t. c. , weisenburger , d. d. , armitage , j. o. , warnke , r. , levy , r. , wilson , w. , grever , m. r. , byrd , j. c. , botstein , d. , brown , p. o. , and staudt , l. m. ( 2000 ) , `` distinct types of diffuse large b - cell lymphoma identified by gene expression profiling , '' _ nature _ , 403 , 503511 .
allison , d. , cui , x. , page , g. , and sabripour , m. ( 2006 ) , `` microarray data analysis : from disarray to consolidation and consensus , '' _ nature review genetics _
, 7 , 5565 .
archer , k. , dumur , c. , joel , s. , and ramakrishnan , v. ( 2006 ) , `` assessing quality of hybridized rna in affymetrix genechip experiments using mixed effects models , '' _ biostatistics _ , 7 , 198212 .
barczak , a. , rodriguez , m. , hanspers , k. , koth , l. , tai , y. , bolstad , b. , speed , t. , and erle , d. ( 2003 ) , `` spotted long oligonucleotide arrays for human gene expression analysis , '' _ genome research _ , 1 , 17751785 .
beissbarth , t. , fellenberg , k. , brors , b. , arribas - prat , r. , boer , j. , hauser , n. c. , scheideler , m. , hoheisel , j. d. , schutz , g. , poustka , a. , and vingron , m. ( 2000 ) , `` processing and quality control of dna array hybridization data , '' _ bioinformatics _ , 16 , 10141022 .
bild , a. , yao , g. , chang , j. , wang , q. , potti , a. , chasse , d. , joshi , m. , harpole , d. , lancaster , j. , berchuck , a. , olson , j. j. , marks , j. , dressman , h. , west , m. , and nevins , j. ( 2006 ) , `` oncogenic pathway signatures in human cancers as a guide to targeted therapies , '' _ nature _ , 439(7074 ) , 353357 .
bolstad , b. ( 2003 ) , `` low level analysis of high - density oligonucleotide array data : background , normalization and summarization , '' ph.d .
thesis , university of california , berkeley , http://bmbolstad.com .
bolstad , b. , collin , f. , brettschneider , j. , simpson , k. , cope , l. , irizarry , r. , and speed , t. ( 2005 ) , `` quality assessment of affymetrix genechip data , '' in _ bioinformatics and computational biology solutions using r and bioconductor _ , eds .
gentleman , r. , carey , v. , huber , w. , irizarry , r. , and dudoit , s. , new york : springer , statistics for biology and health , pp .
bolstad , b. , collin , f. , simpson , k. , irizarry , r. , and speed , t. ( 2004 ) , `` experimental design and low - level analysis of microarray data , '' _ int rev neurobiol _
, 60 , 2558 .
bolstad , b. , irizarry , r. , astrand , m. , and speed , t. ( 2003 ) , `` a comparison of normalization methods for high density oligonucleotide array data based on variance and bias , '' _ bioinformatics _ , 19 , 185193 , evaluation studies .
buness , a. , huber , w. , steiner , k. , sultmann , h. , and poustka , a. ( 2005 ) , `` arraymagic : two - colour cdna microarray quality control and preprocessing , '' _ bioinformatics _ , 21 , 554556 .
bunney , w. , bunney , b. , vawter , m. , tomita , h. , li , j. , evans , s. , choudary , p. , myers , r. , jones , e. , watson , s. , and akil , h. ( 2003 ) , `` microarray technology : a review of new strategies to discover candidate vulnerability genes in psychiatric disorders , '' _ am j psychiatry _ , 160 , 657666 .
cho , r. , campbell , m. , winzeler , e. , steinmetz , l. , conway , a. , wodicka , l. , wolfsberg , t. , gabrielian , a. , landsman , d. , lockhart , d. , and davis , r. ( 1998 ) , `` a genome - wide transcriptional analysis of the mitotic cell cycle , '' _ mol cell _ , jul 2(1 ) , 6573 .
cui , x. , hwang , j. , qiu , j. , blades , n. , and churchill , g. ( 2005 ) , `` improved statistical tests for differential gene expression by shrinking variance components estimates , '' _ biostatistics _ , 6 , 3146 .
dobbin , k. , beer , d. , meyerson , m. , yeatman , t. , gerald , w. , jacobson , w. , conley , b. , buetow , k. , heiskanen , m. , simon , r. , minna , j. , girard , l. , misek , d. , taylor , j. , hanash , s. , naoki , k. , hayes , d. , ladd - acosta , c. , enkemann , s. , viale , a. , and giordano , t. ( 2005 ) , `` interlaboratory comparability study of cancer gene expression analysis using oligonucleotide microarrays , '' _ clin cancer res _ , 11 , 565572 .
draghici , s. , khatri , p. , eklund , a. , and szallasi , z. ( 2006 ) , `` reliability and reproducibility issues in dna microarray measurements , '' _ trends in genetics _ , 22 ( 2 ) , 101109 .
dudoit , s. , shaffer , j. , and boldrick , j. ( 2003 ) , `` multiple hypothesis testing in microarray experiments , '' _ statistical science _ , 18 , 71103 .
dudoit , s. , yang , y. , speed , t. , and mj , c. ( 2002 ) , `` statistical methods for identifying differentially expressed genes in replicated cdna microarray experiments , '' _ statistica sinica _
, 12 , 111139 .
dumur , c. , nasim , s. , best , a. , archer , k. , ladd , a. , mas , v. , wilkinson , d. , garrett , c. , and ferreira - gonzalez , a. ( 2004 ) , `` evaluation of quality - control criteria for microarray gene expression analysis , '' _ clin chem _ , 50 , 19942002 .
efron , b. , tibshirani , r. , storey , j. , and v , t. ( 2001 ) , `` empirical bayes analysis of a microarray experiment , '' _ j am stat ass _ , 96 , 11511160 .
finkelstein , d. ( 2005 ) , `` trends in the quality of data from 5168 oligonucleotide microarrays from a single facility , '' _ j biomol tech _ , 16 , 143153 . finkelstein , d. , ewing , r. , gollub , j. , sterky , f. , cherry , j. , and somerville , s. ( 2002 ) , `` microarray data quality analysis : lessons from the afgc project , '' _ plant molecular biology _ , 48 , 119131 .
gassman , j. , owen , w. , kuntz , t. , martin , j. , and amoroso , w. ( 1995 ) , `` data quality assurance , monitoring , and reporting , '' _ controlled clinical trials _ , 16(2 suppl ) , 104s136s .
gautier , l. , cope , l. , bolstad , b. , and irizarry , r. ( 2004 ) , `` affy - analysis of affymetrix genechip data at the probe level , '' _ bioinformatics _ , 20(3 ) , 307315 .
gautier , l. , moller , m. , friis - hansen , l. , and knudsen , s. ( 2004 ) , `` alternative mapping of probes to genes for affymetrix chips , '' _ bmc bioinformatics _
, 5 , e111 . gcos ( 2004 ) , _ genechip expression analysis data analysis fundamentals _ , affymetrix , inc , santa clara , ca .
gentleman , r. , carey , v. , huber , w. , irizarry , r. , and dudoit , s. e. ( 2005 ) , _ bioinformatics and computational biology solutions using r and bioconductor _ , springer .
groves , r. ( 1987 ) , `` research on survey data quality , '' _ public opinion quaterly _ , 51 , part 2 , suppl .
, s156s172 .
hautaniemi , s. , edgren , h. , vesanen , p. , wolf , m. , jrvinen , a. , yli - harja , o. , astola , j. , kallioniemi , o. , and monni , o. ( 2003 ) , `` a novel strategy for microarray quality control using bayesian networks , '' _ bioinformatics _ , 19 , 20312038 .
hoheisel , j. ( 2006 ) , `` microarray technology : beyond transcript profiling and genotype analysis , '' _ nature review genetics _ , 7 ( 3 ) , 200210 .
hu , p. , greenwood , c. , and beyene , j. ( 2005 ) , `` integrative analysis of multiple gene expression profiles with quality - adjusted effect size models , '' _ bmc bioinformatics _
, 6 , e128 .
hwang , k. , kong , s. , greenberg , s. , and park , p. ( 2004 ) , `` combining gene expression data from different generations of oligonucleotide arrays , '' _ bmc bioinformatics _
, 5 , e159 .
irizarry , r. , bolstad , b. , collin , f. , cope , l. , hobbs , b. , and speed , t. ( 2003 ) , `` summaries of affymetrix genechip probe level data , '' _ nucleic acids res _
, 31 , e15 .
irizarry , r. , warren , d. , spencer , f. , kim , i. , biswal , s. , frank , b. , gabrielson , e. , garcia , j. , geoghegan , j. , germino , g. , griffin , c. , hilmer , s. , hoffman , e. , jedlicka , a. , kawasaki , e. , martinez - murillo , f. , morsberger , l. , lee , h. , petersen , d. , quackenbush , j. , scott , a. , wilson , m. , yang , y. , ye , s. , and yu , w. ( 2005 ) , `` multiple - laboratory comparison of microarray platforms , '' _ nat methods _ , 2 , 345350 .
jarvinen , a .- k . ,
hautaniemi , s. , edgren , h. , auvinen , p. , saarela , j. , kallioniemi , o .- p . , and monni , o. ( 2004 ) , `` are data from different gene expression microarray platforms comparable ? '' _ genomics _ , 83 , 11641168 .
jones , l. , goldstein , d. , hughes , g. , strand , a. , collin , f. , dunnett , s. , kooperberg , c. , aragaki , a. , olson , j. , augood , s. , faull , r. , luthi - carter , r. , moskvina , v. , and hodges , a. ( 2006 ) , `` assessment of the relationship between pre - chip and post - chip variables for affymetrix genechip expression data , '' _ bmc bioinformatics _ , 7 , e211 .
kerr , m. ( 2003 ) , `` design considerations for efficient and effective microarray studies , '' _ biometrica _ , 59 , 822828 .
kluger , y. , yu , h. , qian , j. , and gerstein , m. ( 2003 ) , `` relationship between gene co - expression and probe localization on microarray slides , '' _ bmc genomics _ , 4 , e49 .
kong , s. , hwang , k. , kim , r. , zhang , b. , greenberg , s. , kohane , i. , and park , p. ( 2005 ) , `` crosschip : a system supporting comparative analysis of different generations of affymetrix arrays , '' _ bioinformatics _ , 21 , 21162117 .
kuo , w. , jenssen , t. , butte , a. , ohno - machado , l. , and kohane , i. ( 2002 ) , `` analysis of matched mrna measurements from two different microarray technologies , '' _ bioinformatics _ , 18 , 405412 .
li , c. and wong , h. ( 2001 ) , `` model - based analysis of oligonucleotide arrays : expression index computation and outlier detection , '' _ pnas _ , 98 , 3136 .
lipshutz , r. , fodor , s. , gingeras , t. , and lockhart , d. ( 1999 ) , `` high density synthetic oligonucleotide arrays , '' _ nat genet _ , 21 , 2024 .
lockhart , d. , dong , h. , byrne , m. , follettie , m. , gallo , m. , chee , m. , mittmann , m. , wang , c. , kobayashi , m. , horton , h. , and brown , e. ( 1996 ) , `` expression monitoring by hybridization to high - density oligonucleotide arrays , '' _ nat biotechnol _ , 14 , 16751680 .
loebl , a. ( 1990 ) , `` accuracy and relevance and the quality of data , '' in _ data quality control , theory and pragmatics _ , eds .
liepins , g. and uppuluri , v. , new york : marcel dekker , inc .
112 in statistics : textbooks and monographs , pp . 105144 .
lnnstedt , i. and speed , t. ( 2002 ) , `` replicated microarray data , '' _ statistica sinica _
, 12 ( 1 ) , 3146 .
marinez , y. , mcmahan , c. , barnwell , g. , and wigodsky , h. ( 1984 ) , `` ensuring data quality in medical research through an integrated data management system , '' _ stat med _ , 3 , 101111 .
mason , r. and young , j. ( 2002 ) , _ multivariate statistical process control with industrial applications _ , philadelphia , pennsylvania : asa - siam .
mclachlan , g. , do , k. , and ambroise , c. ( 2004 ) , _ analyzing microarray gene expression data _ , hoboken , new jersey : wiley .
mecham , b. , klus , g. , strovel , j. , augustus , m. , byrne , d. , bozso , p. , wetmore , d. , mariani , t. , kohane , i. , and szallasi , z. ( 2004 ) , `` sequence - matched probes produce increased cross - platform consistency and more reproducible biological results in microarray - based gene expression measurements , '' _ nucleic acids res _ , 32 , 74 .
mehta , t. , tanik , m. , and allison , d. ( 2004 ) , `` towards sound epistemological foundations of statistical methods for high - dimensional biology , '' _ nature genetics _
, 36 , 943947 .
mitchell , s. , brown , k. , henry , m. , mintz , m. , catchpoole , d. , lafleur , b. , and stephan , d. ( 2004 ) , `` inter - platform comparability of microarrays in acute lymphoblastic leukemia , '' _ bmc genomics _
, 5 , e71 . model , f. , knig , t. , piepenbrock , c. , and adorjan , p. ( 2002 ) , `` statistical process control for large scale microarray experiments , '' _ bioinformatics _ , 18 suppl 1 , 155163 , evaluation studies .
morris , j. , yin , g. , baggerly , k. , wu , c. , and zhang , l. ( 2004 ) , `` pooling information across different studies and oligonucleotide microarray chip types to identify prognostic genes for lung cancer , '' in _ methods of microarray data analysis iii _ , eds . shoemaker , j. and lin , s. , new york : springer , pp .
naef , f. , socci , n. , and magnasco , m. ( 2003 ) , `` a study of accuracy and precision in oligonucleotide arrays : extracting more signal at large concentrations , '' _ bioinformatics _ , 19 , 178184 .
nimgaonkar , a. , sanoudou , d. , butte , a. , haslett , j. , kunkel , l. , beggs , a. , and kohane , i. ( 2003 ) , `` reproducibility of gene expression across generations of affymetrix microarrays , '' _ bmc bioinformatics _ , 4 , e27 .
novikov , e. and barillot , e. ( 2005 ) , `` an algorithm for automatic evaluation of the spot quality in two - color dna microarray experiments , '' _ bmc bioinformatics _ , 6 , e293 .
perou , c. , sorlie , t. , eisen , m. , van de rijn , m. , jeffrey , s. , rees , c. , pollack , j. , ross , d. , johnsen , h. , akslen , l. , fluge , o. , pergamenschikov , a. , williams , c. , zhu , s. , lonning , p. , borresen - dale , a. , brown , p. , and botstein , d. ( 2000 ) , `` molecular portraits of human breast tumours , '' _ nature _ , 406 , 747752 .
qian , j. , kluger , y. , yu , h. , and gerstein , m. ( 2003 ) , `` identification and correction of spurious spatial correlations in microarray data , '' _ biotechniques _ , 35 , 4244 , evaluation studies .
ramaswamy , s. and golub , t. ( 2002 ) , `` dna microarrays in clinical oncology , '' _ j. clin
_ , 20 , 19321941 .
redman , t. ( 1992 ) , _ data quality : management and technology _
, new york : bantam books .
reimers , m. and weinstein , j. ( 2005 ) , `` quality assessment of microarrays : visualization of spatial artifacts and quantitation of regional biases , '' _ bmc bioinformatics _ , 6 , e166 .
ritchie , m. , diyagama , d. , neilson , j. , van laar , r. , a , d. , a , h. , and gk , s. ( 2006 ) , `` empirical array quality weights in the analysis of microarray data , '' _ bmc bioinformatics _
, 7 , e261 .
rogojina , a. , orr , w. , song , b. , and geisert , e. j. ( 2003 ) , `` comparing the use of affymetrix to spotted oligonucleotide microarrays using two retinal pigment epithelium cell lines , '' _ mol vis _ , 9 , 482496 .
ross , m. , zhou , x. , song , g. , shurtleff , s. , girtman , k. , williams , w. , liu , h. , mahfouz , r. , raimondi , s. , lenny , n. , patel , a. , and downing , j. ( 2003 ) , `` classification of pediatric acute lymphoblastic leukemia by gene expression profiling , '' _ blood _ , 102 , 29512959 .
schoor , o. , weinschenk , t. , hennenlotter , j. , corvin , s. , stenzl , a. , rammensee , h. , and stevanovic , s. ( 2003 ) , `` moderate degradation does not preclude microarray analysis of small amounts of rna , '' _ biotechniques _
, 35 , 11921196 .
shewhart , w. ( 1939 ) , _ statistical method from the viewpoint of quality control _ ,
lanceser , pennsylvania : lancester press , inc .
shippy , r. , sendera , t. , lockner , r. , palaniappan , c. , kaysser - kranich , t. , watts , g. , and j , a. ( 2004 ) , `` performance evaluation of commercial short - oligonucleotide microarrays and the impact of noise in making cross - platform correlations , '' _ bmc genomics _ , 5 , e61 .
smith , k. and hallett , m. ( 2004 ) , `` towards quality control for dna microarrays , '' _ j comput biol _ , 11 , 945970 .
smyth , g. ( 2002 ) , `` print - order normalization of cdna microarrays , '' tech .
rep . , genetics and bioinformatics , walter and eliza hall institute of medical research , melbourne , available at www.statsci.org/smyth/pubs/porder/porder.html .
smyth , g. , yang , h. , and speed , t. ( 2003 ) , `` statistical issues in cdna microarray data analysis , '' in _ functional genomics , methods and protocols _ ,
brownstein , m. j. and khodursky , a. b. , totowa , new jersey : humana press , no . 224 in methods in molecular biology ,
111136 .
speed , t. ( 2003 ) , _ statistical analysis of gene expression of gene expression microarray data _ , boca raton ,
florida : chapman and hall / crc .
stevens , j. and doerge , r. ( 2005 ) , `` combining affymetrix microarray results , '' _ bmc bioinformatics _ , 6 , e57 .
storey , j. ( 2003 ) , `` the positive false discovery rate : a bayesian interpretation and the q - value , '' _ annals of statistics _ , 31 , 20132035 .
surez - farias , m. , haider , a. , and wittkowski , k. ( 2005 ) , `` `` harshlighting '' small blemishes on microarrays , '' _ bmc bioinformatics _ , 6 , e65 .
subramanian , a. , tamayo , p. , mootha , v. , mukherjee , s. , ebert , b. , gillette , m. , paulovich , a. , pomeroy , s. , golub , t. , lander , e. , and mesirov , j. ( 2005 ) , `` gene set enrichment analysis : a knowledge - based approach for interpreting genomewide expression profiles , '' _ pnas _ , 102 , 15554550 .
tai , y. and speed , t. ( 2006 ) , `` a multivariate empirical bayes statstic for replicated microarray time course data , '' _ ann statist _ , 34 , 23872412 .
thach , d. , lin , b. , walter , e. , kruzelock , r. , rowley , r. , tibbetts , c. , and stenger , d. ( 2003 ) , `` assessment of two methods for handling blood in collection tubes with rna stabilizing agent for surveillance of gene expression profiles with high density microarrays , '' _ j immunol methods _ , 283 , 269279 .
the chipping forecast ( 1999 ) , _ the chipping forecast i _ , vol . 21 - 1s
, nature genetics suppl .
( 2002 ) , _ the chipping forecast ii _ , vol .
32 - 4s , nature genetics suppl . ( 2005 ) , _ the chipping forecast iii _
37 - 6s , nature genetics suppl .
thompson , k. , rosenzweig , b. , pine , p. , retief , j. , turpaz , y. , afshari , c. , hamadeh , h. , damore , m. , boedigheimer , m. , blomme , e. , ciurlionis , r. , waring , j. , fuscoe , j. , paules , r. , tucker , c. , fare , t. , coffey , e. , he , y. , collins , p. , jarnagin , k. , fujimoto , s. , ganter , b. , kiser , g. , kaysser - kranich , t. , sina , j. , and sistare , f. ( 2005 ) , `` use of a mixed tissue rna design for performance assessments on multiple microarray formats , '' _ nucleic acid research _ , 33 ( 2 ) , e187 .
tom , b. , gilks , w. , brooke - powell , e. , and ajioka , j. ( 2005 ) , `` quality determination and the repair of poor quality spots in array experiments , '' _ bmc bioinformatics _
, 6 , e234 .
tomanck , p. , beaton , a. , weiszmann , r. , kwan , e. , shu , s. , lewis , s. , richards , s. , ashburner , m. , hartenstein , v. , celniker , s. , and rubin , g. ( 2002 ) , `` systematic determination of patterns of gene expression during drosophila embryogenesis , '' _ genome biol _ , 3 , 114 .
vawter , m. , evans , s. , choudary , p. , tomita , h. , meador - woodruff , j. , molnar , m. , li , j. , lopez , j. , myers , r. , cox , d. , watson , s. , akil , h. , jones , e. , and bunney , w. ( 2004 ) , `` gender - specific gene expression in post - mortem human brain : localization to sex chromosomes , '' _ neuropsychopharmacology _ , 29 , 373384 .
wang , h. , he , x. , band , m. , and wilson , cand liu , l. ( 2005 ) , `` a study of inter - lab and inter - platform agreement of dna microarray data , '' _ bmc genomics _ , 6 , e71 .
wang , r. ( 2001 ) , _ data quality _ , boston : kluwer academic publishers .
wang , r. , storey , v. , and firth , c. ( 1995 ) , `` a framework for analysis of data quality research , '' _ ieee transactions of knowledge and data engineering _ , 7 , 623640 .
wang , x. , ghosh , s. , and guo , s. ( 2001 ) , `` quantitative quality control in microarray image processing and data acquisition , '' _ nucleic acids res _ , 29 ( 15 ) ,
wilson , c. and miller , c. ( 2005 ) , `` simpleaffy : a bioconductor package for affymetrix quality control and data analysis , '' _ bioinformatics _ , 21 , 36833685 . wit , e. and mcclure , j. ( 2004 ) , _ statistics for microarrays : design , analysis , inference _ , hoboken , new jersey : wiley .
woo , y. , affourtit , j. , daigle , s. , viale , a. , johnson , k. , naggert , j. , and churchill , g. ( 2004 ) , `` a comparison of cdna , oligonucleotide , and affymetrix genechip gene expression microarray platforms , '' _ j biomol tech _ , 15 , 276284 .
yang , y. , buckley , m. , and speed , t. ( 2001 ) , `` analysis of cdna microarray images , '' _ brief bioinform _ , 2 , 341349 .
yang , y. , dudoit , s. , luu , p. , lin , d. , peng , v. , ngai , j. , and speed , t. ( 2002 ) , `` normalization for cdna microarray data : a robust composite method addressing single and multiple slide systematic variation , '' _ nucleic acids res _
, 30 , e15 .
yauk , c. , berndt , m. , williams , a. , and douglas , g. ( 2004 ) , `` comprehensive comparison of six microarray technologies , '' _ nucleic acid research _ , 32 ( 15 ) , e124 .
yeoh , e. , ross , m. , shurtleff , s. , williams , w. , patel , d. , mahfouz , r. , behm , f. , raimondi , s. , relling , m. , patel , a. , cheng , c. , campana , d. , wilkins , d. , zhou , x. , li , j. , liu , h. , pui , c. , evans , w. , naeve , c. , wong , l. , and downing , j. ( 2002 ) , `` classification , subtype discovery , and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling , '' _ cancer cell _ , 1 , 133143 .
yuen , t. , wurmbach , e. , pfeffer , r. , ebersole , b. , and sealfon , s. ( 2002 ) , `` accuracy and calibration of commercial oligonucleotide and custom cdna microarrays , '' _ nucleic acids res _
, 30 , e48 .
zhang , w. , shmulevich , i. , and astola , j. ( 2004 ) , _ microarray quality control _ ,
hoboken , new jersey : john wiley & sons , inc .
zhu , b. , ping , g. , shinohara , y. , zhang , y. , and baba , y. ( 2005 ) , `` comparison of gene expression measurements from cdna and 60-mer oligonucleotide microarrays , '' _ genomics _ , 85 , 657665 . | quality of microarray gene expression data has emerged as a new research topic . as in other areas ,
microarray quality is assessed by comparing suitable numerical summaries across microarrays , so that outliers and trends can be visualized , and poor quality arrays or variable quality sets of arrays can be identified .
since each single array comprises tens or hundreds of thousands of measurements , the challenge is to find numerical summaries which can be used to make accurate quality calls . to this end ,
several new quality measures are introduced based on probe level and probeset level information , all obtained as a by - product of the low - level analysis algorithms rma / fitplm for affymetrix genechips .
quality landscapes spatially localize chip or hybridization problems .
numerical chip quality measures are derived from the distributions of _ normalized unscaled standard errors _ and of _ relative log expressions . _
quality of chip batches is assessed by _ residual scale factors . _
these quality assessment measures are demonstrated on a variety of datasets ( spike - in experiments , small lab experiments , multi - site studies ) .
they are compared with affymetrix s individual chip quality report . *
_ to be published in technometrics ( with discussion ) _ * * julia brettschneider @xmath0 , franois collin , benjamin m.bolstad , terence p.speed * * quality assessment for * + + keywords : quality control , microarrays , affymetrix chips , relative log expression , normalized unscaled standard errors , residual scale factors . |
the author acknowledges the brazilian agencies funpe and finatec for partial support , and an anonymous referee for improvements .
99 unruh w g 1976 _ phys .
* 14 * 870 ; davies p c w 1975 _ j. phys . a _ * 8 * 609 ; fulling s a 1973 _ phys . rev
. d _ * 10 * 2850 dewitt b s 1979 in _ general relativity _
s. w. hawking and w. israel ( cambridge university press ) p. 680
audretsch j and mller r 1994 _ phys . rev . a _ * 50 * 1755 dalibard j , dupont - roc j. and cohen - tannoudji c. 1982 _ j. phys . _ ( paris ) * 43 * 1617 takagi s 1988 _ prog .
phys . _ * 88 * 1 unruh w g and wald r m 1984 _ phys
* 29 * 1047 bell j s and leinaas j m 1983 _ nucl .
phys . _ * b212 * 131 ; leinaas j m 2001 _ preprint _ hep - th/0101054 audretsch j , mller j r and holzmann m 1995 _ class .
_ * 12 * 2927 de lorenci v a and svaiter n f 1999 _ found .
_ * 29 * 1233 ; de lorenci v a , de paola r d m and svaiter n f 2000 _ class .
* 17 * 4241 trocheries m g 1949 _ phyl . mag .
_ * 40 * 1143 ; takeno h 1952 _ prog .
* 7 * 367 davies p c w , dray t and manogue c a 1996 _ phys .
d _ * 53 * 4382 sciama d w , candelas p and deutsch d 1981 _ adv .
phys . _ * 30 * 327 whittaker e t and watson g n 1963 _ a course of modern analysis _ , cambridge at the university press , p. 369 . stefani h 1993 _ general relativity : an introduction to the theory of the gravitational field _ , cambridge university press , p. 80 - 83 . | the circular noise is important in connection to mach s principle , and also as a possible probe of the unruh effect . in this letter
the power spectrum of the detector following the trocheries - takeno motion in the minkowski vacuum is analytically obtained in the form of an infinite series .
a mean distribution function and corresponding energy density are obtained for this particular detected noise .
the analogous of a non constant temperature distribution is obtained . and
in the end , a brief discussion about the equilibrium configuration is given .
key words : circular unruh effect the non inertial vacuum , and the noise associated to it , has been extensively studied .
it is called fulling - davies - unruh effect , after the discoverers @xcite .
the more simple situation of proper constant acceleration , is already very well understood .
for instance , a free scalar field can be quantized in rindler coordinates .
the vacuum obtained in these coordinates is unitary inequivalent to the minkowski vacuum .
the monopole type detector introduced by dewitt @xcite , can give a meaning to these vacua .
a detector at rest in rindler coordinates is equivalent to an accelerated one in minkowski space . as a consequence of poincar invariance , a detector in inertial motion in vacuum ,
does not get excited .
this can also be understood in the detailed balance context of @xcite .
the authors used the ddc @xcite formalism to obtain the einstein coefficients for spontaneous excitation and emition . for non inertial motion
the balance of the coefficients is upset , resulting in the unruh effect .
the thermal character of the noise in the kms sense , and the verification of other consistencies are reviewed in @xcite .
in @xcite it is shown the connection between the absorptions of particles and the noise to which the detector is submitted .
the rotational motion is different . besides its own importance connected to the concept of inertia ,
this motion is of experimental relevance , because it provides a possible verification of the unruh effect @xcite .
the ddc formalism was also used in the context of rotation to obtain the spontaneous excitation of the detector , the circular unruh effect @xcite . in the very interesting work @xcite ,
de lorenci and svaiter develop the quantization in trocheries - takeno @xcite coordinates , resulting in a non trivial vacuum .
davies _
et al _ also obtain a very interesting result @xcite concerning rotational motion . since now , the de lorenci - svaiter - trocheries - takeno vacuum is the only one available for rotation .
i do not attempt to enter into the vacuum question . in this letter , an analytical expression for the power spectrum in the form of an infinite series
is obtained , a result which has not yet appeared elsewhere . in @xcite the transition rate obtained here , is left over as an integral .
i suppose a detector in minkowski vacuum , constrained to a trocheries - takeno type motion . it should be stressed that by following the procedure outlined in this letter , inside the light cylinder of @xcite , results in a transition rate similar to equation ( [ eq3 ] ) below . it is assumed a monopole type coupling of the detector and field , @xmath0 where @xmath1 has energy levels .
we require both the detector and the field to obey schrdinger s unitary time evolution .
let us assume that in the very early past the detector is in the ground state with energy @xmath2 , @xmath3 and that the field is in the ( minkowski ) vacuum state @xmath4 .
as the detector moves , field and detector will undergo transitions to various states , in a trajectory dependent fashion @xcite .
since one does not care about the field s final state , the unknown final state of the field should be traced out from the transition probability for thetransitions in question .
this process yields , in first order perturbation theory the following expression for the transition rate @xmath5 where @xmath6 is the energy difference between the initial and final states of the detector and @xmath7 is the positive frequency wightman function .
@xmath8 depends on the internal constituency of the detector and a detailed discussion of it is given in @xcite and @xcite .
while the second term in ( [ eq1 ] ) corresponds to the noise this detector is submitted to , that s to say , on the way the field fluctuates as seen by the observer in his trajectory @xmath9 .
it is an easy exercise to show that a detector at rest in the trocheries - takeno coordinates @xmath10 @xmath11 according to ( [ eq1 ] ) , has a transition rate given by @xmath12 ^ 2 \sin[(s/2)\sinh(\omega r)r^{-1}]^2}ds,\ ] ] where @xmath13 is the proper time between two points in the trajectory of the detector . after a change in the integration variable the above transition rate ,
is written as @xmath14dx } { ( x)^2-v^2\sin(x)^2 } , \label{eq2}\ ] ] where @xmath15 the fact that @xmath16 , comes from the type of trajectory that the detector is following . in this sense , this trajectory is more physical , because the detector is not allowed to travel faster than the speed of light . in ( [ eq1 ] )
, the @xmath17 factor is associated to the adiabatic switching of the detector and the usual @xmath18 prescription is related to its size .
as @xmath16 , and the integral is over the real axis , the integrand in ( [ eq2 ] ) can be written as @xmath19}{(x - i\epsilon)^2 } \sum_{n=0}^{\infty}\left(\frac{v\sin(x - i\epsilon)}{x - i\epsilon}\right)^{2n}.\ ] ] after a binomial expansion , this last integral is evaluated using ordinary residue calculus @xmath20^{2n+2k-1 } \theta ( -e + k\omega\gamma ) } { ( -1)^{n-1}(2n+2k-1)(n-1)!(n+2k-1 ) ! } , \label{eq3}\ ] ] where @xmath21 is the heaviside step function , which indicates that the rotational motion is the thermal reservoir .
the time scales of the detector @xmath22 and the ( proper ) period of rotation of the detector @xmath23 are singled out in ( [ eq3 ] ) , a property also shown in @xcite and @xcite . as a spectrum , ( [ eq3 ] ) ,
can be divided by the energy @xmath24 and an _ effective _ distribution function for the scalar particles can be obtained @xmath25^{2n+2k-1 } \theta ( -e + k\omega\gamma ) } { ( -1)^{n-1}(2n+2k-1)(n-1)!(n+2k-1)!}. \label{dist}\ ] ] effective here is in a certain sense , that the distribution function is dependent on the measuring apparatus .
other consequences connected to the motion , should be perceived by other types of detectors .
only the detectable properties are taken into account .
this distribution function is understood as the mean number of particles per volume , with energy @xmath24 perceived by the detector . in the following
, density will be assumed to be per unit volume . in this spirit , the total energy density of the scalar particles is given by the following integral , which can be written in terms of the hypergeometric function @xmath26^{2n+2k-1 } } { ( -1)^{n-1}(2n+2k-1)(n-1)!(n+2k-1)!}\nonumber\\ & & e_t=\frac{\omega^2}{2 \pi v } \sum_{n=1}^{\infty}\sum_{k=1}^{\infty}\int_0^k dx \frac{\left[v\left(-x + k\right)\right]^{2n+2k-1 } } { ( -1)^{n-1}(2n+2k-1)(n-1)!(n+2k-1)!}\nonumber\\ & & e_t=\frac{1}{2 \pi r^2}\sum_{n=1}^{\infty}\sum_{k=1}^{\infty}\frac{(vk ) ^{2(n+k)}}{(-1)^{n-1}(2n+2k)(2n+2k-1)(n-1)!(n+2k-1)!}\label{limite}\\ & & e_t=\frac{1}{2\pi r^2}\sum_{k=1}^{\infty } \frac{(\tanh(\omega r)k)^{2k+2 } { } _ { 2}f_3(k+1/2,k+1;k+3/2,k+2,2k+1;-\tanh(\omega r)^2k^2)}{(k+1/2)(k+1)(2k)!}\label{energia},\end{aligned}\ ] ] where @xmath27 plays a role similar to the temperature , and there is an explicit dependence on the coordinate @xmath28 .
this corresponds to a non constant temperature distribution .
for instance , in the above expression ( [ energia ] ) , at the origin when @xmath29 , the energy density is @xmath30 , which makes sense , because the motion of detector when @xmath29 is inertial .
next , the convergence of the infinite summation ( [ limite ] ) , is discussed .
the second derivative with respect to @xmath31 of ( [ limite ] ) is of the type @xmath32 the last expression is valid asymptotically when @xmath33 and @xmath34 @xcite . in this asymptotic spirit , the summation ( [ converg ] ) can be replaced by the integral over @xmath35 , from @xmath36 to @xmath37 , which results in @xmath38 the original summation ( [ limite ] ) is obtained by integrating the last equation two times in @xmath31 @xmath39 which is plotted numerically in the following + + the above graphic indicates , in a indirect way , that the summation given in ( [ limite ] ) and ( [ energia ] ) converges , except when @xmath40 .
this divergence of the total energy density of the scalar particles perceived by the detector , is expected when the detector is moving ( almost ) with the velocity of light , @xmath40 . by summing the first few terms in ( [ energia ] ) , the following total energy density of scalar particles is obtained at a given @xmath28 , for @xmath41 and @xmath42 + showing that for a same distance @xmath28 from the origin , there is an increase in the energy for larger values of @xmath27 . it should be stressed that in spite that ( [ energia ] ) describes a non constant temperature distribution , it is an equilibrium configuration . this interesting equilibrium concept is well known and is expected if the space being considered is not homogeneous and isotropic as in trocheries - takeno case , see for example @xcite . the external condition that originates
this motion is responsible for this apparent non equilibrium .
the above construction avoids canonical quantization , and it could be usefull to a further understanding of external conditions . |
the measured @xmath3-quark production cross section at hadron colliders exceeds next - to - leading order ( nlo ) qcd predictions by more than a factor of two .
while it is conceivable that this discrepancy is due to higher - order corrections , the disagreement is surprising since nlo calculations have been reliable for other processes in this energy range .
berger et al . have analysed @xmath3-quark production in the context of the minimal supersymmetric standard model ( mssm ) and find that the excess in the cross section could be attributed to gluino pair - production followed by gluino decay into pairs of @xmath3 quarks and @xmath0 squarks , if both the gluino and the @xmath0 squark are sufficiently light @xcite .
in order to reproduce the transverse - momentum distribution of the @xmath3 quarks , the masses of the gluino and light @xmath0-squark mass eigenstate should be in the range @xmath716gev and @xmath85.5gev .
the masses of all other supersymmetric ( susy ) particles are assumed to be large , of order several hundred gev , so as to have evaded detection at lep2 .
interestingly , a renormalization - group analysis in the framework of the unconstrained mssm shows that a light @xmath0 squark is most natural if it is accompanied by a light gluino with mass of order 10gev @xcite .
berger et al . have further observed that a light @xmath0 squark could have escaped direct detection .
for example , the additional contribution to the @xmath9 cross section at large energy would only be about 2% and hence difficult to disentangle .
the pair - production of light scalars would alter the angular distribution of hadronic jets in @xmath10 collisions , but the present data are not sufficiently precise to rule out the existence of this effect @xcite .
on the other hand , there are important @xmath11-pole constraints on the parameters of this model .
most importantly , production of the light @xmath0 squark at the @xmath11 pole has to be suppressed , which implies a stringent constraint on the mixing angle @xmath12 relating the sbottom mass and weak eigenstates @xcite .
more recently , several authors have studied loop effects of the light susy particles on electroweak precision measurements @xcite , finding potentially large contributions to the quantity @xmath13 . however , a conflict with existing data can be avoided by having some of the superpartner masses near current experimental bounds , or by allowing for a new cp - violating phase in the susy sector @xcite .
the null result of a cleo search for the semileptonic decays @xmath14 and @xmath15 of sbottom hadrons implies that the branching ratios for the decays @xmath16 induced by @xmath17-parity violating couplings , or @xmath18 with an ultra - light neutralino @xmath19 , must be highly suppressed @xcite .
however , a light @xmath0 squark would be allowed to decay promptly via hadronic @xmath17-parity violating couplings in the modes @xmath20 or @xmath21 ( with @xmath22 ) .
alternatively , it could be long - lived , forming @xmath0-hadrons
. an interesting consequence of hadronic @xmath17-parity violating decays would be the abundant production of light baryons .
this could significantly alter the thrust - axis angular distribution for continuum events containing baryons at the @xmath4 factories
. a striking manifestation of the light @xmath0-squark scenario would be the production of like - sign charged @xmath4 mesons at hadron colliders , because the majorana nature of the gluino allows for the production of @xmath23 and @xmath24 final states @xcite .
another potential signature at hadron colliders is an enhanced yield of @xmath25 events @xcite .
it has also been pointed out that sbottom pairs would be copiously produced in @xmath26 and @xmath27 decays @xcite .
precise measurements of bottomonium decays could lead to new constraints on the squark and gluino masses .
the presence of light susy particles alters the running of @xmath28 , and it is often argued that this would exclude the existence of light gluinos .
this argument is incorrect .
first , a gluino with mass @xmath2gev would have a relatively small effect on the evolution of @xmath28 .
taking , for instance , @xmath29 ( a value in agreement with all low - energy determinations of the qcd coupling ) and including the contribution of the gluino octet to the @xmath30 function above the scale @xmath31 yields @xmath32 , which is about three standard deviations higher than the canonical value @xmath33 .
however , considering that at leading order only virtual gluino pairs contribute to the @xmath30 function , a more realistic treatment would include the gluino contribution above a scale @xmath34gev , in which case @xmath35 , in good agreement with the standard value .
secondly , it is important to realize that even a value of @xmath36 significantly above 0.118 would not rule out the model , the reason being that the characteristic scale @xmath37 inherent in all determinations of @xmath38 is typically much smaller than the total energy .
this is true , in particular , for the determinations based on event - shape variables . in practice ,
the measurements fix @xmath38 somewhere between a fraction of the @xmath11 mass down to several gev , where the gluino contribution to the @xmath30 function is negligible . using these determinations to quote values of @xmath36 ( as is routinely done )
assumes implicitly that the coupling runs as predicted in the sm . finally ,
a careful analysis of the running of @xmath28 in the presence of light susy particles would have to include , for each observable , the modifications in the theoretical formulae due to virtual and real emissions of the new particles .
these corrections could be significant , and could partially compensate effects arising from the modification of the @xmath30 function .
if we are to take the possibility of a light @xmath0 squark and light gluinos seriously , then the theoretical study of their impact must be extended to the phenomenology of weak decays of the @xmath3 quark .
new sources of flavor violation arise from @xmath39-@xmath0-@xmath40 and @xmath41-@xmath0-@xmath40 couplings .
the overall scale of susy flavor - changing interactions originating from gluino exchange is set by the factor @xmath42 , which is much larger than the corresponding factor @xmath43 for weak decays in the standard model ( sm ) .
consequently , the new flavor - changing couplings must be much smaller than the ckm mixing angles in order for this model to be phenomenologically viable .
the most stringent bounds arise from the radiative decay @xmath5 , which we discuss in the present work .
( contributions of light @xmath0 squarks to kaon decays , @xmath44-@xmath45 mixing , and @xmath46-@xmath47 mixing are strongly suppressed . )
the presence of such tight bounds implies stringent constraints on model building .
if the light @xmath0 squark is sufficiently light to be pair - produced in @xmath3 decays , new unconventional decay channels would be opened up , which could affect the phenomenology of @xmath4 mesons and beauty baryons .
examples of potentially interesting consequences include modifications of beauty lifetime ratios , an enhancement of the semileptonic branching ratio of @xmath4 mesons via production of charmless final states containing @xmath0 squarks , an enhancement of @xmath48 and of the semileptonic cp asymmetry @xmath49 , and wrong - sign kaon production via @xmath50 transitions allowed by the majorana nature of the gluino .
the phenomenology of such effects will be discussed elsewhere .
if the light @xmath0 squark is too heavy to be pair - produced , it would still give rise to potentially large virtual effects in @xmath4 decays .
their study is the main purpose of this letter .
we denote by @xmath52 with @xmath53 the down - squark mass eigenstates , and by @xmath54 and @xmath55 with @xmath56 the interaction eigenstates ( the superpartners of the left - handed and right - handed down quarks ) .
they are related by a unitary transformation @xmath57 and @xmath58 .
we identify @xmath59 with the light sbottom mass eigenstate and define a sbottom - sector mixing angle @xmath12 through @xmath60 and @xmath61 .
the fact that the light @xmath0 squarks are not produced in @xmath11 decays implies @xmath62 the phenomenologically favored range for the @xmath63 coupling is @xmath640.45 @xcite , meaning that the light sbottom is predominantly the superpartner of the right - handed bottom quark . in our numerical analysis
we will assume a vanishing tree - level coupling to the @xmath11 and thus use @xmath65 .
the flavor - changing couplings involving the light @xmath0 and @xmath40 fields can be parameterized by dimensionless quantities @xmath66 where @xmath67 , and @xmath68 or @xmath41 for @xmath69 or @xmath70 transitions , respectively . in general the parameters @xmath71
are complex , which can lead to new cp - violating effects .
these parameters are invariant under a phase redefinition of the light @xmath0-squark state , and they transform in the same way as the products @xmath72 ( with @xmath73 ) of ckm matrix elements under a phase redefinition of the down - type quark fields .
it follows that ratios of the type @xmath74 and @xmath75 are invariant under phase redefinitions , and thus can carry an observable , cp - violating phase .
flavor - changing hadronic processes in the model with a light gluino and a very light @xmath0 squark are most transparently described by means of an effective `` weak '' hamiltonian .
if we neglect effects that are suppressed by inverse powers of the heavy susy scale , the relevant energy scales are the electroweak scale , at which the usual sm flavor - changing operators are generated by integrating out the top quark and the @xmath76 and @xmath11 bosons , and the scale @xmath77 , at which new flavor - changing operators are generated by integrating out the gluinos .
we start by discussing the construction of the effective theory below the gluino scale , focusing on the new interactions proportional to @xmath71 induced by gluino exchange , as illustrated in figure [ fig : effth ] .
susy modifications of the renormalization - group ( rg ) evolution of the standard weak - interaction operators will be discussed later .
the remaining light degrees of freedom in the low - energy theory are the quarks @xmath78 , the photon and gluons , and the light @xmath0 squark .
operators in the effective hamiltonian can be organized in an expansion in inverse powers of the gluino mass . for @xmath79 , it is a good approximation to keep the leading terms in this expansion , which have mass dimension five .
these operators comprise the usual electromagnetic and chromomagnetic dipole operators , and new operators containing two scalar @xmath0 fields .
the effective hamiltonian for @xmath69 transitions is ( here and below , @xmath80 denotes the running gluino mass at the gluino matching scale ) @xmath81 + o(1/{m_{\tilde g}}^2 ) \,,\ ] ] where @xmath82 we work with the covariant derivative @xmath83 , where @xmath84 is the electric charge of a down - type ( s)quark .
( to facilitate comparison with the literature , which usually adopts the opposite sign convention for the couplings , we have included a factor of @xmath85 in the definition of the dipole operators @xmath86 and @xmath87 . ) in addition , there are dimension - five fermion - number violating interactions of the form @xmath88 , which mediate @xmath50 transitions .
they are irrelevant to our discussion here . [ cols="^,^,^ " , ] it is well known that nlo corrections have a significant impact on the @xmath5 decay rate in the sm , which is largely due to nlo corrections to the matrix elements of the operators @xmath89 in the effective weak hamiltonian .
( nlo corrections to the wilson coefficient @xmath90 have a much smaller effect . ) in order to capture the bulk of these corrections , we include the @xmath91 contributions to the matrix elements but neglect susy nlo corrections to the coefficients @xmath92 and @xmath93 .
we also neglect two - loop contributions to the matrix elements involving @xmath0-squark loops .
this is justified because of the relatively large mass of the @xmath0 squark , and because our two - loop anomalous dimension calculation has shown that there is very little mixing of the squark operators into the dipole operators . at nlo
our results become sensitive to the precise definition of the mass parameter @xmath94 , which we identify with the so - called upsilon mass @xcite , for which we take the value @xmath95gev @xcite .
( up to a small nonperturbative contribution , @xmath96 is one half of the mass of the @xmath97 resonance . )
we also introduce a cutoff @xmath98 on the photon energy in the @xmath4-meson rest frame , which is required in the experimental analysis of radiative @xmath4 decays .
we then obtain @xmath99 where @xmath100 is obtained from the formulae in @xcite by obvious modifications to include the effects of the new susy contributions , and by a change in some of the nlo terms due to the introduction of the upsilon mass in ( [ rate ] ) @xcite .
the dependence of the branching ratio on the susy flavor - changing couplings can be made explicit by writing @xmath101 \,.\ ] ] in table [ tab : rcoeff ] , we give results for the coefficients @xmath102 and @xmath103 including the dominant theoretical uncertainties .
following @xcite , we use a running charm - quark mass in the penguin - loop diagrams rather than the pole mass .
this is justified , because the photon - energy cut imposed in the experimental analysis prevents the intermediate charm - quark propagators from being near their mass shell . specifically , we work with the mass ratio @xmath104 , where the running masses are obtained from @xmath105gev and @xmath106gev . in the left - hand plot in figure
[ fig : bsgamma ] , we confront our theoretical result for the @xmath5 branching ratio with the cleo measurement @xmath107 obtained for @xmath108 gev @xcite .
( this result actually corresponds to the sum of @xmath5 and @xmath109 decays .
however , the suppression of the exclusive @xmath110 decay with respect to the @xmath111 mode implies that the dominant contribution to the inclusive decay must come from @xmath112 transitions . in the context of our model
, it follows that the couplings @xmath113 must obey even tighter constraints than the @xmath114 . )
it follows that the maximum allowed values of the parameters @xmath114 are @xmath115 , as is already obvious from the magnitude of the coefficients @xmath103 in table [ tab : rcoeff ] .
however , values larger than @xmath116 would require a fine - tuning of the phase of @xmath117 and are thus somewhat unnatural . note that the ratio shown on the vertical axis in the plot is bound to lie between 1 and @xmath85 , and in the limit @xmath118 corresponds to @xmath119 , where @xmath120 denotes the cp - violating phase of @xmath117 .
the right - hand plot in the figure shows a similar constraint arising from the inclusive charmless decay @xmath6 . at leading order
, the decay rate for this process is obtained from ( [ rate ] ) by the replacements @xmath121 and @xmath122 .
the allowed region corresponds to the cleo upper bound of 8.2% ( at 95% c.l . ) for the @xmath6 branching ratio @xcite .
there are , however , potentially large theoretical uncertainties in this result , because we neglect nlo corrections to the branching ratio .
we therefore refrain from combining the two plots to reduce the allowed parameter space . in the sm
, the direct cp asymmetry in the inclusive decay @xmath5 is very small , below 1% in magnitude @xcite . in the susy scenario , on the other hand
, the phase of the coupling @xmath117 could lead to a large asymmetry . in the approximation where one neglects the susy contributions to the cp - averaged decay rate in ( [ rate ] ) , which is justified in view of the good agreement of the sm prediction with the data , the formulae in @xcite
yield the prediction @xmath123 , where we have neglected the small contribution from the charm - quark loops and the yet smaller contribution from @xmath0-squark loops .
( we use the standard phase convention where @xmath124 is real . in general , the cp asymmetry depends on @xmath125 . )
it follows that even within the very restrictive bounds shown in figure [ fig : bsgamma ] there can be a potentially large contribution to the cp asymmetry , which would provide a striking manifestation of physics beyond the sm .
in fact , the cleo bounds @xmath126 ( at 90% c.l . ) @xcite imply that @xmath127 placing another tight constraint on the flavor - changing coupling @xmath117 .
new supersymmetric contributions to @xmath3-quark production at hadron colliders can account for the long - standing discrepancy between the measured cross sections and qcd predictions if there is a light @xmath0 squark with mass in the range 25.5gev , accompanied by a somewhat heavier gluino @xcite . in this letter , we have explored the phenomenology of rare @xmath4 decays in such a scenario and have found tight bounds on the flavor - changing parameters controlling supersymmetric contributions to @xmath69 and @xmath70 fcnc transitions .
the most restrictive constraints arise from virtual effects of light @xmath0 squarks in @xmath5 decays .
we have analysed this process by constructing a low - energy effective hamiltonian , in which the gluinos are integrated out , while the @xmath0 squarks remain as dynamical degrees of freedom .
we find that the flavor - changing couplings @xmath128 and @xmath117 must be of order few times @xmath129 or less .
( even tighter constraints hold for the analogous @xmath70 couplings . )
this implies that certain off - diagonal entries of the down - squark mass matrix must be suppressed by a similar factor compared to the generic squark - mass squared .
even with such tight constraints on the couplings , this model allows for interesting and novel new physics effects in weak decays of @xmath4 mesons and beauty baryons . as an example
, we have discussed the direct cp asymmetry in @xmath5 decays , which could be enhanced with respect to its standard model value by an order of magnitude .
other possible effects include an enhanced @xmath6 decay rate .
we have not considered here the possibility of @xmath0-squark pair - production , which would be kinematically allowed for very light squark masses .
the new decay modes @xmath130 and @xmath50 would affect the decay widths of @xmath4 mesons and @xmath131 baryons differently , and hence might explain the anomaly of the low @xmath131 lifetime . we will report on this interesting possibility elsewhere . s.b . and m.n .
are supported by the national science foundation under grant phy-0098631 .
is supported by the department of energy under grant de - ac03 - 76sf00515 , and a.k . under grant de - fg02 - 84er40153 .
e. l. berger et al . , phys .
lett . * 86 * , 4231 ( 2001 ) [ hep - ph/0012001 ] .
a. dedes and h. k. dreiner , jhep * 0106 * , 006 ( 2001 ) [ hep - ph/0009001 ] .
m. carena , s. heinemeyer , c. e. wagner and g. weiglein , phys . rev .
lett . * 86 * , 4463 ( 2001 ) [ hep - ph/0008023 ] . j .- j .
cao , z .- h . xiong and j. m. yang , phys .
lett . * 88 * , 111802 ( 2002 ) [ hep - ph/0111144 ] .
g. c. cho , preprint hep - ph/0204348 .
baek , phys .
b * 541 * , 161 ( 2002 ) [ hep - ph/0205013 ] .
v. savinov _ et al . _
[ cleo collaboration ] , phys .
d * 63 * , 051101 ( 2001 ) [ hep - ex/0010047 ] .
a. k. leibovich and d. rainwater , phys .
lett . * 88 * , 221801 ( 2002 ) [ hep - ph/0202174 ] .
e. l. berger and l. clavelli , phys .
b * 512 * ( 2001 ) 115 [ hep - ph/0105147 ] .
e. l. berger and j. lee , phys .
d * 65 * , 114003 ( 2002 ) [ hep - ph/0203092 ] .
m. ciuchini et al .
b * 316 * , 127 ( 1993 ) [ hep - ph/9307364 ] .
k. chetyrkin , m. misiak and m. mnz , nucl .
b * 518 * , 473 ( 1998 ) [ hep - ph/9711266 ] . c. greub , t. hurth and d. wyler , phys .
d * 54 * , 3350 ( 1996 ) [ hep - ph/9603404 ] .
k. chetyrkin , m. misiak and m. mnz , phys .
b * 400 * , 206 ( 1997 ) [ erratum : _ ibid .
_ b * 425 * , 414 ( 1998 ) ] [ hep - ph/9612313 ] .
a. l. kagan and m. neubert , eur .
j. c * 7 * , 5 ( 1999 ) [ hep - ph/9805303 ] .
gambino and m. misiak , nucl .
b * 611 * , 338 ( 2001 ) [ hep - ph/0104034 ] .
p. abreu _ et al .
_ [ delphi collaboration ] , phys .
b * 418 * , 430 ( 1998 ) .
a. h. hoang , z. ligeti and a. v. manohar , phys .
lett . * 82 * , 277 ( 1999 ) [ hep - ph/9809423 ] .
m. neubert and t. becher , phys .
b * 535 * , 127 ( 2002 ) [ hep - ph/0105217 ] . specifically , the expression for the quantity @xmath132 in eq .
( 13 ) of @xcite must be replaced by @xmath133where @xmath134gev@xmath135 and @xmath136gev@xmath135 are hadronic parameters .
s. chen _ et al .
_ [ cleo collaboration ] , phys .
lett . * 87 * , 251807 ( 2001 ) [ hep - ex/0108032 ]
. t. e. coan _ et al .
_ [ cleo collaboration ] , phys . rev .
lett . * 80 * , 1150 ( 1998 ) [ hep - ex/9710028 ] .
a. l. kagan and m. neubert , phys .
d * 58 * , 094012 ( 1998 ) [ hep - ph/9803368 ] .
t. e. coan _ et al .
_ [ cleo collaboration ] , phys . rev . lett . * 86 * , 5661 ( 2001 ) [ hep - ex/0010075 ] . | the presence of a light @xmath0 squark ( @xmath1gev ) and gluino ( @xmath2gev ) might explain the observed excess in @xmath3-quark production at the tevatron .
though provocative , this model is not excluded by present data .
the light supersymmetric particles can induce large flavor - changing effects in radiative decays of @xmath4 mesons .
we analyse the decays @xmath5 and @xmath6 in this scenario and derive restrictive bounds on the flavor - changing quark - squark - gluino couplings . clns 02/1790 + slac - pub-9230 + hep - ph/0205274 + may 24 , 2002 * constraints on light bottom squarks from + radiative
b - meson decays * thomas becher + _ stanford linear accelerator center + stanford university , stanford , ca 94309 , usa _
+ stephan braig , matthias neubert + _ newman laboratory of elementary - particle physics , cornell university + ithaca , ny 14853 , usa _ + alexander l. kagan + _ fermi national accelerator laboratory + batavia , il 60510 , usa _ |
magnetoelectronics , also known as spintronics , is probably the field with the most rapid growth in materials science bringing to the center of scientific research new phenomena.@xcite for researchers dealing with ab - initio calculations one of the most interesting concept in spintronics is the half - metallicity.@xcite research has been focused initially on the explanation of the origin of the half - metallicity and afterwards on the prediction of new half - metallic materials with potential applications in spintronics devices .
half - metals are hybrids between normal metals and semiconductors : the majority - spin band is crossed by the fermi level as in a normal metal while the fermi level falls within a gap in the minority - spin band as in semiconductors leading to a perfect 100% spin - polarization at the fermi level,@xcite maximizing the efficiency of magnetoelectronic devices.@xcite de groot and his collaborators in 1983 were the first to predict the existence of half - metallicity in the case of the intermetallic semi - heusler alloy nimnsb,@xcite and the origin of the gap seems to be well understood .
@xcite there exist several ab - initio calculations on nimnsb reproducing the initial results of de groot and collaborators,@xcite and galanakis _ et al .
_ showed that the gap arises from the hybridization between the @xmath1 orbitals of the ni and mn atoms.@xcite moreover in the latter reference it was shown that the total spin moment for the heusler alloys of the nimnsb type follows a slater - pauling behavior being in @xmath2 the total number of valence electrons in the unit cell minus 18 since there are exactly nine occupied minority - spin states.@xcite theoretical calculations for the interfaces of these materials with the semiconductors are few and all results agree that in general the half - metallicity is lost both at the surfaces@xcite and the interfaces with binary semiconductors.@xcite wijs and de groot have argued than in the case of the nimnsb / cds ( 111 ) contacts the sb / s interface keeps half - metallicity when the s atoms sit exactly on top of sb.@xcite moreover taking into account also the reconstruction at the surfaces and interfaces can have an important effect on their properties.@xcite finally we should note that several other aspects of nimnsb have been studied using first - principles calculations like the exchange constants and curie temperatures,@xcite the quasiparticle states@xcite and the dynamical effects,@xcite the defects and doping,@xcite the structural stability,@xcite the effect of spin - orbit coupling,@xcite the fit of model hamiltonians to describe nimnsb,@xcite orbital magnetism,@xcite the pressure and thermal expansion effects,@xcite the temperature effect@xcite and the magneto - optical properties.@xcite the half - metallic character of nimnsb in single crystals has been well - established experimentally .
infrared absorption@xcite and spin - polarized positron - annihilation@xcite gave a spin - polarization of @xmath0100% at the fermi level .
high quality films of nimnsb have been grown,@xcite but they were found not to reproduce the half - metallic character of the bulk .
values of 58% and 50% for the spin - polarization at the fermi level were obtained by soulen _ _ et al.__@xcite and by mancoff _
et al._,@xcite respectively , and recently zhu _ _ et al.__@xcite found a value of 40% using spin - resolved photoemission measurements on polycrystalline films .
_ et al.__@xcite showed that during the growth of the nimnsb thin films , sb atoms segregate to the surface decreasing the obtained spin - polarization ; they measured a value of @xmath030% at 200k , while at room temperature the net polarization was practically zero .
but when they removed the excess of sb by a flash annealing , they managed to get a nearly stoichiometric ordered alloy surface terminating in mnsb .
inverse photoemission experiments at room temperature revealed that the latter surface shows a spin - polarization of about 67@xmath39% which is significantly higher than all previous values.@xcite there is also experimental evidence that for a temperature of @xmath080 k there is transition from a half metal to a normal ferromagnet,@xcite but these experiments are not yet conclusive .
finally , the effect of stress on the magnetic anisotropy of thin nimnsb films and the role of defects have been explored.@xcite based on the success of first - principles electronic structure calculations to describe the properties of nimnsb , several authors have predicted new half - metallic heusler alloys crystallizing in the @xmath4 structure of semi - heusler compounds like nicrm and nivm ( m= p , as , sb , s , se and te),@xcite and xcral ( x= fe , co , ni ) and nicrz ( z= al , ga , in).@xcite recently , katayama - yoshida and collaborators published a paper including also ab - initio calculations on nimnsi semi - heusler alloy , which was predicted to have a curie temperature of 1050 k,@xcite exceeding even the 730 k shown by nimnsb.@xcite this finding motivated us to study the electronic and magnetic properties of this compound in detail since except the very high curie temperature , it should be easily grown due to the well - known and well - controlled growth of nimnsb and it should be compatible with si ( si crystallizes in a diamond structure and the unit cell of nimnsi is two times the unit cell of si ) .
we decided to expand our study to cover also the closely - related nicrsi and nivsi alloys .
nivsi was found not to be half - metallic at its equilibrium lattice constant and thus we focused our study on nicrsi and nimnsi alloys . in section [ sec2 ]
we present the details of our calculations and in section [ sec3 ] the electronic and magnetic properties of the bulk phase of these alloys . in section [ sec4 ]
we discuss the robustness of half - metallicity and ferromagnetism , and in section [ sec5 ] the properties of ( 100 ) surfaces and of their interfaces with the si , gap and zns semiconductors which have an experimental lattice constant almost identical to the theoretical equilibrium lattice constants of nicrsi and nimnsi .
finally in section [ sec6 ] we discuss the effect of defects and disorder and in section [ sec7 ] we summarize and conclude .
schematic representation of the @xmath4 structure adopted by the semi-(or half-)heusler compounds like nimnsi and of the @xmath5 structure adopted by the full - heusler compounds like co@xmath6mnsi .
the lattice consists of 4 fcc sublattices .
the unit cell in the case of semi - heuslers xyz is that of a fcc lattice with four atoms as basis : x at @xmath7 , y at @xmath8 , a vacant site at @xmath9 and z at @xmath10 in wyckoff coordinates . in the case of full - heusler compounds the vacant site
is also occupied by a x atom . ]
to study the niysi alloys we have employed both the full potential nonorthogonal local
orbital minimum basis band structure scheme ( fplo)@xcite and the full - potential screened korringa - kohn - rostoker ( fskkr ) method.@xcite both methods as we will discuss later give similar results for the bulk alloys .
fplo is suitable to study the disorder within the coherent - potential approximation ( cpa ) .
fskkr is suitable to study surfaces and interfaces since its screened character and the use of decimation method for the inversion of the green function matrix lead to an order @xmath11 scaling of the cpu time , where @xmath11 is the number of inequivalent atoms in the unit cell .
moreover dyson equation makes possible the study of single impurities within the framework of fskkr in real space.@xcite in both cases we employed the local - spin - density approximation ( lsda)@xcite for the exchange - correlation energy within the framework of the density functional theory.@xcite semi - heusler alloys ( also known as half - heusler compounds ) crystallize in the @xmath4 structure which consists of four fcc sublattices and have the chemical formula xyz where x a high - valent transition metal atom , y a low - valent transition metal atom and z a @xmath12 atom . in fig .
[ fig1 ] we present both the @xmath4 structure of semi - heuslers and the @xmath5 structure of the full - heusler compounds like co@xmath6mnsi .
the unit cell is that of a fcc lattice with four atoms as basis : x atoms occupy the @xmath7 site , y atoms the @xmath8 , the @xmath9 site is vacant for the semi - heuslers and occupied by x in the case of full - heuslers , and finally z atoms occupy the @xmath10 site ( we use wyckoff coordinates to denote the sites ) . the zinc - blende structure adopted by a large number of semiconductors , like gap and zns is also consisting of four fcc sublattices . in the case of gap
the @xmath7 site is occupied by a ga atom , the @xmath8 site by a p atom , while the other two sites are unoccupied .
the diamond structure of si is also coherent with the @xmath4 structure of semi - heusler alloys since double the unit cell gives the zinc - blende structure of binary semiconductors .
this close structure similarity makes the heusler alloys compatible with the existing semiconductor technology and thus very attractive for industrial applications .
since the compounds under study have not been grown experimentally , we have to determine their equilibrium lattice constant . in fig . [ fig2 ]
we plot the fplo calculated total energies vs the lattice constant for all three nivsi , nicrsi and nimnsi .
we have assumed that all three are ferromagnets ( which is confirmed by our calculated exchange constants , at least for nicrsi and nimnsi , in section [ sec4 ] ) .
we have scaled the total energy for each compound in a way that the zero of the total energy corresponds to the minimum energy at the theoretical equilibrium lattice constants .
all three compounds have an equilibrium lattice constant very close to 5.4 .
we have decided to use the latter one for all calculations presented in this manuscript . .
( color online)atom - resolved and total spin magnetic moments in @xmath2 calculated using both the fplo and fskkr methods . in the last two columns we present our results on the curie temperature in kelvin units calculated with both the mean - field ( @xmath13 ) and random - phase ( @xmath14 ) approximations .
[ table1 ] [ cols="<,^,^,^,^,^,^ " , ] finally , we have investigated the case of ni - cr and ni - mn atomic swaps .
these defects have the peculiarity that cr(mn ) and ni atoms move to sites of different symmetry with respect to their perfect sites in the nicrsi and nimnsi alloys . in fig . [ fig13 ]
we present for both compounds the total dos for two different concentrations of atomic swaps ( @xmath15=0.1 and 0.2 ) and the atom - resolved dos for the first case .
the ni and cr(mn ) atoms at the perfect sites show a dos similar to the perfect compounds presented in fig .
the impurity ni atoms at the perfect cr(mn ) sites show a double - peak structure and the fermi level falls within this structure . when the concentration is increased to 0.2 , as it is obvious from the total dos , the two peaks overlap creating a wide - band .
the ni impurity atoms carry a very large spin moment with respect to the spin moment of the ni atoms at the perfect ni site as can be seen in table [ table7 ] where we present the spin magnetic moments for both cr- and mn - alloys and for various concentrations of atomic swaps . as we increase the concentration of atomic swaps the spin moments of both the ni impurity atoms and ni atoms at the perfect sites decrease .
the changes are even more important for the cr(mn ) atoms at the ni sites . due to the different symmetry these atoms have a spin moment antiparallel to the spin moment of the other transition metal atoms reducing considerably the total spin moment and leading to ferrimagnetic alloys .
this behavior is present also in the case of cr and mn impurity atoms at perfect co sites in co@xmath6cral(si ) and co@xmath6mnal(si ) full - heusler alloys,@xcite where cr(mn ) impurity atoms have spin moments antiparallel to the spin magnetic moment of the other transition metal atoms .
but in these full - heusler alloys the half - metallic character is kept contrary to the compounds under study here .
this is due to the ni impurity atoms and not to the cr(mn ) impurity atoms which have a very small dos in the minority - spin band which is obviously a very weak image of the ni - impurity dos . to examine this behavior
even further we have plotted in fig .
[ fig11 ] the dos for a single cr impurity atom at a perfect ni site and for a single ni impurity atom at a perfect cr site in nicrsi . in the case of the cr impurity atom
the gap persists and the spin moment is negative ( in table [ table7 ] we present in parenthesis the spin moment of the single impurity atoms ) and close to the values calculated using the fplo - cpa code . the ni impurity atoms at the cr site show a double peak structure at the fermi level with the @xmath16 states just below the gap and the fermi level being pinned exactly at the minority @xmath17 peak .
thus due to symmetry reasons the majority @xmath1-states of ni - impurity atoms are completely occupied leading to large spin moments and the fermi level falls within a double peak - structure leading to a complete destruction of half - metallicity .
thus any kind of ni defects leads to loss of the half - metallic character .
we have studied using both the full potential nonorthogonal local orbital minimum basis band structure scheme ( fplo ) and the full - potential screened korringa - kohn - rostoker ( fskkr ) electronic structure methods the electronic , magnetic and gap - related properties of the niysi compounds and have expanded our study also to cover the case of surfaces , interfaces with semiconductors and defects .
when y stands for v , the ferromagnetism is not very stable due to the weak v - v interactions and these compounds are not suitable for applications .
when y stands for cr or mn , the ferromagnetism is extremely stable leading to very high values of the curie temperature which is predicted to be @xmath0700 k for nicrsi and @xmath01100 k for nimnsi .
both nicrsi and nimnsi are half - metallic at their equilibrium lattice constant with large width of the minority - spin gap ( @xmath01 ev ) and integer values of the total spin moment as predicted by the slater pauling rule .
the gap is created due to the creation of bonding and antibonding @xmath1-hybrids in the minority - spin band .
the width of the gap is marginally affected upon tetragonalization even when we expand or contract the lattice by 5% .
nicrsi ( 001 ) surfaces present a high spin - polarization at the fermi level with respect to the nimnsi alloy due to the large intensity of the cr majority - spin density of states ( dos ) at the fermi level . in the case of interfaces with semiconductors with similar lattice parameter ( gap , zns and si ) ,
ni - based contacts show larger spin - polarization since ni atoms at the interface layer have a more bulk - like environment and in the case of nicrsi interfaces the high cr - dos leads to values of the spin - polarization for the ni - based contacts as high as 90% .
finally , we have shown that there are two kind of defects .
the defects and atomic swaps involving the cr(mn ) and si atoms lead to a broadening of the bands and the gap is slowly shrinking and for a critical value of the defects - concentration it disappears .
ni defects on the other hand lead to a shift of the energy - localized @xmath16 states within the gap and these impurity states completely destroy the half - metallicity . for realistic applications
it seems that nicrsi is more suitable with respect to both nimnsi studied here and the well - known and widely studied nimnsb heusler alloy .
its curie temperature is smaller than the mn - based alloys but it is still high - enough for applications exceeding considerably the room temperature .
moreover the large population of cr majority - spin states at the fermi level ensures the high - spin polarization at the fermi level which is needed for the injection of current in semiconductors .
crucial for the operation of devices based on heusler alloys is the prevention of the creation of ni defects , since they induce impurity states within the gap which can couple to interface states and completely destroy the spin - polarization of the current injected into the semiconductor .
p. j. webster and k. r. a. ziebeck , in _ alloys and compounds of d - elements with main group elements .
_ , edited by h. r. j. wijn , landolt - bornstein , new series , group iii , vol .
19/c ( springer - verlag , berlin 1988 ) . | using state - of - the - art first - principles calculations we study the properties of the ferromagnetic heusler compounds niysi where y stands for v , cr or mn .
nicrsi and nimnsi contrary to nivsi are half - metallic at their equilibrium lattice constant exhibiting integer values of the total spin magnetic moment and thus we concentrate on these two alloys .
the minority - spin gap has the same characteristics as for the well - known nimnsb alloy being around @xmath01 ev .
upon tetragonalization the gap is present in the density of states even for expansion or contraction of the out - of - plane lattice parameter by 5% .
the cr - cr and mn - mn interactions make ferromagnetism extremely stable and the curie temperature exceeds 1000 k for nimnsi .
surface and interfaces with gap , zns and si semiconductors are not half - metallic but in the case of nicrsi the ni - based contacts present spin - polarization at the fermi level over 90% .
finally , we show that there are two cases of defects and atomic - swaps .
the first - ones which involve the cr(mn ) and si atoms induce states at the edges of the gap which persists for a moderate - concentration of defects .
defects involving ni atoms induce states localized within the gap completely destroying the half - metallicity . based on single - impurity calculations we associate these states to the symmetry of the crystal . |
the magneto - rotational instability ( mri ) and magneto - rotational turbulence ( mrt ) provide an elegant framework to study the origin of enhanced angular momentum transport in accretion discs .
much effort has been devoted to understanding the nonlinear development of the mri and the processes that control the saturation amplitude of the instability , since , ultimately this controls the transport efficiency .
because of the difficulties inherent in approaching a strongly nonlinear problem analytically , much of the work on mrt has relied on numerical simulations with all their attendant idealizations and approximations . by far the most popular is the shearing - box approximation in which the computational domain is restricted to a region of small radial extent at a large radius in the disc . under reasonable assumptions
this can be mapped into a cartesian layer with shearing - periodic boundary conditions in the radial direction . because the shearing - box approximation conserves vertical magnetic flux it is important to distinguish two types of simulations : those with finite initial ( vertical ) flux and those with zero initial flux .
if the flux is finite there is a linear instability with a well defined growth rate and wavelength of maximum growth whose values are determined by the amount of flux @xcite . in the nonlinear regime
the amplitude of the maxwell stresses primarily responsible for angular momentum transport is controlled by the amount of magnetic flux , and most crucially , remains finite in the ideal limit of vanishing dissipation .
if , on the other hand , the initial flux is zero , the domain could in principle de - magnetize completely and relax to a state of uniform shear . if after a long time it does not , it must be because the magnetic field is being regenerated by turbulent motions . in this case , the mri does not manifest itself as an exponentially growing linear instability , rather it is a subcritical dynamo process . in this case , the spatial scales of the dominant magnetic structures and the efficiency of the angular momentum transport are determined by the dynamo itself .
two questions naturally arise : what kind of dynamo action can be sustained in a shearing - box , namely small - scale or large - scale , and what happens to the dynamo when the diffusivity , numerical or otherwise , becomes vanishingly small . addressing these issues
has turned out to be a major and complex undertaking , even within the idealized framework of the shearing - box approximation .
the first question is not specific to mri driven dynamos but to dynamos in general . under what circumstances
does a dynamo generate substantial amount of magnetic flux has been a long standing problem in astrophysical dynamo theory .
large - scale dynamos are often associated with flows lacking reflectional symmetry , or incorporating large scale shear , or a net flux of magnetic helicity through the boundaries , or any combination of the above . the second question was originally posed by @xcite within the framework of unstratified , homogeneous shearing boxes and it has since become known as the problem of convergence
. simply stated , a family of solutions of the mri equations does _ not _ converge if the maxwell stresses tend to zero as the dissipation tends to zero .
although , superficially , the convergence problem might seem mostly a matter of numerics , and indeed originally it was framed that way , actually it is not .
understanding why some shearing - box models converge and some do not is a fundamental question about nonlinear dynamo action in centrifugally stable systems .
it is now commonly accepted that homogeneous , unstratified shearing boxes without explicit dissipation these were the cases originally considered by @xcite do not converge ( for recent reviews , see * ? ? ?
* ; * ? ? ?
the reason for the lack of convergence may be related to the small - scale nature of the dynamo operating in these systems , or to the lack of a characteristic outer scale or to a combination of these two factors @xcite .
all other cases are not as clear . in the present paper
we address the problem of convergence , or lack thereof , in the stratified isothermal case without explicit dissipation .
this is the simplest shearing - box model with nontrivial stratification . despite the simplicity of the models ,
the dynamo that operates in these systems is far from simple .
in an isothermal atmosphere with linear gravity reversing in the middle , hydrostatic balance gives rise to a density stratification with an approximately gaussian profile and most of the mass concentrated near the mid - plane .
a seemingly turbulent dynamo operates in this dense , central region while propagating wavelike magnetic activity patterns are observed in the tenuous overlying layers @xcite .
a resolution study by @xcite with resolution up to 128 grid - points per scale height concluded that there was strong evidence for convergence .
this led several authors to declare this case as settled in favor of convergence @xcite . here , we extend this study to 200 grid - points with a similar , but not identical , setup and numerics to that of @xcite and find _ no _ evidence for convergence , at least up to these resolutions .
our conclusion is , therefore , that the problem of convergence for stratified , isothermal shearing - boxes is very much still an open issue .
we perform a convergence study for a three - dimensional compressible , isothermal , stratified shearing box ( for a description of the shearing box model see * ? ? ?
the simulations start from a layer in hydrostatic equilibrium . assuming the vertical gravity of the form @xmath0 , where @xmath1 is the orbital frequency and @xmath2 is the vertical coordinate
, the density distribution takes the form @xmath3 where @xmath4 is the value of density on the equatorial plane , @xmath5 is the scale height given by @xmath6 and @xmath7 is the isothermal sound speed .
taking @xmath8 as the unit of time , @xmath5 as the unit of length and @xmath4 as the unit of density , the ideal mhd equations for a keplerian shearing box , can be written in dimensionless form as @xmath9 @xmath10 @xmath11 where @xmath12 , @xmath13 , @xmath14 and @xmath15 denote , respectively , non - dimensional magnetic field intensity , velocity , density and pressure .
in addition we assume an isothermal equation of state .
note that we absorbed a factor of @xmath16 into the definition of @xmath12 .
the simulation domain covers the region @xmath17 the boundary conditions are periodic in @xmath18the azimuthal direction , shear periodic in @xmath19the radial direction , and impenetrable and stress - free in @xmath2 . on the horizontal planes at @xmath20
we assume hydrostatic balance , and that the magnetic field is purely vertical .
we note that these conditions allow a net flux of magnetic helicity through the boundaries unlike those in @xcite who adopt periodic conditions in @xmath2 .
generically , it is found that in these stratified , isothermal models the vertical " boundary conditions have little effect on the qualitative structure of the solutions @xcite we carried out a series of simulations at different resolutions with , respectively , 32 , 64 , 128 and 200 grid - points per scale height .
the largest grid is , thus , @xmath21 .
initially the magnetic field has the form @xmath22 where @xmath23 is chosen so as to give a ratio between thermal and magnetic pressure of @xmath24 .
clearly there is no net magnetic flux threading the box .
a small random perturbation in the azimuthal component of the velocity is introduced to trigger the instability .
the simulations were carried out with the pluto code @xcite which allows the choice between several different numerical schemes .
for the present work we opted for third order accurate parabolic reconstruction , constrained transport method for the magnetic field evolution and hlld riemann solver @xcite .
it is helpful to introduce the following notation : if @xmath25 is a generic function of space and time , we indicate a volume average by @xmath26 , an average over horizontal planes by @xmath27 and a time average by @xmath28 . the main result of our study is summarized in fig .
[ fig : maxwtime ] , where we show the time history of the volume averaged maxwell stresses , i.e. @xmath29 , for the four simulations with increasing resolution .
as it is usual in these type of simulation , we observe that , after an initial transient that lasts about 50 units of time , the stresses fluctuate around some average value .
the amplitude of the fluctuations strongly decreases in the highest resolution simulation , similarly to what happens in the homogeneous periodic case @xcite .
for the three simulations up to 128 points per scale height the value of maxwell stresses seems to fluctuate around similar values and if we had limited ourselves to these results we would have reached the same conclusion as in @xcite , i.e. that the stratified simulations seem to converge and give an efficiency of the transport independent from resolution .
however , if we look at the curve corresponding to a resolution of 200 points per scale height , it is systematically lower than the other ones . ) , width=377 ] we can get a more precise evaluation on how the efficiency of the transport changes with resolution by considering the horizontal plane time averaged ( computed excluding the initial transient phase ) maxwell stresses , i.e. @xmath30 , shown as a function of @xmath2 in fig .
[ fig : maxwz ] .
the stresses decrease from the resolution of 32 points to 64 points , stay constant at 128 points and decrease again at the highest resolution of 200 points .
in addition , looking at the behavior as a function of @xmath2 , we see that the stresses are concentrated in a region around the equatorial plane and have a strong decrease in the high altitude regions .
.,width=377 ] a major difference between the homogeneous , periodic case and the present one is the presence of an average magnetic field mostly in the toroidal direction . if we represent the distribution of @xmath31 as a function of @xmath32 and @xmath2 , as we do in the top panel of fig .
[ fig : butterfly ] , we can observe cyclic patterns propagating away from the equatorial plane , that become more evident at high altitudes and have been observed in all previous isothermal stratified simulations .
the bottom panel of the same figure shows the r.m.s .
value of the fluctuations of @xmath33 , i.e. @xmath34 , as a function of @xmath32 and @xmath2 , where @xmath35 is defined as @xmath36 indicating that the fluctuation level decreases away from the equatorial region . as a function of @xmath32 and @xmath2 , while the bottom panel shows the distribution of the r.m.s .
value of the fluctuations @xmath34,width=377 ] these considerations can be made more quantitative by comparing the behavior of the total magnetic energy and the magnetic energy of the mean field as a function of height , which we do in fig .
[ fig : magen ] .
the energy of the mean field is negligible in the equatorial region , therefore , in that region , the main contribution comes from fluctuations , on the other hand it becomes comparable to that of the fluctuations or dominant in the high altitude regions . in the highest resolution case
we have a clear decrease both in the fluctuations and the mean . , width=377 ]
we have revisited the problem of mri driven turbulence in isothermal , stratified shearing boxes with zero net ( vertical ) magnetic flux and no explicit dissipation .
we have extended our study to the highest resolution to date and find that , contrary to previously made claims based on lower resolution studies , the solutions do not converge , or at any rate , there is no convincing evidence of convergence .
the average maxwell stresses , principally responsible for the angular momentum transport , continue to decrease with increasing resolution .
this conclusion can be further elaborated in terms of simple models of the types discussed by @xcite and @xcite consisting of coupled dynamo systems operating in different regions .
one dynamo system is confined to the mid - plane where most of the mass is concentrated and gravity reverses , the other operating in the tenuous overlying regions .
the second dynamo is assumed to be of the mean - field type and be responsible for the generation of the magnetic structures that appear in the form of upward propagating dynamo waves . as for the nature of the mid - plane dynamo system
two possibilities readily come to mind .
one is that the motions in the mid - plane are driven by small - scale dynamo action similar to that observed in unstratified shearing - boxes .
the justification for this assumption is that gravity is weak near the mid - plane . in this scenario
the overlying mean - field dynamos are driven by the magneto - rotational turbulence in the mid - plane .
the source of the turbulence is a subcritical dynamo instability .
the other possibility is that the mean - field dynamos generate enough mean toroidal field in the mid - plane to drive an azimuthal mri whose non - linear development drives the turbulence that , in turn , drives the mean - field dynamos .
although the outward manifestation of these two scenarios is the same , the reason for the apparent non - convergence is different . in the first scenario ,
the lack of convergence of the overall system follows from the non - convergence of the small - scale dynamo operating in the mid - plane , which can plausibly be reconstructed to the non - convergence of the unstratified homogeneous cases .
if this analysis is the correct one , in the isothermal case , because most of the mass is concentrated in a region where there is practically no gravity , stratification does not help to resolve the convergence problem .
it is useful to note that the above argument reduces the convergence issue for the stratified case to the convergence issue for the homogeneous case . in other words ,
the former does not converge because the latter does not converge .
contrariwise , it could be argued that if the homogeneous case were to converge so would the stratified one . at present , there is convincing numerical evidence from several different groups that in the absence of explicit dissipation the unstratified , homogeneous case does not converge @xcite .
the case in which dissipation is included explicitly is not so clear cut ( see , for instance , the comments in * ? ? ?
although it is often asserted that the homogeneous case with explicit dissipation converges @xcite , as far as we can tell , all the numerical evidence supporting these assertions originates from a single paper , namely that of @xcite .
and although the simulations described therein remain an impressive numerical tour de force , we would argue that they are not such as to settle the issue of convergence unequivocally .
we hope however that in the near future other attempt will be made to settle the issue of convergence in the presence of explicit dissipation conclusively . in the second scenario
the non - convergence derives from the inability of the mean - field dynamo to operate at high magnetic reynolds numbers be they real or numerical .
this is a well known effect that has received much attention and goes back to the original works by @xcite and @xcite . in this case as the dissipation decreases so does the generated mean toroidal field needed to destabilize the azimuthal mri in the mid - plane regions .
eventually the mean toroidal field is so weak that the system become indistinguishable from the one in the first scenario with all its attendant limitations .
should this analysis turn out to be correct it is interesting that it applies to a case with boundary conditions that allow a net flux of magnetic helicity .
undoubtedly other scenarios can be constructed that agree with the numerical evidence , provide an explanation for the non - convergence and highlight the role of other physical processes .
however , to quote from one of the authors favorite poems whatever the reason his heart or his shoes " @xcite the stratified isothermal shearing - box appears not to converge .
the authors would like to thank the anonymous referee for the careful reading of the manuscript and for several valuable comments .
this work was supported in part by the national science foundation sponsored center for magnetic self organization at the university of chicago .
computations were performed at the jlich supercomputing center thanks to a deisa deci grant . | we consider the problem of convergence in stratified isothermal shearing boxes with zero net magnetic flux .
we present results with the highest resolution to - date up to 200 grid - point per pressure scale height that show no clear evidence of convergence .
rather , the maxwell stresses continue to decrease with increasing resolution .
we propose some possible scenarios to explain the lack of convergence based on multi - layer dynamo systems . |
creation of ultracold hetero - nuclear molecules opens the path towards experimental realization of strongly - interacting dipolar many - body systems .
depending on the constituent atoms , in moderate electric field these molecules can have large dipole moment of @xmath0 debye in their vibrational ground states @xcite .
in particular , fermionic molecules in presence of an optical lattice can be used to simulate various quantum phases , such as quantum magnetism and phases of @xmath1 like models @xcite , various charge density wave orders @xcite , bond - order solids @xcite etc .
one should also stress that in the strongly correlated regime , both in bosonic and fermionic systems the standard descriptions of single - band hubbard model ceases to be valid .
the effect of non - standard terms become important leading to novel phases like pair - superfluidity etc @xcite . while most of the works have dealing with higher bands concentrated on bosonic systems , in this paper , we study dipolar fermions confined in 2d optical lattice @xmath2 + \frac{m\omega^2}{2}z^2 $ ] , where @xmath3 is the lattice depth , @xmath4 is the lattice constant , @xmath5 is the mass of the molecule , and @xmath6 is the frequency of harmonic potential in @xmath7 direction .
the dipoles are polarized along the direction of harmonic trapping .
usually , at low temperature and for low tunneling , the phase diagram consists of different crystal states whose structure depends on the filling @xmath8 @xcite . in this paper
, we derive a fermi - hubbard model for dipolar fermions including the effects of higher bands .
we show that , even for moderate dipolar strength , it is necessary to take into account the excitations along the @xmath7 direction . simultaneously , in this regime , the interaction induced hopping along the lattice give also important contributions .
this changes the phases expected for a spinless hubbard model including only a single band .
near @xmath9 , we find a spontaneous appearance of non - fermi liquid behaviour in the form of smectic metallic phase .
near @xmath10 , we find that the system can be mapped to an extended pseudo - spin @xmath11 hubbard model with different emergent lattice configuration .
we find a regime where chiral @xmath12-wave superconductivity emerges through kohn - luttinger ( kl ) mechanism with transition temperature @xmath13 of the order of tunneling .
this gives rise to an exotic supersolid , with the diagonal long - range order provided by the checkerboard pattern of the lower orbital fermions , while the superfluidity originating from the fermions in the higher band .
the paper is organized as follows : in section ii we have introduced a multi - orbital model to describe dipolar fermions in optical lattices .
we then discuss quantitatively the contributions of different parameters present in the model . in section
iii we have described the energy contribution of different crystal structures in the limit of vanishing tunneling .
we also compare the corresponding energies of such crystal states without taking into account the higher bands and show that it is necessary to take into account the higher band contributions for experimentally realizable parameters . in section iv , we have investigated the ground state properties for filling greater than @xmath14 .
we find that due to the higher band occupation dependent tunneling contributions , within certain parameter regime , there is a spontaneous formation of smectic - metal phase , along with stripe - like phases . in section
v we describe the ground state structures for @xmath15 .
we find that the higher - band tunneling can give rise to sub - lattices which further can give rise to @xmath12-wave superfluidity . in section
vi we present our conclusions followed by acknowledgements in section vii .
the hamiltonian for the dipolar fermions in the second quantized form reads @xmath16 where @xmath17 is a spinless fermion field operator . in the units of recoil energy @xmath18
, the single particle hamiltonian becomes @xmath19 and the long - rage interaction potential @xmath20 , where @xmath21 is a dimensionless dipolar strength , related to the electric dipolar moment @xmath22 .
for krb molecules with a dipole moment of @xmath23 debye confined in the optical lattice with @xmath24 @xcite one gets @xmath25 whereas , for similar lattice parameters , lics molecules can have a dipole moment of @xmath26 debye with @xmath27 .
we decompose the field operator in the basis of wannier functions in the @xmath28 directions and of harmonic oscillator eigenstates in @xmath7 direction . for convenience
we introduce orbital index @xmath29 denoting @xmath12 , @xmath5 and @xmath30 excitations in @xmath31 , @xmath32 , and @xmath7 direction respectively . in this basis the field operator @xmath33 , where @xmath34 is the single - particle wave - function in orbital @xmath35 localized on site @xmath36 ( @xmath37 and @xmath38 are unit vectors in the proper directions ) .
fermionic operator @xmath39 annihilates particle in this state .
the hamiltonian can be rewritten in the following hubbard - like form @xmath40 where [ hhubbard ] @xmath41 parameters @xmath42 and @xmath43 comes from the single particle hamiltonian and denote single - particle energy and nearest - neighbour tunneling in orbital @xmath35 respectively .
the inter - particle interaction has three contributions to the hamiltonian : _
( i ) _ the on - site interaction energy of fermions occupying different orbitals @xmath35 and @xmath44 of the same site @xmath45 , _
( ii ) _ the long - range interaction energy of fermions occupying orbitals @xmath35 and @xmath44 of different sites @xmath46 , _ ( iii ) _ and the tunneling from orbital @xmath44 at site @xmath47 to the orbital @xmath35 at site @xmath48 induced by presence of an additional fermion at site @xmath48 in orbital @xmath49 denoted by @xmath50 .
the hamiltonian is very general . to get a physical understanding of its properties , we start by examining the properties of density - density interactions . we calculate the interactions between few lowest bands : @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 , and @xmath56 .
we find that the on - site interactions @xmath57 is always repulsive .
it means that putting two fermions in @xmath58 and @xmath59 or @xmath60 band simultaneously is energetically unfavorable .
remarkably we find that @xmath61 is always negative .
this surprising attraction stems from the presence of the fermionic exchange term and the shape of the dipolar interactions ( see appendix a. ) .
moreover , as it is seen from fig .
[ interaction]a , this interaction can not be neglected even for stronger confinements in @xmath7 directions . for higher orbitals
we find that @xmath62 .
in addition for long - range interactions we find that @xmath63 . from this analysis
we conclude that for polar molecules there always exists some critical dipolar strength for which single - band approximation breaks down since two particles can occupy the same site .
this critical behavior is controlled by the on - site energy cost @xmath64 where the energy gap between @xmath58 and @xmath65 orbital is given by @xmath66 .
we find that @xmath67 for @xmath68 .
thus it is important to take into account atleast @xmath58 and @xmath65 orbital to describe the dipolar fermions . in this paper
we consider situations when @xmath69 is positive due to which three or more fermions in same site is unfavourable . as functions of the lattice confinement .
( a ) on - site interaction @xmath70 ( blue solid line -1- ) , and nearest - neighbor long - range interactions @xmath71 ( solid black line -2- ) , @xmath72 ( dashed black line -3- ) , and @xmath73 ( dotted black line -4- ) .
the red line -5- shows the ratio @xmath74 .
( b ) magnitudes of the induced tunneling terms @xmath75 ( black solid line -1- ) , @xmath76 ( red solid line -2- ) , and @xmath77 ( blue solid line -3- ) . the dashed black line -4- denotes the ratio @xmath78 for @xmath79 . ]
next we discuss the role of the interaction induced tunnelings in hamiltonian .
counter intuitively the most important contribution does not come from the induced tunneling in @xmath65 band ( @xmath80 ) , but from the inter - band tunneling which changes @xmath65 orbital to the @xmath81 and @xmath82 ones ( @xmath83 ) . note
, that this inter - band tunneling is absent for usual single - particle tunneling due to the properties of single particle hamiltonian @xcite . from properties of @xmath12 orbital states
it follows that @xmath84 .
the relation of this term to other interaction - induced tunnelings is shown in fig .
[ interaction]b . from the above analysis we introduce simplified , but realistic model of polar fermi molecules confined in 2d optical lattice by taking into account effects of interactions between orbitals @xmath85 .
to obtain an idea about the ground state structures , in this section we compare the energies of different possible ground - state crystal configurations for specific filling factors with and without the contributions from the @xmath65-orbitals .
pictorial diagram of the different checker board lattices for @xmath86 .
the blue spheres denote @xmath58-orbital fermions and the smaller black spheres denote @xmath65 orbital fermions .
( a ) ground state crystal phase of hamiltonian .
( b ) @xmath14 checkerboard lattice of @xmath58-band fermions and extra @xmath65 fermions with density @xmath87 .
( c ) density - wave structure of the effective bosons with filling @xmath88 corresponding to the ground state structure of the hamiltonian .
( d ) the energies @xmath89 ( thick - solid line ) , @xmath90 ( dashed line ) , and @xmath91 ( thin - solid line ) as functions of the trap frequency @xmath92 for dipolar strength @xmath79.,title="fig : " ] pictorial diagram of the different checker board lattices for @xmath86 .
the blue spheres denote @xmath58-orbital fermions and the smaller black spheres denote @xmath65 orbital fermions .
( a ) ground state crystal phase of hamiltonian .
( b ) @xmath14 checkerboard lattice of @xmath58-band fermions and extra @xmath65 fermions with density @xmath87 .
( c ) density - wave structure of the effective bosons with filling @xmath88 corresponding to the ground state structure of the hamiltonian .
( d ) the energies @xmath89 ( thick - solid line ) , @xmath90 ( dashed line ) , and @xmath91 ( thin - solid line ) as functions of the trap frequency @xmath92 for dipolar strength @xmath79.,title="fig : " ] for the clarity of discussion , we first neglect the tunneling terms as justified in the strongly coupled regime . without the higher orbital effects at most one fermion can occupy a given site .
the corresponding hamiltonian reads : @xmath93 with the dipolar interaction in the @xmath58 band @xmath94 . by taking the orbital effects into account ,
the corresponding hamiltonian is defined in eq .
( 1a ) and ( 1b ) , @xmath95 where @xmath35 denotes the @xmath58- and @xmath65 orbital fermions .
now we consider the situation , when each occupied site contains two fermions . in this case
we can define a corresponding hardcore bosonic operator at site @xmath48 as @xmath96 and @xmath97 and the bosonic number operator @xmath98 . from this
we can see that @xmath99 as the number of fermions is twice the number of bosons .
subsequently , we can write an effective bosonic hamiltonian as : @xmath100 where @xmath101 . here
@xmath64 is the energy cost of having a composite boson .
is similar to the bosonic dipolar system with modified dipolar interaction and can simulate the crystal phases of dipolar bosons @xcite .
for concreteness we first specifically choose @xmath86 . at filling @xmath86 the ground state of the single band hamiltonian forms a crystal structure in accordance with @xcite
and it is shown in fig .
[ lattice1]a .
its energy is @xmath89 . in the current paper , we analyze other structures as a ground states corresponding to the full hamiltonian .
two such structures are presented in fig .
[ lattice1]b and [ lattice1]c with corresponding energies @xmath90 and @xmath91 . in the @xmath102 structure
the @xmath58-band fermions form a @xmath14 crystal structure and remaining @xmath87 @xmath12-orbital fermions occupy already occupied sites . the third possible ground state candidate fig .
[ lattice1]c comes from the effective bosonic hamiltonian at filling @xmath103 .
we compare energies of these three structures by plotting them as functions of the harmonic trapping frequency for a dipolar strength @xmath79 ( fig .
[ lattice1]d ) .
we find that the energy of the structure @xmath104 is almost insensitive to the trapping frequency @xmath6 .
moreover , the structure is the lowest energy state ( the true ground state of the system ) only for large enough @xmath6 ( @xmath105 for studied case ) . for lower trap frequencies
we find that structure @xmath102 ( @xmath106 ) or @xmath107 ( @xmath108 ) becomes a ground states of the system .
we also note that in the structure @xmath107 , within the bosonic subspace , tunneling can arise in second order processes and it is much lower than the binding energy of the bosons .
we have also checked that , for filling factors between @xmath109 and @xmath86 , the energy of the configuration @xmath102 is lower than the energy of the phase - separated structures of single - band hamiltonian .
similarly we can infer also the ground state structures at filling @xmath110 as the situation is very simmilar to the filling @xmath86 . the ground state of the single band hamiltonian shown in fig .
[ lattice2]a with corresponding energy @xmath111 is a true ground state of the system only for large enough @xmath6 . for lower confinement frequencies the ground state is _ (
i ) _ a @xmath11 checkerboard @xmath58-band crystal with @xmath12-band fermions ( with density @xmath112 ) moving on the occupied sites ( energy @xmath113 and fig .
[ lattice2]b ) or _
( ii ) _ @xmath114 stripe structure of composite bosons ( energy @xmath115 and fig .
[ lattice2]c ) @xcite
. pictorial diagram of the different checkerboard lattices for @xmath110 .
the blue spheres denote @xmath58-orbital fermions and the smaller black spheres denote @xmath65 orbital fermions .
( a ) ground state crystal phase of hamiltonian .
( b ) @xmath11 checkerboard lattice of @xmath58-band fermions and extra @xmath65 fermions with density @xmath112 .
( c ) density - wave structure of the composite bosons with filling @xmath116 corresponding to the ground state structure of the hamiltonian .
( d ) the energies @xmath111 ( thick - solid line ) , @xmath113 ( dashed line ) , and @xmath115 ( thin - solid line ) as functions of the trap frequency @xmath92 for dipolar strength @xmath117.,title="fig : " ] pictorial diagram of the different checkerboard lattices for @xmath110 .
the blue spheres denote @xmath58-orbital fermions and the smaller black spheres denote @xmath65 orbital fermions .
( a ) ground state crystal phase of hamiltonian .
( b ) @xmath11 checkerboard lattice of @xmath58-band fermions and extra @xmath65 fermions with density @xmath112 .
( c ) density - wave structure of the composite bosons with filling @xmath116 corresponding to the ground state structure of the hamiltonian .
( d ) the energies @xmath111 ( thick - solid line ) , @xmath113 ( dashed line ) , and @xmath115 ( thin - solid line ) as functions of the trap frequency @xmath92 for dipolar strength @xmath117.,title="fig : " ] similar results are also obtained for other filling fractions , namely @xmath118 . for these filling fractions
we also find that below a certain critical trapping strength @xmath6 , for critical @xmath119 , it is important to take into account the excited trap states .
in this section we will look into the properties of the ground states near @xmath109 filling .
we show that the presence of higher orbitals not only changes the ground - state crystal structures , it also fundamentally changes the properties of such states .
specifically we show that new forms of matter , like smectic metal phase , can spontaneously form due to the effect of higher orbitals .
first , here we consider the case when @xmath120 , therefore for low filling all fermions occupy only the @xmath58 orbital states . for filling @xmath109 and large enough @xmath119 ( @xmath121 )
there is non vanishing single - particle excitation gap and the system is in the @xmath58-band insulator state ( denoted by blue spheres in fig.[lattice]a ) @xcite .
situation change dramatically for higher fillings .
it can be simply understood with energy arguments .
the energy cost of putting additional particle in the vacant site is given by @xmath122 .
in contrast the cost of putting additional particle to the @xmath65 orbital of an occupied site @xmath123 . for @xmath119 larger than some critical strength
one finds that @xmath124 . as an example
, such conditions are fulfilled for @xmath125 , @xmath79 , and @xmath126 .
consequently , additional particles start to fill @xmath65 band of previously occupied sites . in this scenario energy
conserving dynamics of the system comes from the second - order processes involving tunneling to the next occupied site ( along @xmath31 direction in fig .
[ lattice]a ) . to the leading order ,
this effective tunneling is given by @xmath127 thus , the @xmath65 fermions will only move along one direction chosen by the insulator checkerboard geometry in @xmath58-band , in our case along @xmath37 .
the resulting system can be thought as stacks of one - dimensional chains or stripes placed along @xmath38 without inter - chain tunnelings .
the effective hamiltonian governing the @xmath65 fermions can be written as @xmath128 with intra - chain hamiltonian @xmath129 and inter - chain hamiltonian @xmath130 where @xmath131 and @xmath132 are creation and annihilation operators of @xmath65 fermions on @xmath58-fermion occupied site @xmath133 on chain @xmath30 .
the intra - chain and inter - chain interactions are given by @xmath134\boldsymbol{e}_x)$ ] and @xmath135\boldsymbol{e}_x+[l - l']\boldsymbol{e}_y)$ ] respectively .
the ground state structure of this coupled - chains system is investigated by introducing the bosonized fields @xmath136 related to the fermi operator @xmath137 rewritten in the continuum limit as @xmath138 @xcite . near the left and right fermi momenta @xmath139 ,
we can write @xmath140/\sqrt{2\pi\epsilon}$ ] , where @xmath141 is a cutoff length and @xmath142 are klein factors . the fermi momentum is given by the density of @xmath65 fermions which in terms of total density @xmath8 reads , @xmath143 . by writing the bosonized phase field @xmath144 in terms of its fourier transform @xmath145 along the @xmath38
the lagrangian for the system reads @xmath146.\ ] ] the interaction parameter @xmath147 and sound velocity @xmath148 are determined by the details of the dipolar interactions ( see appendix c ) .
there is an additional inter - chain @xmath65 fermion cdw coupling @xmath149 .
consequently , the scaling dimension of the cdw operator is given by @xmath150 .
when @xmath151 the cdw operator is irrelevant . then the stable phase has properties similar to 1d luttinger liquid with low - energy bosonic collective excitations .
this state preserves the smectic symmetry @xmath152 , with @xmath153 constant on each chain .
this phase is known as _ smectic - metal _
phase @xcite as there metallic behavior along the chain with insulating density wave order along transverse direction .
this phase is a peculiar example of spontaneous emergence of non - fermi liquid behaviour in two - dimensional fermi systems .
in contrast , when @xmath154 then @xmath65 fermions becomes unstable towards formation of stripe crystals . in fig .
[ lattice]b we plot @xmath155 and @xmath156 as functions of total density @xmath8 for @xmath79 and @xmath157 .
it is clear that for @xmath158 there is a _ smectic - metallic _ phase while for @xmath159 , the system goes to a stripe - crystal phase .
in this section let us discuss the case of filling @xmath160 where for low dipolar strength @xmath119 , due to the same reasons as before fermions will occupy only the @xmath58-band and the ground state of the system is the checkerboard insulator ( see supplimentary sec .
d. ) as denoted by filled - blue and open - red spheres in fig .
. pictorial diagram of the different checker board lattices .
the filled - blue and open - red spheres denotes @xmath58-orbital fermions and the smaller black sphere denotes @xmath65 orbital fermions . ( a ) checkerboard lattice at @xmath109 filling .
the @xmath65 fermions will move with effective tunneling @xmath161 ( arrow 1 ) only along the shaded region making a stack of 1d chains .
interaction between neighboring @xmath12-band fermions is equal to @xmath162 ( arrow 2 ) .
( b ) scaling dimensions @xmath155 ( solid - line ) and @xmath156 ( dashed - line ) as functions of the total density @xmath8 for @xmath79 and @xmath157 .
( c ) checkerboard lattice at @xmath160 .
the blue and thick - red lines constitute two different sub - lattices .
they are not coupled via tunneling processes since the tunneling @xmath163 ( arrow 3 ) is much smaller than @xmath161 . ] to look for properties of the system with additional particles , we define deviation from half - filling @xmath164 and we introduce corresponding chemical potential @xmath165 . from energy arguments we find that two scenario can happen .
the additional fermion _ ( i ) _ occupies a vacant site with energy cost @xmath166 or _
( ii ) _ it goes to the @xmath65 orbital of an occupied site with energy cost @xmath167 . consequently , in the second scenario ( when @xmath168 )
, all extra fermions will occupy the @xmath65 orbitals of the already occupied sites . as an example
, such conditions are fulfilled for @xmath125 , @xmath117 , and @xmath169 .
in such a case @xmath170 corresponds to the filling of @xmath65 band fermions .
the parallel tunneling of the @xmath65 fermions between the occupied sites will again arise from the second order processes .
moreover , tunneling to the diagonally occupied site @xmath171 ^ 2/|u_{s , p_z}|$ ] for @xmath172 it is @xmath173 times smaller than @xmath161 .
consequently fermions in the @xmath65 orbitals can move in independent square sub - lattices ( either the thick - red or blue sub - lattice shown in the fig .
[ lattice]b ) .
note , that fermions can not tunnel between different sub - lattices .
thus we can describe the system of the @xmath65 fermions in the blue ( thick - red ) lattice as pseudo - spin up ( down ) . by introducing operators @xmath174 , where @xmath175 the effective hamiltonian can be written as @xmath176 with @xmath177}\hat{n}_{\boldsymbol{i}\uparrow}\,\hat{n}_{\boldsymbol{j}\downarrow},\ ] ] where @xmath178 . for convenience
we introduce @xmath179 and @xmath180 .
note , that now @xmath181 is understood as a nearest - neighbor in a given sub - lattice .
nearest - neighbors between different sub - lattices is denoted by @xmath182 $ ] . the modified lattice constant of the sub - lattices is @xmath183 . in this way we are able to study the system properties with the weak - coupling theory .
we investigate the emergence of triplet superconductivity between the same pseudo - spin fermions , arising via kl mechanism @xcite ( magnetic instabilities are discussed in appendix d. ) .
we look for cooper pairs with chiral @xmath12-wave symmetry .
the effective interaction between fermions in kl mechanism in terms of the scattering momentum @xmath184 can be written as @xmath185 , \nonumber\end{aligned}\ ] ] where @xmath186 , @xmath187 is the fermi distribution function , @xmath188 is the dispersion and @xmath189 and @xmath190 .
the summation in comes from taking into account the second - order terms represented by diagrams shown in fig .
[ diagrams]a .
the two terms inside the first bracket in comes from the top - left diagram in fig [ diagrams]a , while the next two terms comes from the top - right and bottom - left diagrams representing vertex corrections .
the last term in comes from the bottom - right diagram in fig [ diagrams]a denoting exchange interactions . by performing the integration over the momentum in the limit of @xmath191
, we finally get antisymmetric part of effective coupling @xmath192 where @xmath193 . functions @xmath194 and @xmath195
originate in the second - order corrections and their detailed forms are given in the appendix e. the point is that , due to the van - hove singularity in density of states , function @xmath195 contains a logarithmic divergence . at the same time function @xmath194
is analytical due to the dressing of the density of states .
this means that there always exists finite critical @xmath196 above which the interaction is attractive and superfluidity appears . from the bcs theory
one can get an estimate of the transition temperature @xmath13 ( derivation is shown in appendix e. ) . in fig .
[ diagrams]b we plot the transition temperature @xmath13 as a function of deviation @xmath170 for example parameters discussed previously .
for @xmath197 we get @xmath198 ( @xmath199 ) .
thus the ground state has a checkerboard density pattern due to the @xmath58 fermions and @xmath12-wave superfluid @xmath65 fermions at temperature below @xmath13 . ) .
the dashed lines denote interaction and the solid lines denote fermion propagator .
( b ) @xmath12-wave superfluid transition temperature @xmath13 as a function of density @xmath200 .
[ diagrams ] ]
in conclusion , we have derived a generalized hubbard model for dipolar fermions in an optical lattice by taking into account higher orbitals .
we have shown that the effect of these higher orbitals leads to new phenomena .
due to the strong interaction - dependent hopping terms in higher orbitals , these systems can be described by effective weakly - interacting theories .
for particular parameters , near @xmath9 , we found a cross - over to the one - dimensional physics resulting in simultaneous metallic and density wave properties .
for @xmath201 , the @xmath58 fermion checkerboard order is given by the configuration in fig ( 2)a .
as @xmath202 is increased there will be single - particle and dipole excitations at different regions of the @xmath109 checkerboard crystal similar to the one considered in ref .
@xcite for wigner - hubbard crystals due to coloumb interaction .
these excitations can induce inter - chain tunneling at different regions and the resulting model will be subject of future study . for other
set of parameters , @xmath15 , the system can be described by a weakly interacting hubbard model with pseudo - spin originating from the lattice geometry . using the kl theory
, we found a transition to the chiral @xmath12-wave superfluidity due to the @xmath65 fermions without destroying the checkerboard order created by the @xmath58 fermions .
the parameters used here are currently experimentally achievable .
this paper was supported by the eu strep name - quam , ip aqute , erc grant quagatua , spanish micinn ( fis2008 - 00784 and consolider qoit ) , aaii - hubbard , and the national science center grant no .
dec-2011/01/d / st2/02019 .
t.s . acknowledges hospitality from icfo .
99 k. ni , et .
al . , science * 322 * , 231 ( 2008 ) .
j. deiglmayr , et .
al . , faraday discuss . * 142 * , 335 ( 2009 ) .
m. debatin , et .
al . , arxiv:1106.0129 .
j. w. park , et .
al . , arxiv:1110.4552 .
a. v. gorshkov , et .
al . , phys .
lett . * 107 * , 115301 ( 2011 ) .
k. a. kuns , a. m. rey , and a. v. gorshkov , phys .
a * 84 * , 063639 ( 2011 ) .
k. mikelsons , and j. k. freericks , phys . rev .
a * 83 * , 043609 ( 2011 ) .
gadsbolle , and g. m. bruun , phys . rev .
a * 85 * , 021604 .
s. g. bhongale , et .
arxiv : 1111.2873 .
luehmann , o. juergensen , and k. sengstock , new j. phys .
* 14 * , 033021 ( 2012 ) .
a. mering , and m. fleischhauer , phys .
rev . a , * 83 * , 063630 ( 2011 ) .
u. bissbort , f. deuretzbacher , and w. hofstetter , arxiv:1108.6047 .
s. will , et .
al . , nature * 465 * , 197 ( 2010 ) .
o. dutta , et .
al . , new j. phys . * 13 * , 023019 ( 2011 ) .
t. sowiski , et .
al . , phys .
. lett . * 108 * , 115301 ( 2012 ) s. kotochigova and e. tiesinga , phys .
a * 73 * , 041405 ( 2006 ) .
b. capogrosso - sansone , et.al , phys .
* 104 * , 125301 ( 2010 ) .
v. j. emery , e. fradkin , s. a. kivelson , and t. c. lubensky , phys .
lett . * 85 * , 2160 ( 2000 ) .
a. vishwanath and d. carpentier , phys .
lett . * 86 * , 676 ( 2001 ) .
w. kohn , and j. m. luttinger , phys .
lett . * 15 * , 524 ( 1965 ) .
s. fratini and j. merino , phys .
b * 80 * , 165110 ( 2009 ) .
kagan , k. i. kugel , and d. i. khomskii , jetp * 93 * , 415 ( 2001 ) .
khomskii , preprint of the p.n .
lebedev physics institute no .
105 ( 1969 ) .
r. micnas , j. ranninger , s. robaszkiewicz , rev .
phys . * 62 * , 113 ( 1990 ) .
in this section we represent the on - site interaction term @xmath61 and interaction - induced hopping terms @xmath203 and @xmath204 in terms of the single - particle wave - function @xmath205 in orbital @xmath35 localized on site @xmath48 .
orbital index @xmath29 denoting @xmath12 , @xmath5 and @xmath30 excitations in @xmath31 , @xmath32 , and @xmath7 direction respectively .
then the @xmath58 orbitals can be written as @xmath206 where @xmath207 are the lowest band one - dimensional wannier functions and @xmath208 is the ground state wave - function of the harmonic oscillator in the @xmath7 direction .
similarly we can write @xmath209 and @xmath210 . here
@xmath211 is the wannier functions in the first band and @xmath212 is the first excited state of the harmonic oscillator in the @xmath7 direction . for simplicity we took @xmath213 .
from this we can write various parameters as , [ u1 ] @xmath214 the integrations over @xmath215 can be done analytically using convolution theorem in the momentum space . consequently in the momentum space
we get @xmath216,\ ] ] where @xmath217 denotes fourier transform , @xmath218 , @xmath219 is a natural harmonic oscillator length unit , and @xmath220 where @xmath221 denotes complementary error function .
it is important to note that @xmath222 is always negative for any @xmath223 .
this explains the appearance of the attractive on - site interaction for any value of the confinement along the @xmath7 direction .
( a ) pictorial diagram for @xmath224 .
the blue spheres denote @xmath58-orbital fermions and the smaller black spheres denote @xmath65 orbital fermions .
the @xmath58-orbital fermions constitute the underlying @xmath14 checkerboard structure .
the @xmath65 fermions move with the effective tunneling @xmath161 ( arrow 1 ) only along the shaded regions .
interaction between neighboring @xmath12-band fermions is equal to @xmath162 ( arrow 2 ) .
( b ) density - wave structure at filling @xmath225 .
the dark and light shadings denote higher and lower density of the @xmath65 fermions respectively . ] as we explained in the main text , for filling @xmath224 and parameters @xmath125 , @xmath79 , and @xmath226 , the ground state structure is given by @xmath14 checkerboard structure formed by @xmath58-fermions . the @xmath65 fermions ( with density @xmath227 )
move in the occupied sites along @xmath228 direction ( fig .
[ smectic]a ) .
the resulting system can be thought as stacks of one - dimensional chains or stripes placed along @xmath229 without inter - chain tunnelings .
the effective hamiltonian governing the @xmath65 fermions can be written as ( see the main text ) : @xmath230 the bosonized form of the intra - chain lagrangian reads [ lag0 ] @xmath231,\ ] ] where the luttinger liquid parameter @xmath232(2-\cos2\tilde{k}_{1})}{2\pi t_{\mathtt{eff}}^{\parallel}\sin\tilde{k}_1+[v_{p_z , p_z}(2\boldsymbol{e}_x)+\ldots]\cos2\tilde{k}_{1}}\right]^{1/2},\ ] ] and the sound velocity @xmath233\right)^2\nonumber\\ & -&[v_{p_z , p_z}(2\boldsymbol{e}_x)+\ldots]^2(1-\cos2\tilde{k}_{1})^2.\end{aligned}\ ] ] next we include the bosonized form of the inter - chain hamiltonian which results in the the total lagrangian @xmath234 where @xmath235.\ ] ] here the modified luttinger parameter is given by @xmath236\cos(q_y)+[v_{p_z , p_z}(4\boldsymbol{e}_y)+\ldots]\cos(2q_y)}{2\pi t_{\mathtt{eff}}^{\parallel}\sin\tilde{k}_f+[v_{p_z , p_z}(2\boldsymbol{e}_x)+\ldots](2-\cos(2\tilde{k}_{f1}))}\right]^{1/2},\ ] ] and sound velocity @xmath237 .
inter - chain interactions induce additional charge - density wave ( cdw ) perturbation , @xmath238 with @xmath239\sum_{l } \cos[2\sqrt{\pi}(\theta_l-\theta_{l+1})],\\ { \cal l}_{\rm cdw,2}&= & \frac{1}{u}\sum_n v_{p_z , p_z}(2n\boldsymbol{e}_x+4\boldsymbol{e}_y ) \cos[(2n)\tilde{k}_{1 } ] \sum_{l } \cos[2\sqrt{\pi}(\theta_l-\theta_{l+2})].\end{aligned}\ ] ] at half - filling , i.e. @xmath240 , we see that @xmath241 .
it means that the charge - density wave instability is induced by the next - nearest neighbour inter - chain interaction .
we checked that this interaction is much weaker than the tunneling @xmath161 .
it means that the smectic - metal phase discussed in the paper will be stable till low enough temperature .
to look into the ground state of @xmath58-orbital fermions at @xmath160 , we express the average density @xmath242 , where @xmath243 is the order parameter .
we also define the single - particle imaginary time green functions @xmath244 , where @xmath245 denotes time - ordering . by following the procedure described in @xcite
we find the following equations for @xmath246 in the momentum space [ g1g2 ] @xmath247g_1(\boldsymbol{k},\omega)-\epsilon_{\boldsymbol{k}}g_2(\boldsymbol{k},\omega ) & = 1 , \\ \left[\omega+\mu -2v(1+\delta')\right]g_2(\boldsymbol{k},\omega)-\epsilon_{\boldsymbol{k}}g_1(\boldsymbol{k},\omega ) & = 0,\end{aligned}\ ] ] where the kinetic energy @xmath248 , the effective potential @xmath249 , and @xmath250 . in the position space
@xmath251 ( @xmath252 ) is equal to @xmath253 for @xmath254 even ( odd ) .
these mean field equations for @xmath255 and @xmath256 are similar to the ones found for extended hubbard model , with a renormalized nearest neighbour interaction and density imbalance @xcite .
then , by solving equations , we find that in the strong coupling limit @xmath257 ( @xmath258 for @xmath117 and @xmath259 ) .
thus our assumption of a checkerboard lattice with alternative sites occupied ( like the one in fig .
2(c ) in the main text ) is justified .
in this section we discuss the appearance of stoner ferromagnetism and charge - density wave ( cdw ) instability of the @xmath65 fermions . to do this we transform to momentum space and introduce charge fluctuations @xmath261 and spin fluctuations @xmath262 operators ( @xmath263 ) .
subsequently we rewrite eq .
( 4 ) from the main text in the momentum space as @xmath264 where @xmath189 and @xmath190 .
the system with interactions described by can manifest three possible magnetic instabilities : charge - density wave ( cdw ) , spin - density wave ( sdw ) , and ferromagnetic instability . at @xmath265
( each sub - lattice is half - filled with @xmath65-orbital fermions ) , as @xmath266 for nesting vector @xmath267 , interaction in the spin channel becomes repulsive and therefore sdw order is absent . in the spin channel ,
the onset of an stoner ferromagnetism is given by the divergence of susceptibility with momentum @xmath268 .
this condition can be written as @xmath269 , where @xmath270 $ ] , and the bare susceptibility @xmath271 with @xmath272 . here
@xmath187 is the fermi distribution function and @xmath188 is the dispersion relation .
consequently , in the limit of @xmath273 , @xmath274 where the two - dimensional density of states @xmath275 with @xmath276 being an elliptic integral of first kind . substituting the density of states we get , @xmath277 .
as @xmath278 , or density of @xmath65 fermions are near @xmath14 , due to the logarithmic divergence of @xmath279 , the transition temperature for the stoner ferromagnetism is given by @xmath280.\ ] ] the case @xmath125 , around @xmath281 , @xmath282 , corresponds to a very low stoner temperature @xmath283 .
next we discuss the checkerboard charge - density wave structure due to the @xmath65 fermions at @xmath225 .
due to nesting , each fermionic component will be unstable towards cdw . in the density channel ,
the onset of an cdw is indicated by the divergence of susceptibility with momentum @xmath284 .
the condition for @xmath65 fermion cdw can be written as @xmath285 , where @xmath286=-2v_{\uparrow\uparrow},\ ] ] and the bare susceptibility @xmath287 subsequently , the transition temperature to the cdw is given by @xmath288 .
for example , when @xmath125 , @xmath289 , and @xmath117 we find that @xmath290 ( @xmath291 ) .
one should note that due to the relative shift of sub - lattices the resulting density modulation in this phase looks like stripes rather than the standard checkerboard structure as shown in fig .
[ smectic]b .
in this section we derive the @xmath12-wave interaction from the kohn - luttinger effective interaction in terms of the scattering momentum @xmath292 . for this purpose
we rewrite eq .
( 6 ) from the original paper @xmath293 , \nonumber\\ & & \end{aligned}\ ] ] where @xmath294 with @xmath295 being the fermi distribution function .
first we put the expression of @xmath296 and @xmath297 back to .
as we are interested in @xmath12-wave interaction , after expanding in terms of the momenta @xmath298 , @xmath299 , @xmath300 , @xmath301 , we keep terms proportional to @xmath302 . in this way
we get @xmath303 \right ) ( \sin k_x\tilde{a}\sin k'_x\tilde{a}+\sin k_y\tilde{a}\sin k'_y\tilde{a } ) \nonumber \\ & + 4(v_{\uparrow\uparrow})^2 \sum_{\boldsymbol{p}}q_{\boldsymbol{q},\boldsymbol{p } } ( \sin k_x\tilde{a}\sin k'_x\tilde{a}\sin^2 p_x\tilde{a}+\sin k_y\tilde{a}\sin k'_y\tilde{a}\sin^2p_y\tilde{a})\end{aligned}\ ] ] where @xmath304 . by converting sums to integrals in
we have to compute terms of the form @xmath305 , @xmath306 , and @xmath307 . in the limit of vanishing temperature @xmath308
we approximate all integrals @xmath309 for arbitrary function @xmath310 .
the effective density of states reads @xmath311 .
we see that the only first term inside the third bracket in is not dressed by @xmath312 or @xmath313 .
hence the effective density of states for this term contains van - hove singularity .
all other terms with in , due to the dressing by @xmath312 or @xmath313 , are analytic . for convenience , we re - express @xmath314 , where @xmath315 .
the functions @xmath194 and @xmath195 are given by where @xmath317 with @xmath276 being the elliptic integral of the first kind and @xmath318 is the elliptic integral of the second kind .
then we can write the bcs equation for the transition temperature @xmath13 as @xcite @xmath319 ^ 2 \right)^{1/2}\frac{4t_{\mathtt{eff}}^{\parallel}}{t_c } \right |=1,\ ] ] where the effective density of state is given by @xmath320 as discussed in the main text , the ground state has a checkerboard density pattern due to the @xmath58 fermions and @xmath12-wave superfluid @xmath65 fermions at temperature below @xmath13 . from previous section
, we see that at @xmath225 ( half filled @xmath65 fermions for the red and blue lattices ) , the transition temperatures for the @xmath65 fermion cdw and @xmath65 fermion superfluidity are similar .
this will result in a competition or co - existance between both instabilities for the @xmath65 orbital fermions .
a detailed account of such scenario is beyond this scope of this paper . | we study a system of polar dipolar fermions in a two - dimensional optical lattice and show that multi - band fermi - hubbard model is necessary to discuss such system . by taking into account both on - site , and long - range interactions between different bands , as well as occupation - dependent inter- and intra - band tunneling
, we predict appearance of novel phases in the strongly - interacting limit . |
in t hooft s large @xmath2 limit @xcite gauge theories are expected to be drastically simplified .
thus , in this limit the gauge theory diagrams are organized in terms of riemann surfaces , where each extra handle on the surface suppresses the corresponding diagram by @xmath3 .
the large @xmath2 expansion , therefore , resembles perturbative expansion in string theory . in the case of four - dimensional gauge theories
this connection can be made precise in the context of type iib string theory in the presence of a large number @xmath2 of d3-branes @xcite .
thus , we consider a limit where @xmath4 , @xmath5 and @xmath6 , while keeping @xmath7 fixed , where @xmath8 is the type iib string coupling .
note that in this context a world - sheet with @xmath9 handles and @xmath10 boundaries is weighted with @xmath11 once we identify @xmath12 , this is the same as the large @xmath2 expansion considered by t hooft . note that for this expansion to make sense we must keep @xmath13 at a small value @xmath14 . in this regime
we can map the string diagrams directly to ( various sums of ) large @xmath2 feynman diagrams .
note , in particular , that the genus @xmath15 planar diagrams dominate in the large @xmath2 limit , then no matter how large @xmath2 is , for sufficiently many boundaries the higher genus terms become relevant , and we lose the genus expansion .
in fact , in this regime one expects an effective supergravity description to take over as discussed in @xcite . ] .
if the space transverse to the d3-branes in the setup of @xcite is @xmath16 , then we obtain the @xmath17 supersymmetric @xmath18 gauge theory on the d3-branes , which is conformal . on the other hand
, we can also consider orbifolds of @xmath16 , which leads to gauge theories with reduced supersymmetry versions of these orbifold theories via the compactifications of type iib on ads@xmath19 ( where @xmath20 is the orbifold group ) were originally discussed in @xcite . ] .
as was shown in @xcite , if we cancel all twisted tadpoles in such models , in the large @xmath2 limit the corresponding @xmath21 gauge theories are conformal .
moreover , in the planar limit the ( on - shell ) correlation functions in such theories are the same as in the parent @xmath17 gauge theory . in this paper
we discuss non - conformal gauge theories within the setup of @xcite .
such theories can be obtained by allowing some twisted tadpoles to be non - vanishing . in particular
, we can have consistent embeddings of non - conformal gauge theories if we allow logarithmic tadpoles , which correspond to the twisted sectors with fixed point loci of real dimension two .
in particular , even though the corresponding string backgrounds are not finite ( in the sense that we have logarithmic ultra - violet divergences ) , they are still consistent as far as the gauge theories are concerned , and the divergences correspond to the running in the four - dimensional gauge theories on the d3-branes .
regularization of the aforementioned divergences can be conveniently discussed in the context of what we refer to as the brane - bulk duality , which is a consequence of the open - closed string duality . in particular , in certain non - trivial @xmath0 cases in the planar limit the corresponding gauge theories perturbatively are not renormalized beyond one - loop .
in fact , in this limit the ( on - shell ) correlation functions in these theories are the same as in the parent @xmath1 non - conformal gauge theories . in the @xmath1 as well as
the aforementioned @xmath0 cases the brane - bulk duality is particularly simple , and implies that the quantum corrections in the corresponding gauge theories are encoded in classical higher dimensional field theories whose actions contain the twisted fields with non - vanishing tadpoles .
in particular , various quantum corrections can be obtained via integrating out the bulk fields in the corresponding classical action , that is , by considering the self - interaction of the d3-branes via the bulk fields .
we give explicit computations in various @xmath21 examples in this context , including the treatment of divergences .
we also discuss whether the brane - bulk duality can be extended to the non - perturbative level in the aforementioned theories . in the @xmath1 cases we argue that , since we are working in the large @xmath2 limit , the low energy effective action does not receive non - perturbative corrections .
we also conjecture that this should be the case for the corresponding @xmath0 theories as well . in the @xmath22 cases we verify that there are no non - perturbative corrections to the superpotential in these theories in the large @xmath2 limit .
the remainder of this paper is organized as follows . in section
ii we discuss our setup .
in section iii we discuss non - conformal large @xmath2 gauge theories which can be constructed within this setup . in section
iv we discuss the large @xmath2 limit and brane - bulk - duality . in sections v , vi and vii we give details of classical computations that in the context of the brane - bulk duality reproduce quantum results in the corresponding @xmath1 , @xmath22 and @xmath23 gauge theories , respectively . in section viii
we comment on the non - perturbative extension of the brane - bulk duality .
in section ix we give a few concluding remarks . in appendix a we compute the brane - bulk couplings used in sections v , vi and vii .
in this section we discuss the setup within which we will consider four - dimensional large @xmath2 gauge theories in the context of brane - bulk duality .
parts of our discussion in this section closely follow @xcite . thus , consider type iib string theory in the presence of @xmath2 coincident d3-branes with the space transverse to the d - branes @xmath24 .
the orbifold group @xmath25 ( @xmath26 ) must be a finite discrete subgroup of @xmath27 . if @xmath28 , we have @xmath22 ( @xmath1 ) unbroken supersymmetry , and @xmath23 , otherwise
. we will confine our attention to the cases where type iib on @xmath29 is a modular invariant theory .
if @xmath30 , then modular invariance requires that the set of points in @xmath16 fixed under the @xmath31 twist has real dimension 2 . ] .
the action of the orbifold on the coordinates @xmath32 ( @xmath33 ) on @xmath29 can be described in terms of @xmath34 matrices : @xmath35 .
the world - sheet fermionic superpartners of @xmath32 transform in the same way .
we also need to specify the action of the orbifold group on the chan - paton charges carried by the d3-branes .
it is described by @xmath36 matrices @xmath37 that form a representation of @xmath20 .
note that @xmath38 is an identity matrix , and @xmath39 .
the d - brane sector of the theory is described by an oriented open string theory .
in particular , the world - sheet expansion corresponds to summing over oriented riemann surfaces with arbitrary genus @xmath9 and arbitrary number of boundaries @xmath10 , where the boundaries of the world - sheet are mapped to the d3-brane world - volume .
moreover , we must consider various `` twists '' around the cycles of the riemann surface .
the choice of these `` twists '' corresponds to a choice of homomorphism of the fundamental group of the riemann surface with boundaries to @xmath20 .
for example , consider the one - loop vacuum amplitude ( @xmath40 ) .
the corresponding graph is an annulus whose boundaries lie on d3-branes .
the annulus amplitude is given by @xmath41 the one - loop partition function @xmath42 in the light - cone gauge is given by @xmath43 where @xmath44 is the light - cone hamiltonian , @xmath45 is the fermion number operator , @xmath46 is the real modular parameter on the cylinder , and the trace includes sum over the chan - paton factors .
the states in the neveu - schwarz ( ns ) sector are space - time bosons and enter the partition function with weight @xmath47 , whereas the states in the ramond ( r ) sector are space - time fermions and contribute with weight @xmath48 .
the elements @xmath49 acting in the hilbert space of open strings act both on the left end and the right end of the open string .
this action corresponds to @xmath50 acting on the chan - paton indices .
the partition function ( [ partition ] ) , therefore , has the following form : @xmath51 where @xmath52 are characters corresponding to the world - sheet degrees of freedom .
the `` untwisted '' character @xmath53 is the same as in the @xmath17 theory for which @xmath54 .
the information about the fact that the orbifold theory has reduced supersymmetry is encoded in the `` twisted '' characters @xmath52 , @xmath55 . in this subsection
we discuss one - loop tadpoles arising in the above setup . as was pointed out in @xcite ,
if all tadpoles are canceled , then the resulting theory is finite in the large @xmath2 limit .
however , not all tadpoles need to be canceled to have a consistent four - dimensional gauge theory .
in fact , we can obtain non - conformal gauge theories if we allow such tadpoles .
the characters @xmath52 in ( [ gsq ] ) are given by @xmath56^{2+d_a } } \left[{\cal x}_a(e^{-2\pi t})-{\cal y}_a(e^{-2\pi t})\right]~,\ ] ] where @xmath57 is the real dimension of the set of points in @xmath16 fixed under the twist @xmath49 .
the factor of @xmath58 in the denominator comes from the bosonic zero modes corresponding to four directions along the d3-brane world - volume .
two of the @xmath59-functions come from the bosonic oscillators corresponding to two spatial directions along the d3-brane world - volume ( the time - like and longitudinal contributions are absent as we are working in the light - cone gauge ) .
the other @xmath57 @xmath59-functions come from the bosonic oscillators corresponding to the directions transverse to the d - branes untouched by the orbifold twist @xmath49 .
finally , the characters @xmath60 , @xmath61 correspond to the contributions of the world - sheet fermions , as well as the world - sheet bosons with @xmath49 acting non - trivially on them ( for @xmath55 ) : @xmath62~,\\ & & { \cal y}_a = { 1\over 2 } { \rm tr}^\prime \left[g_a ( -1)^f e^{-2\pi t l_0}\right]~,\end{aligned}\ ] ] where prime in @xmath63 indicates that the trace is restricted as described above . for the annulus amplitude
we therefore have @xmath64~,\ ] ] where @xmath65^{2+d_a}}{\cal x}_a(e^{-2\pi t})~,\\ & & { \cal b}_a=\left({\rm tr}\left(\gamma_a\right)\right)^2 \int_0^\infty { dt\over t^3 } { 1\over \left[\eta(e^{-2\pi t})\right]^{2+d_a}}{\cal y}_a(e^{-2\pi t})~.\end{aligned}\ ] ] these integrals .
( the entire partition function vanishes as the numbers of space - time bosons and fermions are equal . ) for consistency , however , we must extract tadpoles from individual contributions @xmath66 and @xmath67 .
thus , for instance , cancellation of certain tadpoles coming from @xmath67 is required for consistency of the equations of motion for the twisted r - r four - form which couples to d3-branes ( see below ) . ]
are generically divergent as @xmath68 reflecting the presence of tadpoles . to extract these divergences we can change variables @xmath69 so that the divergences correspond to @xmath70 : @xmath71 the closed string states contributing to @xmath66 ( @xmath67 ) in the transverse channel are the ns - ns ( r - r ) states with @xmath72 ( and @xmath73 is the number of such states ) .
the massive states with @xmath74 do not lead to divergences as @xmath70 . on the other hand ,
the divergence property of the above integrals in the @xmath70 limit is determined by the value of @xmath57 .
given the orientability of @xmath20 the allowed values of @xmath57 are @xmath75 .
for @xmath76 there is no divergence in @xmath77 , so we have no restriction for @xmath39 . for @xmath78
the corresponding twisted ns - ns closed string sector contains tachyons .
this leads to a tachyonic divergence in @xmath66 unless @xmath79 for the corresponding @xmath49 twisted sector .
finally , the ground states in the r - r sectors are massless , so we get divergences due to massless r - r states in the integral in @xmath67 for large @xmath80 for @xmath81 . in such sectors in non - supersymmetric cases
we can also have tachyonic ns - ns divergences , while in supersymmetric cases we have massless ns - ns divergences . to avoid complications with tachyons , for
now we will focus on supersymmetric theories ( we will discuss non - supersymmetric cases in section vii ) .
we must therefore consider massless tadpoles arising for @xmath81 .
for @xmath82 the corresponding integrals are linearly divergent with @xmath80 as @xmath70 .
to cancel such a tadpole we must require that @xmath83 for the corresponding @xmath49 twisted sector .
on the other hand , if such a tadpole is not canceled , in the four - dimensional field theory language this would correspond to having a _ quadratic _ ( in the momentum ) divergence at the one - loop order .
this would imply that the corresponding four - dimensional background is actually inconsistent in the sense that either four - dimensional supersymmetry contains a twist with @xmath82 , we have @xmath22 supersymmetry . ] and/or poincar invariance must be broken propagate in four dimensions , in the presence of the tadpoles their equations of motion read @xmath84 ( @xmath85 , @xmath86 are the tadpoles , and the ellipses stand for terms of higher power in @xmath87 ) , so that we can _ a priori _ have solutions with broken four - dimensional poincar invariance . ] .
this is related to the fact that in such cases the corresponding twisted closed string states , which propagate only in the d3-brane world - volume as they are supported at the orbifold fixed _ points _ in @xmath16 , have inconsistent field equations ( yet they couple to the gauge / matter fields describing the low energy limit of the d3-brane world - volume theory ) .
in fact , as was pointed out in @xcite , generically uncancelled tadpoles arising in the @xmath82 cases lead to non - abelian gauge _ anomalies _ in the corresponding d3-brane gauge theories .
thus , let us consider the following example .
let @xmath88 , where the action of @xmath89 on the complex coordinates @xmath90 ( @xmath91 ) on @xmath29 is that of the z - orbifold : @xmath92 ( where @xmath9 is the generator of @xmath20 , and @xmath93 ) .
next , let us choose the representation of @xmath20 when acting on the chan - paton charges as follows : @xmath94 ( where @xmath95 , and @xmath96 is an @xmath97 identity matrix ) .
the massless spectrum of this model is the @xmath22 supersymmetric @xmath98 gauge theory with the matter consisting of chiral supermultiplets in the following representations : @xmath99 note that this spectrum is anomalous for non - vanishing @xmath100 ( the non - abelian gauge anomaly does not cancel ) unless @xmath101 . on the other hand , @xmath102 if and only if @xmath101 .
here we should mention that not all the choices of @xmath37 that do not satisfy @xmath103 for @xmath82 lead to such apparent inconsistencies .
thus , consider the same @xmath20 as above but with @xmath104 .
the massless spectrum of this model is the @xmath22 supersymmetric @xmath18 gauge theory with no matter , so it is anomaly free .
in fact , for any orbifold group @xmath20 containing a twist @xmath49 with @xmath82 we obtain an anomaly free theory only if @xmath79 or @xmath105 . in the latter case , however , as we discussed above , the corresponding background is nonetheless inconsistent , so we must require that for such twists @xmath79 . finally , let us discuss @xmath106 cases . for such twists
the corresponding integrals are only logarithmically divergent as @xmath70 . if such a tadpole is not canceled , that is , if the corresponding @xmath107 , in the four - dimensional field theory language this corresponds to having a _ logarithmic _ divergence ( in the momentum ) at the one - loop order .
as we will see in the following , these logarithmic divergences are precisely related to the running in the corresponding gauge theories , which are _ not _ conformal ( even in the large @xmath2 limit ) .
note that the corresponding twisted closed string states now propagate in two extra dimensions , so that the four - dimensional backgrounds are perfectly consistent - the tadpoles for these fields simply imply that these fields have non - trivial ( logarithmic ) profiles in these two extra dimensions , while the four dimensions along the d - brane world - volume are still flat ( and the four - dimensional supersymmetry is unbroken ) . in fact , the presence of such tadpoles does not introduce any anomalies in @xcite in a somewhat more complicated way . ] .
a simple way to see this for general @xmath20 is to note that ( in supersymmetric cases which we are focusing on here ) an individual twist with @xmath106 preserves @xmath1 supersymmetry , so that the corresponding gauge theory is anomaly free .
this then immediately implies that we do not have any non - abelian gauge anomalies in a theory with multiple such twists either . indeed
, such anomalies would have to arise at the one - loop level . in the string theory language
the relevant diagram is a @xmath15 , @xmath108 diagram ( the annulus ) with three external lines corresponding to non - abelian gauge fields attached to a single boundary .
thus , if we attach two external lines to one boundary , while the third one to another boundary , the corresponding diagram will vanish - indeed , the chan - paton structure of such a diagram is given by ( @xmath109 , @xmath110 , are the chan - paton matrices corresponding to the external lines ) @xmath111 which vanishes as for non - abelian gauge fields @xmath112 , so @xmath113 as well for by definition @xmath109 are invariant under the orbifold group action as @xmath109 correspond to the gauge bosons of the gauge group left unbroken by the orbifold ( in particular , note that @xmath109 commute with @xmath37 ) .
as to the aforementioned diagram with all three external lines attached to one boundary , its chan - paton structure is given by @xmath114 for twists with @xmath106 , each diagram of this type is @xmath1 supersymmetric as it does not contain any information about further supersymmetry breaking ( that is , the characters corresponding to the world - sheet degrees of freedom multiplying the chan - paton trace ( [ anomaly ] ) are those of an @xmath1 theory ) , so such diagrams do not introduce any anomalies .
we therefore conclude that a theory with multiple @xmath106 twists with @xmath107 is also anomaly free .
thus , as we see , in supersymmetric cases we can have non - trivial twists with @xmath81 .
consistency of the background requires that @xmath115 while for @xmath106 we do not have such a restriction .
in this section we discuss certain large @xmath2 gauge theories arising in the above setup with some twists @xmath49 with @xmath106 such that @xmath107 .
as we have already mentioned , such theories are not conformal .
the simplest examples of such theories are those with @xmath1 supersymmetry .
perturbatively such theories are not renormalized beyond the one - loop order .
we can also construct non - conformal @xmath22 supersymmetric theories in this way .
in general such theories are rather complicated . however , as we will see in the following , certain non - trivial @xmath22 theories of this type have the property that in the large @xmath2 limit the leading ( that is , planar ) diagrams do not renormalize the gauge theory correlators beyond the one - loop order .
in fact , the corresponding correlation functions in such an @xmath22 theory are ( up to overall numerical coefficients ) the same as in a parent @xmath1 theory .
the simplest example of an @xmath1 theory is obtained if we take @xmath116 , where the action of @xmath117 on the complex coordinates @xmath90 ( @xmath91 ) on @xmath29 is as follows : @xmath118 , @xmath119 ( where @xmath120 is the generator of @xmath20 ) .
next , let us choose the representation of @xmath20 when acting on the chan - paton charges as follows : @xmath121 .
the massless spectrum of this model is the @xmath1 supersymmetric @xmath18 gauge theory with no matter .
the non - abelian part of this theory is not conformal , and is asymptotically free . as to the overall center - of - mass @xmath122
, it is free , and can therefore be ignored .
more generally , if we have twists with @xmath123 , the gauge group generically is a product of @xmath124 factors , and we also have matter , which can be obtained using the quiver diagrams ( see @xcite ) .
there is always an overall center - of - mass @xmath122 , which is free .
other @xmath122 factors , however , run as the matter is charged under them . in the large @xmath2 limit , however , these @xmath122 s decouple in the infra - red , and can therefore also be ignored s are actually anomalous , and are broken at the tree - level via the generalized green - schwarz mechanism .
in particular , in cases with twists with @xmath82 we have mixed @xmath125 anomalies . however , in the cases with matter we are interested in here we do have running @xmath122 s that are anomaly free .
these @xmath122 s decouple in the infra - red in the large @xmath2 limit . ] .
as we have already mentioned , within the above setup we can construct non - trivial @xmath22 supersymmetric theories which are not renormalized beyond the one - loop order in the planar limit .
the idea here is similar to that in @xcite , where it was noticed that in theories with vanishing twisted tadpoles the planar diagrams are the same ( up to overall numerical factors ) as in the parent @xmath17 theory .
this is because the diagrams that contain information about supersymmetry breaking always contain twisted chan - paton traces @xmath126 , which vanish in such theories . in this subsection
we will consider theories where the planar diagrams are the same as in a parent @xmath1 theory in essentially the same way . to obtain such models , consider an orbifold group @xmath20 , which is a subgroup of @xmath127 but is not a subgroup of @xmath128 .
let @xmath129 be a non - trivial subgroup of @xmath20 such that @xmath130 .
we will allow the chan - paton matrices @xmath37 corresponding to the twists @xmath131 ( which have @xmath106 ) not to be traceless , so that the corresponding @xmath1 model is not conformal .
however , we will require that the other chan - paton matrices @xmath37 for the twists @xmath132 be traceless .
the resulting @xmath22 model is not conformal .
however , in the planar limit the perturbative gauge theory amplitudes are not renormalized beyond one loop .
the proof of this statement is straightforward .
thus , consider a planar diagram with @xmath10 boundaries ( but no handles)-loop diagram in the field theory language . ] with all external lines attached to a single boundary , which without loss of generality can be chosen to be the outer boundary .
such a diagram is depicted in fig.1 .
next , we need to specify the twists on the boundaries . a convenient choice ( consistent with that made for the annulus amplitude ( [ partition ] ) ) is given by is non - abelian , and we have to choose base points on the world - sheet to define the twists .
our discussion here , however , is unmodified also in this case . ]
@xmath133 where @xmath134 corresponds to the outer boundary , while @xmath135 , @xmath136 , correspond to inner boundaries .
let @xmath109 , @xmath137 , be the chan - paton matrices corresponding to the external lines .
then the above planar diagram has the following chan - paton group - theoretic dependence : @xmath138 where the sum involves all possible distributions of the @xmath135 twists that satisfy the condition ( [ mono ] ) , as well as permutations of the @xmath109 factors ( note that @xmath109 here correspond to the states that are kept after the orbifold projections , so that they commute with all @xmath37 ) .
note that the diagrams with all twists @xmath139 are ( up to overall numerical coefficients ) the same as in the parent @xmath1 theory , and therefore vanish beyond one loop .
all other diagrams contain at least one twist @xmath140 with @xmath141 .
this follows from the condition ( [ mono ] ) .
this then implies that all such diagrams vanish as @xmath142 in particular , this implies that the non - abelian gauge couplings do not run in the large @xmath2 limit beyond one loop limit compared with the leading one - loop contribution .
this is analogous to what happens in theories discussed in @xcite .
in fact , the techniques used there to prove that the higher loop corrections are subleading are very similar to the one we are using here . ] . the anomalous scaling dimensions ( two - point functions corresponding to the wave - function renormalization ) for matter fields , the corresponding diagrams are ( up to overall numerical factors ) the same as in the parent @xmath1 theory , so they are not renormalized in the planar limit . finally , the yukawa ( three - point ) and quartic scalar ( four - point ) couplings are not renormalized in perturbation theory due to the @xmath22 non - renormalization theorem for the superpotential .
let us now consider a simple example of the aforementioned gauge theories .
thus , let @xmath143 , and let the corresponding twisted chan - paton matrices @xmath144 .
let us assume that @xmath145 act non - trivially on the complex coordinates @xmath146 on @xmath147 , while leaving the coordinate @xmath148 untouched .
moreover , let us assume that the orbifold group @xmath129 is abelian , and its action on the coordinates @xmath146 is diagonal .
next , let @xmath120 be a generator of a @xmath31 group with the following action : @xmath149 , @xmath150 .
note that @xmath120 commutes with @xmath151 .
the full orbifold group is given by @xmath152 , which is also abelian , and @xmath153 .
let the twisted chan - paton matrix @xmath154 .
then the theory is the @xmath22 supersymmetric @xmath155 gauge theory with chiral matter multiplets , call them @xmath156 and @xmath157 , in @xmath158 and @xmath159 , respectively ( in the following we will ignore the @xmath122 factors ) .
note that there is no tree - level superpotential in this theory : @xmath160 this implies that we can break the gauge group down to the diagonal @xmath161 subgroup by giving the appropriate vacuum expectation values to the fields @xmath156 and @xmath157 .
the resulting theory in the ir is then the @xmath1 supersymmetric @xmath161 gauge theory without matter . in the string theory language this corresponds to moving the d3-branes off the @xmath31 orbifold singularity , hence the @xmath1 supersymmetry of the resulting theory .
before we end this section , as an aside we would like to discuss an embedding of pure @xmath22 supergluodynamics in the setup discussed in section ii .
thus , consider the orbifold group @xmath162 with the generators @xmath163 and @xmath164 of the two @xmath31 s acting on the complex coordinates @xmath90 on @xmath116 as follows : @xmath165 where @xmath166 .
next , let the twisted chan - paton matrices be given by @xmath167 . the d3-brane gauge theory is then given by the @xmath22 supersymmetric @xmath161 super - yang - mills theory without matter ( plus an overall @xmath122 which can be ignored ) .
note that the real dimensions of the points fixed under the twists @xmath168 are @xmath169 .
the above embedding shows that at least perturbatively pure supergluodynamics possesses a discrete symmetry , which is not evident in the field theory language .
thus , we have the @xmath170 symmetry as all the interactions involving twisted closed string states must be invariant under the orbifold group .
in addition , we have a @xmath171 symmetry which permutes the closed string states coming from the @xmath168 twisted sectors .
in particular , the generator @xmath172 of this @xmath171 group has the following action : @xmath173 thus , the @xmath171 subgroup does not commute with the @xmath174 subgroup .
thus , @xmath163 and @xmath172 generate a non - abelian discrete subgroup of @xmath175 , namely , the _ tetrahedral _ group @xmath176 .
thus , as we see , the perturbative large @xmath2 pure superglue theory possesses a discrete @xmath171 symmetry .
in fact , this symmetry persists even at finite @xmath2 .
indeed , even the diagrams with handles possess this symmetry as this symmetry is a symmetry of the underlying embedding of the pure supergluodynamics into string theory via the orbifold setup .
in fact , within this setup the @xmath171 symmetry is a discrete _ gauge _ symmetry .
indeed , instead of @xmath170 , the orbifold group can be chosen as the full tetrahedral group @xmath176 , where the generator @xmath172 of the @xmath171 subgroup of @xmath176 acts on the complex coordinates @xmath90 as follows : @xmath177 ( @xmath178 ) .
if we now choose the twisted chan - paton matrix @xmath179 , then the d3-brane gauge theory is still pure superglue .
this embedding makes it evident that the @xmath171 symmetry is indeed a gauge symmetry , so it might be an exact symmetry of pure superglue even non - perturbatively which is a subgroup of @xmath175 but not a subgroup of @xmath128 .
the corresponding gauge theory has @xmath22 supersymmetry .
note that the non - trivial twists in @xmath20 all have @xmath106 , and we can choose the corresponding twisted chan - paton matrices as @xmath105 , so that ( upon dropping the overall @xmath122 ) we obtain the @xmath161 pure super - yang - mills theory .
an appropriate abelian subgroup of @xmath20 is then a symmetry of this theory . ] .
we would like to end this subsection by pointing out one immediate consequence of the aforementioned @xmath171 symmetry - both the one - loop as well as the two - loop - function becomes gauge dependent . ] @xmath180-function coefficients in the pure supergluodynamics are multiples of 3 .
as we mentioned in introduction , perturbative expansion of d3-brane gauge theories simplifies substantially in the large @xmath2 limit .
it is advantageous to work in the full string theory framework , which in a way is much simpler than the feynman diagram techniques .
( at the end of the day we will take the @xmath181 limit , which amounts to reducing the theory to the gauge theory subsector . )
thus , in the string theory language there are two classes of diagrams we need to consider : ( _ i _ ) diagrams without handles ; ( _ ii _ ) diagrams with handles .
the latter correspond to closed string loops , and can be neglected as they are subleading in the large @xmath2 limit .
we will therefore focus on the diagrams without handles .
the latter diagrams can be viewed as _ tree - level _ closed string graphs connecting various boundaries .
this suggests that ( upon taking the @xmath181 limit ) in some cases ( which we will identify in a moment ) we might be able to rewrite the perturbative expansion in the large @xmath2 limit of the corresponding gauge theories in the language where various quantum corrections in the gauge theory are encoded in a _ classical _ higher dimensional _ field _ theory . the effective quantum action for the four - dimensional large @xmath2 gauge theory is then obtained by starting with the corresponding higher dimensional classical action , integrating out the bulk fields , and restricting to the gauge theory subsector .
that is , we consider the d - branes in the background of the bulk fields that is created by the d - branes themselves .
the field theory cut - off in this language arises precisely from the fact that , in the above setup , the twisted closed string states have tadpoles resulting in logarithmic profiles in the extra dimensions , and we need to regularize these profiles at the sources , that is , the d - branes , whose locations in the extra dimensions are given by points in @xmath29 .
the aforementioned proposal that in the large @xmath2 limit various quantum corrections in the d - brane gauge theory should in some cases be encoded in a higher dimensional classical field theory is what we refer to as the _ brane - bulk duality_. in this subsection we would like to identify the cases where the brane - bulk duality is applicable as stated above .
note that the brane - bulk duality is a consequence of the _ open - closed duality _ property of string theory .
in particular , the diagrams without handles in the transverse closed string channel can indeed be viewed as tree - level closed string graphs connecting various boundaries . at the one - loop order the corresponding diagram ( up to external lines ) is an annulus , where two boundaries are connected by a single closed string tube . in particular
, this diagram does _ not _ involve closed string interactions .
moreover , in the field theory limit @xmath181 contributions due to the massive closed string states in the transverse closed string channel are always finite .
that is , these contributions correspond to heavy string _
thresholds_. in the field theory language they translate into subtraction scheme dependent artifacts , which arise due to a particular choice of regularization .
thus , at the one - loop order the contributions due to massive string modes can be absorbed into the subtraction scheme dependence , so that in the field theory limit the brane - bulk duality indeed holds - various one - loop corrections in the gauge theory can be computed by calculating the classical d - brane self - interaction via the massless bulk fields . beyond the one - loop order , however , that is , when the number of boundaries is greater than two , we must include closed string interactions in the transverse closed string channel .
the massive closed string states are now expected to contribute in a non - trivial way .
in particular , _ a priori _ there is no longer a clear interpretation of these contributions in terms of thresholds .
another way of phrasing this is that at higher loop orders we can have mixed contributions coming from both massless as well as massive states .
that is , in general the brane - bulk duality can still be formulated except that at higher loop orders it will involve massless as well as _
infinitely _ many massive states as far as the corresponding higher dimensional classical `` field theory '' is concerned . in particular
, in general this classical `` field theory '' is nothing but the corresponding closed string theory , so we are back to the original open - closed string duality .
there are , however , non - trivial cases where in the large @xmath2 limit the bulk closed string theory can be consistently truncated to a field theory containing only a finite number of massless fields .
these cases are those where in the planar limit the gauge theory perturbatively is not renormalized beyond one loop .
thus , this is clearly the case for @xmath1 theories . in theories without matter
this , in fact , holds even at finite @xmath2 . in theories with matter
, however , we have running @xmath122 s whose decoupling is ensured only in the large @xmath2 limit . in subsection iiia we discussed @xmath22 theories which perturbatively are not renormalized beyond the one - loop order in the planar limit .
it is clear that the aforementioned formulation of the brane - bulk duality also holds in such theories , where , once again , various quantum corrections in the gauge theory are encoded in a higher dimensional classical field theory with a finite number of massless fields .
in fact , in @xmath1 as well as @xmath22 theories the bulk fields entering the relevant part of the corresponding higher dimensional classical action are _ twisted _ closed string states . indeed , at the one - loop order the diagrams involving untwisted closed string states in the transverse closed string channel are @xmath17 supersymmetric , so that they do not contribute into the renormalization of the corresponding ( on - shell ) gauge theory correlators . to clarify the above discussion ,
let us give a schematic description of the above procedure for obtaining quantum corrections in the corresponding gauge theories via the brane - bulk duality .
thus , we start from the classical action @xmath182+s_{\rm{\small bulk}}[\sigma]+ s_{\rm{\small int}}[\phi,\sigma]~,\ ] ] where @xmath183 is a collective notation for the ( fractional ) d3-brane world - volume fields , while @xmath184 is a collective notation for the _ massless twisted bosonic _ ( ns - ns and r - r ) bulk fields ; @xmath185 $ ] is the classical four - dimensional action for the brane fields @xmath183 , while @xmath186 $ ] is the classical higher dimensional action for the bulk fields @xmath184 ( note that some of these fields such as twisted closed string states propagate in less than 10 dimensions ) ; finally , @xmath187 $ ] , which is a classical four - dimensional functional ( with the integral over the ( fractional ) d3-brane world - volume ) , describes the coupling of the bulk fields to the brane fields , as well as to the brane itself ( in particular , it includes all tadpoles ) .
since here we are interested in one - loop corrections in the gauge theory language , the action @xmath186 $ ] contains only the terms quadratic in @xmath184 ( that is , the kinetic terms ) , while @xmath187 $ ] contains only the terms linear in @xmath184 .
moreover , the fields @xmath184 include only the twisted fields with non - vanishing ( logarithmic ) tadpoles .
next , we solve the equations of motion for @xmath184 following from the above action : @xmath188 in particular , we are interested in the solution where the fields @xmath184 do _ not _ explicitly depend on the coordinates @xmath189 along the d3-brane world - volume , but only functionally via @xmath190 ; this solution , however , does in general explicitly depend on the coordinates @xmath191 transverse to the d - branes .
let us denote this solution via @xmath192 .
the effective quantum action for the fields @xmath183 in the large @xmath2 limit is then given by @xmath193= s_{\rm{\small brane}}[\phi]+ s_{\rm{\small int}}[\phi,{\overline\sigma}]~.\ ] ] note that @xmath194 are generically divergent , so we need to further clarify the meaning of the second term in ( [ quantum ] ) .
thus , let @xmath195 note that @xmath196 are independent of @xmath189 , and , in fact , are nothing but the tadpoles for the fields @xmath184 . if all @xmath79 ( for @xmath81 ) , then there are no tadpoles for the corresponding twisted fields ( that is , the corresponding @xmath197 ) , and the large @xmath2 gauge theory is finite @xcite factors can still run , but , as we have already mentioned , in the large @xmath2 limit they decouple in the infra - red , and can therefore be ignored . ] . on the other hand , if some @xmath107 for twists with @xmath106 , some divergences no longer cancel , and we need to regularize @xmath194 in ( [ quantum ] ) .
note , however , that the corresponding divergences are only logarithmic , and , in fact , correspond to the running in the four - dimensional gauge theory , which is no longer conformal . in this subsection
we would like to address one technical point in the context of the brane - bulk duality , namely , the issue of regularizing the aforementioned divergences . here
we will consider a simple toy model which possesses the main ingredients for illustrating the regularization procedure .
in fact , as we will see in the following sections , this model actually captures all the key features of the @xmath1 models as well the @xmath22 models discussed in subsection iiia .
thus , consider the six dimensional theory with the following action : @xmath198 here @xmath199 is a six dimensional scalar field , whose dimension is @xmath200 ; @xmath201 is a four - dimensional operator with dimension @xmath202 localized on the hypersurface @xmath203 ; the couplings @xmath204 ( which are assumed to be non - vanishing ) are dimensionless ; finally , @xmath205 is a parameter of dimension ( length ) . in the above action @xmath156 is an analog of a twisted closed string state with a non - vanishing tadpole ( the last term in ( [ toy ] ) ) ; @xmath203 plays the role of a 3-brane ; @xmath206 is an analog of a dimension - four gauge theory operator such as @xmath207 ; finally , @xmath208 is analogous to @xmath209 .
following the procedure described in the previous subsection , we look for a solution to the equation of motion for @xmath156 : @xmath210 \delta^{(2)}(z)~.\ ] ] here @xmath211 are the coordinates transverse to the 3-brane ( whose location is chosen to be @xmath212 ) ; also , we use the notation @xmath213 .
it is convenient to fourier transform to the momentum space @xmath214 where @xmath215 and @xmath216 are the momenta corresponding to @xmath217 and @xmath218 , respectively .
thus , we have the following solution @xmath219~.\ ] ] this gives @xmath220~.\ ] ] the corresponding effective quantum action on the 3-brane is then given by @xmath221{\cal o}_4(p=0 ) + b^2 l^4\int { d^4 p\over(2\pi)^4}\int { d^2 k\over ( 2\pi)^2 } { { \cal o}_4(p){\cal o}_4(-p)\over { k^2+p^2}}+\dots~,\end{aligned}\ ] ] where the ellipses in the last line stand for a divergent constant piece ( which does not contain the operator @xmath206 , and is proportional to @xmath222 as it is solely due to the presence of the @xmath156 tadpole ) .
the second term in the last line in ( [ toyqu ] ) is a non - local higher dimensional operator , and it disappears in the @xmath223 limit .
the first term containing @xmath224 is the same as that in the classical 3-brane world - volume action except for the corresponding coupling - the quantum coupling in @xmath225 is a renormalized coupling : @xmath226 here @xmath227 is an ultra - violet ( uv ) cut - off , while @xmath228 is the infra - red ( ir ) cut - off . in the four - dimensional 3-brane world - volume field theory language
the ir cut - off is interpreted as the renormalization group ( rg ) scale at which the renormalized coupling @xmath229 is measured .
thus , if in the above toy model we adapt the interpretation of the previous subsection , from the classical dynamics of the bulk field @xmath156 we will obtain the `` one - loop '' renormalization of the 3-brane world - volume field theory .
in fact , in this model the 3-brane world - volume theory is not renormalized beyond the `` one - loop '' order .
in this section we would like to apply the brane - bulk duality discussed in the previous section to gauge theories arising in the orbifold construction discussed in section ii . here
we will focus on the simplest examples of this type , in particular , those with @xmath1 supersymmetry .
perturbatively such gauge theories are not renormalized beyond the one - loop order even for finite @xmath2 .
thus , the brane - bulk duality in such examples can be understood in detail theories in the regime where the effective t hooft coupling is large were discussed in @xcite . ] .
we described the simplest example of an @xmath1 gauge theory in the above context in the beginning of the previous section . in this example
@xmath230 , and the twisted chan - paton matrix @xmath121 .
the gauge group is @xmath18 , and we have no matter fields . in subsection
a we discuss the brane - bulk duality in this model in the case of unbroken gauge symmetry .
we will discuss the case of spontaneously broken gauge symmetry in subsection b. finally , in subsection c we discuss examples with matter . to discuss the brane - bulk duality in this model ,
we need the relevant part of the classical action containing the d3-brane as well as the bulk fields .
this action has the following form @xcite ( we are not including terms containing @xmath231 for we are going to be interested in renormalization of the @xmath232 term ) .
also , in @xmath233 $ ] only the @xmath122 contribution is non - vanishing . ] : @xmath234 + b~\sigma~{\rm tr}\left[f^{\mu\nu}f_{\mu\nu}\right]+ c~\sigma+ d~\epsilon^{\mu\nu\sigma\rho}~c_{\mu\nu}~ { \rm tr}\left[f_{\sigma\rho}\right]\big)\nonumber\\ & & -\int_{{\rm d3}\times { \bf r}^2 } \left({1\over 2}\partial^\mu\sigma~\partial_\mu\sigma+{1\over 12 } h^{\mu\nu\sigma } h_{\mu\nu\sigma}\right)~. \label{eq : actionu}\end{aligned}\ ] ] here @xmath184 is a twisted ns - ns scalar , while @xmath235 is a twisted two - form ( whose field strength is @xmath236 ) .
we have normalized the kinetic terms of the bulk @xmath184 and @xmath235 fields in the standard way .
once this normalization is fixed , the couplings @xmath237 can be determined .
we discuss these couplings in appendix a. the combinations relevant for our discussion here are given by @xmath238 finally , the coupling @xmath239 is given by @xmath240 where @xmath241 is the tree - level yang - mills gauge coupling , and the @xmath18 generators @xmath242 are normalized as @xmath243 .
the yang - mills gauge coupling is related to the string coupling @xmath8 via @xmath244 . as in section iv
, we solve classical equations of motion for the fields @xmath184 and @xmath235 ( the latter is a gauge field , so we must use gauge fixing such as @xmath245 ) , and integrate them out of the classical action ( [ eq : actionu ] ) .
the resulting effective quantum action is given by ( we are using the momentum representation , and drop higher dimensional terms as well as those independent of @xmath246 ) : @xmath247~ { \rm tr}\left[f^{\mu\nu}(p)~f_{\mu\nu}(-p)\right]+\right.\nonumber\\ & & \left.{d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}~ { \rm tr}\left[f^{\mu\nu}(p)\right]~{\rm tr}\left[f_{\mu\nu}(-p ) \right]\right)~.\end{aligned}\ ] ] since only the @xmath122 subgroup contributes into @xmath248 $ ] , we can rewrite the last term in the above expression in terms of the @xmath122 field strength @xmath249 ( note that the corresponding generator is @xmath250 ) : @xmath251~{\rm tr}[f_{\mu\nu}(-p)]= { \rm tr}[f^{\mu\nu}(p)]~{\rm tr}[f_{\mu\nu}(-p)]= n~{\rm tr}[f^{\mu\nu}(p)~f_{\mu\nu}(-p)]~.\ ] ] thus , we have @xmath247~ { \rm tr}\left[{\widehat f}^{\mu\nu}(p)~ { \widehat f}_{\mu\nu}(-p)\right]+\right.\nonumber\\ & & \left.\left[a-{bc\over 2\pi^2}\int { d^2
k\over k^2}+ { n d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}\right]~ { \rm tr}\left[f^{\mu\nu}(p)~f_{\mu\nu}(-p ) \right]\right)~,\end{aligned}\ ] ] where @xmath252 is the @xmath161 field strength .
note that the integrals contributing to the @xmath253 coupling individually are uv divergent .
however , since we have ( [ bcd ] ) , the total contribution is uv finite . on the other hand ,
the first integral is ir divergent , while the second integral is ir finite for @xmath254 ( for this latter integral the @xmath255 term in the denominator plays the role of the ir cut - off ) .
we must therefore introduce an ir cut - off in the first integral .
the fact that the @xmath122 gauge coupling should not be renormalized then dictates that the ir cut - off in the first integral must be chosen as follows gauge coupling can in general receive finite ( string ) threshold corrections . here , however , we ignore such corrections as we are interested in the field theory limit . ] : @xmath256 the effective quantum action therefore reads : @xmath257+a~ { \rm tr}\left[f^{\mu\nu}(p)~f_{\mu\nu}(-p ) \right]\right)~,\ ] ] where @xmath258 this is nothing but the one - loop renormalized yang - mills gauge coupling with the @xmath180-function coefficient @xmath259 . in this subsection
we will discuss the brane - bulk duality in the above model in the case where the gauge symmetry is spontaneously broken : @xmath260 , @xmath261 . in the field theory language this corresponds to the complex adjoint scalar in the @xmath1 vector supermultiplet having an appropriate vacuum expectation value . in the string theory language this corresponds to splitting the @xmath2 d3-branes into two stacks of @xmath262 and @xmath263 d3-branes , and moving them apart by some distance @xmath264 in the two real directions transverse to the d - branes untouched by the @xmath31 orbifold action .
the relevant part of the classical action in this case is given by : @xmath265 + b~\sigma~{\rm tr}[f_1^{\mu\nu}f_{1\mu\nu}]+c_1~\sigma+ d~\epsilon^{\mu\nu\sigma\rho}~c_{\mu\nu}~{\rm tr}[f_{1\sigma\rho}]\nonumber\\ & & -\int_{{\rm d3}_2 } a~{\rm tr}[f_2^{\mu\nu}f_{2\mu\nu}]+ b~\sigma~{\rm tr}[f_2^{\mu\nu}f_{2\mu\nu}]+c_2~\sigma+
d~\epsilon^{\mu\nu\sigma\rho}~c_{\mu\nu}~{\rm tr}[f_{2\sigma\rho}]\nonumber\\ & & -\int_{{\rm d3}\times { \bf r}^2 } \left({1\over 2}\partial^\mu\sigma~\partial_\mu\sigma+{1\over 12 } h^{\mu\nu\sigma } h_{\mu\nu\sigma}\right)~. \label{eq : action}\end{aligned}\ ] ] the couplings @xmath266 are the same as before , while the relevant combinations containing the couplings @xmath267 are now given by @xmath268 in the following we will assume that the d3@xmath269-branes ( that is , the stacks of @xmath270 d3-branes ) are located at @xmath271 , respectively , in the extra two dimensions . here
@xmath191 , @xmath272 , are real coordinates , and in the following we will use the notation @xmath273 , @xmath274 . upon integrating out the bulk fields ,
we obtain : @xmath275~ { \rm tr}\left[f^{\mu\nu}_1(p)~f_{1\mu\nu}(-p)\right]+\right.\nonumber\\ & & \left[a-{bc_2\over 2\pi^2}\int { d^2 k\over k^2}- { bc_1\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}\right]~ { \rm tr}\left[f^{\mu\nu}_2(p)~f_{2\mu\nu}(-p)\right]+\nonumber\\ & & { d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}~\big ( { \rm tr}\left[f^{\mu\nu}_1(p)\right]~{\rm tr}\left[f_{1\mu\nu}(-p)\right]+ { \rm tr}\left[f^{\mu\nu}_2(p)\right]~{\rm tr}\left[f_{2\mu\nu}(-p)\right ] \big)+\nonumber\\ & & \left .
{ 2d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}~e^{ik\cdot x}~ { \rm tr}\left[f^{\mu\nu}_1(p)\right]~{\rm tr } \left[f_{2\mu\nu}(-p)\right]\right)~.\end{aligned}\ ] ] as in the previous subsection , let us separate the @xmath122 contributions : @xmath276~ { \rm tr}\left[{\widehat f}^{\mu\nu}_1(p)~{\widehat f}_{1\mu\nu}(-p)\right]+ \right.\nonumber\\ & & \left[a-{bc_2\over 2\pi^2}\int { d^2 k\over k^2}- { bc_1\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}\right]~ { \rm tr}\left[{\widehat f}^{\mu\nu}_2(p)~{\widehat f}_{2\mu\nu}(-p)\right ] + \nonumber\\ & & \left[a-{bc_1\over 2\pi^2}\int { d^2
k\over k^2}- { bc_2\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}+ { n_1 d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}\right]~ { \rm tr}\left[f^{\mu\nu}_1(p)~f_{1\mu\nu}(-p)\right ] + \nonumber\\ & & \left[a-{bc_2\over 2\pi^2}\int { d^2 k\over k^2}- { bc_1\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}+ { n_2 d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}\right]~ { \rm tr}\left[f^{\mu\nu}_2(p)~f_{2\mu\nu}(-p)\right ] + \nonumber\\ & & \left .
{ 2d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}~e^{ik\cdot x}~ { \rm tr}\left[f^{\mu\nu}_1(p)\right]~{\rm tr } \left[f_{2\mu\nu}(-p)\right]\right)~.\end{aligned}\ ] ] to further simplify this expression , let us explicitly take the traces over the @xmath122 generators .
thus , we have @xmath277 where @xmath278 is a null @xmath97 matrix , and @xmath279 are the @xmath122 field strengths .
we therefore obtain : @xmath276~ { \rm tr}\left[{\widehat f}^{\mu\nu}_1(p)~{\widehat f}_{1\mu\nu}(-p)\right]+ \right.\nonumber\\ & & \left[a-{bc_2\over 2\pi^2}\int { d^2 k\over k^2}- { bc_1\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}\right]~ { \rm tr}\left[{\widehat f}^{\mu\nu}_2(p)~{\widehat f}_{2\mu\nu}(-p)\right ] + \nonumber\\ & & \left[a-{bc_1\over 2\pi^2}\int { d^2 k\over k^2}- { bc_2\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}+ { n_1 d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}\right]~ { 1\over 2}{\overline f}^{\mu\nu}_1(p)~{\overline f}_{1\mu\nu}(-p ) + \nonumber\\ & & \left[a-{bc_2\over 2\pi^2}\int { d^2 k\over k^2}- { bc_1\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}+ { n_2 d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}\right]~ { 1\over 2}{\overline f}^{\mu\nu}_2(p)~{\overline f}_{2\mu\nu}(-p ) + \nonumber\\ & & \left .
{ \sqrt{n_1 n_2 } d^2\over \pi^2}\int
{ d^2 k\over{k^2+p^2}}~e^{ik\cdot x}~ { \overline f}^{\mu\nu}_1(p)~ { \overline f}_{2\mu\nu}(-p)\right)~.\end{aligned}\ ] ] here it is convenient to rotate the above @xmath122 s to the following basis @xmath280 where @xmath281 here @xmath282 is the field strength of the overall center - off - mass @xmath122 in @xmath283 , while @xmath284 is the field strength of the @xmath122 in the breaking @xmath285 .
indeed , we have @xmath286 the matrices multiplying @xmath287 and @xmath288 on the r.h.s . of this equation
are nothing but the properly normalized generators of @xmath289 and @xmath290 , respectively . in the new basis
we have @xmath291~ { 1\over 2}{\overline f}^{\mu\nu}_1(p)~{\overline f}_{1\mu\nu}(-p ) + \right.\nonumber\\ & & \left[a-{bc_2\over 2\pi^2}\int { d^2 k\over k^2}- { bc_1\over 2\pi^2}\int { d^2 k\over k^2}~e^{ik\cdot x}+ { n_2 d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}\right]~ { 1\over 2}{\overline f}^{\mu\nu}_2(p)~{\overline f}_{2\mu\nu}(-p ) + \nonumber\\ & & \left.{\sqrt{n_1 n_2 } d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}~e^{ik\cdot x}~ { \overline f}^{\mu\nu}_1(p)~ { \overline f}_{2\mu\nu}(-p)\right)=\nonumber\\ & & \int{d^4 p\over ( 2\pi)^4}~ \left({\sqrt{n_1 n_2}d^2\over \pi^2}~{{n_1-n_2}\over n}~ \left[\int{d^2 k\over { k^2+p^2}}~ e^{ik\cdot x}-\int{d^2 k\over k^2}~e^{ik\cdot x}\right]~ { \overline f}^{\mu\nu}_+(p)~{\overline f}_{-\mu\nu}(-p)+\right.\nonumber\\ & & \left(a+{2n_1 n_2\over n}~{d^2\over \pi^2}~\left[\int{d^2 k\over { k^2+p^2}}~ e^{ik\cdot x}-\int{d^2 k\over k^2}~e^{ik\cdot x}\right]\right)~ { 1\over 2}{\overline f}^{\mu\nu}_+(p)~{\overline f}_{+\mu\nu}(-p)+\nonumber\\ & & \left.\left(a-{d^2\over n\pi^2}~\left[2n_1
n_2 \int{d^2 k\over { k^2+p^2}}~ e^{ik\cdot x}+\left[n_1 ^ 2+n_2 ^ 2\right ] \int{d^2 k\over k^2}~e^{ik\cdot x}\right]\right)~ { 1\over 2}{\overline f}^{\mu\nu}_-(p)~{\overline f}_{-\mu\nu}(-p)\right)~. \label{u(1)s}\end{aligned}\ ] ] in arriving at this equation we have used ( [ couplings12 ] ) , as well as the fact that , as we discussed in the previous subsection , in the integral @xmath292 the ir cut - off is given by @xmath293 , so that @xmath294 in fact , the r.h.s . of ( [ u(1)s ] )
further simplifies once we go to the field theory limit .
this limit is given by @xmath295 in this limit the two stacks of d3-branes come on top of each other , but the gauge symmetry is still @xmath296 - the original @xmath18 gauge group is broken by the adjoint higgs vacuum expectation value parametrized by @xmath297 . in particular , in this limit we have @xmath298 this implies that on the r.h.s . of ( [ u(1)s ] )
the @xmath299 term goes to zero , the coupling for the @xmath300 term goes to @xmath301 ( that is , this coupling is not renormalized , which is consistent with the fact that the overall center - of - mass @xmath122 should not run ) , while the coupling for the @xmath302 term is renormalized as follows : @xmath303 this coupling , as well as the non - abelian gauge couplings , need to be regularized .
this regularization , however , depends on whether @xmath304 or @xmath305 .
this is because the aforementioned @xmath181 limit must be taken differently depending on @xmath255 .
however , ( [ vanish ] ) holds regardless of how the limit is taken . collecting the above results we see that the effective quantum action is given by : @xmath306 + { \widetilde a}_2(p^2,\phi^2)~ { \rm tr}\left[{\widehat f}^{\mu\nu}_2(p)~{\widehat f}_{2\mu\nu}(-p)\right ] + \nonumber\\ & & { 1\over 2 } a~{\overline f}^{\mu\nu}_+(p)~{\overline f}_{+\mu\nu}(-p ) + { 1\over 2}{\widetilde a}_-(p^2,\phi^2)~ { \overline f}^{\mu\nu}_-(p)~{\overline f}_{-\mu\nu}(-p)\big)~ , \end{aligned}\ ] ] where the @xmath290 coupling is given by ( [ u(1)_- ] ) , while the non - abelian couplings are given by :
@xmath307 as before , the integral ( [ k^2 ] ) is regularized as follows : @xmath308 however , as we have already mentioned , the regularization of the integral @xmath309 depends upon @xmath255 .
for @xmath310 we take the @xmath311 limit in the exponent @xmath312 and regularize the resulting integral as in ( [ k^2reg ] ) . for @xmath310 we therefore have @xmath313 the r.h.s . of this equation
is nothing but the one - loop renormalized @xmath161 gauge coupling with the @xmath180-function coefficient @xmath259 .
all three gauge couplings , that is , those for @xmath314 , @xmath315 and @xmath290 , run together at @xmath310 as at large momenta the effects of the @xmath316 breaking are negligible .
now , at small momenta @xmath305 the @xmath181 limit must be taken differently .
in particular , in the integral ( [ k^2x ] ) we first redefine the integration variables via @xmath317 .
the resulting integral @xmath318 must then be regularized for small @xmath319 .
note that @xmath319 has dimension of @xmath320 , so this is a uv divergence .
the regularized integral ( [ zeta^2 ] ) is then given by @xmath321 where @xmath322 parametrizes the subtraction scheme dependence ( see below ) .
for small momenta @xmath323 we therefore have : @xmath324 where @xmath325 and @xmath326 are the one - loop beta function coefficients for @xmath314 and @xmath315 , respectively . in the above expressions the terms proportional to @xmath327 correspond to the _ threshold _ corrections due to the massive gauge bosons ( that is , the gauge bosons that become heavy in the higgs mechanism ) . as usual , to connect the gauge coupling evolution above and below the threshold , we need to specify a subtraction scheme .
thus , we can choose the subtraction scheme where the gauge couplings are matched at the scale @xmath328 , where @xmath329 is the mass of the heavy gauge bosons .
in particular , this implies that @xmath330 we then have the following gauge coupling running . for @xmath331 @xmath332 where @xmath333 which is the @xmath161 gauge coupling at @xmath331 .
for @xmath334 we have @xmath335 so that below the threshold the @xmath314 and @xmath315 gauge couplings run with the @xmath180-function coefficients @xmath325 and @xmath326 , respectively , while the @xmath290 gauge coupling does not run at all . thus ,
using the brane - bulk duality approach we reproduce the expected perturbative running of gauge couplings in the corresponding @xmath1 gauge theories . in the brane - bulk duality approach , however , we do not perform any loop computations .
rather , the information about the loop corrections in gauge theory is encoded in the corresponding classical higher dimensional field theory .
this is a consequence of the fact that the brane - bulk duality simply follows from the closed - open string duality .
before we end this subsection , the following remark is in order . in the above computations we had to regularize various ( logarithmically ) divergent integrals .
the corresponding regularizations depend upon the four - dimensional momentum squared .
for instance , in this subsection we saw that the regularization of the integrals containing the information about the threshold depends on whether @xmath255 is above or below the threshold .
in particular , the details of the corresponding regularization are somewhat different from what happens in the direct loop computation in the field theory language .
this appears to be a common feature of string theory embeddings of gauge theories as far as computations of , say , gauge coupling running are concerned . in particular , to reproduce the gauge coupling running , it appears to be necessary to introduce an ( ir ) cut - off which is @xmath255 dependent .
this appears to be a consequence of the fact that such computations are typically done within the _ on - shell _ formulation of string theory . for completeness in this subsection we would like to briefly discuss examples of @xmath1 gauge theories with matter .
the simplest examples of this type are obtained via the aforementioned @xmath31 orbifold construction with the twisted chan - paton matrix given by @xmath336 , where @xmath261 . in this case
the gauge group is @xmath296 , and we have matter hypermultiplets in @xmath337 and @xmath338 , where the @xmath122 charges are given in parentheses .
note that all @xmath2 d3-branes are now coincident , and the gauge symmetry is broken due to the non - trivial orbifold action on the chan - paton factors .
the relevant part of the classical action is given by ( for simplicity the matter field contributions are not shown , and we use the notation @xmath339 ) : @xmath234 + b~\sigma~{\rm
tr}\left[\gamma_r~f^{\mu\nu}f_{\mu\nu}\right]+ { \rm tr}(\gamma_r)~{\widehat c}~\sigma+ d~\epsilon^{\mu\nu\sigma\rho}~c_{\mu\nu}~ { \rm tr}\left[\gamma_r~f_{\sigma\rho}\right]\big)\nonumber\\ & & -\int_{{\rm d3}\times { \bf r}^2 } \left({1\over 2}\partial^\mu\sigma~\partial_\mu\sigma+{1\over 12 } h^{\mu\nu\sigma } h_{\mu\nu\sigma}\right)=\nonumber\\ & & - \int_{\rm d3 } \big(a~\left[{\rm tr}\big[f^{\mu\nu}_1 f_{1\mu\nu}\right]+ { \rm tr}\left[f^{\mu\nu}_2 f_{2\mu\nu}\right]\big]+ b~\sigma~\big[{\rm tr}\left[f^{\mu\nu}_1 f_{1\mu\nu}\right]- { \rm tr}\left[f^{\mu\nu}_2 f_{2\mu\nu}\right]\big]\nonumber\\ & & + ( n_1-n_2)~{\widehat c}~\sigma+ d~\epsilon^{\mu\nu\sigma\rho}~c_{\mu\nu}~ \big[{\rm
tr}\left[f_{1\sigma\rho}\right]- { \rm tr}\left[f_{2\sigma\rho}\right]\big]\big)\nonumber\\ & & -\int_{{\rm d3}\times { \bf r}^2 } \left({1\over 2}\partial^\mu\sigma~\partial_\mu\sigma+{1\over 12 } h^{\mu\nu\sigma } h_{\mu\nu\sigma}\right)~ , \label{eq : actionu2}\end{aligned}\ ] ] where the couplings @xmath266 are the same as before , while @xmath340 note that for @xmath341 or @xmath342 we recover the action ( [ eq : actionu ] ) .
integrating out the bulk fields , we obtain : @xmath343~ { \rm tr}\left[f^{\mu\nu}_1(p)~f_{1\mu\nu}(-p)\right]+
\right.\nonumber\\ & & \left[a-{b{\widehat c}\over 2\pi^2}~(n_2-n_1)\int { d^2 k\over k^2 } \right]~ { \rm tr}\left[f^{\mu\nu}_2(p)~f_{2\mu\nu}(-p)\right]+\nonumber\\ & & \left.{d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}~ \left({\rm tr}[f^{\mu\nu}_1(p)]-{\rm tr}[f^{\mu\nu}_2(p)]\right)~ \left({\rm tr}[f_{1\mu\nu}(-p)]-{\rm tr}[f_{2\mu\nu}(-p)]\right)\right)~.\end{aligned}\ ] ] the @xmath122 contributions can be extracted as in the previous subsection , and , in fact , all the corresponding normalizations are exactly the same as before .
we therefore obtain : @xmath344 + { \widetilde
a}_2(p^2)~ { \rm tr}\left[{\widehat f}^{\mu\nu}_2(p)~{\widehat f}_{2\mu\nu}(-p)\right ] + \nonumber\\ & & { 1\over 2 } a~{\overline f}^{\mu\nu}_+(p)~{\overline f}_{+\mu\nu}(-p ) + { 1\over 2}{\widetilde a}_-(p^2)~ { \overline f}^{\mu\nu}_-(p)~{\overline f}_{-\mu\nu}(-p)\big)~ , \end{aligned}\ ] ] where the @xmath290 coupling is given by @xmath345 while the non - abelian couplings are given by @xmath346 where @xmath347 is the @xmath290 one - loop @xmath180-function coefficient , while @xmath348 and @xmath349 are the one - loop @xmath180-function coefficients for the @xmath314 and @xmath315 subgroups , respectively .
thus , as we see , once again , we correctly reproduce the running of the gauge couplings for the @xmath350 subgroup ( the overall center - of - mass @xmath289 does not run as there is no matter charged under it ) .
note that the twisted tadpole for @xmath184 vanishes for @xmath351 , that is , for @xmath352 .
the non - abelian one - loop @xmath180-function coefficients in this case vanish .
the @xmath290 one - loop @xmath180-function coefficient , however , is still non - vanishing , so the @xmath290 gauge coupling runs . in the large @xmath2 limit @xmath290 decouples in the ir , and we are left with an @xmath1 superconformal field theory .
in this section we would like to discuss the brane - bulk duality in @xmath353 supersymmetric theories of the type discussed in subsection iiia .
thus , let us consider the example where @xmath354 , and the action of the generators @xmath163 and @xmath164 of the two @xmath31 s on the complex coordinates @xmath90 on @xmath355 is given by ( here @xmath166 ) @xmath356 let us choose the twisted chan - paton matrices as follows : @xmath357 , @xmath358 , where @xmath261 .
in this case we have the @xmath22 supersymmetric @xmath296 gauge theory with chiral matter in @xmath337 and @xmath338 , where the @xmath122 charges are given in parentheses .
note that all @xmath2 d3-branes are coincident , and the gauge symmetry is broken due to the non - trivial orbifold action on the chan - paton factors .
for @xmath351 we have an @xmath22 theory of the type discussed in subsection iiia .
in particular , in the planar limit the correlation functions in this theory are the same as in the parent @xmath1 supersymmetric gauge theory with @xmath18 gauge group and no matter .
we can therefore discuss this theory in the context of the brane - bulk duality as in the large @xmath2 limit we have the corresponding non - renormalization theorem beyond the one - loop order . for calculational convenience in the following we will keep @xmath262 and @xmath263 arbitrary .
the calculation of the one - loop effective quantum action then gives a correct result even for @xmath359 , but only for @xmath351 do we have the non - renormalization theorem beyond the one - loop order .
the relevant part of the classical action is given by ( for simplicity the matter field contributions are not shown , and we use the notation @xmath339 ) : @xmath234\nonumber\\ & & + { 1\over\sqrt{2 } } \sum_\alpha\big[b~\sigma_\alpha~{\rm tr}\left[\gamma_{r_\alpha}~f^{\mu\nu } f_{\mu\nu}\right]+ { \rm tr}(\gamma_{r_\alpha})~{\widehat c}~\sigma_\alpha+ d~\epsilon^{\mu\nu\sigma\rho}~c_{\alpha\mu\nu}~ { \rm tr}\left[\gamma_{r_\alpha}~f_{\sigma\rho}\right]\big]\big)\nonumber\\ & & -\sum_\alpha \int_{{\rm d3}\times { \cal f}_\alpha } \left({1\over 2}\partial^\mu\sigma_\alpha~\partial_\mu\sigma_\alpha+ { 1\over 12 } h_\alpha^{\mu\nu\sigma } h_{\alpha\mu\nu\sigma}\right)~ , \label{eq : actionu3}\end{aligned}\ ] ] where the couplings @xmath360 are the same as before , and the overall factor of @xmath361 is due to the fact that the @xmath168 twisted fields @xmath362 and @xmath363 now propagate in @xmath364 , where each fixed point set @xmath365 is an orbifold @xmath366 .
integrating out the bulk fields , we obtain : @xmath367~ { \rm tr}\left[f^{\mu\nu}_1(p)~f_{1\mu\nu}(-p)\right]+ \right.\nonumber\\ & & \left[a-{b{\widehat c}\over 4\pi^2}~(3n_2-n_1)\int { d^2 k\over k^2 } \right]~ { \rm tr}\left[f^{\mu\nu}_2(p)~f_{2\mu\nu}(-p)\right]+\nonumber\\ & & { d^2\over 2\pi^2}\int { d^2
k\over{k^2+p^2}}~ \left({\rm tr}[f^{\mu\nu}_1(p)]+{\rm tr}[f^{\mu\nu}_2(p)]\right)~ \left({\rm tr}[f_{1\mu\nu}(-p)]+{\rm tr}[f_{2\mu\nu}(-p)]\right)\nonumber\\ & & \left.{d^2\over \pi^2}\int { d^2 k\over{k^2+p^2}}~ \left({\rm tr}[f^{\mu\nu}_1(p)]-{\rm tr}[f^{\mu\nu}_2(p)]\right)~ \left({\rm tr}[f_{1\mu\nu}(-p)]-{\rm tr}[f_{2\mu\nu}(-p)]\right)\right)~.\end{aligned}\ ] ] the @xmath122 contributions can be extracted as in the previous section , and , in fact , all the corresponding normalizations are exactly the same as before .
we therefore obtain : @xmath344 + { \widetilde a}_2(p^2)~ { \rm tr}\left[{\widehat f}^{\mu\nu}_2(p)~{\widehat f}_{2\mu\nu}(-p)\right ] + \nonumber\\ & & { 1\over 2 } a~{\overline f}^{\mu\nu}_+(p)~{\overline f}_{+\mu\nu}(-p ) + { 1\over 2}{\widetilde a}_-(p^2)~ { \overline f}^{\mu\nu}_-(p)~{\overline f}_{-\mu\nu}(-p)\big)~ , \end{aligned}\ ] ] where the @xmath290 coupling is given by @xmath345 while the non - abelian couplings are given by @xmath346 where @xmath368 is the @xmath290 one - loop @xmath180-function coefficient , while @xmath369 and @xmath370 are the one - loop @xmath180-function coefficients for the @xmath314 and @xmath315 subgroups , respectively .
thus , we correctly reproduce the running of the gauge couplings for the @xmath350 subgroup ( the overall center - of - mass @xmath289 does not run as there is no matter charged under it ) .
note that the twisted tadpole for @xmath371 does not vanish , so that the non - abelian part of the gauge theory is non - conformal even for @xmath372 . in the large @xmath329 limit @xmath290 decouples in the ir , and we are left with the @xmath1 supersymmetric @xmath373 gauge theory with chiral matter in @xmath374 and @xmath375 .
so far we have been focusing on @xmath1 and @xmath22 supersymmetric theories .
however , in the large @xmath2 limit we can also discuss certain non - trivial non - supersymmetric cases as well .
the large @xmath2 property is crucial here .
the reason is that in the cases where the orbifold group @xmath376 , we always have twisted ns - ns closed string sectors with tachyons .
their contributions to the corresponding part of the annulus amplitude ( [ nsns ] ) is then exponentially divergent unless we require that @xmath377 even if this condition is satisfied , we must take the t hooft limit - indeed , otherwise it is unclear , for instance , how to deal with the diagrams with handles , which contain tachyonic divergences .
in fact , the same applies to some non - planar diagrams without handles , that is , diagrams where the external lines are attached to more than one boundaries ( such diagrams are subleading in the large @xmath2 limit ) . to obtain well - defined non - supersymmetric non - conformal models , consider an orbifold group @xmath378 , which is not a subgroup of @xmath127 .
let @xmath129 be a non - trivial subgroup of @xmath20 such that @xmath130
. we will allow the chan - paton matrices @xmath37 corresponding to the twists @xmath131 ( which have @xmath106 ) not to be traceless , so that the corresponding @xmath1 model is not conformal .
however , we will require that the other chan - paton matrices @xmath37 for the twists @xmath132 be traceless .
the resulting non - supersymmetric model is not conformal .
however , in the planar limit the perturbative gauge theory amplitudes are not renormalized beyond one loop ( as usual , various running @xmath122 s decouple in the ir in this limit ) .
the proof of this statement is completely parallel to that we gave in subsection iiia for @xmath22 theories .
let us consider a simple example of such a theory .
let @xmath379 , where the action of the generators @xmath120 and @xmath172 of the @xmath31 respectively @xmath171 subgroups on the complex coordinates @xmath90 on @xmath116 is as follows : @xmath380 , @xmath381 , @xmath382 , @xmath383 , where @xmath384 .
the twisted chan - paton matrices are given by : @xmath385 , @xmath386 .
then the theory is a non - supersymmetric @xmath387 gauge theory with matter consisting of complex scalars in @xmath388 , @xmath389 and @xmath390 , as well as chiral fermions in the above representations plus their complex conjugates .
the gauge coupling renormalization in this model can be discussed in complete parallel with the previous sections . since the @xmath122 s decouple in the ir in the large @xmath2 limit , we will ignore them in the following runnings at one loop receive contributions from non - planar diagrams with the two external lines corresponding to the @xmath122 gauge bosons attached to two different boundaries .
note , however , that these contributions are due to r - r exchanges ( and the r - r sectors do not contain tachyons ) , so that tachyons do not contribute to this . ] .
then the relevant part of the classical action is given by ( for simplicity the matter field contributions are not shown , and we use the notation @xmath391 ) : @xmath392 + { 1\over\sqrt{3 } } \big[b~\sigma~{\rm tr}\left[\gamma_r~{\widehat f}^{\mu\nu } { \widehat f}_{\mu\nu}\right]+ { \rm tr}(\gamma_r)~{\widehat c}~\sigma\big]\big)\nonumber\\ & & -\int_{{\rm d3}\times { \cal f } } { 1\over 2}\partial^\mu\sigma~\partial_\mu\sigma~ , \label{eq : actionu4}\end{aligned}\ ] ] where the couplings @xmath393 are the same as before , and the overall factor of @xmath394 is due to the fact that the @xmath120 twisted field @xmath184 now propagates in @xmath395 , where the fixed point set @xmath396 is an orbifold @xmath397 .
integrating out the bulk fields , we obtain ( @xmath110 ) : @xmath398~ , \end{aligned}\ ] ] where the non - abelian couplings are given by @xmath399 here @xmath400 are the one - loop @xmath180-function coefficients for the @xmath161 subgroups theory has no matter hypermultiplets , regardless of whether the final model has @xmath23 or @xmath22 supersymmetry , in the above construction the one - loop @xmath180-function coefficients for the non - abelian subgroups are always given by @xmath401 , where @xmath402 is the one - loop @xmath180-function coefficient of the non - abelian subgroup in the parent @xmath1 theory . ] .
in the previous sections we discussed the brane - bulk duality in the context of @xmath1 as well as certain @xmath22 and @xmath23 large @xmath2 non - conformal gauge theories . in particular , we saw that in the planar limit perturbatively ( on - shell ) correlators in the corresponding @xmath0 theories are the same as in the parent @xmath1 theories .
moreover , the one - loop effective quantum action in such theories , which in the large @xmath2 limit is not perturbatively renormalized beyond one - loop , can be computed by performing a classical computation in a higher dimensional field theory . in this section
we would like to discuss whether the non - perturbative corrections modify the brane - bulk duality picture in such theories .
one of the key simplifying features here is the large @xmath2 limit .
thus , let us consider the @xmath1 supersymmetric @xmath161 gauge theory without matter .
in this theory the low energy effective action can be described in terms of a prepotential @xmath396 , which perturbatively does not receive corrections beyond one loop .
the non - perturbative corrections come from instantons @xcite : @xmath403 where @xmath404 is the dynamically generated scale of the theory : @xmath405 here @xmath406 is the yang - mills gauge coupling at some high scale @xmath228 , and @xmath259 is the one - loop @xmath180-function coefficient .
thus , the instanton corrections are weighted with @xmath407 where @xmath408 is the effective t hooft coupling .
note that these weights go to zero in the t hooft limit , which implies that the low energy effective action is not renormalized beyond one loop in the large @xmath2 limit .
next consider the @xmath22 and @xmath23 orbifold theories discussed in subsection iiia ( as well as section vi ) and section vii , respectively . due to their underlying @xmath1 structure ,
in these theories we might hope that the non - perturbative corrections to the low energy effective action also vanish in the large @xmath2 limit .
if so , then we have non - trivial statements about infinitely many non - trivial @xmath0 gauge theories , in particular , that in such theories the low energy effective action is not renormalized beyond one loop in the planar limit . checking this conjecture in the non - supersymmetric case is rather non - trivial , but in @xmath22 cases
we can perform some partial checks . in particular
, this conjecture implies that in the large @xmath2 limit the superpotential should not receive non - perturbative corrections , so that the classical superpotential should be exact as the superpotential does not receive any loop corrections in @xmath22 supersymmetric theories .
this statement can indeed be checked explicitly for such theories . instead of being most general , here we will consider the simplest example of such an @xmath22 theory ( other @xmath22 cases can be discussed in a similar fashion ) .
thus , consider the example discussed in section vi .
in this example we have @xmath22 supersymmetric @xmath409 gauge theory with chiral matter supermultiplets in @xmath158 , and @xmath159 . to simplify the discussion ,
let us take the gauge coupling of the second @xmath161 factor to be much smaller than that of the first one .
then the second @xmath161 can be treated as the global symmetry group for the first @xmath161 , and we have the @xmath161 gauge theory with @xmath2 flavors of quarks @xmath410 , @xmath411 , where @xmath412 and @xmath413 transform in the fundamental @xmath414 respectively anti - fundamental @xmath415 of the gauge group @xmath161 .
the low energy dynamics is described in terms of the gauge invariant degrees of freedom given by the mesons @xmath416 and baryons @xmath417 @xcite : @xmath418 the classical moduli space in this theory receives quantum corrections , which can be accounted for via the following superpotential ( here @xmath419 is a lagrange multiplier related to the `` glueball '' field via @xmath420 ) @xcite : @xmath421 where @xmath404 is the dynamically generated scale of the theory , which is given by : @xmath422 once again , @xmath423 goes to zero in the large @xmath2 limit , so that the classical constraint @xmath424 is unmodified in this limit .
thus , in this theory the classical superpotential , which vanishes , is indeed exact in the large @xmath2 limit .
the above discussion suggests that the brane - bulk duality discussed in the previous sections in the context of the aforementioned gauge theories might hold even non - perturbatively , so that the corresponding low energy effective quantum action is not renormalized beyond one loop in the large @xmath2 limit .
we would like to end our discussion with a few concluding remarks .
first , one natural generalization we can consider is to extend the above discussion to the cases containing @xmath425 gauge groups .
this can be done via orientifolding , that is , by including orientifold planes in the setup of section ii in the spirit of @xcite . in the @xmath1 cases we expect no subtleties , but in the @xmath22 cases with twists with @xmath82 some caution is needed @xcite due to the subtleties discussed in @xcite .
another point we would like to comment on is the following .
recently , in the brane world context @xcite , it was pointed out in @xcite that the einstein - hilbert term is generically expected to be induced via loop corrections on a brane as long as the brane world - volume theory is not conformal .
subsequently , it was argued in @xcite that this effect should arise in the context of non - conformal gauge theories from d3-brane , in particular , this is expected to be the case in theories discussed in this paper .
it would be interesting to see whether the brane - bulk duality can be used for simplifying computation of such corrections on the d3-branes .
we would like to thank gregory gabadadze , rajesh gopakumar , martin roek , tom taylor , henry tye , cumrun vafa and slava zhukov for valuable discussions .
this work was supported in part by the national science foundation . z.k .
would also like to thank albert and ribena yu for financial support .
in section v we used various brane - bulk couplings in the @xmath31 orbifold examples . these couplings can be computed within the boundary state formalism , where one computes the couplings of the boundary states to the twisted closed string states in the presence of non - trivial d - brane field backgrounds @xcite . here
, however , since we only need the couplings relevant for the one - loop corrections in the gauge theory language , we will take a shortcut and deduce these couplings using the annulus amplitude in the presence of d - brane field backgrounds
. the annulus amplitude in the light - cone gauge is given by ( here @xmath120 is the generator of the @xmath31 orbifold twist ) : @xmath426~.\ ] ] here we assume that we have a non - trivial _ constant _ background gauge field along the brane . the annulus amplitude @xmath427 then is almost the same as in the case without the background field , which we discussed in section ii , with the difference that some of the open string oscillator modings are modified @xcite . for our purposes here
it will suffice to consider the background field of the form ( @xmath246 , @xmath428 , is the d3-brane gauge field strength ) : @xmath429 where @xmath430 is a cartan generator of the @xmath18 gauge group , which we will take to be hermitian .
let @xmath431 , @xmath432 , be the eigenvalues of @xmath430 .
it is convenient to introduce complex combinations of the world - sheet bosonic and fermionic degrees of freedom corresponding to the directions @xmath433 ( so that instead of two real world - sheet bosons and two real world - sheet fermions corresponding to these directions we have one complex world - sheet boson and one complex world - sheet fermion ) .
then the modings of these complex world - sheet degrees of freedom are modified as follows ( modings of other world - sheet degrees of freedom are unchanged ) : @xmath434 where @xmath435 is given by : @xmath436~. \label{eq : epsilon}\ ] ] here the indices @xmath437 label the two ends of the open string .
the cylinder amplitude is given by @xcite : @xmath438 e^{-{t \over 2\pi\alpha^\prime}~x_{\alpha\beta}^2}~,\ ] ] where @xmath439 is square of the distance ( in the two real extra directions untouched by the orbifold ) between the d3-branes labeled by @xmath440 and @xmath180 .
the factor of @xmath441 in the denominator comes from the bosonic zero modes in the directions @xmath442 .
the factor @xmath443 can be understood from the requirement that in the @xmath444 limit we must reproduce the corresponding answer ( see below ) .
in particular , the untwisted character @xmath445 and the twisted character @xmath446 are given by .
thus , for instance , we could have chosen the phase of the last term in the square brackets in ( [ u ] ) as @xmath48 instead of @xmath47 .
note , however , that these phases do not affect our results here as the terms they multiply are vanishing ( @xmath447 ) . ]
( for simplicity the indices @xmath448 are not shown ) : @xmath449 ^ 3 - z_{-1/2}^\delta ~ [ z_{-1/2}^0]^3 \right.\nonumber\\ & & \left.- z_0^{-1/2+\delta } ~[z_0^{-1/2}]^3 + z_{-1/2}^{-1/2+\delta } ~[z_{-1/2}^{-1/2}]^3\right]~,\\ \label{t } t=&&{1 \over \eta^2(q ) ~z_{-1/2}^{-1/2+\delta } ~[z_0^{-1/2}]^2}\left[z_0^\delta ~z_0 ^ 0 ~ [ z_{-1/2}^0]^2 - z_{-1/2}^\delta ~z_{-1/2}^0 ~[z_{0}^0]^2\right.\nonumber\\ & & \left .- z_0^{-1/2+\delta }
~z_{-1/2}^0 ~[z_{-1/2}^{-1/2}]^2 + z_{-1/2}^{-1/2+\delta } ~z_{-1/2}^{-1/2 } ~[z_{-1/2}^0]^2\right]~.\end{aligned}\ ] ] here the characters @xmath450 are the usual complex fermion characters ( @xmath451 , @xmath452 ) : @xmath453 in particular , note that the character @xmath454 in the denominators in ( [ u ] ) and ( [ t ] ) in the @xmath444 limit becomes : @xmath455 combining this with the aforementioned factor @xmath443 gives the following contribution @xmath456 which precisely corresponds to the bosonic zero modes plus oscillators in the @xmath433 directions in the absence of the gauge field background .
here we would like to extract couplings of the massless closed string states to the d - branes .
this can be done by extracting the leading behavior of the annulus amplitude for @xmath68 . using modular transformation properties of the above characters , we obtain : @xmath457\big|_{{\rm { \small ns - ns } } } - 2\cos(\pi\delta)\big|_{{\rm{\small r - r}}}\right\}~,\\ & & t\sim { 2t\over \sin(\pi\delta ) } \left\{1\big|_{{\rm { \small ns - ns } } } - \cos(\pi\delta)\big|_{{\rm{\small r - r}}}\right\}~,\end{aligned}\ ] ] where individual contributions due to the ns - ns and r - r exchanges are shown .
here we are interested in the brane - bulk couplings involving at most quadratic terms in the gauge field strength @xmath246 .
then the relevant terms in the small @xmath458 limit are given by : @xmath459 \left\{1\big|_{{\rm { \small ns - ns } } } - 1\big|_{{\rm{\small r - r}}}\right\}~,\\ & & t\sim { t\over \pi\delta } \left[1-{1\over 3}(\pi\delta)^2\right ] \left\{\left[2+(\pi\delta)^2\right]\big|_{{\rm { \small ns - ns } } } - 2\big|_{{\rm{\small r - r}}}\right\}~.\end{aligned}\ ] ] using the fact , which follows from ( [ delta ] ) , that to the relevant order @xmath460={{1+(2\pi\alpha^\prime b)^2 q_\alpha q_\beta}\over ( 2\pi\alpha^\prime b ) ( q_\alpha - q_\beta)}~,\ ] ] we obtain the following massless untwisted respectively massless twisted closed string contributions into the annulus amplitude @xmath461 \int_0^\infty dt~t~e^{-{t\over 2\pi\alpha^\prime}~x_{\alpha\beta}^2}~,\\ \label{t1 } { \widetilde { \cal c}}_t=&&{1\over2(4\pi^2\alpha^\prime)^2 } \sum_{\alpha,\beta } \left\{\left[2 + ( 2\pi\alpha^\prime b)^2 \left(q_\alpha^2 + q_\beta^2\right)\right ] \big|_{{\rm { \small ns - ns}}}- \left[2 + 2(2\pi\alpha^\prime b)^2 q_\alpha q_\beta\right ] \big|_{{\rm{\small r - r}}}\right\}\nonumber\\ & & \int_0^\infty { dt\over t}~e^{-{t\over 2\pi\alpha^\prime}~x_{\alpha\beta}^2}~.\end{aligned}\ ] ] note that the integrals in these expressions are related to the corresponding euclidean propagators @xmath462 and @xmath463 : @xmath464 here @xmath465 , @xmath466 , and @xmath467 are the momenta corresponding to the coordinates @xmath468 . as we discussed in section v ,
the relevant part of the classical action is given by ( note that for the field of the form ( [ eq : choicef ] ) terms containing @xmath231 are vanishing ) : @xmath234 + b~\sigma~{\rm tr}\left[\gamma_r~f^{\mu\nu}f_{\mu\nu}\right]+ { \rm tr}(\gamma_r)~{\widehat c}~\sigma+ d~\epsilon^{\mu\nu\sigma\rho}~c_{\mu\nu}~ { \rm tr}\left[\gamma_r~f_{\sigma\rho}\right]\big)\nonumber\\ & & -\int_{{\rm d3}\times { \bf r}^2 } \left({1\over 2}\partial^\mu\sigma~\partial_\mu\sigma+{1\over 12 } h^{\mu\nu\sigma } h_{\mu\nu\sigma}\right)~. \label{eq : actionua}\end{aligned}\ ] ] here @xmath184 is a twisted ns - ns scalar , while @xmath235 is a twisted two - form ( whose field strength is @xmath236 ) .
the twisted chan - paton matrix @xmath121 . note , however , that for generic values of @xmath469 the @xmath18 gauge group is broken down to @xmath470 .
so the d3-branes are not necessarily coincident in the two transverse dimensions untouched by the orbifold action . from the tree - level action ( [ eq : actionua ] ) we obtain the following twisted massless contributions quadratic in the gauge field strength ( here we are using @xmath471 ) : @xmath472 note that this expression contains an overall factor of 2 from exchanging the two ends of the string as we are discussing an _ oriented _ string theory . comparing this expression with ( [ t1 ] )
, we obtain : @xmath473 the coupling @xmath474 can be determined in a similar fashion .
z. kakushadze , g. shiu and s .- h.h .
tye , nucl .
* b533 * ( 1998 ) 25 ; + z. kakushadze , phys . lett . *
b455 * ( 1999 ) 120 ; int . j. mod
* a15 * ( 2000 ) 3461 ; phys . lett . *
b459 * ( 1999 ) 497 ; int . j. mod
a15 * ( 2000 ) 3113 .
dienes , e. dudas and t. gherghetta , phys .
b436 * ( 1998 ) 55 ; nucl . phys .
* b537 * ( 1999 ) 47 ; hep - ph/9807522 ; + z. kakushadze , nucl . phys . *
b548 * ( 1999 ) 205 ; nucl . phys .
* b552 * ( 1999 ) 3 ; + z. kakushadze and t.r .
taylor , nucl . phys .
* b562 * ( 1999 ) 78 . | we discuss non - conformal gauge theories from type iib d3-branes embedded in orbifolded space - times .
such theories can be obtained by allowing some non - vanishing logarithmic twisted tadpoles . in certain cases with @xmath0 supersymmetry correlation functions in the planar limit
are the same as in the parent @xmath1 supersymmetric theories .
in particular , the effective action in such theories perturbatively is not renormalized beyond one loop in the planar limit . in the @xmath1 as well as
such @xmath0 theories quantum corrections in the d3-brane gauge theories are encoded in the corresponding classical higher dimensional field theories whose actions contain the twisted fields with non - vanishing tadpoles .
we argue that this duality can be extended to the non - perturbative level in the @xmath1 theories .
we give some evidence that this might also be the case for @xmath0 theories as well .
= 10000 epsf |
rcw 49 is one of the most luminous and massive regions in the galaxy . at its center
lies the westerlund 2 ( hereafter w2 ; westerlund 1960 ) compact cluster which contains over a dozen ob stars and a wolf - rayet star ; another wolf - rayet star lies several arc minutes away in the extended nebula ( moffat & vogt 1975 ; moffat , shara , & potter 1991 ; carraro & munari 2004 ; van der hucht 2001 ) .
the age of the w2 cluster is estimated to be 2 - 3 myr ( piatti et al . 1998 ) . as discussed in churchwell et al .
( 2004 ; hereafter , c04 ) , distance estimates range from @xmath1 kpc and we follow their adoption of 4.2 kpc . at this distance
, we estimate the cluster luminosity is 1.4@xmath2 @xmath3 based on the iras flux and a relationship derived between far - infrared and bolometric luminosity @xcite .
the stellar mass is estimated to be @xmath4m@xmath5 from the radio - derived ionizing flux @xcite the rcw 49 region was observed with the _ spitzer _ ( werner et al .
2004 ) infrared array camera ( irac ; fazio et al .
2004 ) as part of the @xmath6alactic @xmath7egacy @xmath8nfrared @xmath9id-@xmath10lane @xmath11urvey @xmath12xtraordinaire ( glimpse * ? ? ?
* ) observing strategy validation .
c04 presents irac images of this region that show highly - structured , extended polycyclic aromatic hydrocarbon ( pah ) and gas emission , extending out to @xmath13 from the w2 cluster center .
the pah emission is likely excited by the strong ultraviolet / optical radiation from the central w2 cluster ( leger & puget 1984 ; allamandola et al .
1989 ) , suggesting that large regions of the cloud are optically thin to the central cluster radiation .
furthermore , radio and infrared imaging show that at least the southwest part of the cloud is blown out @xcite ( all directions in this paper are referred to in galactic coordinates ) . without detailed studies of the molecular gas in the rcw 49 region , it is unknown if dense cores of gas and dust remain or if the dust is optically thin .
has star formation ceased and is the cloud in the process of disruption ? or is star formation ongoing , perhaps triggered by winds and shocks from the central cluster ?
this paper presents irac photometry of the rcw 49 region which reveals several hundred sources with large infrared excesses , likely indicating youth . after discussing the observations in 2
, we present images of selected regions in 3 and color - color and color - magnitude diagrams in 4 . in
5 we show spectral energy distributions ( seds ) of two massive ( b2 ) ysos . 6 concludes with a discussion of the current generation of star formation and how it may relate to the w2 cluster .
the observations are described in detail in c04 .
a [email protected]@xmath14 region was imaged ten times with 1.2 s exposures in the four irac bands
. we will refer to these bands by their central wavelength in @xmath15 m , i.e. , [ 3.6 ] , [ 4.5 ] , [ 5.8 ] , and [ 8.0 ] . the data were taken on 2003 december 23 ( spitzer program i d 195 ) , and were processed by the _ spitzer _ science center ( ssc ) pipeline ( version s9.0.1 ) .
positional accuracies are better than 1 ( werner et al . 2004 ) .
point source full - width - half - max resolutions range from @xmath161.6 at [ 3.6 ] to @xmath161.9 at [ 8.0 ] .
the data were further processed by the glimpse pipeline @xcite : point sources were extracted from each frame using a modified version of daophot ( stetson 1987 ) , and cross - referenced using the ssc bandmerger .
we produced a catalog of all sources detected at least 8 out of 10 times in any of the four irac bands .
for this present study , we culled the catalog to include only those sources with signal - to - noise greater than 10 in any band .
we estimate the photometric errors from simulations by placing point sources on a realistic background ( using the rcw 49 diffuse flux after stars are subtracted out ) and comparing the processed photometry to the input values .
the root - mean - square errors are @xmath170.05 mag in bands [ 3.6 ] and [ 4.5 ] for sources brighter than 13 and 12.5 magnitude , respectively ; @xmath170.07 mag in band [ 5.8 ] and @xmath170.15 mag in band [ 8.0 ] for sources brighter than 11.7 magnitude .
only sources brighter than these magnitudes ( in a given band ) are displayed in the color - color and color - magnitude diagrams in 4 .
the flux calibration was checked against five early a - type dwarf stars , and agrees to within 7% in all irac bands with the calculated fluxes ( cohen et al .
2003 , kurucz 1993 ) .
figure [ 3col ] shows a 3-color image mosaic at k@xmath18 , [ 3.6 ] , and [ 4.5 ] @xmath15 m ( the k@xmath18-band image is a mosaic of 2mass images ) . the [ 3.6 ] band diffuse emission ( in green )
is dominated by pahs and small grains ; [ 4.5 ] emission ( red ) is mostly hydrogen br@xmath19 and small grains ( c04 ) ; and k@xmath18-band ( blue ) is likely br@xmath20 and perhaps dust scattering .
main sequence stars appear blue in these images .
the boxed regions in figure [ 3col ] are shown to larger scale in figure [ regions ] .
the top left panel of figure [ regions ] shows the w2 cluster .
this cluster contains five o7v stars , one o6v , and a wn7 star ( moffat et al .
1991 ) . the bright star about 1 northwest of the cluster marked by the arrow is an o7 supergiant ( moffat et al .
1991 ) . winds from the w2 cluster have produced for a radio ring about 4 in radius centered on the cluster @xcite .
the second region ( region 2 ) is 3.4 sw of the w2 center . as discussed in 5 ,
the bright red source and diffuse green source ( marked ) are likely massive ( b3 ) ysos .
this cluster resides ( in projection ) in the `` blow - out '' region of the rcw 49 nebula where the radio ring is disrupted and winds and uv radiation are probably escaping ( whiteoak & uchida 1997 ; c04 , figure 1 ) .
the third highlighted region , at 4.3 se of w2 , contains a column of material similar to the elephant trunks of m16 ( hester et al .
the red star at the tip of the elephant trunk is a strong ir emitter ; and the bright red star to the se is likely a massive yso since it has similar flux and colors to the two sources discussed in 5 . to the north at the edge of the image
lies the second known wolf - rayet star in the rcw 49 nebula ( moffat et al . 1991 ) .
it is thought to be responsible for a second radio bubble @xcite .
the fourth region is located 7.7 se of w2 .
it contains several faint , red sources and dark lanes .
region 5 is located 7.8 e of w2 , and region 6 is located 9.1 south of w2 and each contains a small group of red sources .
region 6 is located on the perimeter of the blow - out region .
these images of clustered red sources and dark lanes suggest that star formation is taking place throughout the rcw 49 nebula .
we next examine stellar photometry of the entire region to determine if this is indeed the case .
figure [ ccd ] shows color - color plots of the rcw 49 region ( green circles ) , and a `` field '' region of similar area about 0.75 to the north ( pink circles ) centered at ( l , b)=(284.18,0.41 ) . the black cross at upper right in each panel shows typical error bars for each color sequence .
the black arrows show reddening vectors derived from optical / near - ir extinction studies of the diffuse ism ( cardelli , clayton , & mathis 1989 ) , and the grey arrows show reddening vectors derived from this dataset ( indebetouw et al .
2004 , in preparation ) .
note that the length of the arrows in the top left panel represent an optical extinction of a@xmath21 whereas the length of the arrows in the other three panels show a@xmath22 .
the black symbols represent colors from models of 1 million year old sources ( using siess et al . 2000 evolutionary tracks for central source temperatures and luminosity ) surrounded by varying amounts of circumstellar gas and dust ( @xmath23 - 10 m@xmath24 ) .
the models were computed using a radiative transfer code that calculates emission from envelopes , disks and outflows surrounding stars @xcite .
the model central star masses are 2.5 , 3.8 , and 5.9 m@xmath24 , corresponding to luminosities of 10 l@xmath24 ( filled circles ) , 100 l@xmath24 ( x s ) , and 730 l@xmath24 ( open triangles ) . the 2mass catalog was used to create the near - ir color sequence in the top left panel of figure [ ccd ] .
only stars brighter than 14th magnitude in all three bands are plotted since the errors increase below this .
the near - ir color sequence is typical of a star formation region .
the field stars ( pink ) follow the reddening line .
the stars that are very red or below the reddening line are usually classified as ir excess sources with disks or envelopes .
however , many ir excess sources lie along the reddening line and are difficult to distinguish from reddened main sequence stars . in contrast , the other panels in figure [ ccd ] using irac data show that infrared - excess sources in rcw 49 are well separated from field stars , which generally have a small color range . comparing the colors of the rcw 49 stars to the field stars
, we find that 357 stars out of 1393 in rcw 49 have colors [ 3.6]-[4.5]@xmath25 mag , [ 4.5]-[5.8]@xmath26 mag , or [ 5.8]-[8.0]@xmath25 mag compared to 21/1114 in the field ( about 2% ) .
we note that due to our conservative brightness limits for this initial study , many of the faint stars apparent in the images in figure [ regions ] are not included in the color - color plots .
even so , it is likely that at least 300 of the sources in this dataset are surrounded by warm circumstellar dust which indicates that they have recently formed or are in the process of formation .
figure [ cmd ] shows color - magnitude diagrams of the rcw 49 region ( green circles ) and the northern field ( pink ) .
the model symbols ( black ) are as described in figure [ ccd ] .
the width of the field star distribution matches reasonably well to the error bars at upper right .
the horizontal dashed line in each panel shows the expected brightness of a b5 main sequence star at a distance of 4.2 kpc ( synthesized from the 2 - 35@xmath15 m spectra of cohen , 1993 ) .
based on the number of ysos more luminous than this ( about 300 ) , and assuming a fairly standard imf @xcite , we estimate the total number of ysos to be @xmath277000 and their total mass to be @xmath274500 m@xmath5 . where is most of the star formation occuring in this cloud ?
figure [ radial ] plots the fraction of red sources as a function of radius from the w2 cluster center , where red is defined as [ 3.6]-[4.5]@xmath25 mag , [ 4.5]-[5.8]@xmath26 mag , and [ 5.8]-[8.0]@xmath28 mag ( based on the color - magnitude diagrams of figure [ cmd ] ) .
this distribution peaks within about 4 ( @xmath29 pc ) from the w2 cluster center , and decreases beyond that .
the diffuse emission has a similar distribution ( c04 , figure 3 ) suggesting the possibility of a systematic bias towards finding red stars due to higher [ 5.8 ] and [ 8.0 ] backgrounds .
however , the density of total stars ( of all colors ) as a function of radius is fairly constant beyond about 2 suggesting that we are seeing the entire population of stars above our magnitude limits .
the correlation between the red stars and the diffuse flux is probably real and due to higher diffuse flux in regions of higher star formation activity .
we conclude from this plot that star formation is occurring preferentially within about 4 ( @xmath30 pc ) from the w2 cluster center and decreases beyond that .
interestingly , this is the approximate outer radius of the radio ring @xcite and the `` plateau region '' of the diffuse flux ( c04 , figure 3 ) where it is thought that winds and radiation have swepted up and compressed gas and dust .
this may have produced conditions for a second generation of star formation .
since the medium is not smooth , the induced star formation likely occurs over a range of radii .
the ysos in region 2 ( figures 1 and 2 ) and the elephant trunk in region 3 fall within this region .
interestingly , the ysos in region 6 lie on the edge of the blow - out cavity .
however , regions 4 and 5 are well beyond the radio ring and may represent a more quiescent form of star formation .
in this section , we present seds of just two of the many interesting sources in the rcw 49 nebula .
these are in region 2 of figure [ regions ] , marked by arrows .
the observed seds include data from 2mass , _ spitzer _ irac , and _
msx_. the source indicated by the left arrow in region 2 is the brightest source in the region , and is mildly saturated at irac [ 5.8 ] and [ 8.0 ] bands . the source shown by the right arrow is a triangular - shaped greenish object .
the seds of both sources are shown in figure [ sed ] .
irac photometry of the bright source was done by our automated pipeline .
we did aperture photometry of the triangular source over a 4 aperture .
the flux at @xmath31 m from these sources is blended at _ msx _
resolution ( 18 ) , though likely dominated by the source at left .
it is clear from the brightness of these sources that they are luminous and likely high - mass .
we computed model seds for the ysos using the code of whitney et al .
alvarez , hoara , & lucas ( 2004 ) modeled near - ir scattered light images of massive ysos using the same geometries we use to fit the seds in figure [ sed ] : rotationally flattened infalling envelopes , disks , and dust - filled outflow cavities . the model that fits the bright source sed best
is powered by a b2 main sequence star , has an envelope with an infall rate of @xmath32 m@xmath24/yr and a disk radius of 500 au ; this gives an envelope extinction along the line - of - sight of a@xmath33 .
the model that fits the triangular source is powered by a b2.5 star , has an infall rate of @xmath34 m@xmath24/yr , disk radius of 100 au , and envelope extinction of 200 along the line - of - sight .
both models are viewed at an inclination of @xmath35(where pole - on is defined as @xmath36 ) , and have foreground extinctions of a@xmath3715 .
the seds could probably be fit with other envelope structures as long as a substantial amount of dust resides close to the star , and therefore the stars are sufficiently young that very little dust clearing has occured .
the large number of infrared excess sources in rcw 49 indicates that star formation is ongoing in this giant region .
we find that at least 300 stars brighter than 13@xmath38 magnitude in the [ 3.6 ] band are pre - main sequence or young stars of mass @xmath39 m@xmath5 . based on integration of a standard imf ,
we conclude that the ysos in the rcw 49 complex comprise a stellar mass of @xmath274500 m@xmath5 .
if the ysos represent a second generation of star formation , this suggests that the mass of the second generation is about 15% of the first .
the presence of ysos of b spectral type does indeed suggest that some star formation is very recent . whether it is continuous or triggered by the w2 cluster
is uncertain . the apparent excess of star formation near the radio / ir ring ( figure [ 3col ] ) and several of the sites of star formation near the ring and blowout region perimeters ( regions 2,3 , and 6 in figure [ regions ] ) suggest that energy input from w2 may have triggered recent star formation . on the other hand ,
regions 4 and 5 appear to be forming stars in a relatively quiescent region of the cloud .
the detailed nature of the star formation in rcw 49whether triggered or continuous , clustered or distributed , and the spatial distribution of mass requires further study by the wider astronomical community .
the images of this field have been released with the opening of the public archive ; the glimpse catalog and mosaic images will be available to the community from the _ spitzer _ science center in the fall of 2004 .
we are grateful to stephan jansen for his invaluable work maintaining the glimpse computing network .
support for this work , part of the spitzer space telescope legacy science program , was provided by nasa through contract numbers ( institutions ) 1224653 ( uw ) , 1225025 ( bu ) , 1224681 ( umd ) , 1224988 ( ssi ) , 1242593 ( ucb ) , 1253153 ( umn ) , 11253604 ( uwy ) , 1256801 ( uww ) by the jet propulsion laboratory , california institute of technology under nasa contract 1407 . | glimpse imaging using the infrared array camera ( irac ) on the _ spitzer space telescope _ indicates that star formation is ongoing in the rcw 49 giant region .
a photometric comparison of the sources in rcw 49 to a similar area to its north finds that at least 300 stars brighter than 13th magnitude in band [ 3.6 ] have infrared excesses inconsistent with reddening due to foreground extinction .
these are likely young stellar objects ( ysos ) more massive than 2.5 @xmath0 suggesting that thousands more low - mass stars are forming in this cloud .
some of the ysos are massive ( b stars ) and therefore very young , suggesting that a new generation of star formation is occurring , possibly triggered by stellar winds and shocks generated by the older ( 2 - 3 myr ) central massive cluster .
the _ spitzer _
irac camera has proven to be ideally suited for distinguishing young stars from field stars , and the glimpse survey of the galactic plane will likely find thousands of new star formation regions . |
within the current framework of hierarchical structure formation ( e.g. white & rees 1978 ) galaxy clusters are formed through the successive mergers of galaxies , galaxy groups , and subclusters .
galaxy clusters form , thus , the largest gravitationnally bound structures in the universe .
interestingly , x - ray observations have shown that most of the baryonic mass of galaxy clusters resides not in galaxies but in their hot intracluster gas ( e.g. jones & forman 1984 ) abell 85 is a rich galaxy cluster located at a redshift of @xmath1 with 305 confirmed cluster members ( durret et al .
abell 85 is a bright x - ray source that has been extensively studied using several x - ray satellites ( e.g. markevitch et al .
1998 ; lima neto et al . 2001 ; sivakoff et al . 2008 ) .
the x - ray emission of abell 85 testifies to an intense past merging activity ( durret et al .
moreover , abell 85 is not fully relaxed and is currently merging with at least two satellite subclusters ( kempner et al .
the complex dynamical state of abell 85 was recently discussed in great detail by ichinohe et al .
( 2015 ) . due to its richness
, abell 85 has also been the target of several studies on the morphology - density relation ( e.g. fogarty et al .
2014 ) . located in the core of galaxy clusters , and formed through a rich merger history , brightest cluster galaxies are , in turn , the most massive and luminous galaxies in the universe ( de lucia & blaizot 2007 ) . recently , lpez - cruz et al .
( 2014 ) reported that the bcg of abell 85 has the largest galaxy core ever discovered .
note that the bcg of abell 85 has also been identified as holm 15a .
the unusually large core in the surface brightness profile of the abell 85 bcg translates into the presence of a supermassive black hole with masses above @xmath2 .
the mass of the black hole is obtained by using scaling relations between galaxy cores and black hole masses ( e.g. kormendy & ho 2013 ) .
the results of lpez - cruz et al .
( 2014 ) were challenged by bonfini et al .
( 2015 ) who find that the abell 85 bcg does not have a depleted core .
in fact , bonfini et al .
( 2015 ) find that a srsic profile plus an outer exponential component provide a good fit to the data .
galaxy cores are defined as a relative light deficit towards the nucleus of the galaxy compared to the inward extrapolation of the surface brightness profile of the outer components of the galaxy .
the physical theory postulated to explain the presence of these cores is the action of binary supermassive black holes that , through three - body interactions , slingshot away stars in the galactic center ( begelman et al .
1980 ) .
due to their possible link to black holes and galaxy formation , the study of cores is an active field with many authors looking at different theoretical and observational aspects .
for instance , through @xmath3-body simulations , milosavljevi & merritt ( 2001 ) modeled the decay of a black hole binary and how it carves galactic cores .
observationally , a major development in the study of galaxy cores came with the analysis of advanced camera for surveys data . this instrument
on - board hst provided both superb resolution and large radial coverage , key factors in deriving an accurate surface brightness profile .
ferrarese et al .
( 2006 ) use a uniform sample of 100 galaxies in virgo imaged with the acs to derive their surface brightness profile .
they find that the surface brightness profiles of most galaxies are well fit by a srsic ( 1968 ) profile .
earlier studies of galaxy cores with the hst include the work of faber et al .
( 1997 ) and laine et al .
( 2003 ) , among many others . in the following sections we present high - resolution gemini observations that have been obtained in order to study in detail the surface brightness profile of the abell 85 bcg , particularly focusing on its nuclear region .
gemini south observations of abell 85 were obtained under the director s discretionary time program gs-2014b - dd-6 .
abell 85 was observed with the gemini multi object spectrograph ( gmos ) on imaging mode with the detector centered on the bcg , as shown in figure 1 .
the data was obtained on 2014 november 15 under stable atmospheric conditions on cerro pachn with a seeing of 0.56 arcseconds .
two exposures of 200 s each were acquired during that night .
the filter in use was @xmath4 , this filter is centered at 630 nm and has a filter width of 136 nm .
we used a @xmath5 binning that gives an effective pixel scale of 0.160 arcseconds per pixel .
we adopt a redshift of abell 85 @xmath6 which yields a distance of 233.9 mpc and a scale of 1.075 kpc / arcsec . the gemini gmos pixel scale for this observation of abell 85
is thus 172 pc / pixel .
data were processed with the standard gemini pyraf package using the tasks described in this section .
we obtained bias , and twilight flats from the observatory .
these calibration files were already processed through the tasks gbias and giflat .
the science images were bias subtracted and divided by the flatfield using the task gireduce .
these data were acquired using the new ccd detectors ( hamamatsu ) recently installed on gmos .
raw science data files have 12 extensions reflecting the fact that the detector has three ccds and four amplifiers per ccd .
a single component image was made for each of the two exposures using the task gmosaic .
the final science image was created combining the two exposures using the task imcoadd .
for the study of the surface brightness profile , the task lucy , within stsdas , is used to deconvolve the image by applying the lucy - richardson algorithm ( lucy 1974 ; richardson 1972 ) .
the task lucy converges after nine iterations yielding a final resolution of 0.45@xmath7 .
the ellipse routine ( jedrzejeswki 1987 ) is applied to the science image in order to extract the 1d luminosity profile of the abell 85 bcg .
this profile , directly obtained from the science image is shown in figure 2 as blue circles . in order to obtain an accurate light profile for the targeted galaxy ,
all nearby sources are properly masked and the task ellipse ran iteratively until the surface brightness profile converges .
we also ran the task ellipse with different initial parameters , while we held some ( or all ) of the parameters ( center , position angle , and ellipticity ) fixed to a constant value to test wether we observe possible variations of the luminosity profile .
the resulting luminosity profile turned out to be robust and shows no dependence on the initial parameters .
also , we did not observe any shift in the center of the isophotes during this experiment .
ellipticity remained mainly constant i.e. @xmath8 within the innermost few arcseconds .
small ellipticity values imply relatively large errors for the position angle .
appropriately removing the sky background is also of great importance in order to obtain an accurate surface brightness profile .
the gmos imager provides a relatively large field of view , at least when compared to hst detectors , this allows us to make a first estimate of the sky background in an area of the detector where the abell 85 bcg has very low emission .
data taken by the canada - france - hawaii - telescope ( cfht ) and surface brightness profile of this galaxy published by donzelli et al . ( 2011 ) were also used to estimate the sky background .
cfht data of abell 85 , in the @xmath0 band , taken under the multi - epoch nearby cluster survey ( meneacs ) ( sand et al .
2011 ) in september 2008 .
the basic assumption we use is that the surface brightness profile of the bcg of abell 85 should be identical at intermediate radii , independent of the telescope in use . once a correct estimate of the background is made , both gemini and cfht surface brightness profiles agree well , with the exception of the galaxy core where seeing effects dominate . ccccccccccc 21.78 & 17.21 & 18.48 & 1.24 & 3.33 & 0.0 & 4.57 & 1.67 & ... & kpno 0.9 m & lpez - cruz et al . + 22.32 & 19.09 & 20.50 & 1.22 & 3.62 & 0.0 & 4.57 & 0.74 & ... & cfht 3.5 m & lpez - cruz et al .
+ 21.05 & 10.70 & 11.56 & 1.90 & 2.29 & 0.14 & 5.02 & 0.56 & 116 & gemini 8 m & this work + @xmath9 & @xmath10 & @xmath10 & @xmath11 & @xmath12 & @xmath13 & @xmath13 & @xmath14 & @xmath15&telescope & reference + ( mag/@xmath16 ) & ( @xmath7 ) & ( kpc ) & & ( mag/@xmath16 ) & ( @xmath7 ) & ( kpc ) & & & + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) & ( 11 ) + + 21.71 & 14.59 & 15.68 & 0.933 & 24.87 & 70.79 & 76.10 & 0.839 & 47 & gemini 8 m & this work + @xmath9 & @xmath10 & @xmath10 & @xmath11 & ... & ... & ... & ... & @xmath15 & telescope & reference + ( mag/@xmath16 ) & ( @xmath7 ) & ( kpc ) & & & & & & + ( 1 ) & ( 2 ) & ( 3 ) & ( 4 ) & ( 5 ) & ( 6 ) & ( 7 ) & ( 8) & ( 9 ) & ( 10 ) & ( 11 ) + + 23.24 & 342.3 & 368.0 & 0.13 & ... & ... & ... & ... & 117 & gemini 8 m & this work +
lpez - cruz et al .
( 2014 ) use the nuker model to fit the surface brightness profile of the bcg of abell 85 derived with data taken by the kpno 0.9 m telescope and a seeing of 1.67@xmath7 .
these data are not publicy available but were given to us by o. lpez - cruz .
as shown in the top panel of figure 2 and in figure 3 , the new gemini data reveals the presence of nuclear emission .
this central and distinct feature is completely absent from the data presented by lpez - cruz et al .
in fact , the surface brightness profile presented by the authors above is featureless within the inner 20 kpc of the center of the galaxy . using hst data , ct et al .
( 2006 ) clearly demonstrates that ground based data with poor seeing underestimates the presence of nucleii in nearby elliptical galaxies .
the gemini data shows that from @xmath176 kpc inwards , the extrapolation of the surface brightness profile results in a light excess not a light deficit as one might believe is the case when looking at the lower quality kpno data .
the above is true regardless of the model chosen to fit the surface brightness profile of the galaxy .
it should be noted that the surface brightness profiles shown in figure 2 were derived using data that received no additional processing beyond basic data reduction . in their analysis of the cfht data
bonfini et al .
( 2015 ) detect a tiny bump " in the light profile within the inner 0.5@xmath7 .
indeed , their core - srsic model fits a light excess rather than a light deficit within the inner @xmath18 . at the faint end of the profile , shown in the bottom panel of fig . 2
, the kpno data becomes noisy beyond @xmath19 kpc from the center of the galaxy .
similar noise levels are only present in the gemini data at @xmath20 kpc .
the presence of a clear light excess within the innermost kiloparsec of the bcg of abell 85 prompts us to discuss its physical origin .
one might attribute this nuclear emission to the presence of an agn given that bcgs are more likely to host a radio - loud agn than other galaxies of similar mass ( best et al .
the detection of x - ray emission co - spatial with the galaxy core can be candidly thought to be proof of the existence of an agn .
the picture for abell 85 is more complex .
in fact , sivakoff , et al .
( 2008 ) exclude the bcg from a census of active galactic nuclei in abell 85 given that its position also corresponds , within a few arcseconds , to the peak x - ray emission of the intracluster medium .
agns also have distinctive radio emission and abell 85 has been observed in the radio ( bagchi et al .
1998 ; slee et al . 2001 ; schenck et al . 2014 ) .
based on the morphology of the radio emission and its spectrum the above authors do not find evidence of strong agn activity for the bcg of abell 85 .
the radio maps of abell 85 do not show jets or lobes that are the clear signatures of strong and current agn activity . on the contrary ,
those radio maps are consistent with the presence of radio relics from shocked gas or from a dead radio galaxy , the latter not obviously cospacial with the bcg ( schenck et al .
2014 ) .
we measured the flux difference in the core of the bcg between the cfht and gemini images .
these images were taken about six years apart : 2008 september ( cfht ) and 2014 november ( gemini ) .
fluxes were measured within an aperture of 1.5@xmath7 for each detector .
this is the aperture at which the integrated magnitudes of a point source converge for both detectors .
we find @xmath21mag = 0.10 @xmath22 0.04 , that is , the nucleus of the bcg has become brighter by 0.10 magnitudes during the last six years .
this type of optical variability is suggestive of the presence of an agn in the core of the bcg .
it is known that all agns vary in short timescales ( e.g. ulrich et al .
we should note that we also measure the flux difference in the core of a dozen random galaxies common to both images and find no difference above the uncertainty level of 0.02 mag .
the variability within the core of the bcg of abell 85 , discussed above , hints to the presence of an agn but can also be of stellar origin particularly in a dense nuclear stellar structure .
variability is indeed a defining property of agns that is often used for their discovery .
for instance , cohen et al . ( 2000 ) search for variable galaxies in the hubble ultra deep field to investigate the presence of agn and find 45 solid candidates .
those agn candidates show characteristic variability of @xmath21mag @xmath170.01 to 0.8 mag . on the other hand ,
dense star clusters are favorite crash sites for binary stars , cataclysmic variables , and classical novae among other stellar exotica ( knigge et al .
classical novae have been found in extragalactic globular clusters and their erupting luminosity is comparable to their entire host ( shara et al .
2004 ) .
we find that the core of the abell 85 bcg is resolved with a fwhm of about 0.85@xmath7 , that is , about 50% larger than the fwhm of a stellar psf . at the distance of abell 85
, the physical size of the central component is thus @xmath170.9 kpc .
this central nuclear component , within the first arcsecond , can be easily modeled as a gaussian function with an integrated magnitude of @xmath23 mag .
the resolved stellar structure in the core of the abell 85 bcg has a physical size that is too large when compared to nuclear star clusters .
indeed , nuclear star cluster sizes are of the order of a few parsecs ( e.g. @xmath173 pc ; bker 2010 ) .
also , the largest nuclear structure found by ct et al . ( 2006 ) in their survey of the virgo cluster has an effective radius of 62 pc .
based on its size , a nuclear stellar disk ( nsd ) is a more compatible candidate for the origin of the nuclear emission of the abell 85 bcg .
for instance , ledo et al . ( 2010 ) compile a census of nuclear stellar disks in early type galaxies .
several of these nsds have sizes of a few hundred parsecs with two of them having sizes larger than 1 kpc . if the central structure of abell 85 is indeed a nuclear stellar disk it would be among the largest reported so far .
to give more context we remark that the catalogue of ledo et al .
( 2010 ) is limited to galaxies within 108 mpc while abell 85 is at more than twice this distance .
also , the abell 85 bcg is brigther than the sample studied by ledo et al .
( 2010 ) .
laine et al . ( 2003 ) study a sample of bcgs with luminosities and distances similar to those of abell 85 and find the presence of two nuclear stellar disks ( abell 261 and abell 1142 ) .
laine et al .
( 2003 ) also found an additional seven bcgs with point - like nucleii that exhibit similar morphology to the abell 85 bcg . with the current data , we favor a nuclear stellar disk as the physical explanation for the nuclear structure present in the core of the abell 85 bcg .
the nuclear variability we measure is , however , more likely associated with an agn .
additional data points are needed in order to build a better sampled light curve and unambiguously identify the origin of the variability .
different models used to fit the radial surface brightness profile of galaxies can be found in the literature . commonly used analytical functions are the srsic ( 1968 ) profile , which is a generalization of the de vaucouleurs ( 1948 ) and exponential profiles , the moffat ( 1969 ) profile , and the gaussian profile .
models that use additional parameters to account for the parametrization of galaxy cores are a blend of two power laws ( ferrarese et al .
1994 ) , the nuker model ( lauer 1995 ) , and the core - srsic profile ( graham et al .
2003 ) . the king ( 1966 )
model is commonly used to fit the radial light profile of globular clusters and small galaxies . in this section ,
the results of fitting different analytical models to the new gemini data are presented .
the results of our best fits and the fits of lopez - cruz et al .
( 2014 ) are shown in table 1 .
the fits below are applied to the deconvolved image .
the fits are carried out to a galactocentric distance of 115 kpc , that is , where the standard deviation of the sky ( @xmath170.3 mag ) equals the uncertainty on the galaxy surface brightness .
schombert ( 1987 ) showed that a de vaucouleurs ( 1948 ) model fails to properly fit the surface brightness profile of the abell 85 bcg .
a de vaucouleurs fit for this galaxy overestimates the flux at the center while it underestimates the flux in the outskirts .
schombert ( 1987 ) , also showed that this result was also true for several other bcgs .
+ we fit a single srsic profile to the new gemini data . for clarity
, we define the srsic profile in its canonical form @xmath24 , where the concentration parameter @xmath25 is the inverse of the srsic index ( srsic 1968 ) : @xmath26\big\}\ ] ] in this equation @xmath27 is the intensity at @xmath28 at the effective radius .
the values for @xmath29 can be calculated using @xmath30 ( caon et al .
1993 ) . the best fit of the srsic model to the gemini data is shown in figure 4 .
we find that a single srsic provides a good fit to the data only over a limited section of the surface brightness profile .
model and data diverge at small radii , that is , for radii below @xmath1718 kpc
. the numerical parameters of the best fit using a single srsic model , such as an exceedingly large effective radius , expose the fact that this model does not provide a good physical representation of the overall surface brightness profile . studying a large sample of elliptical galaxies , kormendy et al .
( 2009 ) also find that a single srsic profile is a good fit to the main section of the radial profile , while the model deviates from the data at small radii .
similarly , lasker et al .
( 2014 ) need to include additional components beyond a single srsic profile when fitting the surface brightness profile of 35 nearby galaxies .
additional models correspond to additional physical components such as bars , nuclei , inner disks , and envelopes .
we also fit a nuker model to the gemini data and present the results in figure 5 .
numerical parameters are given in table 1 .
interestingly , we find an even larger cusp radius , @xmath31 kpc than the one found by lpez - cruz et al .
we find however a break radius of @xmath32 kpc , almost half the value of lpez - cruz et al .
( 2014 ) .
an important caveat to fitting a nuker model to this galaxy is the fact that it does not actually have a flat evacuated core and the nuker model does not identify the presence of the nuclear component .
the existence of this distinct nuclear component naturally changes the results given by the nuker model fit , as shown above .
also , it has been proved that the nuker model is dependent on the radial extent of the fit ( graham et al .
moreover , the nuker model was never intended to fit the entire surface brightness profile but the central region of any given galaxy
see the recent work of bonfini et al .
( 2015 ) and references therein . the nuker model used by lpez - cruz et al .
( 2014 ) underestimates their data beyond @xmath1720 kpc .
the failure of the nuker profile at large radii prompts lpez - cruz et al .
( 2014 ) to fit a de vaucouleurs profile ( i.e. srsic profile with @xmath33 ) .
it should be noted that the de vaucouleurs profile used by lpez - cruz et al .
( 2014 ) overestimates their data in the outskirts of the galaxy . in a recent work bonfini et al .
( 2015 ) carry out a detailed re - analysis of the cfht data of the abell 85 bcg focusing on fitting the core - srsic model .
bonfini et al .
( 2015 ) find that the abell 85 bcg does not have a depleted core as the light profile does not show a light deficit when fitted with the core - srsic model .
in fact , these authors find that the abell 85 bcg is not well adjusted by a core - srsic due to a light excess in its central surface brightness profile
. bonfini et al .
( 2015 ) find that the abell 85 bcg is a core - less galaxy whose surface brightness profile is best fit by an inner srsic profile and an outer exponential halo .
gonzalez et al .
( 2003 , 2005 ) and seigar et al .
( 2007 ) showed that two component models are necessary to accurately fit the surface brightness profiles of bcgs .
the two components pertain to an inner and outer component .
the existence of an outer component reflects the fact that bcgs often have extended envelopes ( schombert 1987 ) .
donzelli et al .
( 2011 ) studied the luminosity profiles of 430 bcgs and found that about half of them required a double srsic model ( srsic @xmath34 exponential ) .
in fact , donzelli et al .
( 2011 ) fit the abell 85 bcg with an inner srsic model and an outer exponential component .
we fit a double srsic profile to the gemini data , and obtain satisfactory results .
the values of our best fit are shown in table 1 and figure 6 .
one component has a srsic index of @xmath35 ( dashed line ) while the second srsic has an index of @xmath36 ( dotted line ) .
the sum of these two components ( a solid line ) is indistinguishable from the data .
models that explain the underlying physical mechanisms that create double srsic components have been postulated by cooper et al .
these authors postulate that for bcgs , the double srsic profile originates from the superposition of two debris components of different progenitors .
the inner profile is associated with relaxed accreted components while the outer profile corresponds with unrelaxed accreted debris ( cooper et al .
2015 ) .
the best fit to the surface brightness of this galaxy is given by a double srsic model .
values for the goodness of fit ( @xmath15 ) are presented in table 1 and a residual image is shown on figure 7 .
we should note that the @xmath15 value given for the single srsic fit pertains to a fit between 20 and 115 kpc , a fit over the whole range of the data would yield @xmath37 .
based on the analysis of new gemini data presented above we conclude that the abell 85 bcg ( holm 15a ) is a nucleated , coreless galaxy .
that is , it does not have an exceptionally large core due to a light deficit in its central region ( lpez - cruz et al .
our results thus nullify the existence of a supermassive black hole based solely on the presence of a depleted core that this galaxy , in fact , does not have . by fitting the nuker model to the surface brightness profile
we find a large cusp and break radius .
we refrain from interpreting the cusp and break radius as representative of an evacuated core created by the scouring action of a binary black hole .
a large cusp radius , derived from the nuker model , does not necesarily imply a downward bend of the inner light profile .
recently , bonfini et al .
( 2015 ) point out that the presence of a singular point in the surface brightness profile of any given galaxy does not imply the presence of a depleted core . in the words of bonfini
( 2015 ) most galaxies have particular values for the negative logarithmic slope of the intensity profile but this is not a sufficient condition for the existence of a depleted core .
moreover , the central brightness profile of the abell 85 bcg is indeed different from the flat , or even decreasing surface brightness profile of , for instance , the bcg of abell 2261 ( postman et al .
the presence of nuclear structure is difficult to reconcile with a core of @xmath38 kpc where other stars within that core are ejected .
the authors would like to thank the referee for a detailed and constructive report that helped us improve this paper .
we are grateful to nancy levenson , gemini head of science , for granting us director discretionary time to carry out these observations under program gs-2014b - dd-6 .
we also thank gemini observer mischa schirmer , and gemini visiting astronomer ricardo de marco for obtaining the data .
the cfht data was kindly shared with us by melissa graham and david sand , likewise , the kpno data was made available to us by omar lopez - cruz .
this research has made use of the nasa astrophysics data system bibliographic services ( ads ) and google .
this work was supported by a grant from secyt - unc , argentina .
based on observations obtained at the gemini observatory which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) . | new high - resolution @xmath0 band imaging of the brightest cluster galaxy ( bcg ) in abell 85 ( holm 15a ) was obtained using the gemini multi object spectrograph .
these data were taken with the aim of deriving an accurate surface brightness profile of the bcg of abell 85 , in particular its central region .
the new gemini data show clear evidence of a previously unreported nuclear emission that is evident as a distinct light excess in the central kiloparsec of the surface brightness profile .
we find that the light profile is never flat nor does it present a downward trend towards the center of the galaxy .
that is , the new gemini data show a different physical reality from the featureless , `` evacuated core '' recently claimed for the abell 85 bcg . after trying different models ,
we find that the surface brightness profile of the bcg of abell 85 is best fit by a double srsic model . |
[ intro ] a key element for understanding the complexity and perplexity in high-@xmath3 cuprates , manganites and colossal magnetoresistance ( cmr ) nanomaterials is the experimental observation of phase separation ( ps ) instabilities at the nanoscale signaled by spin - charge separation and quantum phase transitions ( qpts ) @xcite .
a new guiding principle for the search of new materials with enhanced tc is the proximity to quantum critical points ( qcps ) for spontaneous first order qpts attributed to intrinsic spatial inhomogeneities ( see ref .
@xcite and references therein ) .
strong quantum fluctuations dominate thermal fluctuations and affect the classical properties well above absolute zero temperature @xcite .
the inhomogeneous concentrated system in equilibrium can be well approximated as a quantum gas of decoupled clusters , which do not interact directly but through the grand canonical ensemble , with different electron number per cluster .
our results for possible spatial inhomogeneities are directly applicable to nanoparticles and respective bulk nanomaterials which usually contain an immense number of isolated clusters in contact with a thermal reservoir by allowing electron number per cluster to fluctuate
. the finite - size optimized clusters may be one of the _ few solid _ grounds available to handle this challenging problem in a bottom - up approach @xcite by defining canonical and grand canonical local gap order parameters in the absence of a long - range order , spin or charge density waves @xcite .
the ps instabilities and spin - charge separation effects in bipartite hubbard clusters driven by on - site coulomb interaction @xmath4 display qcps which strongly depend on cluster topology @xcite . in frustrated ( nonbipartite )
geometries spontaneous transitions depend on the sign of the coupling @xmath5 and can occur for all @xmath4 by avoiding qcps ( level crossings ) at finite @xmath4 .
the existence of the _ intrinsic qcps _ and inhomogeneities associated with the ps instabilities , are crucial ingredients of the superconducting ( sc ) and ferromagnetic qpts , providing important clues for understanding the _ incipient microscopic mechanisms _ of pairing instabilities in real space due to coexisting high / low electron ( hole ) or high / low spin up ( down ) densities in high-@xmath3 superconductors ( htscs ) and colossal magnetoresistive ( cmr ) nanomaterials respectively .
however , small systems suffer from finite - size ( edge ) effects , so it is unclear whether the observed instabilities can survive in the thermodynamic limit .
thus , tests on reduced boundary effects are necessary to confirm the picture of local instabilities in larger systems in the so - called optimized " betts building blocks ( finite square lattices ) @xcite .
a square infinite lattice is tiled by identical square unit cells containing @xmath6 sites which periodically repeat in the lattice .
for example , @xmath1-site unit betts cell in fig .
[ fig1 ] is used to fill the whole lattice .
the square units restore basic symmetrical properties of the infinite square lattice and periodicity of clusters partially preserves translational and rotational symmetries of the original lattice . therefore
, betts cells are believed to be the most preferred unit blocks for relieving frustrations over other structures with the same size .
all the betts unit cells can be defined uniquely by two edge vectors @xcite which represent translational vectors between the nearest two clusters .
the lattice generated by betts unit cells provides useful insights into certain physical aspects of the phase diagram in the @xmath7 model @xcite . to our knowledge ,
an exact calculation of phase separation and pairing under doping has not been attempted in the betts lattices applied to the hubbard model either with nearest or next nearest neighbors .
different two - dimensional ( @xmath8 ) square structures , defined by the condition @xmath9 with a linear size @xmath6 ( @xmath10 , @xmath11 are integers ) , can be used as plaquettes to extrapolate the results to the infinite square lattice . here
our primary goal is an exact study of critical instabilities in the two - dimensional @xmath1- and @xmath2-site betts ( generated ) lattices .
-site finite unit cell ( block ) repeating periodically fill the entire ( _ infinite _ ) square @xmath8 lattice .
the cells can be defined by edge vectors @xmath12 and @xmath13 ( arrows in the figure ) as defined in ref .
the lattice in two dimensions can be tiled by periodically repeated betts ( isotropic ) unit cells with reduced edge ( boundary ) effects . in fig .
[ fig1 ] an infinite square lattice has been tiled by @xmath6-site squares with edge vectors @xmath12 and @xmath13 , which represent displacements of one vertex to the equivalent vertex in the neighboring tile with which it shares an edge .
notice that , if the sites are numbered as in fig .
[ fig1 ] , each odd site in the plaquette is surrounded by all the even sites ( as nearest neighbors ) and vice versa .
the betts unit cells take full advantage of the local space group symmetries of the isotropic @xmath8 ( square ) bipartite lattice . the two - orbital hubbard model with p bonding orbitals after elimination of the cu sites can be reduced to a single - orbital hubbard model with the nearest and next - nearest - neighbor ( nnn ) couplings by allowing holes to move within a given oxygen sublattice .
thus , we consider the minimal hubbard model @xmath14 where summation goes through all lattice sites @xmath6 with coupling integral @xmath15 equal to @xmath5 for the nearest and @xmath16 for the next nearest neighbors .
the lattice frustration with @xmath17 allows study of electron pairing in the absence of electron - hole symmetry .
below , an exact diagonalization technique is used to extract the pairing instabilities and qcps in finite @xmath1- and @xmath2-site cluster - based betts square lattices with periodic boundary conditions . in our previous work ( refs .
@xcite ) , we have discussed ps instabilities in selected cluster geometries . a collection of clusters " can be treated at a fixed average number of electrons @xmath18 in a canonical ensemble or a fixed chemical potential @xmath19 in a grand canonical ensemble . in the canonical ensemble
we define a charge gap @xmath20 at a given @xmath4 and temperature @xmath21 as @xmath22 , where @xmath23 and @xmath24 energies are the first derivatives of the many - body ( average ) energy @xmath25 for adding or subtracting one electron in the charge sector for the @xmath26-electron state .
the canonical charge gap also describes electron fluctuations between different many - body cluster configurations , @xmath27 with different electron densities , @xmath28 .
the charge fluctuations closely resemble resonant valence fluctuations @xcite . under certain conditions
the charge gap can be associated with either an excitation ( @xmath29 ) or a binding ( @xmath30 ) energy .
physically , @xmath29 manifests a stable @xmath31 state , while @xmath30 implies a discontinuous spontaneous ps instability , _ i.e. _ , first order transition ( a negative compressibility ) with the generation of @xmath32 and @xmath33 states of an inhomogeneous ( mixture ) of hole - rich and electron - rich phases coexisting with different @xmath11 . here , we focus on one hole off half filling , @xmath34 . at the same time condition
@xmath30 implies delocalization of holes ( electrons ) and pairing , which strongly suggests a possible local ( correlation ) coexistence of pairing with phase separation ( segregation ) rather than their competition . in the negative charge gap region ( @xmath30 ) , the spin pseudogap is nonnegative , _
i.e. _ , @xmath35 , since in the grand canonical ensemble we define it as the separation between the two consecutive peak positions of magnetic susceptibility @xmath36 for a given @xmath19 @xcite . however , when @xmath29 ( @xmath37 ) we define the canonical spin gap @xmath38 as an energy difference between the ground state with spin @xmath39 and the lowest excited state with spin @xmath40 in the spin sector , @xmath41 .
when the excited state has a higher ( lower ) total spin , the spin gap is positive ( negative ) .
in addition , we introduced the grand canonical charge pseudogap @xmath42 as a peak - to - peak distance between two consecutive peak positions of charge susceptibility @xcite , @xmath43 . a key element for the understanding of various electron instabilities is the exact relationship between the charge gap and pseudogap @xmath44 , @xmath45 , and their corresponding spin counterparts @xmath46 , @xmath38 .
the vanishing gaps and pseudogaps ( nodes ) are also crucial . as temperature approaches zero ( @xmath47 ) ,
the possible sign change in canonical gaps signifies the existence of qcps related to first order ( dramatic ) changes , while zeros in the grand canonical pseudogaps , @xmath45 , @xmath46 , describe qcps for continuous qpts into quantum liquids . at finite temperatures , the vanishing gaps and pseudogaps
define the classical critical temperatures ( cct ) for pair condensations @xmath48 , @xmath49 and crossovers @xmath3 , @xmath50 into classical charge and spin liquids respectively . and spin @xmath38 excitation gaps in the 8-site betts lattice for @xmath51 as a function of @xmath4 at @xmath52 and @xmath47 .
the charge and spin gap nodes , @xmath53 , define a number of qcps , @xmath54 , @xmath55 , and @xmath56 for various ps instabilities and complex phases .
phase i , @xmath57 : a ground state describes coherent pairing ( @xmath58 ) of electrons ( holes ) with anti - parallel spins and gapless charge excitations , @xmath59 .
phase ii for all @xmath37 : a stable charge ordered mh insulator , @xmath60 , coexists with a spin liquid , @xmath61 . at @xmath62 : a spin liquid phase with gapless , low spin-@xmath63 excitations . at @xmath64 : a spin liquid phase with gapless , high spin-@xmath65 excitations coexisting with unsaturated ferromagnet . at @xmath66 : a fully saturated spin-@xmath67 ferromagnet .
inset shows an enlarged area in the vicinity of key qcps found for elementary square geometry in ref .
] these pairing scenarios combined together give rise to the fascinating physics of bose - einstein condensation ( bec ) and ferromagnetism due to competing high / low charge and high / low spin densities .
the bound electron ( hole ) pairs with coupled anti - parallel ( singlet ) spins are complied with the bose - einstein statistics .
[ fig2 ] shows variations of gap order parameters @xmath44 , @xmath45 , @xmath46 , and @xmath38 with @xmath4 in the 8-site betts lattice at @xmath47 , where we can see the number of qcps and quantum phases .
the broken line at @xmath68 for a metallic gapless charge excitation ( @xmath69 ) implies a quantum ` cold ' charge liquid behavior .
a finite non - zero field is required to break the coupled anti - parallel spins , @xmath70 , which provides rigidity and stability for bound hole pairs ( @xmath30 ) .
the equal ( pseudo)gap amplitudes of paired charge liquid and coupled anti - parallel spins , @xmath71 , called coherent pairing in ref .
@xcite , are somewhat similar to conventional bcs - type pairing with a unique quasiparticle gap .
this picture of coherent pairing of charge and spin entities in real space in phase i ( fig .
[ fig2 ] ) suggests a full bec and possible superconductivity of electrons .
in contrast , holes in phase ii , localized on separate clusters due to electron - hole pairing , @xmath72 , at @xmath73 behave as a low density mott - hubbard ( mh ) excitonic insulator .
the ` cold ' charge localization at the qcp , @xmath54 , gives rise to a crystallized nonconducting ( mott ) state driven by correlation @xmath4 ( decreasing @xmath5 or pressure , @xmath74 ) .
hence , @xmath54 signifies a qcp for a continuous metal - insulator ( mi ) transition from a charge liquid , @xmath59 , into a mh insulator . notice that this qcp signals also spin - charge separation at @xmath37 @xcite .
the evolution of spin gaps @xmath38 and @xmath46 is also shown in fig . [ fig2 ] .
the isolated qcp , @xmath54 , signifies a ` cold ' quantum melting of anti - parallel spin - insulator , @xmath70 , into a degenerate spin - liquid without rigidity ( @xmath61 ) .
the quantum delocalization of spin as u increases reveals an opposite trend to charge behavior .
the nodes of the spin gap , defined by @xmath75 , are the qcps for the onset of spin ps and spin density ( nanophase ) inhomogeneities due to spontaneous redistribution of the spin liquid among the clusters in the negative gap region .
magnetic field fluctuations lead to segregation of clusters into regions rich in spin @xmath76 and spin @xmath77 , _
i.e. _ , formation of magnetic domains .
the qcps for spontaneously broken symmetry at @xmath78 and @xmath79 in fig .
[ fig2 ] correspond to phase transitions into an unsaturated ferromagnet ( @xmath56 ) and a saturated nagaoka - like ferromagnet @xcite .
phase at @xmath62 in table [ tab1 ] is a degenerate spin-@xmath63 liquid without spin rigidity ( @xmath61 ) .
the sign change in @xmath38 at @xmath64 implies the presence of a qcp between @xmath55 and @xmath56 .
the ps region @xmath80 corresponds to unsaturated ferromagnetism , _
i.e. _ , @xmath81 , while @xmath82 region is a gapless @xmath81 spin liquid ( @xmath61 ) with a finite spin excitation gap ( @xmath82 ) for @xmath83 .
as function of @xmath4 in the 10-site betts lattice for @xmath84 at @xmath52 and @xmath47 .
in contrast to the @xmath1-site lattice , optimized pairing is shifted to larger @xmath4 values . at large @xmath4
, there is a phase transition from a mh insulator into an unsaturated ferromagnetic ground state with qcp , @xmath85 .
inset displays the details in a moderate @xmath4 region , where equal ( pseudo)gaps , @xmath58 , display coherent pairing . ]
.the energy levels in the @xmath1-site cluster for various @xmath39 at @xmath86 . at @xmath87 , a spin-@xmath63 liquid ( @xmath61 )
has a spin-@xmath65 excitation gap , @xmath82 . in @xmath64 range ,
the state with @xmath80 is an unsaturated @xmath81 ferromagnet , while next region with @xmath82 for increasing @xmath83 is a gapless spin-@xmath65 liquid . at @xmath66 , a saturated ferromagnet ( @xmath80 )
behaves as a spin-@xmath67 liquid . [ cols="^,^,^,^,^ " , ] [ tab1 ] and @xmath21 in the @xmath1-site betts lattice at average @xmath88 .
the negative gap region shrinks and totally disappears at about @xmath89 .
the solid contour denotes a second order phase transition for classical critical temperature , @xmath90 , at which the gap vanishes , @xmath91 .
phase i with a shaded area is the metallic charge pairing phase for spontaneous first order ps instabilities and density inhomogeneities . phase ii without a shaded pattern is the stable mh insulator region .
the broken line for onset of coherent pair condensation of anti - parallel spins at @xmath92 is indicative of bec and superconductivity .
below @xmath93 the paired anti - parallel spin stiffness of bound charge describes coherent pairing phase , while above @xmath93 preformed pairs with unpaired spins is characteristics of incoherent phase @xcite . ] versus m in the @xmath1-site betts lattice at three @xmath4 values near @xmath54 in the grand canonical ensemble at @xmath47 .
the mh plateau near @xmath88 at @xmath94 vanishes with abrupt change at @xmath95 and @xmath1 qcps , @xmath96 . in the canonical ensemble at @xmath88
the change from @xmath97 to @xmath98 with unstable @xmath86 clusters suggests a negative compressibility for the first order ps and inhomogeneities in a spinodal region @xcite . ] and grand canonical @xmath99 , denoted by @xmath3 and @xmath50 in the @xmath1-site betts lattice at @xmath100 .
the zero - field spin susceptibility peak @xmath50 terminates at @xmath49 , while charge peak @xmath101 for closed pseudogap , @xmath59 , terminates at the qcp , @xmath102 .
the spin peaks separate the low ( @xmath103 ) and high spin ( @xmath81 ) phases above and below @xmath50 respectively . with the onset of strong spin ( singlet ) pairing below @xmath49
spin collective excitations become gapped .
the charge and spin peaks @xmath3 and @xmath50 emerge and reconcile for the fermi liquid behavior at the close vicinity of @xmath49 .
the increase of the distance between @xmath50 and @xmath3 , peaks above @xmath49 implies a spin - charge separation @xcite .
the charge pseudogap region due to the splitting of @xmath3 into two branches above @xmath49 manifests electron hole ( pairing ) order broken by temperature .
as in @xcite , the formation of incoherent charge pairs at @xmath104 results in a charge pseudogap area above @xmath49 . ]
phase @xmath105 : the region with @xmath80 describes a spin liquid ( @xmath61 ) of fully polarized @xmath106 nagaoka - like ferromagnet coexisting with the mh - like insulator , @xmath29 . in fig .
[ fig3 ] the exact gaps in the @xmath2-site betts lattice reproduce the key features of spontaneous ps instabilities .
the negative charge gap region defines a first order ps , while the positive gap region is a signature for continuous , second order transitions .
the oscillations in the betts @xmath2-site lattice at small and moderate @xmath4 are due to reduced square symmetry similar to the anisotropic @xmath107 hubbard clusters @xcite . in contrast , the @xmath1-site betts lattice and the @xmath108 hubbard cluster both have higher rotational and reflection symmetries . figs . [ fig2 ] and [ fig3 ]
are consistent with our phase diagrams in @xmath108 and @xmath109 geometries @xcite .
next we show that the effect of qcps is also tangible over a wide range of temperatures .
contour isolines along which the charge gap attains a constant value , are illustrated in fig .
[ fig4 ] for @xmath52 .
the contour map defines the charge gap isolines as a function of @xmath21 and @xmath4 . as temperature increases charge gap changes sign and favors electron - hole pairing above the critical temperature @xmath48 defined by @xmath91 .
the vanishing gap describes a smooth ( second order ) transition for the onset of charge pairing below @xmath48 @xcite .
phase at @xmath110 is the mh insulator ( @xmath29 ) , while at low temperature charge paired ( @xmath30 ) phase is a metallic charge liquid , with @xmath111 ( fig .
[ fig6 ] ) .
the continuous correlation driven mi transition , broken by electron hole pairing at @xmath112 , is also somewhat different from the conventional @xmath113 phase boundary between the ordered ( quantum ) phase at low and disordered ( normal ) phase at high temperatures , described by an order parameter , which is non - zero in the ordered phase and zero in the disordered phase .
the charge gap shows pairing below @xmath48 and proximity to electron - hole pairing @xmath29and mh pseudogap behavior ( @xmath114 ) above @xmath48 .
however , unlike in the bcs theory , as @xmath21 increases @xmath44 differs from spin gap @xmath46 , consistent with the existence in fig .
[ fig4 ] of two energy ( gap order parameters or ) scales and two distinct consecutive condensation temperatures for paired charges , @xmath48 , and bound anti - parallel spins , @xmath49 @xcite .
the spin rigidity for bound anti - parallel spins ( @xmath70 ) below @xmath49 manifests the onset of coherent pairing and possible superconductivity @xcite . in fig .
[ fig4 ] the incoherent pairing of bound charge @xmath115 without rigidity @xmath116 exists at @xmath117 .
the grand canonical result for the ( dramatic ) abrupt doping dependent behavior of @xmath18 at @xmath118 in fig .
[ fig5 ] provides strong evidence for the existence of a qcp at equilibrium in @xmath19 space associated with a spontaneous first order ps .
the first order transitions at @xmath95 and @xmath1 at @xmath47 display the qcp , @xmath102 , at optimal doping .
the visible plateau near @xmath88 in fig .
[ fig5 ] for @xmath29 at @xmath94 describes the rigidity of the mh insulating charge gap @xcite . the canonical negative gap at @xmath119 displays also a ps spinodal [ @xmath120 region , @xmath121 for coexisting inhomogeneous n?6 and 8 clusters with electron redistribution @xcite . the effect of quantum criticality can be felt also at low t without even ever reaching the ground state .
[ fig6 ] shows the susceptibility peak crossover @xmath101 and @xmath122 temperatures versus @xmath19 for gapless charge and spin excitations in the grand canonical ensemble .
the peak - to - peak separation for the two splitting branches of @xmath3 beneath @xmath50 in the underdoped regime at @xmath123 defines the so - called pseudogap above @xmath49 .
the distinct pseudogap phase , fundamentally different from pairing , is a precursor to the change of the charge gap sign at low temperature , @xmath124 @xcite .
we find that the pseudogap predominantly originates from @xmath88 ( plateau ) phase at @xmath62 , which is stabilized as @xmath21 , @xmath19 or @xmath4 increases or @xmath74 decreases @xcite . in many real materials ,
the contribution of the @xmath125 term ( frustration ) can be important for breaking the particle - hole symmetry in hole / electron doped systems . as a function of @xmath4 and @xmath125 in the 8-site betts lattice .
as @xmath126 increases , the crossover point shifts to a larger @xmath4 value ( the inset displays the charge gap at @xmath127 in the enlarged area of @xmath4 ) . as @xmath125 approaches @xmath5 or @xmath128 , ps instability region increases or shrinks respectively , by avoiding for @xmath129 of level crossings ( qcps ) at finite @xmath4 . ]
thus , it is necessary to consider the more realistic case , @xmath17 .
the charge gap contour plots for different @xmath125 and @xmath4 values at @xmath119 are shown in fig .
[ fig7 ] by tuning @xmath4 and the next nearest coupling , @xmath125 .
the condition @xmath130 defines the boundary @xmath131 between positive and negative gaps , which separates smooth transitions from first order phase transitions .
the @xmath132 has a detrimental effect , while @xmath126 enhances pairing .
the ps region shown in inset in fig .
[ fig7 ] at @xmath133 manifests a coherent pairing instability for all @xmath4 . as for the tetrahedron @xcite , the vanishing ps region at @xmath134 ,
is signaled by stable @xmath135 mh configurations with @xmath106 ferromagnetism .
the betts lattices , generated by finite clusters , provide strong support for ps instabilities found in generic @xmath108 and @xmath109 clusters @xcite .
many complex critical phenomena , apparent in approximate treatments of concentrated " inhomogeneous systems without long - range order are naturally reproduced in the grand canonical and canonical ensembles of independent clusters , which do interact thermally .
the exact calculations in optimized clusters are a powerful tool for unveiling hidden generic qcps of phase separation , ` hot ' and ` cold ' localization and melting of charge and spin entities .
the theory yields strong evidence for the existence of a qcp at optimal doping , @xmath102 , responsible for a first order ps transition from a sc metal into a mh insulator hidden beneath the sc dome in the htscs @xcite .
the negative canonical gaps are sufficient for spontaneous ps and accompanied inhomogeneous redistribution of the electron charge or spin at energy level crossings .
small variations in model parameters or intrinsic symmetries of the unit block can lead to dramatic qualitative changes in the vicinity of qcps , which control the physics of phase separation over a significant portion of the phase diagram .
the key intrinsic qcps , extracted from exact calculations , decipher the mystery of quantum and classical critical behavior seen in transition metal oxides , cuprates , pnictides and cmrs nanomaterials under variation of external parameters and provide a basic scenario for nagaoka - type ferromagnetism and correlation of electron pairing with phase separation @xcite . building blocks spontaneously organized into ordered structures suggest a bottom - up key paradigm to control qcps and fabricate of novel assembled nanostructures with new superconducting and magnetic properties @xcite .
the authors acknowledge the computing facilities provided by the center for functional nanomaterials , brookhaven national laboratory , which is supported by the u.s .
department of energy , office of basic energy sciences , under contract no.de-ac02-98ch10886 .
this work was performed also , in part , at the center for integrated nanotechnologies , a u.s .
department of energy , office of basic energy sciences user facility at los alamos national laboratory ( contract de - ac52 - 06na25396 ) and sandia national laboratories ( contract de - ac04 - 94al85000 ) .
anderson , science * 288 * ( 2000 ) 480 .
y. kohsaka , _ et al .
_ , science * 315 * ( 2007 ) 1380 .
j. lee , _ et al .
_ , science * 325 * ( 2009 ) 1099 .
s. sachdev , science * 288 * ( 2000 ) 475 .
broun , nature physics * 4 * ( 2008 ) 170 .
kocharian , _ et al .
_ , scanning probe microscopy in nanoscience and nanotechnology , springer , 2010 , pp . 507 - 570 .
z. fisk , science * 325 * ( 2009 ) 1348 .
kocharian , _ et al .
_ , physical review b * 74 * ( 2006 ) 024511 ; + a.n .
kocharian , _ et al .
_ , journal of magnetism and magnetic materials * 300 * ( 2006 ) e585 .
s. sachdev , physica status solidi b * 247 * ( 2010 ) 537 .
kocharian , _ et al .
_ , physics letters a * 373 * ( 2009 ) 1074 ; + a.n .
kocharian , _ et al .
_ , physics letters a * 364 * ( 2007 ) 57 .
kocharian , _ et al .
_ , physical review b * 78 * ( 2008 ) 075431 .
fernando , _ et al .
_ , physical review b * 75 * ( 2007 ) 085109 ; + k. palandage , _ et al .
_ , journal of computer - aided materials design * 14 * ( 2007 ) 103 .
fernando , _ et al .
_ , physical review b * 80 * ( 2009 ) 014525 . j. oitmaa , _ et al . _ , canadian journal of physics * 56 * ( 1978 ) 897 . d.d .
et al . _ , canadian journal of physics * 74 * ( 1996 ) 54 ; + d.d .
et al . _ , canadian journal of physics * 77 * ( 1999 ) 353 .
e. kaxiras , _ et al .
_ , physical review b * 37 * ( 1988 ) 656 .
kocharian , _ et al .
_ , soviet physics jetp * 44 * ( 1976 ) 404
. d. van der marel , nature physics * 7 * ( 2011 ) 10 .
sebastian , _
et al . _ , proceedings of the national academy of sciences of the united states of america * 107 * ( 2010 ) 6175 .
uemura , _ et al .
_ , nature physics * 3 * ( 2007 ) 29 .
e. dagotto , science * 309 * ( 2005 ) 257 .
tranquada , _ et al .
_ , physical review letters * 78 * ( 1997 ) 338 .
k. fang , _ et al .
_ , physics letters a * 376 * ( 2012 ) 538 .
sohn , _ et al .
_ , nano letters * 9 * ( 2009 ) 3392 . | spontaneous phase separation instabilities with the formation of various types of charge and spin pairing ( pseudo)gaps in @xmath0 hubbard model including the _ next nearest neighbor coupling _ are calculated with the emphasis on the two - dimensional ( square ) lattices generated by @xmath1- and @xmath2-site betts unit cells . the exact theory yields insights into the nature of quantum critical points , continuous transitions , dramatic phase separation instabilities and electron condensation in spatially inhomogeneous systems .
the picture of coupled anti - parallel ( singlet ) spins and paired charged holes suggests full bose condensation and coherent pairing in real space at zero temperature of electrons complied with the bose - einstein statistics .
separate pairing of charge and spin degrees at distinct condensation temperatures offers a new route to superconductivity different from the bcs scenario .
the conditions for spin liquid behavior coexisting with unsaturated and saturated nagaoka ferromagnetism due to spin - charge separation are established .
the phase separation critical points and classical criticality found at zero and finite temperatures resemble a number of inhomogeneous , coherent and incoherent nanoscale phases seen near optimally doped high-@xmath3 cuprates , pnictides and cmr nanomaterials .
= 1.7 cm |
the following situation frequently occurs in geometric dynamics .
there is a phase space @xmath0 , a transformation @xmath1 ; and there is a finite decomposition @xmath2 . let @xmath3 be the corresponding alphabet .
a phase point @xmath4 is regular if every element of the orbit @xmath5 belongs to a unique atom of @xmath6 .
suppose that @xmath7 , etc .
the corresponding word @xmath8 is the _ code _ of @xmath9 .
let @xmath10 be the set of words in @xmath11 of length @xmath12 obtained by coding points in @xmath0 .
the positive function @xmath13 is the associated _
complexity_. its behavior as @xmath14 can be bounded , or grow polynomially , or grow exponentially , etc . ]
is an important characteristic of the dynamical system in question .
the following examples have motivated our study .
* example a*. let @xmath15 be a polygon with sides @xmath16 , and let @xmath0 be the phase space of the _ billiard map _
@xmath17 in @xmath18 .
the coding generated by the corresponding decomposition @xmath2 is the traditional coding of billiard orbits by the sides they hit @xcite .
basic questions about its complexity are open @xcite .
* example b*. let @xmath19 be a convex polygon with vertices @xmath16 .
the complement @xmath20 is the phase space of the _ outer billiard _
@xmath21 about @xmath18 .
( it is also called the _ dual billiard_. see @xcite ) .
the conical regions bounded the singular lines of @xmath21 form the natural decomposition @xmath2 . in @xmath22 the mapping @xmath21 is the symmetry about @xmath16 .
the decomposition @xmath6 yields the coding of outer billiard orbits by the vertices they hit .
we will study the complexity of ( @xmath23-dimensional ) _ piecewise convex transformations_. this is a wide class of geometric dynamical systems ; it contains the examples above .
our setting is as follows .
( for simplicity of exposition , we restrict our attention to two dimensions . )
let @xmath0 be a _
geodesic surface _ , and let @xmath24 be a finite _ geodesic graph_. a subset @xmath25 is _ convex _ if for any @xmath26 there is a unique geodesic in @xmath27 with endpoints @xmath26 .
suppose that the closed faces @xmath28 of @xmath29 are convex , and let @xmath2 be the corresponding decomposition .
we say that @xmath6 is a _ convex partition _ of @xmath0 .
a differentiable mapping @xmath30 is convex if it sends geodesics to geodesics .
suppose that @xmath1 is a convex diffeomorphism on @xmath28 , and that @xmath31 also form a convex partition of @xmath0 .
we say that the triple @xmath32 is a _ piecewise convex transformation _ with the _ defining partition _
@xmath6 .
making further assumptions on @xmath32 , we obtain more specialized classes of transformations , e.g. , piecewise isometries , piecewise affine mappings , etc .
the case when the faces of @xmath29 are convex euclidean polyhedra , and @xmath33 is isometric on them arises in task scheduling problems @xcite .
in section [ basic ] we develop geometric and combinatorial techniques to study the complexity of piecewise convex transformations .
in the rest of the paper we apply these results to the inner and outer polygonal billiards on ( simply connected ) surfaces of constant curvature @xmath34 .
one of the goals of this work is to develop a uniform approach to these dynamical systems . while there is a vast literature on the parabolic case ( @xmath35 ) , the elliptic ( @xmath36 ) and the hyperbolic ( @xmath37 ) cases
have been studied only sporadically .
let @xmath38 be a surface of constant curvature , and let @xmath39 be a polygon .
in section [ innerbill ] we cast the ( inner ) billiard map in @xmath18 as a piecewise convex transformation @xmath32 .
see theorem [ bil_piece_prop ] .
we do it simultaneously for all curvatures , and without making additional assumptions on @xmath18 .
thus , we do nt assume that @xmath18 is convex or simple , and have to pay a price for this .
the partition @xmath6 is finer than the natural " one ; the coding it generates is more refined than the standard coding by sides @xcite .
we develop a dictionary between the language of piecewise convex transformations and that of billiard orbits . in the remaining part of the paper
we come back to examples a , b and study the complexity of natural coding . in section [ cases ]
we investigate the complexity of inner billiard orbits in a polygon @xmath18 on a surface of constant curvature @xmath40 . in section [ arbit_sub ] @xmath41 are arbitrary ; later on we specialize to convex @xmath18 but arbitrary @xmath40 , to @xmath35 , and to @xmath42 respectively .
below we formulate the main results .
the side complexity of billiard orbits in any rational euclidean polygon grows at most cubically ; see theorem [ rat_complex_thm ] .
the side complexity of billiard orbits in any spherical polygon grows subexponentially ; see theorem [ spheresubexp ] .
the side complexity of billiard orbits in any hyperbolic ( i.e. , @xmath37 ) polygon grows exponentially ; the exponent in question is the topological entropy of the billiard map ; see theorem [ hyperb_exp ] .
section [ outerbill ] is the outer billiard counterpart of section [ cases ] . here
are some of its results . for @xmath35 and arbitrary polygon ( resp .
rational polygon ) we obtain polynomial bounds from above and below ( resp .
quadratic asymptotics ) for the complexity ; see theorem [ upperbound ] and theorem [ rational ] respectively .
for @xmath36 complexity grows subexponentially ; see theorem [ spheresubexp1 ] . for @xmath37 and arbitrary polygon
, we obtain linear lower bound for complexity , which is sharp : for the so - called large polygons complexity grows linearly ; see theorem [ hyplowbd ] . * notes and references .
* we are grateful to the research in pairs program in oberwolfach and to the shapiro visiting program at penn state for their support .
the second author was partially supported by an nsf grant .
the work @xcite is the predecessor of our section 3 . for additional results on polygonal billiards see @xcite .
we will introduce a class of maps of _ geodesic spaces_. in order to simplify our exposition , we will restrict it to two dimensions , i.e. , to _
geodesic surfaces_. let @xmath38 be one .
then @xmath38 is a topological surface .
it may have a boundary , @xmath43 , and a finite number of _ cone points_. the surface @xmath38 is endowed with a collection of _ geodesics _
, satisfying the standard properties .
in particular , @xmath43 is a finite union of geodesics .
let @xmath44 be any pair of points .
then @xmath38 is _ geodesically convex _ if there is a _
unique shortest geodesic _ @xmath45 joining them . below . ]
* examples . *
i ) the set @xmath46 endowed with a projectively flat ( e.g. , minkowski ) metric is a geodesic surface .
its geodesics are straight lines .
let @xmath47 be a closed , bounded convex region .
then it is a geodesic surface iff @xmath38 is a polygon .
the geodesic @xmath48 is the segment with endpoints @xmath49 .
+ ii ) let @xmath50 be the round sphere , and let @xmath51 be a convex spherical polygon .
then @xmath38 is a geodesic surface iff it does not contain antipodal points .
this holds iff @xmath38 is contained in an open hemisphere . in what follows
we will often say convex geodesic surface " instead of geodesically convex geodesic surface " .
let @xmath0 be a geodesic surface .
a finite graph @xmath29 drawn on @xmath0 is a _ geodesic graph _ if the edges of @xmath29 are geodesics .
we will assume that the cone points of @xmath0 are contained in the set of vertices of @xmath29 , and that the latter are nondegenerate .
( a vertex is degenerate if it has two adjacent edges , and they are colinear . ) let @xmath22 be the ( open ) faces of @xmath29 , and let @xmath52 be their closures .
we will say that @xmath29 is a _
convex geodesic graph _ if the surfaces @xmath52 are geodesically convex
. we will refer to the data @xmath53 as a _ piecewise ( geodesically ) convex geodesic surface_. we will also say that @xmath54 is a _ convex geodesic partition _ of the geodesic surface @xmath0 .
the boundary @xmath55 is spanned by the edges and vertices of @xmath29
. we will also refer to it as the _ support _ of @xmath29 and denote by @xmath56 .
there is a @xmath57 correspondence between convex geodesic graphs drawn on @xmath0 and convex geodesic partitions of @xmath0 .
the latter concept provides an alternative approach to the material below @xcite .
however , the language of convex geodesic graphs is more suitable for our purposes , and we will use it in what follows . let @xmath0 be a convex geodesic surface , and let @xmath27 be an arbitrary geodesic surface . a diffeomorphism ( not necessarily surjective )
@xmath58 is a _ geodesically convex transformation _ if it sends geodesics to geodesics .
then @xmath59 is a convex geodesic surface as well , and the inverse map @xmath60 is a geodesically convex transformation
. we will now introduce a class of dynamical systems which , on one hand , is sufficiently general to include several interesting examples , and on the other , is special enough to allow a common geometric framework .
let @xmath61 be as above .
let @xmath62 be the number of atoms of @xmath6 and write @xmath63 .
suppose that for @xmath64 there is a convex transformation @xmath65 .
set @xmath66 .
suppose that @xmath67 and that the open sets @xmath68 are pairwise disjoint .
then @xmath69 is a convex geodesic partition , and @xmath70 is the support of a convex geodesic graph , @xmath71 , drawn on @xmath0 . these data determine a _ piecewise convex transformation _
@xmath1 with the _ defining partition _ @xmath6 , and we will use the notation @xmath72 for it .
the inverse of @xmath72 is also a piecewise convex transformation .
its defining partition is @xmath73 .
thus in our notation , @xmath74 .
invertible piecewise isometric ( as well as affine , or projective ) maps are examples of piecewise convex transformations .
let @xmath38 be a simply connected surface of constant curvature , and let @xmath39 be a geodesic polygon . in sections [ cases ] and [ outerbill ] we will put the inner and the outer billiard about @xmath18 into the framework of piecewise convex transformations .
denote by @xmath75 the _ full shift space _ on the _ alphabet _ @xmath76 of the faces of @xmath29 .
a point @xmath4 is _ regular _ if @xmath77 belong to open faces of @xmath29 .
let @xmath78 be the set of regular points .
assigning to @xmath79 the sequence of faces of @xmath29 containing the consecutive elements @xmath77 we obtain the _ coding map _
set @xmath81 , and let @xmath10 be the set of words of length @xmath12 that occur in @xmath82 .
the function @xmath13 is the _ complexity _ of @xmath32 _ with respect to the defining partition_. let @xmath83 be two geodesic graphs on @xmath0 .
their _ join _
@xmath84 is the ( unique ) geodesic graph such that @xmath85 we write @xmath86 if @xmath87 .
recall that @xmath88 denotes the join of partitions , and that @xmath89 holds if @xmath90 .
let @xmath91 . if @xmath83 are convex geodesic graphs , then so is @xmath84 , and @xmath92 moreover , @xmath86 iff @xmath89 .
set @xmath93 .
the convex geodesic graphs @xmath94 form an increasing tower with respect to the relation @xmath95 , and we set @xmath96 , @xmath97 . the _ singular set _
@xmath98 is a countable ( at most ) union of geodesics .
note that @xmath99 is a piecewise convex transformation for @xmath100 .
let @xmath101 be a _
finite orbit _ such that the points @xmath102 belong to open faces of @xmath29 .
we say that @xmath102 is a _ regular orbit of length @xmath12 _ and denote by @xmath103 the corresponding word on the alphabet @xmath11 ; it is the _ code of @xmath102_. then @xmath10 is the set of codes of regular orbits of length @xmath12 ; the orbit @xmath101 is regular iff @xmath9 belongs to an open face of the graph @xmath104 .
the proposition below summarizes the discussion .
[ sing_strat_prop ] let @xmath72 be a piecewise convex transformation , and let @xmath105 .
then : + 1 . there is a sequence @xmath106 of convex geodesic graphs in @xmath0 such that @xmath107 2 .
set @xmath108 .
then @xmath109 is a convex geodesic partition , and the @xmath12th iteration of @xmath32 is a piecewise convex transformation with the defining partition @xmath109 , i. e. , @xmath110 .
there is a natural bijection between the set @xmath10 and the set of atoms of the partition @xmath109 .
thus , the complexity @xmath111 is the number of faces of the graph @xmath104 .
we will need a general proposition that concerns the combinatorics of graphs .
let @xmath0 be a piecewise geodesic surface .
let @xmath83 be geodesic graphs drawn on @xmath0 and set @xmath112 .
( we do not assume that the faces of @xmath83 are convex . )
denote by @xmath113 the respective sets of faces , edges and vertices .
let @xmath114 be the sets of faces , edges and vertices of @xmath29 . if @xmath115 intersect non - transversally , then they ( partially ) overlap .
there are @xmath116 ways in which this can happen
. see figure [ fig1 ]
. denote by @xmath117 the number of the overlapping pairs of edges .
[ comb - top_lem ] let @xmath83 be geodesic graphs drawn on a piecewise geodesic surface @xmath0 .
denote by @xmath40 the euler characteristic of @xmath0 .
let @xmath118 be the sets of vertices of @xmath83 that are disjoint from the other graph , and set @xmath119 .
then @xmath120 drawn on @xmath0 satisfies @xmath121 applying ( [ euler_eq ] ) to graphs
@xmath122 , we obtain for the left hand side of ( [ comb - top_eq ] ) @xmath123 the edges of @xmath83 respectively .
let @xmath124 be the number of vertices of @xmath125 that are located in the interior of @xmath126 respectively .
let @xmath127 be the number of times that the interior of @xmath128 transversally intersects the interior of an edge of @xmath125 respectively .
then @xmath126 contribute @xmath129 edges to @xmath130 respectively .
taking the overlapping into account , we obtain @xmath131 let @xmath132 be the set of common vertices of @xmath83 .
let @xmath133 be the sets of vertices of @xmath83 that are in the interior of edges of the other graph . then @xmath134 let @xmath135 be the set of new " vertices of @xmath29 .
then @xmath136 we also have @xmath137),([vertices_eq1]),([vertices_eq2 ] ) and ( [ vertices_eq3 ] ) , we obtain @xmath138 ) , @xmath139),([vertices_eq5 ] ) into ( [ euler_eq1 ] ) , and using ( [ vertices_eq4 ] ) again , we obtain the claim .
the expression for @xmath140 in lemma [ comb - top_lem ] involves information about vertices of @xmath83 which is sometimes not available .
the following corollary of lemma [ comb - top_lem ] is useful .
[ comb_bounds_prop ] let @xmath83 be geodesic graphs drawn on a piecewise geodesic surface @xmath0 of euler characteristic @xmath40 .
let @xmath112 , and let the notation be as above .
then @xmath141 and equation ( [ vertices_eq4 ] ) , the quantity to estimate in ( [ ineq_eq ] ) is equal to @xmath142 .
the following special cases of lemma [ comb - top_lem ] and proposition [ comb_bounds_prop ] will be especially useful .
[ comb_bounds_cor ] let @xmath0 be a geodesic surface homeomorphic to the open disc .
let @xmath83 be geodesic graphs drawn on @xmath0 such that their edges intersect only transversally .
let @xmath112 and let the sets @xmath143 be as above
. then @xmath144 then @xmath145 and @xmath36 .
hence lemma [ comb - top_lem ] and proposition [ comb_bounds_prop ] yield ( [ transv_eq ] ) and ( [ transv_ineq ] ) respectively . if @xmath146 then @xmath147 and ( [ transv_eq ] ) becomes ( [ transv_eq1 ] ) .
let @xmath82 be a language on a finite alphabet @xmath11 , and let @xmath10 be the set of words of length @xmath12 in @xmath82 .
the _ complexity _ of @xmath82 is the function @xmath13 .
set @xmath148 and @xmath149 .
we will refer to the functions @xmath150 as the _ first , second differences of complexity _ , respectively .
our approach is based on cassaigne s formula for @xmath151 @xcite .
denote by @xmath152 the number of left , right one - letter extensions of the word @xmath153 respectively .
we will assume , following @xcite , that @xmath154 for any @xmath155 .
a word is _ bispecial _
if @xmath156 .
let @xmath157 be the number of extensions of the type @xmath158 where @xmath159 .
let @xmath160 be the set of bispecial words , and set @xmath161 .
the _ cassaigne index _ is defined by @xmath162 note that if @xmath153 is not bispecial then @xmath163 .
the cassaigne formula says @xmath164 we define the cumulative index by @xmath165 for @xmath100 , and @xmath166
. set @xmath167 [ complviacass ] the complexity of a piecewise convex transformation satisfies @xmath168 * proof . *
denote by @xmath169 the right hand side of ( [ count_eq1 ] ) . since , by ( [ casform_eq],[casind_eq ] ) , the second differences of @xmath170 are equal , @xmath171 is linear in @xmath12 .
but @xmath172 .
let @xmath32 be a piecewise convex transformation , and let @xmath82 be its coding language .
for @xmath173 let @xmath174 be the corresponding open face of the graph @xmath104 .
we denote by @xmath175 the restriction of @xmath176 to @xmath177 .
let @xmath178 ( resp .
@xmath179 ) be the set of essential vertices ( resp .
pairs of overlappings edges ) for @xmath175 .
[ indexint ] for any @xmath155 we have @xmath180 * proof . *
denote by @xmath83 the restrictions of @xmath181 to @xmath177 respectively .
then in the notation of lemma [ comb - top_lem ] , @xmath182 .
since @xmath177 is contractible , @xmath36 in equation ( [ comb - top_eq ] ) . thus , the left hand side of equation ( [ comb - top_eq ] ) is @xmath183 .
but its right hand side is @xmath184 .
we will use the notation @xmath185 . for @xmath100 set @xmath186 and @xmath187 thus , @xmath188 ( resp .
@xmath189 ) is the number of new " essential vertices ( resp .
new " edge overlappings ) of the graph @xmath176 , while @xmath190 ( resp .
@xmath191 ) is the total number of essential vertices ( resp .
edge overlappings ) of @xmath176 .
note that only bispecial words contribute to these numbers .
[ geomcassform ] let @xmath32 be a piecewise convex transformation and let @xmath192 be the corresponding sequence of convex partitions .
then the complexity of @xmath32 satisfies @xmath193 * proof .
* we have @xmath194 . by lemma [ indexint ] , @xmath195 , hence @xmath196 .
the claim now follows from equation ( [ count_eq1 ] ) .
let @xmath38 be a complete simply connected surface of constant curvature , or @xmath197 or @xmath198 .
thus @xmath38 is either the euclidean plane or the hyperbolic plane or the unit sphere .
we will refer to these geometries as parabolic , hyperbolic and elliptic respectively . ] and let @xmath39 be a geodesic polygon . in order to cast the billiard map in @xmath18 as a piecewise convex transformation simultaneously for all three cases at hand
, we will use the models of the hyperbolic and elliptic geometries where the geodesics are straight lines in the euclidean plane .
( we will refer to them as the _ projective models_. ) for @xmath37 , this is the klein - beltrami model of hyperbolic geometry .
the hyperbolic plane is represented by the open unit disc , the unit circle @xmath199 is the circle at infinity " , and geodesics are the chords of this disc .
let @xmath49 be distinct points in the disc , and let @xmath200 be the intersection points of the line @xmath201 with @xmath199 .
we identify the oriented line @xmath201 with @xmath202 , and hence @xmath203 with real numbers .
the cross - ratio of the four points is given by @xmath204={{(y - a)(b - x)}\over{(x - a)(b - y)}}.\ ] ] the distance between points satisfies @xmath205.$ ] in this model , isometries of the hyperbolic geometry are the projective transformations of the euclidean plane , preserving the unit disc .
projective model of the elliptic geometry ( i.e. , @xmath206 ) is as follows .
we restrict our attention to an open hemisphere ; consider for concreteness the northern hemisphere .
its central projection to the tangent plane at @xmath207 is a surjective diffeomorphism that takes spherical geodesics to euclidean straight lines . in order to endow the _ billiard map phase space _ @xmath208 with the structure of a geodesic surface in a uniform fashion , we will impose a slight restriction on @xmath18 .
let @xmath39 be a geodesic polygon , and let @xmath209 be a side of @xmath18 .
we denote by @xmath210 the geodesic reflection in @xmath38 about @xmath209 , and set @xmath211 .
[ admis_def ] let @xmath39 be a spherical polygon .
then @xmath18 is _ admissible _ if for any side @xmath209 the polygon @xmath212 is contained in an open hemisphere .
note that the admissibility is a restriction only in the elliptic case .
unless we state otherwise , we will consider only admissible polygons , suppressing the qualifier .
the space @xmath208 consists of directed geodesic segments inside @xmath18 both of whose endpoints belong to @xmath213 . since we are using projective models , @xmath15 is a euclidean polygon in any of the three cases at hand .
geodesic segments in @xmath18 are straight ; we refer to their endpoints as the _ beginning _ @xmath214 and the _ end _
the mapping @xmath216 is an embedding @xmath217 ; it induces a topology on @xmath0 .
let @xmath218 be the space of oriented straight lines ( rays ) in the euclidean plane .
endowed with the natural topology , it is a cylinder . besides
, @xmath218 is a geodesic surface : geodesics in @xmath218 are the pencils of rays passing through a point or the pencils of parallel rays .
equivalently , we define this structure via the canonical embedding @xmath219 .
lines in @xmath46 become points of the dual projective plane @xmath220 .
the geodesic surface structure of @xmath218 is thus induced by that of the real projective plane .
denote by @xmath221 the ray containing the chord @xmath9 .
the mapping @xmath222 given by @xmath223 is finite - to - one ; if @xmath18 is convex then it is one - to - one .
it induces the structure of a geodesic surface on @xmath0 .
let @xmath224 ( resp .
@xmath225 ) be given by the condition that @xmath215 ( resp .
@xmath214 ) is a corner of @xmath18 .
let @xmath226 be the set of chords that contain a corner of @xmath18 in the interior .
set @xmath227 by definition , @xmath228 is a geodesic graph in @xmath0 .
the following lemma is crucial .
[ convex_lem ] the faces of the graph @xmath228 are geodesically convex .
* proof*. let @xmath229 , and let @xmath230,$ ] be a path in @xmath231 connecting them .
then the points @xmath232 $ ] ( resp . @xmath233 $ ] ) belong to the interior of a side , @xmath234 ( resp .
@xmath235 ) . up to relabeling and orientation reversal ,
the points in question may form four apriori possible configurations ; the four cases are shown in figure [ figconv1 ] .
cases @xmath236 contradict to our assumptions . indeed , in case @xmath116 the sides @xmath200 intersect at their interior points . in case @xmath237
let @xmath238 be the endpoints of the side @xmath239 .
the two frames @xmath240 and @xmath241 have opposite orientations , thus for some @xmath242 the vectors @xmath243 are colinear , in contradiction to the assumption that @xmath244 . in cases
@xmath245 let @xmath246 be the intersection point of the straight lines @xmath247 .
the two lines may be parallel , but the same argument works .
] let @xmath248 be the curve in @xmath0 obtained by rotating the oriented line @xmath249 about @xmath250 towards @xmath251 by an angle less than @xmath252 .
it is a geodesic . to show that @xmath248 does not intersect @xmath228 , it suffices to prove that the quadrilateral @xmath253 contains no vertices of @xmath18 .
assume that it does , and let @xmath254 be such a vertex .
the frames @xmath255 and @xmath256 have opposite orientations .
hence , for some @xmath257 the vectors @xmath258 are colinear , and thus the segment @xmath259 contains a corner of @xmath18 , in contradiction to the assumptions .
we denote by @xmath260 the billiard map and by @xmath261 its inverse .
they are not defined on all of @xmath0 , and there are choices in defining them on @xmath228 . to avoid the ambiguities , we will now precisely define @xmath262 on their respective domains of definition .
let @xmath263 $ ] be a phase point .
then @xmath264 ( resp .
@xmath265 ) is not defined iff @xmath215 ( resp .
@xmath214 ) is a corner of @xmath18 .
thus , the natural domain of definition for @xmath17 ( resp .
@xmath266 ) is @xmath267 ( resp .
@xmath268 ) .
suppose now that @xmath215 is not a corner , and let @xmath209 be the unique side of @xmath18 that contains @xmath215 .
let @xmath269 be the reflection of @xmath270 about @xmath209 .
let @xmath271\subset l_1(x)$ ] be the longest chord such that @xmath272\in x$ ] .
this defines @xmath17 on @xmath267 .
the definition of @xmath266 on @xmath268 is analogous , and we leave it to the reader .
set @xmath273 [ discont_prop ] the following holds : + 1 .
the subset @xmath274 is the set of discontinuities of the billiard map .
the faces of @xmath29 are geodesically convex .
* proof*. it is elementary to see that @xmath275 belongs to the set of discontinuities .
( the billiard map is not even defined on @xmath275 . )
let @xmath276 .
then arbitrarily close to @xmath9 there are phase points @xmath277 such that @xmath278 belong to distinct , and non - adjacent , sides of @xmath18 .
hence @xmath279 and @xmath280 are not close to each other , and @xmath17 is discontinuous at @xmath9 .
. then @xmath264 is defined .
if however @xmath282 , then arguing as above , we conclude that @xmath283 and @xmath284 are not close to each other , and @xmath17 is discontinuous at @xmath9 .
see figure [ figdiscon ] . on the other hand ,
if @xmath285 , then @xmath9 is a point of continuity of @xmath17 .
this proves the first claim .
let @xmath277 belong to a face of @xmath29 .
arguing as in the proof of lemma [ convex_lem ] , we see that @xmath278 belong to the same side , @xmath209 , of @xmath18
. set @xmath286 , and let @xmath287 .
the reflection " trick associates with any phase point @xmath4 such that @xmath288 a phase point @xmath289 @xcite .
the correspondence @xmath290 maps a curve @xmath291 connecting @xmath277 in @xmath0 into a curve @xmath292 connecting @xmath293 in @xmath27 .
let @xmath294 be the corresponding set of phase points containing corners .
the condition @xmath295 implies that @xmath296 by lemma [ convex_lem ] , phase points @xmath293 are connected by a geodesic , @xmath297 , in @xmath298 .
folding @xmath297 back into @xmath0 by the geodesic reflection @xmath210 , we obtain the geodesic @xmath299 connecting @xmath277 in @xmath300 .
this establishes the second claim . we will now cast the billiard mapping as a piecewise convex transformation .
[ bil_piece_prop ] let @xmath38 be a simply connected surface of constant curvature , and let @xmath39 be a geodesic polygon .
let @xmath208 be the phase space of the billiard map endowed with the structure of a geodesic surface .
let @xmath260 be the billiard map .
let @xmath29 be the graph defined by ( [ graph_eq1 ] ) , and let @xmath301 be the corresponding convex geodesic partition .
then @xmath29 is a convex geodesic graph , and @xmath17 is continuous on the faces of @xmath29 .
let @xmath33 be the map that coincides with @xmath17 on the open faces of @xmath29 and extends to their closures by continuity .
then @xmath32 is a piecewise convex transformation ; the billiard map @xmath17 and the transformation @xmath32 coincide on @xmath300 . * proof*. most of the statements have been established in the course of the preceding discussion .
for instance , proposition [ discont_prop ] asserts that @xmath29 is a convex geodesic graph .
it remains to show that the restriction of @xmath17 to any open face of @xmath29 is a geodesically convex transformation .
the geodesic structure on @xmath0 has been defined via ray focusing .
geodesic reflections send focusing beams into focusing beams .
let @xmath25 be a face of @xmath29 .
as we have seen in the proof of proposition [ discont_prop ] , the endpoints @xmath214 of all phase points @xmath302 belong to the same side , @xmath209 , of @xmath18 .
hence , for all @xmath302 the new phase point @xmath264 is obtained via the geodesic reflection @xmath210 .
hence @xmath17 sends geodesic segments of @xmath27 into geodesic segments .
it has been traditional to study the complexity of billiard orbits in a polygon @xmath18 using the coding by sides of @xmath18 @xcite .
let @xmath32 be as in theorem [ bil_piece_prop ] .
the complexity of billiard orbits with respect to the coding by atoms of @xmath6 coincides with the complexity of @xmath32 . we will study it , and then apply our results to the traditional billiard complexity . first
, we will establish a dictionary between the language of billiard orbits @xcite and that of piecewise convex transformations .
a billiard orbit of ( combinatorial ) length @xmath12 is a sequence @xmath303 of phase points such that @xmath304 .
geometrically , @xmath248 is a sequence of @xmath12 consecutive chords of @xmath18 , where each chord is the reflection of the preceding one . in particular
, the points @xmath305 are not corners .
we will also say that @xmath248 is a _
@xmath12-segment orbit_. if @xmath306 is not a corner , then the _ @xmath198-step forward extention _ is @xmath307 , where @xmath308 . if @xmath309 is not a corner , then the _ @xmath198-step backward extention _ is @xmath310 , where @xmath311 .
we can iterate these extentions in obvious ways .
a billiard orbit is _ regular _ if it does not contain corners .
otherwise , @xmath248 is _
singular_. if @xmath248 does not contain corners of @xmath18 in its interior , but the endpoints @xmath312 are corners , we say , following @xcite , that @xmath248 is a _ generalized diagonal _ of length @xmath12 .
the notion of generalized diagonals works well for the billiard in a convex polygon .
we introduce an extension of this notion which works for arbitrary polygons .
[ sing_def ] an @xmath12-segment billiard orbit @xmath303 is _ strongly singular _ if @xmath313 do not contain corners but @xmath314 and @xmath315 do . a _ family of billiard orbits _ is a one - parameter family @xmath316 where the mapping @xmath317 is injective and continuously differentiable . in particular
, we will consider families of strongly singular billiard orbits .
any such is contained in a unique maximal family , and we will consider only them , suppressing the qualifier _
maximal_. a strongly singular billiard orbit is _ isolated _ if it is not contained in a family of such orbits .
[ transition_prop ] let @xmath39 be a geodesic polygon , and let @xmath32 be the corresponding piecewise convex transformation .
let @xmath318 . then : \1 .
there is a bijection between the set @xmath319 and the union of the sets of isolated @xmath320-segment and @xmath321-segment strongly singular billiard orbits ; \2 .
there is a bijection between the set @xmath322 and the union of the sets of families of @xmath320-segment and @xmath321-segment strongly singular billiard orbits .
* proof*. let @xmath323 . iterating the billiard map , we obtain an @xmath324-segment billiard orbit @xmath325 , where @xmath326 .
let @xmath327 . then @xmath328 iff either @xmath329 or @xmath330 but @xmath331 .
. then @xmath332 iff @xmath333 .
thus , we have obtained a surjective map from @xmath334 to the union of set of @xmath320-segment and @xmath321-segment strongly singular billiard orbits .
( the orbit corresponding to @xmath9 is @xmath335 in the former case , and @xmath336 in the latter . )
the map above is a bijection .
moreover , isolated orbits correspond to the essential vertices of @xmath337 , and families of orbits correspond to overlapping edges of @xmath338 and @xmath339 .
we will now apply the preceding material to the complexity of billiards in geodesic polygons on surfaces of constant curvature . in the beginning of this section
we consider the three cases at hand simultaneously , emphasizing their similarities .
let @xmath38 be a surface of constant curvature @xmath340 , and let @xmath39 be a geodesic polygon .
the operation of _ unfolding _ sends billiard orbits in @xmath18 into geodesics in @xmath38 @xcite , see figure [ fig105 ] .
if @xmath341 is a billiard orbit , then its unfolding is the geodesic @xmath342 where the segments @xmath343 differ by an isometry of @xmath38 .
let @xmath344 be the group generated by the geodesic reflections in the sides of @xmath18 .
denote by @xmath345 the set of elements obtained from products of at most @xmath12 reflections .
( these are the elements of length at most @xmath12 . )
then @xmath346 is an increasing tower of finite sets ; their union is @xmath347 .
[ conj_lem ] 1 .
let @xmath348 .
then all strongly singular billiard orbits in @xmath18 are isolated .
let @xmath349 .
let @xmath341 be a strongly singular billiard orbit in @xmath18 and let @xmath342 be its unfolding .
suppose that @xmath248 extends to a family of strongly singular billiard orbits .
then there are corner points @xmath350 and an isometry @xmath351 of @xmath38 such that @xmath352 and the points @xmath353 either coincide or are antipodal .
* proof*. let @xmath354 be a family of strongly singular billiard orbits . the segments @xmath355 ( resp .
@xmath356 ) pass through the same corner point @xmath357 ( resp .
@xmath358 ) .
we view @xmath359 as a beam of trajectories that emanate from the focusing point @xmath357 and refocus at @xmath358 .
the unfolding @xmath360 transforms directed billiard orbits into directed geodesics in @xmath38 , preserving the length and sending focusing beams of billiard orbits into focusing beams of geodesics@xcite . , the argument applies to arbitrary @xmath40 . ]
the beam @xmath361 emanates from @xmath362 and refocuses at @xmath363 where @xmath364 .
the distance between @xmath353 along any geodesic @xmath361 is the same as the distance between @xmath365 along the orbit @xmath359 , hence positive .
if @xmath348 , this implies @xmath366 ; thus the geodesic beam @xmath361 has two different focusing points .
this is impossible , which proves the first claim . if @xmath349 , i.e. , @xmath38 is the sphere , the focusing points in question are either antipodal or they coincide .
we denote by @xmath367 ( resp .
@xmath368 ) the number of isolated ( resp .
families of ) strongly singular @xmath369-segment billiard orbits in @xmath18 . set @xmath370 [ str_sing_prop ] let @xmath18 be a geodesic polygon on a surface @xmath38 of constant curvature @xmath40 .
let @xmath32 be the associated piecewise convex transformation , and let @xmath371 be the billiard complexity corresponding to the partition @xmath6 .
then there are positive integers @xmath372 depending on @xmath18 , so that : \1 . if @xmath36 , then @xmath373 2 . if @xmath348 , then @xmath374 * proof .
* we use proposition [ transition_prop ] to express the right hand side of ( [ complex_eq ] ) via the numbers of ( families of ) strongly singular billiard orbits .
this yields the first claim . taking into account lemma [ conj_lem ]
, we obtain the second claim .
the factor of @xmath23 in ( [ compl_polyg_eq ] , [ compl_curv_eq ] ) is due to the fact that the contribution of @xmath369-segment strongly singular orbits with @xmath375 is counted twice .
see proposition [ transition_prop ] . there is a correspondence between strongly singular billiard orbits and generalized diagonals .
let @xmath376 @xmath377 be a strongly singular orbit .
let @xmath378 ( resp .
@xmath379 ) be the last ( resp . the first ) corner point .
note that @xmath380 .
set @xmath381,\tilde{x}_n=[b(x_n),q]$ ] .
then @xmath382 is a generalized diagonal .
the mapping @xmath360 is a bijection , and it preserves the combinatorial length
. let @xmath383 ( resp .
@xmath384 ) be the number of @xmath12-segment isolated ( resp .
families of ) generalized diagonals , @xmath385 . set @xmath386 [ gen_diag_thm ] let @xmath18 be a geodesic polygon on a surface @xmath38 of constant nonpositive curvature .
let @xmath32 be the associated piecewise convex transformation , and let @xmath371 be the billiard complexity corresponding to the partition @xmath6 . then there are integers @xmath387 depending on @xmath18 , so that : @xmath388 * proof*. by lemma [ conj_lem ] , strongly singular billiard orbits are isolated .
the same argument shows that generalized diagonals are isolated , as well .
the claim now follows from ( [ compl_curv_eq ] ) and the correspondence between strongly singular billiard orbits and generalized diagonals .
we will now apply the preceding material to the complexity , @xmath389 , of the coding billiard orbits by the sides of @xmath18 . to avoid ambiguities
, we will define this coding now .
let @xmath18 have @xmath62 sides @xmath390 and denote by @xmath391 the set of phase points @xmath9 such that @xmath392 .
as @xmath393 run through the pairs of sides such that @xmath394 has a nonempty interior , the sets @xmath394 form the partition @xmath395 of the phase space .
then the _ side complexity _ , @xmath389 , is the complexity of the billiard map with respect to @xmath395
. it will be useful to have a direct definition of the side complexity .
we denote by @xmath396 the full language on the alphabet of sides of @xmath18 .
let now @xmath397 be a regular @xmath12-segment billiard orbit .
then @xmath398 are interior points of sides @xmath399 , and we set @xmath400 the language @xmath401 is the range of the mapping , and we set @xmath402
. then the _ side complexity _ satisfies @xmath403 to compare complexities , we will use an elementary lemma .
[ compar_lem ] let @xmath32 be a piecewise convex transformation , and let @xmath371 be the corresponding complexity .
let @xmath404 be a partition , such that @xmath405 , and let @xmath406 be the complexity of @xmath32 with respect to @xmath404 .
then @xmath407 .
[ side_cor ] let @xmath18 be a geodesic polygon on a surface @xmath38 of constant curvature @xmath40 , and let @xmath389 be the complexity of coding billiard orbits by the sides of @xmath18 . then the following inequalities hold : \1 . if @xmath36 , then there are constants @xmath372 such that @xmath408 2 .
if @xmath348 , then there are constants @xmath387 such that @xmath409 * proof*. let @xmath6 be the defining partition of the associated piecewise convex transformation . by proposition [ discont_prop ] , the partitions
@xmath410 satisfy @xmath411 .
now the first ( resp .
second ) claim is immediate from proposition [ str_sing_prop ] ( resp .
theorem [ gen_diag_thm ] ) and lemma [ compar_lem ] .
the preceding considerations drastically simplify for convex polygonal billiard tables .
[ convex_prop ] let @xmath39 be a convex geodesic polygon on a surface of constant curvature @xmath40 .
let @xmath389 be the traditional complexity of billiard orbits in @xmath18 coming from the coding of billiard orbits in @xmath18 by the sides they hit .
then there exist constants @xmath387 depending only on @xmath18 such that the following holds : \1 .
let @xmath36 .
then @xmath412 2 .
let @xmath348 .
then @xmath413 * proof*. in the notation of section [ innerbill ] , we have @xmath414 the standard defining partition is @xmath415 , and we have @xmath416 . in view of the above , @xmath417 , and
hence the two complexities coincide : @xmath418 . for a convex polygon ,
the notion of strongly singular billiard orbits and the notion of generalized diagonals coincide .
the argument of proposition [ transition_prop ] works . due to the coincidences we just pointed out
, it establishes a bijection between @xmath419 ( resp .
@xmath322 ) and the set of isolated ( resp .
families of ) @xmath12-segment generalized diagonals in the convex polygon @xmath18 . from this bijection and ( [ complex_eq ] )
, we obtain the first claim .
now the first statement of lemma [ conj_lem ] yields the second claim .
[ low - order_rem ] the preceding proof yields expressions for the constants @xmath387 . from ( [ complex_eq ] )
, we have @xmath420 let the polygon @xmath18 have @xmath62 sides .
the integer @xmath421 ( resp .
@xmath422 ) is the number of types of @xmath198-segment ( resp .
@xmath23-segment ) billiard orbits in @xmath18 . by convexity , @xmath423 .
however , @xmath422 is not determined by @xmath62 alone .
we will now apply the preceding material to the complexity of polygonal billiards .
we will consider the three cases separately .
there is a considerable literature on the billiard dynamics in euclidean polygons .
see @xcite for references .
many basic questions remain open @xcite .
one of them is whether the complexity ( of the coding by sides ) of billiard orbits grows ( at most ) polynomially .
it is known that the growth is subexponential @xcite , implying that the billiard in a euclidean polygon has zero topological entropy . by corollary [ side_cor ] ,
the complexity is bounded from above by the counting function for generalized diagonals .
the latter also grows subexponentially @xcite , and is believed to grow at most polynomially @xcite .
a euclidean polygon is _ rational _ if all of its angles are rational multiples of @xmath252 .
rational polygons play an important role in the subject @xcite .
[ rat_complex_thm ] let @xmath18 be a rational euclidean polygon , and let @xmath389 be the side complexity of billiard orbits in @xmath18
. then there exists @xmath424 such that @xmath425 if @xmath18 is convex , then there exist positive constants @xmath387 such that @xmath426 * proof .
* by a theorem of h. masur @xcite , @xmath427 grows quadratically for rational polygons .
more precisely , there exist positive constants @xmath428 such that @xmath429 the first claim follows from this and the bound ( [ side_eq3 ] ) .
the second claim follows the same way from the formula above and the identity ( [ convex_eq2 ] ) .
[ convex_rem ] we believe that the cubic lower bound on complexity ( [ cub_eq1 ] ) is valid for arbitrary rational polygons .
however , the geometry of the billiard map is much simpler in the convex case , as we saw in section [ convex_sub ] .
figure [ fig2 ] shows that the singular graph of the billiard map in a quadrilateral is much more complicated in the nonconvex case .
the bounds ( [ cub_eq1 ] ) were obtained in @xcite .
the billiard in a spherical polygon has two salient features .
first , the billiard map is a piecewise isometry .
second , a spherical polygon may have non - isolated strongly singular orbits , in particular , non - isolated generalised diagonals .
we will use the notation @xmath430 for the space of oriented geodesics on @xmath38 . let @xmath431 be the standard diffeomorphism .
( see section [ neg_curv_sub ] for a discussion of spherical duality . ) pulling back by @xmath432 the round metric on @xmath433 , we obtain an invariant metric ; the distance between two great circles is the angle between them .
metric geodesics and geodesics of the projective structure coincide .
let now @xmath434 be a polygon .
let @xmath208 be the billiard map phase space , and let @xmath435 be the map introduced in section [ innerbill ] .
pulling back by @xmath436 induces not only the structure of a piecewise convex surface on @xmath0 but also a metric on convex pieces of the phase space .
locally , the billiard map @xmath260 is a geodesic reflection , i.e. , an isometry .
we summarize this discussion as a proposition .
[ piec_isom_prop ] let @xmath18 be an arbitrary spherical polygon .
then the billiard map phase space is a piecewise metrically convex riemannian surface .
the billiard map is a piecewise isometry .
this observation has consequences for complexity of billiard orbits .
[ spheresubexp ] let @xmath18 be any spherical polygon .
is an admissible polygon . ]
then the complexity of the coding of billiard orbits by sides of @xmath18 grows subexponentially .
* let @xmath32 be the associated piecewise convex transformation .
denote by @xmath371 its complexity , and let @xmath389 be the standard complexity of billiard orbits . by the proof of corollary [ side_cor ] , @xmath437 .
hence , it suffices to show that @xmath111 grows subexponentially . by proposition [ piec_isom_prop ]
, @xmath32 is a piecewise isometry on a convex partition . by theorem 4.2 of @xcite ,
its complexity has subexponential growth .
the examples below illustrate peculiarities of spherical polygonal billiards .
* example 1*. let @xmath434 be a polygon such that @xmath347 is a finite group . for instance , @xmath18 may be the fundamental domain of a finite , generated by reflections group of isometries .
then every billiard orbit in @xmath18 is periodic .
moreover , there is a finite number of symbolic codes corresponding to prime periodic orbits , hence complexity is bounded .
* example 2*. let @xmath18 be a
bigon " ; it is bounded by two geodesics , @xmath200 connecting the north and the south poles .
let @xmath299 be the angle between them .
is not an admissible polygon . ]
if @xmath299 is @xmath252-rational , then we are in the situation of example 1 . we will now discuss the case when @xmath299 is @xmath252-irrational .
first , we point out that in any case the set @xmath438 consists of @xmath23 elements : @xmath439 and @xmath440 .
thus , @xmath441 .
there is an obvious periodic orbit .
it corresponds to the intersection of @xmath18 with the equator , and it has @xmath23 segments , perpendicular to @xmath200 .
we will denote this orbit by @xmath442 .
* claim . *
let @xmath299 be @xmath252-irrational .
then @xmath442 is the only prime periodic orbit in @xmath18 .
we will show that any periodic orbit @xmath248 is a multiple of @xmath442 .
we can assume that @xmath248 has an even number , @xmath443 , segments , and that its symbolic code is @xmath444 .
then the element , @xmath445 , of the group @xmath347 that we obtain by tracing @xmath248 is @xmath446 where @xmath447 is the rotation about the vertical axis by the angle @xmath448 .
let @xmath449 be the spherical geodesic corresponding to @xmath248 .
( note that @xmath449 differs , in general , from the unfolding @xmath450 , which is a geodesic segment along @xmath449 . )
the periodicity of @xmath248 implies that @xmath449 is invariant under @xmath445 , which rotates the sphere about the vertical axis by @xmath451 .
the only geodesic invariant under this ( nontrivial ! ) rotation is the equator , which implies the claim . by convention , a periodic billiard orbit in @xmath18
does not pass through its corners .
in particular , it can not trace the boundary of @xmath18 .
it is not known if every euclidean polygon has a periodic orbit @xcite .
below we present a spherical polygon without periodic orbits .
* example 3*. for @xmath452 let @xmath453 be the isosceles spherical triangle with two right angles , and whose third angle is @xmath299 .
if @xmath299 is @xmath252-rational then every billiard orbit in @xmath454 is periodic . if @xmath299 is @xmath252-irrational
, then @xmath454 has no periodic billiard orbits .
we outline a proof below . the triangle @xmath454 is obtained from the bigon @xmath18 of example 2 by folding it about the equator .
in this situation , every billiard orbit , @xmath248 , in @xmath454 uniquely lifts to a billiard orbit @xmath450 in @xmath18 ; the orbit @xmath248 is periodic iff so is @xmath450 .
if @xmath299 is @xmath252-rational , then @xmath454 satisfies the conditions of example 1 .
let @xmath299 be @xmath252-irrational , and let @xmath248 be a periodic orbit in @xmath454 . by preceding remark and example 2 , @xmath450 runs along the equator .
thus , @xmath248 traces the boundary of @xmath454 .
we will now discuss the hyperbolic case .
a positive function , @xmath455 , of natural argument is _ subexponential _ if @xmath456 for any @xmath457 and all sufficiently great @xmath12 .
[ hyperb_exp ] let @xmath458 be a geodesic polygon , and let @xmath371 be the complexity of the coding of billiard orbits by sides of @xmath18 .
denote by @xmath459 the _ topological entropy _ of the billiard map in @xmath18 .
+ then @xmath460 ; there exists a subexponential function @xmath455 such that @xmath461 .
* proof . *
the billiard flow of @xmath18 is ( uniformly ) hyperbolic @xcite .
thus , the metric entropy of the billiard flow with respect to the liouville measure is positive . by abramov s formula ,
the metric entropy of the billiard map in @xmath18 is positive , as well .
the metric entropy is a lower bound on the topological entropy , hence our first claim .
let @xmath410 be our defining partition and the partition by sides of the phase space of the billiard map .
let @xmath462 be infinite billiard orbits that bounce of the same sides of @xmath18 as @xmath463 .
let @xmath464 be their unfoldings
. then @xmath464 are infinite geodesics in @xmath465 such that the distance between them is bounded for @xmath463 . hence @xmath466 and @xmath467 implying @xmath468 .
thus , @xmath469 , and hence @xmath395 is a generating partition . since @xmath411 ,
the convex partition @xmath6 is generating as well .
by @xcite , the complexity of @xmath32 has the form @xmath470 , where @xmath471 is a subexponential function .
the relation @xmath411 implies @xmath472 .
the cardinality of the number of atoms of @xmath473 that partition an atom of @xmath474 grows at most polynomially ; thus @xmath475 for some positive integer @xmath476 .
[ lorenz_rem ] unlike @xmath477 , the spaces @xmath478 and @xmath479 do not have riemannian metrics , invariant under the natural actions of the groups of isometries of @xmath480 and @xmath481 , respectively .
however , there exists an invariant lorentz metric on @xmath478 .
we describe it below .
let @xmath482 be the upper sheet of the hyperboloid @xmath483 in @xmath484 equipped with the pseudo - riemannian metric @xmath485 .
the induced metric on @xmath482 is the metric of constant negative curvature .
isometries of @xmath482 are the restrictions of @xmath351-orthogonal transformations of the ambient space .
geodesics in @xmath482 are its intersections with planes through the origin . given such a plane ,
its @xmath351-orthogonal complement is a line that intersects the hyperboloid of one sheet @xmath486 at two antipodal points ; if the plane is oriented then so is the line , and one can canonically choose one of these intersection points .
this construction identifies @xmath478 with @xmath487 .
the metric @xmath351 induces a pseudo - riemannian metric of signature @xmath488 on @xmath487 .
the identification above yields a lorentz metric on @xmath478 .
let now @xmath489 be a geodesic polygon , and let @xmath0 be the phase space of the billiard map in @xmath18 .
the canonical geodesic space structure in @xmath0 is induced by the natural mapping @xmath490 . pulling back the lorenz metric on @xmath478
, we obtain a metric @xmath491 on @xmath0 . by construction , @xmath491 is locally invariant under the billiard map .
is a lorentz metric with singularities . if @xmath18 is a convex polygon , then @xmath491 is regular . ]
thus , the billiard map in a hyperbolic polygon is a piecewise lorentz isometry .
we do not know of any applications of this observation .
let @xmath32 be a piecewise convex transformation . if the mappings @xmath65 are isometries , then @xmath32 is a _ piecewise convex isometry_. piecewise isometries in one dimension are the interval exchange maps .
they arise , in particular , from the billiard in rational polygons @xcite , and have been much studied .
we will investigate the complexity of a particular class of piecewise convex isometries the outer billiard transformations .
let @xmath38 be a simply connected surface of constant curvature @xmath340 .
for @xmath492 let @xmath493 be the geodesic symmetry about @xmath9 .
let @xmath39 be a convex polygon with @xmath62 vertices @xmath494 listed counterclockwise .
if @xmath495 , set @xmath496 . if @xmath36 , i. e. , @xmath38 is the sphere , we assume that @xmath18 is contained in a hemisphere .
let @xmath497 be the antipodal polygon , and set @xmath498
. for a vertex , say @xmath209 , of @xmath18 , let @xmath499 be the geodesic ray extending the side @xmath500 in the direction of @xmath209 .
the rays @xmath501 partition @xmath0 into convex polygons @xmath502 .
see figure [ fig3 ] .
we denote this partition by @xmath6 .
the statement below defines the protagonist of this section .
let @xmath506 be the partition of @xmath0 defined above .
then @xmath6 is a convex geodesic partition .
the piecewise convex isometry @xmath32 defined by the geodesic symmetries @xmath507 is the _
outer billiard _ about @xmath18 , and the geodesic surface @xmath508 ( it is a topological annulus ) is the _ phase space of the outer billiard_. we will use the notation @xmath509 for the outer billiard , suppressing the subscript if no confusion arises . by _ complexity of the outer billiard _ we mean the complexity of @xmath32 with respect to the partition @xmath6 .
now we introduce notation and terminology that will be used throughout this section . if @xmath510 are two positive sequences then we write @xmath511 if there is a constant @xmath512 such that for all @xmath12 sufficiently large @xmath513 .
if @xmath511 and @xmath514 , then we write @xmath515 ; we will say that the sequences _ have the same growth _ or are in the _ same ( growth ) class_. if @xmath516 then we say that @xmath351 grows at most polynomially with degree @xmath476 , or that @xmath351 is bounded by @xmath517 .
if @xmath518 is a group with a finite set @xmath519 of generators , we denote by @xmath520 the set of elements that can be represented by products of at most @xmath12 elements of @xmath199 and their inverses .
the growth class of the sequence @xmath521 does not depend on the choice of @xmath199 @xcite .
if @xmath522 , then we say that the group @xmath518 _ grows polynomially , with degree @xmath476_. let @xmath523 be the group generated by the geodesic reflections in the vertices of the polygon @xmath18 .
we will relate the growth of @xmath518 and the complexity of the outer billiard about @xmath18 .
we proceed to study separately the three cases at hand .
* proof . * the edges of the graph @xmath104 are parallel to the sides of @xmath18 ; each edge is a segment or a half - line .
assume , for simplicity of exposition , that @xmath18 has no parallel sides .
then there are @xmath62 directions . for each direction
there are @xmath12 parallel half - lines , hence their total number is @xmath525 . since they partition @xmath0 into @xmath525 components ,
the number of faces of @xmath104 is at least @xmath525 .
this yields the linear lower bound on complexity .
now , for the upper bound . * proof . *
the subgroup @xmath529 generated by @xmath530 is a quotient group of @xmath531 ; hence its growth is bounded by @xmath528 . since @xmath482 is a normal subgroup of @xmath518 of index @xmath23 , the two groups have the same growth . * proof . *
the edges of @xmath537 are obtained from the edges of @xmath534 by applying the inverse map @xmath538 .
each time a singularity half - line of @xmath538 intersects an edge of @xmath534 , this edge splits into two , and thus contributes @xmath198 to @xmath539 .
let @xmath540 be the set of straight lines obtained by reflecting at most @xmath12 times in the vertices of @xmath18 the extentions of the sides of @xmath18 .
by lemma [ growthgroup ] , @xmath541 .
each of these lines intersects a singularity half - line of @xmath538 at most once , therefore the total number of intersections of the lines in @xmath540 with the singularity half - lines of @xmath538 is bounded above by @xmath528 .
the edges of @xmath534 belong to the lines from @xmath540 , therefore the total number of intersections of these edges with the singularity half - lines of @xmath538 is bounded above by @xmath528 .
( note that the number of the edges of @xmath534 could be bigger . )
we will now obtain the desired bound on complexity , i.e. , we will estimate the number of faces of @xmath104 .
denote by @xmath542 the number of faces , edges , vertices of the graph @xmath104 respectively . by lemma
[ growthygraph ] , growth of the second difference of the sequence @xmath543 is at most polynomial of degree @xmath544 , hence @xmath545 .
the edges of @xmath104 are parallel to the sides of @xmath18 , thus may have at most @xmath62 possible directions .
therefore , each face of @xmath104 is at most a @xmath546-gon , and the valence of each vertex of @xmath104 is at most @xmath546 .
thus , @xmath547 .
formula @xmath548 implies @xmath549 , hence @xmath550 .
we have obtained our bound , assuming that the rank of the abelian group generated by the sides of @xmath18 is @xmath544 , i.e. , maximal possible .
although generically this is the case , the rank may drop .
our argument proves , in fact , the statement below .
[ bound_cor ] let @xmath18 be a convex euclidean @xmath62-gon , and let @xmath551 be the rank of the abelian group generated by translations in the sides of @xmath18
. then the complexity of the outer billiard about @xmath18 is bounded by @xmath552 .
a polygon is _ rational _ if the rank above is @xmath23 .
rational polygons are dense in the space of all polygons .
we will study complexity of the outer billiard about a rational polygon .
first , we recall preliminaries .
we regard the plane as a vector space , with the center in the interior of the convex @xmath62-gon @xmath18 .
a well known construction @xcite associates with @xmath18 a homothetic family of centrally symmetric convex polygons with at most ( resp .
exactly ) @xmath546 sides ( resp . if @xmath18 is a generic @xmath62-gon ) .
let @xmath454 be a particular polygon in this family .
each of its sides is parallel to a diagonal of @xmath18 .
see figure [ figa ] .
we endow the plane with a minkowski norm such that @xmath454 is the unit disc .
the vector norm @xmath553 , radius , etc , will be understood with respect to it .
we set @xmath554 .
the polygon @xmath454 determines the geometry of orbits of @xmath555 at infinity " @xcite
. we will elaborate .
let @xmath9 be a point in the plane which is sufficiently far from the origin .
let @xmath556 be the circle centered at the origin and passing through @xmath9 .
let @xmath557 be the side containing @xmath9 , and let @xmath476 be the corresponding diagonal of @xmath18 ( parallel to @xmath209 ) .
then @xmath555 translates @xmath9 along @xmath209 by @xmath558 ; this continues until the orbit of @xmath9 overshoots @xmath209 .
let @xmath559 be the corresponding point .
then the recipe above is applied to @xmath560 , etc .
see figure [ figb ] .
let @xmath209 be an arbitrary side of @xmath454 , and let @xmath476 be the corresponding diagonal of @xmath18 .
the polygon @xmath18 is _ quasirational _ if , up to a common factor , the @xmath62 numbers @xmath561 are rational .
* proof . *
every rational polygon is quasirational . by a construction of r. kolodziej @xcite
, there is a nested sequence of @xmath33-invariant polygonal simply connected domains @xmath563 exhausting the plane .-
gons remains an open question for @xmath564 @xcite .
] by @xcite , there exists a constant @xmath565 such that the _ kolodziej domains _ satisfy @xmath566 .
we will need a general lemma .
[ infinity ] let @xmath18 be an arbitrary convex polygon , and let @xmath371 be the complexity of the outer billiard about @xmath18 .
there exists @xmath567 such that the contribution to @xmath111 of the exterior of the disc of radius @xmath568 grows linearly .
consider the @xmath555-orbit of length @xmath12 of an arbitrary point @xmath9 outside of @xmath571 .
it follows a side , @xmath209 , of @xmath454 for @xmath375 iterations , then it jumps " to the adjacent side , @xmath572 , and follows it for @xmath573 iterations . counting the possibilities ( and assuming that @xmath454 is a @xmath546-gon , i.e. , that we are in the generic sitiation ) we obtain @xmath574 types of @xmath555-orbits of length @xmath12 .
but different types mean different contributions to @xmath575 , and vice versa .
since @xmath18 is a rational polygon , the group @xmath576 is discrete .
the graphs @xmath104 are obtained from a finite collection of half - lines by @xmath518-action , hence @xmath577 belongs to a discrete collection of lines .
therefore @xmath578 is a graph , and the sequence @xmath579 stabilizes on compacta .
moreover , there is a finite collection of convex polygons , such that every face of @xmath578 is congruent to a polygon in this collection .
hence the areas of the faces of @xmath578 are bounded away from zero and infinity .
note that the constant @xmath580 in lemma [ infinity ] can be chosen arbitrarily large .
we choose it so that @xmath581
. then for all @xmath12 sufficiently large @xmath582 by lemma [ infinity ] , up to a linear term , @xmath111 is the number of faces of @xmath104 intersecting @xmath583 . by the left inclusion in ( [ inclu_eq ] ) , this is less than or equal to the number of faces of @xmath578 in @xmath584 . by preceding remarks ,
there is @xmath585 such that that number is bounded by @xmath586 . by the right inclusion in ( [ inclu_eq ] ) , @xmath587 is quadratic in @xmath12 .
we have obtained the upper bound @xmath588 .
now for the lower bound .
all regular points in @xmath0 are periodic @xcite .
a face @xmath589 of @xmath590 is _ stable _ if @xmath591 is a face of @xmath578 .
let @xmath592 be the set of points with period at most @xmath12 .
each connected component of @xmath593 is an open , stable face of @xmath104 . by remarks above
, the number of connected components of @xmath593 has the same growth as the area of @xmath593 , thus @xmath594 . by proposition [ period_lem ] below , @xmath595 .
the following proposition is used in the proof of theorem [ rational ] .
it is also of independent interest .
if @xmath596 are positive functions on @xmath597 , the notation @xmath598 means that @xmath599 is bounded as @xmath600 .
the notation @xmath601 means that @xmath602 .
* we assume without loss of generality that @xmath609 . let @xmath556 be the circle through point @xmath9
. the sequence @xmath610 roughly follows @xmath556 . to come back to @xmath9 ,
the sequence has to go around @xmath556 at least once .
let @xmath611 be the largest step " of @xmath555 .
then we need at least @xmath612 steps to return .
since @xmath613 , the first claim follows .
let now @xmath18 be rational , and hence quasirational polygon , and let @xmath614 be the kolodziej domains .
let @xmath615 be such that @xmath616 .
the relations @xmath617 imply that the function @xmath618 satisfies @xmath619 . by inclusion @xmath620
, we have @xmath621 . the point @xmath9 belongs to a unique face , @xmath622 , of @xmath578 , hence @xmath623 . by preceding remarks , @xmath624 , implying @xmath625 , and hence the equivalence @xmath626
. * proof . * for @xmath629 let @xmath630 be the appropriately oriented great circle centered at @xmath9 .
this diffeomorphism @xmath631 is the _ spherical duality _ , and we denote by @xmath632 the inverse diffeomorphism .
let @xmath633 be the vertices of @xmath18 .
the geodesics @xmath634 bound the convex polygon @xmath635 .
the correspondence @xmath636 is an automorphism of the space of convex spherical polygons .
the proof of the following lemma is contained in @xcite .
( see also @xcite . )
[ duality ] let @xmath637 be as above .
let @xmath638 be the phase spaces of the outer billiard about @xmath18 , inner billiard about @xmath635 , and let @xmath639 be the outer billiard , inner billiard maps respectively .
the spherical duality induces a diffeomorphism @xmath640 ; it conjugates @xmath641 and @xmath642 ; it induces an isomorphism of the coding of @xmath21-orbits by corners of @xmath18 and the coding of @xmath17-orbits by sides of @xmath635 .
figure [ figh ] illustrates lemma [ duality ] .
let @xmath643 ( resp .
@xmath644 ) be the corner complexity of the outer billiard about @xmath18 ( resp .
billiard in @xmath635 ) . by lemma [ duality ] , @xmath645 .
the claim now follows from theorem [ spheresubexp ] .
we refer to @xcite for the background on polygonal outer billiards in the hyperbolic plane .
let @xmath458 be a @xmath62-gon , let @xmath646 , and let @xmath1 be the outer billiard map .
it extends to a homeomorphism of the circle at infinity , @xmath647 .
its rotation number satisfies @xmath648 @xcite .
the polygon @xmath18 is _ large _ if @xmath649 and @xmath650 has a hyperbolic @xmath62-periodic orbit .
see figure [ figc ] .
the set of large polygons is open in the natural topology @xcite .
* proof . *
the bound @xmath651 fails iff the sequence @xmath653 stabilizes .
assume this to be the case , and let @xmath654 .
the outer billiard map @xmath1 preserves @xmath578 ; the restriction of @xmath33 to a closed face of @xmath578 is a diffeomorphism onto another one .
since @xmath578 is a finite graph , we find @xmath655 such that every face of @xmath578 is invariant under @xmath656 .
let @xmath591 be a closed face of @xmath578 .
then @xmath657 is either empty , or a vertex , or an edge of @xmath591 .
we will study the latter .
let @xmath658 be the consecutive endpoints of these edges , let @xmath659 ( resp .
@xmath660 ) be the circular arc ( resp .
the geodesic ) with endpoints @xmath661 ( we set @xmath662 ) , and let @xmath663 be the corresponding face of @xmath578 .
the restriction @xmath664 is induced by an isometry , @xmath665 .
the elements @xmath666 are all equal to the identity iff @xmath667 .
* let @xmath669 be a vertex of @xmath18 .
for close points @xmath670 let @xmath671 be their reflections about @xmath669 .
let @xmath672 and let @xmath673 be the angular measure of the arc @xmath674 ; see figure [ fige ] . the triangles @xmath675 and @xmath676 are similar , therefore @xmath677 let @xmath678 be a periodic trajectory of the map @xmath650 consisting of smooth points , and let @xmath679 be the respective ratios . set @xmath680 .
let @xmath681 be a point sufficiently close to @xmath314 , and let @xmath682 be its @xmath650-orbit ; we assume that for both orbits the reflections occur in the same vertices of @xmath18 .
it follows from equation [ cyclicorb ] that @xmath682 is a periodic trajectory iff @xmath683 in particular , if @xmath650 has a periodic interval , then @xmath684 there .
let now @xmath314 cross counter - clockwise a singularity half - line of @xmath33 . in the notation of figure [ figf ] , @xmath685 ( resp .
@xmath686 ) right before ( resp .
after ) this . by @xmath687 , the equality @xmath684
before a singularity half - line implies that @xmath688 immediately after it . by lemma [ nonperiod ] , we can assume without loss of generality that @xmath689 .
then @xmath690 is a ( hyperbolic ) parallel translation with the axis @xmath691 , and @xmath692 is the domain bounded by @xmath691 and @xmath693 .
we will say that @xmath692 is a _ lunar face _ of @xmath578 .
the union of lunar faces of @xmath578 is invariant under @xmath33 .
therefore for any @xmath694 there is @xmath695 such that @xmath696 .
a geodesic @xmath697 can not contain a side of @xmath18 .
if it does , then @xmath663 contains a singular line of @xmath33 in its interior , contrary to the definition of @xmath663 .
see figure [ figf ] where @xmath698 represents now the geodesic @xmath691 .
thus , @xmath691 is not an edge of @xmath699 for any @xmath700 .
this contradiction proves our first claim .
[ degen_rem ] note that the function @xmath371 is bounded below by the complexity of the induced map @xmath647 with respect to the natural partition .
however , the latter may be finite .
see figure [ figg ] . | abstract we introduce the class of piecewise convex transformations , and study their complexity .
we apply the results to the complexity of polygonal billiards on surfaces of constant curvature . |
a muon component of cosmic rays plays an important role in many fields of astroparticle physics .
it provides a basis for verification of our knowledge on primary cosmic ray ( pcr ) spectrum behaviour , high - energy hadronic interactions and for solution of neutrino physics problems . in view of this
, a question of interconsistency of information , gained in the last two decades on primary and muon fluxes , and on high - energy hadronic interactions presents great importance . in well known works of volkova et al .
( 1979 ) @xcite , dar ( 1983 ) @xcite , butkevich et al .
( 1989 ) @xcite , lipari ( 1993 ) @xcite , honda et al .
( 1995 ) @xcite , agrawal et al .
( 1996 ) @xcite and bugaev et al .
( 1998 ) @xcite this problem was outside of the consideration , since the principal aim lied in the estimation of the secondary lepton fluxes .
authors relied on the information on primaries and nuclear interactions , that was available at the papers writing time .
certainly , these data were incomplete and ambiguous , as the consequence the inputs in the calculations also vary rather significantly .
nevertheless , the outputs , i.e. muon fluxes , satisfactory agree with each other ( see comparisons in @xcite ) and muon experimental data .
the given circumstance with necessity shows , that discrepancies in the used pcr spectra in the large part were compensated by different approaches to the treatment of nuclear cascades .
this , in turn , may possibly relate to the fact , that muon flux in most of these works served only for normalization of neutrino fluxes and it was fitted to the muon experimental data via adjustment of not precisely known parameters of hadronic interactions . from these short remarks it is clear , that all these calculations lack some standardization , as it was recently proposed by gaisser and honda @xcite in respect to the pcr spectra .
the current situation provides significantly better capabilities for adequate choice of pcr model and for simulation of nuclear cascades , and in this paper we have tried to take the maximum advantage of this .
but , as our analysis has shown , unfortunately the information on all components of calculations is still very uncertain to provide a firm ground for accurate derivation of muon flux at sea level and , today , it is following ` from top to bottom ' .
direct measurements of primary cosmic ray spectra span up to the energy of @xmath21 pev for protons and up to few hundred tev / n for other groups of nuclei .
most abundant data are collected for @xmath3 tev / n . here proton spectrum is studied in series of recent satellite and balloon experiments with approximately 20% accuracy .
higher energy data of sokol @xcite , mubee @xcite , jacee @xcite and runjob @xcite although have relatively large statistic and systematic errors , also satisfactory agree with each other . at the same time , the measurements of the helium spectrum are much less concordant and still differ by almost a factor of 2 for all energies , and this uncertainty is especially crucial for evaluation of @xmath4 and @xmath5 ratios .
as it is widely known , protons and helium nuclei contribute @xmath6 to the nucleon flux at the top of the atmosphere , which is relevant for derivation of the muon spectrum at sea level , thus the particular emphasis should be given to the accurate description of these spectra .
to avoid simplification , for the simulation of nuclear cascades we have applied widely approved and thoroughly tested monte - carlo code corsika @xcite , that allows to treat hadronic interactions with the use of any of the up - to - date interaction models : qgsjet @xcite , venus @xcite , hdpm @xcite , nexus @xcite or dpmjet @xcite .
overall uncertainty , brought in computations of muon flux by the use of these models , is not very significant and considerably decreased in the last decade .
so , the difference in @xmath7 inelastic cross - sections ` between the models have shrunk from 80 mb to today 20 mb in the region of few pev ' @xcite .
average numbers of muons in cascades from primary protons , obtained with the use of these models , differ not more than by 20% @xcite for all energies of interest .
this means , that integral and differential muon fluxes at sea level will also differ by a close value , and this discrepancy is the smallest , in comparison with the uncertainties of pcr chemical composition and energy spectra for @xmath8 tev / n , and the sea level muon data .
the latter , for @xmath9 gev , are both numerous and , on the whole , quite ambiguous ( see extensive summary in @xcite ) , but the most recent ones of bess @xcite and caprice @xcite are very precise and closely agree with each other .
they give a basis for checking information on primary light nuclei spectra up to the energies @xmath10 gev / n .
the interval of higher muon energies @xmath11 gev , where the spread of the data on muon intensity is still small and amounts to some 20% , allows to examine with nearly the same accuracy behaviour of primary proton and helium spectra for the energies , extending to the upper bound of balloon and satellite measurements @xmath12 pev / n . evidently , that beyond the pcr data , it is not feasible to perform reliable calculations of the muon spectrum .
what is more , the inverse problem , i.e. reconstruction of pcr fluxes from the data on muon spectrum , can not be also solved , because for @xmath13 tev there are only indirect data , in which muon flux at sea level is derived from results of underground measurements .
these data are quite contradictory and have large systematic errors , mainly caused by incomplete information on rock properties and vagueness in question of prompt muon generation mechanisms ( see , for example , @xcite ) . as a consequence of all this
, now there is no possibility to make conclusions neither on the preferability of any model of charm generation , nor on behaviour of pcr spectra for @xmath14 pev / n .
as it is stated above , this paper is devoted to the investigation of compatibility of the present sea level muon flux measurements in energy interval @xmath15 gev with the experimental data on pcr spectra for corresponding primary energies @xmath16 gev / n . in section 2 we briefly review the present data on primaries of h , he , cno , ne - si and fe groups and discuss pcr models , applied in this paper and in the papers of the other authors .
basic characteristics of computations are presented in section 3
. section 4 is devoted to consideration of consistency of our vertical muon flux , derived from the data of direct pcr measurements , with sea level muon experiments in the energy range @xmath15 gev . since in this section
we have revealed a sizable shortage of muons with energies , corresponding to that of primaries , studied with the emulsion chambers , in section 5 we examine possible effects , that may lead to a distortion of the information , obtained in the space and balloon experiments , employing this technique .
our conclusions are given in section 6 .
the extensive compilation of the modern data of space and balloon measurements plotted vs kinetic energy @xmath17 gev / n is shown in figures [ p]-[hefe ] .
the most of the data are gathered for relatively low energies @xmath18 tev / n .
but even here some 15 years ago the spread in experimental results was at least 100% for all groups of nuclei . in spite of series of experiments , performed since then ,
significant improvement was achieved only in the study of the proton spectrum .
so , for energies from @xmath210 gev , where effects of geomagnetic cutoff and solar modulations become negligible , to some hundred gev , data of recent experiments leap @xcite , caprice @xcite , imax @xcite , bess @xcite and ams @xcite are consistent within 20% ( see figure [ p ] ) .
better mutually consistent , within 510% , data of ams and bess , now are considered as the most precise . from @xmath2 200
gev to @xmath19 tev there is a gap between magnet spectrometer and emulsion chamber experimental data , with the only ionization calorimeter measurements of ryan et al . @xcite . since at the energies , overlapping with ams and bess , the given data overestimate proton intensity ( but not for helium ) , probably they should be lowered by 25% , as proposed in @xcite . in any case ,
more accurate information on behaviour of the proton spectrum in this energy interval is required for reliable evaluation of the muon and neutrino fluxes for the energies below 300 gev . from approximately 2 tev to @xmath20
tev sokol , mubee , jacee and runjob measurements provide rather consistent information on proton flux , though with the nearly 20% uncertainty . for understanding of the proton spectrum behaviour beyond the ` knee ' ,
valuable information is provided by the pioneer measurements of the tibet hybrid experiment @xcite , carried out at the mountain altitude .
these first results point at the steepening of the proton spectrum for @xmath21 tev and this is an essential guideline for the extrapolation of the direct measurements .
spread in the data for all heavier nuclei practically for all energies of interest is much larger than that for hydrogen , and equals to 100% and more .
the difference between the helium spectra of sokol , jacee and runjob is of major importance for accurate evaluation of secondary lepton fluxes .
.,title="fig:"].,title="fig : " ] .,title="fig:"].,title="fig : " ] .,title="fig : " ] from the aforesaid it becomes evident , that today no unique fit of all data on primary spectra can be given and considerable arbitrariness in the choice of them still remains .
in such events one may use , for example , upper and lower estimates of all primary fluxes , or the flux , giving some average of the experimental data .
the principal model of the pcr spectrum , employed in our calculations , may be regarded just as the latter case .
it was obtained in frameworks of anomalous diffusion of cosmic rays in fractal interstellar medium @xcite , but here it is considered irrespectively to its validity for description of cosmic ray propagation . altogether in our computations we have accounted for 5 groups of primaries : h , he , cno , ne - si and fe .
mass composition of the applied model is fitted to the direct and eas experimental data on the all - particle primary spectrum ( figure [ allp ] ) . as it it seen from figures [ p]-[hefe ] our model as well satisfactory corresponds to the available experimental information on elemental spectra and for high energies
really gives some average of the data on nuclei , except that for hydrogen it presents rather an upper estimate .
for calculation of muon flux it also matters , that the data on the proton component become insufficiently statistically provided from @xmath22 gev , and that sizeable uncertainty gives @xmath23% spread of the data on helium of jacee and runjob around the values , adopted in our model .
so , though on the whole the situation looks quite well - defined for calculations of vertical muon intensity up to the energies of several tev , for the greater energies the existing pcr data are very ambiguous . in order to estimate the influence of these data spread on the resulting muon flux
, we have additionally performed calculations with the use of another two pcr models , shown in figures [ p]-[hefe ] : first of them presents the jacee fit of their hydrogen and helium measurements @xcite , and the second is the model , suggested by gaisser and honda @xcite . to compute sea level muon flux
we have converted primary nuclei spectra to a spectrum in nucleons in standard manner : @xmath24 here @xmath25 is the differential energy spectrum of the nuclei with atomic number @xmath26 , @xmath27 . in figure [ nucleons ] it is presented , along with our primary nucleon flux ( solid line ) , fluxes , applied in calculations of muon intensity in works @xcite .
these pcr spectra vary both in shapes and values and here it is useful to discuss all of them in order to realize , how they correspond to the modern experimental data .
first of all , one can see that a sizable discrepancy between our , ` gaisser and honda ' and ` jacee fit + 10% ' ( here we add 10% to account a contribution of heavier , than h and he , nuclei ) prevails , and this is a good illustration of the difficulties in the attaining of the unique description of the current experimental data .
behaviour of our model requires some more comments .
enhancement by @xmath28 of intensity of the proton component in the energy range @xmath29 gev is dictated by the fact , that our first test calculations of muon flux with the ` average ' , close to jacee fit , proton flux resulted in a very large shortage of muons at sea level ( see section 4 ) . in order to smooth this contradiction
we have tested several variants of pcr spectra and eventually adopted the presented version .
it gives the maximal increase of the nucleon flux , provided the elemental and all - particle spectra stay consistent with the experiment .
note , that if the muon data did not force us to enlarge the hydrogen flux , then it would reasonably agree with the new data of the tibet as@xmath30 collaboration @xcite .
dip of our nucleon and elemental fluxes for @xmath31 gev / n is an intrinsic feature of the anomalous diffusion propagation model , and , on the other hand , it corresponds to the numerous indications about the change of the pcr spectral index in the ` knee ' region . from the earlier works ,
the most appropriate pcr models were employed by lipari @xcite and agrawal et al .
the nucleon spectrum from the latter paper almost coincides with our one , deviating only from the energy of about @xmath32 gev .
mainly , this discrepancy is caused by the use of excessive , in comparison with today s , flux of protons , obtained from the results of the first 8 jacee flights @xcite . in 1993 jacee
reported , that proton spectrum consists of two parts with different slopes and a spectral break at @xmath33 tev . in paper @xcite this break
was not accounted for and extrapolation of the data for @xmath34 tev was applied for the higher energies . in work of lipari
a simple power primary nucleon spectrum was picked , which is convenient to use in analytic calculations , and no collation with the experiment was presented .
nevertheless , this is a rather appropriate model , closely agreeing with the ` gaisser and honda ' fit .
the three remaining spectra , applied by volkova et al .
@xcite , honda et al .
@xcite and bugaev et al .
@xcite are quite excessive , compared to what is known today . in the oldest of the discussed work of volkova et al .
@xcite , it is applied very large all - nucleon spectrum @xmath35 @xmath36 , obtained from eas data , and no elemental composition had been considered .
spectrum , used by bugaev et al .
@xcite , is taken from the work of nikolsky @xcite , where it was derived from the analysis of fluctuations of muon and electron numbers in eas . for @xmath37 gev proton intensity in this model is essentially overestimated ( as a consequence , nucleon spectrum too ) , in comparison with the experimental data , and precisely such behaviour provides a good agreement of the resulting muon flux in @xcite with the most of experimental data on sea level muon intensity . in the paper of honda
et al . @xcite the proton flux for all energies is also rather high for the following reasons . in the low energy region
the given spectrum relies on the compilation of webber and lezniak @xcite , i.e. mostly on the ryan et al .
@xcite data .
but they are referred in @xcite incorrectly and do not match the original values , overstating them by @xmath28 .
extrapolation of this spectrum to the higher energies reasonably agree with the jacee @xmath38 flights results , but , as it was already said , to some extent overvalues the present data .
from the analysis of these three last models a simple deduction follows : their renormalization down to the experimental values should cause a sizable deficit of calculated sea level muons .
we shall return to this statement again in section 4 .
simulation of cascade processes in the atmosphere has been performed with the use of corsika ( v6.00 , v6.018 ) . in
most of calculations as the model describing hadron - nucleus interaction for energies @xmath39 gev qgsjet model is applied .
this model is in a good agreement with accelerator and eas experimental data ( see e.g. @xcite ) and provides comparatively high calculation speed .
hadronic interactions with energies @xmath40 gev were simulated with gheisha @xcite . 12 and proton@xmath4156 denote the calculation for carbon and iron nuclei in superposition model , i.e. the average number of muons in proton initiated shower , multiplied by 12 and 56 correspondingly .
statistical errors are not shown , since they are less , than the symbol size.,title="fig:"]12 and proton@xmath4156 denote the calculation for carbon and iron nuclei in superposition model , i.e. the average number of muons in proton initiated shower , multiplied by 12 and 56 correspondingly .
statistical errors are not shown , since they are less , than the symbol size.,title="fig : " ] in order to save machine time and accounting , that contribution of all nuclei , heavier than he to muon spectrum at sea level does not exceed 710% , we have used a superposition model , considering nucleus with energy @xmath42 as @xmath26 nucleons with energies @xmath43 .
validity of such approach for calculation of number of secondaries in eas at sea level is widely known ( see , _
e.g. _ , @xcite ) , however , for verification , we have made our own computations of muon numbers in cascades , initiated by primary protons , carbon and iron nuclei , for energies @xmath44 and @xmath45 ev /
n with the qgsjet model .
results are presented in figure [ superp ] .
as it is seen , the agreement between numbers of muons in cascades , simulated with realistic nuclei fragmentation and those , obtained in the frameworks of superposition models is excellent .
this , as well , allows not to make a distinction between showers , initiated by protons and neutrons , and further by nucleons we mean protons . in figure [ depos ]
the relative contribution of all nuclei groups , according to our pcr mass composition , to the integral muon flux at sea level is shown .
since the contribution of he nuclei amounts to @xmath28 throughout the entire energy range , thus mentioned in the previous section @xmath46 spread of experimental results on helium flux around values , taken in our model , brings approximately @xmath47 uncertainty to muon flux at sea level . from this point of view , the uncertainty due to ambiguity of the experimental situation for heavier nuclei is insignificant . to obtain easily the muon spectrum for any model of the primary spectrum ,
it is appropriate to perform calculations of integral muon flux in the following way : @xmath48 here @xmath49 is an average number of muons with energy @xmath50 in shower from primary nucleon with energy @xmath51 , @xmath52 differential primary spectrum converted to spectrum in nucleons , @xmath53 energy , which provides the calculation accuracy @xmath54 .
most of results , presented below , obtained at @xmath55 . for each threshold energy
average muon numbers were computed for 2025 different primary energies ( see figures [ nmu],[imu ] ) with accuracy , generally better than 5% .
vertical lines in these figures show the areas of primary energies , giving 10% , 50% and 95% contributions to the integral muon flux in case of power primary spectrum @xmath35 @xmath36 .
it is seen , that primaries with energies within @xmath56 on 95% determine muon intensity at sea level . totally , with the use of the qgsjet it was simulated about @xmath57 showers . besides , for examination of sensitivity of muon spectrum to hadron - nucleus interaction model , series of calculations with the venus model have been carried out . with energy above threshold , generated in shower from primary proton with energy @xmath58 .
vertical lines show the areas of primary energies , giving 10% , 50% and 95% contributions to the integral muon flux in case of power primary spectrum @xmath35 @xmath36.,title="fig : " ] with energy above threshold , generated in shower from primary proton with energy @xmath58 .
vertical lines show the areas of primary energies , giving 10% , 50% and 95% contributions to the integral muon flux in case of power primary spectrum @xmath35 @xmath36.,title="fig : " ] with energy above threshold , generated in shower from primary proton with energy @xmath58 .
vertical lines show the areas of primary energies , giving 10% , 50% and 95% contributions to the integral muon flux in case of power primary spectrum @xmath35 @xmath36.,title="fig : " ] with energy above threshold , generated in shower from primary proton with energy @xmath58 .
vertical lines show the areas of primary energies , giving 10% , 50% and 95% contributions to the integral muon flux in case of power primary spectrum @xmath35 @xmath36.,title="fig : " ] with energy above threshold , generated in shower from primary proton with energy @xmath58 .
vertical lines show the areas of primary energies , giving 10% , 50% and 95% contributions to the integral muon flux in case of power primary spectrum @xmath35 @xmath36.,title="fig : " ] with energy above threshold , generated in shower from primary proton with energy @xmath58 .
vertical lines show the areas of primary energies , giving 10% , 50% and 95% contributions to the integral muon flux in case of power primary spectrum @xmath35 @xmath36.,title="fig : " ] to @xmath58 to the muon integral flux at sea level @xmath59 versus @xmath60 .
vertical lines have the same meaning as in figure [ nmu].,title="fig : " ] to @xmath58 to the muon integral flux at sea level @xmath59 versus @xmath60 .
vertical lines have the same meaning as in figure [ nmu].,title="fig : " ] to @xmath58 to the muon integral flux at sea level @xmath59 versus @xmath60 .
vertical lines have the same meaning as in figure [ nmu].,title="fig : " ] to @xmath58 to the muon integral flux at sea level @xmath59 versus @xmath60 .
vertical lines have the same meaning as in figure [ nmu].,title="fig : " ] to @xmath58 to the muon integral flux at sea level @xmath59 versus @xmath60 .
vertical lines have the same meaning as in figure [ nmu].,title="fig : " ] to @xmath58 to the muon integral flux at sea level @xmath59 versus @xmath60 .
vertical lines have the same meaning as in figure [ nmu].,title="fig : " ] [ models ] @*6l + & @xmath61 gev&@xmath62 gev&@xmath63 gev&@xmath64 gev + & & 1011 & 318 & 21.0 & 0.605 + & & 1085&316&20.5&0.588 + & & 1008&310&20.9&0.696 + & & 1079&347&23.5&0.679 + & & 1150&349&24.0&0.604 + @*6l
+ & @xmath61 gev&@xmath62 gev&@xmath63 gev&@xmath64 gev + & & 8307&2292&132.3&3.612 + & & 8298&2207&124.9&3.240 + & & 8059&2257&129.8&3.332 + & & 9629&2590&153.5&3.932 + & & 9706&2653&155.0&4.156 + for the verification of our results , we have performed a comparison of numbers of muons in showers from primary protons , obtained by us , with those from papers of corsika authors @xcite for two primary energies @xmath65 gev ( table 1 ) . on the whole , we see good agreement between the data , presented in this table . some difference in muon numbers for @xmath66 gev may be explained by a statistical error , since simulation of 500 showers , made in the above mentioned papers for each primary energy , provides approximately 8% and 4% mean square deviation for energies @xmath67 and @xmath68 gev correspondingly .
obtained integral and differential muon spectra for our model of primary spectrum are presented in figures [ diff ] , [ int ] .
it is appropriate to split the consideration of the situation into three energy intervals .
* gev . * in this region
our calculation of differential muon spectrum is in a good agreement with the most recent and accurate data , obtained by bess 1995,97 - 99 @xcite and caprice 1994 , 1997 @xcite .
this , along with the availability of unambiguous data on primary spectrum and on behaviour of hadronic cross sections , once again evidences in favour of correctness of the applied calculation procedure .
selection of bess and caprice experiments is motivated by the fact , that measured there muon spectra have small statistic and systematic errors and excellently correlate with each other . in these experiments measurements were made with superconducting magnet spectrometers .
earlier data , obtained with iron magnet spectrometers differ from each other not only in values , but also in shapes of measured spectra , that may be due to the influence of some improperly accounted systematic errors .
more detailed discussion of this question may be found , for example , in @xcite .
approaching to 100 gev , the calculated muon flux becomes deficient , conforming only to the caprice data .
two circumstances in this connection are worth noting .
the first is that already from these energies information on muon intensity is not definite enough .
the second is that muons with @xmath69 gev are most effectively produced by the interactions of primaries with @xmath14 tev / n , measured in space and balloon emulsion chamber experiments . to this moment
we shall pay more attention in section 5 .
* gev . *
experimental data on differential proton spectrum for corresponding primary energies @xmath70 gev of sokol , mubee , jacee and runjob show good interconsistency , but from @xmath71 gev they become less definite for technical and natural ( low flux ) reasons .
experimental data on muon component have smaller errors and agree within @xmath28 . in this region along with direct measurements on the surface ( mars @xcite , nottingham @xcite , l3 @xcite ) there are results obtained at underground installations , in which the sea level muon spectrum is reconstructed from the ` depth - intensity ' curve and in other ways .
the latter are the data of the baksan neutrino observatory ( bno ) @xcite , artyomovsk scintillation detector @xcite , msu @xcite , frejus @xcite , lvd @xcite , macro @xcite , kgf @xcite and cosmoaleph @xcite .
most of them are grouped within fit limits , given by macro , and comprise a well correlated data set .
comparison of the obtained differential and integral muon spectra with these data does not support reasonable expectation ` pcr data fit@xmath72nuclear cascades@xmath72muon data fit ' .
the deviation of our results from the data of ` macro zone ' is about 3040% .
this fact requires revising all steps of our calculation in order to realize how real the problem is and what may be done to overcome it .
first of all , one should analyze an adequacy of nucleon cascades description . as it is shown in the previous section
, our computations do not contain methodical errors , and total statistical + interpolation + integration error is less , than 5% . from figure [ int ] it
is also seen , that the maximal difference @xmath73 in predictions of the qgsjet and venus hadronic interaction models is also practically out of significance in this situation . since the venus model was initially chosen by us as the model , providing one of the largest muon numbers in shower ( see @xcite ) , hence 10% should be regarded as the utmost possible increase of muon intensity for any interaction model in comparison with the qgsjet . in this connection the results of the earlier works , not indicating any problems with the description of the muon data , deserve some closer consideration .
we analyze here five works @xcite .
the collation of the pcr spectra with the experimental data , performed in section 2 , allows to split these papers into 2 groups : papers with excessive @xcite and appropriate @xcite fluxes . in order to reveal differences in hadronic interactions and calculation technique ,
we have made computations of muon fluxes for the same pcr spectra , as used in the discussed works .
results are presented in figure [ comp ] . from this comparison
it becomes evident , why no muon deficit was discovered earlier .
in addition to overvalued nucleon fluxes , in the papers from the first group the muon yield is larger or , at least , comparable with that of qgsjet and venus .
muon production in the papers with appropriate primary fluxes is @xmath74 higher , than obtained by us , and this difference is sufficient to mask the muon shortage . besides , pcr flux from the work @xcite becomes excessive from @xmath75 tev , as we have already pointed in section 2
. the data in respect of muon production of the latter two papers closely agree , apparently because in both of these works essentially the same authors participated . in @xcite also an explanation of relatively large muon yield may be found . in this calculations , for simulation of hadronic interactions
target code had been applied , but with the enhancement of pion and kaon production .
this resulted in increase of contribution of kaon decays to muon flux with @xmath76 tev up to a very large value of 50% .
of course , uncertainties in description of hadronic interaction prevail , but , as our calculations show , discrepancies in muon yield at sea level between the widely recognized and extensively tested models , included in corsika , are at the level of @xmath47 .
these models are evidently free from many drawbacks , peculiar to the models from the discussed works , and more thoroughly developed .
hence , likely the largest part of the discrepancy should be attributed to underestimation of primary nucleon flux , but not to the incorrectness of the simulation of the nuclear cascades in the atmosphere .
there is no need to repeat arguments from section 2 in favour of adequacy our pcr model , since it may be regarded just as a valid representation of the current experimental situation .
we should only repeat , that our proton spectrum does not present a fit of the experimental data for @xmath77 gev , but even on @xmath220% overestimates them .
this was done to smooth the discussed problem .
if directly input in calculations fits of primary h and he spectra , obtained by jacee , with addition of 10% contribution of heavier nuclei , or ` standard ' spectrum , proposed by gaisser and honda , then resulting integral muon flux will be even lower , than ours ( see table 2 ) . there is a possibility to compensate some of the muon deficit by increase of intensity of heavier nuclei in our model . but to get 10% addition to muon flux it is necessary to rise helium flux 50% or to double aggregate flux of cno , ne - si and fe groups .
this mechanism is rather ineffective , because limitations , set by the data on all - particle spectrum thus would require to lower fluxes of other , lighter species .
summarizing , we see , that today for this energy region the situation is determined rather rigidly and the stated problem of muon deficit can not be resolved without some extraordinary assumptions .
necessity in rise of all pcr spectra as high , as possible to diminish the muon shortage , indicates , that measurements with emulsion chambers systematically understate fluxes of the primaries . a large spread in the data on pcr nuclei with @xmath78
provides some more ground for such speculations and demonstrates necessity in further improvement of experimental technique .
this question is closely related to the appropriate description of the characteristics of nuclear interactions and in more details is considered in section 5 .
[ other_spectra ] cccccccccc & + & + & & & & & & & & & + @xmath79/i_\mu$]@xcite&@xmath80&@xmath80&0.98&1.07&1.12&1.11&1.04&0.94&0.79 + @xmath79/i_\mu$]@xcite&1.02&1.08&1.12&1.13&1.10&1.03&0.89&0.75&0.58 + * gev . *
analysis of this interval in the present situation looks vain for the following reasons : \i .
obviously , muon deficit takes place for these energies too , and its nature and quantity are unclear ; \ii .
there is no experimental data on chemical composition of pcr spectrum for energies higher than @xmath68 gev , that is why , as in our case , one has to use only theoretical guidelines , describing the origin and transport of cosmic rays in interstellar medium .
behavior of hadron - nucleus cross - sections for @xmath81 gev , despite the decrease of divergence between predictions of different models , remains also not enough studied , that results in additional uncertainty in calculations of the muon spectrum .
all muon component measurements for the given energies are obtained from results of underground experiments and contain quite large systematic errors , that causes their poor mutual agreement .
dominant sources of the mentioned errors are such factors as incomplete information on chemical composition of overburden rock and necessity to input in calculation a ratio @xmath82 of prompt muons to @xmath83-muons as a function of energy , which is not only unknown , but should be determined from the experimental data .
taking aforesaid into account , it seems impossible to make any definite conclusions on behaviour of pcr spectrum and on prompt muon contribution on the basis of sea level data in this energy region .
though from the comparison with the experimental data the vfgs ( volkova , fulgione , galeotti and saavedra ) model @xcite of charm generation looks as the most preferable , however , here it is needless to stress all unreliability of this deduction .
in this section we intend to discuss some problems , that are inherent to the study of high - energy cosmic rays spectra with the application of the emulsion chambers ( ec ) .
in addition to all factors , listed below , one can easily add some more , that experimenters encounter during processing of the ec data , including , for example , not negligible experimental bias .
we are mostly concentrated at the factors , depending on the description of nuclear interactions , and in this aspect the situation may eventually turn out similar to that with the calculations of sea level muons .
the necessity of further improvements in the experimental technique is clearly recognized in connection with the serious disagreements between the current data on different groups of nuclei , and short history of space and balloon ec experiments gives many examples of how the reported data changed with the gain of experience and statistics .
further investigations will inevitably provide information , that improves our knowledge , and , possibly , will differ from it .
the most detailed description of the experimental technique is presented by the runjob collaboration @xcite and we shall mostly rely on their data , bearing in mind that the procedure is generally the same in all ec experiments ( mubee , jacee and runjob ) . the key moment lies in the determination of the energy , transferred to the electromagnetic component @xmath84 in the cascades , initiated by the interaction of a primary particle within ec .
this energy is related to the initial energy @xmath85 via partial inelasticity coefficient @xmath86 ( we shall denote so @xmath87 to avoid further confusion with the spectral index @xmath30 ) . the spectrum of primaries is obtained from the spectrum of electromagnetic cascades ( emc ) with a simple shift in the energy scale by the value @xmath88^{-1/\gamma},\ ] ] here @xmath89 is a distribution function of @xmath86 and @xmath30 is the spectral index of primary spectrum @xmath90 .
the given statement , at the least dating back to as early as 1962 @xcite , holds true , provided @xmath89 does not depend on @xmath85 in a wide energy range .
usually independence of @xmath91 on energy is argued from the evidence , that total and partial @xmath92 inelasticities are constant . but this may prove to be incorrect ( see , e.g @xcite ) , since @xmath93 essentially depends on the contribution of the large values of @xmath86 to the distribution @xmath89 , and even though @xmath92 is constant , ` tail ' of @xmath89 may change with @xmath85 . besides , large body of information on behaviour of the total inelasticity , presented in the literature , points at its slow growth with energy .
another important moment , is the difference in @xmath92 between predictions of various interaction models . in comparison with fritiof , applied by runjob
, venus gives @xmath94 lower value of partial inelasticitiy @xmath92 in a single _ p - a _ , he_-a _ collisions @xcite ( other experiments do not provide information neither on the code used , nor on the single interaction characteristics ) .
note , that 10% correction to the energy conversion from @xmath84 to @xmath85 for power pcr flux @xmath95 is equivalent to @xmath96 correction to the intensity .
thus , at first sight , the use of the venus provides not only the largest number of muons at sea level , but , as well may lead to an increase of the measured proton and helium fluxes .
in fact , the estimation of the influence of the discussed effect on the results of ec measurements is much more complicated .
firstly , one needs to evaluate total energy , transferred to emc in successive interactions of secondary particles , i.e. from the whole shower , not from single interaction , and to obtain distribution function @xmath89 for this fraction @xmath86 of the initial energy .
secondly , this distribution for @xmath86 is rather broad , so the conversion to @xmath85 for individual shower is impossible and it is necessary to derive effective value of @xmath91 , allowing to get pcr spectrum from the emc one .
so , simple knowledge of differences in @xmath92 in the first interaction does not allow to make straightforward qualitative estimate of changes in the final energy shift value .
the analysis becomes even more complicated , if to account , that the choice of the interaction model strongly influences the previous steps of the experimental data processing .
namely , it affects the determination of @xmath84 and detection efficiency . for evaluation of both of these values a complete simulation of nuclear cascades in ec
is required .
for example , in the runjob experiment
the ` actual ' emc energy @xmath97 is obtained from the estimated with the @xmath30-ray core method one @xmath98 via direct monte - carlo simulation of showers with the fritiof code . in the jacee experiment @xcite determination of cascade energies also involves complete monte - carlo calculations of the transition curves , used for calibration of the @xmath84 , derived from the direct electron counting or from the x - ray film densitometry .
analogous calculations are required for evaluation of the detection efficiency ( see detailed description in @xcite ) .
thus , deviations between the interaction models in hadron - nucleus cross - sections , energy spectra , multiplicities and phase space distributions ( this is important for estimation of the most energetic in the lab system and back - scattered in cms numbers of gamma - quanta .
venus predicts the largest quantity of the latter , compared to the other models @xcite ) of seconday particles , in fluctuations of the energy , transferred to emc , etc .
, may significantly influence the interpretation of the ec measurements results .
the mentioned arguments give enough ground for performing of thorough analysis of sensitivity of the ec data to the use of various interaction models , which is still missing .
this as well would allow to make fully consistent estimations of the cosmic ray fluxes , i.e. with the application of the single model , both at the top of the atmosphere and at the sea level .
besides , it is possible that some data distortion may come from changes or anomalies in characteristics of hadronic interactions at very high - energies . at present , these questions are widely discussed in connection with the data of eas experiments , also employing ec technique
. another hypothetical reason , which may cause systematic errors in processing of the ec experiments data , could be the presence of sizable fraction of unusual component in pcr .
this may be regarded equivalent to the ` change ' in hadronic interactions and all aforesaid about that effect holds true in the given case . in this connection , rising interest to the primary antiproton component deserves closer attention .
ratio from the ground - based experiments : @xcite solid arrows ; dashed arrows : @xcite tibet , @xcite l3+c , @xcite macro . ] as it is appropriately supposed , antiprotons present natural compound of pcr , produced mostly in interactions of nuclei ( dominantly protons ) with interstellar medium . other , exotic @xmath99 generation mechanisms , result in softer spectra and they are out of interest for the energies , we are dealing with ( for more information , see references in caption to figure [ antip ] and _ e.g. _
the experimental study of @xmath99 component began in 1970 s from the first balloon measurements ( golden et al .
@xcite and bogomolov et al .
@xcite ) , and today the direct data cover region from few mev to approximately 40 gev ( figure [ antip ] ) .
higher energy region is scarcely studied and is marked by few indirect upper estimates of @xmath100 ratio .
one of them by stephens ( 1985 ) @xcite is derived from the analysis of sea level muon charge ratio .
soon we shall present elsewhere such calculations with the employment of improved since then knowledge on pcr fluxes , hadronic interactions and muon charge ratio .
other three @xmath100 estimates by tibet @xcite , l3+c @xcite and macro @xcite collaborations are made with the use of the moon and the sun shadow effects .
the idea of this approach is the following : positively and negatively charged particles are deflected by the earth s magnetic field to the opposite sides , and when the sun or the moon , during their transit over detector , block particles , deficit of secondary particles should be observed , if antimatter presents in pcr along with the matter , at the both opposite sides of the real sun / moon positions .
since these experiments failed to detect ` antimatter shadow ' , they could report only upper limits of @xmath100 ratio to be equal 22% at 10 tev @xcite , 27.5% at 20 tev @xcite and 15% for muon sample with @xmath101 gev ( primary energy @xmath12 tev ) @xcite .
theoretical predictions are not able to clarify the situation , since there is no generally recognized propagation theory , reproducing satisfactory even low energy data on @xmath100 ratio together with all totality of the data on elemental nuclei abundances , secondary to primary nuclei ratios , and fluxes of positrons and gammas .
standard diffusion / reacceleration models fail to do this , requiring introduction of specific assumptions , generally working for description of some kind of observations , but leading to contradictions in other cases . among such assumptions
one may readily call necessity in ` artificial breaks in diffusion coefficients and in the primary injection spectrum ' @xcite to match both @xmath100 and secondary to primary nuclei ratios , suggestion about harder interstellar proton and/or electron spectra to explain well known ` gev excess ' in diffuse gamma ray flux , etc .
( for more information see , _
@xcite and references therein ) . the detailed discussion of these problems is beyond the scope of our paper and we shall return to it elsewhere . in any case , it is clear , that there is no firm ground to state neither the presence of significant fraction of @xmath99 s for high energies , nor the reverse and that actual @xmath100 ratio is still to be determined via direct and various kinds of indirect observations .
thus , one can not completely exclude probability , that among events , identified in ec as protons , sizable fraction of antiprotons presents .
as it is already said above , it is difficult to estimate the influence of the differences between _ p - a _ and _ @xmath99-a _ interaction characteristics on the results of pcr measurements with ec .
it is known that antiprotons deposit more energy into @xmath30 s , than @xmath102 in a single interaction @xcite , but , we should repeat , the complete analysis of changes , produced in the measured @xmath103 flux requires to perform evaluation of many parameters , that is possible to do only via direct monte - carlo simulations with respect to the specific experimental conditions .
we have made similar estimates for the total muon flux at sea level .
as we have elucidated , actually all primary nucleons in our calculations are protons . replacing them by antiprotons , we have computed integral muon flux for @xmath104 gev and have not found any noticeable changes , in comparison with the calculations for protons . in other words , total vertical sea level muon flux turned out to be insensitive to the fraction of antiprotons in pcr and , possibly , the same may happen with the ec results .
of course , this is not true for @xmath105 and @xmath106 spectra , and @xmath4 ratio , their calculations may help to determine presence of @xmath99 s in pcr . in conclusion
, we would like to present an estimation of the changes in nucleon flux , required to match the muon experimental data . in figure [ nucleon_shift ]
we show , along with initially applied spectrum , its part for @xmath107 tev 8% shifted toward the higher energies . from the magnet spectrometers region to the ec energies we use an interpolation .
so , muon spectra , given in figures [ adiff],[aint ] , are calculated with the use of the combined nucleon flux , shown in figure [ nucleon_shift ] with open circles .
these spectra still present rather a lower estimate of the muon data and may be approximated via the following formula ( energy is in gev , @xmath108 gev ) @xmath109 with parameters , given in table 3 .
the used nucleon spectrum is just a model spectrum , since in practice , correction to the pcr flux hardly may be constant and should depend both on the energy and atomic mass of the projectile .
but it is easy to note , that required 8% shift is significantly smaller , than the quoted accuracy of the energy determination in the ec experiments and , on the other hand , to some part it may be attributed to the effects , discussed in this section . , [ ai nt ] . ] with open circles .
other designations are the same as in figure [ diff].,title="fig : " ] with open circles .
other designations are the same as in figure [ diff].,title="fig : " ] with open circles .
other designations are the same as in figure [ diff].,title="fig : " ] with open circles .
other designations are the same as in figure [ diff].,title="fig : " ] [ param ] @l*5l + + & & & & & + @xmath110100&@xmath111&@xmath112&@xmath113&@xmath114&@xmath115 + @xmath116&@xmath117&@xmath118&@xmath119&@xmath120&@xmath121 + + + & & & & & + @xmath110100&@xmath122&&@xmath123&@xmath124&@xmath125 + @xmath116&@xmath126&&@xmath127&@xmath128&@xmath129 +
for the first time a complete set of the most recent experimental data on pcr nuclei spectra has been used in calculations of muon flux at sea level in a wide energy range from 1 to @xmath130 gev .
the adequate description of cascade processes in the atmosphere has been provided by the application of widely approved code corsika along with qgsjet and venus models .
this makes our results free of the possible errors , caused by a simplified approach to description of hadronic interactions .
the main and a rather unexpected conclusion , which our investigation has led to , is that the data on primaries conform to the muon data only for rather low energies up to several tens of gev . from these energies
a deficit of muons becomes evident , corresponding to a deficit of primary nucleons with @xmath131 tev . in the earlier works of the other authors
this problem was masked by the uncertainties of the data on pcr fluxes and hadronic interactions , existed those days . in a surprising way , refinement and gain of the information
have not approached us to the possibility of the more accurate muon flux derivation , but , on the contrary , have revealed that our today s , improved notions are in some part erroneous .
this , by the way , makes obsolete attempts to standardize the lepton fluxes calculations in respect to the primary spectra , since , as our consideration shows , in the large part the problem lies in underestimation of pcr intensity . in our point of view , to remedy the situation with the least consequences , the re - examination of the ec experimental results with the application of different interaction codes is required .
as we have noted , in this connection venus possesses with two positive features . in comparison with the other models it predicts a larger number of muons at sea level , and in the single hadronic interaction , less fraction of the energy
is deposited into electromagnetic component , but the influence of the latter effect on the measurements results is not so simply to analyze , since it may be compensated at the different stages of the ec data processing .
another fact , that must be taken into account , is the possible presence in pcr of sizable fraction of antiprotons .
if the emulsion chambers data would prove to be insensitive to the given effects , only then the problem should be attributed to some uncertainties or anomalies in the high - energy hadronic interactions . in any case ,
our calculations have demonstrated , that today the fit of the sea level muon data can not be derived directly from the pcr data , and the reasons , leading to this puzzle , look unclear .
we thank the two anonymous referees for their constructive comments , which helped us significantly improve the paper .
this work is supported in part by the ur grants no . 02.01.014 and no . 02.01.001 .
10 volkova l v , zatsepin g t and kuzmichev l a 1979 _ yad .
fiz . _ * 29 * 1252 ( in russian ) knapp j , heck d and schatz g 1996 _ comparison of hadronic interaction models used in air shower simulations and of their influence on shower development and observables _ ( forschungszentrum karlsruhe ) * fzka 5828 * | for the first time a complete set of the most recent direct data on primary cosmic ray spectra is used as input into calculations of muon flux at sea level in wide energy range @xmath0 gev .
computations have been performed with the corsika / qgsjet and corsika / venus codes .
the comparison of the obtained muon intensity with the data of muon experiments shows , that measurements of primary nuclei spectra conform to sea level muon data only up to several tens of gev and result in essential deficit of muons at higher energies . as it follows from our examination , uncertainties in muon flux measurements and in the description of nuclear cascades development
are not suitable to explain this contradiction , and the only remaining factor , leading to this situation , is underestimation of primary light nuclei fluxes .
we have considered systematic effects , that may distort the results of the primary cosmic ray measurements with the application of the emulsion chambers .
we suggest , that re - examination of these measurements is required with the employment of different hadronic interaction models .
also , in our point of view , it is necessary to perform estimates of possible influence of the fact , that sizable fraction of events , identified as protons , actually are antiprotons .
study of these cosmic ray component begins to attract much attention , but today nothing definite is known for the energies @xmath1 gev . in any case , to realize whether the mentioned , or some other reasons are the sources of disagreement of the data on primaries with the data on muons , the indicated effects should be thoroughly analyzed . _
accepted for publication in _ |
during the past several decades there have been significant advances in the accuracy of satellite - based navigation and time transfer systems @xcite .
the most recent example of such a system is the global positioning system ( gps ) , which is run by the u. s. department of defense @xcite .
the gps is based on a network of 24 atomic clocks in space , from which a user can receive signals and compute his position and time .
the order of magnitude of the accuracy of the gps currently is ten meters and twenty nanoseconds @xcite .
future navigation and time transfer systems are being contemplated that will have several orders of magnitude better position accuracy and sub - nanosecond time accuracy @xcite . over satellite to ground distances , precise measurements must be interpreted within the framework of a relativistic theory @xcite .
the design of high - accuracy navigation and time transfer systems depends on detailed modeling of the measurement process .
the comparison of experiment with theory in general relativity is more subtle than in special relativity or other areas of physics for two reasons .
first , due to the possibility of making coordinate gauge transformations , often it is not clear what is a measurable quantity and what is an artifact of the coordinate system @xcite .
second , all measurements must in - principle be interpreted in the comoving frame of the experimental apparatus @xcite . in the case of the gps , even theoretical relativistic treatments
do not interpret the measured observable in the comoving frame of the apparatus , see for example ref .
@xcite .
the principle reason for investigating in detail relativistic effects is to improve the current accuracy of gps and to create future time transfer and navigation systems that have several orders of magnitude better accuracy . at the present time , it is well - known that small anomalies exist in position and time computed from gps data .
the origin of these anomalies is not understood .
in particular , gps time transfer data from the naval observatory indicates that gps time is periodic with respect to the master clock , which is the most accurate source of official time for the department of defense @xcite .
furthermore , other anomalies have been found in air force monitor station data that are not understood at present @xcite .
the work below is a first step in revisiting from a first principles approach some of the physics that has gone into the design of the gps .
a theoretical framework , known as the tetrad formalism , has been developed for relating measurements to relativistic theory@xcite .
however , due to a historical lack of accurate measurements over satellite to ground distances , in most cases it has not been necessary to apply the tetrad formalism in the area of satellite - based systems .
this situation is rapidly changing due to improved measurement technology@xcite . in the tetrad formalism , measurements are interpreted as projections of tensors on the tetrad basis vectors .
consequently , these projections are invariant quantities ( scalars ) under transformation of the space - time coordinates .
however , these projections do depend on the world line of the observer and the choice of local cartesian axes used by the observer@xcite .
the need for the tetrad formalism to relate experiment to theory , and the problem of measurable quantities in general relativity , are well discussed by synge @xcite , soffel @xcite and brumberg @xcite .
the tetrad formalism was initially investigated for the case of inertial observers , which move on geodesics@xcite . however ,
many users of geodetic satellite systems , are terrestrially based observers , or are based in non - inertial platforms , such as missiles .
the theory for the case of non - inertial observers has been investigated by synge@xcite , who considered the case of non - rotating observers moving along a time - like world line , and by others@xcite , who considered accelerated , rotating observers . for arbitrary observer motions ,
the effects are indeed very complicated , and the general theory gives limited physical insight .
for this reason , in this work , i work out in detail a simple model problem .
i treat the particular case of an observer moving in a circle in minkowski space .
i use the tetrad formalism@xcite to obtain the transformation from an inertial frame in minkowski space , to the accelerated frame of reference of an observer moving in a circle .
the observer fermi - walker transports his tetrad basis vectors .
the case presented here corresponds to an observer kept in a circular trajectory by rockets , he is not on a geodesic .
the explicit results presented below form a particular case of the general results obtained by synge@xcite , specialized to zero gravitational field and circular motion .
the situation treated here also corresponds to a reference frame that has a constant orientation with respect to the distant stars@xcite , such as an earth - centered inertial reference frame used in gps satellite computations .
the results are given here as expansions in small velocity of the observer , compared with speed of light .
however , the coordinate transformation equations are not expanded in a power series in time , and hence , they are valid for arbitrary long times .
this work explores the apparent behavior of ideal clocks as observed from a fermi coordinate frame of reference of an observer moving in a circle in flat space - time .
the fermi coordinate frame is the closest type of non - rotating frame to that which is used in experiments and in gps . in particular ,
a fermi coordinate frame is a non - rotating frame of reference , where coriolis forces are absent , and light travels essentially along a straight line@xcite .
the observer s fermi frame in this work is analogous to the earth - centered inertial ( eci ) frame used in gps
. the eci frame moves with the earth but maintains its orientation with respect to the distant stars .
this work neglects all gravitational field effects that are significant for the actual behavior of gps clocks .
as such , this work treats only the kinematic effects associated with the circular motion of the fermi coordinate frame about a central point ( analogous to the sun s position ) . since gravitational fields are neglected in this work , the observer s tetrad
is fermi - walker transported along a time - like world line ( a circle in 3-d space ) . in the case of gps , the origin of the eci frame follows a geodesic path ( the earth s orbit ) .
therefore , the observer s fermi frame in this work is an accelerated , non - rotating frame , while the eci frame is in free fall . as remarked above
, the effects on clocks are in general very complicated , even without the gravitational field .
for this reason , i work out in detail the specific case of an observer in circular motion in minkowski space - time , and i apply the results to parameters that are relevant to the gps .
consider a time - like world line @xmath0 of an observer moving in a circle in 3-dimensional ( 3-d ) space given by @xmath1 where @xmath2 , @xmath3 , @xmath4 , and @xmath5 are constants , @xmath6 is a parameter along the world line , and @xmath7 is the speed of light in an inertial frame in vacuum .
the observer travels in a circle in the @xmath8 plane .
see figure [ observermotionfigure ] .
along his world line , the observer travels with a four velocity @xmath9 and has an acceleration @xmath10 where @xmath11 is the affine connection . from the normalization of the 4-velocity , @xmath12
, i obtain the relation between the arc length , @xmath13 , where @xmath14 is the proper time , and the parameter @xmath6 , @xmath15 where @xmath16 .
the parameter @xmath17 is the dimensionless velocity , given by @xmath18 i use the convention that roman indices take the values @xmath19 and greek indices take values @xmath20 , see the appendix for the conventions used in this work . in minkowski space ,
the affine connection components all vanish , @xmath21 .
hence , the explicit expressions for the 4-velocity and acceleration are : @xmath22 @xmath23 as the observer moves on the time - like world line @xmath0 , he carries with him an ideal clock and three gyroscopes .
i use the notation @xmath24 for minkowski space coordinates . at some initial
coordinate time @xmath25 , the observer is at point @xmath26 at proper time @xmath27 , corresponding to @xmath28 . on his world line
, the observer carries with him three orthonormal basis vectors , @xmath29 , where @xmath20 , labels the vectors and @xmath19 labels the components of these vectors in minkowski coordinates .
these vectors form the basis for his measurements @xcite .
see figure [ observerworldlinefigure ] .
the orientation of each basis vector is held fixed with respect to each of the gyroscope s axis of rotation @xcite .
the fourth basis vector is taken to be the observer 4-velocity , @xmath30 .
the four unit vectors @xmath31 form our tetrad@xcite , an orthonormal set at @xmath26 , which satisfy @xmath32 where the @xmath33 if @xmath34 and zero otherwise . at a later time
@xmath13 , the observer is at a point @xmath35 .
the observer s orthonormal set of basis vectors are related to his tetrad basis at @xmath26 , by fermi - walker transport .
fermi - walker transport preserves the lengths and relative angles of the transported vectors .
furthermore , fermi - walker transport of basis vectors along the observer s world line corresponds to keeping the vectors as parallel as possible by keeping their orientation constant relative to the spin axes of the gyroscopes @xcite .
for a given vector , with contravariant components @xmath36 in minkowski space , its components at @xmath35 are related to its components at @xmath26 by the fermi - walker transport differential equations @xcite @xmath37 where @xmath38 when we use eq .
( [ fwtensor ] ) to transport a vector @xmath36 that is orthogonal to the 4-velocity , @xmath39 , the second term in eq .
( [ fwtensor ] ) does not contribute .
we refer to transport of such space - like basis vectors as fermi transport , and @xmath40 . for an arbitrary vector @xmath36 perpendicular to the 4-velocity , the explicit form of the fermi transport differential equations for the world line in eqs .
( [ worldline0])([worldline3 ] ) is @xmath41 where the components @xmath36 are functions of @xmath42 and the dimensionless proper time is given by @xmath43 the differential equations in eqs .
( [ fdiffeq0])-([fdiffeq3 ] ) are nothing more than the equations that describe the thomas precession of an electron s spin vector as it moves in a circular orbit around the nucleus @xcite .
the solution of eq.([fdiffeq0])([fdiffeq3 ] ) is given by @xcite @xmath44 where @xmath45 , @xmath46 , @xmath47 and @xmath48 are integration constants .
at the point @xmath26 , the tetrad basis vectors satisfy eq.([initialtetrad ] ) .
these equations constitute 12 relations ( since eq .
( [ initialtetrad ] ) is symmetric in @xmath46 and @xmath48 ) for 16 components @xmath49 .
however , the zeroth member of the tetrad is the 4-velocity , which is assumed known @xmath50 and is given by eq .
( [ fourvelocity2 ] ) . equation ( [ initialtetrad ] ) can be rewritten as two equations @xmath51 @xmath52 substituting the explicit form of the minkowski metric ( see the appendix ) into eq .
( [ zerothorthog ] ) , gives a condition on the zeroth components of a tetrad vector if the spatial components are known ( greek indices take values @xmath46,@xmath531,2,3 ) @xmath54 then use of eq .
( [ zerothlambda2 ] ) to eliminate the time components @xmath55 from eq .
( [ spaceorthog ] ) , leads to an orthogonality relation for only the spatial components of the tetrad vectors @xmath56 the general idea is to solve eq .
( [ lambdaspaceorthog ] ) for the spatial components of the tetrad at @xmath28 , and then to substitute the spatial components into eq .
( [ zerothlambda2 ] ) to obtain the time components of each tetrad vector .
thus , i obtain the tetrad vectors at the initial time corresponding to @xmath28 . for each tetrad vector ,
these components are then used as initial conditions in eq .
( [ fcompsolution ] ) to determine the tetrad components at point @xmath35 at finite @xmath57 .
the exact solution of eq . ( [ lambdaspaceorthog ] ) can be obtained by noting that for @xmath58 , the spatial components of the tetrad vectors are orthonormal , and the solution is given by @xmath59 i assume that the tetrad unit vectors are approximately oriented parallel to the observer s local cartesian @xmath60 , @xmath61 and @xmath62 axes . for small @xmath63 , the solution of eq.([lambdaspaceorthog ] ) can be obtained as a triple power series in @xmath64 , by iteration , starting with eq.([initiallambdav ] ) as the first approximation .
having solved for the tetrad components as a power series , it is easy to guess the exact solution to be @xmath65 where the components of dimensionless velocity @xmath64 are given in eq .
( [ fourvelocity2 ] ) .
equations ( [ initialtetrad2a])([initialtetrad2 ] ) give the components of the observer s tetrad basis vectors at point @xmath26 . at point @xmath35 ,
the arc length is @xmath66 , and eqs.([initialtetrad2a])([initialtetrad2 ] ) give the initial components for each tetrad vector .
these initial components are used in the general solution of a fermi - walker transported vector , given in eq .
( [ fcompsolution ] ) .
the tetrad components at point @xmath35 at finite @xmath57 are given by @xmath67 equations ( [ lam0])([lam3 ] ) are valid for all times @xmath57 because they are exact solutions of eqs.([fdiffeq0])([fdiffeq3 ] ) and satisfy the orthogonality relations in eq .
( [ initialtetrad ] ) .
fermi coordinates are defined by the geometric construction shown in figure [ observerworldlinefigure ] . every event @xmath68 in space - time
has coordinates @xmath69 in the global inertial coordinate system .
according to the observer moving in circular motion along a time - like world line , the same event has the fermi coordinates @xmath70 , @xmath71 .
the first fermi coordinate , @xmath72 , is just the proper time ( in units of length ) associated with the event @xmath68 . the proper time for @xmath68
is defined as the value of arc length @xmath57 in space - time at which a space - like geodesic from point @xmath35 ( which defines @xmath57 ) and passes through event @xmath68 .
the definition of @xmath57 requires the tangent vector of this space - like geodesic , @xmath73 , be orthogonal to the observer 4-velocity at @xmath35 , @xmath74 for the simple case of minkowski space , the geodesic is simply a straight line connecting the points @xmath75 @xmath76 where @xmath77 and @xmath69 are the coordinates of @xmath35 and @xmath68 , respectively , in the global inertial frame , and @xmath78 is a normalization constant that makes @xmath73 a unit vector .
the orthogonality condition in eq .
( [ normal1 ] ) is @xmath79 and gives an implicit equation for @xmath57 for a given @xmath68 .
this orthogonality condition gives the first fermi coordinate of the point @xmath68 @xmath80 and eq .
( [ normalcondition ] ) gives @xmath57 as an implicit equation : @xmath81 \label{sdef}\ ] ] in the limit of small speeds of the observer , @xmath82 , eq.([sdef ] ) gives @xmath83 .
for @xmath84 , eq.([sdef ] ) can be solved for @xmath57 by iteration , resulting in a power series in @xmath17 @xmath85 \,\nu - \frac{1}{2}\delta x^0 \,\nu^2 \nonumber \\ & & -\frac{1}{2a } \left [ -a + 2\,\delta x^1 \ , \cos ( \frac{\delta x^0 \,\omega } { c } ) + 2\ , \delta x^2 \ , \sin ( \frac{\delta x^0 \,\omega } { c } ) \right ] \ , \left [ \delta x^2 \ , \cos ( \frac{\delta x^0 \,\omega } { c } ) - \delta x^1 \ , \sin ( \frac{\delta x^0 \,\omega } { c } ) \right ] \ , { \nu } ^3 + o(\nu ) ^4 \label{sexplicit}\end{aligned}\ ] ] where i use the notation @xmath86 for future reference , i label the coordinates of @xmath35 , @xmath68 , and @xmath26 in the global inertial frame by @xmath87 , @xmath88 , and @xmath89 .
the contravariant spatial fermi coordinates , @xmath90 , @xmath20 , are defined to be@xcite @xmath91 where i used the minkowski definition of space - time distance , @xmath92 , along the space - like geodesic between @xmath35 and @xmath68 @xmath93 where @xmath94 is the minkowski metric and @xmath95 is an invariant matrix , which is also equal to the minkowski metric , given in the appendix .
the parameter @xmath57 in eq.([fccoorddef ] ) is a function of @xmath68 coordinates @xmath96 , so @xmath57 must be eliminated using eq.([sexplicit ] ) , so that @xmath90 are explicit functions of the global inertial coordinates @xmath96 .
i have calculated this transformation correct to fourth order in @xmath17 . however , since the expressions are complicated , i write them here correct only to third order in @xmath17 : @xmath97 \ , \nu^2 \nonumber \\ & & + \frac{1}{2a } \delta x^0 \ , \left [ - \delta x^2 + a \,\sin ( \frac{\delta x^0\,\omega } { c } ) \right ] \ , \nu^3 + o(\nu ) ^4 \label{x1 } \\
x^{(2 ) } & = & \delta x^2 - a\,\sin ( \frac{\delta x^0 \,\omega } { c } ) + \frac{1}{4}\left [ \delta x^2 + \delta x^2 \ , \cos ( \frac{2\,\delta x^0\,\omega } { c } ) - \delta x^1 \ , \sin ( \frac{2\,\delta x^0 \,\omega } { c } ) \right ] \ , \nu^2 \nonumber \\ & & + \frac{1}{2 a } \ , \delta x^0 \ , \left [ \delta x^1 - a\,\cos ( \frac{\delta x^0\,\omega } { c } ) \right ] \,\nu^3 + o(\nu ) ^4 \label{x2 } \\
x^{(3 ) } & = & \delta x^3 \label{x3}\end{aligned}\ ] ] the inverse transformation can be obtained from eqs.([x1])([x3 ] ) by iteration resulting in : @xmath98 \nu + \frac{1}{2}\ , x^{(0 ) } \nu^2 + o(\nu ) ^3 \\ % x^{1 \ , \prime } - x^1_0 & = & x^{(1 ) } + a\,\cos ( \frac{\omega x^{(0)}}{c } ) + \frac{1}{4 } \left [ x^{(1 ) } - x^{(1 ) } \,\cos ( \frac{2\ , \omega x^{0)}}{c } ) - x^{(2 ) } \,\sin ( \frac{2\,\omega x^{(0 ) } } { c } ) \right ] \ , \nu^2 + o(\nu ) ^3 \\ % x^{2 \ , \prime } - x^2_0 & = & x^{(2 ) } + a\,\sin ( \frac{\omega x^{(0)}}{c } ) + \frac{1}{4 } \left [ x^{(2 ) } + x^{(2 ) } \,\cos ( \frac{2\ , \omega x^{(0)}}{c } ) - x^{(1 ) } \,\sin ( \frac{2\,\omega x^{(0 ) } } { c } ) \right ] \ , \nu^2 + o(\nu ) ^3 \\ % x^{3 \ , \prime } - x^3_0 & = & x^{(3 ) } \label{inversetransformation}\end{aligned}\ ] ]
the line element in the fermi coordinate system of the observer is written as @xmath99 where @xmath100 are the metric tensor components when the @xmath101 are used as coordinates . a direct calculation of the metric tensor components from the tensor transformation rule @xmath102 using the transformation eq .
( [ x1])([x3 ] ) and the minkowski metric for @xmath94 , gives @xmath103 and @xmath104 where @xmath105 is given by @xmath106 \
, \nu^2 + o(\nu^4 ) \\ \label{tildezeta}\ ] ] using the transformation eqs .
( [ x1])([x3 ] ) to express @xmath105 as a function of the fermi coordinates , i obtain : @xmath107 \ , \nu^2 + o(\nu^4 ) \label{zetadef}\ ] ] the result in eq .
( [ fcmetric ] ) agrees with synge s general theory@xcite of fermi coordinates specialized to flat space .
synge shows that @xmath108 to third order in @xmath17 , eq .
( [ zetadef ] ) agrees with the general theory given by synge@xcite .
the quantities @xmath109 are the covariant fermi coordinates and @xmath110 are the components of the observer s 4-acceleration in the fermi coordinate system .
the 4-acceleration , @xmath111 , @xmath19 , is related to the ordinary acceleration , @xmath112 , @xmath113 , by @xmath114 using this relation , @xmath105 can be written in terms of the global inertial coordinates of @xmath35 and @xmath68 as @xmath115 for the circular motion treated here , from eq .
( [ zetadef ] ) , the three spatial components of the observer s acceleration in fermi coordinates are given by @xmath116 where we have dropped fourth - order terms in @xmath17 . to third order in @xmath17 , the observer s acceleration is just the classical centripetal acceleration for an observer moving in a circle .
the quantity @xmath117 is a potential ( analogous to a gravitational potential ) that determines the rate of proper time with respect to coordinate time in fermi coordinates .
proper time is kept by an ideal clock .
coordinate time depends on the definition of the reference frame and coordinates used within that frame . the quantity @xmath105 , given in eq .
( [ zetadef2 ] ) , depends on acceleration @xmath112 , and the difference in coordinates , @xmath118 , from the spatial origin of the fermi coordinates . away from the spatial origin of fermi
coordinates the acceleration produces an effective gravitational potential field , @xmath119 , which leads to a change in the relation of coordinate time to proper time , through the metric components @xmath100 .
the significance of proper time is that it is the quantity that ideal clocks keep , that it may be related to time on real clocks , and that it is closely related to measurable quantities .
the significance of coordinate time is that it enters into the theory . in order to analyze experiments with clocks
, we must relate what is measured , i. e. proper time intervals between space - time events , to coordinate positions and times of these events .
consider an ideal clock whose world line is given in fermi coordinates by @xmath120 . at coordinate times @xmath121 , and @xmath122 ,
the clock is at spatial positions @xmath123 , and @xmath124 , respectively .
the elapsed proper time between these two events , @xmath125 , is given by the integral along the world line of the clock @xmath126 for the case of a stationary clock , located at constant fermi coordinate position @xmath90 , the relation between proper time @xmath14 and coordinate time , @xmath127 , is given by taking @xmath128 in eq .
( [ fermicoordlineelement ] ) . the _ rate _ of proper time with respect to coordinate time is then given by @xmath129 equation ( [ stationaryclockrate ] ) gives the well - known result that the rate of proper time depends on the location of the clock @xcite .
for the simple case of circular motion of the observer with @xmath105 given by eq .
( [ zetadef ] ) , substitution in eq .
( [ stationaryclockrate ] ) and integration from @xmath121 to @xmath122 gives @xmath130 -x^{(2)}\left [ \cos \left ( \frac{\omega}{c}x^{(0)}_b \right ) - \cos \left ( \frac{\omega}{c}x^{(0)}_a \right ) \right ] \right\ } \label{stationaryclock}\end{aligned}\ ] ] where i have dropped terms of @xmath131 .
equation ( [ stationaryclock ] ) shows that there is a complicated relation between elapsed proper time on a clock located at fermi coordinates @xmath132 , and fermi coordinate time , @xmath127 .
for example , a clock on the @xmath133 axis ( with @xmath134 ) has a periodic time with respect to fermi coordinate time , given by @xmath135 .
the initial conditions in eq.([stationaryclock ] ) determine complicated phase relations between proper time and coordinate time .
now consider an ideal clock that is a distance @xmath136 from the origin of fermi coordinates and the clock executes circular motion in fermi coordinates in the @xmath137 plane .
i take the world line for this motion to be given by @xmath138 the clock moves in the @xmath133-@xmath139 plane , which coincides with the plane defined by the circular motion of the origin of fermi coordinates .
see figure [ observerworldlinefigure ] .
the situation considered here is a simplification of the problem of a gps satellite ( with a clock ) orbiting the earth , which is itself orbiting the sun .
( in the case of gps , the satellites are in orbits that are inclined at 55@xmath140 with respect to the earth s equator . )
however , as mentioned previously , in the case of the earth and gps satellite , both travel on geodesics . since we are neglecting the effects of gravity in the problem considered here , the fermi observer and orbiting clock are both fermi - walker transported , e. g. , with the aid of rockets .
substitution of the satellite clock s world line into eq.([worldlineintegral ] ) , leads to the following relation between elapsed proper time observed on the clock and fermi coordinate time @xmath141 ( x^{(0)}_b - x^{(0)}_a ) -\frac{a \ , b}{c}\frac{\omega^2}{\omega-\omega}\left [ \sin \left(\frac{\omega-\omega}{c}x^{(0)}_b \right ) - \sin\left(\frac{\omega-\omega}{c}x^{(0)}_a \right ) \right ] \label{orbitingclock1}\end{aligned}\ ] ] the _ rate _ of proper time with respect to coordinate time is @xmath142 the proper time has a constant rate offset from coordinate time , represented by the first term in eq .
( [ orbitingclock1 ] ) .
in addition , the proper time has a periodic component with respect to coordinate time , given by the second term in . to demonstrate the magnitude of these effects , i use parameters that correspond to earth - sun distance and gps satellites .
i compute the magnitudes of the following terms with the values shown in table [ constantvalues ] : @xmath143 @xmath144 equation ( [ timedilationrateoffset ] ) gives the constant rate offset of the moving clock with respect to fermi coordinate time , due to time dilation . in gps
, this term leads to a net slowing down of gps clocks by approximately 7 @xmath145s per day@xcite , which results from multiplying the value given in eq .
( [ timedilationoffset ] ) for a single revolution by a factor of 2 ( since the gps satellites make approximately two revolutions per earth day . ) the second term in eq .
( [ orbitingclock1 ] ) is a periodicity of the proper time with respect to fermi coordinate time with amplitude @xmath146 which corresponds to an amplitude in the rate @xmath147 the proper time given by eq .
( [ orbitingclock1 ] ) is periodic with respect to fermi coordinate time with the difference frequency , @xmath148 , which is the difference frequency between satellite and observer rotations .
this is a kinematic effect due to the type of coordinate system used ( fermi coordinates ) and the fact that the observer ( fermi coordinate frame ) is moving along an arc of a circle , and therefore experiences an acceleration .
note that the periodic term vanishes when the observer angular velocity @xmath149 .
if the observer were moving a straight line and hence have zero acceleration , @xmath150 , the periodic effect would also be absent , as can be seen from eq .
( [ zetadef1 ] ) .
sagnac@xcite demonstrated that there exists a phase shift between two counter propagating light beams on a rotating platform and that this phase shift depends on the angular frequency of rotation of the platform . in the fermi
coordinates of the observer moving in a circle , there exists a sagnac - like effect for two portable clocks that are orbiting in opposite directions .
consider a clock that moves eastward along a circle of radius @xmath136 in the equatorial plane of the fermi coordinates with world line given by eq.([circularclockworldline1])([circularclockworldline3 ] ) .
see figure [ observerworldlinefigure ] .
the speed of the clock in fermi coordinates is @xmath151 . assuming that the clock begins its journey at point @xmath45 at time @xmath152 and ends it at @xmath153 , which corresponds to one revolution , the proper time that elapses on the clock is given by eq .
( [ orbitingclock1 ] ) as @xmath154 \frac{2 \pi}{\omega}c + \frac{a \ , b}{c}\frac{\omega^2}{\omega-\omega } \sin \left(2 \pi \frac{\omega}{\omega } \right ) \label{rightclock}\ ] ] the first term in eq .
( [ rightclock ] ) is the time dilation due to the speed @xmath155 of the clock .
the second term depends on the path traversed .
next , consider an identical clock that starts at point @xmath45 at time @xmath152 but moves westward .
the world line of this clock is then given by eq.([circularclockworldline1])([circularclockworldline3 ] ) but with @xmath156 .
the proper time on this clock is given by @xmath157 \frac{2 \pi}{|\omega|}c - \frac{a \
, b}{c}\frac{\omega^2}{|\omega|+\omega } \sin \left(2 \pi \frac{\omega}{|\omega| } \right ) \label{leftclock}\ ] ] when the east - moving and west - moving clocks have each made one revolution and they have returned to their starting positions at point @xmath45 , the difference in their times is given by @xmath158 .
for the case applicable to gps , @xmath159 and @xmath160 , the difference on the two clocks is given by @xmath161 where in the last line i have dropped terms of @xmath162 .
this sagnac - type effect is due to the non - inertial nature of the fermi coordinate system .
the origin of this effect can be understood from the point of view of the global inertial coordinate system .
if the fermi coordinate origin moved along a straight line in minkowski space ( rather than a circle ) then there would be no acceleration and @xmath163 . in this case
, the east - moving and west - moving clocks would have zero time difference after one revolution .
however , since the origin of fermi coordinates moves along a circle , the acceleration picks out a direction in the fermi coordinate space - time and the symmetry between the two clock world lines is broken .
consequently , the elapsed proper time is different on the two clocks when they are compared after one revolution .
note that the time difference on the two clocks increases with increasing angular velocity of rotation @xmath164 , analogous to the optical sagnac effect@xcite .
this increase is due to the fact that the breaking in symmetry is due to the magnitude of the acceleration , @xmath165 , see eq.([zetadef1 ] ) . using the values of the parameters appropriate to gps satellites , given in table 1
, i find the magnitude of the time differernce on the two clocks after one revolution to be @xmath166 this is a small effect that vanishes when the origin of fermi coordinates moves along a straight line or when @xmath149 .
the three dimensional space of minkowski space - time is homogeneous and isotropic , and consequently , the speed of light is a constant .
however , for the case of fermi coordinates of an observer moving in a circle , the symmetry of the 3-d space is broken by an acceleration . consequently , the space is not homogeneous and isotropic and the coordinate speed of light depends on position in the 3-d space . for a general space - time
metric @xmath94 , the coordinate speed of light @xmath167 in the direction of a unit vector @xmath168 is given by@xcite @xmath169 where @xmath170 and the unit vector satisfies the relation @xmath171 , where the 3-d spatial metric is given by@xcite @xmath172 equation ( [ lightspeed ] ) shows that for a general space - time metric the speed of light depends on direction @xmath168 and on position and time , through the metric components .
a non - zero off - diagonal metric component @xmath173 leads to an anisotropy ( directional dependence ) of the coordinate speed of light .
the metric for fermi coordinates given in eq .
( [ fcmetric ] ) has vanishing off - diagonal terms,@xmath174 , and consequently the coordinate speed of light is isotropic .
however , the metric component @xmath175 has a non - trivial position and time dependence , and eq .
( [ lightspeed ] ) gives the coordinate speed of light : @xmath176 where @xmath105 is a function of fermi coordinate position and time , given by eq .
( [ zetadef ] ) .
therefore , i can write the coordinate dependence of the speed of light as @xmath177 + o(\nu^4 ) \label{fclightspeedvariation}\ ] ] this variable speed of light is a kinematic effect due to the non - inertial nature of the fermi coordinate system .
note that at any given position @xmath178 the speed of light depends on time in a periodic way , with the orbital period of the reference frame , @xmath164 .
this expression can be rewritten as @xmath179 where the unit vector @xmath180 is in the observer s orbital plane and points in the direction of acceleration , toward the center of the circle .
see figure [ speedoflightfigure ] .
as the fermi reference frame moves around the circle , the vector @xmath181 always points toward the circle s center in fermi coordinates . at any given time @xmath127 , eq .
( [ fclightspeedvariation2 ] ) shows that to second order in @xmath17 , in the plane given by @xmath182 the speed of light does not differ from it s vacuum inertial frame value @xmath7 . in the 3-d half - space containing positions @xmath183 closer to the center of the circle
, the speed of light is increased .
for the 3-d half - space containing positions @xmath183 that are further from the center of the circle , the speed of light is decreased . as time @xmath127 increases , the dividing plane , given by eq .
( [ plane ] ) , rotates at angular frequency @xmath164 around the circle , with the origin of fermi coordinates remaining the point at which the plane is tangent to the circle . the magnitude of the inhomogeneous variation in the speed of light in eq .
( [ fclightspeedvariation ] ) is small . for the kinematics of gps , using the numbers
given in table 1 , with the coordinates @xmath133 , @xmath184 , so the effect has a magnitude @xmath185 and has a complicated position and time dependence given by eq.([fclightspeedvariation ] ) .
recently , gps data have been used to obtain an upper bound on the _ anisotropy _ of the speed of light to be @xmath186 , see ref .
@xcite and references contained therein .
given the present accuracy of gps , the small inhomogeneity of the speed of light , given by eq.([fclightspeedvariation])is probably not measurable .
i have considered the problem of an observer moving in a circle around a central point in flat space ( no gravitational fields ) .
the observer carries with him three gyroscopes and uses them to define a set of non - rotating axes .
he keeps his axes at a constant orientation with respect to the gyroscopes spin axes .
the observer s axes are fermi transported along his world line and fermi coordinates are defined .
the fermi time coordinate is just the observer s proper time , according to the construction shown in figure [ observermotionfigure ] .
fermi coordinates play the role of eci frame coordinates in gps .
i investigate the kinematic effects seen on ideal clocks from the point of view of these fermi coordinates .
specifically , i consider two cases : stationary clocks ( with respect to fermi coordinates ) and satellite clocks in circular orbit with respect to fermi coordinates . the proper time interval between two events on the stationary clock , given by eq.([stationaryclock ] ) , is a complicated periodic function of the clock s fermi coordinate time and coordinate position in the plane perpendicular to the rotation axis .
i considered next a clock that orbits the fermi coordinate origin ( similar to a satellite ) in the equatorial plane of rotation and computed the relation of proper time to coordinate time .
i find that the proper time on the clock has time dilation ( constant rate offset ) plus a periodic dependence on fermi coordinate time at the difference frequency , @xmath187 , between satellite clock orbital frequency , @xmath188 , and fermi coordinate frame orbital frequency , @xmath164 .
next , i looked at the time difference of two counter - orbiting clocks in the equatorial plane of rotation , one west - moving and the other east - moving .
after one revolution , the difference in proper time on the clocks shows a small sagnac - like effect , increasing with the angular frequency of revolution of the fermi frame , @xmath164 , given by eq .
( [ eastwesttimedifference ] ) .
finally , i computed the coordinate speed of light in the fermi coordinates .
the speed of light is isotropic ( same in all directions ) , but is not homogeneous . specifically , the speed of light depends on fermi coordinate position and time , but not on direction .
equation ( [ fclightspeedvariation2 ] ) gives the change in the speed of light from @xmath7 , its vacuum value in an inertial frame .
i have computed the magnitudes of these effects for numbers relevant to gps satellite orbits and orbital frequency .
the effects are all small for gps , except for the 12 ns periodicity for an orbiting clock , given in eq .
( [ orbitingclock1 ] ) , that occurs at the difference frequency , @xmath187 , between satellite orbital frequency @xmath188 and earth orbital frequency @xmath164 .
this effect is reminiscent of the periodic effect in gps described in the introduction .
however , the effects here are all kinematic since gravitational fields are neglected .
the inclusion of gravitational fields is expected to significantly modify the results . in particular , from the work of ashby and bertotti , this 12 ns periodic effect is expected to cancel when gravitational fields are included @xcite .
work is in progress to extend these results to include the effects of gravitational fields . while the effects discussed in this work are small for the present gps , these effects are quite significant for future navigation and time transfer systems that are expected to have sub - nanosecond time transfer accuracy @xcite .
the author thanks tom van flandern for useful discussions .
i use the notation and conventions of synge@xcite , except that my space - time indices on coordinates , @xmath24 , take values @xmath19 .
specifically , i use the convention that roman indices take the values @xmath19 , and greek indices @xmath20 .
i use the summation convention for repeated indices , when the same index appears in a lower and upper position then summation is implied over the range of the index .
if @xmath24 and @xmath189 are two events along the world line of an ideal clock , then the square of the proper time interval between these events is @xmath190 , where @xmath191 is given in terms of the space - time metric as @xmath192 and i choose @xmath94 to have the signature @xmath1932 .
when @xmath94 is diagonalized at any given space - time point , the elements can take the form of the flat space minkowski metric given by @xmath194 , @xmath195 , and @xmath196 .
b. w. parkinson and j. j. spilker , eds .
, _ global positioning system : theory and applications _ , vol .
i and ii ( p. zarchan , editor - in - chief ) , progress in astronautics and aeronautics , vol .
163 and 164 , amer .
astro . , washington , dc ( 1996 ) .
committee on the future of the global positioning system , _ the global positioning system : a shared national asset , recommendations for technical improvements and enhancements _ , commission on engineering and technical systems , national research council , national academy press , washington , dc ( 1995 ) .
m. weiss , apparent diurnal effects in the global positioning system " , p. 33 of proceedings of the nineteenth annual pprecise time and time interval ( ptti ) applications and planning meeting , redondo beach , california , december 1 - 3 , 1987 . | coordinate transformations are derived from global minkowski coordinates to the fermi coordinates of an observer moving in a circle in minkowski space - time .
the metric for the fermi coordinates is calculated directly from the tensor transformation rule .
the behavior of ideal clocks is examined from the observer s reference frame using the fermi coordinates .
a complicated relation exists between fermi coordinate time and proper time on stationary clocks ( in the fermi frame ) and between proper time on satellite clocks that orbit the observer .
an orbital sagnac - like effect exists for portable clocks that orbit the fermi coordinate origin .
the coordinate speed of light is isotropic but varies with fermi coordinate position and time .
the magnitudes of these kinematic effects are computed for parameters that are relevant to the global positioning system ( gps ) and are found to be small ; however , for future high - accuracy time transfer systems , these effects may be of significant magnitude . |
the r cra association , at a distance of about 130 pc ( marraco & rydgren 1981 ) , is one of the nearest star - forming associations to the sun .
the association is projected against a prominent dark cloud 18@xmath0 below the galactic plane ( see figure 1 of knacke _
et al_. 1973 ) .
the total mass of the cloud is estimated to be between 3000 and 10,000 m@xmath1 ( dame _ et al_. 1987 ; loren 1979 ) .
et al_. ( 1994 ) comment that the morphology of the cloud bears a resemblance to the @xmath2 oph cloud . however , the number of known and suspected members of the stellar association is only 10% of that known or suspected in @xmath2 oph .
knacke _ et al_. ( 1973 ) discovered a group of 11 young stars near r cra .
members of the association include the intermediate - mass stars r , t , and ty cra and the classical t tauri stars s , vv , and dg cra .
nine of the 11 stars had near - ir continuum excesses .
et al_. also identified 10 h@xmath3-emitting stars spread over the region , at projected distances up to a half degree from the dark cloud .
glass & penston ( 1975 ) undertook near - ir photometry of 44 stars in this region , and noted that the only stars with near - ir continuum excesses were the known variable stars .
they failed to detect near - ir excesses for 3 of the h@xmath3-emitting stars identified by knacke _
et al_. , and commented that h@xmath3 emission and 2@xmath4 m excesses do not necessarily correlate .
marraco & rydgren ( 1981 ) reexamined the association .
they found h@xmath3 emission in only 4 of the 10 knacke _
et al_. sources , but discovered another 6 h@xmath3-emitting objects . wilking _
et al_. ( 1992 ) surveyed this region using the iras data , finding 16 iras sources associated with young stellar objects ( ysos ) .
they identified 24 ysos from their near- and far - ir spectral energy distributions . wilking _
et al_. ( 1994 ) presented results of near - ir imaging of the cloud core , and concluded that there are approximately 30 members of this association .
knacke _ et al_. ( 1973 ) estimated an association age of less than 10@xmath5 years ( 1 myr ) , assuming that the late - b stars of the association have just arrived on the main - sequence . wilking _
et al_. ( 1992 ) estimate an age of between 1.5 myr ( the contraction age of the zero - age main sequence b8 star ty cra ) and 6 myr ( the contraction time for the pre - main sequence a5 star r cra ) .
while infrared observations can be useful in identifying low - mass pre - main - sequence ( pms ) stars , experience over the past decade has shown that the population sampled in the infrared is not necessarily representative of the true population of young associations .
x - ray imaging observations ( e.g. , feigelson & kriss 1981 ; montmerle _ et al_. 1983 ; feigelson _ et al_. 1993 ; mundt _ et al_. 1983 ; strom & strom 1994 ; walter _ et al_. 1988 , 1994 ) of young associations reveal a population of x - ray - luminous young stars which generally lack infrared continuum excesses .
walter ( 1986 ) found 3 x - ray - luminous pms stars near the cra cloud .
here we report on the complete einstein observatory imaging proportional counter ( ipc ) observations of the cra dark cloud .
in addition to the 3 sources previously identified , there are another 8 x - ray sources . in this paper
we discuss the identifications of these x - ray sources .
the x - ray data consist of the two overlapping ipc observations listed in table [ tbl-1 ] and shown in figure [ fig1 ] .
the ipc ribs , which support the detector window , obscure @xmath630% of the nominal 1 square - degree field of the ipc .
there is significant vignetting in the ipc ; effective exposure times at the edge of the field ( 30 arcmin from the center ) are about 50% of those on - axis .
the ipc is sensitive to photons with energies between about 0.1 and 4.5 kev .
active stellar coronae have characteristic temperatures of 10@xmath7k , or about 1 kev , and are well matched to the ipc response .
the energy resolution of the ipc is about 100% , and only gross spectral information is available for weak sources .
the 90% confidence uncertainty on the x - ray source positions is typically @xmath830 - 40 arcseconds .
details of the ipc are given in giacconi _
et al_. ( 1979 ) .
there are three x - ray sources in ipc sequence 3501 , centered on @xmath9 cra ( hd175813 ) .
two of these were found in the standard rev 1b processing of sequence .
the mdetect ( map detect ) source detection algorithm had not run as a part of the standard processing .
we requested that mdetect be run off - line , and it yielded the third x - ray source . in ipc sequence 4512 ,
centered on the cra dark cloud , the standard processing yielded 10 x - ray sources .
all of the x - ray sources had potential optical counterparts on the poss sky survey .
the w uma star @xmath9 cra ( cruddace & dupree 1984 ) was detected in both ipc observations .
the other x - ray sources , their j2000 coordinates , and aliases , are presented in table [ tbl-2 ] .
we obtained low - dispersion spectra of these stars at the kitt peak national observatory in june 1983 , april 1986 , and may 1987 , using the intensified image dissector scanner ( iids ) on the 2.1-m telescope .
all observations were made through the 7-arcsecond aperture .
observations in the red covered the range 6150 - 7200 at about 3 resolution , and were obtained with grating 36 and the gg495 order - sorting filter .
blue observations , taken using grating 36 and the cuso@xmath10 order - sorting filter , covered the range 3750 - 4200 with about 1.5 resolution .
the spectra were fully reduced and wavelength- and flux - calibrated using the mountain data reductions and the ipps software package .
flux calibration relied on the irs standard stars ( barnes & hayes 1982 ) .
subsequent data analysis was performed using software written in idl .
low - dispersion spectra were also taken on 18 and 23 july 1995 , and on 21 july 1996 , at the european southern observatory ( eso ) using the 1.52 m telescope equipped with a boller and chivens spectrograph .
we used grating # 5 and ccd # 24 ( in 1995 ) or ccd # 39 ( in 1996 ) , achieving a mean resolution of about 2.5 ( fwhm ) in the 4600 - 7000 spectral range .
this instrumental set - up allows us to resolve the lithium i absorption line at 6707 from the ca i absorption line at 6717 as well as to carry out an accurate spectral type classification . for the wavelength calibration
, a he - ar exposure was taken after each science exposure .
the standard stars ltt 4816 , ltt 7987 , and eg 274 were observed each night with the same instrumental set - up to perform the flux calibration .
we used the midas software ( versions 95may and 95nov ) to reduce these spectra .
bias and dark subtraction was first performed on each frame .
the 2-d science frames were then divided by a mean flat - field and then calibrated in wavelength .
the sky subtraction , the extraction of one dimensional spectra and the flux calibration using a mean response function , were finally performed .
high - dispersion spectra were obtained using the cerro tololo inter - american observatory ( ctio ) 4-m echelle on 4 - 6 april 1987 and 23 - 27 may 1989 . using the red air schmidt camera with the 31.6 l / mm echelle , cross - disperser # 3 , the 200@xmath4 m slit , and the gg495 order sorting filter
, we obtained 17 spectral resolution from 5600 to 7000 .
we used decker 1 in april 1987 and decker 2 in 1989 .
the ccd detector was epi # 9 in 1987 and epi # 12 in 1989 .
we subtracted the overscan region and the ccd bias and trimmed the frames using iraf .
subsequent reductions were undertaken using an echelle data reduction package written in idl ( walter 1992 ) .
the data were flux - calibrated using kopff 27 and ltt 3864 as flux - standard stars .
most observations consisted of 3 sequential integrations ; cosmic rays were filtered out by applying a median filter to the multiple observations of each target .
the 1987 observations were compromised by unusually large flexure within the spectrograph . in some cases
flexure shifted the spectra off the region illuminated by the projector flat ( taken using decker 2 ) , resulting in lower s / n than indicated by counting statistics alone .
radial velocities obtained during this run are unreliable , and have not been used .
we obtained multiple high - dispersion single - order spectra of each star with the echelle spectrographs of the whipple observatory 1.5-m reflector and the mmt between 1986 and 1993 for the purpose of measuring radial velocities .
all the spectra were centered at @xmath115187 and provided about 50 coverage .
radial velocities were obtained via cross - correlation , following the procedures discussed by nordstrom _
et al_. ( 1994 ; see also latham 1992 ) .
cross - correlations were made against a grid of synthetic templates with a separation of 250k in temperature , 0.5 in log g , and 10 in v _ sin i_. for each star the radial velocity was adopted for that template providing the highest value for the average of the correlation peak . the typical precision of a single radial - velocity measurement is 1 , based on the dispersions of repeated measurements .
rotational velocities were determined through an interpolation based on correlation peak strength , with a measurement lower limit of @xmath12 10 .
we estimate an uncertainty of 5 .
ubvri photometry was obtained in may 1986 , june 1987 , and may 1988 at ctio .
the data were obtained with the 1.0-m telescope using a hamamatsu gaas photomultiplier and apertures ranging from 15 - 20 arcseconds .
all ubv observations are on the johnson ( 1963 ) system , while all r and i observations are on the cousins ( 1980 ) system .
we observed secondary standards from landolt ( 1983 ) .
additionally , a few ubv obsevations were obtained for us by dr .
robert millis during may 1986 with the ctio / lowell 0.4-m telescope using an emi6256 photomultiplier .
multiple , independent observations were obtained of each star during these observing runs with no obvious systematic differences in results between the various systems employed .
further ubvri photometry was obtained from ccd images acquired at ctio on 1993 may 5 and 1994 june 3 using the 0.9 meter telescope .
the detector used was the tek 2k-1 ccd with the standard bessell filter set .
ccd debiasing and flat - fielding were performed using the iraf ccdred package .
instrumental magnitudes were determined using the iraf apphot package .
in the case of the close pair crapms 6 , photometry was performed on the individual stars after first subtracting off the other star using routines in the iraf daophot package .
the photometry was placed on the standard johnson - kron - cousins system using calibration stars from landolt ( 1992 ) .
mean residual errors are less than 3% except in the case of the crapms 6 pair , which individually have mean residual errors of about 5% .
we also obtained near - infrared jhkl photometry during june 1986 using the ctio 4-m telescope and during may 1989 using the ctio 1.5-m telescope . both observing runs
employed a monolithic insb detector and used standards from elias _
et al_. ( 1982 ) to place the observations onto the standard system . at the 4-m telescope
a 6-arcsecond aperture was used , while at the 1.5-m telescope a 14-arcsecond aperture was used .
one to five independent observations were obtained of each program star .
we obtained additional jhk observations on 19 - 21 february 1997 , using the cirim imaging detector on the ctio 1.5-m telescope .
each image consisted of a raster of 6 exposures , with the telescope offset by 15 arcsec between exposures .
we observed standard stars from elias _
et al_. ( 1982 ) .
the non - linearity correction was applied using the iraf task irlincor .
all subsequent data reductions were undertaken using idl software written for the task .
the images were flattened using dome - flats .
the local sky was determined by median - filtering the six images , and was subtracted off .
the frames were then registered and coadded .
we used the standard star observations to determine the extinction correction for each night and place the photometry on the standard cit system .
uncertainties in the photometric solution amount to @xmath80.02 mag for single observations .
target fluxes were measured in an 8 arcsec radius aperture . for close binaries we measured the relative fluxes in smaller apertures , and then corrected to the total flux of the system .
we provide identifications for the low - mass stars following the iau convention ( dickel _ et al_. 1987 ) in table [ tbl-2 ] .
the identifications are the truncated j2000 coordinates .
we also provide crapms numbers which are shorter and easier to work with .
note that , by iau convention , the proper designation for crapms 1 is ntt j190134 - 3700 . in this designation ,
the `` ntt '' prefix is retained for continuity with other papers in this series , and does not necessarily imply anything about the evolutionary status of the stars . finding charts for newly - identified stars are presented in figure [ fig2 ] . a finding chart for crapms 7
is given in marraco & rydgren ( 1981 ) , and finding charts for crapms 1 , 2 , and 3 are in walter ( 1986 ) .
optical spectra are shown in figures [ fig3 ] and [ fig4 ] .
low dispersion spectra of crapms 1 , 2 , and 3 are in walter ( 1986 ) .
comments on the individual sources follow .
crapms 1 .
: : this star has many aliases : cd -37@xmath0 13022 , anon 1 ( knacke _ et al_. 1973 ) , i2 ( glass & penston 1975 ) , vss13 ( vrba , strom , & strom 1976 ) , cra 1 ( walter 1986 ) and hbc 676 ( herbig & bell 1988 ) .
marraco & rydgren ( 1981 ) noted weak ca ii h&k emission cores and 0.2 mag variability at v , and called it a pms star with extremely weak emission .
crapms 2 .
: : this is cd -37@xmath0 13029 , a2 ( glass & penston 1975 ) , vss1 ( vrba _ et al_. 1976 ) , cra 2 ( walter 1986 ) , and hbc 678 ( herbig & bell 1988 ) .
crapms 3 .
: : this is star w ( glass & penston 1975 ) , vss24 ( vrba _ et al_. 1976 ) , cra 3 ( walter 1986 ) , and hbc 679 ( herbig & bell 1988 ) .
the faint optical companion visible to the northeast is very red , and is about 1.8 mag fainter than crapms 3 at k ( j k = [email protected] , j
h = [email protected] ) .
crapms 4 .
: : this source was found by running the mdetect source - detection algorithm off - line .
two stars , designated 4nw and 4se and separated by 43 arcsec , are within the x - ray error circle
. both are pms stars .
crapms 6 .
: : this 3 arcsec pair of m3 stars is h@xmath3 11 of marraco & rydgren .
the stars appear of comparable brightness on the palomar observatory sky survey prints and on our ccd images , but on 3 dates ( 1987 june 5 , 1988 may 24 , and 1989 may 27 ) the sw component appeared faint . on 1988
may 24 the sw component was fainter than the ne component by [email protected] magnitudes at v. in 1993 may the v magnitudes were identical within the errors .
6sw may be a large amplitude variable .
6ne is also variable , with a 0.3 mag range at v in 3 observations . in 1997
the k magnitudes were comparable .
crapms 7 .
: : the counterpart is h@xmath3 6 of marraco & rydgren .
h@xmath3 5 is on the edge of the 90% confidence error circle , but it is fainter , bluer , and does not have h@xmath3 emission ( marraco & rydgren 1981 ) , and is unlikely to be pms .
h@xmath3 6 is a classical t tauri star , with an infrared excess ( wilking _ et al_. 1992 ) and a strong h@xmath3 emission line ( w@xmath13(h@xmath3)=34 ) .
crapms 9 .
: : there are three stars in the 43 arcsec error circle .
the star to the north is the pms star ; the close visual pair to the south shows neither evidence for h@xmath3 emission nor li absorption .
crapms 10 .
: : this x - ray source is coincident with the pair of late - b stars hd 176269 ( = hr7169 = sao 210815 ) and hd 176270 ( = hr7170 = sao 210816 ) . the resolution of the ipc is insufficient to split the pair .
crapms 11 .
: : this source is coincident with the b9 star ty cra , in the ngc 6726/7 nebulosity .
ty cra is a multiple system ( e.g. , casey _ et al_. 1993 ) .
the b9 star is part of a double - lined spectroscopic binary with a 2.9@xmath14 orbital period ( casey _ et al_. 1995 ) .
the secondary is a 1.6 m@xmath1 g star .
there is also a tertiary in the system .
we identified no spectroscopic binaries among the 6 stars observed at high dispersion .
four of the low - mass stars are in two visual pairs .
we consider both to be physical pairs .
the members of the wide crapms 4 pair ( @xmath65000 au projected separation ) have essentially identical radial velocities , and the crapms 6 pair consists of similar stars with a projected separation of 300 au .
the spatial association of each pair seems secure , but that they are gravitationally bound is not established , especially in the case of crapms 4 .
each of the higher mass objects , crapms 10 and crapms 11 , are multiple .
the binary fraction exceeds 56% ( 9 stars of 16 in multiple systems ) or 36% ( 4 multiple systems of 11 ) , depending on the definition .
the binary fraction is higher ( 59% or 46% ) if we assume that the companion to crapms 3 is physical .
the number of stars per system exceeds 1.4 .
seven of these 11 low - mass pms stars were first noted through their x - ray emission , 5 here and 2 by walter ( 1986 ) ; only the h@xmath3 11 pair ( crapms 6 ) , h@xmath3 6 ( crapms 7 ) , and anon 1 ( crapms 1 ) were first identified using optical techniques .
once again , x - rays prove to be a powerful search technique for low - mass pms stars .
the photometric magnitudes and colors are given in tables [ tbl - pht ] and [ tbl - irpht ] .
the second column of these tables gives the number of independent optical or near - infrared observations the listed photometry is based upon .
errors of a single observation in vri were typically .01 mag at v @xmath15 15 mag , increasing to 0.03 mag at v = 16 mag , while at b and u the errors increased rapidly above 0.02 mag for v @xmath16 14 , due to the red nature of these stars . however , the uncertainties in the optical photometry listed in table [ tbl - pht ] are dominated by the fact that all stars with sufficient numbers of observations showed evidence for variability .
the range of observed variability is typically 0.15 , 0.08 , 0.04 , and 0.02 mag for ubvr , respectively . for stars with many ( n @xmath17 15 ) observations ,
the variability appears consistent with rotational modulation by starspots with additional flaring visible in the u and b filters . in the near - ir ,
the uncertainties on a single observation are typically 0.02 mag ( jhk ) or 0.03 mag ( l ) .
only crapms 3 showed significant near - ir variability ( table [ tbl - irpht ] ) ; part or all of this may be attributable to its companion .
the measured radial ( v@xmath18 ) and rotational ( v@xmath19 velocities and equivalent widths w@xmath13 , the estimated spectral types and extinctions a@xmath20 , and the derived luminosities and masses are presented in table [ tbl - par ] .
we determine the spectral types by visual comparison of the spectra with a grid of mk standards .
we use the spectral types and the colors jointly to estimate the extinctions and luminosity classes .
temperatures are estimated from the unreddened r@xmath21i@xmath21 and v
k colors ( laird 1985 , carney 1983 , veeder 1974 ) , and agree with the spectral type temperature calibration of de jager & nieuhenhuijzen ( 1987 ) , with a mean difference of 39k . for stars with v
[email protected] , we use the color - t@xmath23 calibration of kirkpatrick _
et al_. ( 1993 ) .
stauffer ( 1997 ) has shown that this color - t@xmath23 calibration calibration brings the lower main sequence of the pleiades into agreement with the dantona & mazzitelli ( 1994 ) evolutionary tracks .
masses and ages are determined from the star s location on the h - r diagram , and are dependent upon the choice of evolutionary track .
further details of the techniques may be found in walter _
et al_. ( 1994 ) .
we assume a distance of 130 pc to the association . on the basis of spatial association with the r cra cloud , location above the main - sequence , luminosity classes , radial velocities , and the presence of li i @xmath116707 absorption ,
all these stars are pms .
the dispersion in radial velocities is consistent with the 2 km s@xmath24 velocity dispersion generally observed among pms stars in the central parts of star - forming regions ( mathieu 1986 ) . the lithium abundances log n(li ) ( table [ tbl - par ] ) are derived using the nlte analysis of pavlenko & magazz ( 1996 ) .
these curves of growth are valid for t@xmath253500k ; we extrapolate their curves of growth for our coolest stars ( 3300k@xmath15t@xmath263500k ) .
the abundances determined for the hotter stars are consistent with the abundances seen in t tauri stars ( basri , martin , & bertout 1991 ) .
these are presumably the initial li abundances , as depletion is not yet expected in stars of these masses . for the cooler stars ,
the abundances are seen to be significantly lower .
pre - main sequence lithium burning ( e.g. , pinsonneault , kawaler , & demarque 1990 ) is expected among the m stars , but the expected depletion of less than half a dex for an age of @xmath1510 myr is significantly less than we measure .
this discrepancy may be accounted for by the fact that the presence of an active chromosphere / transition region will over - ionize li , giving an apparent underabundance ( houdebine & doyle 1995 ) .
note that the m stars with the largest h@xmath3 equivalent widths have the smallest apparent li abundances .
there is evidence of a near - ir excess attributable to warm circumstellar dust only in h@xmath3 6 ( crapms 7 ) .
x - ray count rates and derived luminosities , surface fluxes and flux ratios are given in table [ tbl - xray ] .
none of the targets were sufficiently bright to make extraction of the spectrum practical .
we assume that the x - ray emission arises in a 10@xmath7k corona for conversion to luminosities and flux ratios .
all derived quantities in table [ tbl - xray ] are corrected for extinction , assuming r=3.1 .
we divide the observed counts equally between the stars in the two pairs .
this is a good assumption for crapms 6 , which consists of two similar stars , but probably not for the crapms 4 pair .
the luminosities and flux ratios are similar to those observed in other young associations ( walter _ et al_. 1994 ) .
the luminosity - temperature diagram for the x - ray - selected stars is presented in figure [ fig5 ] .
the two panels show the stars plotted against theoretical evolutionary tracks from dantona & mazzitelli ( 1994 ) , and from swenson _
et al_. ( 1994 ) .
the dantona & mazzitelli tracks employ the canuto & mazzitelli ( 1990 ) treatment of convection and alexander _
et al_. ( 1989 ) opacities .
both sets of tracks are for solar abundances . in the absence of other information
, we must rely on the evolutionary tracks to estimate stellar masses and ages .
we note the following in figure [ fig5 ] : 1 .
the x - ray - selected stars are significantly older than the ir - excess stars discussed by knacke _
et al_. ( 1973 ) and by wilking _
et al_. ( 1992 ) .
note that the x - ray and ir samples have few stars in common .
ages estimated from the swenson _
et al_. tracks are systematically older than those from the dantona & mazzitelli tracks , by about a factor of 2 .
the mean ages are 7 and 15 myr , relative to the dantona & mazzitelli and the swenson _
et al_. tracks , respectively .
2 . with the exception of crapms 1 , which appears significantly younger than the rest of the stars , the stars lie along a narrow locus .
3 . there is no evidence that the h@xmath3-emitting stars h@xmath3 6 and h@xmath3 11 are significantly younger than the other x - ray - selected stars .
two of the x - ray sources are coincident with late - b stars .
ty cra ( b9v ) is an eclipsing triple system ( casey _ et al_. 1995 ) . with a mass of 1.6 m@xmath1 and an age of @xmath156 myr ,
the secondary will be a g / k star similar to hde283572 ( walter _ et al_. 1987 ) , and can easily account for all the observed x - ray flux .
the tertiary mass is less certain : casey _ et al_. ( 1995 ) find 2.4 m@xmath1 , while corporon , lagrange , & beust ( 1996 ) find a mass of 1.2 - 1.4 m@xmath1 . in either case
the tertiary is likely to be an x - ray source . both hr7169 ( b9v ) and hr7170 ( b8iv - v ) , a 13 arcsec visual pair ,
are listed as spectroscopic binaries in the yale bright star catalog ( hoffleit 1982 ) ; further observations are needed to determine if there are low - mass stars associated with the hr7169/70 system .
are late - b stars in star - forming regions bright x - ray sources ?
a basic conclusion to be drawn from the einstein and rosat stellar x - ray observations is that all stars are x - ray sources _ except the late - b to mid - a stars_. schmitt _ et al_. ( 1993 ) observed the young lindroos ( 1986 ) pairs , consisting of a near - zero - age main sequence b star and a young late - type visual companion . using the rosat hri , with 3 arcsec positional accuracy
, they showed that in 4 cases _ both _ the b star and the late - type companion ( ages less than 5@xmath2710@xmath7 years ) were x - ray sources .
however , caillault , gagn , & stauffer ( 1994 ) have argued that the fraction of late - b / early - a stars in the orion nebula which are detected as x - ray sources is just that fraction expected to have late - type ( f5 or later ) binary companions .
our data clearly do not help resolve the question of whether the late - b stars can be x - ray sources .
the classical t tauri stars s cra and dg cra , and the intermediate - mass stars r cra and t cra were within the area viewed by the einstein observatory .
none were detected .
the 2@xmath28 upper limit to the ipc count rates is given in table [ tbl-3 ] .
the conversion from flux to counts depends on the assumed extinctions , which are poorly known .
we assumed a coronal plasma at a temperature of 10@xmath7k .
for a wide range of reasonable extinctions ( a@xmath29 10 mag ) , the count rate limits correspond to coronal luminosities l@xmath305@xmath2710@xmath31 erg s@xmath24 .
star formation in this cloud has been going on for a long time .
the x - ray - selected stars , with ages extending upward to over 10 myr , appear significantly older than the b stars and imbedded sources .
the narrow locus occupied by the stars in the luminosity - temperature diagram may be evidence for a burst of star formation about 7 - 10 myr ago , since younger stars will be more luminous , both at optical and x - ray wavelengths ( that the imbedded ir sources are not detected is a consequence of their high extinction and the relatively high limiting flux in these short observations ; koyama _ et al_. ( 1996 ) and neuhuser & preibisch ( 1997 ) , using asca and rosat , respectively , have reported the x - ray detection of imbedded class i protostars ) .
however , given the small numbers , we can not rule out continuous star formation at a constant rate over the last 10 myr . given this population of older stars
, it appears that the r cra cloud is not simply a lower mass analog of the @xmath2 oph cloud ( wilking _ et al_. 1992 ) , but may represent a more evolved state of such a cloud .
we observe 11 low - mass stars ( counting visual pairs as 2 stars , and neglecting the probable low - mass companions of the b stars ) in these 9 x - ray error circles . in previous studies of the x - ray emitting populations of star - forming associations ( walter _
et al_. 1988 ; 1994 ) , we argued that the completeness of the x - ray sample is @xmath630% .
this incompleteness has been verified in the taurus region by neuhuser _
et al_. ( 1995 ) .
the same arguments hold here , implying that the total number of x - ray emitting low - mass pms stars in the @xmath61.5 square - degrees observed is about 30 . therefore the total number of low - mass stars in this association must be about 60 , or double that previously known .
this is an underestimate of the total population of the r cra association .
the x - ray observations sampled only about 1.5 square degrees near the center of the cloud . over 10% ( 2 of 16 ) of those ir sources whose spectral energy distributions are consistent with their being pms stars ( wilking _ et al_. 1992 see figure [ fig1 ] ) lie in the dark streamers to the east of the r cra complex , near the b cloud ( rossano 1978 ) and outside the einstein ipc fields .
if the x - ray sources are spatially distributed in the same way , the total number of x - ray emitting low - mass pms stars must be increased by over 10% .
note that neuhuser _
et al_. ( 1997 ) identified gsc 7916 0050 , 1.6@xmath0 to the southwest of r cra and well outside the ipc fields , as a pms star .
in addition , some fraction of the x - ray bright population will lie behind the dark cloud , and will not be detected because their x - ray fluxes lie below our flux limit . since 4 pms systems are projected against the near side of the cloud , we would expect about the same number , or about 30% of the total population , to be hidden from view .
these corrections would increase the inferred x - ray bright population ( seen and unseen ) to about 40 stars , and the total population to about 70 stars .
wilking _ et al_. ( 1994 ) claimed , based on ir source counts , that the total stellar population in the core 0.24 pc@xmath32 of the cra complex is small ( @xmath630 stars ) , assuming a 2 myr age for the association . if the age is 10 myr , their ir counts permit up to 78 association members in this small region ( wilking , private communication ) .
given the large uncertainties in the estimates and the differences in the volumes samples , there is no significant difference in the total population estimates from the ir or x - ray samples .
the x - ray - bright population is observed to be more widely dispersed in space than the ir - selected population .
this is to be expected if the x - ray - bright population is older , and has some non - zero velocity dispersion .
some fraction of the total population may have dispersed outside the area we observed . to estimate the fraction of an old , dispersed population that should be visible within the volume sampled by the einstein observations
, we modelled the association as a population with an age of 7 myr and a velocity dispersion of 1.7 km s@xmath24 , and neglect gravitational decelerations .
we assume that the stars all form within 1@xmath0 of the position of r cra .
if the star formation rate has remained constant over the past 7 myr , then about 70% of the members of the association will lie within the volume sampled .
if the x - ray bright sample represents a short burst of star formation , with an age spread of 1 myr , then only 45% of the population lies within the area sampled . identifying the ir - bright sample with the youngest sources , and the x - ray bright sample with the non -
imbedded older population , then the correction for the stars outside the ipc fields leads to an estimate of a total population of about 90 stars ( in the case of continuous star formation , [email protected] ; in the case of 2 episodes of star formation , n=30 + [email protected] ) , with a likely uncertainty of @xmath820 stars .
we are attempting to verify this estimate using an x - ray - selected sample of stars from observations with the rosat observatory ( walter _ et al_. , in preparation ) .
wilking _ et al_. ( 1992 ) noted that 25% of the stars in this association are of intermediate - mass ( spectral types b8 to a5 ) , an unusually large number .
other star formation regions ( e.g. , upper sco ; walter _ et al_. 1994 ) show initial mass functions ( imf ) similar to the field star mass function .
if the r cra imf is consistent with the miller - scalo ( 1979 ) field star mass function , and the intermediate mass stars all have masses greater than [email protected] m@xmath1 , then one expects a total population of 92@xmath33 stars with [email protected]@xmath1 .
given the uncertainties in both the extrapolation of the imf from 6 high mass stars and in our estimate of the low mass population , we see no evidence that the the mass function is abnormal . we conclude that the cra clouds have spawned of order 100 low - mass stars over the past 10 myr , and that there is no compelling evidence for an excess of intermediate - mass stars .
we have identified 11 low - mass pre - main - sequence x - ray emitting stars associated with the r cra star - forming region . with the exception of h@xmath3 6 , none of these stars have near - ir excesses attributable to warm circumstellar dust .
none of the four classical t tauri stars in the field were detected .
x - ray emission was detected from the vicinity of the late - b stars ty cra and hr 7169 , but the emission may be attributable to lower mass companions
. only 4 of the low - mass stars ( 2 in a close visual pair ) were previously known from optical work to be pms .
the low - mass stars are significantly older than the 1 - 2 myr age inferred for the infrared sources and intermediate - mass stars .
the narrowness of the locus occupied by the x - ray - selected stars suggests an episode of star formation some 7 myr ago , but we can not exclude continuous star formation at a roughly constant rate over the past 10 myr .
we estimate that the older ntt population dominates the total number of stars , and that the total pms population amounts to about 90 stars . given 6 intermediate - mass stars ,
the number of low - mass stars is consistent ( within the large uncertainties ) with the field initial mass function .
this research has been supported by nsf grant ast89 - 96308 , nasa grants nag8 - 884 , nag5 - 1663 ( adp ) , and nagw-1365 ( origins of solar systems ) to suny , and nsf grant ast8814986 to the university of wisconsin .
rdm acknowledges support from the presidential young investigator program .
the paper is based in part on observations collected at the european southern observatory , la silla , chile .
we thank the folks at cfa , especially f. seward and j. mcsweeney , for their assistance with obtaining and examining the ipc images and using the measuring engine .
we thank r. millis for contributing a portion of the ubv photometry , and j. stauffer for bringing to our attention an improved color
t@xmath23 relation for m stars .
we thank r. davis , e. horine , j. peters , d. latham , and a. milone for assistance in the acquisition of the cfa radial velocity measurements , and juan alcal of mpe garching and the eso staff at la silla observatory .
this research made use of the simbad database , operated by cds , strasbourg , france , and of the digitized sky survey , produced at the space telescope science institute under grant nag w-2166 .
we thank the referee for constructive criticisms and for bringing to our attention the near - ir brightness of the faint optical companion to crapms 3 . 185717 - 3643
& 4nw & 7421 0665 & 18 57 17.8 & -36 42 35.9 & + 185720 - 3643 & 4se & 7421 1242 & 18 57 20.7 & -36 43 00.4 & + 185801 - 3655 & 5 & 7421 1040 & 18 58 01.7 & -36 53 45.0 & + 185914 - 3711ne & 6ne & & 18 59 14.7 & -37 11 30 & h@xmath3 11 + 185914 - 3711sw & 6sw & & & & h@xmath3 11 + 190001 - 3637 & 7 & & 19 00 01.5 & -36 37 07 & h@xmath3 6 + 190028 - 3656 & 8 & & 19 00 28.9 & -36 56 02 & + 190039 - 3648 & 9 & & 19 00 39.1 & -36 48 11 & + & 10&7421 2294 & 19 01 03.1 & -37 03 38.7 & hr7169 + & " & 7421 2295 & 19 01 04.2 & -37 03 41.8 & hr7170 + 190134 - 3700 & 1 & 7421 1890 & 19 01 34.9 & -37 00 55.8 & cra 1 + & 11&7421 1126 & 19 01 40.8 & -36 52 36.3 & ty cra + 190201 - 3707 & 2 & 7421 0493 & 19 02 01.9 & -37 07 43.2 & cra 2 + 190222 - 3655 & 3 & 7421 1899 & 19 02 22.1 & -36 55 40.8 & cra 3 + lrrrrrrr [ tbl - pht ] 1 & 20 & 11.34 & 0.21 & 0.65 & 1.11 & 0.71 & 0.73 + 2 & 18 & 10.44 & 0.15 & 0.29 & 0.79 & 0.47 & 0.46 + 3 & 18 & 13.79 & 0.18 & 1.04 & 1.43 & 0.99 & 1.02 + 4nw & 5 & 13.35 & 0.02 & 1.16 & 1.38 & 0.93 & 0.96 + 4se & 13 & 10.95 & 0.07 & 0.44 & 0.88 & 0.51 & 0.49 + 5 & 15 & 12.30 & 0.20 & 0.85 & 1.11 & 0.70 & 0.67 + 6ne & 3 & 15.14 & 0.36 & 1.03 & 1.48 & 1.11 & 1.56 + 6sw & 1&15.34 & & & 0.96 & 1.26 & 1.82 + 7 & 1&13.93 & & & 1.26 & 0.85 & 1.00 + 8 & 3 & 15.68 & 0.14 & 1.20 & 1.47 & 1.30 & 1.66 + 9 & 1 & 15.88 & & 1.47 & 1.65 & 1.26 & 1.58 + lrrrrr [ tbl - irpht ] 1 & 4 & 7.80 & 0.92 & 0.68 & 0.13 + 2 & 3 & 8.34 & 0.52 & 0.43 & 0.08 + 3&2 & 9.18 & 1.06 & 0.80 & 0.16 + 3&2 & 9.05 & 1.11 & 0.83 & + 3&1 & 9.21 & 1.12 & 0.84 & + 4nw & 2 & 9.42 & 0.84 & 0.67 & + 4se & 2 & 8.70 & 0.57 & 0.47 & + 5 & 4 & 9.20 & 0.73 & 0.59 & 0.11 + 6ne & 1 & 10.69 & 0.77 & 0.54 & + 6ne & 1 & 10.56 & 0.84 & 0.55 & + 7&- & 9.35 & 1.43 & 0.93 & 0.66 + 8 & 2 & 9.59 & 1.12 & 0.81 & + 1 & k1iv & 1.2 & -1.0 & 58 & 8 & 0.39 & 3.0 & 0.3 & 4702 & 0.19 & 1.0 + 2 & g5iv & 0.3 & -1.2 & 23 & 20 & 0.28 & 3.1 & -0.7 & 5281 & 0.10 & 1.2 + 3 & k2iv & 2.3 & -1.2 & @xmath1510 & 4 & 0.41 & 3.2 & -0.9 & 4523 & -0.31 & 0.9 + 4nw & m0.5v & 0.2 & -2.2 & @xmath1510 & 5 & 0.45 & 2.0 & -1.1 & 3886 & -0.66 & 0.6 + 4se & g5iv & 0.5 & -2.0 & @xmath1510 & 6 & 0.36 : & 3.5 & 1.0 & 5302 & -0.03 & 1.1 + 5 & k5iv & 0.1 & -0.8 & 22 & 7 & 0.44 & 2.4 & -0.8 & 4111 & -0.51 & 0.7 + 6ne & m3v & 0.2 & & & - & 0.70 & 2.2 & -5.9 & 3525 & -1.14 & 0.3 + 6sw & m3.5v & 0.3 & & & - & 0.44 & 1.7 & -9.8 & 3332 & -1.06 & 0.2 + 7 & m1v & 0.6 & & & - & 0.36 : & 1.5 & -33.5 & 3973 & -0.66 & 0.7 + 8 & m3v & 1.6 & & & - & 0.57 & 2.1 & -3.9 & 3631 & -0.73 & 0.4 + 9 & m2v & 1.7 & & & - & 0.5 : & 1.9 & -9.2 & 3704 & -0.93 & 0.5 + lrccr [ tbl - xray ] 1 & 0.034 @xmath80.005 & 2.9e30 & 1.4e7 & 4.5e-4 + 2 & 0.038 @xmath80.006 & 2.2e30 & 1.8e7 & 4.0e-4 + 3 & 0.012 @xmath80.003 & 1.7e30 & 1.5e7 & 7.2e-4 + 4nw & 0.014 @xmath80.005 & 3.9e29 & 3.7e6 & 3.7e-4 + 4se & & 4.6e29 & 5.1e6 & 1.1e-4 + 5 & 0.014 @xmath80.005 & 6.3e29 & 8.9e6 & 5.2e-4 + 6ne & 0.011 @xmath80.004 & 2.9e29 & 5.2e6 & 7.1e-4 + 6sw & & 3.1e29 & 3.7e6 & 7.3e-4 + 7 & 0.017 @xmath80.004 & 1.1e30 & 2.9e7 & 1.1e-3 + 8 & 0.010
@xmath80.003 & 9.6e29 & 9.2e6 & 1.1e-3 + 9 & 0.009 @xmath80.003 & 8.9e29 & 1.4e7 & 1.5e-3 + | we report on optical identifications of x - ray sources in the vicinity of the r cra association .
we identify 11 low - mass pre - main - sequence stars as counterparts of 9 x - ray sources .
we also find x - ray emission from coordinates consistent with the position of 3 bright late - b stars , although the emission may come from lower mass companions .
the x - ray - selected stars lie along a narrow locus in the luminosity - temperature diagram at an age of about 7 myr , which is considerably older than the estimated ages of the higher mass stars or of the ir - excess stars .
we determine the physical characteristics of these stars , including masses , ages , and lithium abundances .
we estimate that the complete membership of the r cra association amounts to about 90 stars , mostly older naked t tauri stars , and that the population is consistent with the standard imf . |
* the background and the problem . * broadcasting and gossiping are fundamental communication tasks in networks . in broadcasting ,
one node of a network , called the _ source _ , has a message that must be learned by all other nodes . in gossiping
, every node has a ( possibly different ) input message , and all messages must be learned by all nodes .
we study these well - researched tasks in a very weak communication model , called the _ beeping model_. communication proceeds in synchronous rounds . in each round
, a node can either listen , i.e. , stay silent , or beep , i.e. , emit a signal .
a node hears a beep in a round , if it listens in this round and if one or more adjacent nodes beep in this round .
the beeping model has been introduced in @xcite for vertex coloring and used in @xcite to solve the mis problem .
the beeping model is widely applicable , as it makes small demands on communicating devices , relying only on carrier sensing .
in fact , as mentioned in @xcite , beeps are an even weaker way of communicating than using one - bit messages , as the latter ones allow three different states ( 0,1 and no message ) , while beeps permit to differentiate only between a signal and its absence .
the network is modeled as a simple connected undirected graph .
initially all nodes are dormant .
the adversary wakes up the source in the case of broadcasting and some nonempty subset of nodes , at possibly different times , in the case of gossiping .
a woken up node starts executing the algorithm .
a dormant node is woken up by a beep of any neighbor .
our aim is to provide fast deterministic algorithms for broadcasting and gossiping in the beeping model .
the time of broadcasting is defined as the number of rounds between the wakeup of the source and the round in which all nodes of the network acquire the source message .
the time of gossiping is defined as the number of rounds between the wakeup of the first node and the round in which all nodes acquire all messages .
messages are considered as binary strings and the size of a message is the length of this string . in the case of broadcasting ,
our algorithm does not assume any information about the network , and it does not require any labeling of nodes . in the case of gossiping , we assume that all nodes have different labels from the set @xmath0 and that they know @xmath7 .
moreover , we assume that all nodes know the same upper bound @xmath1 on the size of the network and the same upper bound @xmath4 on the size of all input messages . without loss of generality
we may assume that @xmath8 .
indeed , the parameter @xmath7 known to nodes is an upper bound on the size of the network , as all nodes have different labels .
let @xmath2 be the diameter of the network , initially unknown to the nodes .
* our results .
* for the task of broadcasting we give an algorithm working in time @xmath5 for arbitrary networks , where @xmath3 is the size of the source message .
this complexity is optimal .
for the task of gossiping we give an algorithm working in time @xmath6 for arbitrary networks . due to space restrictions
several proofs are moved to the appendix
. * related work . * broadcasting and gossiping have been studied in various models for over four decades .
early work focused on the telephone model , where in each round communication proceeds between pairs of nodes forming a matching , and nodes that communicate exchange all previously acquired information .
deterministic broadcasting in this model has been studied , e.g. , in @xcite and deterministic gossiping in @xcite . in @xcite
the authors studied randomized broadcasting . in the telephone model studies focused on the time of the communication task and on the number of messages it uses .
early literature on communication in the telephone and related models is surveyed in @xcite .
fault tolerant aspects of broadcasting and gossiping are surveyed in @xcite . more recently , broadcasting and gossiping have been studied in the radio model .
while radio networks model wireless communication , similarly as the beeping model , in radio networks nodes send entire messages of some bounded , or even unbounded size in a single round , which makes communication drastically different from that in the beeping model .
the focus in the literature on radio networks was usually on the time of communication .
deterministic broadcasting in the radio model was studied , e.g. , in @xcite and deterministic gossiping in @xcite .
randomized broadcasting was studied in @xcite and randomized gossiping in @xcite .
the book @xcite is devoted to algorithmic aspects of communication in radio networks .
randomized leader election in the radio and in the beeping model was studied in @xcite .
deterministic leader election in the beeping model was studied in @xcite .
the authors showed an algorithm working in time @xmath9 in networks of diameter @xmath2 with labels polynomial in the size @xmath10 of the network .
in this section we consider the simpler of our two communication tasks , that of broadcasting .
even for this easier task , the restrictions of the beeping model require the solution of the basic problem of detecting the beginning and the end of the transmitted message . the naive idea would be to adapt the method of beeping waves , used in @xcite in a different context , and transmit a message by coding bit 1 by a beep and bit 0 by silence , other nodes relaying these signals after getting them .
however , in this coding there is no difference between message @xmath11 and message @xmath12 because both these messages are coded by a single beep .
hence we need to reserve some sequence of beeps to mark the beginning and end of a message , and code bits by some other sequences of beeps and silent rounds .
one way of defining such a coding is the following .
consider the message @xmath13 that has to be transmitted by some node @xmath14 .
let @xmath15 denote a round in which @xmath14 beeps , and let @xmath16 denote a round in which @xmath14 is silent .
the beginning and end of the message are marked by the sequence @xmath17 , bit @xmath18 is coded by @xmath19 , and bit @xmath20 is coded by @xmath21 .
hence message @xmath22 is transmitted as the sequence @xmath23 , where + @xmath24 , for @xmath25 , and + @xmath26 and @xmath27 , if @xmath28 , for @xmath29 , + @xmath30 and @xmath31 , if @xmath32 , for @xmath29 .
a node @xmath33 hearing the sequence @xmath23 of beeps and silent rounds can correctly decode message @xmath22 as follows . upon hearing two beeps in the first two rounds it divides the successive rounds into segments of
length 2 and records all beeps and silent rounds until a segment with two consecutive beeps .
each segment between the two segments @xmath17 is either of the from @xmath19 or of the form @xmath21 .
the node @xmath33 decodes each segment @xmath19 as 1 and each segment @xmath21 as 0 . in this way , message @xmath22
is correctly reconstructed .
note that the time of transmitting this message is @xmath34 and hence linear in its size .
we will call the above sequence of beeps and silent rounds , chosen by a node @xmath14 for the message @xmath22 , the _ canonical sequence _ for @xmath22 transmitted by node @xmath14 . the way of reconstructing message @xmath22 by node @xmath33
is called the _ canonical decoding_. using canonical sequences we formulate our broadcasting algorithm that works for arbitrary networks .
* algorithm broadcast * let @xmath35 be the round in which the source is woken up . given a source message @xmath22
, the source transmits the canonical sequence for @xmath22 in rounds @xmath36 , for @xmath37 , and stops . more precisely , if @xmath38 , then the source beeps in round @xmath36 , and if @xmath39 , then the source is silent in this round . in all other rounds the source is silent .
every other node @xmath33 that is woken up by a beep in round @xmath40 , beeps in round @xmath41 .
if it hears a beep in a round @xmath42 , for some positive integer @xmath43 , it beeps in round @xmath44 . in all other rounds
node @xmath33 is silent .
it divides all rounds @xmath42 , for @xmath45 , into segments of length 2 .
after the second segment when it hears @xmath17 , the node decodes the source message using the canonical decoding of the sequence received in rounds @xmath42 , for @xmath45 . then it beeps in the next round and stops .
@xmath46 [ broad - trees ] algorithm broadcast is a correct broadcasting algorithm working in any network of diameter @xmath2 in time @xmath5 , where @xmath3 is the size of the source message .
the time complexity of this algorithm is optimal .
consider any network and define its @xmath47th layer as the set of nodes at distance @xmath47 from the source .
consider any node @xmath33 in the @xmath47th layer of this network , other than the source , and suppose that @xmath33 is woken up in round @xmath40 .
hence nodes in layer @xmath48 beep only in rounds @xmath42 , nodes in layer @xmath47 beep only in rounds @xmath44 , and nodes in layer @xmath49 beep only in rounds @xmath50 , for some integer @xmath43 .
consequently , in rounds @xmath42 the node @xmath33 hears a beep in exactly these rounds in which nodes from the layer @xmath48 beep .
by induction on the distance of @xmath33 from the source we get that @xmath33 gets correctly the canonical sequence for the source message and hence decodes the source message correctly . if the source is woken up in round @xmath35 and @xmath33 is at distance @xmath47 from the source , it starts receiving transmissions in round @xmath51 and stops in round @xmath52 .
since @xmath53 , it follows that the algorithm works in time @xmath5 . in order to prove that this time complexity is optimal , it is enough to show that every algorithm requires at least time @xmath54 and at least time @xmath55 .
the first fact is immediate because the farthest node from the source must be at distance at least @xmath56 from it , and no signal can get to this node from the source faster than after @xmath56 rounds .
the second fact holds even in the two - node network .
suppose that some broadcasting algorithm transmits every message of size @xmath3 from one node of this network ( the source ) to the other , in time @xmath57 .
the number of sequences of length @xmath58 with terms from the set @xmath59 is @xmath60 .
since the number of possible source messages of size @xmath3 is @xmath61 , it follows that for two distinct source messages the source must behave identically , and hence the input of the other node is identical .
hence the other node must decode the same message in both cases , which contradicts the correctness of the algorithm . @xmath62
in this section we investigate the more complex task of gossiping . one way to accomplish
this task is to have each node broadcast its input message .
however , in the highly contrived beeping model , periods of broadcasting by different nodes should be disjoint , otherwise messages , transmitted as series of beeps and silent rounds , risk to become damaged , when a node receives simultaneously a beep being part of the transmission of one message and should hear a silent round being part of the transmission of another message .
the receiving node will then just hear the beep and the transmission of the message requiring this round to be silent becomes scrambled . in order to broadcast in disjoint time intervals
, nodes must establish an order between them and reserve the @xmath47th time interval to the broadcast of the @xmath47th node in this order .
this yields the following high - level idea of a gossiping algorithm .
first we establish a procedure that finds the node with the largest label .
this is done in such a way that all nodes learn the largest label .
( notice that we can not use the leader election algorithm from @xcite because this algorithm works only under an additional strong assumption that all nodes that are woken by the adversary and not by hearing a beep are woken simultaneously in the first round of the algorithm execution . ) next , using this elected leader , all nodes are synchronized : they agree on a common round , and hence can simultaneously start the rest of the algorithm execution .
then the procedure of finding the largest - labeled node is repeated at most @xmath1 times , where @xmath1 is an upper bound on the size of the network , each time the currently found largest - labeled node withdrawing from the competition . in this way , after at most @xmath1 repetitions all nodes know the order between them , and subsequently they broadcast their values in disjoint time intervals , in this order .
we assume that all nodes know the size @xmath7 of the label space .
let @xmath63 and let @xmath64 be the binary representation of the integer @xmath65 .
we assume that all sequences @xmath64 are of length @xmath66 , the representations being padded by a prefix of 0 s , if necessary .
hence the representation of a smaller integer is lexicographically smaller than the representation of a larger integer .
we also assume that all nodes know a common upper bound @xmath1 on the number of nodes in the network .
our first procedure finds the largest label among a set @xmath67 called the set of _ participating nodes_. * procedure find max * when a node is woken up in round @xmath68 , either by the adversary or by a beep , it defines the following round numbers : @xmath69 , for @xmath70 .
then , for @xmath71 , the node defines the time interval @xmath72 $ ] .
note that these intervals are pairwise disjoint .
the node beeps in round @xmath73 .
the rest of the procedure is divided into @xmath66 stages , corresponding to time intervals @xmath74 , for @xmath71 .
first assume that the node is participating .
let @xmath75 be the binary representation of the node s label , of length @xmath66 .
in the beginning of the first stage the node is _
active_. if the node is still active at the beginning of the @xmath43th stage , then it behaves as follows .
if @xmath76 , the node listens in all rounds of the time interval @xmath74 until it hears a beep . if it does not hear any beep , it remains active and proceeds to stage @xmath77 .
if it hears a beep for the first time in some round @xmath35 , then it beeps in round @xmath78 and becomes non - active .
if @xmath79 , the node listens until it hears a beep or until round @xmath80 , whichever comes earlier .
if it hears a beep in some round @xmath81 , it beeps in round @xmath78 , listens till the end of time interval @xmath74 and remains active .
otherwise , it beeps in round @xmath80 , listens till the end of time interval @xmath74 and remains active .
if the node is non - active at the beginning of the @xmath43th stage , then it listens in all rounds of the time interval @xmath74 until it hears a beep .
if it does not hear any beep , it remains non - active and proceeds to stage @xmath77 .
if it hears a beep for the first time in some round @xmath35 , then it beeps in round @xmath78 , listens till the end of time interval @xmath74 and remains non - active .
a non - participating node is never active .
it listens in all rounds of the time interval @xmath74 until it hears a beep .
if it does not hear any beep , it remains non - participating and proceeds to stage @xmath77 .
if it hears a beep for the first time in some round @xmath35 , then it beeps in round @xmath78 , listens till the end of the stage and remains non - participating . at the end of stage @xmath66 ,
the ( unique ) participating node that remained active till the end of this stage is the node with the largest label among participating nodes .
@xmath46 [ find - max ] at the end of the execution of procedure find max , there is exactly one active participating node .
this node has the largest label among participating nodes .
all nodes know the label of this node .
procedure find max takes time @xmath82 .
the goal of our next procedure is synchronizing all processors .
the procedure will be used upon completion of procedure find max , and hence we assume that the largest label is known to all nodes .
let @xmath83 be the node with the largest label . upon completion of procedure synchronization
, each node declares a specific round to be _ red _ , and this round is the same for all nodes . *
procedure synchronization * each node other than @xmath83 has an integer variable @xmath84 initially set to 0 .
let @xmath85 .
for every integer @xmath86 , let @xmath87 be the binary representation of @xmath47 of length @xmath88 , padded by a prefix of 0 s , if necessary .
a string @xmath87 will be transmitted by nodes of the network , using the canonical sequence , as it was done for broadcasting in section 2 .
we briefly recall this coding .
let @xmath15 denote a round in which @xmath14 beeps , and let @xmath16 denote a round in which @xmath14 is silent .
in @xmath89 rounds , node @xmath14 transmits the following message @xmath90 . the beginning and end of the message
are marked by the sequence @xmath17 , every bit @xmath18 of @xmath87 is coded by @xmath19 , and every bit @xmath20 of @xmath87 is coded by @xmath21 . at the beginning of the procedure node @xmath83
transmits @xmath91 starting in round @xmath78 , where @xmath35 is the round in which @xmath83 completed the execution of procedure find max . after completing this transmission
, node @xmath83 waits till round @xmath92 and declares it to be _
red_. every node other than @xmath83 that is at @xmath84 0 waits until it hears two consecutive beeps .
then it partitions the following rounds into consecutive segments of length 2 , and decodes each segment of the form @xmath19 as 1 and each segment of the form @xmath21 as 0 .
as soon as it hears a segment @xmath93 consisting of two beeps , it considers the previously decoded string of bits as the binary representation of an integer @xmath43 .
it changes the value of its variable @xmath84 to @xmath77 , transmits @xmath94 in @xmath89 rounds , starting in the round @xmath40 following the segment @xmath93 , then waits till round @xmath95 and declares it to be _
red_. @xmath46 [ synch ] all nodes declare the same round to be _
red_. procedure synchronization takes time @xmath96 .
we prove the following invariant by induction on @xmath97 . 1 . in time interval
@xmath98 $ ] the only message transmitted is @xmath99 and it is transmitted by all nodes at distance @xmath97 from node @xmath83 and only by these nodes ; 2 .
a node at distance @xmath97 from the node @xmath83 declares round @xmath92 as _
red_. the invariant is clearly satisfied for @xmath100 .
suppose that it holds for some @xmath101 .
the only nodes that hear the beeps transmitted in time interval @xmath98 $ ] are those at distance at most @xmath102 from node @xmath83 .
the only nodes among them that have value of @xmath84 0 are nodes at distance exactly @xmath102 from node @xmath83 .
since all the nodes at distance @xmath97 from @xmath83 transmit the same message @xmath99 in this time interval , they all beep exactly in the same rounds of the interval .
hence the value of @xmath97 is correctly decoded by all nodes at distance @xmath102 from node @xmath83 .
these nodes , and only these nodes , transmit @xmath103 in the time interval @xmath104 $ ] .
this proves the first part of the invariant .
all these nodes set their value of @xmath84 to @xmath102 and declare as _ red _ the round @xmath105 , where @xmath40 is the last round of the preceding time interval , i.e. , @xmath106 .
hence the declared round is @xmath107 , which proves the second part of the invariant .
this implies , in particular , that part 2 of the invariant is true for nodes at any distance @xmath97 from node @xmath83 , and hence all nodes of the network declare the same round as _
red_. since there are at most @xmath1 time intervals used in the procedure and each of them has length @xmath108 , the entire procedure takes time @xmath96 .
@xmath62 as we mentioned at the beginning of this section , we want to use our broadcasting algorithm many times , each time starting from a different node . in order to take advantage of the efficiency of broadcasting , which depends on the diameter and not on the size of the network , all nodes need to have a linear upper bound on the diameter of the network .
note , that in order to accomplish one execution of this algorithm , from one source node , no such upper bound was needed .
it becomes needed for multiple broadcasts , as we want to execute each of them in a separate time interval , and thus we need a good estimate of the time of each execution . recall , that we assume knowledge of a bound @xmath1 on the size of the network but no knowledge of any such bound on the diameter .
clearly @xmath1 is an upper bound on the diameter as well , but may be vastly larger than the diameter .
the following procedure is devoted to obtaining a linear upper bound on the diameter @xmath2 of a network .
it will be executed after the execution of procedure find max and procedure synchronization .
hence we may assume that the largest of all labels is known to all nodes .
let @xmath83 be the node with this label .
we may also assume that all nodes declared the same round @xmath40 as _ red_. moreover , each node other than @xmath83 has its variable @xmath84 set to its distance from @xmath83 ( this is done in procedure synchronization ) .
* procedure diameter estimation * each node defines consecutive time intervals @xmath109 $ ] , for positive integers @xmath43 . in time interval @xmath110 each node at level @xmath111 beeps in round @xmath111 of this interval . for @xmath112 ,
if a node at @xmath84 @xmath111 heard a beep in round @xmath113 of interval @xmath114 , then it beeps in round @xmath111 of interval @xmath115 . in the first round @xmath16 of the form @xmath116
in which @xmath83 does not hear a beep , it sets @xmath117 .
let @xmath22 be the binary representation of the integer @xmath118 and let @xmath3 be the size of message @xmath22 .
all nodes execute algorithm broadcast with node @xmath83 as the source and message @xmath22 as the source message . in this execution
the role of round @xmath35 in which @xmath83 is woken up is played by round @xmath119 .
every node decodes the integer @xmath120 as an upper bound on the diameter @xmath2 of the network .
all nodes declare round @xmath121 as _ blue_. @xmath46 [ diamest ] upon completion of procedure diameter estimation all nodes have the same linear upper bound @xmath122 on the diameter of the graph .
they all declare the same round as _ blue _ , and procedure diameter estimation is completed by this round .
procedure diameter estimation takes time @xmath123 .
if nodes know a linear upper bound @xmath122 on the diameter of the network , procedure find max can be modified to work faster .
the modifications are detailed below .
* procedure modified find max * in procedure find max replace the wakeup round @xmath68 by some round @xmath124 , given as input in its call .
round @xmath125 will be called the _ starting round _ of the procedure .
let @xmath126 instead of @xmath69 , for @xmath70 .
let @xmath127 $ ] instead of @xmath72 $ ] , for @xmath71 .
the rest of procedure find max remains unchanged .
@xmath46 the proof of the following proposition is the same as that of lemma [ find - max ] , using the above modifications .
[ prop1 ] at the end of the execution of procedure modified find max , there is exactly one active participating node .
this node has the largest label among participating nodes .
all nodes know the label of this node .
procedure modified find max takes time @xmath128 .
using procedure modified find max we now establish the order between all nodes as follows . the procedure will be called after the execution of procedure find max , procedure synchronization and procedure diameter estimation .
hence we assume that @xmath83 is the node with the largest label , found by procedure find max , and that @xmath15 is the common _ blue _ round found by all nodes in procedure diameter estimation .
all nodes start procedure ordering in round @xmath129 .
let @xmath130 be an upper bound on the duration of procedure modified find max , established in proposition [ prop1 ] .
* procedure ordering * @xmath131 the set of all nodes except node @xmath83
node @xmath83 assigns itself rank 0 * for * @xmath132 * to * @xmath1 * do * + execute procedure modified find max in the time interval + @xmath133 $ ] , with the set @xmath134 of participating nodes ; + the node found in the current execution of procedure modified find max + removes itself from @xmath134 and assigns itself @xmath135 .
@xmath46 [ ranks ] ranks assigned to nodes in the execution of procedure ordering define a strictly decreasing function of the node labels .
procedure ordering is completed in round @xmath136 and takes time @xmath137 .
since after each execution of procedure modified find max , the node with the largest label among participating nodes is found , and this node stops participating in the following executions , the rank assigned to the @xmath43th largest node is @xmath138 .
since there are @xmath1 time intervals , each of length @xmath139 , the time estimate follows .
@xmath62 we are now ready to formulate a gossiping algorithm working for arbitrary networks .
let @xmath4 be the upper bound on the size of all input messages , known to all nodes .
let @xmath140 be the upper bound on the duration of algorithm broadcast established in theorem [ broad - trees ] , for the value @xmath122 of the diameter and for the size @xmath4 of the message .
* algorithm gossiping * 1 .
execute procedure find max ; let @xmath83 be the node with the largest label ; 2 .
execute procedure synchronization ; let @xmath40 be the _ red _ round found in this procedure ; 3 .
execute procedure diameter estimation starting in round @xmath41 ; let @xmath122 be the upper bound on the diameter of the network found in this procedure ; let @xmath15 be the _ blue _ round found in procedure diameter estimation ; 4 .
execute procedure ordering starting in round @xmath129 ; 5 .
let @xmath74 be the time interval @xmath141 $ ] ; 6 . in time
interval @xmath74 execute algorithm broadcast with node of rank @xmath43 , found in procedure ordering , as the source , and the input message of this node as the source message .
( for each @xmath43 , the role of the round @xmath35 when the source is woken up is played by the first round of interval @xmath74 . ) @xmath46 * remark . *
note that , in the first step of the algorithm , we have to use procedure find max instead of the more efficient procedure modified find max because at this point the only estimate on the diameter , known to nodes , is @xmath142 .
however , since this procedure is executed only once , it does not increase the time complexity of the entire algorithm .
algorithm gossiping is a correct gossiping algorithm working in any network of diameter @xmath2 with at most @xmath1 nodes in time @xmath6 , where @xmath4 is an upper bound on the size of all input messages , known to all nodes . by lemma [ find - max ] ,
procedure find max correctly finds the node @xmath83 with the largest label . by lemma [ synch
] , all nodes compute the same round @xmath40 , and hence start procedure diameter estimation in the same round @xmath41 . by lemma
[ ranks ] , there is at most one node with any rank @xmath143 . by lemma [ diamest ]
, @xmath122 is a linear upper bound on the diameter of the network .
hence , in view of theorem [ broad - trees ] , the upper bound @xmath144 is indeed an upper bound on the time of execution of algorithm broadcast starting from any source node .
this implies that all nodes broadcast their messages in pairwise disjoint time intervals , and hence all broadcasts are correct , in view of theorem [ broad - trees ] .
this proves the correctness of algorithm gossiping .
it remains to estimate the execution time of the algorithm .
procedure find max takes time @xmath82 .
procedure synchronization takes time @xmath96 .
procedure diameter estimation takes time @xmath145 .
procedure ordering takes time @xmath137 . at most @xmath1 executions of algorithm broadcast in step 6 of algorithm gossiping
take time @xmath146 .
hence algorithm gossiping takes time @xmath6 .
@xmath62 we conclude with the following lower bound on the time of gossiping that holds even for the class of trees of bounded diameter .
[ lb ] assume that all input messages have size @xmath147 for some constant @xmath148 .
then there exist bounded diameter trees of size @xmath149 for which every gossiping algorithm takes time @xmath150 .
we considered two basic communication tasks , broadcasting and gossiping , in the arguably weakest communication model , in which in every round each node has only the choice to beep or to listen .
for the task of broadcasting , we proposed an optimal algorithm working in time @xmath5 for arbitrary networks of diameter @xmath2 , where @xmath3 is the message size .
for the task of gossiping , we presented an algorithm working in time @xmath6 for arbitrary networks of diameter @xmath2 with at most @xmath1 nodes . here
@xmath4 denotes an upper bound on the size of all input messages , known to all nodes .
it remains open if this complexity can be improved in general .
note however , that our gossiping algorithm has optimal time for networks of diameter bounded by a constant , if the following two assumptions are satisfied : the size of all input messages is the same , it is known to all nodes , and satisfies @xmath147 for some constant @xmath148 , and the size @xmath7 of the label space is polynomial in @xmath1 . indeed , in this case @xmath151 and @xmath152 . since for bounded @xmath2 , we have @xmath153 , algorithm gossiping works in time @xmath154 in this case , which matches the lower bound @xmath150 , shown in proposition [ lb ] .
the two above assumptions do not seem to be overly strong .
indeed , in most applications , we want messages exchanged by gossiping nodes to be large enough to hold at least the node s label and some other useful information , which justifies the assumption @xmath147 for some constant @xmath148 . on the other hand , labels of nodes are often assumed to be polynomial in the size of the network , as there is usually no need of larger labels .
notice , moreover , that if these assumptions are satisfied , the complexity of our gossiping algorithm is @xmath155 , i.e. , it exceeds the lower bound @xmath150 only by a factor of @xmath2 , for any value of the diameter @xmath2 .
thus , our gossiping algorithm is not only optimal for networks of bounded diameter , but it is close to optimal for `` shallow '' networks , e.g. , networks whose diameter is logarithmic in their size , such as the hypercube .
let @xmath156 be the earliest round in which some node is woken up by the adversary .
( there may be several nodes woken up in round @xmath156 ) . since each node beeps in the round following its wakeup ( either by the adversary or by a beep ) , all nodes are woken up until round @xmath157 . in order to prove the claim , first note that stage @xmath43 of node @xmath14 starts after round @xmath158 . for any node @xmath159 and any non - negative integer @xmath47 ,
let @xmath160 denote the round @xmath161 computed by node @xmath159 ( relative to its wakeup round ) .
since non - active nodes beep in some round @xmath162 only if they heard a beep in round @xmath163 , it follows that if @xmath14 heard a beep in some round @xmath35 of its stage @xmath43 , then an active node @xmath33 must have beeped in some round @xmath164 , such that no node beeped in round @xmath165 . according to the procedure ,
the only reason for such a beep is that , for some @xmath166 , @xmath167 , node @xmath33 is active in its stage @xmath166 , and that @xmath168 , where @xmath75 is the binary representation of the label of @xmath33 .
suppose that @xmath169 .
this implies @xmath170 , hence @xmath171 .
hence the wakeup rounds of nodes @xmath14 and @xmath33 differ by more than @xmath1 , which is a contradiction
. thus @xmath166 can not be smaller than @xmath43 . for similar reasons
, @xmath166 can not be larger than @xmath43 .
this leaves the only possibility of @xmath172 , which proves the claim .
next , we prove that the node @xmath83 with the largest label among participating nodes remains active till the end of its stage @xmath66 .
let @xmath173 be the binary representation of this label .
suppose , for contradiction , that @xmath83 becomes passive in some stage @xmath174 .
according to the procedure , this implies that it heard a beep in stage @xmath43 and that @xmath175 . in view of the claim ,
there is a node @xmath33 active in stage @xmath43 , with the binary representation @xmath176 of its label , such that @xmath79 .
consider any index @xmath169 . if @xmath177 then also @xmath178 .
otherwise , since node @xmath83 is active in stage @xmath166 , it beeps in stage @xmath166 and hence node @xmath33 would become passive in stage @xmath166 , which contradicts the fact that it became passive only in stage @xmath43 .
this proves that the sequence @xmath173 is lexicographically smaller than the sequence @xmath176 , which contradicts the assumption that @xmath83 has the largest label among participating nodes .
further , we prove that no node other than the node @xmath83 with the largest label among participating nodes remains active till the end of its stage @xmath66 .
let @xmath33 be such a node with the binary representation @xmath176 of its label , and let @xmath16 be the first index where @xmath179 . since @xmath83 is active in stage @xmath16 and @xmath180 ,
node @xmath83 beeps in stage @xmath16 , and hence @xmath33 hears it and becomes passive ( at the latest ) in stage @xmath16 .
it follows that @xmath83 is the only participating node that is active at the end of the execution of procedure find max .
node @xmath83 knows that it remained active at the end , so it knows that it has the largest label .
every other node @xmath33 ( it is passive at the end of the procedure execution ) deduces the binary representation @xmath173 of the label of @xmath83 as follows : @xmath181 , if and only if @xmath33 heard a beep in stage @xmath47 . indeed , if @xmath181 then it beeped in stage @xmath47 , because it was active in this stage , and hence @xmath33 heard a beep at most @xmath1 rounds later , hence still in its stage @xmath47 .
if @xmath182 , then no node beeped in stage @xmath47 because all nodes that have 1 as the @xmath47th term of the binary representation of their label must be already passive in stage @xmath47 , as this representation is lexicographically smaller than @xmath173 . let @xmath184 be the largest distance of any node from @xmath83 .
for every node other than @xmath83 , the value of the variable @xmath84 is its distance from @xmath83 .
hence in time interval @xmath115 exactly nodes at distance at most @xmath185 from @xmath83 beep .
it follows that @xmath186 is the first round of the form @xmath116 in which @xmath83 does not hear a beep
. consequently @xmath187 . by the correctness of algorithm broadcast
, all nodes correctly decode the integer @xmath120 . since @xmath156 is the largest distance of any node from @xmath83 ,
the diameter @xmath2 of the network satisfies inequalities @xmath188 . thus @xmath122 is a linear upper bound on the diameter of the network . after decoding the integer @xmath122 ,
all nodes know @xmath1 , @xmath40 , @xmath156 and @xmath3 .
hence the round declared as _ blue _ is the same for all nodes .
it was follows from the proof of theorem [ broad - trees ] that algorithm broadcast takes time at most @xmath189 . since @xmath83 starts broadcasting in round @xmath119 , by round _
blue _ the procedure is completed .
it takes time @xmath190 .
consider the star @xmath67 with @xmath191 leaves , i.e. , a tree with one node @xmath14 of degree @xmath191 and with @xmath191 nodes of degree 1 .
@xmath67 is a tree of diameter 2 .
let @xmath33 be any leaf .
consider any algorithm @xmath192 accomplishing gossiping in @xmath67 in time @xmath35 .
the leaf @xmath33 can obtain information only from node @xmath14 . in time
@xmath35 node @xmath14 can transmit @xmath193 sequences with terms in the set @xmath59 , where @xmath15 denotes a beep and @xmath16 denotes silence .
consider the family @xmath194 of possible sets of input messages initially held by the @xmath191 nodes of @xmath67 other than @xmath33 , assuming that each node has a different message of size @xmath4 .
if @xmath195 , then node @xmath14 executing algorithm @xmath192 must generate the same sequence with terms in the set @xmath59 for two distinct sets of messages initially held by the @xmath191 nodes of @xmath67 other than @xmath33 , and consequently @xmath33 can not correctly deduce the set of messages held by other nodes of @xmath67 .
this implies that @xmath196 . | broadcasting and gossiping are fundamental communication tasks in networks . in broadcasting ,
one node of a network has a message that must be learned by all other nodes . in gossiping
, every node has a ( possibly different ) message , and all messages must be learned by all nodes .
we study these well - researched tasks in a very weak communication model , called the _ beeping model_. communication proceeds in synchronous rounds . in each round
, a node can either listen , i.e. , stay silent , or beep , i.e. , emit a signal .
a node hears a beep in a round , if it listens in this round and if one or more adjacent nodes beep in this round .
all nodes have different labels from the set @xmath0 .
our aim is to provide fast deterministic algorithms for broadcasting and gossiping in the beeping model .
let @xmath1 be an upper bound on the size of the network and @xmath2 its diameter .
let @xmath3 be the size of the message in broadcasting , and @xmath4 an upper bound on the size of all input messages in gossiping .
for the task of broadcasting we give an algorithm working in time @xmath5 for arbitrary networks , which is optimal .
for the task of gossiping we give an algorithm working in time @xmath6 for arbitrary networks .
* keywords : * algorithm , broadcasting , gossiping , deterministic , graph , network , beep . at the time of writing this paper
we were unaware of the paper + a. czumaj , p. davis , communicating with beeps , arxiv:1505.06107 [ cs.dc ] which contains the same results for broadcasting and a stronger upper bound for gossiping in a slightly different model . |
the design of effective numerical methods for solving structured generalized eigenvalue problems has recently attracted a great deal of attention .
palindromic matrix polynomials arise in many applications @xcite .
an @xmath0 matrix polynomial of degree @xmath1 @xmath2 , @xmath3 , @xmath4 , @xmath5 , is said to be t - palindromic if @xmath6 for @xmath7 .
it is well - known @xcite , @xcite that the palindromic structure induces certain spectral symmetries : in particular if @xmath8 is an eigenvalue of @xmath9 then @xmath10 is also an eigenvalue of @xmath9 .
numerical solution methods are generally asked to preserve these symmetries .
the customary approach for polynomial eigenproblems consists in two steps : first @xmath9 is linearized into a matrix pencil @xmath11 , @xmath12 , and then the eigenvalues of @xmath13 are computed by some iterative solver .
the usual choice of the matrix qz algorithm applied to a companion linearization @xcite of @xmath9 is implemented in the matlab function _ polyeig_. an alternative solver based on the ehrlich - aberth root finding algorithm is proposed in @xcite for dealing with certain structured linearizations .
specifically the method of @xcite is designed to solve generalized tridiagonal eigenvalue problems but virtually , as shown below , it can be extended to several other rank structures . a generalization for tridiagonal quadratic eigenvalue problems
is presented in @xcite . a similar strategy using newton s iteration directly applied to compute the zeros of @xmath9
is pursued in @xcite .
modified methods for palindromic eigenproblems which are able to preserve their spectral symmetries have been proposed in several papers .
the construction of t - palindromic linearizations of palindromic eigenproblems is the subject of @xcite and @xcite , whereas numerical methods based on matrix iterations have been devised in @xcite , @xcite , @xcite and @xcite for computing the eigenvalues of these linearizations by maintaining the palindromic structure throughout the computation . to date , however , the authors are not aware of any specific adaptation of the root - finding based methods to palindromic structures .
the contribution of this paper is to fill the gap by developing a root finder specifically suited for t - palindromic matrix polynomials , with particular emphasis on the case of large degree .
t - palindromic polynomials of large even degree arise as truncation of fourier series in several applications such as spectral theory , filtering problems , optimal control and multivariate discrete time series prediction @xcite .
the polynomial root - finding paradigm is a flexible , powerful and quite general tool for solving both structured and unstructured polynomial eigenproblems . in its basic form it proceeds in four steps : 1 .
the matrix polynomial is represented in some convenient polynomial basis .
2 . the transformed polynomial is linearized .
the linearization is reduced in the customary hessenberg - triangular form .
a root - finding method is applied for approximating the eigenvalues of the ( reduced ) pencil .
this scheme has some degrees of freedom concerning the choice of the polynomial basis at step 1 and the choice of the linearization at step 2 which can be used to exploit both structural and root properties of the matrix polynomial .
the complexity heavily depends on the efficiency of the polynomial zero - finding method applied to the determinant of the pencil .
steps 2 and 3 are optional but can substantially improve the numerical and computational properties of the method .
some caution should be used at step 1 since the change of the basis could modify the spectral structure of the matrix polynomial .
the key idea we propose for the implementation of step 4 is the use of the jacobi formula @xcite .
we emphasize that , although in this paper we focus on palindromics and on a version of the method that is able to extract the palindromic spectral structure , this strategy may be used to address the most general case of an unstructured matrix polynomial eigenproblem , for instance by applying it to the companion linearization .
an analysis of the application of the method to a generic matrix polynomial will appear elsewhere .
in this paper we consider the polynomial root - finding paradigm for solving t - palindromic eigenproblems .
in particular , we address the main theoretical and computational issues arising at steps 1 , 2 and 4 of the previous scheme applied to t - palindromic matrix polynomials , and also we indicate briefly how to carry out the reduction at step 3 .
the proposed approach relies upon the representation and manipulation of t - palindromic matrix polynomials in a different degree - graded polynomial basis @xmath14 , namely the dickson basis , satisfying a three - term recurrence relation and defined by @xmath15 , @xmath16 and @xmath17 for @xmath18 . for the given t - palindromic polynomial @xmath9 of degree
@xmath19 we determine a novel polynomial @xmath20 , @xmath21 , @xmath22 , @xmath23 , with the property that if @xmath24 and @xmath25 are two distinct ( i.e. @xmath26 ) finite semi - simple eigenvalues of @xmath9 with multiplicity @xmath27 , then @xmath28 is a semi - simple eigenvalue for @xmath29 with multiplicity @xmath30 .
moreover , we find that @xmath31 ^ 2 = p(y ) \cdot p(y),\ ] ] where @xmath32 is a polynomial of degree @xmath33 at most .
solving the algebraic equation @xmath34 is at the core of our method for t - palindromic eigenproblems .
our computational experience in polynomial root - finding indicates that the ehrlich - aberth method @xcite , @xcite for the simultaneous approximation of polynomial zeros realizes a quite good balancing between the quality of convergence and the cost per iteration .
the main requirements for the effective implementation of the ehrlich - aberth method are both a fast , robust and stable procedure to evaluate the newton correction @xmath35 and a reliable criterion to stop the iteration .
concerning the first issue it is worth noting that @xmath36 and , therefore , the computation immediately reduces to evaluating the newton correction of @xmath37 .
a suitable structured linearization @xmath38 of @xmath29 can be obtained following @xcite which displays a semiseparable structure . in this way , in view of the celebrated jacobi formula@xcite @xmath39 , the newton correction can be evaluated by performing a qr factorization of @xmath38 , say @xmath40 , at low computational cost and fulfilling the desired requirements of robustness and stability . also , since @xmath41 for @xmath42 we obtain at no additional cost a reliable stop condition based on an estimate of the backward error given by higham and higham @xcite .
if @xmath1 , the degree of the matrix polynomial , is large with respect to @xmath43 , the size of its matrix coefficients , our approach looks appealing since with a smart choice of the starting points it needs @xmath44 operations , whereas the qz method makes use of @xmath45 operations .
the unpleasant factor @xmath46 in our cost estimate depends on the block structure of the linearization used in our current implementation and can in principle be decreased by performing the preliminary reduction of the linearization in hessenberg - triangular form as stated at step 3 of the basic scheme .
the reduction can be carried out by a structured method exploiting the semiseparable structure of the block linearization to compute a rank - structured hessenberg - triangular linearization .
incorporating the structured method in our implementation would finally lead to a fast method that outperforms the qz algorithm for large degrees and is comparable in cost for small degrees . the paper is organized as follows .
the theoretical properties of the considered linearizations of t - palindromic matrix polynomials expressed in the dickson basis are investigated in section [ teo ] and [ teo1 ] .
the derivation of the proposed eigenvalue method for t - palindromic eigenproblems is established in section [ main1 ] and [ main2 ] .
the complete algorithm is described in section [ alg ] .
numerical experiments are presented in section [ exp ] to illustrate the robustness of our implementation and to indicate computational issues and possible improvements of our algorithm compared with other existing methods .
finally , conclusion and future work are discussed in section [ end ] .
this preparatory section recalls some basic definitions , background facts and notations used throughout the paper . for @xmath47 let @xmath48 , @xmath49 , be constant matrices and consider the matrix polynomial @xmath50 .
the generalized polynomial eigenproblem ( pep ) associated to @xmath51 is to find an eigenvalue @xmath52 and a corresponding nonzero eigenvector @xmath53 satisfying @xmath54 in this paper , we will always suppose that @xmath51 is _ regular _
, i.e. its determinant does not identically vanish .
a _ linearization _ of @xmath51 is defined as a pencil @xmath55 , with @xmath56 , such that there exist unimodular polynomial matrices @xmath57 and @xmath58 for which @xmath59 moreover , if one defines the reversal of a matrix polynomial as rev@xmath60 , the linearization is said to be _ strong _ whenever rev@xmath61 is a linearization of rev@xmath62 . following the work of mackey , mackey , mehl and mehrmann @xcite , in the paper @xcite higham , mackey , mackey and tisseur study the two ( right and left ) _ ansatz vector linearization spaces _ : having introduced the vector @xmath63 , these spaces are defined as follows : @xmath64 it is shown in @xcite that almost every pencil in these spaces is a linearization , while in @xcite two binary operations on block matrices , called column shifted sum and row shifted sum , are first introduced and then used to characterize the above defined spaces .
on the other hand , in @xcite amiraslani , corless and lancaster consider linearizations of a matrix polynomial expressed in some polynomial bases different than the usual monomial one .
equation ( 7 ) in @xcite resembles closely the defining equation of @xmath65 .
the authors themselves stress this analogy , that suggests an extension of the results of @xcite to the case of different polynomial bases .
let @xmath66 be a basis for the polynomials of degree less than or equal to @xmath1 . in @xcite degree - graded bases that satisfy a three - terms recurrence relation ( for instance ,
orthogonal polynomials always do so ) are considered : @xmath67 the @xmath68 are obviously linked to the leading - term coefficients of the @xmath69 . specifically , calling @xmath70 such coefficients , one has that @xmath71 .
we wish to consider the expansion of the polynomial @xmath51 in this basis : @xmath72 we introduce the vector @xmath73 by generalizing the linearizations studied in @xcite , for each choice of @xmath74 two new ansatz vector linearization spaces can be defined : @xmath75 it is worth noticing that it is not strictly necessary for the new basis to be degree - graded , nor it is to satisfy a three - term recurrence relation .
in fact , it is sufficient that @xmath76 are linearly independent and have degree less than or equal to @xmath77 , so that there exists an invertible basis change matrix @xmath78 such that @xmath79 .
the basis is degree - graded if and only if @xmath78 is lower triangular . in the light of the above definitions it is immediately seen that the main results of @xcite , @xcite remain valid in the case of a more general polynomial basis . in particular the following result holds .
[ prop2 ] let @xmath80 @xmath81 .
the following properties are equivalent : * @xmath82 is a linearization of @xmath83 * @xmath82 is a strong linearization of @xmath83 * @xmath82 is regular _ proof .
_ it is a corollary of theorem 4.3 of @xcite .
in fact , any @xmath84 @xmath85 can be written as @xmath86 @xmath87 for some @xmath88 @xmath89 .
therefore , @xmath82 has each of the three properties above if and only if @xmath90 has the corresponding property .
@xmath91 this proposition guarantees that almost every ( more precisely , all but a closed nowhere dense set of measure zero ) pencil in @xmath92 @xmath81 is a strong linearization for @xmath83 . for a proof , see theorem 4.7 of @xcite .
the eigenvectors of @xmath82 are related to those of @xmath83 .
more precisely , @xmath93 is an eigenpair for @xmath82 if and only if @xmath94 is an eigenpair for @xmath83 .
moreover , if @xmath82 is a linearization then every eigenvector of @xmath82 is of the form @xmath95 for some eigenvector @xmath96 of @xmath83 .
a similar recovery property holds for the left ansatz vector linearizations .
these properties can be simply proved as in theorems 3.8 and 3.14 of @xcite , that demonstrate them for the special case @xmath97 . for the numerical treatment of palindromic generalized eigenproblems a crucial role
is played by the so - called dickson basis @xcite @xmath98 defined by @xmath99 if we consider the mapping @xmath100 ( which we will refer to as the dickson transformation or the dickson change of variable ) then @xmath101 for @xmath102 . for @xmath103 , we obtain that @xmath104 . from @xcite by choosing @xmath105 as the ansatz vector we find a suitable strong linearization of @xmath51 represented as in : in the next section we study the spectral modifications induced by the dickson change of variable that provide the basic link between palindromic matrix polynomials and matrix polynomials expressed in the dickson basis .
let us recall that if @xmath52 is an eigenvalue of @xmath51 then the set @xmath106 , @xmath107 is a jordan chain of length @xmath108 if @xmath109 and the following relations hold @xcite : @xmath110 where @xmath111 denotes the @xmath1-th derivative of @xmath51 evaluated at @xmath112 .
the case @xmath113 corresponds to the definition of an eigenvector .
the notion of a jordan chain can be extended to any matrix function @xmath114 whose determinant vanishes at @xmath52 , as long as @xmath58 is analytic in a neighborhood of @xmath52 .
in particular , the case of laurent polynomials is important for our investigations .
if the principal part of a laurent polynomial @xmath115 is a polynomial of degree @xmath1 in @xmath10 , then @xmath116 is a polynomial .
the following lemma relates the jordan chains of the two .
the proof is a straightforward application of the product differentiation rule .
[ laurent ] let @xmath115 be a ( laurent ) polynomial and @xmath117 for some natural number @xmath1 .
then the set @xmath106 is a jordan chain of length @xmath108 for @xmath51 associated to the eigenvalue @xmath118 if and only if @xmath106 is a jordan chain of length @xmath108 for @xmath115 associated to the same eigenvalue .
roughly speaking , lemma [ laurent ] makes us able to switch between regular and laurent polynomials without worrying about changes in eigenvalues and generalized eigenvectors .
actually , this result can be slightly generalized with the next lemma , which is just an adaptation of a well - known result in @xcite for the case where the four matrix functions that we are going to consider are polynomials . in order to prove the lemma , we recall @xcite that a vector polynomial @xmath119 is called a root polynomial of order @xmath120 corresponding to @xmath52 for the matrix polynomial @xmath51 if the following conditions are satisfied : @xmath121 obviously a root polynomial of order @xmath120 is defined up to an additive term of the form @xmath122 for any suitable vector polynomial @xmath123 .
it is possible to prove that @xmath124 if and only if @xmath125 is a jordan chain of length @xmath108 for @xmath51 at @xmath112 .
when @xmath118 , thanks to lemma [ laurent ] it is possible to extend the concept to laurent polynomials : if @xmath115 is a laurent polynomial whose singular part has degree @xmath1 as a polynomial in @xmath25 , then we say that @xmath119 is a root polynomial for @xmath115 if it is a root polynomial for @xmath126 .
[ analytic ] let @xmath127 , @xmath128 be ( laurent ) polynomials and @xmath129 , @xmath130 be two matrix functions with @xmath131 .
suppose that an open neighborhood @xmath132 of @xmath133 exists such that all the considered functions are analytic in @xmath132 , and also suppose that both @xmath134 and @xmath135 are invertible .
then @xmath52 is an eigenvalue for @xmath136 if and only if it is an eigenvalue for @xmath137 , and @xmath138 is a jordan chain of length @xmath108 for @xmath137 at @xmath52 if and only if @xmath139 is a jordan chain of length @xmath108 for @xmath136 at @xmath52 , where @xmath140 .
_ if @xmath127 and @xmath128 are classical polynomials then the thesis follows as in the proof of proposition 1.11 in @xcite after having represented @xmath129 and @xmath130 by their taylor series expansions . to deal with the laurent case ,
let @xmath141 and @xmath142 be the minimal integers such that @xmath143 and @xmath144 are classical polynomials .
just follow the previous proof for @xmath145 and apply lemma [ laurent].@xmath91 we are now in the position to prove a result for the dickson change of variable @xmath146 .
the following proposition shows that the number of jordan chains and their length at some eigenvalue @xmath147 ( for the sake of brevity , we shall use the expression _
jordan structure at @xmath147 _ ) is related to the jordan structures at @xmath52 and @xmath148 .
[ change ] let @xmath149 and let @xmath29 be a polynomial in @xmath150 , so that @xmath151 is a laurent polynomial in @xmath24 .
let first @xmath152 , @xmath153 , be a finite eigenvalue of @xmath29 .
then the jordan structure of @xmath29 at @xmath147 is equal to the jordan structure of @xmath154 at either @xmath52 or @xmath155 .
if on the contrary @xmath156 , then there is a jordan chain of length @xmath27 at @xmath157 if and only if there is a jordan chain of length @xmath30 at @xmath158 .
_ it is obvious that @xmath159 is an eigenvalue for @xmath29 if and only if both @xmath52 and @xmath148 are eigenvalues of @xmath154 .
let @xmath160 , where @xmath161 is the smith form ( @xcite,@xcite ) of @xmath29 .
define @xmath162 .
if @xmath163 are such that @xmath164 , @xmath165 , @xmath166 and @xmath167 are polynomials in @xmath24 , then we have the relation @xmath168 ; however , in general @xmath169 is not the smith form of @xmath170 .
nevertheless , it has the form @xmath171 where @xmath172 and @xmath173 . in other words ,
the @xmath174-s are palindromic polynomials with no roots at @xmath175 and such that @xmath176 divides @xmath177 for @xmath178 .
moreover , @xmath147 is a zero of multiplicity @xmath43 for @xmath179 if and only if both @xmath52 and @xmath148 are zeros of multiplicity @xmath43 for @xmath174 . to reduce @xmath169 into a smith form ,
we proceed by steps working on @xmath180 principal submatrices . in each step , we consider the submatrix @xmath181 , with @xmath182 .
if @xmath183 , then do nothing ; if @xmath184 , premultiply the submatrix by @xmath185 and postmultiply it by @xmath186 , where @xmath187 while @xmath188 and @xmath189 are such that @xmath190 ; the existence of two such polynomials is guaranteed by bezout s lemma , since @xmath191 is the greatest common divisor of @xmath192 and @xmath193 .
it is easy to check that both matrices are unimodular , and that the result of the matrix multiplications is @xmath194 . by subsequent applications of this algorithm
we thus conclude that the smith form of @xmath169 is @xmath195 .
it follows that the @xmath196th invariant polynomial of @xmath29 has a root of multiplicity @xmath197 at @xmath147 if and only if the @xmath196th invariant polynomial of @xmath169 has a root of multiplicity @xmath198 at @xmath153 and a root of multiplicity @xmath198 at @xmath155 . from lemma [ analytic ] ,
the jordan structures of @xmath170 are equal to those of @xmath169 .
the thesis follows from the properties of the smith form and from lemma [ laurent ] .
mutatis mutandis , a similar argument can be used to analyze the case of @xmath199 : notice in fact that @xmath200 is a factor of the @xmath196th invariant polynomial of @xmath29 if and only if @xmath201 is a factor of the @xmath196th invariant polynomial of @xmath169 .
we will now specialize our analysis to the case of a matrix polynomial with palindromic structure .
[ oddtoeven ] in this section , we will only treat the case of even degree palindromic matrix polynomials .
notice in fact that an odd degree palindromic may always be transformed to an even degree palindromic , either by squaring the variable ( @xmath202 ) or by multiplication by @xmath203 .
potentially , both actions may introduce problems : squaring the variable adds an additional symmetry @xmath204 to the spectrum while multiplying by @xmath205 increases by @xmath43 the multiplicity of @xmath206 as an eigenvalue .
however , the first issue may be solved , after passing to laurent form , by the use of the change of variable @xmath207 .
see also remark [ evenodd ] . regarding the latter issue , since one knows that he is adding n times @xmath206 there is no need to compute it : @xmath43 of the @xmath208 starting points of the ehrlich - aberth iteration shall be set equal to @xmath209 , and there they remain with no further corrections .
the shortcoming is that the jordan structure at @xmath210 changes .
let @xmath211 be a polynomial of even degree .
by lemma [ laurent ] , switching to the laurent form is not harmful for finite nonzero eigenvalues and the corresponding ( generalized ) eigenvectors ; we can therefore consider its laurent counterpart @xmath212 three different kinds of palindromic structure can be defined .
we say that the laurent polynomial is purely palindromic ( resp .
, @xmath213-palindromic , @xmath214 ) if the following relations hold between its matrix coefficients : @xmath215 it is well - known that the palindromic structure induces certain symmetries of eigenvalues and eigenvectors : in particular if @xmath52 is an eigenvalue , @xmath96 is a right eigenvector and @xmath216 is a left eigenvector , then , denoting complex conjugation with the operator @xmath217 @xmath218 in this paper we are primarily interested in the design of an efficient solver for @xmath219palindromic eigenproblems . a numerical method will be presented in subsection [ tpalin ] .
the proposed approach can however be described very easily with purely palindromic polynomials .
thus we first consider this case for the sake of clarity .
the most obvious way to deal with this kind of palindromicity is via introduction of the change of variable @xmath220 , in order to halve the degree of the polynomial .
more explicitly , one can define @xmath221 ; clearly , the purely palindromic structure of @xmath51 guarantees that @xmath222 is itself a polynomial in the new variable @xmath150 . the next proposition is a simple application of lemmas [ laurent ] and [ change ] , and it relates eigenvectors and jordan chains of the two polynomials : when @xmath223 , the jordan structure of @xmath222 at the eigenvalue @xmath224 is equal to the jordan structure of @xmath51 at either @xmath52 or @xmath148 . if @xmath156 , @xmath222 has a jordan chain of length @xmath27 at @xmath225 if and only if @xmath51 has a jordan chain of length @xmath30 at @xmath156 .
in particular , the eigenvectors of @xmath222 at @xmath147 are exactly the same of the eigenvectors of @xmath51 at @xmath52 ( or equivalently at @xmath148 , since they are the same ) .
albeit very attractive , from a numerical point of view this trick is not very suitable as soon as one considers a high degree polynomial .
in fact , the matrix coefficients of @xmath222 need to be computed as linear combinations of the ones of @xmath51 .
since the powers of a binomial are involved , the coefficients of these linear combinations would exponentially grow with the polynomial degree . to circumvent this difficulty , we shall make use of the dickson polynomials .
the polynomial @xmath222 is readily expressed in terms of the @xmath226s since in the dickson basis the coefficients are just the old ones and therefore no computation at all is needed , namely , @xmath227 the associated linearization has several computational advantages with respect to other customary linearizations of @xmath51 .
its size is @xmath228 versus @xmath229 , the spectral symmetries are preserved and , moreover , the linearization displays a semiseparable structure .
more precisely , it is of the form @xmath230 where @xmath231 is identity plus low rank while @xmath232 is hermitian plus low rank .
this kind of structure is preserved under the qz algorithm and it may be exploited for the design of an efficient and numerically robust root - finder applied to the algebraic equation @xmath233 .
consider now a t - palindromic polynomial of even degree @xmath234 .
we will suppose once more that neither @xmath175 nor @xmath235 are eigenvalues , so that we can divide by @xmath236 and consider the laurent form @xmath51 , which is a t - palindromic laurent polynomial of degree @xmath1 both in @xmath24 and in @xmath25 .
since the symmetry @xmath237 is still present in the spectrum , we expect that the dickson basis may still play a role .
however , unlike the purely palindromic case , it is not possible to directly express a t - palindromic polynomial as a polynomial in the variable @xmath150 .
in fact , splitting @xmath51 as the sum of its symmetric part and its skew - symmetric part we obtain that @xmath238.\ ] ] if we introduce the new variables @xmath239 and @xmath240 , then @xmath51 can be expressed as a bivariate polynomial in @xmath241 and @xmath150 which is always linear in @xmath241 , that is , @xmath242 the property follows from by substituting @xmath243 notice moreover that @xmath244 is a symmetric polynomial ( that is to say , every matrix coefficient is symmetric ) , @xmath245 is skew - symmetric , and the operation of transposition corresponds to changing the sign of @xmath241 , that is , @xmath246 in principle one may think of treating @xmath247 with available techniques for the bivariate eigenvalue problem ( see e.g. @xcite and references therein ) , but actually @xmath150 and @xmath241 are not independent .
they are related by the trigonometric dispersion relation @xmath248 .
this suggests that it is possible to obtain a univariate polynomial by doubling the dimensions of the matrix coefficients .
let us define @xmath249 then @xmath29 is a polynomial in @xmath150 of degree @xmath250 at most .
moreover , it has the following property : if @xmath52 and @xmath148 are two distinct ( i.e. @xmath153 ) finite semisimple eigenvalues of @xmath51 with multiplicity @xmath251 , then @xmath224 is a semisimple eigenvalue for @xmath29 with multiplicity @xmath252 . to see this ,
notice first that @xmath253 and @xmath254 hence , we find that @xmath255 since , as long as @xmath256 is defined ( that is to say @xmath257 or @xmath258 ) , @xmath259 then @xmath260 ^ 2 , \quad \forall \ ( y , w)\in \mathbb c\times \mathbb c.\ ] ] therefore , @xmath52 has algebraic multiplicity @xmath251 for @xmath51 if and only if @xmath147 has algebraic multiplicity @xmath252 for @xmath29 .
this gives the factorization @xmath261 for a suitable polynomial @xmath32 having the zero @xmath147 of multiplicity @xmath251 . concerning eigenvectors ,
if @xmath52 is semisimple , then let @xmath262 ( resp .
@xmath263 ) , @xmath264 be the eigenvectors for @xmath51 ( resp .
@xmath265 ) corresponding to @xmath52 : it can be easily checked that @xmath266 , where @xmath267 , are two linearly independent eigenvectors for @xmath29 corresponding to @xmath147 .
thus , geometric multiplicity is also @xmath252 .
indeed , something more can be said in the more general case of jordan chains .
[ jc ] let @xmath224 be an eigenvalue of @xmath29 so that @xmath52 and @xmath148 are eigenvalues for @xmath51 . if @xmath268 then the jordan structure of @xmath29 at @xmath147 is equal to the union of the jordan structures of @xmath51 at @xmath52 and at @xmath148 .
_ since @xmath51 is t - palindromic , it is clear that the jordan structure of @xmath269 at either @xmath52 or @xmath148 is the union of the jordan structures of @xmath51 at @xmath52 and at @xmath148 .
define @xmath270 the matrix function @xmath256 , defined in the previous page , is analytic everywhere in the @xmath241 complex plane but on a branch semiline passing through the origin . since by hypothesis @xmath271 , the branch cut can be always chosen in such a way that @xmath256 is analytic in a neighborhood of @xmath272 , and thus @xmath273 is analytic in a neighborhood of @xmath52
. then we can apply lemma [ analytic ] to conclude that the jordan structures of @xmath274 and @xmath275 are the same .
application of lemma [ change ] completes the proof .
@xmath91 [ ins1 ]
another remarkable property of @xmath29 is that its coefficients are all skew - hamiltonian , that is to say they can be written as @xmath276 where @xmath277 and @xmath278 is some skew - symmetric matrix .
this link between t - palindromic and skew - hamiltonian polynomials is interesting because it may shed more light on the relation between several polynomial structures .
it is known that one can easily transform a palindromic polynomial to an even polynomial by a cayley transformation , and then to a hermitian polynomial via a multiplication by @xmath196 ( if one started from a real polynomial ) or to a symmetric polynomial by squaring the matrix coefficients . on the other hand , hamiltonian polynomials
can lead to skew - hamiltonian polynomials by squaring each coefficients , and multiplication by @xmath279 sends a skew - hamiltonian polynomial to a skew - symmetric polynomial .
the dickson change of variable , followed by doubling the dimension , is able to map t - palindromic polynomials of even degree to a subset of skew - hamiltonian polynomials .
unlike some of the other mentioned maps , this is not a bijection between two classes of structured polynomials , because what is obtained is actually a subset of skew - hamiltonian polynomials .
in fact , since the north - west and south - east coefficients of @xmath29 are the coefficients of @xmath244 they must be symmetric and there is a relation between the north - east and south - west coefficients of @xmath29 .
however , a deeper investigation on this subject is needed in the future .
[ evenodd ] notice that a similar technique can be applied to _ even / odd _ matrix polynomials , that is polynomials whose coefficients alternate between symmetric and skew - symmetric matrices . in this case , on can apply the transformation @xmath280 and use algebraic manipulations , akin to the ones described for the t - palindromic case , in order to build a new polynomial in @xmath281 with double dimensions .
in the case of an odd - degree t - palindromic polynomial , the substitutions @xmath202 and @xmath282 lead to an @xmath29 such that @xmath283 is odd .
therefore , one may apply @xmath284 and build a third polynomial in order to extract the additional structure @xmath204 .
equation and enable the computation of the eigenvalues of @xmath51 to be reduced to solving algebraic equations . from proposition [ jc ]
it follows that possible discrepancies in the jordan structures can be expected for @xmath225 and @xmath285 corresponding to @xmath156 and @xmath286 , respectively . when @xmath156 not only the proof we gave is not valid ( because , since @xmath287 is a branch point , there is no neighborhood of analyticity of the matrix function @xmath288 ) , but in fact the proposition itself does not hold .
as a counterexample , let @xmath289 and consider the polynomial @xmath290 we have that @xmath291 is a jordan chain for @xmath51 at @xmath292 .
the corresponding @xmath29 is @xmath293 which has a semisimple eigenvalue at @xmath294 with the corresponding eigenvectors @xmath295 and @xmath296 . if the leading coefficient of @xmath51 is not symmetric , then @xmath29 has @xmath297 extra infinite eigenvalues , where @xmath43 is the dimension of the matrix coefficients of @xmath51
. these eigenvalues are defective since their geometric multiplicity is only @xmath298 , where @xmath299 is the leading coefficient of @xmath245 . for the numerical approximation of the roots of @xmath32 we can exploit again the properties of the dickson basis to compute the matrix coefficients of @xmath300 .
the code below computes the matrices @xmath301 , @xmath302 , given in input the coefficients @xmath303 of @xmath51 , @xmath304 , defined as in .
* function dickson@xmath305transform * + * input : * @xmath306 + * output : * @xmath307 + @xmath308 ; @xmath309 ; + for @xmath310 + @xmath311 ; @xmath312 ; + end + @xmath313 , @xmath314 ; + for @xmath315 + @xmath316 + @xmath317 ; + end + @xmath318 ; @xmath319 ; @xmath320 + @xmath321 ; @xmath322 ; + for @xmath323 + @xmath324 ; + end + @xmath325 ; @xmath326 ; + for @xmath327 + @xmath328 ; + @xmath329 $ ] ; + end the coefficients of @xmath245 are linear combinations of @xmath330 . as can be seen by the above algorithm , the coefficents of such combinations expressed in the dickson basis remain bounded , the upper bound being @xmath331 . an analogous result , with
upper bound @xmath332 , holds for @xmath333 .
this is in contrast with the exponential growth that would have been seen in the purely palindromic case if one had directly applied the dickson transformation without the use of the dickson basis .
the arithmetic cost is @xmath334 operations .
once the coefficients @xmath335 are determined , a linearization of @xmath29 of the form can be constructed .
the properties of this linearization are investigated in the next section in order to devise a fast and numerically robust method to evaluate the newton correction of @xmath32 defined by .
our aim in this section is to derive a fast , robust and stable method for computing the newton correction @xmath336 , where @xmath32 and @xmath29 are related by , given a structured linearization @xmath337 , with @xmath338 , of @xmath29 of the form , namely , @xmath339 and @xmath340 our approach relies upon the celebrated _ formula of jacobi_@xcite @xmath341 which reduces the evaluation of @xmath342 to computing the trace of @xmath343 . in the sequel
we describe a method for finding the block entries and , a fortiori , the trace of the inverse of @xmath38 from the lq factorization of the matrix
. then we slightly modify the computation to take into account the contribution due to the matrix @xmath288
. it will be clear from what follows that this method is general and can be applied , with only trivial modifications , to any kind of unstructured matrix polynomial , simply by considering for instance the standard companion linearization instead of .
we denote as @xmath344 the @xmath345 unitary givens rotation given by @xmath346 let @xmath347 be the ( block ) lq factorization of @xmath38 obtained by means of givens rotations so that @xmath348 it can be easily checked that the lower triangular factor @xmath349 has the following structure @xmath350 where @xmath351 , @xmath352 . if @xmath353 and , therefore , @xmath38 is invertible then the lq factorization can be used to find a condensed representation of the inverse of @xmath38 .
observe that @xmath354 . in order to take into account the occurrence of the matrix @xmath288 in the jacobi formula
let us introduce the matrix @xmath355 .
then we have the following [ inv ] there exist matrices @xmath356 such that @xmath357 where the blank entries are not specified .
_ the proof basically follows by applying the ( block ) schur decomposition of @xmath358 to the block lower triangular factor @xmath359 . to show it more formally we can proceed by induction .
let us assume that the the @xmath360th block row of @xmath361 can be represented as @xmath362,\ ] ] where @xmath363 is the diagonal entry and the value of the entries in the strictly lower triangular part denoted by @xmath213 is not essential .
then , by applying @xmath364 on the left of the matrix we find that the @xmath365 block row looks like @xmath366,\ ] ] whereas the diagonal entry in position @xmath367 becomes @xmath368 .
@xmath91 this result says that the block diagonal entries of @xmath369 can be determined from the entries in its first ( block ) row .
the computation of this row is equivalent to the solution of the linear system @xmath370 or , equivalently , @xmath371 in the view of the structure of @xmath358 this reduces to @xmath372 let @xmath373 be a block diagonal matrix defined by @xmath374 using the matrix @xmath375 to balance the coefficient matrix yields @xmath376 observe that @xmath377 and , therefore , the solution of @xmath378 gives the desired unknown matrices @xmath379 .
to achieve some computational savings we rewrite the system as @xmath380 and thus we arrive at the following relation @xmath381 which is used to compute the reciprocal of the newton correction .
the function @xmath382 below implements our resulting algorithm at the cost of @xmath383 operations . * function trace * + * input : * @xmath307 , @xmath384 , @xmath385 + * output : * the value of @xmath386 + @xmath387 ; @xmath388 ; + @xmath389 ; @xmath390 ; + @xmath391 ; @xmath392 ; @xmath393 ; + for @xmath310 + @xmath394 $ ] ; @xmath395 ; @xmath396 ; @xmath397 ; + @xmath398 ; @xmath399 ; + @xmath400 ; @xmath401 ; @xmath402 ; @xmath403 ; + @xmath404 ; @xmath405 ; @xmath406 ; + end + for @xmath407 + @xmath408 ; + end + for @xmath409 + @xmath410 ; + end + @xmath411 ; + for @xmath412 + @xmath413 ; @xmath414 ; + end + @xmath415 ; @xmath416 ; + for @xmath417 + @xmath418 ; + end + @xmath419 ; + for @xmath420 + @xmath421 ; + end + @xmath422 ; @xmath423 ; @xmath424 ;
a simple tool for the simultaneous approximation of all the eigenvalues of a polynomial is the ehrlich - aberth method .
bini and fiorentino @xcite showed that a careful implementation of the method yields an efficient and robust polynomial root finder .
the software package mpsolve documented in @xcite is designed to successfully compute approximations of polynomial zeros at any specified accuracy using a multi - precision arithmetic environment .
a root finder for t - palindromic eigenproblems can be based on the ehrlich - aberth method applied for the solution of the algebraic equation @xmath34 , where @xmath32 is related with @xmath29 by and @xmath29 is generated by the function * dickson@xmath305transform * applied to the input coefficients @xmath425 of the t - palindromic matrix polynomial @xmath51 of degree @xmath234 given as in .
the method simultaneously approximates all the zeros of the polynomial @xmath32 : given a vector @xmath426 , @xmath427 , of initial approximations to the zeros of @xmath32 , the ehrlich - aberth iteration generates a sequence @xmath428 , @xmath429 , which locally converges to the @xmath430tuple of the roots of @xmath32 , according to the equation @xmath431 the convergence is superlinear for simple roots and linear for multiple roots . in practice
, the ehrlich - aberth method exhibits quite good global convergence properties , even though no theoretical results are known in this regard .
the main requirements for an efficient implementation of the method are : 1 . a rule for choosing the initial approximations ; 2 . a fast , numerically robust and stable method to compute the newton correction @xmath432 ; 3 .
a reliable stopping criterion .
concerning the first issue it is commonly advocated@xcite that for scalar polynomials the convergence benefits from the choice of equally spaced points lying on some circles around the origin in the complex plane . in the case of matrix polynomials where the eigenvalues are often widely varying in magnitude this choice can not be optimal .
a better strategy using the initial guesses lying on certain ellipses around the origin in the complex plane is employed in our method .
the second task can be accomplished by means of the function * trace * in the previous section . with respect to the third issue ,
it is worth observing that the ql - based method pursued for the trace computation also provides an estimate on the backward error for the generalized eigenvalue problem . from a result in @xcite
it follows that if @xmath433 is not an eigenvalue of @xmath38 then @xmath434 gives an appropriate measure of the backward error for the approximate eigenvalue @xmath433 .
since for @xmath435 we have that @xmath436 in our implementation we consider the quantity @xmath437 as an error measure .
if @xmath438 is smaller than a fixed tolerance then @xmath433 is taken as an approximate eigenvalue and the corresponding iteration is stopped .
the resulting ehrlich - aberth algorithm for approximating finite eigenvalues of @xmath29 and hence obtaining the corresponding eigenvalues of @xmath51 is described below . in the next section we present results of numerical experiment assessing the robustness of the proposed approach
. * function palindromic@xmath305aberth@xmath305zeros * + * input : * @xmath306 , @xmath439 , @xmath440 , initial guesses @xmath441 + * output : * approximations @xmath442 , @xmath443 , of the zeros of @xmath444 + @xmath445=$]*dickson@xmath305transform*@xmath446 + @xmath443 ; @xmath447 ; + @xmath448 ; + for @xmath449 + for @xmath450 + if @xmath451 + @xmath452*trace*@xmath453 ; @xmath454 ; + @xmath455 ; + @xmath456 ; + @xmath457 ; @xmath458 ; + if @xmath459 + @xmath460 ; @xmath461 ; + end ; + end + end + if @xmath462 + break + end + end + for @xmath450 + @xmath463)$ ] ; + @xmath464 ; @xmath465 ; + end the total cost of the algorithm is therefore @xmath466 operations , where @xmath467 is the total number of times that the function @xmath382 is called .
numerical experiments presented in the next section show that @xmath467 heavily depends on the choice of the starting points . with a smart choice
, @xmath467 is of order @xmath468 , which gives a total computational cost of @xmath44 .
since the cost of our method grows as @xmath46 but is only quadratic in @xmath1 , where customary qz - like methods use @xmath469 operations , an ehrlich - aberth approach looks particularly suitable when the matrix polynomial has a high degree and small coefficients so that @xmath470 is large .
it is worth noticing that the case of large @xmath43 can still be treated by means of an ehrlich - aberth method in @xmath45 operations .
the basic observation is that the factor @xmath46 comes from the block structure of the linearization involved in the computation of the trace .
a reduction of the cost can therefore be achieved by a different strategy where the linearization is initially converted into ( scalar ) triangular - hessenberg form : say , @xmath471 where @xmath472 is ( scalar ) triangular and @xmath473 is ( scalar ) hessenberg .
the task can virtually be performed by any extension of the fast structured methods for the hessenberg reduction proposed in @xcite .
these methods preserve the rank structure which can therefore be exploited also in the triangular - hessenberg linearization . once the matrices @xmath472 and
@xmath473 have been determined then the computation of @xmath474 can be performed by the following algorithm which has a cost of @xmath475 operations : * perform a @xmath476 decomposition of the hessenberg matrix @xmath477 , obtaining a unitary matrix @xmath478 represented as product of @xmath479 givens transformations ( schur decomposition ) and a triangular matrix @xmath480 .
* compute the last row of @xmath481 by solving @xmath482 and then computing @xmath483 . *
recover the diagonal entries of @xmath481 from the entries of @xmath484 and the elements of the schur decomposition of @xmath478 .
this alternative road leads to an algorithm of total cost @xmath469 operations .
an efficient implementation exploiting the rank structures of the matrices involved will be presented elsewhere .
the function * palindromic@xmath305aberth@xmath305zeros * for computing the roots of a t - palindromic matrix polynomial @xmath485 , given its coefficients @xmath486 , @xmath304 , has been implemented in matlab and then used for the computation of the zeros of polynomials of both small and high degree . the tolerance is fixed at @xmath487 and for the maximum number of iterations we set @xmath488 .
extensive numerical experiments have been performed to illustrate some basic issues concerned with the efficiency and the accuracy of a practical implementation of our method .
an accurate and efficient root - finder is essential to the success of our algorithm . in practice ,
the cost of each iteration is strongly dependent on the amount of early convergence ( for the sake of brevity , in the following we will refer to this phenomenon using the word _ deflation _ ) occurring for a given problem . in other words , a critical point to assess
the efficiency of the novel method is the evaluation of the total number @xmath467 of calls of the function trace , and of its dependence on the total number @xmath489 of the eigenvalues . when the ehrlich - aberth method is used to approximate scalar polynomials roots ,
experiments show that @xmath467 depends on the choice of the starting points .
if there is not any a priori knowledge about the location of the roots , empirical evidence @xcite shows that choosing starting points distributed on some circles around the origin leads to acceptable performances and/or quite regular convergence patterns .
the class h@xmath490 of t - palindromic polynomials have been used to verify if these properties still hold in the matrix case .
the polynomials are constructed according to the following rules : @xmath491 from @xmath492 we find that most of the eigenvalues lie on the unit circle and for @xmath1 even @xmath210 is a double root of @xmath493 .
figure 1 describes the convergence history for our root finder applied to @xmath494 with starting values equally spaced on the circle centered in the origin with radius 4 .
the curves represented are generated by plotting the sequences @xmath495 , @xmath496 , for @xmath450 .
the convergence is quite regular and very similar to that exhibited in the scalar polynomial case @xcite and theoretically predicted for simultaneous iterations based on newton - like methods @xcite . with this choice of starting points
we have observed that the number of global iterations is typically of order of @xmath497 but there are not enough early deflations , that is , iterations that are prematurely stopped due to early convergence . in order to increase the cost savings due to premature deflation in our program we have employed a slightly refined strategy .
since the method does not approximate directly the eigenvalues @xmath498 but their dickson transform @xmath499 , we have chosen starting points on the dickson transform of the circles @xmath500 , that is points lying on ellipses @xmath501 .
more precisely , this is the algorithm we used to pick the starting points : * input : * number @xmath497 of eigenvalues to approximate and parameters @xmath502 and @xmath503 + * output : * starting points @xmath504 , @xmath505 + @xmath506 ; + @xmath507=randn ; + for @xmath508 + @xmath509=mod(@xmath510 ) ; + @xmath511 ; + @xmath512 ; + @xmath513 ; + @xmath514 + end + -0.6 cm the integer @xmath515 determines the number of ellipses whereas @xmath516 is used to tune the lenghts @xmath141 and @xmath142 , defined as above , of their semiaxes .
we expect that a good choice for the parameters @xmath515 and @xmath516 depends on the ratio @xmath517 : when @xmath518 we expect many eigenvalues to lie on or near to the unit circle , while when @xmath519 we expect a situation more similar to the eigenvalues of a random matrix , with no particular orientation towards unimodularity .
we therefore expect that a small ratio @xmath520 works well in the former case while on the contrary in the latter case @xmath521 should be a better choice .
moreover , we expect that as @xmath228 grows it is helpful to increase the total number @xmath515 of ellipses as well .
we show here some of the results on random t - palindromic polynomials .
figure 2 refers to an experiment on small - dimensional , high - degree polynomials : the value of @xmath43 has been set to @xmath522 while @xmath1 was variable .
the average number of @xmath467 over a set of @xmath523 random polynomials for each value of @xmath443 is shown on the graph .
the parameters satisfy @xmath524 and @xmath525 and they are determined by @xmath526 and @xmath527 , where the integer @xmath528 is defined as @xmath529 .
the graph shows a linear growth of @xmath467 with respect to @xmath443 .
figure 3 refers to an experiment where on the contrary the case of small @xmath1 is explored .
we have considered here @xmath530 and let @xmath43 vary and we show the results for @xmath467 plotted against @xmath228 for several choices of @xmath515 and @xmath516 .
the choice labelled as step function is for @xmath531 and @xmath532 generated by @xmath533 and @xmath534 . once again
the experiments suggest that when the starting points are conveniently chosen @xmath535 for some constant @xmath141 and any @xmath497 in the specified range , and , moreover , the bound still holds for different reasonable choices of the parameters @xmath515 and @xmath516 .
the experimentation with random polynomials gives @xmath536 as an estimate for the constant . in conclusion
, the algorithm can greatly benefit from a smart strategy for the selection of the starting points by increasing the number of early deflations .
the experiments show that as long as the starting points are suitably chosen the value of @xmath467 is proportional to @xmath443 . the other important aspect of our solver based on polynomial root - finding concerns the accuracy of computed approximations . in our experience
the method competes very well in accuracy with the customary qz - algorithm .
the accuracy of the computed non - exceptional roots for the random polynomials was always comparable with the accuracy of the approximation obtained with the @xmath537 method .
the results of other numerical experiments confirm the robustness of the novel method .
figure 4 illustrates the computed eigenvalues for the problem @xmath538 .
figure 5 also reports the plot of the absolute error vector @xmath539 and @xmath540 , where @xmath541 is the vector formed by the eigenvalues computed in high precision arithmetic by mathematica while @xmath542 and @xmath543 are , respectively , the vectors formed from the eigenvalues returned by our routine * palindromic@xmath305aberth@xmath305zeros * and suitably sorted by the internal function @xmath544 . the numerical results put in evidence the following important aspects : 1 .
poor approximations for the exact eigenvalue @xmath210 are in accordance with the theoretical predictions : in fact the reverse transformation from @xmath220 to @xmath545 is known to be ill - conditioned near @xmath546 ( or @xmath199 ) . since in this example @xmath206 is a defective eigenvalue , the approximations returned by @xmath544 have comparable absolute errors of order @xmath547 which are in accordance with the unstructured backward error estimates given in @xcite .
the accuracy of the remaining approximations is unaffected from the occurrence of near - to - critical eigenvalues and is in accordance with the results returned by @xmath544 . for most non - exceptional eigenvalues ,
the accuracy of approximations computed by our method is slightly better .
this kind of behavior is confirmed by many other experiments .
our method performs similarly to the qz for non - exceptional eigenvalues and for defective exceptional eigenvalues , but generally worse than qz and the structure - preserving methods @xcite for exceptional eigenvalues .
in this paper we have shown that the ehrlich - aberth method can be used for solving palindromic and t - palindromic generalized eigenproblems .
the basic idea can be applied to a generic matrix polynomial of any kind ; moreover , as we have shown in this paper , it is possible to adapt it in order to exploit certain structures as the palindromic structure that we have considered here .
the resulting algorithm is numerically robust and achieves computational efficiency by exploiting the rank - structure of the associated linearization in the dickson basis .
the algorithm is quite interesting for its potential for parallelization on distributed architectures and , moreover , can be easily incorporated in the mpsolve package to develop a multiprecision root finder for matrix polynomial eigenproblems . 1 .
the development of an automatic procedure for the selection of starting points is important to attain a low operation count due to the prevalence and ease of deflation .
we have shown that a smart choice could be based on a few parameters to be determined from some rough information on the spectrum localization .
the proposed algorithm is still inefficient with respect to the size of the polynomial coefficients .
the preliminary reduction of the linearized problem into a hessenberg - triangular form is the mean to devise a unified efficient algorithm for both small and large coefficients .
a fast reduction algorithm would be incorporated in our implementation .
the algorithm should be able to exploit the rank structure of the linearization ( for large degrees ) , and , at the same time , the inner structure of the quasiseparable generators ( for large coefficients ) .
3 . regarding the accuracy of the method
there are still some difficulties in the numerical treatment of the critical cases .
our current research is focusing on the issue of a structured refinement of the approximations of such eigenvalues .
* acknowledgements . *
the second author wishes to thank volker mehrmann for the many discussions and the valuable comments and suggestions provided during the gene golub siam summer school 2010 .
we are grateful to the anonymous referees for their hints and suggestions .
w. gander .
zeros of determinants of @xmath24-matrices . in vadim olshevsky and eugene
tyrtyshnikov , editors , _ matrix methods : theory , algorithms and applications , dedicated to the memory of g. golub_. world scientific publisher , 2010 . | an algorithm based on the ehrlich - aberth root - finding method is presented for the computation of the eigenvalues of a t - palindromic matrix polynomial .
a structured linearization of the polynomial represented in the dickson basis is introduced in order to exploit the symmetry of the roots by halving the total number of the required approximations .
the rank structure properties of the linearization allow the design of a fast and numerically robust implementation of the root - finding iteration .
numerical experiments that confirm the effectiveness and the robustness of the approach are provided .
_ ams classification : _
65f15 generalized eigenvalue problem , root - finding algorithm , rank - structured matrix |
+ local - susceptibility measurements via the nmr shifts of @xmath0p and @xmath1v nuclei in the high - pressure phase of ( vo)@xmath2p@xmath2o@xmath3 confirmed the existence of a unique alternating antiferromagnetic chain with a zero - field spin gap of 34 k. the @xmath0p nuclear spin - lattice relaxation rate scales with the uniform spin susceptibility below about 15 k which shows that the temperature dependence of both the static and dynamical spin susceptibilities becomes identical at temperatures not far below the spin - gap energy .
magnetic excitations of a low - dimensional quantum antiferromagnet have been one of the current topics among the condensed matter physicists .
vanadyl pyrophosphate ( vo)@xmath2p@xmath2o@xmath3 had long been believed as a prototype of a spin-@xmath4 two - leg ladder which has a magnetic lattice intermediate between one and two spatial dimensions @xcite . the ladder model , however , has been rejected by an observation of a dominant magnetic interaction perpendicular to the supposed ladder axis via the inelastic neutron scattering ( ins ) measurements @xcite .
a dimerized ( alternating ) chain model has now been becoming accepted as an alternative starting point , although a mechanism of the major exchange interaction between distant pairs of v@xmath5 spins via po@xmath6 tetrahedra is still under study @xcite .
the ins experiments has also revealed the existence of the mode with a gap nearly twice the gap of the lowest excited triplet which can not be explained by a simple alternating - chain model .
this mode has first been assigned as a bound state of two magnons possibly formed via interchain couplings @xcite , but it was difficult to account for the intensity comparable to the fundamental mode .
recent nmr @xcite and high - field magnetization @xcite studies have suggested on this issue that the two structurally - distinguishable chains of v atoms , which were thought to be magnetically identical , have different spin - gap energies .
this gives a natural explanation for the existence of two distinct modes with almost equal spectral weight , and has been supported by the subsequent raman - scattering experiments @xcite and theoretical studies on relevant exchange interactions @xcite .
the above confusion concerning the modelling and interpretation of the experimental results of ( vo)@xmath2p@xmath2o@xmath3 comes not only from the unexpectedly strong v - v exchange via po@xmath6 tetrahedra , but also from the presence of structurally - inequivalent v chains @xcite .
more recently , azuma have found that ( vo)@xmath2p@xmath2o@xmath3 transforms into another phase with different symmetry under pressure @xcite .
all the v atoms occupies a unique crystallographic site in the high - pressure ( hp ) phase , so that the magnetic chains made of v@xmath5 spins are all equivalent .
therefore , hp-(vo)@xmath2p@xmath2o@xmath3 will be a better example of the alternating antiferromagnetic chain with quantum spin @xmath4 . in this letter
, we report microscopic characterization of the magnetic chains in the hp phase of ( vo)@xmath2p@xmath2o@xmath3 via nmr . a single spin component characterized by a zero - field gap of 34
k was found , presenting support for the double - chain scenario for the ambient - pressure ( ap ) phase .
single crystals of the hp phase of ( vo)@xmath2p@xmath2o@xmath3 were grown as described in @xcite . since the crystals were too small to observe an nmr signal , they were crushed into powders and the nmr measurements were made on these powders .
standard spin - echo pulse techniques were utilized for most of the experiments .
an example of the field - swept @xmath0p nmr spectrum in the hp phase of ( vo)@xmath2p@xmath2o@xmath3 is shown in figure [ fig:31pspectrum ] .
the spectrum in the ap phase @xcite is also shown for comparison .
the spectrum in the hp phase consists of a single line as expected from the unique crystallographic site of phosphor in the unit cell .
this is contrasted with the ap phase where the spectrum splits into two groups of lines owing to the presence of two kinds of v chains with different gap energies @xcite .
the line - shape analysis revealed that the symmetry of an nmr - shift tensor at the p site is almost uniaxial . assuming the exact uniaxial symmetry , we determined the two independent principal values @xmath7 and @xmath8 corresponding to the shift with the external field parallel and perpendicular to the local symmetry axis , respectively .
the results are shown in figure [ fig:31kvst ] as a function of temperature .
both @xmath7 and @xmath8 scale the bulk magnetic susceptibility @xmath9 which is corrected by subtracting the contribution of paramagnetic impurities .
following the standard @xmath10@xmath9 analysis , the tensor components of the hyperfine coupling at the p site were determined as @xmath11 t/@xmath12 and @xmath13 t/@xmath12 .
these values yield the isotropic and uniaxial components , @xmath14 t/@xmath12 and @xmath15 t/@xmath12 , respectively .
@xmath16 is larger than and different in sign from that due to the classical dipolar field of v@xmath5 spins @xmath17 t/@xmath12 , indicating that the v@xmath5 spins are transferred not only to the p-@xmath18 orbitals but also to the p-@xmath19 orbital .
the susceptibility of a one - dimensional ( 1d ) gapped spin system at temperatures well below the gap @xmath20 is proportional to @xmath21 @xcite . in order to determine @xmath20
, we fitted the @xmath22 dependence of the isotropic component of the nmr shift @xmath23 below 10 k to the form @xmath24 , where the reduction of @xmath20 by fields is explicitly written .
the result is shown in the inset of figure [ fig:31kvst ] .
the obtained parameters are @xmath25 % , @xmath26 k@xmath27 , and @xmath28 = 31 k which gives @xmath29 = 34 k with the use of the measured @xmath30 factor @xcite .
@xmath31 is in good agreement with that evaluated from the bulk @xmath9 but is larger than the values determined from the critical field of the magnetization process ( @xmath3223 k ) @xcite and the ins on polycrystals ( @xmath3225 k ) @xcite for unknown reasons .
a free - induction - decay ( fid ) signal of @xmath1v has also been observed below about 50 k. the spectrum was obtained by integrating the fid signal while sweeping the external field .
the @xmath22 dependence of the @xmath1v nmr shift @xmath33 determined from the peak position of the spectrum is shown in figure [ fig:51kvst ] .
also shown in the inset is a plot of @xmath23 versus @xmath33 with @xmath22 the implicit parameter .
a linear relation found between @xmath23 and @xmath33 demonstrates that the @xmath22 dependence of the local spin susceptibility is identical for both the sites .
this is a clear sign of hp-(vo)@xmath2p@xmath2o@xmath3 having only one independent spin component .
the @xmath22 dependence of @xmath33 was analyzed in the same way as that of @xmath23 using @xmath20 determined above .
the @xmath22-independent orbital ( van - vleck ) shift was then obtained to be 0.182 % .
the hyperfine coupling constant at the v site determined from the slope of the @xmath34 plot is @xmath35 t/@xmath12 , which is in a reasonable range as a core - polarization field of a @xmath36 transition - metal ion @xcite .
figure [ fig:31wvst ] shows the @xmath22 dependence of the @xmath0p nuclear spin - lattice relaxation rate @xmath37 .
@xmath38 above 8 k was determined as the time constant of the exponential recovery of @xmath0p magnetization @xmath39 . below 8 k where non - exponential recovery appears ,
we analyzed @xmath39 by fitting to the form @xmath40 which incorporates the relaxation rate @xmath41 due to paramagnetic impurities @xcite .
as shown in the inset of figure [ fig:31wvst ] , @xmath37 exhibits activated behavior below about 20 k. the exponential decrease of @xmath37 is , however , masked below @xmath328 k synchronizing the appearance of non - exponential recovery .
the asymptotic value of @xmath37 at low @xmath22 is suppressed by applying fields as expected for the impurity - limited relaxation rate .
@xmath37 depends on @xmath42 at higher temperatures as well where the recovery is exponential , but the @xmath42 dependence roughly follows the 1d diffusive form @xmath43 as observed in ap-(vo)@xmath2p@xmath2o@xmath3 @xcite .
details of the @xmath42 dependence of @xmath37 will be presented in a separated paper .
the activation energy was estimated as @xmath44 k by fitting the data between 8 and 20 k to the form @xmath45 .
as the interbranch ( @xmath46 ) transitions within the lowest excited triplet @xcite are expected to dominate the nuclear - spin relaxation due to the predominantly - isotropic hyperfine fields , the obtained @xmath47 would give an estimate of the zero - field gap .
@xmath47 indeed agrees well with @xmath29 evaluated from the nmr shift .
figure [ fig : t1tkvst ] shows the @xmath22 dependence of @xmath48 divided by @xmath49 .
one of the remarkable features of the result is that the ratio @xmath50 becomes @xmath22 independent below about 15 k. ( an upturn below @xmath327 k is due to the impurity contribution to @xmath37 and is extrinsic . )
it is well known that , while the nmr shift is proportional to the uniform static susceptibility @xmath51 , @xmath37 samples the dissipative part of the dynamical susceptibility @xmath52 at the nuclear larmor frequency @xmath53 @xcite ; @xmath54 here @xmath55 is the fourier transform of the hyperfine coupling . since @xmath55 has a maximum at @xmath56 = 0 , @xmath37 at the p site is most sensitive to @xmath57 which is dominant at low @xmath22 in a gapped 1d spin system @xcite .
the @xmath22-independent behavior of @xmath50 therefore indicates that the @xmath22 dependence of @xmath58 and @xmath51 at low @xmath22 is identical and should be described by a common energy gap .
such a characteristic of the magnetic excitations in a gapped 1d spin system has been predicted theoretically based on a picture of free magnons @xcite , but has rarely been observed experimentally @xcite . to our knowledge , this is the first experimental verification of @xmath48 and @xmath10 having identical @xmath22 dependence at low @xmath22 , not relying on any model - dependent form of these quantities . from the experimental viewpoint , it is worth noting that the scaling between @xmath48 and @xmath10 holds below @xmath59 .
this suggests nearly free propagation of magnons being realized at temperatures not far below @xmath20 .
it is therefore practical to use experimental data in the region @xmath60 for a reliable estimate of @xmath20 , although the activated behavior of physical quantities such as @xmath9 and @xmath37 is theoretically justified only for @xmath61 @xcite . above about 20 k , the scaling breaks down and @xmath62 increases gradually with @xmath22
this means that @xmath63 grows more rapidly than @xmath51 .
as the temperature is now comparable with or higher than @xmath20 , interactions between magnons and/or the @xmath64 component of spin fluctuations will become increasingly important and would enhance @xmath65 over @xmath51 . in conclusion ,
we have measured @xmath0p and @xmath1v nmr in the high - pressure phase of ( vo)@xmath2p@xmath2o@xmath3 .
it was found that the temperature dependence of the local static spin susceptibility at the p site is identical with that at the v site .
the dynamical spin susceptibility @xmath65 near @xmath66 also scales with the static susceptibility at low temperatures below about a half of the spin - gap energy which was estimated to be 34 k at zero field .
all of these observations provides microscopic evidence for a unique kind of magnetic chain existing in the high - pressure phase of ( vo)@xmath2p@xmath2o@xmath3 , as well as for coexistence of magnetically - inequivalent chains in its ambient - pressure phase . |
|
this paper is devoted to the study of mechanical devices called _ hexapods _ , which are also known as _
stewart gough platforms_. as described in @xcite , the geometry of this kind of mechanical manipulators is defined by the coordinates of the @xmath0 platform points @xmath1 and of the @xmath0 base points @xmath2 in one of their possible configurations .
a hexapod is called _ planar _ if both base points and platform points are coplanar , otherwise _ non - planar_. all pairs of points @xmath3 are connected by a rigid body , called _ leg _ , so that for all possible configurations the distance @xmath4 is preserved ( see fig .
[ figure : hexapod ] ) .
we say that a hexapod is _ movable _ , or admits a _ self - motion _ if , once we fix the position of the base points @xmath5 , the platform points @xmath6 are allowed to move in an ( at least ) one - dimensional set of configurations respecting the constraints given by the legs . in this case , each @xmath7 moves on the sphere with center @xmath8 and radius @xmath9 .
hexapod ( -4,0)@xmath10 ( 50,-4)@xmath11 ( 100,2)@xmath12 ( 78,19)@xmath13 ( 34,24)@xmath14 ( 20,13)@xmath15 ( -6,73)@xmath16 ( 50,56)@xmath17 ( 70,69.5)@xmath18 ( 98,80)@xmath19 ( 45,85)@xmath20 ( 12,86)@xmath21 one can associate to each movable hexapod an algebraic curve ( if the hexapod has mobility one ; otherwise it is an algebraic variety of higher dimension ) contained in the algebraic group @xmath22 of direct isometries of @xmath23 ; it parametrizes the set of possible positions of the platform and because of this we call it the _ configuration curve _ ( or the _ configuration set _ in case the mobility is larger than one ) .
then one may define various invariants of a hexapod , such as the degree or the genus of its configuration curve .
clearly the degree depends on the embedding of the algebraic group @xmath22 in projective space .
we mainly use the so - called conformal embedding @xmath24 described in @xcite because it is most practical for the study of hexapods ; in this way we define the _ conformal degree _ of a hexapod . in the literature ( see for example @xcite ) it is more common to project the configuration curve to @xmath25 and use the well - known embedding @xmath26 determined by quaternions . in this case
we call the degree of the configuration curve the _ euler degree _ , since the coordinates of @xmath27 are called _
euler parameters_. the classification of hexapods with mobility one is still an open problem .
there is a family of planar hexapods , discovered by duporq ( see @xcite ) , with a configuration curve of conformal degree @xmath28 , euler degree @xmath29 , and genus @xmath30 ( see section [ degree ] ) ; these are the largest possible degrees and genus .
the family we introduce in this paper is non - planar and has conformal degree @xmath31 and euler degree @xmath32 .
we show that this is the maximal degree among non - planar hexapods if we exclude some degenerate cases . by the way , in this case
the genus of the configuration curve is @xmath33 , but we do not know whether this is maximal or not .
our linkages are constructed in the following way .
we start with @xmath0 points @xmath34 in @xmath23 , forming the base .
then we employ basic facts from liaison theory a method coming from algebraic geometry to construct a system of equations for the @xmath0 points @xmath35 forming the platform .
we conjecture that this system admits solutions for a general choice of @xmath36 .
once @xmath37 is fixed ( up to scaling ) , we derive linear equations for a scalar @xmath38 and a vector @xmath39 implying that the hexapod @xmath40 has mobility one , with a configuration curve of degree @xmath31 .
we conjecture that this system of equations has a unique solution for @xmath38 and a three - dimensional set of solutions for the vector @xmath41 .
we call a hexapod constructed by this procedure a _ liaison hexapod_. the nature of the two conjectures is so that if they are false , then they are falsified by a general choice of base points .
we tested the conjectures against many random choices , so there is quite a strong experimental evidence in their favor .
in addition , we found two particular subfamilies ( see proposition [ proposition : family_lines ] and proposition [ proposition : family_order3 ] ) for which we can prove that the properties predicted by the conjectures hold .
the paper is organized as follows . in section
[ degree ] , we prove our main theorem ( theorem [ thm : bound_degree ] ) concerning the degree bound for configuration curves of general non - planar hexapods of mobility one ; this is achieved by bounding the degree by the number of intersections of two algebraic curves , and then by using some facts from algebraic geometry to control such number .
the remaining sections deal with the construction of examples . in section [ liaison_hexapods ] , we describe liaison hexapods : given a general @xmath0-tuple @xmath36 , we construct up to a scaling factor a candidate platform @xmath37 applying liaison theory and mbius photogrammetry , a technique used by the authors in @xcite to establish necessary conditions for the mobility of pentapods . after
that we determine the right scaling factor for the platform points and the leg lengths so that the hexapod we obtain is movable ; both these tasks are achieved by inspection of tangency conditions for the configuration curve of the hexapod . in section
[ study ] , we prove with the aid of symbolic computation that the properties predicated by our conjectures hold in the case of two particular subfamilies of hexapods , and we exhibit an example of a liaison hexapod .
the goal of this section is to associate to each hexapod @xmath42 a projective curve @xmath43 , called the _ configuration curve _ of @xmath42 .
its degree , called the _ conformal degree _ of @xmath42 , is an invariant of the hexapod , and in this paper we are interested in understanding what are its possible maximal values .
in particular we report well - known examples of hexapods attaining high values for the conformal degree , and eventually we prove a bound for the degree of a hexapod satisfying some non - degeneracy conditions ( theorem [ thm : bound_degree ] ) .
we recall some definitions from @xcite that we need in our discussion .
following many other authors , we describe the set of admissible configurations of a given hexapod @xmath42 by prescribing the set of direct isometries ( from @xmath23 to itself ) mapping the initial position of the platform to an admissible one . by this
we mean that we consider the base @xmath36 as fixed , and if @xmath37 is the initial position of the platform points , we look for all @xmath44 ( where @xmath22 denotes the group of direct isometries ) such that @xmath45 where @xmath9 are the given leg lengths of the hexapod @xmath42 . the constraints imposed on isometries @xmath44 by eq . are called _
spherical conditions_. as shown in ( * ? ? ?
* section 2 , subsection 2.1 ) , it is possible to construct a projective compactification in @xmath46 for ( the complexification of ) the group @xmath22 ; we denote it by @xmath47 .
it turns out that @xmath47 is a projective variety of dimension @xmath0 and degree @xmath28 .
we call the map @xmath48 the _ _ conformal embedding _ _ of @xmath22 .
if we write the spherical condition in the coordinates of @xmath46 , we obtain six linear conditions , namely we determine a linear subspace @xmath49 of codimension @xmath0 . the intersection @xmath50 is defined to be the _ complex configuration set _ of the hexapod @xmath42 .
let @xmath42 be a hexapod .
the _ mobility _ of @xmath42 is defined to be the dimension of @xmath51 , where @xmath52 is the complexification of @xmath22 embedded in @xmath47 . for a hexapod of mobility one
we define the _ conformal degree _ to be the degree of the projective curve @xmath43 . from the fact that @xmath47 is a variety of degree
@xmath28 and @xmath43 is always a linear section of @xmath47 we immediately see that the conformal degree is always smaller than or equal to @xmath28
. one can also attach another degree to a hexapod , by considering only the rotational part of the isometries determining its configuration .
this is classically done by embedding the ( complexification of the ) group of rotations @xmath25 as an open subset of @xmath27 using the quaternionic description of rotations . in this way we get for a mobility one hexapod
@xmath42 a curve in @xmath27 , and its degree is called the _ euler degree _ of @xmath42 .
note that we obtain a point in @xmath27 if the corresponding self - motion is a pure translation . in this case
we set the euler degree to zero .
if @xmath53 is the map sending a direct isometry to its rotational part , by analyzing the definition of the conformal embedding one sees that there exists a linear projection @xmath54 such that the following diagram is commutative : @xmath55 \ar[d]_{\rho } & & x \ar@{}[r]|{\subseteq } \ar@{-->}[d ] & { \mathbb{p}}^{16}_{{\mathbb{c } } } \ar@{-->}[d ] \\ { \mathrm{so}_3}\ar[r ] & { \mathbb{p}}^3_{{\mathbb{c } } } \ar[r]^-{v_{3,2 } } & v_{3,2 } \ar@{}[r]|{\subseteq } & { \mathbb{p}}^9_{{\mathbb{c } } } } \end{array}\ ] ] where @xmath56 is the veronese embedding of @xmath57 by quadrics and @xmath58 is its image in @xmath59 .
if we have a closer look at @xmath47 , we discover that it can be written as the disjoint union @xmath60 where @xmath61 is a special hyperplane section of @xmath47 , called the _
boundary_. the boundary can be , in turn , decomposed into @xmath62 subsets : vertex : : the only real point in @xmath61 , a singular point with multiplicity @xmath29 ; it is never contained in the complex configuration set of a hexapod ; collinearity points : : if @xmath43 contains such a point , then either the platform points or the base points are collinear ; similarity points : : if @xmath43 contains such a point , then there are normal projections of platform and base to a plane such that the images are similar ; inversion points : : if @xmath43 contains such a point , then there are normal projections of platform and base to a plane such that the images are related by an inversion ; butterfly points : : if @xmath43 contains such a point , then there are two lines , one in the base and one in the platform , such that any leg has either a base point in the base line or a platform point in the platform line . [ remark : center_projection ] the center of the projection @xmath63 sending @xmath47 to the veronese variety @xmath58 is the linear space spanned by similarity points , which contains also the collinearity points and the vertex .
inversion points form an open subset of @xmath61 , while all other sets of points form proper ( quasi - projective ) subvarieties of @xmath61 .
the points of the configuration set @xmath43 of a hexapod @xmath42 that lie on the boundary @xmath61 are called the _ bonds _ of @xmath42 . as we see from the previous description , although bonds do not represent configurations , their presence constrains the geometry of the corresponding hexapod .
we now discuss a few well - known special cases of hexapods attaining high values for the conformal and the euler degree , mainly in order to exclude them later , when stating and proving our main theorem ( theorem [ thm : bound_degree ] ) . for a general planar pentapod ,
it is possible to add an additional leg without changing the configuration set ( see @xcite ) , hence obtaining a movable hexapod .
its configuration curve is the intersection of @xmath47 with a linear subspace of codimension @xmath62 , hence it has conformal degree @xmath28 , and its euler degree is @xmath29 .
the genus can easily be computed by the hilbert series of @xmath47 , and it turns out to be @xmath30 is half - canonical , namely embedded by a linear system @xmath64 such that @xmath65 is canonical . ] . a hexapod is called _ equiform _ if there is a similarity @xmath66 mapping base points to platform points .
non - planar equiform hexapods are discussed in @xcite .
equiform hexapods of mobility one admit configuration curves of degree @xmath67 .
the assumption on the existence of the similarity implies that for every direction in @xmath23 , the projections of both base and platform points along that direction are similar , hence every direction in @xmath23 determines a similarity point in @xmath43 .
therefore the intersection of @xmath43 with the boundary @xmath61 is a curve , and one can show that it is actually a conic .
thus the configuration curve of such a hexapod has two components , one given by the zariski closure of @xmath68 , and the other by the intersection of @xmath43 with @xmath61 .
if we assume that @xmath42 is neither planar nor equiform , then @xmath69 intersects @xmath61 in at most a finite number of points ( see @xcite ) . for this case , we show a useful proposition relating the conformal degree and the euler degree . [ proposition : euler_conformal ] let @xmath42 be a mobility one hexapod , neither planar nor equiform , then its euler degree is at most half its conformal degree .
consider the linear projection @xmath63 from diagram .
then the configuration curve @xmath43 is mapped by such projection to a curve @xmath70 in @xmath58 such that the preimage of @xmath70 under the veronese embedding @xmath56 determines the euler degree of @xmath42 .
let @xmath71 be the number of intersection points , counted with multiplicity , of @xmath69 with the center of the projection . since by remark [ remark : center_projection ] this is the cardinality of similarity bonds plus the cardinality of collinearity bonds of @xmath42 , the number @xmath71 is finite
. by a well - known formula for the degree of a projection , the degree of the image of @xmath69 is a factor of @xmath72 , where @xmath73 is the conformal degree , with equality if and only if the restriction of the projection to @xmath69 is birational to the image . on the other hand ,
the veronese map just doubles the degree , hence the statement follows .
we aim at bounding the conformal degree of a hexapod @xmath42 that satisfies some genericity assumptions ( for example , we will assume @xmath42 to be non - planar and non - equiform ) , and we will determine it by means of two estimates : first of all we connect the conformal degree to the intersection number of two algebraic curves defined by base and platform points only , and then the latter will be bounded using some arguments from algebraic geometry .
to do so we need to introduce a new tool , called _ mbius photogrammetry _ , and this is the content of the next subsection .
notice that a necessary condition for the mobility of a hexapod is the existence of bonds of some type : in fact if a hexapod @xmath42 is movable , then its configuration set @xmath43 intersects the boundary @xmath61 non - trivially . in particular , when the mobility is one , the bigger the degree of @xmath43 , the bigger the number of bonds counted with multiplicity .
if we assume that @xmath42 is neither planar nor equiform nor does it satisfy a butterfly condition , but has mobility one , then by the characterization of boundary points we get that there exist planar projections of base and platform such that their images are related by either a similarity or an inversion . in the second case
, we can replace the inversion also by the composition of the inversion with a reflection , by considering the projection from the opposite direction .
then , in both cases , we have projections of base and platform points that are mbius equivalent , i.e. equivalent up to mbius transformations .
hence , if we are interested in hexapods with high conformal degree , then we should look for hexapods with a high number of bonds . since the presence of inversion or similarity bonds corresponds to mbius equivalent orthogonal projections of base and platform , we are going to study such projections . in @xcite
the authors introduced the concept of _ mbius photogrammetry _ for @xmath62-tuples of points in @xmath23 , and this technique allowed to deduce some results about pentapods with mobility two .
the idea behind it is to take `` pictures '' of a configuration of @xmath62 points in @xmath23 , namely consider its orthogonal projection onto @xmath74 along all possible directions , and try to deduce properties of this configuration in @xmath23 from the knowledge of its projections .
we briefly recall the construction of the photographic map , adapting it to our setting , namely considering @xmath0-tuples of distinct points for definitions and proofs we refer to @xcite .
we start with a vector @xmath75 of @xmath0 distinct points in @xmath23 . for each direction @xmath76 in @xmath23 ,
where @xmath77 denotes the unit sphere , we consider an orthogonal projection @xmath78 along @xmath79 .
hence for each @xmath80 we have a tuple @xmath81 of projected points of @xmath82 , namely we have a @xmath0-tuple of elements of @xmath74 . by identifying @xmath74 with @xmath83 and embedding the latter in @xmath84
, we can think of @xmath85 as of an element of @xmath86 . since mbius transformations form the automorphism group of @xmath84 , the following two concepts are equivalent : * a @xmath0-tuple @xmath85 in @xmath87 , considered up to mbius transformations ; * a @xmath0-tuple @xmath85 in @xmath86 , considered up to automorphism of @xmath84 .
let @xmath88 denote the moduli space of @xmath0 points in @xmath84 , and denote by @xmath89 the corresponding quotient map , namely the function sending a configuration of @xmath0 points to its class modulo automorphisms of @xmath84 .
the previous notation encodes the fact that @xmath90 is not defined on the whole @xmath91 , since we know from geometric invariant theory that we need to restrict to the open set @xmath92 of @xmath93 of @xmath0-tuples where no four points coincide in order to build a well - behaved theory .
the previous equivalence allows us to think of a @xmath0-tuple @xmath85 of points in @xmath74 , considered up to mbius transformations , as a point in @xmath88 , in particular as the image of @xmath85 under the quotient map @xmath90 . in order to give an explicit formulation of the quotient map @xmath90
we follow , as in @xcite , the graphical approach provided in @xcite : * consider a convex hexagon in the plane and construct all plane undirected multigraphs without loops whose set of vertices coincides with the set of vertices of the hexagon and that satisfy the following conditions : * * edges are given by segments ; * * any two edges do not intersect ; * * the valency of every vertex is @xmath94 .
+ there are exactly @xmath62 of these graphs , as shown in fig .
[ figure : graphs ] .
+ ccc & & + [ 4 mm ] & & + * associate to each graph @xmath95 with set of edges @xmath96 a homogeneous polynomial in the coordinates @xmath97 of @xmath91 in the following way : @xmath98 for example the polynomial associated to the first graph in fig .
[ figure : graphs ] is @xmath99 * these five polynomials determine a rational map @xmath100 . *
consider the open set @xmath101 * the image of @xmath102 , namely the restriction of @xmath90 to @xmath92 , gives an embedding of @xmath88 in @xmath103 .
summing up , once we fix a @xmath0-tuple @xmath82 of distinct points in @xmath23 , we can associate to each direction @xmath76 the image ( under the quotient map @xmath90 ) of the projected @xmath0-tuple @xmath85 . in this way
we construct a map from @xmath77 to @xmath88 . in order to make it a morphism between projective varieties
, we identify @xmath77 with the conic @xmath104 in @xmath105 .
we obtain a map @xmath106 .
if we write @xmath107 and we set for all @xmath108 with @xmath109 @xmath110 then the components of @xmath111 have the following structure : @xmath112 from this explicit description we see that in particular this map is algebraic . [ definition : photographic_map ] the regular map @xmath113 we obtain is called the _ photographic map _ of @xmath82 .
[ definition : moebius_curve ] let @xmath82 be a @xmath0-tuple of distinct points in @xmath23 from now on we will omit the adjective `` distinct '' , but throughout the paper we will always consider this situation .
the image of the photographic map @xmath111 is a rational curve in @xmath88 , that we call the _ mbius curve _ of @xmath82 .
as noticed in @xcite , the map @xmath111 is a morphism of real varieties , namely it respects the real structures of @xmath114 and @xmath88 induced by the standard ones of @xmath115 and @xmath103 , respectively .
one can prove that @xmath88 is embedded in @xmath103 as a cubic threefold , called the _
segre cubic primal _ , whose equation is @xmath116 the threefold @xmath88 contains exactly @xmath117 planes , corresponding to equivalence classes of configurations @xmath118 where @xmath119 for some @xmath120 .
furthermore , the maximum possible number of nodes for a cubic hypersurface in @xmath103 , namely @xmath121 , is attained by the segre cubic .
the nodes correspond to configurations of points @xmath118 where three points coincide , and are the only singular points of @xmath88 .
there are @xmath122 such configurations , but the following phenomenon happens in @xmath88 : if we partition @xmath123 , then configurations of points for which @xmath124 are sent by the quotient map @xmath125 to the same point of @xmath88 as configurations of points for which @xmath126 .
the interested reader can find proofs and explanations for the stated facts in ( * ? ? ?
* chapter 9 , subsection 9.4.4 ) . by construction ,
whenever a hexapod @xmath42 admits an inversion or similarity bond , then the mbius curves of its base and platform intersect .
however , due to the construction of the moduli space @xmath88 , this is not the only situation when we get intersections between the two curves : suppose in fact that @xmath42 admits three collinear base points , say @xmath127 and @xmath12 , and three collinear platform points , say @xmath128 and @xmath21 , so that we have a butterfly bond ; then the mbius curves intersect in a node of @xmath88 , since the projections along the lines @xmath129 and @xmath130 determine configurations in @xmath74 where three points coincide .
the following two lemmata , in analogy with lemma 3.7 and 3.8 in @xcite , describe the possible behavior of the photographic map .
the proof technique is similar to the one in @xcite , but we report the proofs for self - containedness purposes .
[ lemma:6_space ] let @xmath131 be a @xmath0-tuple of points in @xmath23 . if the @xmath132 are not coplanar , then the photographic map @xmath133 is birational to a rational curve of degree @xmath0 or @xmath134 in @xmath88 .
we notice that if the lines @xmath135 and @xmath136 are parallel , then the factors @xmath137 and @xmath138 in eq .
differ only by a scalar .
this means that in particular situations , when some of the @xmath139 are collinear , the components of @xmath111 can have factors in common that can be simplified , reducing the degree of the map .
hence we partition the set of possible configurations in two cases , depending on the number of common factors of the components of @xmath111 .
notice that , since we assume that the points are not coplanar , it can never happen that @xmath62 or all @xmath0 points are collinear .
case ( a ) : : here no @xmath134 points are collinear . in this case
the components of @xmath111 do not have any factor in common , so @xmath140 thus there are only four possibilities : either @xmath111 is @xmath141 to a line , or it is @xmath142 to a conic , or it is @xmath143 to a cubic curve , or @xmath144 to a sextic curve .
we prove that the first three situations can never happen .
first of all , notice that there are exactly two directions in @xmath23 for which the images of @xmath139 and @xmath145 under the projection coincide , namely the directions of the oriented lines @xmath146 and @xmath147 . denote by @xmath148 the plane in @xmath88 of classes of @xmath0-tuples @xmath149 where @xmath119 ( see the fact in subsection [ degree : photogrammetry : construction ] )
. then there are exactly @xmath150 points in @xmath114 that are mapped to the plane @xmath148 and they are complex conjugate , since complex conjugation in @xmath114 corresponds to the antipodal map in the unit sphere @xmath151
. + if @xmath111 is @xmath142 or @xmath141 , then those two points have to be branching points of @xmath111 .
setting @xmath107 one can check that @xmath152 where we recall that @xmath153 .
if the points in the preimage of @xmath148 are branching points , then the line @xmath154 should intersect @xmath114 tangentially at those points .
however , this is impossible , since both @xmath114 and the line are real varieties , so if they are tangent their intersection point is real , but @xmath114 has no real points . in this way we rule out the @xmath142 and the @xmath141 case .
+ suppose now that @xmath111 is @xmath143 .
we are going to show that @xmath82 should be planar , contradicting the hypothesis .
in fact , by assumption @xmath111 factors through an involution of @xmath114 the one swapping the fibers of @xmath111 ; one can show that such involution is a real automorphism of @xmath114 , and hence corresponds to a rotation by @xmath155 degrees of the unit sphere @xmath77 along some axis @xmath156 .
the fibers @xmath157 are invariant under the involution , so they are contained in the subset of @xmath77 invariant under the rotation , namely the intersection @xmath158 together with the maximal circle in @xmath77 lying on a plane orthogonal to @xmath156 .
a case - by - case analysis proves that all direction @xmath146 belong to such maximal circle , this implying that @xmath82 is planar .
hence the birational case is the only possible .
case ( b ) : : here exactly @xmath134 points are collinear . in this case
the components of @xmath111 have one factor in common , leading to @xmath159 we have three possibilities : either @xmath111 is @xmath160 to a line , or it is @xmath143 to a conic , or @xmath144 to a quartic curve . arguing as in case ( a ) we prove the statement . [ lemma:6_plane ]
let @xmath131 be a @xmath0-tuple of points in @xmath23 .
if the @xmath132 are coplanar , but not collinear , then the photographic map @xmath113 is @xmath143 to a rational curve of degree @xmath161 , @xmath150 or @xmath94 in @xmath88 .
since @xmath82 is planar , then all directions @xmath146 belong to a maximal circle of the unit sphere @xmath77 .
let @xmath156 be the line through the origin orthogonal to the plane spanned by such maximal circle . reversing the argument in the proof of lemma [ lemma:6_space ] ,
case ( a ) one can prove that the map @xmath111 factors through the @xmath143 involution @xmath162 of @xmath114 determined by the rotation of @xmath77 by @xmath155 degrees around @xmath156 .
then we have @xmath163 . from now on
the proof works as in lemma [ lemma:6_space ] , and the following cases arise : case ( a ) : : suppose that no @xmath134 points of @xmath82 are collinear .
then the components of @xmath111 do not have any factor in common , and @xmath164 is birational to a cubic .
case ( b ) : : suppose that @xmath134 points of @xmath82 are collinear , but no @xmath62 points are so .
then the components of @xmath111 have exactly one factor in common , and @xmath164 is birational to a conic .
case ( c ) : : suppose that @xmath62 points of @xmath82 are collinear
. then the components of @xmath111 have exactly two factors in common , and @xmath164 is birational to a line .
we conclude this section by showing that the photographic map of some @xmath0-tuples of points extends to a morphism defined on the whole plane @xmath115 . from this
we deduce geometric constraints for mbius curves .
we hope that this can be the first step towards a complete geometric characterization of mbius curves as curves in @xmath88 .
we focus on the case when mbius curves are of degree @xmath0 , which is the most interesting for our application .
[ lemma : moebius_extend ] let @xmath82 be a @xmath0-tuple of points in @xmath23 and let @xmath165 be its photographic map , where the resulting mbius curve is of degree @xmath0
. then @xmath111 extends to a morphism @xmath166 .
we only need to prove that the map @xmath167 does not have base points in @xmath115 .
recall the structure of the map @xmath111 described by eq .
: we infer that a base point has to vanish on at least one polynomial @xmath137 for each component of @xmath167 .
a direct inspection shows that this would imply that at least @xmath134 points of @xmath82 are collinear , but this is impossible because by lemma [ lemma:6_space ] the curve @xmath168 would not have degree @xmath0 , contradicting the hypothesis . [ proposition : characterization_moebius ] let @xmath168 be a smooth mbius curve of degree @xmath0 .
then * @xmath168 can be defined by real polynomials , but has no real points ; * @xmath168 is contained in a linear projection , defined by real polynomials , of the third veronese embedding of @xmath115 . by construction @xmath168
is a real variety , since @xmath111 is a real map and @xmath114 is a real variety ; since @xmath168 is smooth and using lemma [ lemma:6_space ] we have that @xmath111 is an isomorphism , hence @xmath168 has no real points , because this holds for @xmath114 .
eventually from lemma [ lemma : moebius_extend ] we get that @xmath168 is contained in a linear projection , defined by real polynomials , of the third veronese embedding of @xmath115 , which is real by construction .
one notices that such projections is the complete intersection of @xmath88 with another cubic hypersurface .
we have now all the tools needed to prove the bound on the conformal degree of a hexapod ( theorem [ thm : bound_degree ] ) .
we split the argument in two parts : first we prove that the conformal degree of a hexapod @xmath42 is less than or equal to twice the number of intersections of the mbius curves of the base and platform of @xmath42 ( theorem [ theorem : bound_degree ] ) ; then we show that such intersection can not be composed by more than @xmath32 points ( proposition [ proposition : bound_intersection_moebius ] ) . for our proofs to work we need to exclude some degenerate cases planarity , equiformity and a certain kind of collinearity .
the reader should be warned about the length of these two proofs .
[ theorem : bound_degree ] let @xmath42 be a mobility one hexapod .
let @xmath169 and @xmath170 be the mbius curves of its base and platform .
suppose that @xmath42 is non - planar , non - equiform and no @xmath134 base or platform points are on a line .
then @xmath171 where the intersection @xmath172 is meant scheme - theoretically , namely the points of intersection are counted with multiplicity .
informally , the proof goes as follows : the degree of @xmath43 is computed by intersecting it with a hyperplane , and we choose the hyperplane defining the boundary @xmath61 of @xmath47 .
this intersection , as we will see , should be counted twice , and this means that the degree of @xmath43 is bounded by twice the number of bonds of @xmath42 .
eventually , bonds correspond to common images of the photographic maps , so the statement follows . in order to make the argument precise ,
we need to rephrase the analysis of bonds carried out in @xcite in a scheme - theoretical way .
we start by noticing that the hyperplane @xmath173 defining the boundary @xmath61 is totally tangential to @xmath47 , namely the scheme - theoretical intersection @xmath174 has a non - reduced structure , and we have @xmath175 ; moreover , as divisors on @xmath47 , it holds @xmath176 .
all these results can be proved via a direct computation using , for example , grbner bases .
therefore , if @xmath177 denotes the top - dimensional part of @xmath43 namely @xmath177 is an equidimensional scheme of dimension one we have that : @xmath178 the last inequality follows from the general fact that if @xmath179 and @xmath180 are two divisors of @xmath47 satisfying @xmath181 , then for any subvariety @xmath182 of @xmath47
we have @xmath183 .
the scheme @xmath184 is the scheme of bonds of @xmath42 , and we are going to prove that the number @xmath185 is less than or equal to the degree of the intersection @xmath172 of the two mbius curves of @xmath42 . to reach this goal ,
we need to connect the boundary @xmath61 of @xmath47 to the curve @xmath186 used in definition [ definition : photographic_map ] by means of a rational map @xmath187 .
as we already reported in subsection [ degree : configuration ] , to every inversion , similarity and butterfly point we can associate two directions in @xmath23 , and since we use the curve @xmath114 as a model for the unit sphere @xmath151 , this means that to every such point we can associate two elements @xmath188 .
notice that our assumption on the hexapod @xmath42 rules out the existence of collinearity bonds , and so it is harmless to exclude them from the picture . using the notation introduced in ( * ? ? ?
* sections 2.1 and 2.3 ) , we can write @xmath189 if we define @xmath190 then we get @xmath191 if @xmath192 , then there exist @xmath188 and @xmath193 such that @xmath194 and from the definition of @xmath195 we get that either @xmath196 or @xmath197 .
we define the algebraic map @xmath198 notice that the fiber of @xmath199 over a point @xmath200 is parametrized by @xmath201 this allows to conclude that @xmath195 forms an open subvariety of a @xmath27-bundle over @xmath202 . as remarked before ,
our hypotheses imply that the scheme of bonds @xmath203 is a closed subscheme of @xmath195
. * claim . * _ the map @xmath204 is an isomorphism . _ in order to prove the claim , we need to rephrase the so - called _ pseudo spherical condition _ , introduced in ( * ? ? ? * definition 3.6 ) .
the pseudo spherical condition imposed by a pair @xmath205 on a bond @xmath206 reads as @xmath207 using eq . , and setting @xmath208 and @xmath209 yields @xmath210 the scheme @xmath211 is cut out by these @xmath0 pseudo spherical conditions for @xmath212 .
hence , if we define the @xmath213 matrix @xmath214 then the scheme @xmath211 is locally defined by @xmath215 where @xmath216 and @xmath217 are suitable open subvarieties of @xmath114 and @xmath27 respectively , and @xmath218 .
we restate our previous claim .
* claim . * _ the map @xmath199 maps @xmath211 isomorphically to the scheme in @xmath202 cut out by the @xmath219 minors of @xmath220 .
_ we prove that the rank of the matrix @xmath221 is always at least @xmath161 .
in fact , if the rank is @xmath94 then the collinearity hypothesis is violated ( all @xmath222 would be equal , and the same for the @xmath223 ) ; if the rank is @xmath150 , then the planarity condition is violated ( this can be deduced by a direct computation , imposing that the second column of @xmath220 is a linear combination of the third and the fourth ) .
hence the rank of @xmath221 is greater than of equal to @xmath161 .
notice that column operations on @xmath220 correspond to projective transformations in the fibers of @xmath199 , while row operations on @xmath220 correspond to the choice of a different system of generators for the ideal of @xmath211 .
hence we can reduce to the case when @xmath220 has the form @xmath224 where @xmath225 are rational functions on @xmath202 .
the local description of @xmath211 becomes @xmath226 while the zero locus of the @xmath219 minors of @xmath220 is locally given by @xmath227 and one sees that these two schemes are isomorphic .
the proof is then complete once we are able to show the following
. * claim . * _ the image of @xmath211 under @xmath199 is contained in the pullback of the two photographic maps @xmath228 . _
notice that the fact that @xmath42 is supposed to be non - equiform implies that the mbius curves of @xmath37 and @xmath36 are different .
the pullback of the two maps is @xmath229 the coordinates of @xmath230 are obtained by substituting each term @xmath137 by @xmath231 in the general formula from eq . ; for @xmath232 one just needs to consider @xmath233 instead .
then the pullback is the scheme cut out by the @xmath234 minors of the following @xmath235 matrix : @xmath236 where @xmath237 and similarly for @xmath238 .
a direct computation ( for example , with the aid of grbner bases ) shows that the ideal generated by such @xmath234 minors is contained in the ideal of the @xmath219 minors of @xmath220 .
this settles the claim and hence concludes the proof .
[ remark : pullback ] we can not hope for equality in the last claim in the proof of theorem [ theorem : bound_degree ] .
in fact , consider one of the @xmath117 planes in @xmath88 ( see the fact in subsection [ degree : photogrammetry : construction ] ) , and suppose it parametrizes classes of tuples @xmath118 where @xmath119 ; then the projection from such a plane maps @xmath88 to @xmath84 , and the latter has a modular interpretation as @xmath239 , namely the moduli space of @xmath134-tuples @xmath240 obtained by removing @xmath241 and @xmath242 .
then the images of @xmath230 and @xmath232 under the projection from the plane of classes with @xmath243 coincide if and only if @xmath244 where @xmath245 . on the other hand , the left - hand - side of eq .
is a @xmath219 minor of @xmath220 .
in fact , if we select the submatrix with rows of index @xmath246 and @xmath247 and we perform column operations we obtain @xmath248 and a direct computation proves the equality . as it is shown in the second claim of the proof , the zero locus of these minors is image of @xmath211 under @xmath199 . on the other hand , there are points in the pullback of @xmath249 and @xmath250 for which not all the cross - ratios of eq .
are equal : suppose in fact the points @xmath251 and @xmath18 are collinear along the direction @xmath252 , and the points @xmath127 and @xmath12 are collinear along the direction @xmath253 ; then @xmath230 and @xmath232 coincide with one of the nodes of @xmath88 , independently of the projections of the other points ; hence it is possible to have situations in which not all the @xmath117 cross - ratios coincide , but still we have an intersection of the two mbius curves .
next , we estimate the number of intersections of two mbius curves of degree @xmath0 that satisfy some non - degeneracy conditions , and we show that this number is always less than or equal to @xmath32 ( see proposition [ proposition : bound_intersection_moebius ] ) . to clarify why such a bound should hold ,
let us consider @xmath254 and @xmath255 , two mbius curves of degree @xmath0 .
as we will point out later , there always exist at least two quadrics passing through each of them .
in general , the quadrics passing through @xmath254 will be different from the ones passing through @xmath255 ; up to swapping the roles of @xmath254 and @xmath255 , in general there will be a quadric @xmath256 containing @xmath254 , but not @xmath255 . in this case
@xmath257 instead , as we are going to see , we reach @xmath32 intersections when @xmath254 and @xmath255 are the two components of a curve of degree @xmath258 , complete intersection of two quadrics and @xmath88 . unfortunately , proving that the bound holds in all cases requires the analysis of several particular situations , so we break the proof of proposition [ proposition : bound_intersection_moebius ] into a sequence of lemmata .
these intermediate results are essentially of two kinds : either they predicate some property of subvarieties of @xmath103 containing a mbius curve , or they prove that the bound holds in some particular sub - case .
recall that we denote by @xmath148 the plane in @xmath88 parametrizing the classes of @xmath0-tuples @xmath118 in @xmath84 such that @xmath259 .
[ lemma : moebius_no_plane ] let @xmath168 be a mbius curve of degree @xmath0
. then @xmath168 can not be contained in a plane .
suppose @xmath260 , where @xmath96 is a plane in @xmath103 .
we distinguish two cases : * the plane @xmath96 is completely contained in @xmath88 .
then @xmath96 is one of the @xmath117 planes @xmath148 .
write @xmath261 , then the projections of the points @xmath139 and @xmath145 along any direction coincide , and this is possible only if @xmath262 , but we only consider @xmath0-tuples @xmath82 of distinct points , so we get an absurd . notice that in this case we did not make use of the assumption on the degree of @xmath168 . *
the plane @xmath96 is not contained in @xmath88 .
then @xmath263 is at most a cubic curve , but this is in contrast with the assumption on the degree of @xmath168 .
therefore a degree @xmath0 mbius curve lies on at most one hyperplane in @xmath103 .
let @xmath254 and @xmath255 be two distinct degree @xmath0 mbius curves , and suppose that @xmath254 lies on a hyperplane @xmath264 . if @xmath255 is not contained in @xmath264 , then @xmath265 . by assumption @xmath266 . from the discussion
so far we see that if one of the two mbius curves lies on a hyperplane , then the bound of @xmath32 intersections holds provided that the other is not contained in the same hyperplane .
now we consider the latter case , and prove that the bound still holds .
[ lemma : moebius_no_degree2 ] let @xmath168 be a degree @xmath0 mbius curve
. then @xmath168 can not be contained in a surface of degree @xmath150 completely contained in @xmath88 . the proof is based on a photographic description of surfaces of degree @xmath150 in @xmath88 ( see also remark [ remark : pullback ] ) .
let @xmath267 be a surface of degree @xmath150 .
then @xmath268 spans a hyperplane @xmath264 in @xmath103 , and by degree reasons @xmath269 , where @xmath96 is a plane .
hence @xmath96 is one of the @xmath117 planes @xmath148 in @xmath88 .
consider the rational map @xmath270 defined by the hyperplanes through @xmath96 : its restriction to @xmath88 has a photographic meaning , namely it sends the class of a tuple @xmath118 to the class of the tuple @xmath271 i.e. we remove the points @xmath241 and @xmath242 in @xmath239 , identified with @xmath84 .
moreover , by construction @xmath268 is a fiber of such map .
therefore a mbius curve @xmath261 of degree @xmath0 can not lie in @xmath268 , because otherwise the cross - ratio of the projections of the points @xmath272 along any direction in @xmath23 would stay constant ; this is only possible if four points in @xmath82 are collinear , which is excluded by lemma [ lemma:6_space ] .
[ lemma : moebius_hyperplane_irreducile ] let @xmath168 be a degree @xmath0 mbius curve contained in a hyperplane @xmath264 . then @xmath273 is an irreducible cubic surface .
suppose that @xmath274 is reducible , then either @xmath275 where @xmath96 is a plane and @xmath268 is a surface of degree @xmath150 , or @xmath276 where the @xmath277 are planes .
in both cases a mbius curve , which is irreducible , can not lie in any of the components of @xmath278 because of lemma [ lemma : moebius_no_plane ] and lemma [ lemma : moebius_no_degree2 ] . from now on , for a mbius curve @xmath168 contained in a hyperplane @xmath264 we define @xmath279 .
we know that all @xmath280 are lines ( and not planes ) , because by lemma [ lemma : moebius_hyperplane_irreducile ] the hyperplane @xmath264 can not contain any of the planes @xmath148 .
consider a mbius curve @xmath261 such that @xmath281 for pairwise distinct @xmath282 , i.e. @xmath283 .
then , by construction of the photographic map , the lines @xmath135 and @xmath136 are parallel .
we say that a @xmath0-tuple @xmath82 is _ non - parallel _ if there do not exist pairwise distinct @xmath284 such that the lines @xmath135 and @xmath136 are parallel .
we say that a mbius curve @xmath261 is non - parallel if @xmath82 is so .
[ lemma : lines_distinct ] let @xmath168 be a degree @xmath0 mbius curve contained in a hyperplane @xmath264 , and suppose that @xmath168 is non - parallel . then all @xmath117 lines @xmath285 are distinct .
suppose @xmath286 .
first of all , notice that this can happen only if all @xmath287 are distinct , since if @xmath288 with @xmath289 and @xmath290 , then @xmath148 and @xmath291 meet only in a point .
then @xmath292 since @xmath261 meets every @xmath148 in two ( not necessarily distinct ) points given by the projections long the directions @xmath135 and @xmath293 then @xmath168 would meet @xmath280 and @xmath294 , but this contradicts the assumption that @xmath168 is non - parallel .
[ lemma : intersection_lines ] let @xmath168 be a degree @xmath0 mbius curve contained in a hyperplane @xmath264 , and suppose that @xmath168 is non - parallel
. then a. [ lemma : intersection_lines:4 ] if @xmath295 , then @xmath296 ; b. [ lemma : intersection_lines:3 ] if @xmath297 with @xmath120 and @xmath290 , then @xmath298 if and only if the hyperplane @xmath264 passes through the node of @xmath88 corresponding to configurations @xmath299 for which @xmath300 ; otherwise @xmath301 .
the statement follows from the definition of the lines @xmath280 , from the fact that they are all distinct ( lemma [ lemma : lines_distinct ] ) and from the fact that in case [ lemma : intersection_lines:4 ] the planes @xmath148 and @xmath291 intersect in a line , while in case [ lemma : intersection_lines:3 ] the planes @xmath148 and @xmath291 intersect in a point , a node of @xmath88 .
let @xmath168 be a degree @xmath0 mbius curve contained in a hyperplane @xmath264 , and suppose that @xmath168 is non - parallel .
then @xmath264 can pass through at most one node of @xmath88 .
in fact , if @xmath264 passes though two nodes , then @xmath264 contains a line of the form @xmath302 ; this implies that @xmath168 intersects @xmath302 , which contradicts the hypothesis that @xmath168 is non - parallel .
we are now going through all the possible cases for an irreducible cubic surface @xmath303 obtained by intersecting @xmath88 with a hyperplane containing a degree @xmath0 non - parallel mbius curve @xmath168 .
first we analyze the case when @xmath268 is a singular cubic with non - isolated singularities ( see for example ( * ? ? ?
* case e ) ) : here @xmath268 is either a cone or a projection of a cubic scroll in @xmath103 . in the first case ,
all lines of @xmath268 meet , but this contradicts lemma [ lemma : intersection_lines ] , which asserts that some lines in @xmath268 do not meet . in the second case
there exists a pencil of pairwise disjoint lines , each of which intersects two other special lines on the surface ; this case is again ruled out by lemma [ lemma : intersection_lines ] , which implies the existence of three lines intersecting pairwise ( for example , @xmath304 , @xmath305 and @xmath306 ) .
hence none of these cases can appear in our context .
we are left with the case when @xmath268 is smooth , or has isolated singularities .
we start with the smooth situation .
[ lemma : cubic_smooth ] let @xmath254 , @xmath255 be two distinct degree @xmath0 non - parallel mbius curves .
suppose that both @xmath254 and @xmath255 are contained in a hyperplane @xmath264 and that @xmath307 is a smooth cubic surface
. then @xmath308 .
let @xmath168 be a mbius curve as in the hypothesis .
since @xmath268 is smooth , we can express it as the blowup of @xmath115 at @xmath0 points @xmath309 in general position such that its picard group is generated by @xmath310 where each @xmath277 is the class of the exceptional divisor over @xmath311 , and @xmath173 is the class of the strict transform of a line in @xmath115 , so we have the relations @xmath312 moreover , denoting by @xmath313 $ ] the class in @xmath314 of a divisor , we can put ourselves in the situation where @xmath315 \ ; = \ ; l - e_i - e_j.\ ] ] there exist integers @xmath316 and @xmath317 such that @xmath318 \ ; = \ ; k { \thinspace}l - ( e_1 { \thinspace}e_1 + \dotsb + e_6 { \thinspace}e_6).\ ] ] since @xmath168 intersects each @xmath280 in @xmath150 points , @xmath319 from this we deduce that @xmath320 since @xmath168 is effective , then @xmath321 \cdot e_i \geq 0 $ ] for all @xmath322 , and @xmath321 \cdot ( 2l - e_1 - \dotsb - e_5 ) \geq 0 $ ] , so @xmath323 .
we are going to exclude the cases @xmath324 and @xmath325 . if @xmath324 , then @xmath321 = 2l$ ] .
recall from lemma [ lemma : moebius_extend ] that @xmath168 is contained in a projection of a veronese surface , which is the complete intersection of @xmath88 with another cubic hypersurface @xmath326 .
then @xmath327 is a cubic surface in @xmath27 containing @xmath168 , therefore @xmath328 $ ] is effective where @xmath329 is the canonical divisor on @xmath268 ; recall in fact that @xmath268 is anticanonically embedded in @xmath27 , so @xmath330 = -3k_{s}$ ] in @xmath314 .
but @xmath331 & = 3 { \thinspace}(3l - e_1 - \dotsb - e_6 ) - 2l \\ & = 7l - 3(e_1 + \dotsb + e_6).\end{aligned}\ ] ] on the other hand @xmath332 and this is absurd , since both divisors are effective .
similarly for @xmath325 .
hence @xmath333 .
we are ready to prove the statement . by what we said so far , for @xmath334
we have @xmath335 = ( 2m_i + 2)l - m_i(e_1 + \dotsb + e_6)$ ] for some @xmath336 .
one computes @xmath337 \cdot [ d_2 ] \ ; = \ ; -2(m_1 - 2)(m_2 - 2 ) + 12.\ ] ] then it follows that @xmath338 \cdot [ d_2 ] \leq 14 $ ] , so the statement is proved .
we are left with the situation when the cubic surface @xmath303 has isolated singularities .
taking into account that at least @xmath117 different lines lie on @xmath268 ( see lemma [ lemma : lines_distinct ] ) , by the classification of cubic surfaces ( see for example @xcite ) the only possibilities are : a. a cone over a cubic plane curve ( infinitely many lines ) ; b. one singularity of type @xmath339 ( @xmath340 lines ) ; c. two singularities of type @xmath339 ( @xmath341 lines ) ; d. one singularity of type @xmath342 ( @xmath117 lines ) .
we consider these cases one by one .
let @xmath254 , @xmath255 be two distinct degree @xmath0 non - parallel mbius curves .
then it can not happen that both @xmath254 and @xmath255 are contained in a hyperplane @xmath264 and that @xmath303 is a cone over a cubic plane curve .
this case is ruled out as before by the existence of non - intersecting lines on @xmath268 ( see lemma [ lemma : intersection_lines ] ) .
let @xmath254 , @xmath255 be two distinct degree @xmath0 non - parallel mbius curves .
then it can not happen that both @xmath254 and @xmath255 are contained in a hyperplane @xmath264 and that @xmath303 is a singular cubic surface with two singularities of type @xmath339 .
this case is ruled out by the following argument . if @xmath268 is such a cubic surface , then there are exactly @xmath62 lines passing through each of the two @xmath339 singularities .
in fact such a surface can be realized as the blowup of @xmath115 at @xmath0 points @xmath309 where both @xmath343 and @xmath344 are collinear , but involving different lines , and @xmath345 is in general position with respect to the other points followed by blowing down the ( strict transforms of the ) lines @xmath346 and @xmath347 , which get contracted to the two singularities . in this case , using the same notation as in lemma [ lemma : cubic_smooth ] , the classes of the ( strict transforms of the ) @xmath341 lines of @xmath268 in the blowup of @xmath115 at @xmath309 are @xmath348 in this way the number of lines that pass through a singular point is given by counting how many of the previous ones intersect one of the @xmath349-curves @xmath350 however , in our case from lemma [ lemma : intersection_lines ] we deduce that either we have @xmath0 lines passing through a singularity ( and this happens when the hyperplane @xmath264 passes though one of the nodes of @xmath88 ) or there are at most @xmath150 lines passing through a point .
hence @xmath268 can not have two @xmath339 singularities .
let @xmath254 , @xmath255 be two distinct degree @xmath0 non - parallel mbius curves .
suppose that both @xmath254 and @xmath255 are contained in a hyperplane @xmath264 and that @xmath307 is a singular cubic surface with one singularity of type @xmath339
. then @xmath351 . in this case
we can think of @xmath268 as obtained in the following way : we blow up @xmath115 at @xmath0 points @xmath309 lying on a conic , and then we blow down the @xmath349-curve @xmath352 , which gets contracted to the singular point .
the classes of the ( strict transforms of the ) lines on @xmath268 are @xmath353 the latter @xmath117 lines are the classes of the lines @xmath280 .
the computation for @xmath338 \cdot [ d_2]$ ] goes exactly as in the smooth case because none of the lines @xmath280 passes through the singular point , as one can check by computing the intersection product of @xmath354 with the @xmath349-curve @xmath355 . from this
we conclude that @xmath356 \cdot [ l_{ij } ] = 2 $ ] for all @xmath357 and for @xmath358 , and so we can proceed as in the smooth case ( see lemma [ lemma : cubic_smooth ] ) .
however , we can not directly infer from @xmath338 \cdot [ d_2]$ ] the number of intersections of @xmath254 and @xmath255 .
in fact , if @xmath254 and @xmath255 intersect in the singular point of @xmath268 , such intersection is counted as a contribution by @xmath359 and not @xmath94 , as usual in @xmath338 \cdot [ d_2]$ ] .
luckily in this situation we can exclude that @xmath360 or @xmath361 equals @xmath161 .
in fact , for such a value we obtain the class @xmath362 , which should hence be effective ; on the other hand also the class @xmath363 is effective since the points @xmath311 are arranged on a conic .
but @xmath364 and this is absurd .
hence @xmath365 .
we analyze the possible cases . * either @xmath366 or @xmath367 , then @xmath338 \cdot [ d_2 ] = 12 $ ] . by computing the intersection product with the @xmath349-curve @xmath368
we see that in this case either @xmath254 or @xmath255 does not pass through the singularity , so we can conclude that @xmath369 . * or @xmath370 , then @xmath338 \cdot [ d_2 ] = 10 $ ] . in this case both @xmath254 and @xmath255 pass through the singular point , and moreover both have a node at that point . from this
we conclude that @xmath371 .
let @xmath254 , @xmath255 be two distinct degree @xmath0 non - parallel mbius curves .
suppose that both @xmath254 and @xmath255 are contained in a hyperplane @xmath264 and that @xmath307 is a singular cubic surface with one singularity of type @xmath342
. then @xmath372 . in this case
we obtain @xmath268 by blowing up @xmath115 at @xmath0 points @xmath373 such that @xmath343 are collinear and @xmath374 are collinear on another line , and then blowing down the ( strict transforms of the ) lines @xmath375 and @xmath376 the latter get contracted to the unique @xmath342 singularity of @xmath268 . in this case
the classes of the ( strict transforms of the ) @xmath117 lines of @xmath268 , which coincide with the lines @xmath280 , are @xmath377 notice that the only lines @xmath280 passing through the singular point are the ones whose class is an exceptional divisor @xmath277 , as the computation of the intersection product with the two @xmath349-classes @xmath378 and @xmath379 confirms . here
the computation of @xmath338 \cdot [ d_2]$ ] does not work as in the smooth case , due to the fact that in the blowup of @xmath115 at the points @xmath380 the ( strict transforms of ) the curves @xmath381 may not intersect the ( strict transforms of ) the lines @xmath280 this depends whether or not the curves @xmath381 pass through the singular points .
let @xmath168 be a mbius curve with the properties of @xmath254 and @xmath255 .
let us suppose that @xmath168 does not pass through the singular point of @xmath268 .
then we know that in the blowup of @xmath115 at the points @xmath380 we have @xmath321 \cdot [ l_{ij } ] = 2 $ ] for all @xmath120 . if we write @xmath318 \ ; = \ ; k { \thinspace}l - ( e_1 { \thinspace}e_1 + \dotsb + e_6 { \thinspace}e_6),\ ] ] then the previous conditions translate into @xmath382 \cdot e_i = 2 & \text{for all } i \in \{1 , \dotsc , 6\ } , \\ { }
[ d ] \cdot ( l - e_i - e_j ) = 2 & \text{for all } i \in \{1,2,3\ } , \ , j \in \{4,5,6\}. \end{array}\ ] ] this forces @xmath383 and @xmath384 for all @xmath322 , so @xmath321 = 6l - 2(e_1 + \dotsb + e_6)$ ] .
suppose now that @xmath168 passes through the singular point .
notice that the fact that both @xmath168 and each @xmath280 are real implies that their intersection is real ; moreover , the fact that @xmath261 where @xmath111 is a real map and @xmath114 is a real variety without real points implies that it can not happen that @xmath168 intersects an @xmath280 transversely at the singular point and then in another different point , or tangentially at the singular point .
the only possibility is that @xmath168 has an ordinary node at the singular point . as we pointed out before , the @xmath280 passing through the singular point are the ones whose class is some @xmath277
this is compatible with the situation when the hyperplane @xmath264 passes through a node of @xmath88 , since as pointed out in case iii .
here we have @xmath0 lines passing through the node . therefore @xmath318
\cdot e_i \ ; = \ ; 0 \quad \text{for all } i \in \{1 , \dotsc , 6\}.\ ] ] moreover in this case the strict transform of @xmath168 meets the two @xmath349-curves in two points , so @xmath318 \cdot \bigl ( ( l - e_1 - e_2 - e_3 ) + ( l - e_4 - e_5 - e_6 ) \bigr ) \ ; = \ ; 2.\ ] ] the result is that @xmath321 = l$ ] . summing up
, we have the following scenarios : * both @xmath254 and @xmath255 do not pass through the singular point .
then @xmath338 \cdot [ d_2 ] = 12 $ ] , so @xmath372 . * only @xmath254 ( or @xmath255 ) passes through the singular point .
then @xmath338 \cdot [ d_2 ] = 6 $ ] and so @xmath265 . * both @xmath254 and @xmath255 pass through the singular point , and have a node there
. then @xmath338 \cdot [ d_2 ] = 1 $ ] , and @xmath385 .
the discussion so far proves that the bound on the intersection of two non - parallel degree @xmath0 mbius curves holds if one of them is contained in a hyperplane of @xmath103 .
hence from now on we can suppose that both curves @xmath254 and @xmath255 do not lie on any hyperplane .
riemann - roch predicts that there are at least @xmath150 quadrics in the ideal of a smooth rational curve of degree @xmath0 in @xmath103 ( for more details , see the proof of lemma [ lemma : photo_gen_prop ] ) ; this is true also in the singular case , because in that situation we have an injective homomorphism @xmath386 .
thus each @xmath381 is contained in a pencil of quadrics .
[ lemma : base_locus_quadrics ] let @xmath168 be a degree @xmath0 non - parallel mbius curve not contained in any hyperplane .
let @xmath387 be a pencil of quadrics containing @xmath168 .
then the base locus of @xmath387 is a quartic surface .
if the base locus of @xmath387 were of dimension @xmath161 , then it would be a component of each of the quadrics in @xmath387 , hence all of them would split into two hyperplanes .
but by assumption @xmath168 is not contained in any hyperplane , so the base locus has dimension @xmath150 . then the latter is a complete intersection , and from bezout s theorem it has degree @xmath134 .
in lemma [ lemma : bound_complete_intersection ] and lemma [ lemma : quartic_reducible ] we discuss the situation when the base locus @xmath268 of a pencil of quadrics passing through a mbius curve is irreducible ; in particular we prove that the bound holds when @xmath268 is not contained in @xmath88 , while the case @xmath267 can not occur .
[ lemma : bound_complete_intersection ] let @xmath254 and @xmath255 be two distinct degree @xmath0 non - parallel mbius curves not contained in any hyperplane .
suppose that there is a pencil @xmath387 of quadrics containing @xmath254 whose base locus @xmath268 is irreducible and not contained in @xmath88
. then @xmath351 .
here @xmath254 is a component of the curve @xmath388 , which is a complete intersection of degree @xmath258 by lemma [ lemma : base_locus_quadrics ] and bezout s theorem .
if @xmath255 coincides with the other component of @xmath182 , then the second part of lemma [ lemma : curves_intersection ] proves that @xmath389 .
if this is not the case , then there exists at least a quadric @xmath256 passing through @xmath254 , but not passing through @xmath255 . hence @xmath390 [ lemma : quartic_reducible ] let @xmath168 be a degree @xmath0 non - parallel mbius curve not contained in any hyperplane .
let @xmath387 be a pencil of quadrics containing @xmath168 and suppose that its base locus @xmath268 is contained in @xmath88 .
then @xmath268 is reducible .
suppose instead that @xmath268 is irreducible .
pick any quadric @xmath256 in the pencil @xmath387 : then the intersection @xmath391 is the union of @xmath268 and a surface @xmath392 of degree @xmath150 .
we claim that it is always possible to choose @xmath256 such that @xmath392 splits in the union of two planes .
in fact , each @xmath392 spans a hyperplane @xmath264 , so that @xmath393 , where @xmath96 is a plane .
since the set of planes in @xmath88 is discrete , by continuity we obtain that the plane @xmath96 is always the same , regardless of which @xmath394 we start with . therefore the one - dimensional family of surfaces @xmath392 is obtained by cutting @xmath88 with the pencil of hyperplanes through the plane @xmath96 .
a direct computation for example taking the plane @xmath395 , where the @xmath396 are coordinates in @xmath103 shows that in such one - dimensional family there are always reducible members .
hence we can select @xmath394 such that , after a possible rearrangement of the indices , @xmath397 we intersect both sides of the previous equality with the plane @xmath398 : on the left we obtain either a conic ( if @xmath398 is not contained in @xmath256 ) or the plane @xmath398 itself ( if @xmath399 ) , while on the right we get the union of @xmath400 , @xmath401 ( a line ) and @xmath402 ( another line ) . since @xmath268 is irreducible by assumption , then @xmath398 can not be contained in @xmath268 , so @xmath400 can be either a curve , or a finite set of points .
this forces the left hand side @xmath403 to be a conic . in turn
, this implies that @xmath400 has to be contained in the union of the two lines @xmath401 and @xmath404 . on the other hand ,
since @xmath405 and @xmath168 intersects @xmath398 in two points , then @xmath168 should intersect one of the two lines @xmath401 and @xmath406 , but this contradicts the assumption that @xmath168 is non - parallel .
hence @xmath268 can not be irreducible .
hence we are left with the case when the base locus @xmath268 is reducible .
we notice that it can not happen that @xmath268 is contained in @xmath88 and at the same time splits into the union of two surfaces of degree @xmath150 , because by lemma [ lemma : moebius_no_degree2 ] no mbius curve of degree @xmath0 can lie on a degree @xmath150 surface contained in @xmath88 .
the only remaining cases are when @xmath407 with both @xmath392 and @xmath408 surfaces of degree @xmath150 , but @xmath409 , or @xmath410 where @xmath96 is a plane and @xmath392 is a cubic surface .
let @xmath254 and @xmath255 be two distinct degree @xmath0 non - parallel mbius curves not contained in any hyperplane .
suppose that there is a pencil @xmath387 of quadrics containing @xmath254 whose base locus @xmath268 is not contained in @xmath88 and splits into the union @xmath407 of two surfaces of degree @xmath150
. then @xmath351 .
since @xmath254 is irreducible , then @xmath411 or @xmath412 .
from now on we will suppose @xmath411 .
since @xmath254 can not lie in a degree @xmath150 surface contained in @xmath88 , then @xmath392 is not contained in @xmath88 and hence by degree reasons @xmath413 .
suppose that there exists a quadric @xmath394 not passing through @xmath255 ; then @xmath257 otherwise @xmath414 .
it can not happen that @xmath415 , because otherwise we would have @xmath416 , contradicting the hypothesis .
thus @xmath417 , and so @xmath418 .
therefore @xmath254 and @xmath255 are the two components of the degree @xmath258 complete intersection @xmath388 , and then the second part of lemma [ lemma : curves_intersection ] concludes the proof .
let @xmath254 and @xmath255 be two distinct degree @xmath0 non - parallel mbius curves not contained in any hyperplane .
suppose that there is a pencil @xmath387 of quadrics containing @xmath168 , whose base locus @xmath268 splits into the union of a plane @xmath96 and a cubic surface @xmath392
. then @xmath419 . by lemma [ lemma : moebius_no_plane ] ,
the curve @xmath254 can not be contained in the plane @xmath96 , so @xmath420 .
suppose that @xmath392 is not contained in @xmath88 .
there are several possibilities .
* the intersection @xmath421 is a curve of degree @xmath422 .
then @xmath254 is a component of such curve .
this implies that there is a quadric @xmath423 not passing through @xmath255 , because otherwise we would have @xmath424 .
hence @xmath257 * the cubic @xmath392 splits into a plane @xmath425 and a surface @xmath408 of degree @xmath150 , and @xmath408 is contained in @xmath88 .
this case can not happen , since @xmath254 neither can lie on a plane , nor on a degree @xmath150 surface contained in @xmath88 . *
the cubic @xmath392 splits into a plane @xmath425 and a surface @xmath408 of degree @xmath150 , and @xmath408 is not contained in @xmath88 .
then @xmath426 .
this implies that there exists a quadric @xmath394 not passing through @xmath255 , because otherwise we would have @xmath416 .
hence @xmath257 the last case that needs to be treated is the one where @xmath392 is contained in @xmath88 .
then @xmath392 is irreducible , because @xmath427 can not lie on planes or surfaces of degree @xmath150 contained in @xmath88 ( see lemma [ lemma : moebius_no_plane ] and [ lemma : moebius_no_degree2 ] ) .
hence @xmath392 is a cubic scroll maybe singular , namely a cone over a rational cubic plane curve .
thus @xmath392 admits a determinantal representation as the zero set of the @xmath234 minors of a @xmath428 matrix of linear forms .
we consider the intersection of @xmath392 with the planes @xmath148 : first of all , each @xmath148 is not contained in @xmath392 , because otherwise @xmath392 would be reducible . by restricting the determinantal representation of @xmath392 to @xmath148
we see that @xmath429 is defined by three quadratic equations in @xmath148 , so it is either a finite set of points , or a line , or a conic . on the other hand if @xmath430 then there exists a hyperplane @xmath264 in @xmath103 such that @xmath431 .
therefore @xmath432 since @xmath392 is not contained in @xmath264 ( because @xmath254 does not lie in any hyperplane ) we have that @xmath433 is a cubic curve .
suppose that one among @xmath280 , @xmath294 and @xmath434 , say @xmath280 , is a finite number of points .
then , by eventually rearranging the indices , @xmath294 is a line and @xmath434 is a conic .
this implies that @xmath435 , but this contradicts the hypothesis on @xmath254 of being non - parallel .
therefore we conclude that all @xmath280 are lines .
moreover all lines @xmath280 are distinct , because otherwise this would violate the non - parallel assumption .
we consider now the mutual position of the lines @xmath280 .
first of all we notice that analogously as in lemma [ lemma : intersection_lines ] if @xmath436 and @xmath280 meets @xmath294 , then they intersect at the node of @xmath88 given by the class of @xmath0-tuples @xmath299 for which @xmath300 .
this rules out the case where @xmath392 is a cone , because in that case all lines in @xmath392 meet in a single point , but on the other hand the points @xmath437 and @xmath438 are different , an absurd .
so @xmath392 is smooth , then @xmath392 admits a pencil of mutually disjoint lines , all of which intersect one line @xmath156 .
we distinguish two cases : * suppose that @xmath156 does not appear among the lines @xmath280
. then in particular @xmath304 , @xmath305 and @xmath306 are disjoint . on the other hand , as mentioned before , @xmath439 for some hyperplane @xmath264 ; from ( * ? ? ?
* chapter iii , corollary 7.9 ) we know that a hyperplane section of @xmath392 is connected , so this is absurd . *
suppose that @xmath156 appears among the lines @xmath280 .
after a possible rearrangement of the indices , we can suppose that @xmath440 .
then @xmath441 , @xmath442 and @xmath443 are disjoint , but we can repeat the argument of case a. and see that this leads to an absurd . this concludes the proof of the statement .
we sum up the previous discussion in the following proposition . [
proposition : bound_intersection_moebius ] let @xmath254 and @xmath255 be two distinct degree @xmath0 non - parallel mbius curves
. then @xmath351 .
now we can prove the main result of this section .
[ thm : bound_degree ] the conformal degree of a non - planar and non - equiform hexapod @xmath42 such that both base and platform are non - parallel is at most @xmath31 . since @xmath42 is non - planar , then the mbius curves @xmath254 and @xmath255 of base and platform points are birational by lemma [ lemma:6_space ]
; by the same result , using the non - parallel hypothesis we infer that that no @xmath134 points are on a line , and so we conclude that both @xmath254 and @xmath255 have degree @xmath0 . since @xmath42 is non - equiform , then @xmath254 and @xmath255 are distinct .
theorem [ theorem : bound_degree ] states that the conformal degree of @xmath42 is bounded by @xmath444 , and proposition [ proposition : bound_intersection_moebius ] asserts that @xmath351 , so the statement is proved .
the goal of this section is to provide a construction for a family of movable hexapods , that we will call _ liaison hexapods_. this will be accomplished in two stages : first for each general choice a base we design a candidate for the platform ( subsection [ liaison_hexapods : platform ] ) , then we compute a dilation of the latter and leg lengths that guarantee mobility for the corresponding hexapod ( subsection [ liaison_hexapods : leg_lengths ] ) .
we believe that the family created in this way is maximal among movable hexapods , namely , if we consider the set of movable hexapods as an algebraic variety , it is not contained in any irreducible component of strictly larger dimension .
we begin with the definition of the candidate platform , once we are given a base constituted of @xmath0 general points .
the idea for the construction is the following : in subsection [ degree : photogrammetry ] we associated to any @xmath0-tuple of distinct points in @xmath23 its mbius curve in the moduli space @xmath88 . at this point , we know from bond theory and from the discussion in section [ degree ] that the more the mbius curves associated to base and platform of a hexapod intersect , the more the hexapod has the chance to be movable and the higher will be its conformal degree .
inspired by this , we start from the curve @xmath168 associated to a given general @xmath0-tuple @xmath36 of points in @xmath23 , and we construct from it , applying _ liaison _
techniques , another curve @xmath427 , for which we give evidence to be the curve associated to another tuple @xmath37 of @xmath0 points in @xmath23 ( unfortunately we are not able to exhibit a complete proof of this ) , and that intersects @xmath168 in @xmath32 points . at this stage , the tuple @xmath37
is only determined up to similarities , namely rotations , dilations and translations , and the right scaling factor will be fixed later in subsection [ liaison_hexapods : leg_lengths ] .
we briefly glance at the concept of liaison via an example .
let @xmath445 be a projective curve .
we know that @xmath446 , the homogeneous ideal of @xmath168 , can not be generated by less than @xmath150 polynomials .
hence we can always pick two homogeneous polynomials @xmath447 such that the corresponding surfaces @xmath448 and @xmath449 intersect in a one - dimensional projective set . since we took @xmath450 and @xmath451 in the ideal of @xmath168
, we have that @xmath452 .
the inclusion may be strict ( unless @xmath168 is a so - called _ complete intersection _ ) , and in that case the intersection @xmath453 is the union of @xmath168 and another curve @xmath427 .
we say that the curves @xmath168 and @xmath427 are _ linked _ by @xmath454 .
this procedure can be applied not only to curves in @xmath27 , but also to curves ( and higher dimensional varieties ) in any projective space , as we are going to do for the case of @xmath103 .
linked curves share many properties , and in particular we are interested in a relation between their degrees and genera , expressed by the following result ( see ( * ? ? ?
* chapter 3 , corollary 5.2.14 ) ) : [ proposition : linked_genera ] let @xmath168 and @xmath427 be two projective curves in @xmath103 linked by a complete intersection @xmath455 and let @xmath456 and @xmath457 be their arithmetic genera . then @xmath458 where @xmath459 .
proposition [ proposition : linked_genera ] implies in particular that if the curves @xmath168 and @xmath427 have the same degree , then they have the same genus .
this property will be used in the proof of lemma [ lemma : curves_intersection ] . as it will be clear from several proofs in this section , the construction we are going to propose works only if the base points to start with are sufficiently general .
we would like to make this condition precise , in order to provide later a conjecture that is easily falsifiable .
we say that a @xmath0-tuple of points in @xmath23 is _ mbius - general _ if its mbius curve is smooth , the ideal of the mbius curve contains only two linearly independent quadratic forms , and these two quadratic forms cut out a one - dimensional set from @xmath88 consisting of two smooth curves .
the following lemma , together with remark [ remark : residual_smooth ] , shows that mbius - general @xmath0-tuples form an open subset of the variety of @xmath0-tuple of points in @xmath23 .
[ lemma : photo_gen_prop ] let @xmath82 be a general @xmath0-tuple of points in @xmath23 .
then the mbius curve of @xmath82 is a smooth curve of degree @xmath0 contained in the complete intersection of @xmath88 and two quadric hypersurfaces .
let @xmath168 be the mbius curve of @xmath82 .
since @xmath82 is general , we can suppose that it is non - planar , so from lemma [ lemma:6_space ] the degree of @xmath168 is @xmath0 ( because we have degree @xmath134 only for a special choice of @xmath82 ) .
we prove the smoothness of @xmath168 with the following argument .
what we showed so far is that for a general @xmath82 the curve @xmath168 is rational and of degree @xmath0 ; therefore , it can be thought as a point @xmath321 $ ] in the hilbert scheme @xmath460 of subschemes of @xmath103 with hilbert polynomial @xmath461 .
hence we have a map @xmath462 , where @xmath216 is a suitable zariski - open subset of @xmath463 : the map @xmath464 associates to a @xmath0-tuple @xmath82 the image @xmath321 $ ] of its photographic map @xmath111 .
since smooth rational sextics form an open subset of @xmath460 , if we are able to show that for a particular @xmath0-tuple @xmath82 the curve @xmath168 is smooth , then for all @xmath82 belonging to a ( possibly smaller ) open set @xmath465 the mbius curve of @xmath82 is smooth .
one can check that if we take @xmath82 with @xmath466 then the obtained curve @xmath168 is a smooth sextic . since @xmath168 is smooth and rational
, riemann - roch implies that @xmath467 , i.e. there is at most a @xmath468-dimensional family of quadrics cutting @xmath168 in finitely many point .
there is a @xmath117-dimensional family of quadrics in @xmath103 , so there are at least two linearly independent quadrics passing through @xmath168 .
since in the example we just gave we have exactly two quadrics , then this holds for a general @xmath82 , and the same is true for the fact that the two quadrics form a complete intersection with @xmath88 .
let @xmath82 be a mbius - general @xmath0-tuple of points in @xmath23 .
let @xmath168 be the mbius curve of @xmath82 and let @xmath278 be the complete intersection of degree @xmath258 whose existence is ensured by lemma [ lemma : photo_gen_prop ] .
the curve @xmath427 such that @xmath469 is called the _ residual curve _ of @xmath168 .
[ remark : residual_smooth ] the condition that also the residual curve @xmath427 is smooth is an open condition , and this can be proved as in lemma [ lemma : photo_gen_prop ] , namely showing that this is true in one example , because smoothness is an open property ; one can check that the same @xmath0-tuple @xmath82 that we chose in lemma [ lemma : photo_gen_prop ] yields a curve whose residual one is smooth .
this concludes the proof , initiated in lemma [ lemma : photo_gen_prop ] , that a general @xmath0-tuple @xmath82 is mbius - general .
[ lemma : curves_intersection ] let @xmath82 be a mbius - general @xmath0-tuple of points in @xmath23 .
let @xmath168 be the mbius curve of @xmath82 and let @xmath427 be its residual curve .
then @xmath427 is rational and of degree @xmath0 .
moreover @xmath168 and @xmath427 intersect in @xmath32 points . since @xmath82 is mbius - general , then the curve @xmath427 has degree @xmath0 . from proposition [ proposition : linked_genera ]
we obtain @xmath470 . from (
* remark 5.2.7 ) we have the following exact sequence : @xmath471 & \omega_{d}(-2 ) \ar[r ] & { \mathscr}{o}_{d \cup d ' } \ar[r ] & { \mathscr}{o}_{d ' } \ar[r ] & 0}\ ] ] where @xmath472 denotes the canonical sheaf of @xmath168 . taking the associated long exact sequence in cohomology and using the fact that @xmath473 , one can prove that @xmath474 , namely @xmath427 is connected .
since @xmath427 is smooth by the assumption of mbius - generality , it is irreducible , and so rational .
we are left to prove that @xmath168 and @xmath427 intersect in @xmath32 points .
since @xmath475 is a complete intersection whose ideal is generated by two quadrics and a cubic , the ideal sheaf @xmath476 admits a graded free resolution given by the koszul complex : @xmath477 & { \mathscr}{o}_{{\mathbb{p}}^4}(-7 ) \ar[r ] & { \mathscr}{o}_{{\mathbb{p}}^4_{{\mathbb{c}}}}(-4 ) \oplus { \mathscr}{o}_{{\mathbb{p}}^4_{{\mathbb{c}}}}(-5)^2 \ar[r ] & { \mathscr}{o}_{{\mathbb{p}}^4_{{\mathbb{c}}}}(-2)^2 \oplus { \mathscr}{o}_{{\mathbb{p}}^4_{{\mathbb{c}}}}(-3 ) \ar[r ] & { \mathscr}{i}_{d \cup d ' } \ar[r ] & 0}\ ] ] a computation shows that the euler characteristic @xmath478 equals @xmath468 , which implies @xmath479 because of the exact sequence @xmath471 & { \mathscr}{o}_{d \cap d ' } \ar[r ] & { \mathscr}{o}_{d } \oplus { \mathscr}{o}_{d ' } \ar[r ] & { \mathscr}{o}_{d \cup d ' } \ar[r ] & 0}\ ] ] this proves that @xmath480 is constituted of @xmath32 points .
[ remark : reality_linked ] suppose @xmath168 is the mbius curve of a @xmath0-tuple @xmath82 , and @xmath427 is its residual curve
. then @xmath427 inherits a real structure from @xmath168 .
in fact , by construction @xmath168 is a real curve , so the generators of its ideal can be taken to be all real polynomials ; hence the degree @xmath258 complete intersection @xmath278 is also a real curve , thus by construction @xmath427 is a real curve .
lemma [ lemma : photo_gen_prop ] , together with lemma [ lemma : curves_intersection ] and remark [ remark : reality_linked ] , shows that if we start from a mbius - general @xmath0-tuple , then the residual curve @xmath427 that we obtain satisfies many of the properties that a mbius curve needs to have ( see also proposition [ proposition : characterization_moebius ] ) .
unfortunately , we are not able to establish theoretically that @xmath427 is actually a mbius curve . on the other hand
, we have strong experimental evidence that this holds ; this leads us to formulate the following conjecture . [ clemma : points ] let @xmath82 be a mbius - general @xmath0-tuple of points in @xmath23 .
let @xmath168 be the mbius curve of @xmath82 and let @xmath427 be its residual curve . then @xmath427 is a real variety without real points and is contained in a linear projection , defined by real polynomials , of the third veronese embedding of @xmath115 .
moreover there exists a @xmath0-tuple @xmath481 in @xmath23 ( unique up to translations , rotations and dilations ) such that the mbius curves of @xmath82 and @xmath481 intersect in @xmath32 points .
[ example : georg ] if we take @xmath82 as in the proof of lemma [ lemma : photo_gen_prop ] , then the residual curve @xmath427 is a mbius curve . in fact
, @xmath427 is rational and real without real points , so we can compute a real isomorphism @xmath482 , where @xmath483 .
one notices then that @xmath450 extends to a real morphism @xmath484 .
the preimages of the planes @xmath148 are lines in @xmath115 , and their normal vectors give the difference vectors @xmath485 , for a @xmath0-tuple @xmath481 of candidate points , where @xmath486 with @xmath120 . in this way
it is not difficult to get possible coordinates for the @xmath487 , for instance : @xmath488 eventually one checks that @xmath489 , namely @xmath450 is a photographic map , so @xmath427 is indeed a mbius curve .
[ remark : faith ] if the above conjectured lemma is true for a generic choice of base points , then it is true for all mbius - general @xmath0-tuples of base points , because within the mbius - general instances , the statement that the linked curve is contained in a linear projection of a veronese is a closed property .
despite we do not know whether the conjecture is true or false , we have tested it using random numerical instances of base points , and all the results confirmed it
. it could be that we were always lucky ( or unlucky ) to hit a special case , but we think that this is not very likely , hence our belief in the conjecture . by applying conjectured lemma [ clemma : points ] to a mbius - general @xmath0-tuple @xmath36 of base points
we obtain a candidate platform @xmath37 for the construction of a movable hexapod . from the definition of the photographic map it follows that the candidate platform can by scaled by any non - zero factor without losing any of the properties ensured by conjectured lemma [ clemma : points ] . in the next subsection we are going to determine the right scaling factor , together with leg lengths , leading to a movable hexapod . here
we determine the scaling factor for the @xmath0-tuple of platform points obtained in subsection [ liaison_hexapods : platform ] and the leg lengths so that the resulting hexapods is movable .
this is achieved using the following technique .
as introduced in subsection [ degree : configuration ] , it is possible to assign to every hexapod @xmath42 a projective subvariety @xmath43 of a variety @xmath47 in @xmath46 , the latter being a compactification of the algebraic group of direct isometries of @xmath23 .
moreover , if @xmath43 is a curve we can study its intersection with a particular hyperplane @xmath173 of @xmath46 ; such intersection is constituted in general by a finite number of points , called bonds , contained in the intersection @xmath490 , called the boundary . by imposing tangency conditions at the bonds between the curve @xmath43 and this hyperplane @xmath173
we derive linear conditions on the scaling factor and on the ( squares of the ) leg lengths ensuring that the hexapod we obtain is movable .
in subsection [ liaison_hexapods : platform ] , given general base points @xmath36 , we constructed a candidate for the platform points @xmath37 , unique up to rotations , dilations and translation . from now on ,
we fix one such candidate @xmath37 , and we denote by @xmath491 the vector of points obtained by scaling @xmath37 by a non - zero factor @xmath492 .
the goal is to find a scalar @xmath38 and leg lengths @xmath41 such that the hexapod @xmath493 is movable . by construction , the mbius curves @xmath168 and @xmath427 of @xmath36 and @xmath494 intersect in @xmath32 points , and these points
do not depend on @xmath38 .
as pointed out in remark [ remark : pullback ] , it is not always the case that all intersections correspond to boundary point .
however , this is true if we exclude the situation when both mbius curves pass through one of the nodes of @xmath88 .
hence from now on we will impose another condition on mbius - general @xmath0-tuples , namely we suppose that no three points are collinear . in this way according to theorem [ theorem : bound_degree ] to each of the intersections of @xmath168 and @xmath427 we can associate a boundary point @xmath495 in the following way : if for @xmath188 we have @xmath496 , then @xmath252 and @xmath253 represent directions in @xmath23 such that the orthogonal projection of @xmath36 along @xmath253 and the orthogonal projection of @xmath491 along @xmath252 are mbius equivalent . if @xmath497 is the mbius transformation sending the projection of @xmath36 to the projection of @xmath491 , then by the description of boundary points we gave in subsection [ degree : configuration ] we know that the triple @xmath498 defines a unique boundary point @xmath499 . by a direct computation one sees that as @xmath492 varies , then @xmath500 moves on a line contained in @xmath61 . by repeating this procedure for all @xmath32 points of intersection of the two mbius curves we get @xmath32 boundary points @xmath501 , with @xmath502 . by construction ,
the @xmath503 have the property that if @xmath41 are leg lengths such that @xmath504 has no empty configuration set @xmath505 , then @xmath506 for all @xmath502 .
we now state the tangency conditions we are interested in . to do so , we recall the concept of _ pseudo spherical condition _ ( already used in theorem [ theorem : bound_degree ] ) , which is nothing but the restriction of the spherical condition in eq . to the hyperplane
@xmath173 defining the boundary .
in contrast with the spherical condition , the pseudo spherical condition does not depend on the leg lengths .
the pair of @xmath0-tuples given by @xmath36 and @xmath491 imposes @xmath0 pseudo spherical conditions , which determine a linear space that we denote by @xmath507 .
notice that if we fix a vector @xmath41 of leg lengths , and we create the hexapod @xmath508 , denoting by @xmath509 the linear space cut out by the spherical conditions determined by @xmath510 , then for all boundary points @xmath511 @xmath512 since @xmath513 , where @xmath173 is the hyperplane in @xmath46 determining the boundary @xmath61 . in particular this holds for all @xmath32 points @xmath514 .
[ definition : t2 ] let @xmath36 and @xmath37 be two @xmath0-tuples whose mbius curves intersect in @xmath32 points giving rise to @xmath32 boundary points . following the notation introduced before , denote by @xmath514 the @xmath32 boundary points determined by @xmath36 and @xmath491 , and by @xmath507 the linear space cut out by the @xmath0 pseudo spherical conditions imposed by @xmath36 and @xmath491 .
suppose furthermore that @xmath507 and @xmath47 intersect properly , so that each of the @xmath514 is an irreducible component of @xmath515 .
we say that @xmath516 satisfies the tangency condition @xmath517 if and only if for each of the @xmath32 points @xmath514 the intersection multiplicity @xmath518 of @xmath507 and @xmath47 at @xmath514 is greater than or equal to @xmath150 .
we notice that , with the notation previously introduced , @xmath519 if and only if @xmath520 for every @xmath41 for which @xmath509 and @xmath47 intersect properly . to prove this it is enough to show that @xmath521 this follows from the fact that the projective tangent space @xmath522 of @xmath47 at @xmath514 is contained in @xmath173 ( we used already this fact in theorem [ theorem : bound_degree ] ; this can be proved by a direct computation , using for example grbner bases ) and that @xmath523 and similarly for @xmath509 .
the condition @xmath517 is affine - linear in @xmath38 . in fact , pick one of the @xmath32 boundary points @xmath514 ; from now on we will denote it by @xmath500 .
the condition @xmath517 is equivalent to the condition that the dimension of the intersection of the projective tangent space @xmath524 and the linear space @xmath507 is greater than or equal to @xmath94 .
after possibly reparametrizing the projective line on which @xmath500 lies as @xmath38 varies , we can write @xmath525 , since the line passes through the vertex of the boundary @xmath61 . one can show that @xmath524 is spanned by the rows of the following matrix : @xmath526 where @xmath527 , together with @xmath253 , span the tangent line in @xmath115 of @xmath114 at @xmath253 , and the same for @xmath528 , subject to the condition that @xmath529 .
in particular , the projective tangent space does not depend on @xmath38 .
the @xmath0 pseudo spherical conditions defining @xmath507 determine a linear map @xmath530 , where we identify @xmath531 with the vector space associated to @xmath524 .
its kernel is the vector space associated to the intersection @xmath532 .
we are going to show that the condition @xmath533 is affine - linear in @xmath38 . to do this , we pick coordinates so that ( see ( * ? ? ?
* subsection 2.3 ) ) @xmath534 then a direct inspection of the matrix of @xmath535 proves the statement .
[ example : georg_continued_gamma ] consider the pair of @xmath0-tuples computed in example [ example : georg ] : one finds that each of the @xmath32 affine - linear equations for @xmath38 is a multiple of the equation @xmath536 .
[ definition : t3 ] let @xmath537 with @xmath36 and @xmath37 as before , and suppose that @xmath38 satisfies @xmath517 .
let @xmath509 be the linear space defined by the @xmath0 spherical conditions determined by @xmath538 .
suppose that @xmath509 and @xmath47 intersect properly .
we say that @xmath41 satisfies the tangency condition @xmath539 if and only if for each of the @xmath32 points @xmath514 the intersection multiplicity @xmath540 is greater than or equal to @xmath161 .
[ example : georg_continued_legs ] recall example [ example : georg_continued_gamma ] .
naively one expects that the intersection @xmath541 is greater than or equal to @xmath161 if we are able to find a solution of the system of equations given by @xmath542 in @xmath543/(t^3)$ ] . for a solution of the form @xmath544 , the coefficients @xmath545 and @xmath546
are determined by @xmath514 itself and by a tangent vector in @xmath547 , which is unique up to scaling . for @xmath548
, we obtain a system of affine - linear equations in @xmath549 and @xmath548 .
the solvability of these equations with respect to @xmath548 is equivalent to another system of affine - linear equations in @xmath549 .
these equations are : @xmath550 [ thm : movability ] assume that @xmath36 is a @xmath0-tuple of points in @xmath23 .
assume that @xmath37 is another 6-tuple such that the mbius curves of @xmath36 and of @xmath37 intersect in @xmath32 points and do not intersect in a node of @xmath88 .
assume that @xmath38 and @xmath41 satisfy the conditions @xmath517 and @xmath539 .
then the hexapod @xmath493 is movable . since the two mbius curves do not intersect in a node
, then from the discussion at the beginning of the section we obtain @xmath32 boundary points . as
before , denote by @xmath551 the linear space defined by the @xmath0 spherical conditions determined by @xmath42 .
suppose that @xmath551 and @xmath47 intersect properly in finite number of points , namely that @xmath42 is not movable . by assumption , the intersection multiplicity of @xmath47 and @xmath552 at each of the @xmath32 bonds of @xmath42 corresponding to the @xmath32 common images of the mbius curves is at least @xmath161 .
hence the intersection count gives @xmath553 , which is absurd by ( * ? ? ?
* appendix a , axiom a6 ) .
therefore @xmath42 is movable .
[ remark : liaison_hexapods ] if @xmath36 is a mbius - general @xmath0-tuple of points in @xmath23 , then conjectured lemma [ clemma : points ] predicts the existence of a @xmath0-tuple @xmath37 satisfying the hypothesis of theorem [ thm : movability ] .
[ clemma : legs ] for a mbius - general @xmath0-tuple @xmath36 of base points , consider the @xmath0-tuple @xmath37 of platform points predicted by conjectured lemma [ clemma : points ] .
there is a unique @xmath38 such that the tangency condition @xmath517 is fulfilled .
moreover , the set of leg length vectors @xmath41 fulfilling the tangency condition @xmath539 is of dimension @xmath161 .
[ definition : liaison_hexapods ] let @xmath36 be a mbius - general @xmath0-tuple of points in @xmath23 .
then conjectured lemma [ clemma : points ] and conjectured lemma [ clemma : legs ] predict the existence of @xmath37 , @xmath38 and @xmath41 satisfying the hypothesis of theorem [ thm : movability ] .
we call the resulting hexapod @xmath42 a _ liaison hexapod_. as for conjectured lemma [ clemma : points ] , conjectured lemma [ clemma : legs ] is formulated such a way that if it is false , then it is falsified by a generic choice of base points .
we tested the conjecture against random numerical examples , and constructed many movables hexapods in this way .
but , as in remark [ remark : faith ] , we could have been lucky ( or unlucky ) in all these experiments and the conjecture may be false .
finally , we would like to point out that , in case the two conjectured lemmata hold , the family of liaison hexapods is maximal among movable hexapods .
in fact , the possible bases of such hexapods form an open subset in @xmath554 ; moreover , if the conjectures are true , there exists exactly a @xmath161-dimensional collection ( parametrized by the leg lengths ) of movable hexapods of maximal conformal degree @xmath31 for each such base . this family can not be contained in a larger family with smaller conformal degree since the latter is upper - semicontinuous .
in this section we exhibit two positive - dimensional families of base points @xmath36 for which we can explicitly compute the candidate platform points @xmath37 coming from the liaison procedure in concrete instances as explained in subsection [ liaison_hexapods : platform ] , and such that there exists a unique scaling factor @xmath38 satisfying the condition @xmath517 and a three - dimensional set of leg lengths @xmath41 satisfying the condition @xmath539 . we show that the hexapods corresponding to such families are movable .
[ proposition : family_lines ] consider a @xmath0-tuple @xmath36 of points in @xmath23 such that the lines @xmath555 , @xmath556 and @xmath557 meet in one point ( see fig .
[ figure : concurrent_lines ] ) . then consider the following setting : * take the candidate platform @xmath37 to be : @xmath558 * the scaling factor @xmath38 is given by @xmath559 . * the condition on the leg lengths @xmath560 ensuring mobility one is : @xmath561 then the hexapod that we obtain is movable .
[ proposition : family_order3 ] consider a @xmath0-tuple @xmath36 of points in @xmath23 such that there exists an isometry @xmath562 of @xmath23 of order @xmath161 acting on @xmath36 in the following way : @xmath563 then consider the following setting : * take the candidate platform @xmath37 to be : @xmath564 * the scaling factor @xmath38 is given by @xmath94 . * the condition on the leg lengths @xmath560 ensuring mobility one is given by a system linear in the @xmath565 admitting a three - dimensional solution set .
then the hexapod that we obtain is movable . family ( -3,27)@xmath10 ( 3.5,20)@xmath11 ( 23,20)@xmath12 ( 23,12)@xmath13 ( 60,24.5)@xmath14 ( 52.5,18)@xmath15 ( 8,79)@xmath16 ( 2,72)@xmath17 ( 44,82)@xmath18 ( 70,69.5)@xmath19 ( 26,88)@xmath20 ( 26,77.5)@xmath21 we are going to prove proposition [ proposition : family_lines ] and proposition [ proposition : family_order3 ] using study parameters
. moreover we report some interesting properties observed within this approach , and finally we compute the configuration curve of an example of a liaison hexapod .
due to a result of husty ( see @xcite ) , we use study parameters @xmath566 to compute the configurations of a hexapod this is also known as the forward kinematics problem .
note that the first four homogeneous coordinates @xmath567 are the euler parameters already mentioned in the introduction .
all real points of the study parameter space @xmath568 that are located on the so - called study quadric given by @xmath569 , where @xmath570 correspond to a direct isometry , with exception of the 3-dimensional subspace @xmath571 , as its points do not fulfill the condition @xmath572 with @xmath573 .
the translation vector @xmath574 and the rotation matrix @xmath575 of the direct isometry @xmath576 corresponding to a point in the study quadric are given by : @xmath577 and @xmath578 if the normalizing condition @xmath579 is fulfilled . by using the study parametrization of direct isometries , the condition that the point @xmath7 is located on a sphere centered in @xmath8 with radius
@xmath9 is a quadratic homogeneous equation according to husty ( see @xcite ) . for the explicit formula of the used spherical condition @xmath580 we refer to ( * ? ? ?
* eq . ( 2 ) ) .
the solution for the direct kinematics over @xmath83 of a hexapod can be written as the algebraic variety whose ideal is spanned by @xmath581 , @xmath582 , with the condition @xmath579 . in general this variety consists of a discrete set of points , corresponding to the ( at most ) @xmath28 solutions of the forward kinematic problem . in the case of liaison linkages
it is @xmath94-dimensional as they have a self - motion .
we consider the polynomials @xmath583 for @xmath584 and @xmath585 , which are affine - linear in @xmath586 and therefore they can be written in the form : @xmath587 the system of polynomials @xmath588 can be grouped into the following six sets : @xmath589 denote by @xmath590 the ideal generated by the set @xmath591 , together with the polynomial @xmath581 .
it can easily be seen that each of these ideals is generated by five linear polynomials in @xmath592 : @xmath593 a necessary condition for the solvability of a system of five linear equations in @xmath592 is that the determinant of the extended coefficient matrix is equal to zero .
this condition is denoted by @xmath594 ; e.g. @xmath595 is given by : @xmath596 it is well known ( see @xcite ) that @xmath597 factors into @xmath598 and a quartic factor @xmath599 that has @xmath600 terms . in this way we get six quartic equations @xmath601 in the euler parameter space
. now it can easily be checked by direct computations that @xmath602 holds , i.e. the @xmath603 are linearly dependent .
therefore we can restrict to the equations @xmath604 .
based on this preparatory work we prove the mobility of two classes of liaison examples .
we assume that two cartesian frames called the moving and the fixed frame are rigidly attached to the platform and the base of the hexapods , respectively .
without loss of generality we can choose these frames in a way that the base and platform points have the following coordinates ( with respect to the corresponding frames ) : @xmath605 plugging these coordinates and leg lengths relations @xmath606 , @xmath607 and @xmath608 into our above calculated expressions shows that @xmath599 factors into a linear expression @xmath609 in the euler parameters and a common cubic factor @xmath268 with @xmath610 terms .
this already proves that the mobility is one , because the dimension of the variety of the ideal @xmath611 is at least @xmath150 , as the latter is generated by : @xmath612 since there is only one equation left ( any equation @xmath580 can be taken ) we get at least mobility one over @xmath83 .
note that @xmath613 can split up into several components .
an example for this is the case @xmath614 with pairwise distinct @xmath615 . in this case
the cubic surface @xmath613 splits up into a quadric and a plane . for @xmath616 , @xmath617 and the additional relations @xmath618 , @xmath619 and @xmath620 we get even 3 planes , namely : @xmath621 without loss of generality we can choose the fixed frame and the moving frame in a way that the base and platform points have the following coordinates ( with respect to the corresponding frames ) : @xmath622 moreover with respect to this choice of coordinates the leg lengths can be expressed as @xmath623 with @xmath624 with respect to these coordinates and leg lengths the numerator of @xmath599 splits up into the factor platform and base are congruent . ]
@xmath625 , a linear expression @xmath609 in the euler parameters and a common cubic factor @xmath268 with @xmath626 terms .
this proves that the mobility is one for the same reasons as in the last proof .
finally it should be mentioned that the two families given in propositions [ proposition : family_lines ] and [ proposition : family_order3 ] also contain geometries that are excluded from the conjectures formulated in section [ liaison_hexapods ] ( e.g. planar or congruent hexapods , or hexapods with @xmath134 collinear points ) .
based on random examples ( see section [ study : generic ] ) and the families given in propositions [ proposition : family_lines ] and [ proposition : family_order3 ] we provide the following interesting observations , which hopefully could lead to a simpler ( or even explicit ) computation of the linked @xmath0-tuple in the future .
a. each @xmath599 factors in a linear expression @xmath609 and a common cubic factor @xmath268 for @xmath627 .
every plane defined by a linear equation @xmath628 belongs to a bundle of planes with vertex @xmath629 in the euler parameter space .
therefore there exists a @xmath150-parametric set of linear combinations : @xmath630 moreover , the vertex @xmath629 belongs to the common cubic surface @xmath613 .
b. the vertex @xmath629 does not depend on the remaining leg lengths .
there exists a bijection between the remaining three leg lengths and the translation vector ; i.e. a self - motion can be started from every pose of the platform , if it has this special orientation .
+ therefore this manipulator can be seen as an translational singular manipulator ( or cartesian - singular manipulators ; see ( * ? ? ?
* section 5 ) ) with respect to this orientation .
c. moreover @xmath629 corresponds to an orientation of the manipulator , where the five difference vectors @xmath631 only span a plane @xmath632 with normal vector @xmath633 .
+ now we consider the orthogonal projection of the points @xmath634 and @xmath635 on @xmath632 which yields @xmath636 , @xmath637 .
there exist three pairs of centers @xmath638 in @xmath632 , in a way that the two pencils of six lines @xmath639 , \ldots , [ q_i , p_6^{\prime } ] \quad \text{and } \quad [ q_i , p_1^{\prime } ] , \ldots , [ q_i , p_6^{\prime}]\ ] ] can be mapped onto each other by a congruence sending @xmath640 $ ] to @xmath641 $ ] for @xmath642 .
moreover there is an orientation reversing equiform transformation with @xmath643 for @xmath644 .
observation iii is illustrated in fig .
[ fig0 ] with respect to the example given in section [ study : generic ] .
+ @xmath645 + + [ width=115mm]properties_liaison ( 12,102)base ( 58,102)platform + ( 2.1,21)@xmath646 ( 6.7,16.1)@xmath647 ( 10,25)@xmath648 ( 15.2,19.7)@xmath649 ( 6.3,5.2)@xmath650 ( 11.0,6.5)@xmath651 ( 20.6,1)@xmath652 ( 26.2,25.2)@xmath653 ( 23,37)@xmath654 ( 25.8,43)@xmath655 ( 27.5,32)@xmath656 ( 23.8,29)@xmath657 ( 18,36.5)@xmath658 ( 56,2)@xmath659 ( 44,16.5)@xmath660 ( 46.8,18.2)@xmath651 ( 50.2,24.5)@xmath661 ( 46,36)@xmath662 ( 56.5,41.1)@xmath663 ( 66,37)@xmath664 ( 73.5,34.5)@xmath665 ( 58,30)@xmath666 ( 80.5,60)@xmath667 ( 70.8,32)@xmath668 ( 67.2,29)@xmath669 ( 61,36.5)@xmath670 ( -0.2,65)@xmath671 ( 8,65)@xmath672 ( 16,65)@xmath559 ( 25.7,65)@xmath673 ( 43,65)@xmath671 ( 51,65)@xmath672 ( 59,65)@xmath559 ( 68.7,65)@xmath673 ( 77.1,65)@xmath94 ( 2,82.2)@xmath674 ( 6.4,73)@xmath675 ( 10,93)@xmath676 ( 15.8,84)@xmath677 ( 26,85.8)@xmath678 ( 22.5,75)@xmath679 ( 27,75)@xmath680 ( 27.5,79.5)@xmath681 ( 23,70.2)@xmath682 ( 18,75)@xmath683 ( 70.8,79.5)@xmath684 ( 67,70.2)@xmath685 ( 61,75)@xmath686 ( 55.8,84)@xmath687 ( 50.2,73)@xmath688 ( 44.2,82.4)@xmath689 ( 56.5,93.4)@xmath690 ( 70.5,73)@xmath691 ( 73.7,85.8)@xmath692 ( -1.5,73)@xmath673 ( -1.5,81.3)@xmath94 ( -1.5,89.6)@xmath150 ( -1.5,97.9)@xmath161 ( 40.5,73)@xmath673 ( 40.5,81.3)@xmath94 ( 40.5,89.6)@xmath150 ( 40.5,97.9)@xmath161 ( -3.5,2)@xmath693 ( -3.5,10.3)@xmath694 ( -3.5,18.6)@xmath695 ( -3.5,26.8)@xmath696 ( -3.5,35.1)@xmath697 ( -3.5,43.3)@xmath698 ( -3.5,51.6)@xmath699 ( -3.5,59.7)@xmath700 ( 38.5,2)@xmath693 ( 38.5,10.3)@xmath694 ( 38.5,18.6)@xmath695 ( 38.5,26.8)@xmath696 ( 38.5,35.1)@xmath697 ( 38.5,43.3)@xmath698 ( 38.5,51.6)@xmath699 ( 38.5,59.7)@xmath700 d. the cubic surface @xmath613 in the euler parameter space contains a line @xmath156 through the vertex @xmath629 . the points of these line @xmath156 correspond to the rotation of the platform about a fixed line orthogonal to @xmath632 . in each configuration of the @xmath150-parametric set , obtained from the composition of this rotation and a translation of the platform in direction of @xmath633 , a self - motion can be started .
+ [ remark : schoenflies ] the planar hexapod with platform @xmath701 and base @xmath702 is even schnflies - singular ( see @xcite ) with respect to the direction orthogonal to the @xmath632-parallel carrier planes of the planar platform and planar base .
e. there exists a regular projectivity mapping @xmath8 to @xmath7 for @xmath703 .
we choose the following set of base points with respect to the fixed frame @xmath704 : @xmath705 for @xmath706 with @xmath139 given in eqs . and .
according to example [ example : georg ] the liaison technique explained in section [ liaison_hexapods ] yields the following platform with respect to the moving frame @xmath707 : @xmath708 for @xmath706 with @xmath487 given in eqs .
to . note that this coordinatization of the platform already corresponds with the special orientation @xmath629 .
moreover the leg lengths are given by eqs . to .
for the computation of the self - motion we express @xmath709 from @xmath710 , @xmath711 , @xmath712 , @xmath569 for pairwise distinct @xmath713 and insert them into @xmath714 .
we denote the numerator of the resulting expression by @xmath715 with pairwise distinct @xmath716 .
this is an octic expression in the euler parameter space , where @xmath717 appears maximally to the power of @xmath0 . due to the involved powers of @xmath717
, we eliminate this euler parameter by computing the resultant of @xmath268 and @xmath715 , which yields the expression @xmath718 of degree @xmath719 in @xmath720 .
the greatest common divisor @xmath721 of all @xmath718 corresponds to the self - motion , which is in the generic case of degree @xmath258 .
it is always possible to choose special values for the remaining leg lengths @xmath722 in a way that @xmath268 is linear in @xmath717 and that all @xmath715 are of degree @xmath62 in @xmath717 . for our example
the overdetermined system of equations resulting from the coefficients of @xmath723 , @xmath724 , @xmath725 of @xmath268 and the coefficients of @xmath726 , @xmath727 , @xmath728 , @xmath729 , @xmath730 , @xmath731 of @xmath715 has the following solution : [ width=120mm]euler @xmath732 in this case @xmath721 is only of degree @xmath121 .
this decic @xmath733 is illustrated in fig .
[ fig1 ] and it consists of two components , a black and a red colored one . on the red component a point is highlighted , whose corresponding configuration is illustrated in fig .
[ fig2 ] together with the trajectories corresponding to the red component of the self - motion .
finally it should be noted that the number of @xmath32 bonds can also be verified within the study parameter approach according to the method presented in @xcite .
the first - named and third - named author s research is supported by the austrian science fund ( fwf ) : w1214-n15/dk9 and p26607 - `` algebraic methods in kinematics : motion factorisation and bond theory '' .
the second - named author s research is funded by the austrian science fund ( fwf ) : p24927-n25 - `` stewart gough platforms with self - motions '' .
( 23.7,3)@xmath19 ( 6,18)@xmath18 ( 26,53)@xmath17 ( 36.5,56.5)@xmath16 ( 52,14)@xmath20 ( 88.7,16)@xmath21 ( 41,41)@xmath13 ( 47.2,30.5)@xmath12 ( 60,64.8)@xmath14 ( 67,77)@xmath15 ( 88,27)@xmath11 ( 97,32)@xmath10 ernest duporcq , _ sur la correspondance quadratique et rationnelle de deux figures planes et sur un dplacement remarquable _ , comptes rendus des sances de lacadmie des sciences * 126 * ( 1898 ) , 14051406
. manfred husty and adolf karger , _ self motions of stewart - gough platforms , an overview _ , proceedings of the workshop on fundamental issues and future research directions for parallel mechanisms and manipulators ( quebec city , canada ) , 2002 . | the complete classification of hexapods also known as stewart gough platforms of mobility one is still open . to tackle this problem , we can associate to each hexapod of mobility one an algebraic curve , called the configuration curve . in this paper
we establish an upper bound for the degree of this curve , assuming the hexapod is general enough .
moreover , we provide a construction of hexapods with curves of maximal degree , which is based on liaison , a technique used in the theory of algebraic curves . |
quasicrystals are by now well - established in solid state physics .
they exhibit non - crystallographic symmetries together with a number of unusual physical properties .
recent reviews may be found in refs .
@xcite .
there has recently been much interest in atomic self - diffusion in quasicrystals , mostly triggered by a paper by kalugin and katz@xcite where a diffusion mechanism specific to quasicrystals was proposed .
the elementary process in this _ flip mechanism _ consists of certain quasicrystal - specific rearrangements of atoms , where the initial and final configurations are energetically almost degenerate . in quasicrystals which are decorations of quasiperiodic tilings ,
the flip mechanism consists of a reshuffling of certain local tile configurations , along with their decorations @xcite . up to now experiments
have not given a clear answer to the diffusion behaviour of quasicrystals .
coddens and his group @xcite found atomic hopping of cu and fe in alcufe quasicrystals . the process can be regarded as flips but it behaves differently from the proposal of kalugin and katz .
tracer diffusion has been studied by nakajima et al .
their results indicate an unusually low diffusivity of mn in alpdmn quasicrystals as compared to crystals of similar composition . to shed more light on these results one has to use simulations to find the elementary diffusion moves and their energetical properties .
flip diffusion has so far been studied in pure tiling models , without bothering about a specific atomic decoration of the tilings @xcite .
while such an approach indeed proves that elementary flip processes do add up to diffusive behaviour , the physical feasibility of the flip mechanism and the magnitude of flip diffusion remain much less certain . in particular , activation energies of elementary flips can not be estimated without a concrete atomic structure , and in a reshuffling of tiles atoms may have to move only much smaller distances than the vertices of the tiles .
those questions can not be studied without having a specific atomic structure in mind .
some preliminary work treating this problem has been published in ref .
@xcite .
we propose to study the feasibility of flip diffusion in a concrete atomic model quasicrystal , by means of molecular dynamics ( md ) simulations .
for such a simulation , not only a realistic model structure , but also ( short range ) interatomic potentials stabilizing the model structure are needed .
fortunately , this has become available : dzugutov @xcite has found a one - component system with a simple potential , which solidifies into a quasicrystalline structure known already before as a realistic model of dodecagonal quasicrystals @xcite . in section 2 , we shall describe this model structure in more detail , along with some ideas on potential flip moves , before we present our md simulation results in section 3 . in section 4 activation energies and frequencies for the flips are given .
we finally conclude in section 5 .
the structure found by dzugutov @xcite in his md simulation is a layered structure which , apart from some defects , is periodic in one direction , but quasiperiodic and 12-fold symmetric in the plane perpendicular to it .
it basically is of frank - kasper type , i.e. , it is mostly tetrahedrally close - packed , and can be described as a periodic stacking @xmath0 of a dodecagonal layer @xmath1 and two hexagonal layers , @xmath2 and @xmath3 , which are just rotated by @xmath4 with respect to each other .
the atoms in layer @xmath1 form the vertices of a simple tiling of squares , triangles , asymmetric hexagons and @xmath4 rhombi .
the whole structure is , in fact , a decoration of such a tiling @xcite .
the tiles occurring in the structure , together with their decorations , are shown in fig .
[ basictile ] .
layer atoms is replaced by a slightly jagged column of @xmath2/@xmath5 layer atoms , or vice versa .
these columns are shown on the right .
a whole column thus has to move a quarter of a period length upwards or downwards .
, width=529 ] all structures built on a tiling with squares and triangles only are perfectly tetrahedrally close - packed .
such structures are therefore very rigid , and there are also very few small groups of tiles that can be reshuffled . if hexagons or rhombi are present , however , some atoms are not close - packed : there are some octahedra occurring in the interior of hexagons , and near the obtuse corners of rhombi . near those octahedra ,
the structure is much softer .
moreover , in such structures there are also many local tile configurations which can easily be reshuffled .
some of these tile flips are shown in fig .
[ basictile ] . in all the moves ,
a whole column of atoms has to move upwards or downwards ( fig .
[ basictile ] ) , whereas other atoms move very little .
obviously , such a move is very unlikely , since the vertical distance of atoms is already rather small .
if one introduces a vacancy in the column , however , the move becomes much easier .
since between two atom positions in a column there is another good atom position , a neighboring atom can then move half - way into the vacancy , effectively splitting it into two half - vacancies , which then can move up and down the column , thereby transforming it .
such half - vacancies therefore efficiently catalyze the moves shown in fig .
[ basictile ] .
they are responsible for the breaking of periodicity in z - direction .
note that the movement of atoms in the xy - plane is much smaller than that of the vertices . in our simulations , structures built on several different tilings
are used , both perfectly ordered as well as disordered ones .
two different geometries have been studied : a cubic box with each sample containing about 8000 atoms and a flat tetragonal box with aspect ratio of about 7.5:1 containing 14040 atoms .
our md simulations were all carried out with a standard 4th - order predictor - corrector algorithm@xcite , at constant temperature and constant pressure@xcite using the constraint method . to keep atoms sufficiently mobile , a rather high simulation temperature of @xmath6 ( in reduced units )
was used , compared to a melting temperature of @xmath7 for the idealized structures , and @xmath8 for dzugutov s structure , which was obtained in a cooling simulation @xcite .
since dzugutov s structure contains many defects of all kinds , we decided to use , in our simulations , the idealized structures described in the last section , and to introduce defects in a controlled way where necessary .
we will start with the description of the results of the cubic case , where the simulation box contains 10 double layers along the periodic direction .
the pure square - triangle tiling structures turned out to be very stable , as predicted . the same is true for structures containing asymmetric hexagons .
regions with isolated rhombi are found to have a clear tendency to transform into hexagons , and then remain immobile . only more disordered structures , where also pairs of rhombi occur , are somewhat more mobile .
after very long simulations , these latter structures even developed vacancies , whereas the other structures only developed half - vacancies . in order to accelerate this process , and since dzugutov s structure also contains many holes , we decided to introduce some vacancies artificially , in a controlled way , on different classes of sites .
vacancies are most effective when they are introduced on sites in the interior of hexagons , which are involved in the formation and disappearance of rhombi inside hexagons . when introduced on other sites , vacancies had a somewhat lesser effect . as expected , mobility is primarily in the periodic direction , whereas in the quasiperiodic plane atoms move much less ( fig .
[ compconf ] ) .
it can clearly be seen , however , that the tiling is reshuffled , and periodicity is broken . in agreement with our theoretical picture ,
atoms indeed make discrete jumps in our simulations ( fig .
[ histdist ] ) .
these distances covered are much bigger in the periodic direction than in the quasiperiodic direction since in the first case an atom may jump several times during the simulation . in the latter case jumps
are still discernible , although only few atoms jump and most do nt move at all .
the behaviour of the structural changes is different in the flat tetragonal sample containing only three double layers along the periodic direction .
now the flips occur in _ all _ configurations and they do not terminate , and we observe an equilibrium structure containing squares , triangles , hexagons and single rhombi . even in the pure square - triangle tiling the atoms become mobile but only if we increase the simulation temperature up to @xmath9 .
several mobility processes can be distinguished at the tiling level ( fig .
[ flatsim ] ) : pairs of rhombi are eliminated immediately .
the reason for this will be discussed in the next section .
if there are untilable defects then they disappear after about 16,000 time steps .
the equilibrium density of hexagons and single rhombi , established by their transition into one another , is reached after about 70,000 time steps , but the number of hexagons keeps showing large fluctuation and a very slow relaxation during the whole simulation run ( up to 700,000 time steps ) . on the atomic scale
we observe phason hopping .
this represents the motion of the atoms corresponding to the tile flip .
the hopping is in almost all cases only local , back and forth . it does not lead to long range diffusion .
furthermore nearest neighbour jumps typical for ordinary vacancy diffusion have been found .
a few long - range diffusion processes in the periodic plane have been found , the overwelming part of diffusion however is along the periodic direction as seen already in the cubic cell simulations .
therefore the diffusion in this model seems to be maximally anisotropic .
for the flat sample it is possible to describe the flipping process more quantitatively . a jump of an atom chain along the periodic direction occurs every 110,750 time steps .
the cosine of the angle between the directions of two successive jumps of the same atom is -0.169 on average , which means that successive jumps are nearly uncorrelated .
the frequencies for the jumps are the following : a jump of any tile occurs every 12,283 time steps , a jump of a rhombus every 21,123 and a jump of a hexagon every 29,288 steps . for rhombus - hexagon - rhombus
the branching ratio between the backward and the two forward jumps is 4:1:1 , and for a hexagon - rhombus - hexagon jumps it is 13:4 . to determine the activation energies we have optimized the potential energy of small samples containing the flipable configurations .
the energy of the configuration with the pairs of rhombi ( bottom of fig .
[ basictile ] ) is about 5.4 ( in reduced units ) per period higher than the energy of the square - triangle configuration .
the energy of the hexagon with the symmetric interior is nearly identical to the hexagon containing a rhombus ( it is 0.1 per period lower ) .
if there are three rhombi in the hexagon the energy increases by 6.4 per period . therefore this configuration does not occur .
the barrier energies can be found by molecular static calculations @xcite .
the jumping atoms are slowly dragged from one configuration to the other , while monitoring the energy along this path . in this way it is possible to find the saddle point configurations and their energies .
special care is needed , however , to avoid introducing artificial constraints , which might increase the energy . for the configurations in fig .
[ basictile ] it turns out that the pairs of rhombi actually display saddle point configurations .
that is the reason why they disappear immediately at the beginning of the simulation .
the barrier for the flipping process shown in the upper part of fig .
[ basictile ] is too low to be determined accurately , the barrier for the opposite process ( not shown ) transforming a hexagon into a second hexagon via a common central rhombus is also very low : 0.5 .
if we take the geometry of the simulation cell into account we find that there are percolating clusters of hexagons sharing a rhombus
. therefore long range tile diffusion should be possible .
a careful analysis of the final states has shown that flip moves indeed occur in our model quasicrystals .
such flip moves have also been observed by dzugutov@xcite .
their activation energy is prohibitively high if the periodic direction is long , unless there are other defects present as well , which can catalyze the flips .
half - vacancies and vacancies are particularly efficient such catalyzers .
we have found that a sizeable density of such half - vacancies and vacancies ( up to 2% ) is always present in equilibrium , which therefore makes flip diffusion possible .
unfortunately , due to the ( necessary ) presence of vacancies and half - vacancies , the contributions of flip diffusion and classical vacancy diffusion are very hard to separate .
we should emphasize , however , that the flip mechanism , although catalyzed by vacancies , is qualitatively different from vacancy diffusion , in that the passage of a vacancy without flips associated with it leaves the structure unchanged , whereas with the flips the structure is left completely reshuffled after the passage of the vacancy .
if a flat geometry is chosen it is possible to study tiling flips and their dynamics explicitly . if the starting configuration contains untilable defects , it is possible to observe tiles move finite distances . in most of the cases , however , the tiles and the associate atoms move only back and forth as they do in a two - well - potential . for geometrical reasons
it is not clear why this should happen , since it is possible for the tiling moves to percolate .
maybe the long - range diffusion process is too slow to be seen in molecular dynamics . since we have determined all the energetical parameters for the flips it should be possible to use them in a monte carlo simulation to clarify these questions and to calculate the diffusion constants accurately .
this is planned for the future .
our model quasicrystals are somewhat untypical in that they are periodic in one direction , which causes problems because isolated flips have to break periodicity and thus lead to larger mismatches
. moreover , our models are mostly tetrahedrally close - packed , which makes them very rigid , and which is also not very typical for quasicrystals .
still , we believe that our results are relevant also for ( stable ) icosahedral quasicrystals , which are completely non - periodic and not frank - kasper like .
the authors would like to thank to m. dzugutov for making available the structure obtained in his simulation , and c. beeli for suggesting the atomic flip moves shown in fig . 1 .
part of this work was supported by the swiss nationalfonds , deutsche forschungsgemeinschaft , doe grant de - fg02 - 89er-45405 , and packard foundation grant fdn 89 - 1606 during a common stay of the authors at lassp , cornell university , ithaca , new york .
the computations were performed on facilities of cornell materials science center and of the institut fr theoretische und angewandte physik .
f. hippert and d. gratias ( eds . ) , _ lectures on quasicrystals _ , ( les ditions de physique , les ulis 1994 ) ; c. janot , _ quasicrystals : a primer _ , ( clarendon press , oxford 1994 ) .
p. a. kalugin and a. katz , _ europhys .
* 21 * , 921 ( 1993 ) .
g. coddens , c. soustelle , r. bellisent , y. calvayrac , _ europhys .
* 23 * , 33 ( 1993 ) , s. lyonnard , g. coddens , y. calvayrac , d. gratias , _ phys .
rev . _ * b53 * , 3150 , ( 1996 ) .
h. nakajima , j. asai , k. nonaka , i. shinbo , a. tsai , t. masumoto , _ mat .
* a181 / a182 * , 738 ( 1994 ) .
d. joseph , m. baake , p. kramer and h .-
trebin , _ europhys .
* 27 * , 451 ( 1994 ) ; d. joseph and f. ghler , in _
aperiodic 94 _ , eds .
g. chapuis , w. paciorek ( world scientific , singapore 1995 ) , p. 188
, m. v. jari and e. s. srensen , _ phys . rev . lett . _ * 73 * , 2464 ( 1994 ) .
f. ghler , j. roth , in _
aperiodic 94 _ , eds
. g. chapuis , w. paciorek ( world scientific , singapore 1995 ) , p. 183 .
m. dzugutov , _ phys .
lett . _ * 70 * , 2924 ( 1993 ) .
f. ghler , in _ quasicrystalline materials _ , ed .
janot and j. m. dubois ( world scientific , singapore , 1988 ) , p. 272 . c. beeli , f. ghler , h .- u .
nissen and p. stadelmann , _
j. phys .
france _ * 51 * , 661 ( 1990 ) .
allen and d. j. tildesley , _ computer simulations of liquids _
( clarendon press , oxford , 1987 ) .
f. lanon , l. billard , _ j. phys .
france _ * 49 * , 249 ( 1988 ) .
m. dzugutov , _ europhys .
_ * 31 * , 95 ( 1995 ) .
j. r. beeler , _ radiation effects computer experiments _ , ( north holland , amsterdam 1983 ) | we present a molecular dynamics study on atomic self - diffusion in frank - kasper type dodecagonal quasicrystals .
it is found that the quasicrystal - specific flip mechanism for atomic diffusion as predicted by kalugin and katz@xcite , indeed occurs in this system .
however , in order to be effective , this mechanism needs to be catalyzed by other defects such as half - vacancies if the structure is truly three - dimensional . for this reason ,
flip diffusion is difficult to distinguish from standard vacancy diffusion . in a quasi - two - dimensional setup ,
however , the flips may occur without other defects .
activation energies and flip frequencies are also determined . |
estimates for the ambient neutrino flux are an important input parameter to experiments analyzing underground neutrino interactions and upward going muons to understand the time evolution of a neutrino beam .
different directions provide neutrino fluxes with varying source distance so observation of the angular distribution is an essential tool in the study of neutrino oscillations .
since the overall flux normalization is uncertain , experiments frequently place a greater emphasis on the shape of the distribution than the absolute event rate .
this note points out _ depth dependent _ effects that can also provide a directional modulation .
these effects are modest , but predominantly effect the higher end of the neutrino spectrum .
upward going muons @xcite are usually attributed to muon neutrino interactions in the rock surrounding the detector . in order to be detected as an upward going muon , the lepton produced in a neutrino interaction must propagate through the rock to be recorded in the detector .
if we approximate the muon energy loss as muon energy independent then the range increases linearly with the muon energy .
so the target rock surrounding the detector has a larger effective mass for neutrino interactions at high energy , scaling roughly as @xmath0 . over a substantial range of neutrino energies
the cross section rises linearly . so that a constant mass detector will have more high energy neutrino interactions than low energy neutrino interactions at the same flux .
these two factors suggest that the neutrino induced muon flux is sensitive to the third power of the neutrino energy .
small neutrino flux differences at high energies are substantially amplified by this @xmath1 factor .
we present a one dimensional model to show that the atmospheric decay path length is a function of the detector depth .
detectors which are above sea level will see neutrinos with a higher decay path length than detectors below sea level . to first order
the high energy part of the neutrino flux is proportional to the decay path length .
figure [ geom ] illustrates the geometry . this figure is not to scale .
we take @xmath2 to be the radius of the earth , 6380 km and @xmath3 to be the radius at which neutrino source particles are produced .
@xmath4 will be one interaction length below @xmath3 .
most decays will occur between @xmath3 and @xmath4 .
@xmath5 represents the detector depth .
if the detector is above sea level @xmath5 will be negative .
we take as the decay length the difference in length for ray originating at @xmath5 and ending at a point along the ray at radius @xmath3 or @xmath2
. it should be clear from figure [ geom ] , with its disproportionate scale that the decay length will depend on detector depth .
for muons we take @xmath6 , the surface of the earth .
a particle initiating in the upper atmosphere will travel a distance s before being absorbed .
@xmath7 very near the horizon , the particle path may not intersect @xmath4 , if the detector is above sea level ( @xmath8 ) . in that case
we approximate the decay length by the distance from the upper atmosphere to the detector .
@xmath10 is the zenith angle .
note for upward going neutrinos @xmath11 . in figure [ maxrel ]
we illustrate this effect for muons , where we take @xmath6 and @xmath12 km .
the figure shows the maximum decay length for muons below the horizon . in most cases
this maximum is obtained at the horizon .
but when the detector is above sea level , @xmath13 , the maximum decay path length is achieved near the horizon . notice that the decay path length increases with the detector height
. common detector depths , @xmath5 are in the range of -1 km @xmath14 2 km .
detectors located in the mountains tend to be deep underground but well above sea level .
in figure [ murel ] we explore the relative flux , as a function of neutrino parent particle energy , for 4 different detector depths .
this figure plots the contribution to the neutrino flux of a detector at @xmath15 - 2 , -1 , 1 or 2 km .
relative to the contribution to the neutrino flux for a detector located at sea level ( @xmath16 ) .
the flux is averaged over the solid angle region of @xmath17 , the angular bin just below the horizon .
variations are about 5% to 10% . in figure [ murel ]
one sees that the enhancement is not present at low energies , where all particles will decay .
there is a transition region at moderate energies where the decay length is comparable to the absorption length . at the highest energies the decay probability scales linearly with the available decay length
so the flux differences directly reflect the path length differences due to detector depth .
the shape differences for muons , pions and kaons are due to the differences in masses , lifetimes and absorption lengths .
figure [ murel2 ] is similar to figure [ murel ] except that now the solid angle region @xmath18 is considered .
variations are now of the order of 0.6% to 1.2% .
the much lower flux modification away from the horizon indicates that standard neutrino flux calculations , that do not include the detector depth , will not correctly represent the angular distribution .
figure [ mpkprob ] plots the muon , pion and kaon decay probability as a function of decay particle energy .
the contribution at high energies is dominated by the particle with the shortest lifetime .
figure [ shower ] illustrates the depth effect in hadronic shower evolution .
the multiple curves are for hadrons initiating at greater depth into the atmosphere .
each curve is 2 absorption lengths deeper in than the one to its right .
the depth modulation is maintained at approximately the same amplitude , but as showers go deeper into the atmosphere the curves move to lower energies since the absorption length drops .
higher energy hadrons are more likely to be absorbed than to decay when the are formed lower in the atmosphere .
this paper has made no attempt to quantitatively sum the various contributions to the atmospheric neutrino flux .
muons , pions and kaons all contribute but their relative contributions depend on their initial fractions and the fate of the other contributions . at very high energies `` prompt '' sources of neutrinos such as
charm decay become important . at the highest energies muons produced via decays of the form @xmath19 and
@xmath20 , or drell - yan processes constitute a significant source of neutrino flux .
our analysis suggests that there is a modest location dependent contribution to the high energy atmospheric neutrino flux .
detectors above sea level ( but still underground ) will see enhancements of the high energy flux in the vicinity of the horizon .
this would be manifest as an angular distortion and an increased rate .
detectors below sea level would expect to see the opposite effect , with fewer high energy events and fewer events near the horizon . while these effects are of the order of 5% to 10% they contribute a systematic distortion when comparing data taken at different locations or when comparing observations from a deep detector to a flux estimate for a sea level location .
todor stanev has recently reached similar conclusions about the influence of detector depth .
i would like to thank o. ryazhskaya for pointing out the significance of muon production via hadronic resonance decay ( @xmath19 ) as a major source of neutrinos at high energies .
this work was supported in part by the us department of energy under grant de - fg02 - 00er41145 .
long m. ambrosio _ et al .
_ , phys . lett . * b434 * , 451 ( 1998 )
. s. hatakeyamma _ et al .
_ , phys .
lett . * 81 * , 2016 ( 1998 ) .
r. becker - szendy _ et al .
_ , phys .
lett . * 69 * , 1010 ( 1992 ) .
mcgrath , `` a treatise on high energy muons in the imb detector '' , ph.d .
thesis , university of hawaii at manoa ( 1993 ) .
y. oyama , `` experimental study of upward - going muons in kamiokande '' , ph.d .
thesis , institute for cosmic ray research , university of tokyo ( 1989 ) . | we note that detector depth can influence the decay path length available for the primary and secondary particles that are the source of atmospheric neutrinos . as a consequence
there is a location dependent modulation to the neutrino flux , which could be as large as 5 - 10% in some directions . |
the inability of the visible mass components in disk galaxies to account for the observed rotation curves is usually interpreted as evidence for the existence of an additional , invisible mass component .
other theories suggest that this mass discrepancy is an indication of a breakdown of classical newtonian dynamics .
it is difficult to evaluate these theories , as only a few make specific and testable predictions .
one of the exceptions is the modified newtonian dynamics ( mond ) , advocated by milgrom ( 1983 , 1989 ) and sanders ( 1990 , 1996 ) .
this theory postulates that newton s law of gravity should be modified for very small accelerations , with the result that any need for dark matter disappears .
fits to rotation curves of hsb galaxies using mond are of equal quality as the fits made using a dark matter halo ( see sanders 1996 ) .
mond is however also able to satisfactorily explain observations of the dynamics of e.g. dwarf galaxies and dwarf spheroidals ( see the discussion in milgrom 1995 , and also mcgaugh & de blok ( 1998b ) [ hereafter paper ii ] ) . for a complete description of mond , its predictions , and observational results
we refer to milgrom ( 1983 , 1989 ) , sanders ( 1990 ) , begeman , broeils and sanders ( 1991 ) , bekenstein & milgrom ( 1984 ) and sanders ( 1996 ) . an extensive description of mond results in the context of lsb galaxies is given in paper ii .
mond assumes that the force law changes from the conventional newtonian form when the acceleration of a test particle is much smaller than a limiting acceleration @xmath0 , where @xmath0 is a universal constant .
thus , while the normal newtonian acceleration @xmath1 which a mass @xmath2 exerts on a test particle at distance @xmath3 is identical to the true test - particle acceleration @xmath4 for accelerations @xmath5 , in the mond limit ( i.e. , @xmath6 ) the implied newtonian acceleration is related to the true test - particle acceleration @xmath4 by @xmath7 .
the acceleration @xmath0 is a fundamental parameter in the mond theory . from rotation curve fitting to high - quality rotation curves ,
begeman et al . (
1991 ) determined a value of @xmath8 m s@xmath9 ( for @xmath10 km s@xmath11 mpc@xmath11 , which we adopt throughout this paper ) .
as described in paper ii , lsb galaxies provide a strong test of mond .
their low surface densities imply accelerations @xmath12 , which means that these galaxies should be almost completely in the mond regime .
milgrom ( 1983 , 1989 ) made a number of testable predictions on the shapes of rotation curves , and noted that low surface density galaxies should have slowly rising rotation curves .
this expectation of mond is confirmed by the observed rotation curves . in newtonian terms
this translates in these galaxies having large mass discrepancies ( mcgaugh & de blok 1998 , hereafter paper i ) .
this brings us to one of the more pronounced differences between mond and classical newtonian dynamics , which is the explanation of the tully - fisher ( tf ) relation . as is described in detail in paper i ( see also zwaan et al .
1995 ) , the fact that lsb galaxies are observed to obey the same tf relation as hsb galaxies implies a strong coupling between the central surface brightnesses of the disks of galaxies and their total mass - to - light ratios ( which include dark matter ) . assuming standard newtonian dynamics this implies that lsb galaxies have a higher total mass ( within the disk radius ) than hsb galaxies of the same asymptotic velocity .
it is hard to derive this result in the standard context without a lot of fine - tuning .
mond _ predicts _ that all galaxies should fall on one _ mass_-velocity relation , which takes the form @xmath13 , where @xmath14 is the asymptotic velocity and @xmath2 is the total mass of the galaxy ( that is , the mass of stars and gas ) .
once the value of @xmath0 is fixed , this relation becomes absolute and can be tested and falsified .
we use the rotation curves of 15 lsb galaxies to do a mond analysis .
section 2 describes the fitting procedure .
section 3 presents the results . in sect .
4 we discuss whether mond can fit any rotation curve , and we present our conclusions in sect .
in this paper we fit the rotation curves of the sample of 15 lsb galaxies presented in van der hulst et al .
( 1993 ) , de blok , mcgaugh & van der hulst ( 1996 ) and de blok & mcgaugh ( 1997 ) using the mond prescription .
we refer to these papers for a description of the properties of lsb galaxies and their rotation curves .
the rotation curves were fitted following the procedure outlined in begeman et al .
( 1991 ) and sanders ( 1996 ) . to be consistent with the results presented in sanders ( 1996 ) we have assumed that the stars and the gas are both in an infinitesimally thin disk ( for our purposes this assumption has no appreciable impact on the stellar rotation curve
see broeils 1992 ) .
the newtonian rotation curves of the visible components ( stars and hi ) were determined first .
the rotation curve of the stars was computed assuming that the mass - to - light ratio of the stars @xmath15 is constant with radius .
the rotation curve of the gas was computed using the radial hi surface density profile , after multiplying this by 1.4 to take the contribution of he into account .
we ignore any molecular gas : observations suggest that lsb galaxies contain only small amounts of molecular gas ( schombert et al . 1990 , knezek 1993 , de blok & van der hulst 1998 ) .
with the newtonian accelerations known , the mond acceleration can be computed ( see paper ii ) and the observed rotation curves fitted using a least - squares program .
the fitting procedure has three free parameters : the distance @xmath16 to the galaxy ; the mass - to - light ratio of the stellar disk @xmath15 ; and the value of the constant @xmath0 .
as @xmath0 is supposed to be a universal constant we fix it at the value determined in begeman et al .
as the lsb galaxies investigated here are sufficiently far away that the redshift is a fair indicator of the distance , we take the distance @xmath16 to be fixed at the values given in de blok et al .
( 1996 ) , assuming @xmath10 km s@xmath11 mpc@xmath11 .
this leaves @xmath15 as the _ only _ free parameter . for very gas
rich galaxies , there are effectively no free parameters , as the dynamics are completely dominated by the gas component , and the contribution of the stars can be ignored for any reasonable value of @xmath15 .
the rotation curves and mond fits are presented in fig . [ rotcurves ] and
table 1 . of the sample presented in de blok & mcgaugh ( 1997 )
we have not used f571 - 8 and f571-v2 .
f571 - 8 is an edge - on galaxy , and as the fits depend crucially on the photometric profile , the simple analysis done in de blok et al .
( 1996 ) does not suffice here ( due to uncertainties in the extinction ; see barnaby & thronson 1994 ) . for f571-v2
no photometry is available .
the results presented in table 1 differ slightly from those presented in paper ii .
this is because we now fit the entire shape of the rotation curve , rather than just the asymptotic velocity . as can be seen in fig .
[ rotcurves ] a good fit to the rotation curve was readily found for the majority of the galaxies .
these fits are shown in the two left - most columns in fig .
[ rotcurves ] .
the ease with which these one - parameter fits were obtained is remarkable , especially if one keeps in mind that in lsb galaxies the dynamics are often dominated by the gas rather than the stars .
this severely limits the possible values for @xmath15 and the freedom the mond recipe has to produce a good fit . for
6 galaxies a good fit did not occur trivially .
the corresponding `` bad '' fits are shown in the third column in fig .
[ rotcurves ] . were we to stop here , we would have to claim to have falsified mond in spite of many good fits and confirmed predictions .
it is therefore necessary to examine the uncertainties in the data before making such a radical step . for each galaxy
there are two parameters which might affect our ability to obtain a fit .
one is the distance to the galaxy @xmath16 and the second is the inclination of the galaxy @xmath17 ( @xmath0 is supposed to be a universal parameter , and varying it from galaxy to galaxy would be against the spirit of mond ) .
we have made two - parameter fits to the 6 `` bad '' curves with @xmath15 and @xmath16 free .
the results are shown in table 2 .
expect for ugc 5999 we find that a two - parameter fit prefers smaller distances which are typically @xmath18 per cent of the original distance .
as the galaxies are not near virgo and have a sufficiently large redshift that this is a fair indicator of the distance , we think it is unlikely that these fits show a realistic picture .
a closer study of the fits shows that the distance factor is almost entirely caused by the fit overestimating the rotation velocity in the outer parts ( cf . third column in fig .
[ rotcurves ] ) .
this is caused by the gas - rotation curve having relatively high velocities in the outer parts . in order to accomodate the rotation curve of the stellar disk ,
the fitting procedure tries to reduce the amplitude of the gas - rotation curve by reducing the gas mass , i.e. , it brings the galaxy in closer . with @xmath16 free reasonable values of @xmath15 between 0.5 and 3
are obtained .
however , one should keep in mind that the amplitude of the gas - rotation curve in the outer parts essentially forces us to adopt a particular value of @xmath16 , and consequently also forces us to adopt a value of @xmath15 .
therefore we can not use @xmath16 to improve the fits .
the inclinations of lsb galaxies are difficult to constrain from the shapes of their outer isophotes ( mcgaugh & bothun 1994 ; de blok et al .
1995 , 1996 ) . yet
this is essentially the only information we have which is independent of the kinematics . the total rotation curve and the rotation curves of the stars and the gas each depend differently on the inclination .
the total rotation velocity can be derived from the observed rotation velocity by @xmath19 .
the rotation velocity of the gas and stars needs to be derived from the true surface density , which in turn is determined from the observed surface density @xmath20 , so that @xmath21 .
the different dependencies on @xmath17 will change the ratios between the total rotation curve and the stellar and gas rotation curves .
changes of only a few degrees can have quite noticeable effects for relatively face - on galaxies .
if the inclination is under - estimated , the inclination - corrected observed velocities are too high .
the gas contribution , which dominates in the outer parts , has no scaling factor built in , and is therefore unable to fit these velocities .
this can be compensated for by using a larger value of @xmath15 , but this will result in a model rotation curve with the wrong shape , as it will overestimate the velocity in the inner parts by a large factor .
the effects of over - estimating the inclination are the reverse . the inclination - corrected observed velocities are too low , so that the rotation curve of the gas alone suffices to fit the data in the outer parts . consequently only a very modest or even zero or negative value of @xmath15 is needed .
this again results in a model curve with the wrong shape : too low velocities in the inner parts , and too high in the outer parts .
there is only a narrow range in inclination where both the inner and outer parts are fit equally well . in deriving the best value for the inclination we first changed the inclinations in a systematic way starting from the initial value from de blok et al .
( 1996 ) , and rederived the mass models for each value of the inclination .
we used the @xmath22 value to select a range of inclinations that yielded good fits .
this range was then narrowed by inspecting these fits by eye , and choosing those fits that most closely mimicked the behaviour of the high - resolution fits presented in sanders ( 1996 ) .
this means that the newtonian rotation curve of the stellar disk should be able to describe the innermost one - third of the rising part of the observed rotation curve . in practice the values for the inclination that were determined in that way
are constrained to typically within 4 degrees .
we now discuss on a case - by - case basis the validity of changing the inclination in these 6 galaxies .
the two right - most columns in fig .
[ rotcurves ] show the original and revised fits of these two galaxies .
figure [ pics ] shows slightly smoothed optical @xmath23-band pictures of the 6 galaxies with ellipses superimposed whose axis ratios correspond to the original assumed inclination and the derived mond inclination .
the mond inclination values are also given in table 1 .
[ [ f565-v2 ] ] f565-v2 at the original inclination of 60 degrees no fit with @xmath24 could be made .
the inclination needs to be changed to 44 degrees before positive @xmath15 values are produced by the fitting procedure .
the best fitting mond value is 31 degrees .
this might mean that we are missing a face - on lsb disk underlying the elongated structure we now observe as the galaxy .
finding such a lsb disk will be made difficult by the presence of a bright star just outside the region shown ( cf .
fig . 2 in de blok et al .
1996 ) .
[ [ f571-v1 ] ] f571-v1 the inclination had to be changed from 35 to 30 degrees , which is still fully consistent with the optical picture ( fig .
[ pics ] ) .
[ [ ugc-1230 ] ] ugc 1230 only a very minor change was needed . the original value derived in van der hulst et al .
( 1993 ) is 22 degrees .
mond needs a value of 17 degrees in order to produce a good fit .
note that the quality of the fit for this face - on system depends critically on the inclination .
using mond to fit the rotation curve of a face - on galaxy could actually tie down its inclination to within a few degrees .
this is impossible with a classical dark halo fit .
[ [ ugc-5005 ] ] ugc 5005 the original value for the inclination of this galaxy is 41 degrees .
mond demands an inclination in the range 26 to 30 degrees for the best fits .
here we adopt a value of 30 degrees .
[ pics ] shows that this value is still consistent with the observed axis ratio .
[ [ ugc-5750 ] ] ugc 5750 good solutions are only found for inclinations between of 38 and 42 degrees . the value of 64 degrees from van der hulst et al .
( 1993 ) is ruled out .
the new low value can only be consistent with the optical axis ratio if there is a lsb face - on disk underlying the observed bar - like structure .
this might be consistent with the hi column density map in van der hulst et al .
( 1993 ) where one can derive an inclination of @xmath25 degrees for the outermost column density contours .
deeper observations of this galaxy should be able to test this .
[ [ ugc-5999 ] ] ugc 5999 mond demands a value for the inclination of 22 degrees instead of 14 degrees . as shown in fig .
[ pics ] this results in only a very minor change in the axis ratios .
it is clear that in most cases the changes needed in the inclination to produce good fits are only minor .
the only two challenging objects are f565-v2 and ugc 5750 . to justify their inclinations one
needs in both cases a more face - on disk underlying the observed part of the galaxy . in both cases
, deeper observations should provide more definite determinations of the inclinations .
if these new values are still inconsistent with the mond values , then this may be a problem for the theory .
it is striking that the fraction of barred galaxies in figure 2 is much higher ( @xmath26 per cent ) than in the entire lsb sample ( @xmath27 per cent ) .
this may play some role in the inclination uncertainties .
also , note that the general preference is for slightly more face - on disks .
this is expected for systematic errors introduced by finite thickness disks or deviations from purely circular shapes .
in the rest of this paper we will adopt the new mond inclinations and the corresponding results , but we will distinguish them from the `` good '' fits by using different symbols . in practice the conclusions will not be affected .
-band tully - fisher relation for hsb and lsb galaxies .
the open symbols show the data from sanders ( 1996 ) .
the filled circles lsb galaxies where the inclination was not changed .
the filled triangles represent the lsb galaxies where @xmath17 had to be changed in order to get a good fit . ]
lrrrrrcccl name & @xmath28 & @xmath29 & @xmath30 & @xmath16 & @xmath31 & @xmath32 & @xmath33 & @xmath34 & @xmath35 $ ] + f563 - 1 & 111 & 0.135 & 23.6 & 45 & 0.385 & 2.97 & 3.99 & 5.82 & 25 + f563-v2 & 111 & 0.302 & 22.1 & 61 & 0.321 & 1.81 & ( 2.38 ) & 2.87 & 29 + f565-v2 & 51 & 0.023 & 24.7 & 48 & 0.084 & 2.18 & 2.70 & 5.85 & 60 [ 31 ] + f568 - 1 & 119 & 0.275 & 23.8 & 85 & 0.557 & 3.01 & 3.96 & 5.03 & 26 + f568 - 3 & 120 & 0.331 & 23.1 & 77 & 0.394 & 1.33 & 1.62 & 2.52 & 40 + f568-v1 & 124 & 0.224 & 23.3 & 80 & 0.343 & 2.96 & 3.80 & 4.49 & 40 + f571-v1 & 73 & 0.102 & 24.0 & 79 & 0.164 & 0.66 & 0.84 & 2.26 & 35 [ 30 ] + f574 - 1 & 100 & 0.372 & 23.3 & 96 & 0.485 & ( 0.54 ) & 0.71 & 1.85 & 65 + f583 - 1 & 85 & 0.063 & 24.1 & 32 & 0.243 & 1.68 & 1.96 & 5.53 & 63 + f583 - 4 & 67 & 0.093 & 23.8 & 49 & 0.077 & ( 0.24 ) & 0.31 & 1.07 & 55 + u128 & 131 & 0.513 & 23.2 & 60 & 0.882 & 0.84 & 1.06 & 2.56 & 55 + u1230 & 102 & 0.324 & 23.3 & 51 & 0.812 & 1.16 & 1.71 & 3.67 & 22 [ 17 ] + u5005 & 99 & 0.204 & 23.8 & 52 & 0.406 & ( 3.65 ) & 4.80 & 5.64 & 41 [ 30 ] + u5750 & 75 & 0.468 & 23.5 & 56 & 0.140 & ( 0.68 ) & 0.89 & 0.98 & 64 [ 39 ] + u5999 & 155 & 0.170 & 23.5 & 45 & 0.252 & ( 0.53 ) & 0.70 & 2.01 & 14 [ 22 ] +
the main result is that mond is able to explain the slowly - rising , non - flat rotation curves of galaxies with low surface brightnesses , where a large mass discrepancy is inferred in the classical , newtonian case .
the resulting stellar and total mass - to - light ratios enable us to derive some additional results which we will discuss below . in this discussion
, we will also make use of the results of the mond fits to the rotation curves of a large sample of hsb galaxies , presented in sanders ( 1996 ) .
mond predicts a unique and exact _
mass_-velocity relation @xmath36 , where @xmath37 denotes the total mass , that is , @xmath38 . converting this to more directly observable quantities yields @xmath39 the exact @xmath40 relation thus translates into an identical @xmath41 relation _ if _
@xmath42 is constant .
any scatter in @xmath43 will show up as scatter in the @xmath41 relation ( that is , as scatter in the tf relation ) .
mond thus implies that there is a unique tf relation , where any scatter must is caused by scatter in @xmath42 .
the small observed scatter in the tf relation thus implies there is only a small scatter in @xmath42 . as discussed extensively in papers i and ii , in the classical dark - halo tf relation the only way to keep galaxies on the tf relation
is by having a fine - tuned relation between the optical surface brightness of a galaxy and its total mass - to - light ratio ( i.e. , including the dark matter zwaan et al .
in mond all that is needed is that the scatter in @xmath42 is small enough to be consistent with the observed scatter in the tf relation . in fig .
[ tf ] we show the @xmath44-band mond tf relation for hsb galaxies as derived by sanders ( 1996 ) , with the lsb data superimposed .
it is clear that there is no systematic offset between the hsb and the lsb galaxies .
we now study the implied small scatter in @xmath45 a bit more closely .
lrrr galaxy & @xmath16 & @xmath46 & @xmath47 + f565-v2 & 48 & @xmath48 & @xmath49 + f571-v1 & 79 & @xmath50 & @xmath51 + u1230 & 51 & @xmath52 & @xmath53 + u5005 & 52 & @xmath54 & @xmath55 + u5750 & 56 & @xmath56 & @xmath57 + u5999 & 45 & @xmath58 & @xmath59 + figure [ ml ] shows the stellar mass - to - light ratio @xmath15 and the total mass - to - light ratio @xmath42 as a function of maximum rotation velocity @xmath28 , central surface brightness @xmath60 and color @xmath61 .
@xmath15 overall is clearly increasing with @xmath28 , consistent with the color - magnitude relation .
the lsb galaxies are indistinguishable from the other galaxies ; a situation which is in sharp contrast with the maximum disk results ( de blok & mcgaugh 1997 ) .
there is no clear trend with central surface brightness , although the scatter increases towards lower surface brightnesses , which probably just reflects the large importance of the gas in these galaxies and the large effects which small amounts of star formation can have on the blue luminosity of lsb galaxies .
this increase in scatter in @xmath15 is also apparent at the bluest @xmath61 colors .
the values of @xmath15 are reasonable and vary between @xmath62 and @xmath63 ( see fig .
[ mlhist ] ) .
the distribution peaks at @xmath64 .
the median of the hsb data equals 1.4 , while that of the lsb data equals 1.3 .
most fall within the range of @xmath15 predicted by population synthesis models ( see sanders 1996 ) .
no negative values of @xmath15 were found .
mond therefore does not over - correct the mass - discrepancy .
measurements in e.g. the near - infrared should yield in a much smaller range in @xmath15 values , as tentatively indicated by sanders ( 1996 ) and sanders & verheijen ( 1998 ) . as found from fitting of rotation curves .
the dotted histogram indicated the lsb data superimposed on the hsb data ( full line).the range and shape of both distributions is similar .
both peak between @xmath65 value of 0 and 1 .
the median of the hsb distribution is 1.4 , that of the lsb distribution is 1.3 .
both distributions are consistent with values normally derived from population synthesis models.,scaledwidth=45.0% ] the relevant parameter for the tully - fisher relation is however the _ total _ mass - to - light ratio , defined as @xmath66 .
this quantity is also shown in fig .
a comparison with the @xmath15 data points shows that inclusion of the gas mass has made the increasing trend with @xmath28 disappear .
@xmath42 does not show any systematic trend with @xmath28 , @xmath61 and @xmath60 .
the gas mass thus makes a large contribution for low luminosity and low surface brightness galaxies .
the average value of @xmath42 is 2.5 .
the dispersion in @xmath67 is 0.30 dex , which compares to a dispersion in @xmath68 in the blue tf relation of 0.33 dex .
the scatter in the tf relation is thus entirely consistent with , and can be explained by , the scatter in the values of @xmath42 .
the lsb galaxies in all respects confirm the predictions made by milgrom ( 1983 ) .
these galaxies are deep in the mond regime , with the lowest accelerations currently known .
they provide the clearest and strongest test of mond made to date .
one of the myths surrounding mond is that it is specifically designed to fit rotation curves and is therefore guaranteed to do so .
reality is a bit more complicated than this .
mond _ predicts _ what the rotation curve will be on the basis of solely the _ observed _ matter distribution .
this is a fundamental difference from the dark matter theory where a dark halo is fit to whatever is left after the subtraction of the luminous component .
the dark matter theory does not predict rotation curves , whereas in mond there is a unique and unambiguous correspondence between the observed matter distribution and the rotation curve .
this results in mond having much less freedom in getting the `` right '' answer than the dark matter rotation curve fits .
this is already clear from fitting rotation curves of hsb galaxies .
these show that the inner parts of hsb galaxies are still in the newtonian regime , and that therefore a newtonian ( i.e. , `` classical '' ) fit in these parts fixes @xmath15 .
once @xmath15 is fixed , there is no more freedom in the fit .
if the rotation curve in the outer parts does not exactly behave in the way mond wants it to , then no good fits can be made .
we illustrate all of this by the following example . the hsb / lsb pair ngc2403/ugc 128
are two galaxies that occupy identical positions on the tully fisher relation , with indistinguishable luminosities and flat rotation amplitudes , but with very different central surface brightnesses .
this is extensively described in de blok & mcgaugh ( 1996 ) . in the classical , dark matter picture ,
the fact that these galaxies are on the same position on the tf relation means their mass distributions must be related .
whether this means that their halos are identical or not is unclear . we will now construct a hybrid model galaxy , by taking the rotation curve of ngc 2403 , but where we will replace the optical photometry of ngc 2403 by that of ugc 128 .
this results essentially in a low surface brightness version of ngc 2403 .
the similarity of the two galaxies maximizes the opportunity for both mond and dark matter theories to achieve a fit to the hypothetical hybrid .
we have fitted this hypothetical galaxy with both the mond and dark matter fitting procedures .
we present the results in fig .
[ hybrid ] .
it is clear that in the mond case no good fit can be made .
the observed mass distribution is not related to the rotation curve , hence the bad fit .
the implied stellar mass - to - light ratio is not reasonable , and both galaxies have rather high , well determined inclinations so that adjusting the inclination within the errors will not have an impact on the procedure . in the dark matter case
one _ can _ reach a very tolerable fit , which is comparable in quality with the fit of the real ngc 2403 ( see begeman , broeils and sanders 1991 , their fig . 1 ) .
the conclusion we can reach from this is that mond can _ not _ fit any arbitrary rotation curve , dispelling the myth which asserts it was designed to do just this .
mond was designed to produce an asymptotically flat rotation curve , but this is no guarantee that the application of the mond force law to the observed luminous mass distribution will obtain the correct shape for just any rotation curve .
there must be a very close relation between the observed mass distribution and the rotation curve ( both of which can be determined independently ) , and this relation must be the mond recipe if one wants to derive the latter solely from the former .
the good fits to lsb galaxies rotation curves support mond , especially so as these are galaxies with large mass discrepancies in their observed disks , where the mond effects are strongest .
the mond fits furthermore _
test _ the mond theory .
this is not the case with dark matter fits , where the properties of the dark matter discrepancy are _ derived _ from the fits .
these fits _ define _ the properties of the mass discrepancy . it is therefore not possible to falsify nor confirm the dark matter hypothesis from rotation curves alone , as can be done with the mond hypothesis .
we show this by fitting a hypothetical non - physical galaxy : the mond theory is unable to produce a good fit ( which in this case is a good thing ) , whereas we can get a good fit with the dark matter theory ( showing it is more flexible in fitting non - physical models ) . in principle , there are an infinite number of things the rotation curves of galaxies could do given the presence of an invisible mass component .
there are many plausible predictions for what they are expected to do given various ideas about dark matter . in the case of mond
, there is one and only one thing rotation curves can do .
this is precisely what they do .
one last remarkable result is the great efficiency with which mond can describe the rotation curves of galaxies of very different types .
even if mond is ultimately falsified , it would still be worth knowing why it works so well . until a mond cosmology can be derived ( see sanders 1998 )
it remains a recipe for describing rotation curves , but as a recipe it is far superior to the dark halo recipe , and this raises the question of _ why _ mond works so well .
mond is after all only a simple analytic formula which gets @xmath69 correct based solely on the luminous mass .
if the mond phenomenology arises as the result of dark matter , then this would imply that the dark matter must `` know '' about both the distribution of light _ and _ the mond formula , and arrange itself appropriately , right down to amplifying the bumps and wiggles in the stellar and gas rotation curves by the appropriate amount .
barnaby , d. , thronson , h.a .
, 1994 , , 107 , 1717 begeman , k.g .
, broeils , a.h . , sanders , r.h .
1991 , , 240 , 523 bekenstein , j. , milgrom , m. , , 1984 , 286 , 7 broeils , a.h . , 1992 , phd thesis , university of groningen de blok , w.j.g . ,
mcgaugh , s.s .
1996 , , 469 , l89 de blok , w.j.g . , mcgaugh , s.s .
1997 , , 290 , 533 de blok , w.j.g . , van der hulst j.m . ,
bothun , g.d .
, 1995 , , 274 , 235 de blok , w.j.g . , van der hulst j.m .
, 1998 , , in press de blok , w.j.g . , mcgaugh , s.s , & van der hulst , j.m .
1996 , , 283 , 18 knezek , p.m. 1993 , phd thesis , university of massachusetts mcgaugh , s.s .
, bothun , g.d . , 1994 , , 107 , 530 mcgaugh , s.s . , de blok , w.j.g . , 1998a , , in press ( paper i
) mcgaugh , s.s . , de blok , w.j.g . , 1998b ,
, in press ( paper ii ) milgrom , m. , 1983 , , 270 , 365 milgrom , m. , 1989 , , 338 , 121 milgrom , m. , 1994 , ann .
phys . , 229 , 384 milgrom , m. , 1995 , , 455 , 439 sanders , r.h .
1990 , , 2 , 1 sanders , r.h .
1996 , , 473 , 117 sanders , r.h .
1998 , , in press sanders , r.h . ,
verheijen , m.a.w . , 1998 , , in press schombert j.m . ,
bothun g.d .
, impey c.d . ,
mundy l.g . , 1990 , , 100,1523 van der hulst j.m . , skillman e.d . ,
smith t.r .
, bothun g.d . ,
mcgaugh s.s .
, de blok w.j.g . , 1993 , , 106 , 548 zwaan m.a . , van der hulst j.m
, de blok w.j.g . ,
mcgaugh s.s . , 1995 , , 273 , l35 | we present mond fits to 15 rotation curves of lsb galaxies .
good fits are readily found , although for a few galaxies minor adjustments to the inclination are needed .
reasonable values for the stellar mass - to - light ratios are found , as well as an approximately constant value for the total ( gas and stars ) mass - to - light ratio .
we show that the lsb galaxies investigated here lie on the one , unique tully - fisher relation , as predicted by mond .
the scatter on the tully - fisher relation can be completely explained by the observed scatter in the total mass - to - light ratio .
we address the question of whether mond can fit any arbitrary rotation curve by constructing a plausible fake model galaxy .
while mond is unable to fit this hypothetical galaxy , a normal dark halo fit is readily found , showing that dark matter fits are much less selective in producing fits .
the good fits to rotation curves of lsb galaxies support mond , especially as these are galaxies with large mass discrepancies deep in the mond regime . |
iib matrix model @xcite is a candidate of the constructive definition of superstring theory .
it is defined by the action @xmath8[a_\mu , a_\nu ] + \frac{1}{2}\bar{\psi}\gamma^\mu[a_\mu,\psi ] \right ) , \label{eq : action}\end{aligned}\ ] ] where @xmath9 and @xmath10 are @xmath11 hermitian matrices .
@xmath9 is a ten - dimensional vector and @xmath10 is a ten - dimensional majorana - weyl fermion .
one of the important aspects of this model is the dynamical generation of the spacetime and the gauge group . in this model
the eigenvalue distribution represents spacetime geometry . from this point of view , if a four - dimensional eigenvalue distribution has smaller free energy than the other configurations , the spacetime will be compactified to four - dimensions . on the other hand
the stability of @xmath2 coincident d - branes may indicate the dynamical generation of @xmath12 gauge group .
so far , various attempts on this have been made .
several such attempts using branched polymer @xcite , complex phase effect @xcite and especially improved mean - field approximation @xcite were successful . in these studies ,
four - dimensional configurations were found to have smaller free energy .
another interesting approach is to compare the free energies on noncommutative backgrounds . in string theory ,
noncommutative gauge theories on flat or curved backgrounds are realized with constant or nonconstant @xmath13 @xcite .
iib matrix model on flat fuzzy space backgrounds can be mapped to noncommutative super yang - mills theory ( ncsym ) @xcite , and the system can be analyzed by perturbative calculations .
noncommutative gauge theories on curved manifolds have recently been studied extensively by deforming iib matrix model @xcite .
however , flat fuzzy space backgrounds can not be realized as classical solutions at finite @xmath14 .
therefore , in order to see the @xmath14-dependences of the free energies , we must study compact noncommutative spaces .
fuzzy @xmath15 , @xmath16 and @xmath17 backgrounds have already been studied and @xmath16 has been shown to be stable under certain deformations @xcite . in this paper , we study the free energies of iib matrix model on fuzzy torus backgrounds . the relation to flat fuzzy space is more transparent in the case of the fuzzy torus than in the case of fuzzy sphere .
we expand the action of iib matrix model around a fuzzy torus background and sum up all 1pi diagrams .
then , we search the minimum of the free energy . especially , we compare @xmath0 and @xmath1 backgrounds . in principle , in order to find the true vacuum , we must calculate free energies on all possible backgrounds and find the minimum . here , however , we consider only fuzzy torus .
although fuzzy torus is not a classical solution , however , as we will show later , fuzzy torus locally looks like flat fuzzy space , which is a classical solution .
therefore , we expect our calculation captures some qualitative feature of flat fuzzy spaces . if the true vacuum is similar to a flat fuzzy space , we may find some informations such as dimensionality of the spacetime on it .
in present study we will see that @xmath0 has smaller free energy than @xmath1 .
this implies that if the true vacuum is four - dimensional , it is not a simple noncommutative space considered here .
although both fuzzy sphere and fuzzy torus break supersymmetry , the breaking effect is soft and the loop corrections to their free energies almost vanish .
this is because their local structures are similar to flat fuzzy spaces . in the case of fuzzy @xmath15 , @xmath16 and @xmath17 , such cancellations
have been confirmed to 2-loop level @xcite . for fuzzy torus
we can see such cancellation at any loop level .
furthermore , in this case the calculations become simpler and we can study larger parameter region .
for example , on @xmath1 , we vary the ratio @xmath18 of the noncommutativity parameters of two tori and their ratio @xmath19 of the matrix size . , @xmath20 was studied @xcite . ]
we can also see the effect of the global topologies on the free energies by comparing the results for fuzzy tori and fuzzy spheres .
on the other hand , fuzzy torus has a few drawbacks .
we embed fuzzy @xmath0 into @xmath21 , while fuzzy @xmath15 can be embedded into @xmath22 .
therefore , in ten dimensions we can realize only two- and four - dimensional configurations using fuzzy tori ( @xmath0 and @xmath1 ) , while a six - dimensional configuration can be constructed using fuzzy spheres ( @xmath17 ) . in @xcite
it is shown that @xmath16 has a smaller free energy than @xmath23 . in our case , such comparison is impossible .
we also have subtlety in the perturbative calculation , due to the presence of a few tachyons on fuzzy torus .
we will discuss this problem in detail later .
the organization of the paper is as follows . in section 2
, we will construct the fuzzy torus backgrounds and show the correspondence between fuzzy torus and flat fuzzy space . in section 3
, we calculate the free energy up to two loop level analytically , and in section 4 , we compare the free energies numerically . in appendices ,
we show the detail of the calculations .
although we are interested in @xmath1 , we will mostly show the feynman diagrams and calculations of @xmath0 as @xmath1 calculations are very lengthy .
in order to perform a perturbative calculation , we decompose @xmath9 and @xmath10 into a background @xmath24 and the fluctuation @xmath25 around it : @xmath26 adding a gauge fixing and the corresponding ghost term @xmath27 ^ 2+[p_\mu , b][a^\mu , c ] \right ) \end{aligned}\ ] ] to ( [ eq : action ] ) , we obtain the full action @xmath28 + \frac{1}{2}{\bar{\varphi}}\gamma^\mu[a_\mu,{\varphi } ] + ( { { \cal p}}_\mu b)[a^\mu , c ] \right\}\nonumber\\ & & -\frac{1}{4g^2}\ { { \rm tr}}[a_\mu , a_\nu]^2 , \end{aligned}\ ] ] where @xmath29,\ f_{\mu\nu}\equiv i[p_\mu , p_\nu]$ ] and @xmath30 $ ] . if the background is not a classical solution , e.g. fuzzy sphere or fuzzy torus , the @xmath31 term does not vanish .
first , we introduce the @xmath11 clock and shift matrices @xmath32 and @xmath33 : @xmath34 where @xmath32 and @xmath33 satisfy @xmath35 using them , we construct a 2-dimensional fuzzy torus as follows : @xmath36 here , @xmath37 and @xmath38 are the noncommutativity parameters , which have the dimension of mass .
their product , @xmath39 , corresponds to the noncommutativity parameter @xmath40 of fuzzy plane .
the relation of this background to fuzzy plane is discussed more precisely in section [ sec : relation_of_torus_to_d_brane ] .
we can expand the fluctuations in terms of @xmath32 and @xmath33 .
for example , @xmath41 can be expressed as @xmath42 and the hermiticity condition is given by @xmath43 we will show that this expansion corresponds to the ordinary fourier expansion in section [ sec : relation_of_torus_to_d_brane ] . using these fourier modes , we can construct the feynman rules , which we summarize in appendix [ sec : feynman_rule ] .
we can similarly construct fuzzy @xmath1 using the tensor products of the fuzzy tori constructed above : @xmath44 here , @xmath45 , and @xmath46 are the @xmath47 clock and shift matrices . for simplicity , we set the radii of the two circles in the first @xmath0 to the same value @xmath48 , and similarly in the second @xmath0 to @xmath49 . in the following sections , we first fix the value of @xmath14 , and calculate the free energy as a function of @xmath50 , @xmath51 and @xmath52 , and
minimize it with respect to them .
we also consider a system of @xmath2-coincident fuzzy tori . to study @xmath2-coincident fuzzy @xmath0 ,
we replace @xmath53 and @xmath54 with @xmath55 and @xmath56 , where @xmath53 and @xmath54 stand for the @xmath11 clock and shift matrices , respectively .
then , the fourier modes @xmath57 and @xmath58 become @xmath59 matrices .
similarly , in the case of fuzzy @xmath1 , we can replace @xmath60 with @xmath61 , and so on , to construct a @xmath2-coincident @xmath1 .
in this case , we have @xmath62 . in this section ,
we discuss the relation between the fuzzy torus and flat fuzzy space @xcite .
to be specific , we consider the case of fuzzy @xmath0 and flat fuzzy plane ( flat d1-brane ) .
generalization to higher dimensions is straightforward . in order to avoid confusion
, we denote fuzzy plane background as @xmath63 and that of fuzzy torus as @xmath24 .
the fuzzy plane is characterized by the constant commutators : @xmath64=-ib , \quad q_3=\cdots = q_{10}=0 . \end{aligned}\ ] ] let us write @xmath32 and @xmath33 as @xmath65 then , if we consider the region where eigenvalues of @xmath66 and @xmath67 are close to zero , the condition @xmath68 is equivalent to @xmath69=-ib$ ] . therefore , @xmath70 in this region . on the other hand , if we compare ( [ t2 ] ) with @xmath71 we have the identifications in the embedded space as @xmath72 . in general , tangent spaces of fuzzy torus look like fuzzy planes
. iib matrix model on fuzzy torus can be mapped to ncsym as in the case of flat fuzzy space @xcite .
if we introduce the spacetime coordinates as @xmath73 we have @xmath74 where momenta @xmath75 are defined by @xmath76 .
hence the expansion([eq : fourier ] ) becomes the ordinary fourier expansion in the commutative limit .
we calculate the free energies in this section . we ignore the tachyons and zero modes on the fuzzy torus in our calculations .
the vector- and fermion - kinetic terms on fuzzy torus have tachyons and extra zero - modes besides those coming from the u(1)-part .
( see appendix [ sec : tachyon ] for the detail . )
therefore , it is not clear whether perturbative calculation is possible or not .
however , the number of tachyons remains the same in the large-@xmath14 limit , and their mass squared is of @xmath77 and vanishes in the large-@xmath14 limit .
hence we expect that the tachyons give a relative correction of @xmath78 to the free energy . on the other hand , in the case of @xmath1 and @xmath79
, we will see that the leading contributions cancel out due to the supersymmetry , but the terms of @xmath77 do not vanish .
they are expected to be larger than the effect of the tachyons , and we can take the large-@xmath14 limit without paying much attention to the tachyons .
we start with the tree - level free energy .
it is given by @xmath80 ^ 2 & = & \frac{16l^2n^3}{g^2}\sin^2\left(\frac{\pi}{n}\right ) \nonumber\\ & \to & \frac{16n\pi^2l^2}{g^2 } \quad \mbox{(in the large-$n$ limit)}\ ] ] for @xmath0 , and @xmath80 ^ 2 & = & \frac{16l^2n^3}{g^2}\sin^2\left(\frac{\pi}{n}\right ) \left(r^2+\frac{1}{r^2}\right ) \nonumber\\ & \to & \frac{16n\pi^2l^2}{g^2 } \left(r^2+\frac{1}{r^2}\right ) \quad \mbox{(in the large-$n$ limit ) } \nonumber\\\end{aligned}\ ] ] for @xmath1 , respectively . the 1-loop free energy can be easily evaluated as @xcite @xmath81 the terms for @xmath82 cancel out , and the first nonzero contribution comes from @xmath83 . in the case of @xmath1 and @xmath79 ,
it is @xmath84 and can be neglected compared with the tree - level and 2-loop level contributions . in the case of @xmath0 , truncation at
some order of @xmath85 is not a good approximation because of infrared divergence . instead
, we can numerically evaluate the 1-loop free energy without expanding in @xmath85 .
we calculate all the eigenvalues of the kinetic terms , and sum up their logarithms .
we find that the 1-loop free energy on @xmath0 background is proportional to @xmath86 and hence negligible compared to the tree level and 2-loop contributions .
next we evaluate the 2-loop free energy on the fuzzy @xmath1 .
since it is difficult to obtain analytic forms of the vector and fermion propagators , we expand them in powers of @xmath85 as shown in appendix [ sec : propagators ] , and calculate the free energy order by order in @xmath85 .
a naive power counting shows that each power of @xmath85 is associated with @xmath87 and @xmath88 for @xmath0 and @xmath1 , respectively , if there is no infrared divergence .
in fact , this expansion is good on the @xmath1 background , while it is not on the @xmath0 background .
we discuss the validity of this expansion in appendix [ sec : f - expansion ] . here
we show the result for the @xmath1 background .
we have four kinds of 1pi 2-loop vacuum diagrams ( see fig.[fig:2loop_diagrams ] ) , and their contributions can be approximated as ( see appendix [ twoener ] for the detail ) @xmath90 where @xmath91 , and @xmath92 .
+ @xmath93 are analogous to the ordinary momentum variables in the continuum limit and given in appendix [ twoener ] . summing up these approximated values ( [ eq:2loop_v3fs])@xmath94([eq:2loop_v4bs ] ) , we have @xmath95 this cancellation is not exact , but we can see numerically that the exact contribution of @xmath89 terms is smaller than that of @xmath96 terms .
the @xmath97 contribution is zero , because integrands are odd function of the momenta .
the calculation of @xmath96 contribution is very tedious .
here we show only the simplest case , that is , the planar part of fig.[fig:2loop_diagrams ] ( b ) . in this case
, the contribution can be writen as @xmath98 where @xmath99 here @xmath100 , and @xmath101 and @xmath102 vary from @xmath103 to @xmath104 . because the ghost propagators do not contain @xmath85 , this expression is rather simple .
the contributions from the other diagrams ( fig.1 ( a),(b),(c ) ) also can be written as @xmath105 , @xmath106 and @xmath107 .
but the expressions of @xmath108 and @xmath109 become much more complicated , because both vector and fermion propagators contain @xmath110 corrections . summing up all of the diagrams , we find that the leading behavior of the @xmath96 contribution can be expressed as @xmath111 the function @xmath112 is dimensionless and has a symmetry under the exchange of two tori @xmath113 the total free energy on the fuzzy @xmath1 background can be written as @xmath114 where we have neglected subleading terms in @xmath14 .
the @xmath115 terms come from higher - loop diagrams .
if we neglect them and minimize the free energy with respect to @xmath116 , we obtain the minimum value @xmath117 at @xmath118 the situation does not change qualitatively , even if we take the full order of the loop expansion into account . in general , the @xmath119-loop diagrams are proportional to @xmath120 . however , the @xmath89 contributions cancel due to the supersymmetry for the detail . ] , and the @xmath97 contributions vanish because they are given by integrals of odd functions . therefore , if we include the contribution from the higher - loop diagrams , the free energy can be expressed as @xmath121 where @xmath122 are of @xmath123 .
the minimum value is again of @xmath124 .
here we consider the system of @xmath2-coincident fuzzy tori .
this system can be mapped to u(@xmath2 ) ncsym .
the free energy in this case becomes @xmath125 therefore , the single fuzzy torus has the lowest free energy , and @xmath126 gauge group is favored in this case .
as the analytic expressions are very complicated , here we try to compare the free energies numerically .
first , we consider @xmath127 . it is given by an 8-fold integral , and we have evaluated it approximately using the quadrature by parts .
we divide the integration domain into @xmath128 bins in each direction , with @xmath129 .
let @xmath130 be the value of @xmath127 evaluated by @xmath128 bins .
we have fitted the plot of @xmath130 versus @xmath131 by a polynomial , and estimated @xmath132 extrapolating it up to @xmath133 . in fig.[fig :
fitting ] , we plot @xmath130 against @xmath131 for @xmath134 and @xmath135 .
the curve shows a nice polynomial fitting .
we can evaluate the @xmath132 replacing @xmath136 in the polynomial .
once we have @xmath127 , we can readily obtain @xmath137 from eqn .
( [ totenergy ] ) .
we plot @xmath138 as a function of @xmath18 for @xmath19 fixed to 1 in fig.[fig : r=1 ] .
we see that @xmath139 is the minimum .
therefore , the fuzzy @xmath1 is favored under the variation of the ratio @xmath18 of the noncommutativity .
fig.[fig : r=1 ] shows @xmath140 as a function of @xmath19 for @xmath18 fixed to 1 , and this time @xmath139 is the maximum .
therefore , we find that the fuzzy @xmath1 is _ unstable _ under the variation of the ratio @xmath19 of the matrix sizes . to compare with the result for the fuzzy @xmath16 in @xcite , we plot @xmath141 as a function of @xmath18 in fig.[fig : r = r^2 ] .
the curve @xmath142 is the same as the one studied for the fuzzy @xmath16 in @xcite . on this curve
, @xmath139 is the minimum , hence the fuzzy @xmath1 is favored .
this observation is similar to the fuzzy @xmath16 . in @xcite , it is pointed out that although the 2-loop contribution ( @xmath143 in our notation ) is maximum at @xmath139 , the tree - level contribution , @xmath144 , is minimum at @xmath139 and this makes the total free energy minimum at @xmath139 .
we have exactly the same situation here .
these results show that although the fuzzy @xmath1 is favourable in some directions , it is unstable under the variation of @xmath19 . because the limit @xmath145 corresponds to the fuzzy @xmath0 ,
we conclude that the fuzzy @xmath0 is more favourable than the fuzzy @xmath1 .
since our observations along @xmath142 is completely analogous to that of @xcite , it is plausible that the fuzzy @xmath16 also has the same kind of instability . a similar phenomenon is observed in a 6 dimensional bosonic matrix model with chern simons term , where @xmath15 is indeed shown to be more favourable than @xmath146 @xcite .
we can compare the free energy of the fuzzy @xmath1 and the fuzzy @xmath16 . at @xmath139 ,
the free energy of @xmath1 is @xmath147 . according to @xcite ,
the free energy of @xmath16 with the same matrix size and noncommutativity is @xmath148 .
therefore , @xmath16 is more favorable than @xmath1 .
we have also computed the 2-loop free energy of the fuzzy @xmath0 up to @xmath96 as a function of the ratio of radii , @xmath149 .
we found that it has the minimum at @xmath150 and is proportional to @xmath14 .
however , because the @xmath85 expansion is not good in this background , this result is not reliable . the free energy on the @xmath15 background is calculated in @xcite .
in this case , the tree - level action dominates and the free energy is @xmath103 at @xmath151 .
in this paper , we have calculated the free energies of iib matrix model on the fuzzy @xmath1 backgrounds up to 2-loop level .
the free energy on @xmath1 is a function of the ratio @xmath18 of the noncommutativity and @xmath19 of the matrix size . at @xmath139 ,
the free energy takes minimum when @xmath18 is varied , but maximum when @xmath19 is varied .
this indicates that @xmath0 is more stable than @xmath1 .
it is plausible that @xmath16 has the same sort of instability .
if so , a four - dimensional vacuum might not be realized as a direct product of two noncommutative surfaces .
commutative backgrounds seem to be more plausible .
another possibility is that new terms like myers term arise dynamically ( for example via large-@xmath14 rg ) and stabilize noncommutative spaces .
we also have compared the free energies of @xmath16 and @xmath1 , and found that the latter is about 10 times as large as the former .
it implies that iib matrix model favors configurations with higher symmetry .
furthermore , we calculated the free energy of the system of @xmath2-coincident fuzzy @xmath1 , and found that the @xmath152 system has the smallest free energy . in terms of ncsym , this means that u(1 ) gauge group is favored .
this result is the same as that of the fuzzy sphere in iib matrix model @xcite and in bosonic matrix models with chern simons term @xcite .
it seems that in general noncommutative backgrounds favor u(1 ) gauge group .
acknowledgments the authors would like to thank y.kimura , y.matsuo and a.miwa for fruitful discussions .
s.b . was supported in part by the grants - in - aid for scientific research ( no .
p02040 ) the ministry of education , culture , sports , science and technology of japan . m.h . would like to thank the japan society for the promotion of science for financial support .
this work was also supported in part by a grant - in - aid for the 21st century coe `` center for diversity and universality in physics '' .
in this section , we summarize the feynman rules of iib matrix model on the @xmath0 background .
generalization to the @xmath1 background is straightforward .
* boson ( vector and scalar ) * + + the kinetic term for @xmath153 is given by @xmath154 and the propagator is expressed as @xmath155 we expand it with respect to @xmath85 and truncate it at some order : @xmath156 note that this expansion is not good in this case but a similar expansion in the @xmath1 background is reliable .
+ + * fermion * + + the fermion kinetic term is given by @xmath157 and the propagator is expressed as @xmath158 here we also expand @xmath159 in powers of @xmath85 : @xmath160 + + * ghost * + the ghost kinetic term and propagator are given , respectively , by @xmath161 and @xmath162 a direct calculation shows that @xmath164 & = & -2i\sqrt{n}l_1\sum_{m}u^{m_2}v^{-m_1}\sin\left(\frac{\pi m_1}{n}\right ) \left\ { { \tilde{a}}_\mu(m_1,m_2 - 1)e^{\pi i m_1/n } - { \tilde{a}}_\mu(m_1,m_2 + 1)e^{-\pi i m_1/n } \right\ } , \nonumber\\ \ [ p_2,a_\mu ] & = & -2\sqrt{n}l_1\sum_{m}u^{m_2}v^{-m_1}\sin\left(\frac{\pi m_1}{n}\right ) \left\ { { \tilde{a}}_\mu(m_1,m_2 - 1)e^{\pi i m_1/n } + { \tilde{a}}_\mu(m_1,m_2 + 1)e^{-\pi i m_1/n } \right\ } , \nonumber\\ \ [ p_3,a_\mu ] & = & -2i\sqrt{n}l_2\sum_{m}u^{m_2}v^{-m_1}\sin\left(\frac{\pi m_2}{n}\right ) \left\ { { \tilde{a}}_\mu(m_1 + 1,m_2)e^{-\pi i m_2/n } - { \tilde{a}}_\mu(m_1 - 1,m_2)e^{\pi i m_2/n } \right\ } , \nonumber\\ \ [ p_4,a_\mu ] & = & -2\sqrt{n}l_2\sum_{m}u^{m_2}v^{-m_1}\sin\left(\frac{\pi m_2}{n}\right ) \left\ { { \tilde{a}}_\mu(m_1 + 1,m_2)e^{-\pi i m_2/n } + { \tilde{a}}_\mu(m_1 - 1,m_2)e^{\pi i m_2/n } \right\}. \nonumber\\\end{aligned}\ ] ] from these equations , we obtain @xmath165 and so on .
this expression is not very simple , but @xmath166 can be written in the following simple form : @xmath167 there are four types of interaction vertices ( fig.[fig : vertices ] ) . here ,
the solid , dotted and wavy lines represent the fermions , ghosts and bosons , respectively . + + * ( a ) fermion - fermion - boson coupling @xmath168 * @xmath169 note that @xmath170 ( both vectors and scalars couple to the fermion ) .
+ + * ( b ) ghost - ghost - vector coupling @xmath171 * @xmath172 where @xmath173 note that only vectors couple to the ghosts and scalars do not .
+ + * ( c ) boson - boson - vector coupling @xmath174 * @xmath175 + * ( d ) four - boson coupling @xmath176 * @xmath177
first , we consider the cancellation of the free energy on flat fuzzy space backgrounds . on this background , the free energy should be zero at least formally .
( the meaning of the word formally " is explained shortly . ) for example , the sum of the two - loop diagrams is proportional to @xcite @xmath178 naively this is zero , because @xmath179 .
however , if the size @xmath14 of the matrices is finite , there should be some momentum cutoff , and ( [ eq:2loop_d_brane ] ) is not necessarily zero .
for example , in the region where @xmath2 and @xmath180 are allowed but @xmath181 is outside the cutoff , the second term of ( [ eq:2loop_d_brane ] ) does not make sense .
however , such problem occurs only in the large - momentum region . on the other hand in the small - momentum region , the free energy vanishes due to the supersymmetry .
therefore , it is natural to expect that the free energy is formally " zero at any order of the loop expansion , if we use @xmath179 and shift the integration momentum not taking into account the cutoff .
we assume this property .
next , we consider the fuzzy torus background . note that in this case the cutoff procedure is completely determined and there is no ambiguity . to be concrete ,
we consider the @xmath0-background here .
if we use the approximations such as @xmath182 , we have @xmath183 where @xmath184 + under this approximation , at @xmath89 ,
the only difference from fuzzy plane is that the momentum @xmath185 is replaced with @xmath186 .
+ the @xmath186 s satisfy the following set of equations : @xmath187 which is an analogue of @xmath188 . using this
, we can express the free energy only by @xmath189 s .
since the momentum space is completely periodic in the fuzzy torus background , we can shift the integration variables without any obstruction , and find that the free energy is zero at this level .
if we evaluate the free energy without using the approximations , there appear @xmath77-corrections .
in this section , we discuss the validity of the @xmath85-expansion .
first we consider the case of @xmath1 .
@xmath85 s are expressed in the momentum basis such as @xmath190 and @xmath166 is expressed as @xmath191 hence , in the generic region , where @xmath192 etc , one @xmath193-insertion amounts to multiplying @xmath194 . however ,
if the arguments in @xmath195 are small , the situation changes .
for example , if we take @xmath196 and consider the region where all arguments in @xmath195 are of @xmath197 , then @xmath198 .
this shows that in the ir region convergence of the @xmath85-expansion becomes worse . in the case
that @xmath196 , the free energy does not have ir divergence .
therefore the generic region dominates , and the convergence of this expansion is good .
indeed , from the numerical calculations of the 2-loop free energy , we can see that one insertion of @xmath193 decreases the power of @xmath14 by @xmath199 . on the other hand ,
in the case of @xmath0 background , the free energy has ir divergence . therefore ir region dominates and the insertion of @xmath193 does not decrease the power of @xmath14 . in this case
@xmath85-expansion is not reliable .
on fuzzy torus , there exist zero and tachyonic modes , which we list below .
to be concrete we discuss the @xmath0-case here .
we use the following 4-dimensional gamma matrices : @xmath200 where @xmath201 are the pauli matrices and @xmath202 . + for the left - handed spinor , the dirac operator is given by @xmath203 & l_1[u^{-1},\ \cdot\ ] \\
l_1[u,\ \cdot\ ] & -l_2[v^{-1},\ \cdot\ ] \end{array } \right ) .
\end{aligned}\ ] ] hence , there are two left - handed zero modes , @xmath204 similarly for the right - handed spinors we have two zero modes given by @xmath205 these zero - modes are not doublers , because their momenta become zero in the large-@xmath14 limit .
although the numbers of zero and tachyonic modes depend on @xmath14 , they become constant when @xmath14 is sufficiently large . if the ratio @xmath206 , there are 4 zero modes and 5 tachyons . on the other hand , if @xmath207 , there are no zero modes and 7 tachyons .
a direct calculation shows that if @xmath135 , @xmath208 is an eigenstate with the smallest eigenvalue @xmath209 .
this mode corresponds to the deformation of the ratio @xmath18 .
here we evaluate the @xmath89 and @xmath96 contributions to the 2-loop free energy on the fuzzy @xmath1 .
for brevity of notation , we introduce the following notations @xmath210 where @xmath211 .
we further introduce the abbreviation @xmath212 , which is the analogue of the inner product of momenta .
we have four kinds of 1pi 2-loop vacuum diagrams ( fig [ fig:2loop_diagrams ] ) .
we list the contributions from each diagram : + + * ( a ) diagram with two fermion - fermion - boson vertices @xmath168 * @xmath213 here we have used @xmath214 .
+ + * ( b ) diagram with two ghost - ghost - vector vertices @xmath171 * + the exact result is given by @xmath215 shifting @xmath216 s and introducing approximations , @xmath217 and so on , we obtain @xmath218 + * ( c ) diagram with two boson - boson - vector vertices @xmath174 * + the exact result is given by @xmath219 again by shifting @xmath220 and using the approximations ( [ simeq : approximation ] ) , we can rewrite it as @xmath221 + * ( d ) diagram with two 4-boson vertex @xmath176 * @xmath222 summing up ( a)@xmath94(d ) , we find that the 2-loop vacuum diagrams almost cancel . actually , if we use the approximate values ( [ eq:2loop_v3g_approximated ] ) and ( [ eq:2loop_v3b_approximated ] ) , we have @xmath223 the calculation of the @xmath96 contribution is very tedious . here
we show only the simplest case , that is , the planar part of fig.[fig:2loop_diagrams ] ( b ) : @xmath224 here , @xmath225 , and @xmath101 and @xmath102 vary from @xmath103 to @xmath104
. 3 n.ishibashi,h.kawai,y.kitazawa and a.tsuchiya,_a large n reduced model as superstring _ , hep - th/9612115 .
+ m. fukuma , h. kawai , y. kitazawa , a. tsuchiya , _ string field theory from iib matrix model _ , hep - th/9705128 ; + n. ishibashi , s. iso , h. kawai , y. kitazawa _ wilson loops in noncommutative yang mills_;hep - th/9910004 + s. iso , h. kawai , _ spacetime and matter in iib matrix model - gauge symmetry and diffeomorphism - _ , hep - th/9903217 + for review , see h.aoki,s.iso,h.kawai,y.kitazawa,a.tsuchiya and t.tada , _ iib matrix model _ , hep - th/9908038 .
j. nishimura , g. vernizzi , _ spontaneous breakdown of lorentz invariance in iib matrix model _ hep - th/0003223 ; + k.n .
anagnostopoulos , j. nishimura , _ a new approach to the complex - action problem and its application to a nonperturbative study of superstring theory _ hep - th/0108041 ; + j.f .
wheater , g. vernizzi , _ rotational symmetry breaking in multi - matrix models_,hep - th/0206226 + j. nishimura , _ exactly solvable matrix models for the dynamical generation of space - time in superstring theory _ , hep - th/0108070 j.nishimura and f.sugino , _ dynamical generation of four - dimensional space - time in the iib matrix model _ , hep - th/0111102 ; + h.kawai , s.kawamoto , t.kuroki , t.matsuo and s.shinohara , _ mean field approximation of iib matrix model and emergence of four dimensional space - time _ , hep - th/0204240 ; + h.kawai , s.kawamoto , t.kuroki and s.shinohara , _ improved perturbation theory and four - dimensional space - time in iib matrix model _ , hep - th/0211272 .
+ j. nishimura , t. okubo , f. sugino , _ gaussian expansion analysis of a matrix model with the spontaneous breakdown of rotational symmetry _
, hep - th/0412194 + similar method has also applied to matrix quantum mechanics : d.kabat and g.lifschytz ,
_ approximations for strongly - coupled supersymmetric quantum mechanics _ , hep - th/9910001 a. connes , m. r. douglas , a. schwarz _ noncommutative geometry and matrix theory : compactification on tori _ ; hep - th/9711162 ; + n. seiberg , e. witten _ string theory and noncommutative geometry _
hep - th/9908142 ; + r. c. myers,_dielectric - branes _ , hep - th/9910053 ; + a.yu .
alekseev , a. recknagel , v. schomerus , _ brane dynamics in background fluxes and non - commutative geometry _ , hep - th/0003187 ; + r. j. szabo , _ quantum field theory on noncommutative spaces _ , hep - th/0109162 s.iso , y.kimura , k.tanaka , k. wakatsuki , _ noncommutative gauge theory on fuzzy sphere from matrix model _ ,
hep - th/0101102 ; + s. bal , h. takata , _ interaction between two fuzzy spheres _ , hep - th/0108002 ; + y. kitazawa , _ matrix models in homogeneous spaces _ , hep - th/0207115 ; + y. kimura , _ on higher dimensional fuzzy spherical branes _ , hep - th/0301055 ; + y. kimura , _ myers effect and tachyon condensation _
, hep - th/0309082 ; + y. kimura _
nonabelian gauge field and dual description of fuzzy sphere _
, hep - th/0402044 ; + y. kitazawa , y. takayama , d. tomino , _ correlators of matrix models on homogeneous spaces _ , hep - th/0403242 u. c. watamura , s. watamura , _ noncommutative geometry and gauge theory on fuzzy sphere _ , hep - th/9801195 ; + d. p. jatkar , g. mandal , s. r. wadia , k.p .
yogendran , _ matrix dynamics of fuzzy spheres _ ,
hep - th/0110172 + p. valtancoli , _ stability of the fuzzy sphere solution from matrix model _ , hep - th/0206075 ; + peter austing , john f. wheater , _ adding a myers term to the iib matrix model_,hep - th/0310170 ; + p.castro-villarreal , r.delgadillo-blando , b ydri , _ a gauge - invariant uv - ir mixing and the corresponding phase transition for u(1 ) fields on the fuzzy sphere _
, hep - th/0405201 t. azuma , s. bal , k. nagao and j. nishimura , _ nonperturbative studies of fuzzy spheres in a matrix model with the chern - simons term _ , hepth/0401038 ; + _ absence of a fuzzy @xmath228 phase in the dimensionally reduced 5d yang - mills - chern - simons model , _ hep - th/0405096 ; + _ dynamical aspects of the fuzzy @xmath229 in the large-@xmath14 reduced model with a cubic term _ , hep - th/0405277 . | we have calculated the free energy up to two loop to compare @xmath0 with @xmath1 in iib matrix model .
it turns out that @xmath0 has smaller free energy than @xmath1 .
we have also discussed the generation of the gauge group by considering @xmath2-coincident fuzzy tori and found that in this case u(1 ) gauge group is favored .
this means that if the true vacuum is four - dimensional , it is not a simple fuzzy space considered here .
@xmath3 [ cols= " < " , ] + + + fuzzy torus in matrix model + subrata bal@xmath4 , masanori hanada@xmath5 , hikaru kawai@xmath6 , fukuichiro kubo@xmath7 @xmath4 _ theoretical physics laboratory , riken ( the institute of physical and chemical research ) , + wako , saitama 351 - 0198 , japan _ + @xmath7 _ department of physics , kyoto university , kyoto 606 - 8502 , japan _ |
in wireless networking , determining the sets of links that can be active simultaneously is a cornerstone optimization task of combinatorial nature . for a link to be active , a given signal - to - interference - and - noise ratio ( sinr ) threshold
must be met at the receiver , according to the physical connectivity model @xcite . within this domain
, previous analyses assume that the communication system employs single - user decoding ( sud ) receivers that treat interference as additive noise . for interference - limited scenarios ,
it is very unlikely that all links can be active at the same time .
hence , it is necessary to construct transmission schedules that orthogonalize link transmissions along some dimension of freedom , such as time .
the schedule is composed by link subsets , each of which is a feasible solution to the link activation ( la ) problem .
thus , for scheduling , repeatedly solving the la problem becomes the dominant computational task .
intuitively , with sud , a solution to the la problem consists in links being spatially separated , as they generate little interference to each other .
thus , scheduling amounts to optimal spatial reuse of the time resource .
for this reason , scheduling is also referred to as spatial time - division multiple access ( stdma ) @xcite .
optimal la has attracted a considerable amount of attention .
problem complexity and solution approximations have been addressed in @xcite .
a recent algorithmic advance is presented in @xcite .
research on scheduling , which uses la as the building block , is extensive ; see , e.g , @xcite and references therein .
in addition to scheduling , la is an integral part of more complicated resource management problems jointly addressing scheduling and other resource control aspects , such as rate adaptation and power control , as well as routing , in ad hoc and mesh networks ; see , e.g. , @xcite . in the general problem setting of la , each link
is associated with a nonnegative weight , and the objective is to maximize the total weight of the active links . the weights may be used to reflect utility values of the links or queue sizes @xcite .
a different view of weights comes from the column generation , method proposed in @xcite , which has become the standard solution algorithm for scheduling as well as for joint scheduling , power control , and rate adaptation @xcite .
the algorithm decomposes the problem to a master problem and a subproblem , both of which are much more tractable than the original . solving the subproblem constructs a feasible la set . in the subproblem ,
the links are associated with prices coming from the linear programming dual , corresponding to the weights of our la problem .
a special case of the weights is a vector of ones ; in this case , the objective becomes to maximize the cardinality of the la set .
all aforementioned previous works on optimal la have assumed sud , for which interference is regarded as additive noise . in this work ,
we examine the problem of optimal la under a novel setup ; namely when the receivers have multiuser decoding ( mud ) capability @xcite . note that , unlike noise , interference contains encoded information and hence is a structured signal .
this is exploited by mud receivers to perform interference cancellation ( ic ) .
that is , the receivers , before decoding the signal of interest , first decode the interfering signals they are able to and remove them from the received signal . for ic to take place , a receiver acts as though it is the intended receiver of the interfering signal .
therefore , an interfering signal can be cancelled , i.e. , decoded at the rate it was actually transmitted , only if it is received with enough power in relation to the other transmissions , including the receiver s original signal of interest . in other words ,
the `` interference - to - other - signals - and - noise '' ratio ( which is an intuitive but non - rigorous term in this context ) , must meet the sinr threshold of the interfering signal . with mud
, the effective sinr of the signal of interest is higher than the original sinr , with sud , since the denominator now only contains the sum of the residual , i.e. , undecoded , interference plus noise .
clearly , with mud , concurrent activation of strongly interfering links becomes more likely , enabling activation patterns that are counter - intuitive in the conventional stdma setting .
the focus of our investigation is on the potential of ic in boosting the performance of la .
because la is a key element in many resource management problems , the investigation opens up new perspectives of these problems as well .
the topic of implementing mud receivers in real systems has recently gained interest , particularly in the low sinr domain using low - complexity algorithms ; see , e.g. , @xcite . technically , implementing ic is not a trivial task .
a fundamental assumption in mud is that the receivers have information ( codebooks and modulation schemes ) of the transmissions to be cancelled .
furthermore , the transmitters need to be synchronized in time and frequency .
finally , the receivers must estimate , with sufficient accuracy , the channels between themselves and all transmitters whose signals are trying to decode . for our work ,
we assume that mud is carried out without any significant performance impairments , and examine it as an enabler of going beyond the conventionally known performance limits in wireless networking . hence
, the results we provide effectively constitute upper bounds on what can be achievable , for the considered setup , in practice .
the significance of introducing mud and more specifically ic to wireless networking is motivated by the fundamental , i.e. , information - theoretic , studies of the so - called interference channel , which accurately models the physical - layer interactions of the transmissions on coupled links .
the capacity region of the interference channel is a long - standing open problem , even for the two - link case , dating back to @xcite . up to now , it is only known in a few special cases ; see , e.g. , @xcite for some recent contributions .
two basic findings , regarding optimal treatment of interference in the two - link case , can be summarized as follows .
when the interference is very weak , it can simply be treated as additive noise .
when the interference is strong enough , it may be decoded and subtracted off from the received signal , leaving an interference - free signal containing only the signal of interest plus thermal noise . from a physical - layer perspective
, the simple two - link setting above corresponds to a received signal consisting of @xmath0 , where @xmath1 is the signal of interest , with received power @xmath2 and encoded with rate @xmath3 , @xmath4 is the interference signal with received power @xmath5 encoded with rate @xmath6 , and @xmath7 is the receiver noise with power @xmath8 .
assuming gaussian signaling and capacity achieving codes , the interference @xmath4 is `` strong enough '' to be decoded , treating the signal of interest @xmath1 as additive noise , precisely if @xmath9 where @xmath10 denotes the sinr threshold for decoding the interference signal @xmath4 .
if condition ( [ eq : intro : example1 ] ) holds , i.e. , the `` interference - to - other - signal - and - noise '' ratio is at least @xmath11 , @xmath4 can be decoded perfectly and subtracted off from @xmath12 .
then , decoding the signal of interest @xmath1 is possible , provided that the interference - free part of @xmath12 has sufficient signal - to - noise ratio ( snr ) @xmath13 where @xmath14 denotes the sinr threshold for decoding signal @xmath1 .
by contrast , if the interference is not sufficiently strong for ( [ eq : intro : example1 ] ) to hold , then it must be treated as additive noise .
in such a case , decoding of signal @xmath1 is possible only when @xmath15 this way of reasoning can be extended to more than one interfering signals . towards this end , we examine the effect of ic in scenarios with potentially many links in transmission .
our study has a clear focus on performance engineering in wireless networking with arbitrary topologies . in consequence , a thorough study of the gain of ic to la is highly motivated , in view of the pervasiveness of the la problem in resource management of many types of wireless networks . in the multi - link setup that we consider , the optimal scheme is to allow every receiver perform ic successively , i.e. , in multiple stages . in every stage ,
the receiver decodes one interfering signal , removes it from the received signal , and continues as long as there exists an undecoded interfering link whose received signal is strong enough in relation to the sum of the residual interference , the signal of interest , and noise .
this scheme is referred to as _ successive ic ( sic)_. from an optimization standpoint , modeling sic mathematically is very challenging , because the order in which cancellations take place is of significance .
clearly , enumerating the potential cancellation orders will not scale at all .
thus compact formulations that are able to deliver the optimal order are essential , especially under the physical connectivity model , which quantifies interference accurately .
alternatively , a simplified ic scheme is to consider only the cancellations that can be performed concurrently , in a single stage . in this scheme , when determining the possibility for the cancellation of an interfering link , all remaining transmissions , no matter whether or not they are also being examined for cancellation , are regarded as interference .
we refer to this scheme as _
parallel ic ( pic)_. it is easily realized that some of the cancellations in sic may not be possible in pic ; thus one can expect that the gain of the latter is less than that of the former .
a further restriction is to allow at most one cancellation per receiver .
this scheme , which we refer to as _ single - link ic ( slic ) _ , poses additional limit on the performance gain . however , it is the simplest scheme for practical implementation and frequently captures most of the performance gain due to ic . in comparison to sic ,
pic and slic are much easier to formulate mathematically , as ordering is not relevant . in @xcite
, we evaluated the potential of slic in the related problem of sinr balancing .
that is , we considered as input the number of active links , let the transmit powers be variables and looked for the maximum sinr level that can be guaranteed to all links . in @xcite ,
we exploited link rate adaptation to maximize the benefits of ic to aggregate system throughput . in parallel ,
another set of authors has made a relevant contribution in the context of ic @xcite .
they considered a sic - enabled system and introduced a greedy algorithm to construct schedules of bounded length in ad - hoc networks with mud capabilities .
there though , the interference is modeled using the protocol - based model of conflict graphs @xcite , which simplifies the impact of interference , in comparison to the more accurate physical connectivity model that we are using .
the overall aim of our work is to deliver a comprehensive theoretical and numerical study on optimal link activation under this novel setup in order to provide insight into the gains from adopting interference cancellation .
this is achieved through the following contributions : * first , we introduce and formalize the optimization problems of la in wireless networks with pic and sic , focusing on the latter most challenging case . *
second , we prove that these optimization problems are np - hard . *
third , we develop ilp formulations that enable us to approach the global optimum for problem sizes of practical interest and thus provide an effective benchmark for the potential of ic on la . *
fourth , we present an extensive numerical performance evaluation that introduces insight into the maximum attainable gains of adopting ic . we show that for some of the test scenarios the improvement is substantial .
the results indicate that strong interference can indeed be taken as a great advantage in designing new notions for scheduling and cross - layer resource allocation in wireless networking with mud capabilities .
the remainder of the paper is organized as follows . in section [
sec : preliminaries ] , we introduce the notation and formalize the novel optimization problems . in section
[ sec : complexity ] , we prove the theoretical complexity results . in section [ sec :
singlestage ] , we propose a compact ilp formulation for the la problem with pic having quadratic size to the number of links .
for the most challenging problem of la with sic , we devote two sections . in section [ sec : sicuniform ] , we treat sic under a common sinr threshold . by exploiting the problem structure ,
we show that the order of cancellations can be conveniently modeled and derive an ilp formulation of quadratic size .
then , in section [ sec : sicvariable ] , we consider individual sinr thresolds . for this case , we give an ilp formulation of cubic size . in section
[ sec : simulation ] , we present and discuss simulation results evaluating the performance of all proposed ic schemes . in section [ sec : conclusion ] , we give conclusions and outline perspectives .
consider a wireless system of @xmath16 pairs of transmitters and receivers , forming @xmath16 directed links .
the discussions in the forthcoming sections can be easily generalized to a network where the nodes can act as both transmitters or receivers .
let @xmath17 denote the set of links .
the gain of the channel between the transmitter of link @xmath18 and the receiver of link @xmath19 is denoted by @xmath20 , for any two @xmath21 .
the noise power is denoted by @xmath8 and , for simplicity , is assumed equal at all receivers .
the sinr threshold of link @xmath19 is denoted by @xmath22 .
each link @xmath19 is associated with a predefined positive activation weight @xmath23 , reflecting its utility value or queue size or dual price .
if a link is activated , its transmit power is given and denoted by @xmath24 , for @xmath25 .
a la set is said to be feasible if the sinr thresholds of the links in the set are met under simultaneous transmission .
all versions of the la problems we consider have the same input that we formalize below .
+ * input : * a link set @xmath26 with the following parameters : transmit powers @xmath27 , sinr thresholds @xmath28 , and link weights @xmath29 , @xmath30 , and gain values @xmath20 , @xmath31 .
consider first the la problem with the conventional assumption of sud , where the interference is treated as additive noise .
this is the baseline version of the la problem in our comparisons ; its output is formulated as follows .
_ optimal link activation with single - user decoding .
_ + * output : * an activation set @xmath32 , maximizing @xmath33 and satisfying the conditions : + @xmath34 this classical version of the la problem can be represented by means of an ilp formulation ; see , e.g. , @xcite .
a set of binary variables @xmath35 , is used to indicate whether or not each of the links is active .
the activation set is hence @xmath36 . in order to ease comparisons to the formulations that are introduced later , we reproduce below the formulation of la - sud : [ eq : prelim : model]@xmath37 \text{s . t.}~~ & \frac{p_k g_{kk } + m_{k}(1 - x_{k})}{\sum_{m \not = k}\limits p_m g_{mk } x_m + \eta } \geq \gamma_k \qquad \forall k \in \k , \label{eq : prelim : sinr}\\ & x_k \in \{0 , 1\ } \qquad \forall k \in \k .
\label{eq : prelim : x_k01}\end{aligned}\ ] ] the objective function aims to maximize the total weight of the la set .
the constraints formulate the sinr criteria .
if @xmath38 , indicating that link @xmath19 is active , the @xmath19th inequality constrains the sinr of link @xmath19 to be at least @xmath22 . for the case that link
@xmath19 is not active , @xmath39 , the @xmath19th inequality in is always satisfied , i.e. , it has null effect , if parameter @xmath40 is set to a sufficiently large value . by construction ,
an obvious choice is @xmath41 .
note that the size of the formulation , both in the numbers of variables and constraints , is of @xmath42 .
now , consider the same system but with receivers having mud capability that enable cancellation of strongly interfering links .
we distinguish between ic in a single stage ( pic ) and in multiple stages ( sic ) . in the former , to cancel the transmission of an interfering link , all other signals of active links , including the signal of interest , are considered to be additive noise , independent of other cancellation decisions at the same receiver .
a formal definition of the output is given below .
_ optimal link activation with parallel interference cancellation . _ + * output : * an activation set @xmath32 and the set @xmath43 of cancelled transmissions for each @xmath44 , maximizing @xmath33 and satisfying the conditions : @xmath45 & \frac{p_k g_{kk}}{\sum_{m\in\a \setminus\{k,\c_k\ } } \limits p_m g_{mk } + \eta } \geq \gamma_k \qquad \forall k \in \a .
\label{eq : sinr1}\end{aligned}\ ] ] the set of conditions ensures that the specified cancellations can take place . for the receiver of link @xmath19 to cancel the transmission of link @xmath18 , the receiver of @xmath19 acts as if it was the receiver of @xmath18 .
hence , the `` interference - to - other - signals - and - noise '' ratio incorporates in the numerator the received power @xmath46 of the interfering link @xmath18 and in the denominator the received power @xmath47 of own link @xmath19 .
this ratio must satisfy the sinr threshold of the signal @xmath18 to be decoded .
the set of conditions formulates the sinr requirements for the signals of interest , taking into account the effect of ic in the sinr ratio .
that is , the cancelled terms are removed from the sum in the denominator , determining the aggregate power of the undecoded interference which is treated as additive noise . for sic , the output must be augmented in order to specify , in addition to the cancellations @xmath48 , by the receiver of @xmath44 , the order in which they take place .
a formal definition of the output is given below .
_ optimal link activation with successive interference cancellation .
_ + * output : * an activation set @xmath32 and the set @xmath49 of cancelled transmissions along with a bijection for each @xmath50 , maximizing @xmath33 and satisfying the conditions : @xmath51 @xmath52 similarly to la - pic , the set of conditions formulates the requirement for sic and the set the requirement for decoding the signals of interest , taking into account the effect of ic in the sinr ratio . in the output ,
the cancellation sequence for each @xmath19 in the activation set is given by the bijection @xmath53 ; the bijection defines a unique mapping of the link indices in the cancellation set @xmath48 to the ic order numbers in the cancellation sequence .
that is , @xmath54 defines the stage at which link @xmath18 is cancelled by the receiver of link @xmath19 .
the bijection is used in the ic conditions , in order to exclude from the sum in the denominator , the interference terms that have been cancelled in stages prior to @xmath18 .
the baseline problem , la - sud , is known to be np - hard ; see , e.g. , @xcite . for
a combinatorial optimization problem , introducing new elements to the problem structure may change the complexity level , potentially making the problem easier to solve . hence , without additional investigation , the np - hardness of la - sud does not carry over to la with ic . in this section , we provide the theoretical result that problems la - pic and la - sic remain np - hard , using a unified proof applicable to both cases . [
theorem : hardness ] problem la - pic is np - hard . we provide a reduction from la - sud to la - pic . considering an arbitrary instance of la - sud
, we construct an instance of la - pic as follows . for each link @xmath25
, we go through all other links in @xmath55 one by one .
let @xmath18 be the link under consideration .
the power of link @xmath19 is set to @xmath56 where @xmath57 is a small positive constant . by ,
the power of @xmath19 is either kept as before , or grows by an amount such that @xmath58 .
therefore , link @xmath19 is not able to decode the signal of @xmath18 , i.e. , the ic condition of la - pic can not be satisfied , even in the most favorable scenario that all other links , apart from @xmath18 and @xmath19 , are inactive . after any power increase of link @xmath19
, we make sure that this update does not have any effect in the application of to the other links .
this is achieved by scaling down the channel gain @xmath59 as @xmath60 meaning that for any @xmath18 , the received signal strength from @xmath19 remains the same as in the original instance of la - sud . as a result , even though ic is allowed in the instance of la - pic , no cancellation will actually take place , since none of the ic conditions holds due to the scalings in and . by the construction above , for each link @xmath25 the total interference that is treated as noise in the instance of la - pic equals that in the instance of la - sud .
thus , the denominator of the sinr of the signal of interest does not change . on the other hand ,
the numerator may have grown from @xmath24 to @xmath61 . to account for this growth ,
the sinr threshold @xmath22 is set to @xmath62 in effect , the increase of the power on a link , if any , is compensated for by the new sinr threshold .
note that , because @xmath63 , @xmath64 prohibits cancellation of the @xmath19th signal by any receiver other than the @xmath19th one . from the construction
, one can conclude that a la set is feasible in the instances of la - sud , if and only if this is the case in the instance of la - pic .
in addition , the reduction is clearly polynomial . hence the conclusion .
[ th : sichardness ] problem la - sic is np - hard .
the result follows immediately from the fact that , in the proof of theorem [ theorem : hardness ] , the construction does not impose any restriction on the number of links to be cancelled , nor to the order in which the cancellations take place .
in this section , we propose a compact ilp formulation for la - pic .
in addition to the @xmath65 , variables in , we introduce a second set of binary variables , @xmath66 .
variable @xmath67 is one if the receiver of link @xmath19 decodes and cancels the interference from link @xmath18 and zero otherwise .
the output of la - pic is then defined by @xmath68 and @xmath69 , for each @xmath44 .
the proposed formulation for la - pic is [ eq : ssic_model ] @xmath70 & y_{mk } \leq x_k \qquad \forall m , k \in \k , ~ m \neq k , \label{eq : ssic_activ_cond_k } \\[1ex ] & \frac{p_k g_{kk } + m_{k}(1 - x_{k})}{\sum_{m \neq k}\limits p_m g_{mk } ( x_m -y_{mk } ) + \eta } \geq \gamma _ k \qquad \forall k \in \k , \label{eq : ssic_sinr_k } \\[1ex ] & \frac{p_m g_{mk } + m_{mk}(1 - y_{mk})}{\sum_{n \neq m}\limits p_n g_{nk } x_n + \eta } \geq \gamma_m \qquad \forall m , k \in \k , ~ m \neq k , \label{eq : ssic_isnr_m } \\[1ex ] % two - column version % & \frac{p_m g_{mk } + m_{mk}(1 - y_{mk})}{\sum_{n \neq m}\limits p_n g_{nk } x_n + \eta } \geq \gamma_m \nonumber\\ % & \qquad \qquad \qquad \qquad \qquad ~~ \forall m , k \in \k , ~ m \neq k , \label{eq : ssic_isnr_m } \\[1ex ] & y_{mk } \in \{0 , 1\ } \qquad \forall m , k
\in \k , ~ m \neq k,\\[1ex ] & x_{k } \in \{0 , 1\ } \qquad \forall k \in \k.\end{aligned}\ ] ] the objective function is the same as for la - sud .
the first two sets of inequalities , and , pose necessary conditions on the relation between the variable values .
namely , a cancellation can take place , i.e. , @xmath71 , only if both links @xmath19 and @xmath18 are active , i.e. , @xmath72 .
the set of inequalities formulates the sinr requirements for decoding the signals of interest , in a way similar to for la - sud , with the difference that here the cancelled interference terms are subtracted from the denominator using the term @xmath73 .
note that , without , the formulation will fail , as in it would allow to reduce the denominator of the ratio by subtracting non - existing interference from non - active links .
the next set of constraints formulates the condition for pic : @xmath67 can be set to one only if the interference from link @xmath18 , @xmath74 , is strong enough in relation to all other active signals , including the signal of interest .
if the ratio meets the sinr threshold @xmath75 for link @xmath18 , cancellation can be carried out .
setting @xmath67 to zero is always feasible , on the other hand , provided that the parameter @xmath76 is large enough .
a sufficiently large value is @xmath77 .
the construction of reflects the fact of performing all cancellations in a single stage , as in cancelling the signal of link @xmath18 , other transmissions being cancelled in parallel are treated as additive noise . note that the model remains in fact valid even if is removed .
doing so would allow the receiver of an inactive link to perform ic .
however , since an inactive link does not contribute at all to the objective function , this is a minor `` semantic '' mismatch that can be simply alleviated by post - processing . for practical purposes , each receiver may be allowed to cancel the signal of at most one interfering link .
the resulting la problem , denoted la - slic , can be easily formulated by adapting the formulation for la - pic .
the only required change is the addition of the set of constraints @xmath78 that restricts each receiver to cancel at most one interfering transmission .
the size of the formulation , both in the numbers of variables and constraints , is of @xmath79 .
thus , the formulation is compact and its size grows by one magnitude in comparison to .
in fact , to incorporate cancellation between link pairs , one can not expect any optimization formulation of smaller size .
when implementing the formulation , two pre - processing steps can be applied to reduce the size of the problem and hence speed - up the calculation of the solution .
first , the links that are infeasible , taking into account only the receiver noise , are identified by checking for every receiver whether the received snr meets the sinr threshold for activation .
if the answer is `` false '' , i.e. , @xmath80 , then link @xmath19 is removed from consideration by fixing the @xmath81 variable to zero .
second , the link pairs for which cancellation can never take place are found by checking , for every receiver and interfering signal , whether the `` interference - to - signal - of - interest - and - noise '' ratio meets the sinr threshold for decoding the interference signal .
if the answer is `` false '' , i.e. , @xmath82 , then link @xmath19 can not decode the interference from @xmath18 and this option is eliminated from the formulation by setting the respective variable @xmath67 to zero .
incorporating the optimal ic scheme , sic , to the la problem is highly desired , since it may activate sets that are infeasible by pic . however , using ilp to formulate compactly the solution space of la - sic , is challenging .
this is because the formulation has to deal , for each link , with a bijection giving the cancellation sequence .
we propose an ilp approach and present it in two steps . in this section ,
we consider la - sic under the assumption that all links have a common sinr threshold for activation , i.e. , @xmath83 . in the next section ,
we address the general case of individual sinr thresholds . for sic under common sinr threshold ,
we exploit the problem structure and show that the optimal cancellation order can be handled implicitly in the optimization formulation . as a result , we show that la - sic can in fact be formulated as compactly as la - pic , i.e. , using @xmath79 variables and constraints .
the idea is to formulate an optimality condition on the ordering of ic . to this end , consider an arbitrary link @xmath19 and observe that meeting the sinr threshold for decoding the signal of interest is equivalent to having in the receiver , after ic , a total amount of undecoded interference and noise at most equal to @xmath84 .
we refer to this term as the _ interference margin _ @xmath85 . similarly , the interference margin that allows cancellation of the interference from link @xmath18 at the receiver of link @xmath19 is @xmath86 .
consider any two interfering links @xmath18 and @xmath87 , and suppose @xmath88 .
note that , because the sinr threshold is common , the condition is equivalent to @xmath89 , i.e. , the receiver of link @xmath19 experiences stronger interference from @xmath18 .
if the condition holds , the cancellation of @xmath18 should be `` easier '' in some sense .
thus , one may expect that if @xmath19 can decode both @xmath18 and @xmath87 , the decoding of @xmath18 should take place first . in the following ,
we prove a theorem , stating that this is indeed the case at the optimum there exists an optimal solution having the structure in which if a weaker interference signal can be cancelled , then any stronger one is cancelled before it .
[ th : order ] if @xmath88 and the receiver of link @xmath19 is able to cancel the signal of @xmath87 , then it is feasible to cancel the signal of @xmath18 before @xmath87 and there exists at least one optimum having this structure in the cancellation sequence .
let @xmath90 denote the total power of undecoded interference and noise when the receiver of @xmath19 decodes the signal from @xmath87 .
assume that @xmath18 has not been cancelled in a previous stage .
then , @xmath74 is part of @xmath90 .
successful cancellation of @xmath87 means that @xmath91 .
since @xmath92 , it holds that @xmath93 .
consider now decoding the signal of @xmath18 immediately before @xmath87 .
thus for this cancellation , the total power of the undecoded interference and noise incorporates the interference of @xmath87 , but not that of @xmath18 , i.e. , @xmath94 .
because @xmath88 implies @xmath89 , it holds that @xmath95 . since @xmath93 ,
the cancellation condition @xmath96 is satisfied . after cancelling @xmath18
, ic can still take place for @xmath87 , because the new @xmath90 is decreased by @xmath74 .
consequently , both @xmath18 and @xmath87 can be cancelled .
obviously , doing so will not reduce the number of active links and the theorem follows . by theorem
[ th : order ] , for each link @xmath19 , we can perform a pre - ordering of all other links in descending order of their interference margins .
sic at link @xmath19 can be restricted to this order without loss of optimality . at the optimum
, the cancellations performed by @xmath19 , for interfering links that are active , will follow the order , until no more additional cancellations can take place . in this optimal solution ,
when considering the cancellation condition of interfering link @xmath18 , interference can only originate from links appearing after @xmath18 in the sorted sequence .
we propose an ilp formulation based on theorem [ th : order ] .
the formulation uses the same variables of for la - pic , as there is no need to formulate the cancellation order explicitly .
the sorted sequence is denoted by , for each link @xmath97 , a bijection @xmath98 , where @xmath99 is the position of link @xmath18 in the sorted sequence .
the sorting results in @xmath100 if @xmath101 . in case of @xmath102
, the tie can be broken arbitrarily without affecting the optimization result .
in addition , let @xmath103 denote the number of links appearing after @xmath18 in the sorted sequence for @xmath19 .
the proposed formulation for la - sic under common sinr threshold is [ eq : modelic ] @xmath104 & y_{mk } \leq x_k \qquad \forall m , k \in \k , ~ m \neq k , \label{eq : cond_k } \\[1ex ] & \frac{p_k g_{kk } + m_{k}(1 - x_{k})}{\sum_{m \neq k}\limits p_m g_{mk } ( x_m -y_{mk } ) + \eta } \geq \gamma \qquad ~ \forall k \in \k , \label{eq : sinr_k } \\[1ex ] & \frac{p_m g_{mk } + m_{mk}(1 - y_{mk})}{\sum _ { n \neq k , ~ i_{k}(n ) >
i{_{k}(m ) } } \limits p_n g_{nk } x_n + p_kg_{kk } + \eta } \geq \gamma \qquad \forall m , k \in \k , ~ m \neq k , \label{eq : sinr_mk } \\[1ex ] & \sum _ { n \neq k , ~ i_{k}(n ) >
i{_{k}(m ) } } y_{nk } \leq c_{mk } ( 1 - x_m + y_{mk } ) \qquad \forall m , k \in \k , ~ m \neq k , \label{eq : ordering } \\[1ex ] % two - column version % & \frac{p_m g_{mk } + m_{mk}(1 - y_{mk})}{\sum _ { n \neq k , ~ i_{k}(n ) >
i{_{k}(m ) } } \limits p_n g_{nk } x_n + p_kg_{kk } + \eta } \geq \gamma \nonumber\\ % & \qquad \qquad \qquad \qquad \qquad \forall m , k \in \k , ~ m \neq k , \label{eq : sinr_mk } \\[1ex ] % & \sum _ { n \neq k , ~ i_{k}(n ) >
i{_{k}(m ) } } y_{nk } \leq c_{mk } ( 1 - x_m + y_{mk } ) \nonumber \\ % & \qquad \qquad \qquad \qquad \qquad \forall m , k \in \k , ~ m \neq k , \label{eq : ordering } \\[1ex ] & y_{mk } \in \{0 , 1\ } \qquad \forall m , k \in \k , ~ m \neq k , \\[1ex ] & x_{k } \in \{0 , 1\ } \qquad \forall k \in \k .\end{aligned}\ ] ] the first three constraint sets have the same meaning with for la - pic ; see section [ sec : singlestage ] . the constraint set ( [ eq : sinr_mk ] )
formulates the conditions for sic , making use of theorem [ th : order ] . consider the condition for cancellation of signal @xmath18 from receiver @xmath19 in stage @xmath99 . then
, in the denominator of the ratio , the sum of undecoded interference is limited to the transmissions coming after @xmath18 in the sorted sequence of @xmath19 , since all other active links with higher interference margin than @xmath18 have already been cancelled .
the formulation is however not complete without .
this set of constraints , in fact , ensures the optimality condition set by theorem [ th : order ] and utilized in .
that is , if both @xmath18 and @xmath87 are active , @xmath105 , and @xmath87 is cancelled by @xmath19 , then @xmath18 is cancelled by @xmath19 as well .
equivalently speaking , if @xmath18 is active but not cancelled by @xmath19 , then none of the other links after @xmath18 in the sequence of @xmath19 may be cancelled .
examining , we see that it has no effect as long as @xmath106 equals @xmath67 . if link @xmath18 is active but not cancelled , corresponding to @xmath107 and @xmath108 , the right - hand side of becomes zero , and therefore no cancellation will occur for any @xmath87 having position after @xmath18 in the ordered sequence .
also , note that the case @xmath109 but @xmath71 can not occur , because of .
given a solution to the formulation , the cancellation sequence of each active link @xmath19 , i.e. , the bijection @xmath53 in the definition of la - sic in section [ sec : preliminaries ] , is easily obtained by retrieving from the predefined bijection @xmath110 the elements with @xmath111 .
the compactness of the formulation is manifested by the fact that its size , in both the numbers of variables and constraints , is of @xmath79 .
thus , provided that there is a common sinr threshold for activation , we have formulated la - sic as compactly as la - pic .
when implementing the formulation , similar pre - processing steps with for la - pic can be applied to reduce the size of the problem . first , the infeasible links are removed for consideration by fixing @xmath81 to zero when @xmath112 .
second , the infeasible ic options are eliminated from the formulation by fixing @xmath67 to zero when @xmath113 .
in this section , we consider the la - sic problem under the most general setup ; namely when the links have individual sinr thresholds . differently from the common sinr case , treated in section [ sec : sicuniform ] ,
a pre - ordering of the sequence of potential ic does not apply .
the reason is that the interference margin @xmath114 does not depend anymore only on the received power @xmath46 but also on the link - specific sinr threshold @xmath75 . to see this point , consider a scenario where link @xmath19 attempts to cancel the signal of two interfering links @xmath18 and @xmath87 in two consecutive stages .
denote by @xmath4 the sum of the remaining interference , other than @xmath18 or @xmath87 , the received power of link @xmath19 s own signal , and noise .
assume a mismatch between the relation of interference margin and that of received power : @xmath115 but @xmath88 because @xmath116 . if @xmath19 cancels @xmath18 and then @xmath87 , the cancellation conditions are @xmath117 and @xmath118 . reversing the cancellation order leads to the conditions @xmath119 and @xmath120 .
for our example , we let @xmath121 .
consider two sets of values for the other parameters .
the values in the first set are : @xmath122 , @xmath123 , @xmath124 , @xmath125 and in the second set are : @xmath126 , @xmath123 , @xmath127 @xmath125 . for the first set , both interfering links can be cancelled only if cancellation applies to @xmath18 first , whereas the opposite order must be used for the second set .
hence the interference margin ( or received power ) does not provide a pre - ordering for cancellation . in the following ,
we propose an ilp formulation for la - sic under individual sinr thresholds , that explicitly accounts for the cancellation order .
our approach is to introduce for each pair of links , @xmath128 , a set of binary variables @xmath129 and represent the cancellation stage by the superscript @xmath130 .
variable @xmath129 is one if and only if the receiver of link @xmath19 cancels the interference from link @xmath18 at stage @xmath130 .
the effect is that , for each link @xmath19 , the solution values of @xmath129 order the feasible cancellations ; hence , they have a direct correspondence to the output bijection @xmath53 of la - sic , defined in section [ sec : preliminaries ] .
it is apparent that the index @xmath130 ranges between one and @xmath131 . in practice , due to computational considerations
, we may want to restrict the maximum number of cancellation stages .
to this end , we define , for each @xmath97 , the integer parameters @xmath132 and the sets @xmath133 . the proposed formulation of the general la - sic problem , under individual sinr thresholds and restricted cancellation stages , is [ eq : finalmodel ] @xmath134 & \text{s . t. } \nonumber \\ & \sum _ { t=1}^{t_k } y^{t}_{mk}\leq x_m , \qquad \forall m , k
\in \k , ~ m \neq k , \label{eq : finalactiv}\\[1ex ] & \sum_{m \neq k } y^{t}_{mk } \leq x_k , \qquad \forall k \in \k , ~\forall t\in\t_k , \label{eq : finalstage}\\[1ex ] & \frac{p_k g_{kk } + m_{k}(1 - x_{k})}{\sum_{m \not = k}\limits p_m g_{mk } \big(x_m -{\sum _ { t=1}^{t_k}\limits y^{t}_{mk}}\big ) + \eta } \geq \gamma_k \qquad \forall k \in \k , \label{eq : finalsinr } \\[1ex ] & \frac{p_m g_{mk } + m_{mk } ( 1-y_{mk}^t)}{\sum_{n \not = m , k } \limits p_n g_{nk } \big(x_n - { \sum_{t'=1}^{t-1 } \limits y^{t'}_{nk } } \big ) + p_k g_{kk } + \eta } \geq \gamma_m \qquad \forall m , k \in \k , ~ m \neq k , ~\forall t\in\t_k , \label{eq : finalisnr } \\[1ex ] % two - column version % & \frac{p_m g_{mk } + m_{mk } ( 1-y_{mk}^t)}{\sum_{n \not = m , k } \limits p_n g_{nk } \big(x_n - { \sum_{t'=1}^{t-1 } \limits y^{t'}_{nk } } \big ) + p_k g_{kk } + \eta } \geq \gamma_m \nonumber \\ % & \qquad\qquad\qquad\qquad\quad \forall m , k \in \k , ~ m \neq k , ~\forall t\in\t_k , \label{eq : finalisnr } \\[1ex ] & \sum_{m \neq k } \limits y^{t}_{mk } \leq \sum_{m \neq k}\limits y^{t-1}_{mk } \qquad \forall k \in \k , ~\forall t \in \t_k \setminus\{1\ } , \label{eq : finalblock}\\[1ex ] & y^{t}_{mk } \in \{0 , 1\ } \qquad \forall m , k \in \k , ~ m \neq k , ~\forall t\in\t_k , \label{eq : finaly}\\[1ex ] & x_{k } \in \{0 , 1\ } \qquad \forall k \in \k.\label{eq : finalx}\end{aligned}\ ] ] the conditions have similar role with .
namely , only when links @xmath19 and @xmath18 are active , the receiver of @xmath19 can consider to cancel the transmission of @xmath18 .
in addition , the summation over @xmath130 in ensures that link @xmath18 is cancelled in at most one stage .
furthermore , the summation over @xmath18 in enforces each receiver to perform at most one cancellation per stage .
this removes equivalent solutions , without compromizing optimality , to enhance computational efficiency .
the sinr requirements for decoding the signals of interest are set in , similarly to . in the denominator ,
all cancelled links , regardless of the stage the cancellation is performed , are removed from the sum of undecoded interference .
the next set of constraints formulates the requirement for the cancellation of link @xmath18 by link @xmath19 at stage @xmath130 .
the active interfering transmissions that have been cancelled before stage @xmath130 are excluded from the sum of undecoded interference in the denominator of the ratio .
the constraints are formulated with the convention that the sum within the parenthesis in the denominator of the ratio is zero for @xmath135 .
note that , even though for each receiver @xmath19 and interfering link @xmath18 , @xmath136 constraints are formulated , due to , all but at most one will be trivially satisfied by the respective @xmath129 variabls being equal to zero . the constraints are not mandatory for the correctness of the formulation , but their role is to enhance the computational efficiency .
these constraints ensure that the cancellations are performed as `` early '' as possible , i.e. , there are no `` idle '' stages at which no cancellation takes place and which are followed by later stages where cancellation takes place .
otherwise , if @xmath137 and @xmath138 are cancelled by @xmath19 , and the cancellation of the former takes place first , the cancellations can be performed at any two stages @xmath139 and @xmath140 , as long as @xmath141 .
clearly , such solutions are all equivalent to each other , and the presence of them slows down the computational process .
+ since formulates the most general la - sic problem , it also applies to the common - sinr case of section [ sec : sicuniform ] .
its computational efficiency though is significantly lower than the respective of formulation .
the reason is that its size is one magnitude larger than , i.e. , the numbers of variables and constraints grow from @xmath79 to @xmath142 . however
, we note that the formulation remains compact . in order to deal with the scalability issue
, one may resort to restrict the maximum number of cancellations , @xmath136 , to a constant being considerably lower than @xmath131 .
doing so has little impact on the solution quality , because most of the performance gain from ic is due to the first few cancellations .
also , when implementing the formulation , similar pre - processing steps with can be applied , see section [ sec : sicuniform ] , to reduce the size of the problem .
this section presents a quantitative study of the effect that ic has on the optimal la problem in wireless networking .
the ilp formulations , proposed in sections [ sec : singlestage][sec : sicvariable ] , are utilized to conduct extensive simulation experiments on randomly generated network instances with various topologies , densities , cardinalities , and sinr thresholds .
nodes are uniformly scattered in square areas of @xmath143 m and @xmath144 m , in order to create sparse and dense topologies , respectively .
two types of datasets are generated .
the first one takes an information - theoretic viewpoint and is henceforth denoted dataset i. to this end , the transmitter - receiver matchings are arbitrarily chosen @xcite , with the sole criterion of feasible single - link activation .
thus , the links have arbitrary length within the test area , provided that their snr is larger than the sinr threshold required for activation .
the second dataset provides a rather networking - oriented approach @xcite and is henceforth denoted dataset n. in this dataset , the length of the links is constrained to be from @xmath145 m up to @xmath146 m , with the rationale to produce instances resembling a multihop network .
the networks considered have cardinality @xmath16 ranging from @xmath147 up to @xmath148 links .
[ fig : instances ] illustrates instances of a 20-link network ; figs .
[ fig : setup20spit ] and [ fig : setup20deit ] correspond to the sparse and dense topology , respectively , of dataset i , whereas figs .
[ fig : setup20spnw ] and [ fig : setup20denw ] correspond to the sparse and dense topology , respectively , of dataset n. + the input parameters are chosen to be common for all links ; specifically , the transmit power @xmath24 , @xmath149 , is set to @xmath148dbm , the noise power @xmath8 to @xmath150dbm , and the channel gains @xmath20 follow the geometric , distance - based , path loss model with an exponent of @xmath151 .
the major difference between the datasets is the distribution of the link lengths , which effectively determines the snr distribution of the links .
the input parameters yield minimum snr approximately equal to @xmath151db , @xmath152db , and @xmath153db for dataset sparse i , dense i , and n , respectively .
the histograms in fig .
[ fig : snr ] illustrate the snr distribution of each dataset ; as in fig . [
fig : instances ] , left and right sub - figures are for dataset i and n , respectively , whereas upper and lower sub - figures are for sparse and dense topologies , respectively . for dataset
i , the links in the sparse topology have on average lower snr than in the dense topology ; the mass of the snr distribution is roughly for @xmath154@xmath155db in the sparse and for @xmath156@xmath157db in the dense topology .
this is because in the sparse topology the test area is enlarged , allowing generation of longer links which have lower snr values . on the contrary , the snr distribution of dataset
n is invariant to the network density ; this is by construction , since the distribution of the link lengths is not affected by the size of the test area .
+ for each dataset and network cardinality , @xmath148 instances are generated and the performance of la with ic is assessed by two simulation studies . in the first study ,
all links are assumed to require for activation a common sinr threshold @xmath158 , @xmath149 , taking values from @xmath159db up to @xmath160db , and have equal activation weights , e.g. , @xmath161 , @xmath149 .
the goal is to evaluate the performance gain due to single - link , parallel , and successive ic schemes on the la problem over the baseline approach without ic . for this purpose
, we implemented the formulations , , , and , for la - sud , la - slic , la - pic , and la - sic , respectively .
[ fig : activation ] illustrates exemplary activation sets for an instance of a 10-link network , drawn from dataset dense i , when the sinr threshold is @xmath159db .
it is evidenced that performance increases with problem sophistication : figs .
, , , and , show that la - sud , la - slic , la - pic , and la - sic activate @xmath145 , @xmath147 , @xmath160 , and @xmath154 links , respectively .
the optimal solutions are found by an off - the - shelf solver , implementing standard techniques such as branch - and - bound and cutting planes @xcite .
the simulations were performed on a server with a quad - core amd opteron processor at 2.6 ghz and 7 gb of ram .
the ilp formulations were implemented in ampl 10.1 using the gurobi optimizer ver .
3.0 . regarding the computational complexity of the proposed ilp formulations for ic ,
an empirical measure is the running time of the solution process .
we have observed that it is not an obstacle for practical instance sizes . + in the following a selection of the simulation results is presented .
[ fig : sinr-6 ] shows the average , over @xmath148 instances , number of activated links versus the total number of links in the network , achieved by all versions of the la problem when the sinr threshold is @xmath162db .
the results in the four sub - figures correspond to the datasets exemplified in fig .
[ fig : instances ] .
the major observation is that all la schemes with ic clearly outperform la - sud and in particular la - sic yields impressive performance . comparing figs .
[ fig : sinr-6is ] and [ fig : sinr-6id ] , it is concurred that the results for dataset i are density invariant .
as the number of links in the network increases , the performance of la - sud improves , due to the diversity , almost linearly but with very small slope .
la - sic though improves significantly , activating two to three times more links than the baseline .
when the network has up to about @xmath163 links , nearly all of them are activated with la - sic . on the other hand
, la - pic has a consistent absolute gain over la - sud , activating one to two links more .
furthermore , la - slic has almost as good performance with la - pic , i.e. , it captures most of the gain due to single - stage ic .
[ fig : sinr-6nd ] shows that the la schemes have similar performance in dataset dense n as in dataset i. fig .
[ fig : sinr-6ns ] shows that la is easier for dataset sparse n , even without ic .
the curves of all schemes linearly increase with network cardinality , but with ic the slopes are higher , so that the absolute gains , differences from the baseline , broaden .
maximum gains are for @xmath148 links , where la - sud , la - slic , la - pic , and la - sic activate about @xmath164 , @xmath165 , @xmath166 , and @xmath148 links , respectively . for the tested network cardinalities , la - sic achieves the ultimate performance activating all links . + as seen in fig .
[ fig : sinr-6 ] , the performance gains due to ic are very significant when the sinr threshold for activation is low . however ,
for high sinr thresholds , the gains are less prominent .
for example , fig .
[ fig : sinr3 ] shows the performance of the la schemes when @xmath167 is set to @xmath145db .
[ fig : sinr3is ] and [ fig : sinr3ns ] are for the sparse datasets i and n , respectively ; for the dense topologies the results are similar to fig .
[ fig : sinr3is ] .
it is evidenced that ic schemes activate one to two links more than the baseline and that most of this gain can be achieved with single - stage ic . + the fact that the ic gains diminish with increasing the sinr threshold is clearly illustrated in fig . [
fig : size30 ] , which compares , for networks of 30 links , the average performance of all la schemes for various sinr thresholds .
the relative gain of sic is more prominent in the case of dataset i , which is more challenging for the baseline problem .
for dataset i , when the sinr threshold is low , around @xmath159db , sic activates nearly all links , whereas sud activates less than a third of them . for sparse and dense dataset
n , sic activates effectively all links when the sinr threshold is lower than @xmath162db and @xmath159db , respectively , whereas sud activates less than two thirds and less than half of them , respectively . for mid - range sinr thresholds ,
up to about @xmath145db , sic has an exponentially decreasing performance , but nevertheless still significantly outperforms sud . on the other hand , for sinr thresholds up to about @xmath168db , pic yields a relatively constant performance improvement of roughly two to five links , depending on the dataset .
pic is effectively equivalent to its simpler counterpart slic , for sinr higher than @xmath162db .
the performance of all ic schemes converges for sinr thresholds higher than 3db .
the interpretation is that if ic is possible , it is more likely that it will be restricted to a single link .
for very high sinr thresholds , it becomes rarely possible to perform ic .
+ in the second simulation setup , the performance of the general la - sic problem , under individual sinr thresholds , is evaluated .
the sinr threshold @xmath22 for each link is taking , with equal probability , one of the values in the set @xmath169db and the activation weights @xmath23 are set equal to the data rates , in bits per second per hertz , corresponding to the respective sinr thresholds . the formulation is implemented varying the maximum number of cancellation stages @xmath170 , @xmath149 , from @xmath168 , corresponding to the baseline case without ic , up to @xmath147 . fig .
[ fig : individual ] shows the average , over 30 instances , throughput of all activated links versus the network cardinality , for all the datasets .
for dataset i , the network throughput is almost doubled with ic ; roughly half of this increase is achieved by the first cancellation stage and most of the rest by the next two to three stages . for dataset
n , it is seen that the first cancellation stage yields a significant gain of about @xmath171 b / s / hz and that it only pays off to have more than two cancellation stages for large and sparse networks .
in this paper , we have addressed the problem of optimal concurrent link activation in wireless systems with interference cancellation .
we have proved the np - hardness of this problem and developed integer linear programming formulations that can be used to approach the exact optimum for parallel and successive interference cancellation . using these formulations ,
we have performed numerical experiments to quantitatively evaluate the gain due to interference cancellation .
the simulation results indicate that for low to medium sinr thresholds , interference cancellation delivers a significant performance improvement .
in particular , the optimal sic scheme can double or even triple the number of activated links .
moreover , node density may also affects performance gains , as evidenced in one of the datasets . given these gains and the proven computational complexity of the problem , the development of approximation algorithms or distributed solutions incorporating ic
are of high relevance .
concluding , the novel problem setting of optimal link activation with interference cancellation we have introduced here provides new insights for system and protocol design in the wireless networking domain , as in this new context , strong interference is helpful rather than harmful .
thus , the topic calls for additional research on resource allocation schemes in scheduling and routing that can take the advantage of the interference cancellation capability .
indeed , the la setup studied herein assumes fixed transmit power for active links .
this can lead to increased interference levels , since the sinrs can be oversatisfied .
incorporating power control to the la problem with ic will bring another design dimension which can yield additional gains .
furthermore , it may enable ic for high sinr thresholds ic .
v. s. annapureddy and v. v. veeravalli , `` gaussian interference networks : sum capacity in the low interference regime and new outer bounds on the capacity region , '' _ ieee trans .
theory _ , vol .
55 , no . 6 , pp .
30323050 , 2009 .
a. capone , g. carello , i. filippini , s. gualandi , and f. malucelli , `` routing , scheduling and channel assignment in wireless mesh networks : optimization models and algorithms , '' , _ ad hoc netw .
_ , vol . 8 , pp . 545 - 563 , 2010
. a. capone , l. chen , s. gualandi , and d. yuan , `` a new computational approach for maximum link activation in wireless networks under the sinr model , '' _ ieee trans .
wireless commun .
, _ vol .
10 , no . 5 , pp .
13681372 , 2011 .
l. tassiulas , and a. ephremides , `` stability properties of constrained queueing systems and scheduling for maximum throughput in multihop radio networks , '' _ ieee trans . on automat .
1936 - 1949 , 1992 . | a fundamental aspect in performance engineering of wireless networks is optimizing the set of links that can be concurrently activated to meet given signal - to - interference - and - noise ratio ( sinr ) thresholds .
the solution of this combinatorial problem is the key element in scheduling and cross - layer resource management .
previous works on link activation assume single - user decoding receivers , that treat interference in the same way as noise . in this paper
, we assume multiuser decoding receivers , which can cancel strongly interfering signals . as a result , in contrast to classical spatial reuse , links being close to each other are more likely to be active simultaneously .
our goal here is to deliver a comprehensive theoretical and numerical study on optimal link activation under this novel setup , in order to provide insight into the gains from adopting interference cancellation .
we therefore consider the optimal problem setting of successive interference cancellation ( sic ) , as well as the simpler , yet instructive , case of parallel interference cancellation ( pic ) .
we prove that both problems are np - hard and develop compact integer linear programming formulations that enable us to approach the global optimum solutions .
we provide an extensive numerical performance evaluation , indicating that for low to medium sinr thresholds the improvement is quite substantial , especially with sic , whereas for high sinr thresholds the improvement diminishes and both schemes perform equally well . |
electroweak baryogenesis @xcite provides a predictive framework for the computation of the baryon asymmetry of the universe @xcite .
perhaps the most attractive feature of this mechanism is that it relies on anomalous baryon number violation processes which are present in the standard model @xcite . at temperatures far above the electroweak phase transition scale ,
these anomalous processes are unsuppressed and , in the absence of any @xmath0 asymmetry , they lead to the erasure of any baryon or lepton number generated at high energy scales @xcite . these baryon number violation processes are , instead , exponentially suppressed in the electroweak symmetry broken phase , at temperatures far below the electroweak phase transition @xcite . at the electroweak phase transition , non - equilibrium processes may generate a non - vanishing baryon number which may efficiently diffuse into the broken phase @xcite .
the generated baryon number depends on the cp - violating currents present in the model .
the mechanism of electroweak baryogenesis may become effective if the cp - violating sources are strong and , at the electroweak phase transition temperature , the baryon number violation processes in the broken phase are sufficiently suppressed , leading to a baryon number density in the broken phase consistent with observations .
this , in turn , demands a strongly first order electroweak phase transition .
it has been long ago realized that in the standard model the cp violating sources are too weak to lead to an acceptable baryon number density @xcite
. moreover , even if the sources were strong enough to lead to a reasonable baryon number generation , the phase transition is weakly first order , leading to a strong exponential suppression of the baryon number generated in the broken phase @xcite .
a minimal supersymmetric extension of the standard model ( mssm ) , instead , has all the necessary ingredients to improve on both problems @xcite .
first , there are additional sources of cp - violation , associated with the cp - violating phases of the supersymmetry breaking parameters .
second , in the presence of a light stop , the phase transition may become strong enough to allow the preservation of the baryon number generated at the electroweak symmetry breaking scale @xcite-@xcite .
although there is general agreement on both the existence and nature of the new sources and also on the parameter space leading to a strongly first order phase transition , the results regarding the strength and specific form of the cp - violating sources are still controversial .
most groups working on the subject have found that , for values of the cp - odd higgs mass of the order of the weak scale , the sources are dominated by a term proportional to @xmath1 , where @xmath2 denote the expectation value of the neutral components of the two higgs doublets present at low energies @xcite-@xcite .
recently , however , it has been claimed that this contribution is absent , leading to a suppressed result for the baryon asymmetry contribution within the mssm proportional to @xmath3 @xcite . in this article
, we proceed to perform a critical re - analysis of the diffusion equations leading to the baryon asymmetry and of the sources for those diffusion equations . in particular , we demonstrate that the sources can be expressed in terms of appropriate differential operators times the cp - violating currents we computed in a very recent work .
we find no suppression of the dominant sources proportional to @xmath4 . indeed , we show that the suppression claimed the authors of refs .
@xcite is an artifact of the approximation they used in order to compute the currents .
the organization of the article is as follows . in section 2
we provide the derivation of the diffusion equations . in section 3
we review the derivation of the cp - violating currents and we show explicitly where our results differ from the ones of ref . @xcite . in section 4
we discuss the results for the baryon asymmetry within the mssm and section 5 contains our conclusions .
finally some useful formulae concerning the chargino sector of the mssm are summarized in appendix a.
we will start considering a system of particles propagating in a non - trivial cp - violating background localized in the bubble wall , where the bubble is expanding with a speed @xmath5 in the plasma frame .
the presence of the cp - violating background and the particle number changing reactions perturb the distribution functions for particles and antiparticles around the equilibrium .
the corresponding distribution functions @xmath6 satisfy the boltzmann equations , @xmath7\ ] ] where we are using the notation @xmath8 , @xmath9 , @xmath10 is the four - velocity and @xmath11 is the force generated by the non - trivial background .
the term on the right - hand side @xmath12 encodes the effect on the distribution of the particle number changing reactions and elastic collisions . in the bubble - wall frame and neglecting the curvature of the bubble wall the distribution functions can be written as , @xmath13}\pm 1}+\delta f_i^p\ ] ] where @xmath14 is the chemical potential , @xmath15 and @xmath16 are the energy and momentum of the particle , the momentum - dependent part @xmath17 describes departure of the system from kinetic equilibrium and we have introduced a slowly varying cp - even function of @xmath16 and @xmath18 , @xmath19 , which tends to 1 in the absence of forces .
considering the effective distribution functions @xmath20 , describing the difference between particles and anti - particles of a given species , and applying ( [ fwall ] ) to eq .
( [ boltzmann0 ] ) gives @xmath21 + \left(f_i^0+v_\omega f_i^z \right ) \bar{g}(p_z , z ) \frac{\partial f_0}{\partial e}=c_i\ ] ] where @xmath22 is the cp - violating component of the force , we have linearized with respect to the small perturbations , @xmath23 , @xmath24 , in the spirit of the chapman - enskog first order approximation @xcite , @xmath25 is the equilibrium distribution function in the plasma frame , and the @xmath26 denotes derivative with respect to @xmath18 . in this process , due to the assumed smoothness of the cp - even function @xmath19 , we have ignored terms proportional to derivatives of this function . following ref .
@xcite , we rewrite the above equation in the plasma frame by performing a galilean transformation on the velocity @xmath27 to @xmath28 , @xmath29 + \left(f_i^0+v_\omega f_i^z \right ) g(p_z , z ) \frac{\partial f_0}{\partial e}=c_i\ ] ] where @xmath30 is assumed to be an even function of momenta in this frame .
we now define @xmath31 and using the decomposition @xmath32 and the definition of the current @xmath33 we can relate the cp - violating current with the chemical potential and the function @xmath34 as , @xmath35 where , again , we have neglected terms proportional to derivatives of @xmath30 .
the function @xmath34 is such that @xmath36 and the number density is entirely provided by the chemical potential @xmath23 . in this way we can choose , with all generality , the function @xmath34 as odd with respect to @xmath27 and therefore satisfying the property @xmath37 .
we now multiply eq .
( [ boltzplasma ] ) by @xmath38 and integrating over momenta we get the two equations , @xmath39 where @xmath40 is the total interaction rate and the diffusion coefficient @xmath41 is defined as @xmath42 and we have again ignored the smooth momentum dependence of @xmath30 in the numerator and denominator integral functions .
observe that the ratio ( [ diffusion ] ) remains constant up to derivatives of @xmath43 which we have consistently ignored in our treatment .
the terms on the right - hand sides of eqs .
( [ ecua1 ] ) and ( [ ecua2 ] ) can be decomposed , using the decomposition ( [ decomp ] ) as @xmath44=c_i^1[\mu]+c_i^2[\delta f]\ ] ] where @xmath45 and @xmath46 are linear in @xmath23 and @xmath34 , respectively .
in fact @xmath47\right>$ ] is entirely provided by inelastic collisions @xmath48\right>\simeq -\gamma_{ij } \frac{n_j}{k_j}\ ] ] where @xmath49 ( @xmath50 ) for bosons ( fermions ) are statistical factors , @xmath51 ( @xmath52 ) is the averaged interaction rate for the inelastic reaction channel ( @xmath53 ) and @xmath54 implies a sign over all particles participating in the reaction with appropriate sign assignments : for an inelastic reaction corresponding to the decay @xmath55 with decay width @xmath56 , its contribution to the right - hand side of ( [ c1 ] ) would be @xmath57 .
furthermore @xmath58\right>=0 $ ] from the oddness of @xmath34 . on the other hand , @xmath59\right>=0 $ ] . within the present framework ,
the cp - violating forces are provided by the interaction of the different fields with the higgs profiles , which vary along the bubble wall .
these cp - violating forces induce cp - violating currents for the different fields , which lead to @xmath60 after elastic and inelastic interactions with the other fields present in the plasma .
the resulting currents have then two components . the first one , @xmath61 , the background component , is the one that would be obtained in the presence of only the interaction with the background higgs field , and the second one , the `` collision '' component comes from the interaction with the plasma fields and lead to the diffusion process .
the integral @xmath62\right>$ ] is contributed by all collisions and dominated by the elastic ones .
correspondingly with the two components , background and collision , of the currents there can be defined two @xmath63 components .
the part of @xmath63 that is induced by the interactions with the plasma fields , the collision component , is governed by the same interactions as the ones governing @xmath64 and is therefore relevant in the same regime of momenta in which the collision term becomes important . on the other hand
the @xmath63 component coming from the interaction with the background fields , @xmath65 , is not correlated with the collision terms and leads naturally to a negligible contribution to @xmath66\right>$ ] .
therefore , considering that the rate of interactions is approximately constant in the momentum regime in which the collision contribution to the currents become relevant , and negligible anywhere else , we get @xmath67\right > \simeq \gamma_i^t\ , \left(j_i^z - j_i^{(b)z } \right ) .\ ] ] using now ( [ c1 ] ) and ( [ c2 ] ) in ( [ ecua1 ] ) and ( [ ecua2 ] ) we obtain the diffusion equations , @xmath68 a current @xmath69 in the presence of a cp - violating higgs background ( or cp - violating force @xmath70 ) but where no interactions with the plasma fields were considered was computed in ref . @xcite .
it should satisfy the set of boltzmann equations @xmath71 we can see from these equations , and in particular from ( [ ecua3b ] ) that the roles played by the cp - violating forces and currents @xmath72 are equivalent .
in fact by subtracting ( [ ecua3b ] ) from ( [ ecua3 ] ) one obtains an equation @xmath73 where the force terms have been replaced by the divergence of the cp - violating current . by subtracting now ( [ ecua4b ] ) from ( [ ecua4 ] )
we obtain , @xmath74\simeq d_i \left(n_i - n_i^{(b)}\right)'\ ] ] where the last approximation comes from the requirement @xmath75 ( where @xmath76 is the wall thickness ) , which is necessary for the validity of the derivative expansion we will use in the calculation of the cp - violating currents in section [ sources ] . eq .
( [ fick ] ) represents the well - known empirical law of diffusion or fick s law .
replacing now eq .
( [ fick ] ) into ( [ ecua3f ] ) we obtain the final expression for the diffusion equation as , @xmath77\ ] ] where the source is given by @xmath78 and the last approximation comes from the fact that @xmath79 .
notice that eq .
( [ ecfinal ] ) has the correct boundary conditions since in the absence of inelastic reactions ( for @xmath80 ) it provides the trivial solution @xmath81 as required .
we can compare this result with the source obtained in ref .
these authors work in a wkb approximation with quasi - particles . within their semiclassical approximation
the energy is a ( constant ) label of wkb - states and therefore the time component of the force @xmath82 vanishes , while they have assumed that @xmath83 .
moreover their source , @xmath84 , is given by @xcite , @xmath85 which appears in eq .
( [ ecua4b ] ) above and can therefore be related to @xmath86 and @xmath87 as they also appear in ( [ ecua4b ] ) . although we have obtained the sources @xmath88 in a way independent of the form of the cp - violating forces @xmath89 and @xmath90 , for vanishing @xmath90 we obtain a relation between the sources @xmath88 and @xmath84 .
indeed , using eq .
( [ ecua4b ] ) , we obtain @xmath91 where the last approximation relies on the relation @xmath92 .
the latter relation holds for the case of charginos in the mssm , analyzed in section [ sources ] .
in this section we will make contact between the source in ( [ source ] ) , and in particular the current @xmath93 , and the green function and will apply the formalism to the sector of charginos in the mssm .
first consider a chiral fermion ( say a right - handed one ) in the presence of the non - trivial background .
its green function @xmath94 can be considered as function of the center - of - mass coordinate @xmath95 and the relative coordinate @xmath96 , @xmath97 since we will make a gradient expansion with respect to the coordinate @xmath18 we perform a fourier transformation with respect to the relative coordinate @xmath98 as @xmath99 which is called the wigner representation of the green function @xmath100 . by making the general decomposition @xmath101
one can define the corresponding background current @xmath93 as @xmath102 and making the inverse wigner transformation ( [ wigner ] ) it can be cast as @xmath103 the current can be written in an alternative equivalent form , that makes connection with the discussion in the previous section , @xmath104 under small perturbations the relevant component of the integrand is approximately dominated by @xmath105 functions imposing the dispersion relation between @xmath106 and the energy of the particle .
indeed , for a free particle in the presence of a non - trivial chemical potential , the green functions ( [ relations ] ) and ( [ prop ] ) lead to @xmath107 where @xmath108 and , since the momentum integral of the first term vanishes , the charge density is obtained from the integration of the second term . under these conditions ,
( [ cor ] ) and ( [ cor2 ] ) are consistent with the definitions of the particle density @xmath109 and current @xmath110 in ( [ corriente ] ) . for fermions of opposite chirality ( say left - handed ones ) the previous expressions hold just changing @xmath111 . on the other hand for several flavors , the mass eigenstates are an admixture of the weak eigenstates , with a mixing that depends on the value of the varying higgs background .
the expression of the diffusion equations and the sources , eqs .
( [ ecfinal ] ) and ( [ source ] ) , are related in a very simple way ; namely the sources are trivially obtained by demanding a self - consistency relation in the absence of particle changing interactions .
this property should hold , in first approximation , for any flavor structure of the theory .
the relevant sources for the diffusion equations should be obtained by generalizing the trace in ( [ cor2 ] ) to flavor space and eventually including a projection operator inside the trace , @xmath112 , when a given contribution is to be picked up . as stressed above
, this formalism makes contact with the one proposed earlier by the authors of ref .
@xcite and should lead to a more realistic estimate of the baryon asymmetry of the universe than the one presented in previous analyses of this subject .
although , as we have already outlined , we agree with the results of ref .
@xcite in the formal definition of the sources , we shall now show that the structure of these sources differs from the one presented by the authors of ref .
we shall try to clarify the origin of the discrepancy with those authors results . for the chargino sector of the mssm we define the right - handed and left - handed fermions as @xmath113 and
expand the mass matrix to first order in derivatives around the point @xmath18 as @xmath114 where we use the notation @xmath115 , and the mass matrix @xmath116 is defined in ( [ masach ] ) .
we now consider the first term in ( [ mexp ] ) as part of the unperturbed lagrangian , and the second term as the perturbation @xmath117 or , in the basis of mass eigenstates , @xmath118 @xmath119 we denote by @xmath120 , @xmath121 , @xmath122 and @xmath123 the left - left , right - right , left - right and right - left green functions of free fermions in the mass eigenstate basis with mass @xmath124 as given in ( [ relations ] ) . in this basis
the right- and left - handed green functions get modified to @xmath125 and @xmath126 by the presence of the interaction term ( [ philag ] ) and can be defined as solutions of schwinger - dyson equations . to first order in the insertion given by the interaction term ( [ philag ] ) the solution to the schwinger - dyson equations
is @xmath127 and a similar solution for @xmath126 .
the cp - violating green function @xmath128 defined in ( [ cgreen ] ) can be written as , @xmath129 where the notation @xmath130 has been used . working out eq .
( [ inicio ] ) it can be cast as , @xmath131 the calculation of the coefficients @xmath132 , @xmath133 and @xmath134 is straightforward
. we will only give the relevant pieces that contribute to the cp violating current ( [ cor ] ) .
@xmath135= & -\frac{\text{im}(m_2\mu)}{\lambda}\left ( u_1 u_2^\mu+u_2 u_1^\mu\right)\nonumber\\ \text{im } \left[\left(m_2 u_1+\mu\ , u_2\right)\mathcal{a}_{12}^\mu\right]= & -\delta\frac{\text{im}(m_2\mu)}{\lambda}\left ( u_1 u_2^\mu+u_2 u_1^\mu\right)\nonumber\\ \text{im}\left [ \left(m_2 u_1+\mu\ , u_2\right)\mathcal{b}_{12}^\mu\right ] = & -\ \text{im}(m_2\mu ) \left ( u_1 u_2^\mu - u_2 u_1^\mu\right)\end{aligned}\ ] ] we can see from ( [ coeficientes ] ) that the combination @xmath136 appears in the off - diagonal coefficients of @xmath137
. had we neglected the latter we would have obtained only the orthogonal combination @xmath138 as claimed in ref .
however , contrary to the claim made by these authors , we do not find that the off - diagonal terms lead to a contribution of second - order to the currents relevant for electroweak baryogenesis . since this is a crucial difference between our results and those obtained in refs .
@xcite , we will be more explicit about the origin of the discrepancy with those authors at the end of this section . returning to the derivation of the currents , the last step would be going from the green function ( [ inicio2 ] ) to the cp - violating current as defined in ( [ cor2 ] ) . to this end
we must rotate to the weak eigenstate basis and project onto the higgsino component .
we must then define the green function @xmath139 and the cp - violating background current ( [ cor2 ] ) is now given by @xmath140 where @xmath141 with @xmath142 being the two by two pauli matrices in flavor space and @xmath143 being the two by two identity matrix .
the current @xmath144 can be given by the expression @xmath145\,\text{im}\,\mathcal{a}_{11}^\mu
\right.\nonumber\\ + & ( g_1 g_2+g_2 g_1)\ \text{im } \left[\left(m_2 u_1+\mu\ , u_2\right)\mathcal{a}_{12}^\mu\right ] \nonumber\\ + & \left .
2\,p^\mu\left(g_{1\,\nu } g_2-g_1 g_{2\,\nu}\right)\ \text{im}\left
[ \left(m_2 u_1+\mu\ , u_2\right)\mathcal{b}_{12}^{\,\nu}\right ] \phantom{\frac{1}{1}}\right\}\end{aligned}\ ] ] where the coefficients @xmath146 , @xmath147 and @xmath148 of the green function in ( [ coeficientes ] ) are made explicit .
the final expression can be written as , @xmath149 ( u_1 u_2^\mu+ u_2 u_1^\mu ) \right.\nonumber\\ & \nonumber\\ + & \left .
2\,p^\mu \ \left ( g_{1\,\nu } g_2-g_1 g_{2\,\nu}\right)\ ( u_1 u_2^\nu- u_2 u_1^\nu)\right\}\end{aligned}\ ] ] the left - handed current @xmath150 can be obtained in the same way .
the corresponding result can be read off from ( [ cor5 ] ) after making the changes @xmath151 and @xmath152 . from them
the vector and axial currents can be computed and the result obtained from ( [ cor5 ] ) agrees with that presented in eqs .
( 3.14 ) and ( 3.15 ) of ref .
@xcite . to conclude this section we would like to make some final comments concerning the origin of our discrepancy with the results presented in ref
. @xcite . in ref .
@xcite it is correctly argued that , in the absence of particle - changing interactions and in the mass eigenstate basis at a given point @xmath18 , terms proportional to @xmath153 only appear in the off - diagonal elements of the two - by - two matrix of green functions .
this is in agreement with the result obtained in eq .
( [ inicio2 ] ) . from here , the authors of ref .
@xcite conclude that the sources receive no contribution proportional to @xmath154 at first order in the derivative expansion .
this conclusion is not correct as we will now explain .
first of all , let us emphasize that the particle changing interactions are essential in order to convert the original higgsino density into a left - handed quark one . in the absence of these interactions with the plasma the net result for
the baryon asymmetry will be much smaller than the one required for consistency with big - bang nucleosynthesis .
the higgsino states can not be associated with the mass eigenstates at any point @xmath18 along the bubble wall , where the higgs background is non - vanishing . in order to treat the particle changing interactions
one has to transform to the weak interaction basis .
second , observe that the currents are completely determined by @xmath155 ( @xmath156 ) .
the dominant , diagonal contribution @xmath121 ( @xmath120 ) plays no role in the definition of the cp - violating densities at this order .
the cp - violating densities receive contributions from both the diagonal and off - diagonal terms after rotating to the weak eigenstate basis as in ( [ sb ] ) and both contributions appear at first order . therefore , one should compute both the diagonal and off - diagonal contribution to the green functions .
the fact that the diagonal terms have no dependence on the combination @xmath157 does not imply that the sources share such an independence , since the off - diagonal terms in the mass eigenstate basis play an important role when interactions are included . actually , as shown in ref .
@xcite , depending on the parameter space , the contributions proportional to @xmath158 may be the dominant ones in the generation of a relevant baryon asymmetry of the universe .
similar arguments can be used to show that the considerations of the authors of ref .
@xcite , leading to the absence of a dependence of the sources on @xmath157 can not be sustained . indeed ,
@xcite tries to provide a formal proof to these considerations .
moreover , the authors of ref .
@xcite neglect the chargino mixing in the interaction terms , in spite of the fact that mixing is essential for a non - vanishing source .
therefore , their treatment relies on an approximation that is invalid in the interesting region where @xmath159 is close to @xmath160 . as we showed in ref .
@xcite it is precisely in this region of parameters where the sources proportional to @xmath161 become most relevant in the baryon asymmetry generation ( see discussion below ) .
the detailed evaluation of the currents necessary to compute the sources for the diffusion equations presented in the last section was already provided in a previous article @xcite . in ref .
@xcite , the resulting baryon asymmetry was computed under the assumption that the connection between the sources for the diffusion equations and the cp - violating currents was that presented in refs .
@xcite , namely @xmath162 . in section 2
we showed that , contrary to this assumption , the real sources , eq . ( [ source ] ) , have a form that depends on derivatives of the cp - violating currents computed in the presence of the background fields .
therefore , the resulting value of the baryon asymmetry is different from the one computed in ref .
@xcite .
the baryon asymmetry resulting from the solution of the diffusion equations depends on integral functions along the wall of the temperature dependent value of the sources times exponential factors which are slowly varying along the bubble wall @xcite .
therefore , the new values can be simply obtained from the earlier ones by integration by parts , taking care of the proper boundary conditions . as discussed in the introduction ,
the total baryon asymmetry depends on two different contributions .
the first one is proportional to @xmath163 this expression , eq .
( [ asymm ] ) , is proportional to the variation of the angle @xmath164\ ] ] at the wall of the expanding bubble , which tends to zero for large values of @xmath165 .
furthermore , independently of @xmath165 , for large values of @xmath166 , @xmath167 varies only slightly from its value @xmath168 and therefore the baryon asymmetry tends to be suppressed . finally , this contribution to the sources of the baryon asymmetry has a resonant behaviour for @xmath169 .
therefore , it becomes more relevant for moderate values of @xmath165 and @xmath166 and for values of @xmath170 .
the second contribution to the baryon asymmetry depends on @xmath171 similarly to ( [ asymm ] ) , the first term in the expresion ( [ symm ] ) is suppressed for large values of @xmath165 and/or @xmath166 .
although the second term is also suppressed for large values of @xmath166 , it is unsuppressed for large values of @xmath165
. therefore , we expect this contribution to become the dominant one for large values of @xmath165 , particulary in the non - resonant regions with values of @xmath159 very different from @xmath172
. finally , let us remark that the final baryon asymmetry is obtained by solving the diffusion equation for the total baryon number .
this equation may be easily derived by summing up the diffusion equations for the thirty six chiral quarks of the standard model ( three generations , two flavors of quark per generation , two chiralities and three colors per quark ) .
the sum of the quarks densities is just three times the total baryon number density ( each quark carries a baryon number @xmath173 ) .
the variation of the baryon number is governed by the sphaleron processes , which affect only the left - handed chiral quarks and leptons .
the left - handed quark chemical potentials receive two contributions .
the dominant one may be obtained by solving the diffusion equation for the different colors of quarks in the presence of gauge , yukawa , mass and strong sphaleron interactions . since no baryon number violating processes
are included , the solutions to these diffusion equations lead to an equal number of baryons of a given chirality and antibaryons of the opposite chirality . these dominant densities
may be considered as approximately constant during the characteristic long times in which the weak sphaleron processes take place .
the left - handed densities receive also a subdominant contribution coming from the weak sphaleron interactions .
this contribution is associated with a net - baryon number which , considering effective mixing between the different flavors and colors of quarks , is shared in approximately equal parts by all 36 of them . finally , there is also a net lepton number created which due to charge conservation is equal to the net baryon number . taking all these considerations into account , and considering the right - handed leptons out of equilibrium , one arrives to the equation @xmath174 where @xmath175 ( @xmath176 is the weak coupling constant ) , with @xmath177 @xcite being the weak sphaleron rate , while @xmath178 is equal to the sum over the three generations of the left - handed up and down quark chemical potentials associated with a given color of quarks , as obtained by solving their diffusion equations in the absence of the slow sphaleron interactions .
in the numerical simulations , we have considered heavy all squarks and sleptons except for the lightest stop , in which case the coefficient @xmath179 , which differs only slightly from the sm result @xmath180 ( see also ref .
@xcite ) . in fig .
[ figure1 ] we show the comparison of the baryon number to entropy ratio @xmath181 computed within the present model with the mean value of the one consistent with big bang nucleosynthesis ( bbn ) , @xmath182 @xcite .
we have chosen soft supersymmetry breaking parameters in the stop sector such that they lead to a value of the higgs boson mass consistent with the present experimental constraints , and the lightest stop light enough so that the phase transition becomes strongly first order , @xmath183 @xcite .
since the dominant component of the sources comes from the chargino sector @xcite , the results of the baryon asymmetry depend only indirectly on the stop parameters , through the values of @xmath184 and @xmath185 .
different choices of the parameters lead to variations of the final result by a factor of order one .
since the method of computation has implicit uncertainties of similar order , so far @xmath186 , the specific choice of the parameters does not affect the results for the baryon asymmetry in any significant way .
we have chosen a maximal value of the phase of the @xmath160 parameter , @xmath187 .
therefore , the inverse of the values shown in the figure can be interpreted as the value of @xmath188 necessary to obtain a prediction consistent with bbn .
observe that due to the uncertainty in the value of @xmath189 @xcite , the value of @xmath189 may be a factor 2 smaller than the above quoted mean value and therefore the value of the phase may be a factor 2 smaller than the value obtained by the procedure described above .
hence , from fig .
[ figure1 ] we see that for @xmath165 consistent with the present experimental constraints , values of @xmath190 are preferred . in fig .
[ figure2 ] we present the values of the ratio @xmath191 as a function of @xmath160 , for a fixed value of @xmath192 gev . for small values of @xmath165
the dominant contribution comes from the resonant behavior at @xmath169 .
however , as @xmath165 increases the maximum of @xmath193 is obtained for values of @xmath160 much larger than @xmath159 .
the dominant , non - resonant component has opposite sign to the resonant one and for large values of @xmath160 compared to @xmath159 , it becomes only slightly dependent on @xmath165 , in agreement with our discussion above .
we see that even for large values of @xmath165 , values of the baryon asymmetry consistent with bbn predictions may be obtained for phases of order one .
observe that the baryon asymmetry tends to zero for large values of @xmath160 , since the higgsino component of the lightest chargino , present in the plasma , tends to zero in this limit .
the above results are important in view of the constraints coming from electric dipole moments @xcite .
it has been recently observed that , even in the presence of very heavy squarks and leptons , the electric dipole moment contributions may not be small , due to the existence of two loop diagrams involving chargino loops and the non - sm like higgs bosons @xcite .
these contributions are enhanced with increasing @xmath166 , but become rapidly small for large values of @xmath165 and small mixing in the chargino sector .
the results of fig .
[ figure2 ] show that a relevant baryon asymmetry may be obtained even in this particular regime of parameters .
in this article , we have derived the diffusion equations governing the result for the baryon density generated by the passage of the expanding wall at temperatures close to the electroweak phase transition temperature .
important for this result was the identification of the sources for the diffusion equations in terms of the currents generated by the non - trivial higgs backgrounds , which vary along the bubble wall .
we demonstrated that there are two different contributions to these sources , one proportional to @xmath194 and the other proportional to @xmath195 .
contrary to previous claims , we showed that none of these contributions to the sources vanish .
we clarified the origin of the discrepancy of our results with those of refs .
@xcite .
the two contributions to the sources have a different dependence on the parameters of the minimal supersymmetric standard model .
while the former contribution has a resonant behaviour for @xmath169 and goes to zero for large values of @xmath165 , the latter has no resonant behaviour and remains non - vanishing for large values of @xmath165 .
the final result for the baryon asymmetry shows that consistency with big bang nucleosynthesis demands the relative phase between @xmath160 and the gaugino mass parameter @xmath159 ( or @xmath196 , if we consider the neutralino contribution ) to be sizeable .
our calculations lead to a bound on the value of the cp - violating phase , @xmath197 , but due to the natural uncertainties associated with our simplified treatment of the baryon asymmetry calculation , and the experimental error in the determination of @xmath189 , we can not reliably rule out the possibility that somewhat smaller values of the phase @xmath198 may lead to consistency with the bbn predictions .
on the other hand , for sizeable values of the cp - violating phase @xmath199 , cancellation of the potentially large electric dipole moment contributions is required .
these contributions are dominated by one - loop effects with scalar particles and gauginos participating in the loop .
these corrections may be efficiently suppressed for large values of the scalar masses of the first and second generation @xcite . under these conditions ,
non - vanishing corrections still appear at the two - loop level @xcite .
the most relevant two - loop corrections are enhanced for large values of @xmath166 , but are suppressed for large values of the cp - odd mass parameter @xmath165 .
it is reassuring to observe that even for large values of the cp - odd mass parameter a baryon asymmetry consistent with bbn may be obtained . in summary within the framework of the minimal supersymmetric extension of the standard model , the realization of the electroweak baryogenesis scenario requires the presence of a light higgs boson , with mass smaller than about 120 gev , and a light stop , with mass smaller than the top quark mass @xcite-@xcite , @xcite .
moreover it has been found that this baryogenesis mechanism does not guarantee a new cp violating signal at the @xmath200-factories @xcite .
therefore , apart from the experimental constraints coming from electric dipole moments , a definitive test of the electroweak scenario withih the mssm will come from higgs @xcite and stop @xcite searches at the tevatron and at the lhc .
we acknowledge discussions with t. prokopec and s. weinstock . m.q . and m.s .
would like to thank the theory division of the fermi national accelerator laboratory , where part of this work was done , for hospitality .
m.c . and c.w .
also wish to thank the cern theory division , where part of this work was done , for hospitality .
m.s . would like to thank the theory group of instituto de estructura de la materia , where this work was partly done , for hospitality .
m.q . would also like to thank the physics and astronomy department of the johns hopkins university for the hospitality extended to him as bearden visiting professor .
the work of c.w . is supported in part by the us doe , div .
of hep , contract w-31 - 109-eng-38 .
fermilab is operated by universities research association inc . under contract
de - ac02 - 76ch02000 with the doe .
the work of m.s . is supported by the us doe contract de - a505 - 89er40518 .
in this appendix we summarize some useful formulae which concern the chargino sector of the mssm .
the chargino mass matrix is given by @xmath201 where we have defined @xmath202 .
the diagonalizing matrices are @xmath203 where field redefinitions have been made in order to make the higgs vacuum expectation values , as well as the weak gaugino mass @xmath159 , real , @xmath204 and the mass eigenvalues are given by @xmath205 99 for reviews , see : a.g .
cohen , d.b .
kaplan and a.e .
nelson , _ ann . rev
. nucl . part .
* 43 * ( 1993 ) 27 ; m. quirs , _ helv .
acta _ * 67 * ( 1994 ) 451 ; v.a
. rubakov and m.e .
shaposhnikov , _ phys .
* 39 * ( 1996 ) 461 ; m. carena and c.e.m .
wagner , hep - ph/9704347 ; a. riotto , m. trodden , _ ann .
nucl . part .
* 49 * ( 1999 ) 35 ; m. quirs and m. seco , _ nucl .
81 * ( 2000 ) 63 , hep - ph/9703274 .
sakharov , _ jetpl _ * 91b * ( 1967 ) 24 .
g. thooft , _ phys .
* 37 * ( 1976 ) 8 ; _ phys .
_ * d14 * ( 1976 ) 3432 .
d. bodeker , _ phys .
* b426 * ( 1998 ) 351 [ arxiv : hep - ph/9801430 ] ; p. arnold and l.g . yaffe , hep - ph/9912306 ; p. arnold , _ phys . rev . _ * d62 * ( 2000 ) 036003 ; g.d .
moore and k. rummukainen , _ phys .
rev . _ * d61 * ( 2000 ) 105008 ; g.d .
moore , _ phys .
* d62 * ( 2000 ) 085011 . n.s .
manton , _ phys .
rev . _ * d28 * ( 1983 ) 2019 ; f.r .
klinkhamer and n.s .
manton , _ phys .
_ * d30 * ( 1984 ) 2212 .
a. g. cohen , d. b. kaplan and a. e. nelson , _ phys .
_ * b336 * ( 1994 ) 41 [ arxiv : hep - ph/9406345 ] .
farrar and m.e .
shaposhnikov , _ phys .
* 70 * ( 1993 ) 2833 , ( * e * ) : * 71 * ( 1993 ) 210 and _ phys .
* d50 * ( 1994 ) 774 .
p. hernndez , j. orloff , o. pne and c. quimbay , _ nucl .
_ * b430 * ( 1994 ) 382 ; p. huet and e. sather , _ phys .
* d51 * ( 1995 ) 379 .
k. jansen , _ nucl .
phys . _ * b proc
47 * ( 1996 ) 196 , hep - lat/9509018 ; k. rummukainen , m. tsypin , k. kajantie , m. laine and m. shaposhnikov , _ nucl . phys . _
* b532 * ( 1998 ) 283 ; k. rummukainen , k. kajantie , m. laine , m. shaposhnikov and m. tsypin , hep - ph/9809435 .
giudice , _ phys .
_ * d45 * ( 1992 ) 3177 ; s. myint , _ phys .
* b287 * ( 1992 ) 325 .
espinosa , m. quirs and f. zwirner , _ phys .
_ * b307 * ( 1993 ) 106 .
a. brignole , j.r .
espinosa , m. quirs and f. zwirner , _ phys .
* b324 * ( 1994 ) 181 .
m. carena , m. quirs and c.e.m .
wagner , _ phys .
lett . _ * b380 * ( 1996 ) 81 .
d. delepine , j.m .
grard , r. gonzlez felipe and j. weyers , _ phys .
_ * b386 * ( 1996 ) 183 .
moreno , d.h .
oaknin and m. quirs , _ nucl .
_ * b483 * ( 1997 ) 267 ; _ phys . lett . _ * b395 * ( 1997 ) 234 . j. cline and k. kainulainen , _ nucl
. phys . _ * b482 * ( 1996 ) 73 ; _ nucl .
phys . _ * b510 * ( 1998 ) 88 . m. laine , _ nucl . phys . _
* b481 * ( 1996 ) 43 ; m. losada , _ phys .
rev . _ * d56 * ( 1997 ) 2893 ; g. farrar and m. losada , _ phys .
lett . _ * b406 * ( 1997 ) 60 .
espinosa , _ nucl .
phys . _ * b475 * ( 1996 ) 273 .
b. de carlos and j.r .
espinosa , _ nucl .
* b503 * ( 1997 ) 24 .
d. bodeker , p. john , m. laine and m.g .
schmidt , _ nucl .
* b497 * ( 1997 ) 387 .
m. carena , m. quirs and c.e.m .
wagner , _ nucl .
phys . _ * b524 * ( 1998 ) 3 .
m. laine , k. rummukainen , _ nucl .
* b535 * ( 1998 ) 423 .
moreno , m. quirs and m. seco , _ nucl .
phys . _ * b526 * ( 1998 ) 489 .
j.m . cline and g.d .
moore _ phys .
* 81 * ( 1998 ) 3315 .
p. john , _ phys .
_ * b452 * ( 1999 ) 221 ; s.j .
huber , p. john , m. laine and m.g .
schmidt , _ phys
_ * b475 * ( 2000 ) 104 .
m. losada , _ nucl .
_ * b537 * ( 1999 ) 3 and _ nucl .
_ * b569 * ( 2000 ) 125 ; m. laine and m. losada , _ nucl .
b582 * ( 2000 ) 277 . f. csikor , z. fodor , p. hegedus , a. jakovac , s.d .
katz and a. piroth , _ phys .
* 85 * ( 2000 ) 932 . m. laine and k. rummukainen , _ nucl
. phys . _ * b597 * ( 2001 ) 23 [ arxiv : hep - lat/0009025 ] .
p. huet and a.e .
nelson , _ phys .
* b355 * ( 1995 ) 229 ; _ phys .
* d53 * ( 1996 ) 4578 .
m. carena , m. quirs , a. riotto , i. vilja and c.e.m .
wagner , _ nucl
_ * b503 * ( 1997 ) 387 . j. cline , m. joyce and k. kainulainen , _ phys .
* b417 * ( 1998 ) 79 .
t. multamaki , i. vilja , _ phys .
lett . _ * b411 * ( 1997 ) 301 .
a. riotto , _ int .
phys . _ * d7 * ( 1998 ) 815 , _ nucl . phys . _ * b518 * ( 1998 ) 339 and _ phys . rev . _
* d58 * ( 1998 ) 095009 .
worah , _ phys .
rev . _ * d56 * ( 1997 ) 2010 and _ phys
. rev .
* 79 * ( 1997 ) 3810 .
h. davoudiasl , k. rajagopal and e. westphal , _ nucl .
* b515 * ( 1998 ) 384 .
k. enqvist , a. riotto and i. vilja , _ phys .
* b438 * ( 1998 ) 273 .
m. trodden , _ rev .
* 71 * ( 1999 ) 1463 .
n. rius and v. sanz , _ nucl .
_ * b570 * ( 2000 ) 155 .
m. carena , j. m. moreno , m. quiros , m. seco and c. e. wagner , _ nucl .
* b599 * ( 2001 ) 158 .
k. hagiwara et al , particle data group , _ phys .
* d66 * ( 2002 ) 010001 . j. m. cline and k. kainulainen , _ phys .
* 85 * ( 2000 ) 5519 ; j.m .
cline , m. joyce and k. kainulainen , _ jhep _ * 07 * ( 2000 ) 018 [ arxiv : hep - ph/0110031 ] .
k. kainulainen , t. prokopec , m. g. schmidt and s. weinstock , _ jhep _ * 0106 * ( 2001 ) 031 [ arxiv : hep - ph/0105295 ] ; k. kainulainen , t. prokopec , m. g. schmidt and s. weinstock , arxiv : hep - ph/0202177 .
m. quiros , _ nucl
. phys . _ * b proc
. suppl . 101 * ( 2001 ) 401 [ arxiv : hep - ph/0101230 ] .
see , for example , s. r. de groot , w. a. van leeuwen and c. g. van weert , `` relativistic kinetic theory .
principles and applications '' , north - holland ( 1980 ) .
t. ibrahim and p. nath , _ phys .
_ * b418 * ( 1998 ) 98 ; _ phys .
* d57 * ( 1998 ) 478 ; erratum ibid . *
d58 * ( 1998 ) 019901 ; _ phys .
rev . _ * d58 * ( 1998 ) 111301 ; erratum ibid .
* d60 * ( 1999 ) 099902 ; m. brhlik , g.j .
good and g.l .
kane , _ phys .
rev . _ * d59 * ( 1999 ) 115004 ; a. pomarol and d. tommasini , _ nucl .
_ * b466 * ( 1996 ) 3 ; a.g .
cohen , d.b .
kaplan and a.e .
nelson , _ phys .
_ * b388 * ( 1996 ) 588 ; j. feng , k. matchev and f. wilczek , _ phys .
lett . _ * b482 * ( 2000 ) 388 ; d. chang , w .- y . keung and a. pilaftsis ,
* 82 * ( 1999 ) 900 ; a. pilaftsis , _ phys .
_ * b471 * ( 1999 ) 174 ; d. chang , w .- f . chang and w
keung , _ phys .
lett . _ * b478 * ( 2000 ) 239 ; a. pilaftsis , _ phys .
* d62 * ( 2000 ) 016007 .
d. chang , w. f. chang and w. y. keung , arxiv : hep - ph/0205084 ; a. pilaftsis , arxiv : hep - ph/0207277 .
h. murayama and a. pierce , arxiv : hep - ph/0201261 .
m. carena _ et al .
_ , arxiv : hep - ph/0010338 ; d. cavalli _ et al .
_ , arxiv : hep - ph/0203056 ; m. carena and h.e .
haber , `` review on _ higgs boson theory and phenomenology _ '' , preprint fermilab - pub-02/114-t , scipp 02/07 .
h. baer , j. sender and x. tata , _ phys .
_ * d50 * ( 1994 ) 4517 ; r. demina , j. d. lykken , k. t. matchev and a. nomerotski , _ phys .
rev . _ * d62 * ( 2000 ) 035011 ; c. l. chou and m. e. peskin , _ phys .
_ * d61 * ( 2000 ) 055004 ; s. abel _ et al .
_ [ sugra working group collaboration ] , arxiv : hep - ph/0003154 ; a. djouadi , m. guchait and y. mambrini , _ phys .
_ * d64 * ( 2001 ) 095014 ; m. carena , d. choudhury , r. a. diaz , h. e. logan and c. e. wagner , arxiv : hep - ph/0206167 . | electroweak baryogenesis provides a very attractive scenario to explain the origin of the baryon asymmetry .
the mechanism of electroweak baryogenesis makes use of the baryon number anomaly and relies on physics that can be tested experimentally .
it is today understood that , if the higgs mass is not larger than 120 gev , this mechanism may be effective within supersymmetric extensions of the standard model . in this work ,
we reconsider the question of baryon number generation at the electroweak phase transition within the context of the minimal supersymmetric extension of the standard model .
we derive the relevant diffusion equations , give a consistent definition of the sources , and compare our results with those appearing in the recent literature on this subject . |
study of non - perturbative pair creation from a classical electric field , which is known as the schwinger mechanism @xcite , has a long history and wide range of applications ( see ref .
@xcite for a recent review ) .
one of those applications can be found in studies of relativistic heavy - ion collisions , where the schwinger mechanism has been used as a mechanism of matter formation from a color flux tube @xcite .
the color flux - tube model assumes that a strong color electric field is formed in a beam direction just after two nuclei collide and pass through each other @xcite .
formation of longitudinal color electric fields is also predicted in the framework of color glass condensate @xcite .
therefore , particle production due to the schwinger mechanism attracts renewed interest @xcite . under these circumstances , getting an understanding of how an initial electric field and created particles evolve in time is of prime importance . to properly describe the time evolution ,
calculating vacuum persistence probability or pair creation probability , which were first derived by schwinger , is not sufficient @xcite , and an electric field should be treated as a dynamical variable rather than a background field controlled by hand , i.e. back reaction should be taken into account .
there have been considerable numbers of studies treating back reaction ; the ones based on a kinetic theory @xcite and the others on quantum field theory @xcite . to our knowledge , however , field theoretical treatment of the back reaction problem under a _ color _ electric field has been lacking .
therefore , in this paper we investigate the pair creation of quarks under a color electric field incorporating back reaction .
in studies of physics under non - abelian electromagnetic fields , su(2 ) theory has been often used for simplicity . in the case of su(3 ) , however , a new feature arises : anisotropy in color space .
it has been shown that an su(3 ) color electric field has two independent directions and it is characterized by two gauge invariant parameters : one of them is determined by its field strength and the other is related with the color direction of the field @xcite . more generally , an su(@xmath0 ) color vector has @xmath1-independent directions in color space , and physical contents can generally depend on a color direction of an electric field @xcite . in this paper
, we deal with su(3 ) color electric fields and examine the color direction dependence . not only new features which arise in non - abelian fields
, we also analyze phenomena whose essence is common to the abelian case
. collective motion of created particles which couples to an electric field shows plasma oscillation . during this evolution ,
several phenomena are observed : suppression of pair creation or annihilation of the particles due to the pauli blocking , damping of the electric field , and rapid oscillations in the momentum distribution of the created particles due to interference .
we shall give an analysis of these phenomena to advance an understanding of physics in pair creation .
we take a uniform color electric field as an initial state .
pressure of this initial state is quite anisotropic : the longitudinal pressure is negative and the transverse pressure is positive .
therefore , if local thermalization is achieved starting from the flux - tube initial condition , isotropization of pressure should be needed during the time evolution .
however , the full understanding of a thermalization process in heavy - ion collisions has not been obtained . in this paper
, we examine the role of pair creation for the isotropization of pressure as a first step to understand a mechanism of thermalization in heavy - ion collisions .
one of remarkable differences of the color flux tube given by the color glass condensate from that in the original flux - tube model is the existence of a longitudinal color magnetic field in addition to an electric field @xcite .
it has been shown that a longitudinal magnetic field enhances pair creation of fermions and speeds up the decay of an electric field in the previous paper @xcite .
we extend it to the quark pair creation under a longitudinal color electric and magnetic field .
furthermore , we study induction of chiral charge due to pair creation under a magnetic field .
since the chiral anomaly is a semi - classical effect where the quantum aspect of a gauge field is unnecessary , we can also apply our framework to study the chiral anomaly due to pair creation .
the relation between pair creation and the chiral anomaly has been also studied in refs .
emergence of a nonzero chirality in heavy - ion collisions attracts interest in the context of the chiral magnetic effect @xcite .
the remainder of this paper is organized as follows . in the next section
, we shall explain the abelianization of a color electromagnetic field , and introduce the parameter characterizing the color direction of the field .
although this formalism is essentially the same as that given in ref .
@xcite , we make the existence of color direction dependence clearer with the help of the method in refs .
@xcite . in section [ sec : canonical ] , we introduce time - dependent particle picture to describe the time evolution of the system
. then , we shall show our numerical results in section [ sec : ele ] .
time evolution of momentum distribution functions of created quarks , color current density , electric field strength and pressure of the system are displayed and discussed .
color direction dependence of the results is also examined there . in section [ sec : mag ] , effects of a longitudinal magnetic field , i.e. enhancement of pair creation and induction of chiral charge , are discussed .
quark pair creation incorporated with back reaction is described by the following lagrangian density @xmath2 where @xmath3 is a quark field and color indices @xmath4 are omitted .
we assume for simplicity that each flavor has the same mass @xmath5 , and flavor indices are also omitted .
the number of flavor is set to be @xmath6 throughout this paper .
the covariant derivative and the field strengths are defined in terms of a background gauge field @xmath7 as @xmath8 where @xmath9 is the generator in the fundamental representation of gauge group su(@xmath10 ) , and @xmath11 is the anti - symmetric structure constant ( @xmath12 ) .
the equations of motion for @xmath13 and @xmath3 now read @xmath14 because we treat the gauge field @xmath7 as a classical background field , the charge current operator @xmath15 in eq .
is replaced by its expectation value @xmath16 in the following .
these coupled equations govern the pair creation and its back reaction .
we restrict the background field to spatially homogeneous and abelian - like one , which is expressed as @xmath17 where @xmath18 is an abelianized field strength and is independent of the space coordinates .
@xmath19 is a constant vector indicating a color orientation of the electromagnetic field , and is normalized so that @xmath20 .
this field strength is given by the gauge field @xmath21 the relation between @xmath18 and @xmath22 is just the same as that in the abelian case : @xmath23 under the abelianized gauge field , the covariant derivative reads @xmath24 .
because @xmath25 is an @xmath26 hermitian matrix , we can diagonalize it by a unitary transformation acquire space dependence , and the method used in this paper is no longer applicable.](global gauge transformation ) : @xmath27 .
then , the equations of motion and can be rewritten as @xmath28 \psi ^i = 0 \hspace{20pt } ( i=1,2,\cdots , n_c ) .
\label{dirac2}\end{gathered}\ ] ] these equations have the same form with those in the abelian theory , except the presence of the weight vector @xmath29 .
hence , we can solve the coupled equations and as in the abelian case @xcite .
each quark field @xmath30 couples with the background field @xmath22 via a coupling constant @xmath31 . and
is an usual weight diagram for a fundamental representation of su(3 ) .
each @xmath29 is given by the @xmath32 coordinate of each vertex of the @xmath33-rotated triangle .
, title="fig : " ] + --3pt in the case of su(3 ) , a diagonalized @xmath34 may be expressed as @xmath35 because @xmath36 and @xmath37 are diagonal .
( we represent @xmath38 as @xmath39 where @xmath40 is the gell - mann matrix . )
then , @xmath29 are expressed in terms of @xmath33 as follows @xmath41 the angle @xmath33 is related with the casimir ( gauge ) invariant @xcite . in our parametrization in this paper and
those in refs.@xcite is @xmath42 . ] , the relation is @xmath43 where @xmath44 ^ 2 $ ] is the second casimir invariant for su(3 ) and characterizes a direction of the color electromagnetic field in a gauge invariant way .
fig.[fig : triangle ] is a graphical representation of eqs .. each @xmath29 is given by the @xmath32 coordinate of each vertex of the rotated triangle ( weight diagram ) . owing to the symmetry of the diagram , it is sufficient to take the angle @xmath33 restricted to @xmath45
to describe the time evolution of the system , we introduce an instantaneous particle picture and quantize the field using that particle picture .
of course , a definition of particle in the presence of a pair - creating background is rather ambiguous .
thus , the instantaneous particle picture should be regarded as our working definition to describe the system evolution . to calculate field quantities such as current and energy density , we need to identify and subtract the contribution from the dirac sea at each time .
we can do this just by the normal ordering in terms of the instantaneous particle basis , and then energy conservation is automatically guaranteed @xcite .
therefore , one can interpret defining a particle picture at each time as a means to properly regularize field quantities , and can consider particle number at intermediate time as a byproduct .
nevertheless , we will show time evolution of a particle number or momentum distribution function defined through the instantaneous particle picture in the following , because it behaves in a physically reasonable manner , and helps us understand the dynamics of this system .
first , we consider only the electric fields .
we will introduce the magnetic fields in section [ sec : mag ] .
we set an initial state with no electric field and switch it on at time @xmath46 .
an initial field strength is @xmath47 and its direction is along the @xmath48-axis .
the gauge @xmath49 is chosen so that @xmath50 and @xmath51 .
after switching on , @xmath52 evolves according to eqs . and .
a quantized quark field may be expanded as @xmath53 , \ ] ] where @xmath54 [ @xmath55 is the annihilation operator of a particle [ antiparticle ] with momentum @xmath56 and spin-@xmath57 satisfying the anti - commutation relation @xmath58 , and @xmath59 are classical solutions of the dirac equation .
the superscript in distinguishes the initial condition for @xmath59 : at @xmath60 , @xmath61 [ @xmath62 is identical to the positive [ negative ] energy solution of the free dirac equation .
we set the state to be in - vacuum @xmath63 , where no particle exists initially and which is defined by @xmath64 . at @xmath65 ,
@xmath59 evolve under influence of the electric field and become superposition of a positive and negative energy ( frequency ) state . to describe the pair creation dynamically ,
we introduce a time - dependent particle picture by decomposing the field operator @xmath66 into positive and negative frequency instantaneously : @xmath67 , \label{psi}\ ] ] where @xmath68 [ @xmath69 is a positive [ negative ] frequency solution of the dirac equation under the pure gauge @xmath70 .
instantaneous particle picture is defined by @xmath71 and @xmath72 .
of course , @xmath71 and @xmath72 agree with @xmath73 and @xmath74 at @xmath75 , respectively . the particle picture at time @xmath76 and that of the in - state are related by the time - dependent bogoliubov transformation : @xmath77 b^{i\ , \dagger } _
{ -\mathbf{p } s } ( t ) = \alpha ^{i\ , * } _ { \mathbf{p } s } ( t ) b^{i\ , \mathrm{in } \dagger } _ { -\mathbf{p } s } -\beta ^{i\ , * } _ { \mathbf{p } s } ( t ) a^{i\ , \mathrm{in } } _ { \mathbf{p } s } , \end{matrix } \label{bogo}\ ] ] of which coefficients satisfy @xmath78 and are given by @xmath79 \beta ^i _ { \mathbf{p } s } ( t ) \delta ^3 ( \mathbf{p } -\mathbf{q } + w_i g\bar{a } _ 3 ( t ) ) = \displaystyle{\int } \ !
d^3 x { } _ + \ !
\psi ^{i\ , ( t ) \dagger } _ { \mathbf{p } s } ( x ) { } _ - \ ! \psi ^{i\ , \mathrm{in } } _ { \mathbf{q } s } ( x ) .
\end{matrix}\ ] ] a quark pair distribution function is defined by @xmath80 where @xmath81 is the volume of the system . with the help of eqs .
, we can rewrite @xmath82 in terms of the bogoliubov coefficients : @xmath83 the expectation of the charge current operator , which is regularized by the normal ordering is @xmath84 where the transverse mass @xmath85 and the anomalous distribution @xmath86 |0,\mathrm{in } \rangle \frac{(2\pi ) ^3 } { v } = \mathrm{re } \left [ e^{-2i\omega _ p t } \alpha ^i _ { \mathbf{p } s } ( t)\beta ^i _ { \mathbf{p } s } ( t ) \right ] \label{anom}\ ] ] are introduced . in the anomalous distribution is canceled out in total because both @xmath87 and @xmath88 contain the factor @xmath89 . ] the first term of the last expression of eq . is a conduction current , which is caused by movement of real particles and the second is a polarization current , which is generated by variation of electric dipole density .
we have solved the coupled equation and numerically with the help of eqs . , and .
the results are shown in figs.[fig : ldis1 ] , [ fig : tdis ] and [ fig : current1 ] .
the time evolution of longitudinal momentum distributions with fixed transverse momentum is exhibited in fig.[fig : ldis1 ] and that of transverse momentum distributions with fixed longitudinal momentum is in fig.[fig : tdis ] .
fig.[fig : current1 ] shows the time evolution of color current density , electric field strength and quark number density .
the parameters are set to be @xmath90 or @xmath91 , and @xmath92 . to concentrate our attention at first on general features which are common to the abelian case , only the distributions of blue quarks ,
whose effective coupling to the background is @xmath93 when @xmath94 , are presented .
hereafter , all figures are shown in the dimension - less unit scaled by @xmath95 . [ cols="^,^ " , ] fig.[fig : current_mag1 ] presents the magnetic field dependence of the current , the electric field and the particle number density ( @xmath96 ) . the color current density and the quark number density
are indeed enhanced by the magnetic field , and as a consequence the frequency of the plasma oscillation increases .
we can estimate @xmath97 , which is a typical time scale of back reaction , in the same way as that in section [ subsec : pp ] , and obtain @xmath98 this is a decreasing function of @xmath99 . because only the lowest @xmath100 mode mainly contributes under a strong magnetic field modes . ] , the results in the massless case ( @xmath91 ) and those with nonzero mass ( @xmath101 ) show drastic difference [ fig.[fig : current_mag2 ] ] , whereas in the absence of a magnetic field , there is no considerable difference between them [ fig.[fig : current1 ] ] .
the particle number density is especially remarkable : its time evolution becomes nearly periodic in the massless case .
similar behavior of a particle number has been obtained by iwazaki @xcite in a different way .
emergence of this periodic behavior is consistent with our contention that the polarization current and the interference of distributions are the source of apparent irreversibility of time , because as noted in sections [ subsec : damping ] and [ subsec : interference ] , the zero mode contributes to neither the polarization current nor the interference term .
the magnetic field has another effect : induction of chiral charge . because the lowest landau level ( @xmath102 ) is occupied only by particles with spin parallel to the magnetic field ( @xmath103 ) , the balance between spin-@xmath104 and @xmath105 is broken and thus chiral charge is induced .
we consider the chiral charge density given by the following equation @xmath106 in the chiral limit ( @xmath107 ) , the chiral charge equals the difference between the number of right - handed particles and that of left - handed particles . substituting eq . into eq . and subtracting a divergent part by the normal ordering , we can express the chiral charge density in terms of the distribution functions : @xmath108 .\ ] ] note that only the lowest mode contributes to @xmath109 .
contribution from the higher modes is totally canceled out because they are occupied with both the spin-@xmath104 and @xmath105 modes .
when @xmath110 , the anomalous distribution also contributes to the chiral charge .
the time evolution of the chiral charge obeys the adler - bell - jackiw anomaly equation @xcite .
the expectation of the anomaly equation extended to the qcd case now reads @xmath111 where the term @xmath112 describes explicit breaking of the chiral symmetry due to the mass term , and @xmath113 .
\end{split}\ ] ] ccc . ] & . ] & . ] + ( a)@xmath91 & ( b)@xmath101 & in fig.[fig : chiral ] , each term of the time integral of eq . is plotted .
equation is indeed satisfied . due to plasma oscillation
, the chiral charge density shows oscillation .
as noted in the previous subsection , time evolution is nearly periodic in the chiral limit [ fig.[fig : chiral](a ) ] .
this curve of @xmath114 becomes exactly a sine one in the limit of a strong magnetic field @xmath115 @xcite .
in contrast , if @xmath110 , the term @xmath116 grows , and the periodicity seen in @xmath114 is destroyed .
we have studied the time evolution of the system where quantum quark fields and a classical color electric field are dynamically coupled with each other . by the procedure of abelianization , we have reduced the equations of motion to the same form as in the abelian case .
effective coupling strength of each quark with the abelianized electromagnetic field depends on the direction of the field in color space .
therefore , the longitudinal momentum distributions of created quarks greatly depend on the color direction of the gauge field [ fig.[fig : ldis2 ] ] .
however , we have found that the color current density and total particle density are rather insensitive to the color orientation of the field since those quantities are obtained after summing up all the color components [ fig.[fig : current2 ] ] .
we have obtained the oscillating behavior of the longitudinal momentum distributions , and estimated its typical time scale @xmath97 [ eq . ] using the empirical analytic expression for the quark distributions .
this @xmath97 also gives a time scale of net particle production .
the increase of the quark number density is concentrated at initial time and gets saturated after @xmath97 [ fig.[fig : current1](c ) ] .
let us roughly estimate @xmath97 with the parameters motivated by the color glass condensate ; @xmath117 and @xmath118 , where @xmath119 is a saturation scale . then eq . yields @xmath120 note in passing that the @xmath33 dependence of @xmath97 is also small , 2% effect , at most , in the present case .
this @xmath97 is not drastically small compared with the typical time scale of the initial stage of heavy - ion collisions , such as thermalization time expected from hydrodynamic calculations @xmath121 fm ( e.g. @xcite ) .
however , this estimation incorporates only the effect of quark pair creation .
if gluon pair creation is taken into account , @xmath97 would become much smaller .
pair creation of gluons is stronger than that of quarks because ( i ) gluons are boson , so that they are bose - enhanced rather than pauli - blocked , and ( ii ) effective coupling of gluons to an electric field is larger than that of quarks . for gluons ,
a rotating hexagon is obtained , of which vertices are on a circle with a radius of 1 being bigger than @xmath122 in the quark case .
] however , applying the method in the present paper to gluon pair creation , we encounter a difficulty of infrared divergence . furthermore , under a magnetic field
, one particle energy of a gluon in the lowest landau level becomes imaginary and instability occurs , which is known as the nielsen olesen instability @xcite .
thus , we need to resolve these difficulties in order to treat both the quark pair and gluon pair creations in our framework in a unified manner .
in addition to the plasma oscillation , we have pointed out that the response of vacuum to the electric field involves several phenomena such as the pauli blocking , damping of the electric field and rapid oscillations in the quark distribution . in particular
, it has been revealed that the cause of the rapid oscillations is the interference between the particles .
because of this interference , the distribution becomes sensitive to the phase of the bogoliubov coefficients .
furthermore , we have found that apparent irreversibility of time evolution is induced by nonzero transverse mass modes through the polarization current and the interference .
also the time evolution of pressure of the system has been calculated .
we have shown that the initial anisotropy in pressure is moderated by pair creation and back reaction even though our treatment is a mean field one .
if effects of collisions among particles are taken into account , in other words , if quantum gauge fields are introduced , the system would be more isotropized .
it is hoped to investigate these effects , as well as effects of non - uniformity of electromagnetic fields .
finally , we have discussed the effects of a magnetic field on quark pair creation .
we have shown that a magnetic field enhances pair creation because of the emergence of the landau level and the spin - magnetic field interaction . as a result ,
the longitudinal magnetic field speeds up the decay of the electric field .
this mechanism may have significance in the context of heavy - ion collisions .
the author would like to thank professors t. matsui and h. fujii for enlightening discussions and comments on the manuscript .
he also acknowledges professor a. iwazaki for helpful discussions , and professor t. hirano for useful comments .
this work is supported by jsps research fellowships for young scientists . | the time evolution of a system where a uniform and classical su(3 ) color electric field and quantum fields of quarks are dynamically coupled with each other is studied focusing on non - perturbative pair creation and its back reaction .
we characterize the color direction of the electric field in a gauge invariant way , and investigate its dependence .
momentum distributions of created quarks show plasma oscillation as well as quantum effects such as the pauli blocking and interference .
pressure of the system is also calculated , and we show that pair creation moderates degree of anisotropy of pressure . furthermore , enhancement of pair creation and induction of chiral charge under a color magnetic field which is parallel to the electric field are discussed .
ut - komaba/10 - 1 |
the case for a @xmath0 black hole ( bh ) coincident with the radio source sagittarius a * in the galactic center ( gc ) is now compelling ( e.g. , schdel et al .
2002 ; ghez et al .
this only emphasizes the long - standing puzzle that the luminosity from the gc is remarkably low given the presence of a massive black hole .
the resolution of this puzzle must lie in how gas from the ambient medium accretes onto the central bh . in these proceedings
i review accretion models and their application to sgr a*. a unique feature of the galactic center is our ability to constrain the dynamics of gas quite close to the black hole ( relative to other systems ) , thus providing additional boundary conditions on , and much less freedom for , theoretical models .
a canonical formulation of these constraints is the bondi accretion estimate for the rate at which the bh gravitationally captures surrounding gas ( bondi 1952 ; see melia 1992 for an early application to sgr a * ) .
given relatively uniformly distributed matter with an ambient density @xmath4 and an ambient sound speed @xmath5 , the sphere of influence of a bh of mass @xmath6 extends out to @xmath7 .
the accretion rate of this gas onto the central bh , in the absence of angular momentum and magnetic fields , is then @xmath8 .
_ chandra _ observations of the gc detect extended diffuse emission within @xmath9 of the bh ( baganoff et al .
this emission likely arises from hot gas produced when the stellar winds from massive stars in the gc collide and shock ( e.g. , the he i cluster ; krabbe et al .
1991 ) . interpreted as such
, the inferred gas density and temperature are @xmath10 @xmath11 and @xmath12 kev on @xmath13 scales , and @xmath14 @xmath11 and @xmath15 kev on @xmath16 scales ( see also fig .
[ fig:3 ] ) .
the corresponding bondi accretion radius is @xmath17 and the bondi accretion rate is @xmath18 yr@xmath19 . ; najarro et al .
1997 ) , implying that there should also be a global outflow of hot gas from the central parsec ( a gc wind ) . ]
if gas were accreted at this rate onto the bh via a geometrically thin , optically thick accretion disk ( shakura & sunyaev 1973 ) , a model that has been extensively and successfully applied to luminous accreting sources ( e.g. , kortakar & blaes 1999 ) , the expected luminosity would be @xmath20 ergs s@xmath19 , larger than the observed luminosity by a factor of @xmath21 .
this is the strongest argument against a thin disk in sgr a*. an additional argument is the absence of any disk - like blackbody emission component in the spectrum of sgr a*.
if the putative disk were to extend all the way down to the bh , its accretion rate would have to be @xmath22 yr@xmath23 to satisfy infrared limits ( e.g. , narayan 2002 ; see his fig .
one possible caveat to the bondi analysis is that there is far more ( by mass ) cold molecular gas than hot x - ray emitting gas in the central @xmath24 parsecs of the gc ( e.g. , herrnstein & ho 2002 ) .
it is unclear how close to the bh the molecular gas extends and whether it is important for the dynamics of gas accreting onto the bh . in what follows
i ignore this component , but see nayakshin ( 2003 ) for a different view . as emphasized above , the inferred low efficiency of sgr a * is the strongest argument against accretion proceeding via a thin accretion disk . instead
, the observations favor models in which very little of the gravitational potential energy of the inflowing gas is radiated away .
i will refer to such models as radiatively inefficient accretion flows ( riafs ) . in the next section ( 2 )
i summarize the properties of riafs .
i then apply these models to the gc ( 3 ) , emphasizing the interpretation of radio and x - ray observations of sgr a*. finally , i conclude with a brief summary ( 4 ) .
riafs describe the dynamics of rotating accretion flows in which @xmath25 , i.e. , very little energy generated by accretion is radiated away ( e.g. , ichimaru 1977 ; rees et al .
1982 ; narayan & yi 1994 ) .
instead , the gravitational potential energy released by turbulent stresses in the accretion flow is stored as thermal energy . as a result , the accreting gas is very hot , with a characteristic thermal energy comparable to its gravitational potential energy ; close to the bh this implies @xmath26 k. at such temperatures , and for gas densities appropriate to systems like the gc , the coulomb collision time is much longer than the time it takes gas to flow into the bh .
the accretion flow then develops a two - temperature structure with the protons likely hotter than the electrons : @xmath27 k @xmath28 k. the precise electron temperature is uncertain but important since electrons produce the radiation that we see .
the electron temperature depends on how and to what extent they are heated by processes such as shocks , mhd turbulence , and reconnection ( see , e.g. , quataert & gruzinov 1999 ) .
note that because collisions are unimportant one would not expect the electron distribution function to be thermal .
advection - dominated accretion flows ( adafs ) are a simple analytical model for the dynamics of riafs ; they predict that the structure of the flow is in some ways similar to spherical bondi accretion , despite the fact that angular momentum and viscosity are important ( e.g. , ichimaru 1977 ; narayan & yi 1994 ) . in adaf models
the gas rotates at @xmath29 , where @xmath30 is the keplerian rotation rate . because the flow is hot , pressure forces are also important and the inflowing gas is geometrically quite `` thick , '' with a scale height @xmath31 at every radius .
the radial velocity in the flow is given by @xmath32 where @xmath33 is the dimensionless viscosity parameter , @xmath34 is the sound speed in the flow and @xmath35 .
conservation of mass on spherical shells then implies that the density scales as @xmath36 , the characteristic scaling for spherical accretion .
_ adaf models also predict that , even in the presence of rotation , the rate at which gas accretes onto the bh from an ambient medium is comparable to the bondi accretion rate : @xmath37_. ( e.g. , narayan 2002 ) .
the factor of @xmath33 arises because the inflow velocity of gas at the bondi accretion radius is @xmath38 in adaf models while it is @xmath39 in bondi models .
thus for a fixed density in the ambient medium , the accretion rate in an adaf will be smaller by a factor of @xmath33 . ] _ thus in adaf models the low luminosity of sgr a * is due to a very low radiative efficiency @xmath40 .
_ with the advent of global , time - dependent , numerical simulations of accretion flows , it has become possible to numerically simulate riafs and test the adaf predictions . note that riafs are , in one sense , the easiest flows to simulate since ( 1 ) no treatment of radiation or radiative transfer is needed and ( 2 ) the flow is `` thick '' with @xmath41 , so there is no difficult - to - simulate separation of scales like in a thin disk . on the other hand , for a system like the gc ,
the proton - electron collision time close to the bh is @xmath42 orders of magnitude longer than the inflow time of the gas .
thus the fluid approximation used by all simulations to date is suspect ( a kinetic treatment should be used ; see quataert et al . 2002 ) . whether this introduces qualitative or merely quantitative errors in the results
is unknown .
i suspect the latter .
the key result from nearly all simulations to date is that _ @xmath43 , i.e. , very little mass available at large radii actually accretes onto the black hole _ ( e.g. , stone , pringle , & begelman 1999 ; igumenshchev & abramowicz 1999 ; 2000 ; igumenshchev et al .
2000 ; stone & pringle 2001 ; hawley & balbus 2002 ; igumenshchev et al .
another way to state this result is that the radial density profile in the flow is much `` flatter '' than the adaf predictions : for a given gas density at large distances from the black hole ( e.g. , measured by _
chandra _ on 1 scales in the gc ) , the density close to the bh is much _ less _ than the adaf or bondi predictions . following a proposal due to blandford & begelman ( 1999 ) , we can parameterize the density profile of riafs with a parameter @xmath44 , where @xmath45 . with this parameterization ,
the rate at which gas is actually accreted into the bh is @xmath46 , where @xmath47 is the inner radius of the flow and @xmath48 is the outer radius .
values of @xmath49 , rather than @xmath50 predicted by adaf models , are favored by the simulations .
_ this implies that , within the context of riaf models , a low accretion rate , @xmath51 , rather than just a low efficiency , is a major factor in the faintness of sgr a*. _
since the focus of this review is the application of riaf models to sgr a * , i will not dwell on why the simulations differ so significantly from the adaf predictions .
a brief discussion is , however , in order .
the following material will not be used in later sections so uninterested readers can move directly to 3 . in riafs ,
the inflowing gas is heated at a rate @xmath52 , as required by the release of gravitational potential energy in a differentially rotating accretion flow . in adaf models
this energy is stored as thermal energy and carried into the bh . as noted above
, the inflowing gas is then very hot with a sound speed at any radius comparable to the escape speed from the bh s potential well .
adaf models are therefore prone to developing outflows ( see narayan & yi 1994 or blandford & begelman 1999 for a more formal discussion ) .
this led blandford & begelman ( 1999 ) to propose that , in the absence of radiation , the gravitational binding energy of the accreted gas must be lost through some other ( non - radiative ) means .
otherwise the inflowing gas is not sufficiently bound to the bh to accrete .
the numerical simulations to date are broadly consistent with this hypothesis ; e.g. , the gas temperature in the simulations is generally a factor of few-5 less than in adaf models implying that the gas is indeed more strongly bound to the bh
. the non - radiative energy loss can take one of two forms : ( 1 ) efficient turbulent transport of energy through the accretion flow to large radii or ( 2 ) a global outflow ( wind ) that carries away the binding energy of the accreted matter .
in hydrodynamic simulations ( e.g. , stone et al .
1999 ; igumenshchev & abramowicz 1999 ) or in mhd simulations with relatively weak magnetic fields ( @xmath53 , where @xmath54 is the ratio of the gas pressure to the magnetic pressure ) , convective transport of energy and angular momentum dominates the dynamics of the accretion flow ( narayan et al . 2002
; igumenshchev et al . 2003 ; see quataert & gruzinov 2000a and narayan et al .
2000 for such convection - dominated accretion flow models ) .
the convective luminosity through a spherical shell at radius @xmath55 is @xmath56 where @xmath57 is the turbulent convective velocity . since the flow is hot , @xmath58 .
a constant flow of gravitational binding energy from small to large radii therefore requires @xmath59 ; i.e. , @xmath60 instead of @xmath50 as in adaf models . or from the surface layers of the accretion flow .
thus there is not necessarily a clean distinction between global energy transport by turbulence and mass outflow as mechanisms for non - radiative energy loss .
both processes are related and , in particular , the former can drive the latter . ]
in contrast to the hydrodynamic results , in mhd simulations with strong magnetic fields ( @xmath61 ) , mhd turbulence dominates the flow dynamics and convection is unimportant ( e.g. , stone & pringle 2001 ; hawley & balbus 2002 ; igumenshchev et al .
stone & pringle ( 2001 ) and hawley & balbus ( 2002 ) find that most of the inflowing gas is lost to a magnetically driven wind .
igumenshchev et al .
( 2003 ) , on the other hand , find a much more complex flow configuration , though still with a very small accretion rate .
a number of authors have used riaf models , and in particular adaf models , to explain the observed properties of sgr a * ( e.g. , narayan et al . 1995 , 1998 ; manmoto et al . 1997 ) .
these calculations have shown that an adaf model accreting at the observationally inferred rate , @xmath62 , can roughly account for the observed luminosity and spectrum of sgr a*. the key constraint is that the fraction of the turbulent energy that heats the electrons , @xmath63 , must be small @xmath64 , so as to not overproduce the observed luminosity .
equivalently the electron temperature close to the bh must be @xmath65 k @xmath66 k. an example of such a model in shown by the dotted line in figure [ fig:1 ] .
the model roughly reproduces the observed sub - mm emission , satisfies the ir limits , and produces an x - ray luminosity comparable to that seen by _
chandra _ in the quiescent ( non - flaring ) state ( i discuss the _ chandra _ observations in more detail below ) .
it does , however , significantly underproduce the lower frequency radio emission .
since the lower frequency radio emission in sgr a * is phenomenologically similar to that of other agn , a natural interpretation is that a jet is present and produces the radio emission ( e.g. , falcke & markoff 2000 , yuan et al .
2001 , and references therein ) . alternatively
, the results in figure [ fig:1 ] assume purely thermal electrons , for which there is no good justification . as figure [ fig:2 ] shows , even a small population of nonthermal electrons in the accretion flow can produce the radio emission ( e.g. , mahadevan 1998 ; ozel et al .
2000 ) . and @xmath67 .
there is also a power law distribution of electrons with @xmath68 and @xmath69 of the electron thermal energy .
power - law electrons produce the low freq .
radio emission not accounted for in fig .
[ fig:1 ] . _ dashed line _ : inverse - compton model for the _ chandra _ x - ray flare ( see text for details ) .
_ solid line : _ total emission .
x - ray error bars are ( from top to bottom ) : oct
. 2000 flare , average flare , & quiescent emission .
this figure is based on yuan et al .
( 2003 ) . ]
quataert & narayan ( 1999 ) showed that a much broader class of riaf models could also account for the observed properties of sgr a*. specifically , the low accretion rate models favored theoretically ( 2 ) can reproduce the observations as well .
the key requirement is that the electrons must be hotter so as to produce more emission even though @xmath70 and the gas density are lower .
an example of such a model is shown by the solid line in figure [ fig:1 ] : @xmath71 from @xmath72 down to @xmath73 , implying that the accretion rate into the bh is much smaller than the bondi rate ( @xmath74 ) .
the electron temperature close to the bh is @xmath75 k , rather than @xmath76 k as in adaf models .
to first approximation , the two models in figure [ fig:1 ] reproduce the observed spectrum of sgr a * equally `` well '' ( or poorly , depending on one s vantage point ) .
_ however , aitken et al .
( 2000 ) and bower et al.s ( 2003 ) detection of @xmath77 linear polarization in the sub - mm ( 230 ghz ) emission from sgr a * argues strongly for low accretion rate models with @xmath78 . _ beckert & falcke ( 2002 , 2003 ; see also ruszkowski & begelman 2002 ) present detailed models for the radio polarization of sgr a * ( both linear and circular ) . here
i summarize the constraints imposed by the linear polarization detection : accretion at @xmath79 implies a much higher gas density and magnetic field strength close to the black hole than accretion at @xmath80 .
faraday rotation is therefore much stronger ( e.g. , quataert & gruzinov 2000b ; agol 2000 ) . in models with @xmath81 ,
the rotation measure is @xmath82 in the region of the flow ( @xmath83 ) where the sub - mm emission is produced .
this leads to a faraday rotation angle @xmath21 rad at @xmath84 ghz .
this large rotation angle implies that intrinsically linearly polarized synchrotron emission would be depolarized propagating to the observer over most of the radio to infrared spectrum . by contrast , bower et al .
( 2003 ) find that @xmath85 in their linear polarization detection at @xmath86 ghz .
models with @xmath87 can satisfy this constraint because the density and magnetic field strength close to the bh are much smaller .
for example , for @xmath88 , the net rotation measure through the accretion flow is @xmath89 , consistent with the observational constraint .
thus the observed detection of linear polarization in the sub - mm emission of sgr a * argues for a low accretion rate @xmath90 .
one way out of this conclusion is to posit that the magnetic field undergoes so many reversals along the line of sight that the net faraday rotation is @xmath91 times smaller than these simple estimates ( ruszkowski & begelman 2002 ) .
i regard this as very unlikely , but am not aware of a direct observational argument against this possibility at the present time . in light of the above arguments the rest of my discussion centers on models with @xmath87 :
these concordance models are both theoretically favored and satisfy the rotation measure constraint from the linear polarization detection .
i focus on interpreting the _ chandra _ x - ray observations of sgr a*. much of this material is based on quataert ( 2002 ) and yuan , quataert , & narayan ( 2003 ) .
_ chandra _ observations reveal that there are two components to the x - ray emission coincident with sgr a * ( baganoff et al .
2001 , 2003ab ) : ( 1 ) a baseline x - ray flux with a nearly constant luminosity @xmath92 ergs s@xmath19 and a soft spectrum ( photon index @xmath93 , where @xmath94 ) .
this component is clearly extended with a size of @xmath95 and does not vary in time .
( 2 ) x - ray flares occurring at a rate of @xmath96 day@xmath19 and lasting @xmath97 s. the luminosity increases by a factor of few@xmath98 during the flare and the spectrum is quite hard ( @xmath99 ) .
the flares are not extended ; in fact , the observed timescales argue that they arise close to the bh , at @xmath83 .
the x - ray flares are the most dramatic result from the _ chandra _ observations .
markoff et al . (
2001 ) showed that the flares are probably due to electron heating or acceleration , rather than a change in the accretion rate onto the bh ( otherwise there is too much variation in other wavebands ) .
an obvious analogue is magnetic reconnection in solar flares , which one could readily imagine occurring in the inner part of the accretion flow close to the bh .
given an injection of energy into the electrons , there are three emission mechanisms that could , _ a priori _ , give rise to flares : ( 1 ) bremsstrahlung , ( 2 ) synchrotron , and ( 3 ) inverse compton .
bremsstrahlung is attractive because it can naturally explain the very hard spectrum of the flares .
the problem is that , to produce a luminosity of @xmath100 ergs s@xmath19 from a sphere of radius @xmath55 , the gas density must be @xmath101 @xmath11 , where @xmath102 . for comparison ,
the ambient density in the inner @xmath103 for a model with @xmath104 is @xmath105 @xmath11 .
equally importantly , bremsstrahlung emission in riafs is dominated by very large radii @xmath106 , not small radii ( e.g. , quataert & narayan 1999 ; see below ) .
thus bremsstrahlung appears unlikely to be responsible for the x - ray flares .
two possible ways out of this conclusion are ( 1 ) to posit that the gas density increases by a factor of @xmath107 during the flare , and does so preferentially in the very inner part of the accretion flow ( e.g. , liu & melia 2002 ) .
it is unclear , however , what would drive such large density changes , particularly since the cooling time is so long that thermally driven instabilities are unlikely to be important , ( 2 ) perhaps the accreting gas is a two - phase medium , with a cooler , denser phase giving rise to the x - ray flares ( yuan et al .
2003 ) .
in contrast to bremsstrahlung , synchrotron emission can readily account for the observed flares : if @xmath108 of the electrons ( by energy ) are accelerated into a power - law tail in the inner @xmath109 , the radio - ir emission is essentially unchanged while electrons with lorentz factors @xmath110 can produce the x - ray emission ( e.g. , markoff et al . 2001 ) .
moreover , there is a natural explanation for the hard x - ray spectrum seen . for riaf models with @xmath104 ,
the magnetic field strength close to the bh is @xmath111 g. the associated synchrotron cooling time for electrons emitting in the _ chandra _ band is @xmath112 s. thus , unless @xmath113 g in the emitting region , the cooling time is less than the duration of the flare and there should be a cooling break in the electron distribution function below the _ chandra _ band . for an injected distribution of power - law electrons with @xmath114 , where @xmath115 , the distribution function for the population of cooled electrons is @xmath116 .
this implies a synchrotron spectrum with @xmath117 , consistent with the typical flare observed by _
( baganoff et al .
2003b ) .
synchrotron self - compton emission can also explain the x - ray flares observed by _
the dashed line in figure [ fig:2 ] shows a concrete example in which most of the electrons in a @xmath118 volume are accelerated into a power law distribution .
this population of electrons produces synchrotron emission and also upscatters synchrotron photons to produce a hard x - ray flare .
note that there is very little change to the radio or ir emission ( the dot - dashed line in fig .
[ fig:2 ] shows the baseline non - flaring model in which power - law electrons have only @xmath69 of the electron thermal energy ) . finally , it is important to stress that in the models discussed here , the duration of the flare is set by a dynamical or viscous timescale in the inner @xmath119 of the accretion flow .
by contrast , there is no explanation for the mean time between flares , @xmath96 day .
in addition , it is difficult to apply the ideas considered here to the week - long , factor of few , sub - mm flares observed by tsuboi et al .
( 1999 ) and zhao et al .
in particular the duration of these flares is inconsistent with the characteristic timescales in the sub - mm emitting region ( @xmath120 ) .
one possibility is that they are not directly related to the _ chandra _ flares but are instead due to small fluctuations in @xmath70 set by dynamics in the accretion flow at larger radii .
the steady quiescent emission observed by _
chandra _ is qualitatively different from the flaring emission . in particular , it is softer ( @xmath93 ) and extended ( @xmath95 ) .
the latter fact implies that it is a different emission component since synchrotron and inverse compton emission are produced at small radii .
riaf models naturally predict that the thermal bremsstrahlung emission is dominated by large radii in the flow ( e.g. , quataert & narayan 1999 ; ozel & di matteo 2000 ) . given a density profile of the form @xmath45 and a temperature profile @xmath121 ( valid at large radii ) , the bremsstrahlung luminosity is dominated by large radii : @xmath122 , assuming photon energies @xmath123 .
the resulting spectrum , adding up all radii , is @xmath124 , i.e. , @xmath125 .
thus a natural interpretation of the quiescent flux coincident with sgr a * is that it is bremsstrahlung from hot gas in the outer part of the accretion flow that is resolved by _
( e.g. , yuan et al .
2001 ; quataert 2002 ) .
this can account for the size of the source , its lack of variability , and the possible presence of a thermal x - ray line ( baganoff et al .
the above expression for the spectrum of the thermal emission would imply that @xmath126 is required to explain the spectrum .
there is , however , an important problem with this straightforward interpretation : _ chandra _ spectra are extracted in a @xmath16 region around sgr a*. this is comparable to the bondi accretion radius , @xmath127 ( 1 ) .
it is thus incorrect to assume that observations of the extended emission directly probe the `` accretion flow . '' instead , they probe the complex `` transition region '' between the accretion flow and the ambient medium . and @xmath128 for hot gas around sgr a * ; models are extrapolated to @xmath129 using bondi accretion .
_ right : _ x - ray spectra based on the density and temperature profiles in the left panel .
the @xmath130 prediction is consistent with the quiescent x - ray emission coincident with sgr a*.,title="fig : " ] and @xmath128 for hot gas around sgr a * ; models are extrapolated to @xmath129 using bondi accretion .
_ right : _ x - ray spectra based on the density and temperature profiles in the left panel .
the @xmath130 prediction is consistent with the quiescent x - ray emission coincident with sgr a*.,title="fig : " ] figure [ fig:3 ] illustrates the effects of a finite observing beam relative to the bondi accretion radius ; the left panel shows a toy model for the density and temperature as a function of radius around sgr a * ( in units of @xmath131 ) .
the temperature is measured to be @xmath96 kev at large radii . the density profile on @xmath132 scales
is adjusted to roughly reproduce the radial variation of the observed diffuse x - ray emission ( see quataert 2002 for details ) . for @xmath133 gas accretes as a bondi flow , with asymptotic ( @xmath134 ) scalings @xmath36 ( i.e. , @xmath50 ) and @xmath135 .. ] figure [ fig:3]b shows the x - ray spectra that would be seen by a telescope with a beam - size @xmath136 . for large beams , @xmath137 ,
the spectrum is very soft and is dominated by the ambient medium that has @xmath138 kev . for small beams , @xmath139 ,
the spectrum is dominated by the accretion flow and is quite hard , consistent with the above scalings . for the case
applicable to sgr a * , @xmath140 , the emission is still relatively _ soft _ , consistent with the _ chandra _ observations .
this is in spite of the fact that the underlying accretion flow produces a hard x - ray spectrum .
the upshot of figure [ fig:3 ] is that the extended quiescent x - ray source observed by _
chandra _ appears broadly consistent with the emission produced by gas on @xmath131 scales .
this is gas in the transition region between the ambient medium and the accretion flow , rather than the accretion flow itself . using these results to constrain the dynamics of the accreting gas , e.g. , the radial density profile @xmath44 , will require ( 1 ) better theoretical models for the dynamics of the x - ray emitting gas on @xmath141 scales , and ( 2 ) tighter observational constraints , such as spectra as a function of radius .
the galactic center represents a unique opportunity to probe the dynamics of gas accreting onto a massive black hole , from the large scales on which gas is gravitationally captured by the bh to the small scales close to the bh s horizon . in this review ,
i have tried to argue that radiatively inefficient accretion flow ( riaf ) models can provide a reasonably coherent picture of accretion onto sgr a * on all of these scales . to summarize : hot gas in the central @xmath96 pc of the gc
is produced by the ( shocked ) winds from massive stars .
this gas is gravitationally captured by the black hole on scales of @xmath142 pc .
detection of an extended soft x - ray component coincident with sgr a * may be direct evidence for this gravitationally captured gas ( 3.2 ) .
the net rate at which gas accretes through the bh s horizon is , however , much less than the canonical bondi estimate for the rate at which gas is gravitationally captured by the bh ( @xmath143 ) .
the bondi estimate neglects the angular momentum of the accreting gas , which is likely to be a very poor assumption .
numerical simulations ( and analytical models ) of _ rotating _ riafs find that @xmath51 ( 2 ) .
this conclusion is theoretically attractive because it implies that the radiative efficiency of sgr a * need not be as low as @xmath40 as in bondi and adaf models ( which has always been difficult to reconcile with the expectation that electron heating and acceleration would be important in the collisionless magnetized plasma close to the bh ) .
a low accretion rate is also strongly suggested by the detection of linear polarization in the sub - mm emission from sgr a * : faraday rotation constrains the gas density and magnetic field strength close to the bh and argues for @xmath144 , in good agreement with the inference from riaf models .
finally , riaf models with @xmath145 can explain the basic spectral properties of sgr a * ( see , e.g. , figs . [ fig:1 ] & [ fig:2 ] ) .
in particular , the x - ray flares seen by _
chandra _ may be due to synchrotron or inverse - compton emission produced by relativistic electrons accelerated in the inner @xmath119 of the accretion flow ( 3.1 ) .
it is important to stress that , although @xmath87 is both theoretically and observationally favored , all of the models considered here are still `` radiatively inefficient , '' and have efficiencies much less than the canonical thin disk value of @xmath146 ; e.g. , for @xmath145 , @xmath147 .
in fact , all of the physics highlighted in 2 that suppresses the accretion rate with respect to the bondi estimate requires a relatively low efficiency and would not operate in a thin disk .
there are several important issues that i have not addressed in this review . to cite two that clearly require further study : ( 1 ) are the _ chandra _ flares due to , e.g. ,
reconnection or turbulence in the accretion flow , or are they telling us something more fundamental about the dynamics close to the bh ? ( 2 ) both jet and riaf models can explain the basic spectral properties of sgr a*. how can we distinguish between these two components , both of which are almost certainly present ?
e.g. , is the linear polarization detection , which requires a relatively coherent magnetic field , consistent with the magnetic field seen in riaf simulations ? or does it instead require an additional jet component ? 10 aitken , d. k. et al .
, 2000 , apj , 534 , l173 agol , e. , 2000 , apj , 538 , l121 baganoff , f. et al . , 2001 ,
nature , 413 , 45 baganoff , f. et al .
, 2003a , apj , in press baganoff , f. et al . , 2003b , these proceedings beckert , t. & falcke , h. , 2002 , a&a , 388 , 1106 beckert , t. & falcke , h. , 2003 , these proceedings blandford , r.d .
& begelman , m.c . , 1999 ,
mnras , 303 , l1 bondi , h. , 1952 , mnras , 112 , 195 bower , g. , wright , m. c. h. , falcke , h. , & backer , d. , 2003 , apj , in press ( astro - ph/0302227 ) falcke , h & markoff , s. , 2000 , a&a , 362 , 113 ghez , a. et al .
2003 , these proceedings hawley , j. f. & balbus , s. a. , 2002 , apj , 573 , 738 herrnstein , r. m. & ho , p. t. , 2002 , apj , 579 , l83 ichimaru , s. 1977 , apj , 214 , 840 igumenshchev , i. v. , & abramowicz , m. a. 1999 , mnras , 303 , 309 igumenshchev , i. v. , & abramowicz , m. a. 2000 , apjs , 130 , 463 igumenshchev , i. v. , abramowicz , m. a. , & narayan , r. 2000 , apj , 537 , l27 igumenshchev , i. v. , abramowicz , m. a. , & narayan , r. 2003 , apj submitted ( astro - ph/0301402 ) koratkar , a. & blaes , o. 1999 , pasp , 111 , 1 krabbe , a. , genzel , r. , drapatz , s. , & rotaciuc , v. , 1991 , apj , 382 , l19 liu , s. & melia , f. , 2002 , apj , 566 , l77 mahadevan , r. , 1998 , nature , 394 , 651 manmoto , t. , mineshige , s. , & kusunose , m. , 1997 , apj , 489 , 791 markoff , s. , falcke , h. , yuan , f. , & biermann , p. l. , 2001 , a&a , 379 , l13 melia , f. , 1992 , apj , 387 , l25 narayan , r. 2002 , in _ lighthouses of the universe _
m. gilfanov , & r. sunyaev ( heidelberg : springer - verlag ) , in press , astro - ph/0201260 narayan , r. , igumenshchev , i. v. , & abramowicz , m. a. 2000 , apj , 539 , 798 narayan , r. , quataert , e. , igumenshchev , i. v. , & abramowicz , m. a. 2002 , apj , 577 , 295 narayan , r. et al .
, 1998 , apj , 492 , 554 narayan , r. , yi , i. , & mahadevan , r. , 1995 , nature , 374 , 623 narayan , r. , & yi , i. 1994 , apj , 428 , l13 najarro , f. et al .
, 1997 , a&a , 325 , 700 nayakshin , s. , 2003 , these proceedings ozel , f. & di matteo , t. , 2001 , apj , 563 , 276 ozel , f. , psaltis , d. , & narayan , r. , 2000 , apj , 541 , 234 quataert , e. , 2002 , apj , 575 , 855 quataert , e. , dorland , w. , & hammett , g. , 2002 , apj , 577 , 524 quataert , e. & gruzinov , a. , 1999 , apj , 520 , 248 quataert , e. & gruzinov , a. , 2000a , apj , 539 , 809 quataert , e. & gruzinov , a. , 2000b , apj , 545 , 842 quataert , e. & narayan , r. , 1999 , apj , 520 , 298 rees , m. j. , begelman , m. c. , blandford , r. d. , & phinney , e. s. , 1982 , nature , 295 , 17 ruszkowski , m. & begelman , m. c. , 2002 , apj , 573 , 485 schdel , r. et al . , 2002 ,
nature , 419 , 694 shakura , n. i. , & sunyaev , r. a. , 1973 , a&a , 24 , 337 stone , j. m. , pringle , j. e. , & begelman , m. c. 1999 , mnras , 310 , 1002 stone , j. & pringle , j. e. , 2001 , mnras , 322 , 461 tsuboi , m. , miyazaki , a. , & tsutsumi , t. , 1999 , in _ the central parsecs of the galaxy , asp conference series _ , ed .
heino falcke et al . , vol , 186 , p. 105
yuan , f. , markoff , s. , & falcke , h. , 2002 , a&a , 383 , 854 yuan , f. , quataert , e. , & narayan , r. , 2003 , in prep .
zhao , j. et al . , 2003 , apj letters in press ( astro - ph/0302062 ) | i review radiatively inefficient accretion flow models for the @xmath0 black hole ( bh ) in the galactic center .
i argue for a concordance model of sgr a * : both theory and observations suggest that hot ambient gas around the bh is accreted at a rate @xmath1 , much less than the canonical bondi rate .
i interpret _ chandra _ observations of sgr a * in the context of such a model : ( 1 ) the extended quiescent x - ray emission is due to thermal bremsstrahlung from gas in the vicinity of the bondi accretion radius , and ( 2 ) the @xmath2 second long x - ray flares are due to synchrotron or inverse - compton emission by non - thermal electrons accelerated in the inner @xmath3 schwarzschild radii of the accretion flow . |
the variability of grbs provided the main evidence for the internal - external shocks scenario .
external shocks can not produce efficiently such variability @xcite .
internal shocks can produce such temporal structure provided that there are two time scale within the inner engine " - a short time scale that produces the variability and a long time scale that determines the duration of the burst .
so far variability was shown only for long bursts .
it is an open question whether short bursts arise in internal shocks as well . using a new algorithm @xcite we study their variability .
we also present some new results on the temporal structure of long bursts .
our results provide further support for the internal shocks scenario and show that three different time scales operate within the inner engine " .
we analyze the distribution of @xmath4 ( where @xmath5 is the duration of the shortest pulse in a burst , and @xmath6 is the total duration of the burst ) in a sample of the brightest 33 short bursts ( peak flux in [email protected]@xmath8 ) with a good tte data coverage ( for batse data types review @xcite ) .
the tte data is binned into 2msec time bins .
we compare the results to a sample of 34 long bursts with the same peak flux , using the 64ms concatenate data to which we have added a poisson noise so that the signal to noise ratio of both samples would be similar .
we call this later sample the _ noisy long _ set .
fig [ dt / t in shorts ] depicts @xmath4 in both data sets : _
short _ and _
noisy long_. in the _ short _ set the median of @xmath4 is 0.25 .
35% of bursts have @xmath9 and 35% of the bursts show a smooth structure ( @xmath10 ) .
this result could mislead us to the conclusion that a significant fraction of the short bursts have a smooth time profile . but a comparison with the _
noisy long _ results show that also in this group more than 20% of the bursts are single pulsed , while there were no such bursts in the original ( without the added noise ) _
long _ set .
we conclude that short bursts are variable and hence are most likely produced in internal shocks .
while the observed variability is not as large as seen in long bursts one has to remember that when studying variability of short bursts we are approaching the instrumental limitations both in terms of the time scales and of the signal to noise ratio .
it is possible that 10%-20% of the short bursts are produced by external shocks .
according to the internal shocks model @xcite the source ejects relativistic shells with different velocities and shocks arise when faster shells catch slower ones .
we show in @xcite that both the pulses width @xmath11 and the intervals between pulses @xmath12 are proportional to the same parameter - the separation between two following shells , namely the variability time scale of the inner engine " . therefore both distributions should be similar .
moreover , any interval should be correlated to the width of its neighboring pulses .
we have applied our algorithm to a sample of the 68 brightest long bursts in batse 4b catalog ( peak flux in [email protected]@xmath8 ) .
this resulted in 1330 pulses ( 1262 intervals ) .
our null hypothesis was that both @xmath13 and @xmath14 , have lognormal distributions .
the @xmath15 test gives a probability of @xmath16 that the pulses width were taken from a lognormal distribution with @xmath17 ( @xmath18 ) and @xmath19 ( @xmath20 corresponds to @xmath21 between 0.5 and 2.3sec ) .
the @xmath12 distribution shows , however , an excess of long intervals relative to a lognormal distribution .
the @xmath15 probability for a lognormal distribution is * @xmath22*. mcbreen@xcite and li & fenimore@xcite suggest that this deviation is due to the limited resolution ( 64ms ) . however , fitting the intervals above the median with a half gaussian fails .
the inconsistency is not due to the resolution .
many of the long intervals are dominated by a quiescent time : periods within the burst with no observable counts above the background noise . when excluding _ all _ the intervals that contained a quiescent time the @xmath15 probability that the data is lognormal is * @xmath23 * , with @xmath24 ( @xmath25 ) and @xmath26 ( @xmath20 corresponds to @xmath27 between 0.53 and 3.1sec ) .
the similarity between the parameters of both distributions is remarkable .
moreover , we find , as predicted by the internal shocks model , a linear correlation , @xmath28 , between intervals and the following pulses .
the average @xmath28 is 0.48 , showing a strong correlation .
for most short bursts @xmath2 .
this suggests that these bursts are produced by internal shocks .
if , later , the ejecta encounters a surrounding ism then we expect it to produce an external shock and emit an afterglow . for some ( 30% of our sample ) short bursts @xmath29 . however , a comparison with the _
noisy long _ set , shows that this feature could very well be due to the noise .
we can not rule out the possibility that 10%-20% of the short bursts are produced by external shocks or by a single internal collision .
the distribution of interval between pulses shows an excess of long intervals relative to a lognormal distribution . after removing intervals that include quiescent times
the distribution is consistent with a lognormal distribution with comparable parameters to the pulse width distribution .
this result suggests that the @xmath12 distribution is made from the sum of two different distributions : a lognormal distribution that is also compatible with the @xmath11 distribution and the quiescent times distribution . as @xmath30 reflects the central engine behavior
, this suggests that there are two different mechanisms operating within the source .
a short time scale mechanism , with a lognormal distribution and a longer time scale mechanism that turns the central engine on and off and is responsible for the quiescent times .
results that support this suggestion were obtained by ramirez & merloni @xcite .
the correlation between the interval and the following pulse confirms this suggestion and is in an excellent agreement with the internal shocks model .
1 sari , r. & piran , t. , 1997 , apj 485 , 270 nakar , e. & piran , t. , 2001a in preparation scargle , j. d. 1998 , apj v.504 , p.405 rees . m. j. & meszaros , p.,1994 , apj , 430 , l93 nakar , e. & piran , t. , 2001b in preparation mcbreen b. personal communication mcbreen , b. , et .
1994 , mnras , 271 , 662 li , hui ; fenimore , e. 1996 , apj 469 , l115 ramirez - ruiz , e. , merloni , a. , 2001 , mnras 320 l25 | we analyze the temporal structure of long ( @xmath0 ) and short ( @xmath1 ) batse bursts .
we find that : ( i ) in many short bursts @xmath2 ( where @xmath3 is the shortest pulse ) .
this indicates that short bursts arise , like long ones , in internal shocks .
( ii ) in long bursts there is an excess of long intervals between pulses ( relative to a lognormal distribution ) .
this excess can be explained by the existence of _ quiescent times _ , long periods with no signal above the background that arise , most likely , from periods with no source activity .
the lognormal distribution of the intervals ( excluding the _ quiescent times _ ) is similar and correlated with the distribution of the pulses width , in agreement with the predictions of the internal shock model . |
recent surveys of the central region of our galaxy with the * h*igh * e*nergy * s*teroscopic * s*ystem ( h.e.s.s . ) have uncovered a number of @xmath0ray sources in the tev regime ( aharonian et al .
2002 , 2005a , 2005b , 2006a ) .
different from the cases of pulsar wind nebulae ( pwne ) and young supernova remnants ( snrs ) , some of these sources have no non - thermal x - ray counterpart yet been identified . among them , hess j1745 - 303 is one of the most enigmatic objects .
hess j1745 - 303 was firstly discovered by the h.e.s.s .
galactic plane survey ( aharonian et al .
2006a ) and was subsequently investigated in details with dedicated follow - up observations ( aharonian et al .
the tev @xmath0ray image shows that it consists of three spatial components ( i.e. regions a , b and c in fig . 1 of aharonian et al .
2008 ) . owing to the lack of spectral variability and the insignificant dip among these regions in the existing data ,
it was argued that they are originated from a single object ( aharonian et al .
this inference suggests that hess j1745 - 303 as one of the largest unidentified tev sources which has an angular size of @xmath8 .
searches for the possible non - thermal diffuse x - ray component in region a of hess j1745 - 303 have been conveyed with _
xmm - newton _ and _ suzaku _ ( aharonian et al .
2008 ; bamba et al .
none of these observations have resulted in any evidence for the diffuse x - ray emission .
this imposes a tev - to - x - ray flux ratio larger than @xmath9 ( bamba et al .
2009 ) which is larger than the typical value of pwne and snrs ( i.e. less than 2 ) ( cf .
matsumoto et al . 2007 ; bamba et al . 2007 , 2009 ) .
because of the non - detection of counterparts in x - ray / radio , hess j1745 - 303 is dubbed as a dark accelerator " ( bamba et al . 2009 ) . while no non - thermal diffuse x - ray emission has yet been found , a possible excess of neutral iron line emission
was discovered in the direction toward the region a of hess j1745 - 303 ( bamba et al .
2009 ) . together with its proximity to the galactic center and the positional conicidence of a molecular cloud ( aharonian et al .
2008 ) , the line emission is suggested to be the reflected x - rays originated from the previous activity in the galactic center ( bamba et al .
this molecular cloud can be interacted with the shock from a nearby snr g359.1 - 0.5 ( bamba et al . 2000 , 2009 ; lazendic et al . 2002 ; ohnishi et al .
2011 ) and produce the observed @xmath0rays through the acceleration of protons and/or leptons ( see bamba et al . 2009 ) .
nevertheless , this proposed scenario can not be confirmed unambiguously . in view of the presence of many surrounding objects ( see fig . 1 in aharonian et al .
2008 ) , including the mouse " pulsar ( i.e. psr j1747 - 2958 ) , one can not rule out these objects as the source of energetic particles simply based on the tev results ( aharonian et al . 2008 ) . furthermore
, with the current information , it is not possible to discriminate hadronic model and leptonic model ( see fig . 5 & 6 in bamba et al . 2009 ) .
in order to do so , investigations in lower energy regime are required .
it is interesting to note that hess j1745 - 303 is positionally coincident with an unidentified egret source 3eg j1744 - 3011 ( hartman et al .
different from hess j1745 - 303 , 3eg j1744 - 3011 was suggested to demonstrate long - term variability ( torres et al . 2001 ) .
also , based on the mev - gev spectrum observed by egret , the extrapolated flux in the tev regime overshoots that observed by h.e.s.s .. therefore , aharonian et al .
( 2008 ) considered that 3eg j1744 - 3011 is unrelated to hess j1745 - 303 .
after the commence of the large area telescope ( lat ) onboard _ fermi _ gamma - ray space telescope , a detailed investigation of this dark accelerator in the mev@xmath10gev regime is now feasible with its much improved spatial resolution and sensitivity .
however , among 1451 objects detected by lat during the first 11 months , we do not identify any source corresponding to hess j1745 - 303 / 3eg j1744 - 3011
( abdo et al .
2010 ) . very recently , in an analysis of the @xmath0rays from the galactic center with first 25 months lat data , a new serendipitous source was found to coincide spatially with hess j1745 - 303 / 3eg j1744 - 3011 ( chernyakova et al . 2011 ) . in this paper
, we report a detailed analysis of this source with lat observation in the first @xmath1 months .
in this analysis , we used the data obtained by lat between 2008 august 4 and 2010 december 23 .
the _ fermi _ science tools v9r18p6 package is used to reduce and analyze the data in the vicinity of hess j1745 - 303 . only the events that are classified as class 3 or class 4 are adopted .
the post - launch instrument response functions p6_v3_diffuse " were used throughout this investigation . with the aid of the task _ gtlike _
, we performed unbinned maximum - likelihood analysis for a circular region - of - interest ( roi ) with a @xmath11 diameter centered on the nominal position of hess j1745 - 303 ( i.e. ra=@xmath12@xmath13@xmath14 dec=@xmath15@xmath16@xmath17 ( j2000 ) ) .
the size of roi have been chosen to avoid the surrounding bright sources so as to reduce the systematic uncertainties due to the inaccurate background subtraction in this complex region .
for subtracting the background contribution , we included the galactic diffuse model ( gll_iem_v02.fit ) and the isotropic background ( isotropic_iem_v02.txt ) , as well as 41 sources in the first _ fermi_/lat catalog ( 1fgl ; abdo et al .
2010 ) within @xmath11 from the aforementioned center . to begin with
, we compared the spectral properties inferred by lat and egret . for a consistent comparison with hartman et al .
( 1999 ) , we used events with energies @xmath18 mev for our initial analysis .
we assumed a power - law ( pl ) spectrum for hess j1745 - 303 as well as all 1fgl sources in our consideration .
all the sources are assumed to be point sources throughout this investigation .
the best - fit model yields a photon index of @xmath19 . ] and a test - statistic ( ts ) value of 499 which corresponds to a significance of @xmath20 .
this is consistent with the significance reported in the preliminary analysis by chernyakova et al .
the photon index is found to be consistent with 3eg j1744 - 3011 ( i.e. @xmath21 ) . to further compare with the egret results , we compute the integrated photon flux in @xmath22 gev which is found to be @xmath23 @xmath24 s@xmath25 .
this is @xmath26 times smaller than that of 3eg j1744 - 3011 ( i.e. @xmath27 @xmath24 s@xmath25 ) .
the discrepancy can be due to the improved spatial resolution of lat , and hence the estimation of background contribution from the nearby sources is more accurate than egret .
considering the background , as hess j1745 - 303 is located close to the galactic center where the diffuse @xmath0ray emission is very intense , the contamination can possibly be large at lower energies .
also , the point spread function ( psf ) of lat is narrower at higher energies . while the @xmath28 containment radius at 100 mev is @xmath29 , it is only @xmath30 at 1 gev .
therefore , the contamination due to the psf wings of the nearby sources can also be minimized by limiting the analysis at higher energies . in order to minimize the systematic uncertainty in the background modeling so as to obtain robust results ,
we restricted the subsequent analysis in @xmath4 gev . in this adopted band , the best - fit pl model results in a ts value of 332 which corresponds to a detection significance to be @xmath2 . to examine the robustness of the detection , we have also considered the systematic uncertainty of the galactic diffuse emission background . following abdo et al .
( 2010 ) , we repeat the analysis by varying the slope and the normalization of the galactic diffuse model in @xmath31 and @xmath32 respectively . within the uncertainty of the background model ,
the detection significance remains over @xmath33 .
this best - fit pl model yields a photon index of @xmath34 and an energy flux of @xmath35 ergs @xmath24 s@xmath25 in @xmath4 gev . the best - fit pl model and the @xmath0ray spectrum as seen by _
fermi _ lat is shown in figure [ spec ] .
besides a simple pl model , we have also examined if a broken power - law ( bkpl ) or an exponential cutoff power - law ( ple ) can describe the spectrum better .
the fittings with bkpl and ple results in the ts values of 335 and 333 respectively . based on the likelihood ratio test ,
the additional spectral parameters in bkpl / ple are not strongly required which suggests that a single pl is statistically sufficient to describe the data .
for the ple model , the best - fit photon index and cut - off energy are @xmath36 and @xmath37 gev respectively . on the other hand , for the bkpl , the initial fitting resulted in a break energy and the photon indicies of @xmath38 gev , @xmath39 and @xmath40 respectively . however , different from the best - fit solutions inferred from pl and ple models , we found that the solution inferred from the bkpl fit is unstable subjecting to the perturbations in the parameter space .
in view of the problem of the convergence , we will not further consider bkpl model in all subsequent analysis .
although the extra parameter , @xmath41 , in the ple model is not statistically required , one is not able to completely rule out it in view of its capability in depicting the observed data . as a ple model typically describes the @xmath0ray spectrum of a pulsar ( abdo et al . 2010b , 2011 ) , we speculate if there is any hidden pulsar in hess j1745 - 303 . to test this hypothesis ,
we have performed a blind search for coherent pulsation .
the arrival times of each photon were barycentric corrected with the nominal position of hess j1745 - 303 adopted . to minimize the impact due to the ignorance of the spin - down rate
, we have divided the full time span of the adopted data into 5 segments and run the fourier analysis in each segment independently with the aid of the tool _
gtpspec_. from each computed power spectrum , we picked out 10 peaks and investigated if there is any correlation among different segments .
however , we did not identify any promising periodicity candidate in this data .
hence , we conclude that there is no evidence for a hidden pulsar in hess j1745 - 303 . as 3eg j1744 - 3011 was reported to be a variable ( torres et al .
2001 ) , we also examine the variability with lat data .
first , we extracted the light curve obtained from the data within @xmath42 from the nominal position of hess j1745 - 303 with a binning factor of 10 days in the energy range of @xmath4 gev . by fitting a horizontal line to the light curve
, we obtain a reduced chi - square of @xmath43 for 87 degrees of freedom .
hence , there is no strong evidence for any variability or flaring . for a further investigation of the possible spectral and flux variability
, we divided the whole data into five segments of equal time span and performed an unbinned likelihood analysis on each segment .
the results are summarized in table [ spec_var ] .
a simple pl was adopted for all the fittings . within the tolerance of the statistical uncertainties ,
we conclude that neither the spectral shape nor the flux varies among these segments .
ccc time segment & @xmath44 & @xmath45 ( @xmath4 gev ) + mission elapsed time ( s ) & & @xmath46 ergs @xmath24 s@xmath25 + [email protected] & @xmath47 & @xmath48 + [email protected] & @xmath49 & @xmath50 + [email protected] & @xmath51 & @xmath52 + [email protected] & @xmath53 & @xmath54 + 299733661.8@xmath10314777723 & @xmath55 & @xmath56 + [ spec_var ] we have computed the @xmath57 ts map in @xmath4 gev centered at the nominal position of hess j1745 - 303 by using the tool _ gttsmap_. this is displayed in figure [ ts]a . utilizing _
gtfindsrc _ , we determined the best - fit postion in @xmath4 gev to be ra=@xmath12@xmath58@xmath59 dec=@xmath15@xmath60@xmath61 ( j2000 ) with a @xmath62 error radius of @xmath63 .
the position locates between regions a and c of hess j1745 - 303 .
given that the psf has a 68% containment radius of @xmath30 at 1 gev , the source extent inferred in this band is consistent with that of a point source .
the relatively wide psf do not allow us to determine whether gev emission is associated with any particular tev feature . to further examine the spatial nature
, we also computed the ts map in @xmath5 gev which is displayed in figure [ ts]b . since the 68% containment radius at 10 gev is @xmath9 times smaller than that at 1 gev , the feature can be better resolved .
we found that the peak ts value found in this band is 41 which corresponds to a significance of @xmath3 . within the systematic uncertainty of the galactic diffuse background
, we found that the detection significance of the source remains over @xmath64 in this band .
the best - fit position in @xmath5 gev is found to be ra=@xmath12@xmath65@xmath66 dec=@xmath15@xmath67@xmath68 ( j2000 ) with a @xmath62 error radius of @xmath63 .
this differs from that inferred in @xmath4 gev by @xmath69 .
we note that the gev feature found in this hard band is apparently peaked at region c and possibly extended to region b. although it appears to be extended in the hard band , the relatively low detection significance in this energy range does not allow a firm conclusion .
in this paper , we have reported a detailed study of hess j1745 - 303 with _ fermi _ lat data which provides a missing piece in understanding the nature of this dark accelerator . in view of the putative variability of 3eg j1744 - 3011 , aharonian et al .
( 2008 ) argued that the gev counterpart is unlikely to be associated with hess j1745 - 303 .
however , we do not find any evidence for the spectral / flux variability from the lat data .
the discrepancy of the lat result and that inferred from egret is most likely due to the differences between their instrumental performance . with the improved angular resolution and sensitivity of lat ,
many previously unknown sources in the proximity of hess j1745 - 303 are now detected .
as several sources within @xmath70 are found to be variable , including 1fgl j1747.2 - 2958 which is the @xmath0ray counterpart of psr j1747 - 2958 ( abdo et al .
2010 ) , the source confusion in the egret data can possibly lead to the apparent variability .
it is interesting to compare the spectral properties inferred in the gev regime with those obtained by h.e.s.s ..
first , the photon index inferred from the lat data ( i.e. @xmath34 ) is similar to that inferred in tev ( cf . tab . 2 of aharonian et al .
furthermore , the extrapolated photon flux in @xmath7 tev with the best - fit pl model is found to be @xmath71 @xmath24 s@xmath25 which is consistent with any individual spatial component observed by h.e.s.s . within @xmath62 uncertainties ( cf . tab . 2 of aharonian et al .
2008 ) .
since the tev spectral data of region a is avaliable ( cf . fig . 2 of aharonian et al .
2008 ) , we further compare this particular region with the gev spectrum by constructing a spectral energy distribution ( sed ) which is display in figure [ sed ] .
both data can be simultaneously fitted with a single pl of @xmath72 .
this clearly demonstrates that the tev spectrum of this spatial component can be smoothly connected to the gev spectrum . for regions
b and c , despite that no tev spectral data is currently available , we note that their spectral shapes reported by aharonian et al .
( 2008 ) are similar to that of region a. within the statistical uncertainties of the spectral properties inferred in both gev and tev regimes , the tev spectra of these regions can also be possibly connected to that of gev .
further investigation by h.e.s.s .
is strongly encouraged , particularly for region c as most of the @xmath0ray emission in @xmath5 gev apparently coincides with this spatial component . as both pl and ple model can describe the gev spectrum equally well
, we can not discriminate these competing models based on the current data .
in view of the exponential cut - off , the spectral connection between the gev and tev regimes can not be established with the ple model .
however , we would like to point out that the best - fit cut - off energy , @xmath37 gev , falls in the highest energy bin of the lat spectrum ( cf .
[ spec ] and fig .
[ sed ] ) .
owing to the small photon statistics in the hard band , the statistical uncertainty of the highest energy bin is rather large . in view of this , it remains unclear whether the inferred cut - off is genuine as this particular data point is sensitive to the systematic uncertainty of the background .
analysis of the lat data in higher energies ( e.g. @xmath73 gev ) with sufficient photon statistic in the future can help to discriminate these two models .
the detection of @xmath0rays provides a strong evidence for the particle acceleration .
it should be noted that there are two snrs and two pulsars around hess j1745 - 303 which can be the potential high energy particle injector ( cf .
fig . 1 of aharonian et al .
based on the small distances of g359.0 - 0.9 and psr b1742 - 30 , bamba et al .
( 2009 ) argued that these two sources are foreground objects which are unrelated to hess j1745 - 303 . with the gev counterpart revealed by lat
, we can now safely rule out the possibility that psr b1742 - 30 is the major contributor as its spin - down flux , @xmath74 erg @xmath24 s@xmath25 , is lower than the energy flux observed in the gev regime ( cf .
manchester et al .
on the other hand , for psr j1747 - 2958 , the sum of @xmath0ray flux of 1fgl j1747.2 - 2958 and hess j1745 - 303 observed by lat only consumes @xmath75 of its spin - down flux ( @xmath76 erg @xmath24 s@xmath25 ) .
therefore , it is energetically possible to be the source for the high energy particles .
nevertheless , with its eastward proper - motion ( hales et al . 2009 ) , its backward extrapolated position by its spin - down age is found to have an offset of @xmath77 and @xmath30 from the best - fit positions inferred in @xmath4 gev and @xmath5 gev respectively .
if the feature found by h.e.s.s . and lat is indeed a pwn , then it is one of the most peculiar system because of its large positional offset with respect to the pulsar position .
while the contribution of psr j1747 - 2958 is uncertain , the interaction between the shock from g359.1 - 0.5 and a molecular cloud in its neighbourhood is considered to be a more viable means to produce the observed @xmath0rays . for both hadronic and leptonic scenarios , bamba et al .
( 2009 ) have modeled the broadband spectrum of the region a of hess j1745 - 303 .
however , comparing fig .
[ spec ] in this paper and fig . 5 & 6 in bamba et al .
( 2009 ) , the gev flux predicted in all the scenarios considered in their work are at least an order of magnitude lower than that observed by lat . revised theoretical investigation with constraints provided by lat is therefore required .
moreover , while the other known systems of snrs interacting with molecular clouds have their remnant shells found in other wavelengths coincided with the @xmath0ray emission ( e.g. castro & slane 2010 ; abdo et al .
2010c ) , there is no such evidence for hess j1745 - 303 .
we should point out that all the previous multiwavelength campaign were targeted only at the bright component of hess j1745 - 303 , namely the region a. since the lat observation suggested that the gev emission can possibly be originated from the regions c / b , observational investigations aim at these regions are encouraged for a further understanding of this mysterious object .
based on the above discussion , we have to admit that the energy injection source of hess j1745 - 303 remains unclear .
it is instructive to compare it with other nearby high energy sources .
aharonian et al . (
2006b ) have reported observations of an extended region of very high energy ( vhe , @xmath18 gev ) @xmath78-ray emission correlated spatially with a complex of giant molecular clouds in the central 200 pc of the milky way .
it appears that tev emissions from the molecular clouds in the vicinity of the galactic center are quite common phenomena .
in addition , similar to the case of hess j1745 - 303 , 6.4 kev lines have been commonly detected from many molecular clouds near the galactic center , e.g. sgr b2 ( koyama et al . 1996 ; murakami et al . 2000 ) , sgr c ( murakami et al . 2001 ) and others ( bamba et al .
2002 ; predehl et al .
2003 ; nobukawa et al . 2008 ; nakajima et al .
2009 ; koyama et al .
it has been speculated that these neutral iron lines arise from the reflection by the dense molecular clouds which are irradiated by the nearby x - ray sources ( e.g. koyama et al .
2007 ; inui et al .
on the other hand , dogiel et al . (
2009 ) argued that these 6.4 kev lines from the molecular clouds are excited by a background intensity of subrelativistic protons coming from the escaped part of a past captured star by the galactic supermassive black hole sgr a*. the periodic stellar capture events may explain the recent observed _ fermi _ bubble ( cheng et al .
this subrelativistic proton wind can also form shock by hitting the clouds and produce relativistic protons .
it is shown that the decay of neutral pions produced by hadronic collisions between the accelerated relativistic protons in the clouds can emit a power - law @xmath0ray spectrum from 30 mev to 10 tev without any spectral break ( cheng et al .
2011b in preparation ) .
if this is true , then the high energy emission from various molecular clouds in the vicinity of the galactic center are correlated to the past activities of the galactic center and their intensity might be correlated with the propagation history of the injection of these subrelativistic protons escaped from the capture past events .
abdo , a. a. , et al .
2011 , apjs , 193 , 22 abdo , a. a. , et al .
2010 , apjs , 188 , 405 abdo , a. a. , et al .
2010b , apjs , 187 , 460 abdo , a. a. , et al .
2010c , science , 327 , 1103 aharonian , f. , et al .
2002 , a&a , 393 , l37 aharonian , f. , et al . 2005a , science , 307 , 1938 aharonian , f. , et al .
2005b , a&a , 439 , 1013 aharonian , f. , et al . 2006a , apj , 636 , 777 aharonian , f. , et al .
2006b , nature , 439 , 695 aharonian , f. , et al .
2008 , a&a , 483 , 509 bamba , a. , et al .
2009 , apj , 691 , 1854 bamba , a. , et al .
2007 , pasj , 59 , s209 bamba , a. , et al .
2002 , proceedings of the symposium new visions of the x - ray universe in the xmm - newton and chandra era " , noordwijk - nl bamba , a. , et al .
2000 , pasj , 52 , 259 castro , d. , & slane , p. 2010 ,
apj , 717 , 372 cheng , k. s. , et al .
2011 , apj , 731 , l17 chernyakova , m. , malyshev , d. , aharonian , f. a. , crocker , r. m. , & jones , d. i. 2011 , apj , 726 , 60 dogiel , v. , et al .
2009 , pasj , 61 , 901 hales , c. a. , gaensler , b. m. , chatterjee , s. , van der swaluw , e. , & camilo , f. 2009 , apj , 706 , 1316 hartman , r. c. , et al . 1999 ,
apjs , 123 , 79 inui et al .
2009 , pasj , 61 , s241 koyama , k , et al .
2009 , pasj , 61 , 255 koyama , k , et al .
2007 , pasj , 59 , s221 koyama , k , et al .
1996 , pasj , 48 , 249 lazendic , j. s , et al .
2002 , mnras , 331 , 537 manchester , r. n. , hobbs , g. b. , teoh , a. , & hobbs , m. 2005 , aj , 129 , 1993 matsumoto , h. , et al .
2007 , pasj , 59 , s199 murakami , h. , et al .
2001 , apj , 550 , 297 murakami , h. , et al .
2000 , apj , 534 , 283 nakajima , h. , et al .
2009 , pasj , 61 , 233 nobukawa , m. , et al .
2008 , pasj , 60 , 191 ohnishi , t et al .
2011 , pasj in press ( arxiv:1103.0348 ) predehl , p. , costantini , e. , hasinger , g. , & tanaka , y. 2003 , an , 324 , 73 torres , d. f. , et al .
2001 , a&a , 370 , 468 | we present a detailed analysis of the @xmath0ray emission from hess j1745 - 303 with the data obtained by the _ fermi _ gamma - ray space telescope in the first @xmath1 months observation .
the source can be clearly detected at the level of @xmath2 and @xmath3 in @xmath4 gev and @xmath5 gev respectively .
different from the results obtained by the _
compton _
gamma - ray observatory , we do not find any evidence of variability .
most of emission in @xmath5 gev is found to coincide with the region c of hess j1745 - 303 .
a simple power - law is sufficient to describe the gev spectrum with a photon index of @xmath6 .
the power - law spectrum inferred in the gev regime can be connected to that of a particular spatial component of hess j1745 - 303 in @xmath7 tev without any spectral break .
these properties impose independent constraints for understanding the nature of this dark particle accelerator " . |
wannier function@xcite is a powerful tool in the study of the chemical bonding , dielectric properties , excited electronic states , electron transport , and many body correlations in materials . in particular , the modern theory of bulk polarization
relates the vector sum of the centers of the wannier functions to the macroscopic polarization of a crystalline insulator.@xcite however , the intrinsic nonuniqueness in the wannier function definition , and the difficulty in defining their centers within a periodic cell calculation , limited their practical use .
fortunately , an elegant method has been recently proposed by marzari and vanderbilt to obtain a unique set of maximally - localized wannier functions ( mlwfs).@xcite by transforming the occupied electronic manifold into a set of mlwfs , it becomes possible to obtain an enhanced understanding of chemical bonding properties and electric polarization via an analysis of the mlwfs . beside the above points ,
the mlwfs are now also being used as a very accurate minimal basis for a variety of algorithmic or theoretical developments , with recent applications ranging from linear - scaling approaches@xcite to the construction of effective hamiltonians for the study of ballistic transport,@xcite strongly - correlated electrons,@xcite self - interaction corrections , metal - insulator transitions,@xcite and photonic lattices.@xcite in the seminal work of marzari and vanderbilt , first a ground state calculation was carried out to obtain the occupied delocalized canonical orbitals , then a sequence of unitary transformations were performed to obtain mlwfs which minimize the spread function.@xcite using the exponential representation for the unitary transformation , berghold _
et al._@xcite derived an iterative scheme to obtain mlwfs in large supercells of arbitrary symmetry .
also a simple jacobi orbital rotation scheme was found to be remarkably efficient.@xcite a simultaneous diagonalization algorithm similar to the jacobi diagonalization method , was used by gygi _
_ to compute mlwfs.@xcite zicovich - wilson _
proposed a wannier - boys scheme to obtain well localized wannier functions in linear combination of atomic orbital periodic calculations .
@xcite however , all methods mentioned above for calculating mlwfs are nearly o(n@xmath0 ) scaling ( n is the number of electrons ) , which prohibits their applications to large systems containing hundreds or thousands of atoms .
the unfavorable scaling comes from two steps in these methods : the conventional methods for getting ground state wavefunctions is o(n@xmath0 ) or o(n@xmath1n ) , and the localization step in the above localization algorithms is also o(n@xmath0 ) .
usually , the traditional ground state calculation will cost more than the localization step .
however , for large systems the computing amount of the localization step is also time - consuming . in this work
, we propose a simple order - n algorithm for effectively calculating mlwfs . the demanding ground state calculation
is circumvented by using o(n ) density matrix purification methods .
after adopting o(n ) method for the ground state calculation , the conventional o(n@xmath0 ) localization step will become time - dominant for large systems . to obtain mlwfs in linear scaling ,
we first get initial localized orbitals from the density matrix , then an o(n ) localization method which uses the jacobi rotation scheme is utilized to maximally localize the orbitals .
the linear scaling behavior of the new method is demonstrated by computing the mlwfs of boron nitride ( bn ) nanotubes .
this paper is organized as follows : in sec .
[ theory ] , we present our new o(n ) method for calculating mlwfs . in sec .
[ impl ] , we describe the details of the implementation and perform some test calculations to illustrate the rightness , robustness , and linear - scaling behavior of our methods .
we discuss some possible extensions and generalizations of our method in sec .
[ dis ] . finally , our concluding remarks are given in sec .
the wannier functions are defined in terms of a unitary transformation of the occupied bloch orbitals . however , they are not uniquely defined , due to the arbitrary freedom in the phases of the bloch orbitals .
marzari and vanderbilt @xcite resolve this indeterminacy by minimizing the total spread function of the wannier functions @xmath2 @xmath3 where @xmath4 , and @xmath5 . here since we aim at large systems , the @xmath6-point - only sampling of the brillouin zone ( bz ) is used throughout this work .
the method of calculating mlwfs for supercells of general symmetry is proposed by silvestrelli _
et al._.@xcite for the sake of simplicity , considering the case of a simple - cubic supercell of side @xmath7 , it can be proved that minimizing the total spread @xmath8 is equivalent to the problem of maximizing the functional @xmath9 where @xmath10 and similar definitions for @xmath11 and @xmath12 apply .
the coordinate @xmath13 of the @xmath14th wannier - function center ( wfc ) is computed using the formula @xmath15 with similar definitions for @xmath16 and @xmath17 .
our new o(n ) method consists of four o(n ) steps :
first we obtain the density matrix , secondly we find out a set of linear independent nonorthogonal orbitals which span the occupied manifold , thirdly , a modified lwdin orthogonalization is used to orthogonalize these nonorthogonal orbitals , finally , the jacobi rotation scheme is utilized to maximally localize the orbitals . in principles , any localized orbitals or density matrix based linear scaling methods can be used to obtain initial localized orbitals in linear scaling.@xcite here we use the o(n ) trace - correcting density matrix purification ( tc2)@xcite method to get the density matrix since it is very simple , robust , and efficient .
the use of some other linear scaling methods based on localized orbitals will be discussed in sec .
[ dis ] . here , the essence of the tc2 method is briefly outlined . in the begining ,
the hamiltonian @xmath18 represented in the orthogonal basis is normalized to an initial matrix @xmath19 with all its eigenvalues mapped onto [ 0,1 ] : @xmath20 , where @xmath21 and @xmath22 are lowest and highest eigenvalues of @xmath18 , respectively .
then we correct the trace of the density matrix while purifying it using the following iteration : @xmath23 where @xmath24 is the total number of electrons in a close - shell system .
given an atomic orbital , one can project out its occupied component using the density matrix operator @xmath25 : @xmath26 where @xmath27 is the overlap matrix , @xmath28 and @xmath29 denote the atomic basis orbitals , @xmath30 denotes the density matrix in the atomic orbital basis . by applying the density matrix operator on @xmath31 atomic basis orbitals
, we can get @xmath31 localized orbitals @xmath32 , among which only @xmath33 ( the number of occupied states ) localized orbitals are linear independent
. one must select out @xmath33 linear independent localized orbitals among these localized orbitals @xmath32 before performing localization steps .
we have implemented two algorithms to achieve this goal .
one of the algorithms is similar to that proposed by maslen _
et al._,@xcite who used cholesky decompositions for detecting the linear dependence .
we note that in this method the total demanding for the construction of all the overlap matrices is o(n ) since the nonorthogonal orbitals are localized .
furthermore , the total computing amount for performing all cholesky decompositions is almost the same as that for performing a sparse cholesky decomposition with the matrix dimension @xmath33 due to the nature of the cholesky decomposition . in the second algorithm ,
we select these atomic basis orbitals to be projected according to the physical intuition . for example , in bn nanotubes , there are one 2s and three 2p basis orbitals for each b and n atoms when using pseudopotentials . since some electrons will transfer from b to n atoms , we can get @xmath33 linear independent localized orbitals by just projecting the density matrix on all atomic basis orbitals of n atoms in despite of the large covalency in these systems . for systems
where the bonding properties are known , the algorithm is found to be very efficient and the resulting orbitals are very sparse . in cases where the second algorithm does nt apply , we will resort to the first algorithm . since mlwfs
are orthogonal to each other , the @xmath33 linear independent nonorthogonal localized orbitals must be orthogonalized .
it is well known that the orthogonalized orbitals produced by the lwdin orthogonalization are closest to initial nonorthogonal orbitals in the sense of least squares . to obtain orthogonal localized orbitals in linear scaling ,
we carry out a modified lwdin orthogonalization adopted by stephan and drabold.@xcite in this approach , we perform repeated first - order lwdin iterations @xmath34 the functions after every orthonormalization cycle have to be renormalized .
typically , we can obtain well orthogonal localized orbitals in about five o(n ) orthonormalization cycles .
now we will discuss how to maximally localize these orthogonal orbitals to get mlwfs in linear scaling .
our target is maximizing @xmath35 to obtain mlwfs in linear scaling . the key point for the successful o(n )
localization step is that these orbitals are localized in the whole localization procedure .
we use the simple jacobi rotation method to maximize @xmath35 since it does nt require o(n@xmath0 ) diagonalization in contrast to the unitary transformation method.@xcite this method is a traditional method in quantum chemistry for computing localized molecular orbitals first introduced by edmiston and ruedenberg.@xcite the basic idea of the method is to tackle the problem of maximizing @xmath35 by carrying out several jacobi sweeps . in a traditional jacobi sweep ,
we perform @xmath36 consecutive two - by - two rotations among all pairs of orbitals .
the elementary step consists of a plane rotation where two orbitals are rotated through an angle and all other orbitals are fixed . in our o(n ) method , due to the localization of the orbitals , the computing amount of an orbital ratation is o(n ) . moreover ,
each orbital overlaps with only o(1 ) orbitals , thus the number of the jacobi rotations in a jacobi sweep is of order n. assuming the number of the jacobi sweeps does nt change ( it is really the case for systems with similar characters ) , the total computing amount is o(n ) .
our newly developed method has been implemented in siesta,@xcite a standard kohn - sham density - functional program using norm - conserving pseudopotentials and numerical atomic orbitals as basis sets . in siesta ,
periodic boundary conditions are employed to simulate both isolated and periodic systems .
the details about the implementation of the tc2 method can be found in ref . [ 18 ] .
all our calculations reported in this work are done in the local density approximation ( lda).@xcite unless otherwise stated , the double-@xmath37 plus polarization functions ( dzp ) basis set is used in the calculations .
first we calculate the mlwfs ( i.e. , boys orbitals@xcite ) of a water molecule .
it is well known that there are four mlwfs for a h@xmath38o molecule : two covalent o - h @xmath39 bonds and two lone - pair orbitals .
the distance between the centroids of these four mlwfs and the position of the oxygen ion is 0.52 , 0.52 , 0.30 , and 0.30 respectively .
the results agree well with those reported by berghold _
et al._.@xcite as a second check of the validity of our method , we calculate the piezoelectric constant of bulk zno .
both the berry phase method@xcite and our new o(n ) method are used to calculate piezoelectric constant @xmath40 of bulk wurtzite zno . in our o(n ) calculation , we use a @xmath41 zno supercell since we use the @xmath6-only sampling .
the results from these two methods agree very well : the computed values are @xmath42 and @xmath43 c / m@xmath44 , respectively . and
both results accord with others result @xcite ( @xmath45 c / m@xmath44 ) computed through the density functional perturbation theory .
then we test our method by applying it to calculate the mlwfs of bn(5,5 ) armchair nanotubes . fig .
[ fig1 ] shows four mlwfs of bn(5,5 ) nanotubes computed using the supercell containing 200 atoms .
we can clearly see that the three mlwfs in fig .
[ fig1](a)-(c ) are b - n @xmath39 bonds . among these mlwfs ,
the mlwf in fig .
[ fig1](a ) has exactly the same character as that shown in fig .
[ fig1](c ) due to the mirror plane symmetry in armchair bn nanotubes .
the other mlwf in fig .
[ fig1](d ) is a @xmath46 orbital , which almost centers at a nitrogen atom .
it is interesting that in bn nanotubes mlwfs preserve @xmath47 separation .
we will return to this point later . to see the efficiency of our new o(n ) method
, we perform a series of calculations using the linear scaling method or the traditional jacobi rotation method which does nt take advantage of the localization property of the orbitals .
two different basis sets ( single-@xmath37 ( sz ) , dzp ) are employed in the calculations using the new o(n ) method , and only the sz basis set is used for the traditional method . the cpu time is shown in fig .
we clearly see the perfect linear scaling behavior of our new method . and
the traditional method displays a nearly o(n@xmath0 ) scaling as expected .
the computing saving of our method with respect to the traditional method is dramatically large , especially when the size of systems exceeds 400 atoms .
in addition , we note that the ratio between the time for calculating mlwfs using dzp basis and that using sz basis is smaller than the case for the ground state calculation.@xcite
we have also implemented the method for obtaining mlwfs from the localized orbitals produced by the o(n ) mauri - ordejn ( mo ) @xcite or kmg@xcite methods . in case of the mo method ,
the number of localized orbitals is equal to the number of occupied states , and all localized orbitals produced by the o(n ) mo method are linear independent .
thus the projection step is unnecessary .
however , in the kmg energy functional , the number of localized orbitals is larger than the number of occupied states . in this case , we first get the density matrix using equation ( 87 ) in the paper written by soler et al .. @xcite once the density matrix is available , the steps to get mlwfs are the same as those described in sec .
[ theory ] . in our implementation , numerical localized atomic orbitals are used as basis sets . other localized basis such as gaussian orbitals or real space methods@xcite
can also be used .
however , the use of non - local plane - wave basis which was commonly adopted in previous calculations of mlwfs will not result in linear scaling behavior .
our previous discussions mainly focus on mlwfs in periodic systems .
we should note that our method could also be adopted to obtain boys@xcite localized orbitals of isolated molecular systems with or without using periodic boundary conditions . in this case
, one can directly minimize the spread function instead maximize the functional @xmath35 . besides boys orbitals , edmiston - ruedenberg ( er)@xcite and pipek - mezey ( pm)@xcite localized orbitals are also popular among chemists .
the benefit of these localized orbitals is that , unlike boys orbitals , er and pm orbitals always preserve @xmath47 separation .
our method can be used to get pm orbitals in linear scaling .
however , due to the long range character of the operator @xmath48 , our method is unable to reduce significantly the computing amount in the calculation of er orbitals .
we notice that an efficient method which reduces the scaling from o(n@xmath49 ) to o(n@xmath44)-o(n@xmath0 ) has been proposed by subotnik _
et al._.@xcite although we only discuss spin - restricted systems up to now , combined with the spin - unrestricted linear scaling electronic structure theory,@xcite our o(n ) method can be straightforwardly applied to insulating magnetic systems .
this method can be used to study the large multiferroic ( simultaneously ( ferro)magnetic and ferroelectric ) materials .
to summarize , a linear scaling algorithm for calculating mlwfs has been proposed for the first time .
the numerical atomic orbital basis instead of commonly adopted plane wave sets is used .
first we perform a linear scaling ground state calculation using the tc2 purification method . from the density matrix
, we get the initial non - orthogonal localized orbitals . through a modified lwdin orthogonalization
, we obtain the initial orthogonal localized orbitals . due to the localization property of these initial orbitals , the computing requirement of the subsequent jacobi sweeps for getting mlwfs is also linear scaling .
our results for water molecule and bulk wurtzite zno agree well with others results .
the o(n ) behavior of the proposed method is clearly demonstrated by computing the mlwfs of bn nanotubes .
our o(n ) method provides a very efficient way for obtaining mlwfs which have many possible applications .
w. ku , h. rosner , w. e. pickett , and r. t. scalettar , phys .
lett . * 89 * , 167204 ( 2002 ) ; i. schnell , g. czycholl , and r. c. albers , phys .
b * 65 * , 075103(2002 ) ; k. nakamura , r. arita , y. yoshimoto , s. tsuneyuki , cond - mat/0510425 . | we have developed a linear scaling algorithm for calculating maximally - localized wannier functions ( mlwfs ) using atomic orbital basis .
an o(n ) ground state calculation is carried out to get the density matrix ( dm ) . through a projection of the dm onto atomic orbitals and a subsequent o(n ) orthogonalization
, we obtain initial orthogonal localized orbitals .
these orbitals can be maximally localized in linear scaling by simple jacobi sweeps .
our o(n ) method is validated by applying it to water molecule and wurtzite zno .
the linear scaling behavior of the new method is demonstrated by computing the mlwfs of boron nitride nanotubes . |
to solve complicated problems in economics , engineering and environment , we can not successfully use classical methods because of different kinds of incomplete knowledge , typical for those problems .
there are four theories : theory of probablity , fuzzy set theory ( fst ) @xcite , interval mathematics and rough set theory ( rst ) @xcite , which we can consider as mathematical tools for dealing with imperfect knowledge .
all these tools require the pre - specification of some parameter to start with , e.g. probablity density function in probablity theory , membership function in fst and an equivalence relation in rst .
such a requirement , seen in the backdrop of imperfect or incomplete knowledge , raises many problems . at the same time
, incomplete knowledge remains the most glaring characteristic of humanistic systems
systems exemplified by economic systems , biological systems , social systems , political systems , information systems and , more generally , man - machine systems of various types . noting problems in parameter specification molodtsov
@xcite introduced the notion of soft set to deal with problems of incomplete information .
soft set theory ( sst ) does not require the specification of a parameter , instead it accommodates approximate descriptions of an object as its starting point .
this makes sst a natural mathematical formalism for approximate reasoning .
we can use any parametrization we prefer : with the help of words , sentences , real numbers , functions , mappings , and so on .
this means that the problem of setting the membership function or any similar problem does not arise in sst .
sst has seminal links with rough set technique of automated knowledge discovery .
soft set being collection of information granules , bears a close resemblance with rough sets .
a rough set @xcite is defined to be a set given by an upper and a lower approximation sets from a universe of information granules .
aktas and cagman @xcite have shown that , both , an arbitrary rough set or an arbitrary fuzzy set may be expressed as a soft set .
hence soft set theory is more general a set up than rst and/or fst .
links between soft sets and information systems and hence to rough set theory , have been further studied in @xcite . on the other hand , techniques from rst
have been found applicable to sst , due to the affinity of both approaches .
maji , biswas and roy @xcite applied the technique of knowledge reduction to the information table induced by a soft set .
another parametrization reduction of soft set was proposed in @xcite .
recently z. kong _ et.al .
_ has also proposed yet another novel method of parameter reduction in sfst08konb .
applications of soft set theory in other disciplines and real life problems are now catching momentum .
molodtsov @xcite successfully applied the soft set theory into several directions , such as smoothness of functions , riemann - integration , perron integration , theory of probability , theory of measurement and so on .
kovkov _ et.al .
_ @xcite has found promising results by applying soft sets to optimization theory , game theory and operations research .
maji and roy @xcite applied soft sets in a multicriteria decision making ( mcdm ) problem .
it is based on the notion of knowledge reduction of rough sets .
mushrif and sengupta sfst06mus based their algorithm for natural texture classification on soft sets .
this algorithm has a low computational complexity when compared to a bayes technique based method for texture classification .
zou and xia sfst08zou have exploited the link between soft sets and data analysis in incomplete information systems . in this paper , we first introduce the notion of mapping on soft classes .
soft classes are collections of soft sets ( definition [ df_sftclass ] ) .
we also define and study the properties of soft images and soft inverse images of soft sets , and support them with examples and counterexamples .
finally , these notions have been applied to the problem of medical diagnosis in medical expert systems .
first we recall basic definitions and results .
@xcite[st - softset ] a pair @xmath0 is called a soft set over @xmath1 , where @xmath2 is a mapping given by @xmath3 in other words , a soft set over @xmath1 is a parametrized family of subsets of the universe @xmath4 for @xmath5 @xmath6 may be considered as the set of @xmath7-approximate elements of the soft set @xmath0 .
clearly a soft set is not a set in ordinary sense .
@xcite[st - subset ] for two soft sets @xmath0 and @xmath8 over @xmath1 , we say that @xmath0 is a soft subset of @xmath9 if@xmath10 @xmath11 and@xmath12 @xmath13.we write @xmath14 .
@xmath0 is said to be a soft super set of @xmath8 , if @xmath8 is a soft subset of @xmath0 .
we denote it by @xmath15 .
@xcite[st - union ] union of two soft sets @xmath0 and @xmath8 over the common universe @xmath1 is the soft set @xmath16 where @xmath17 and @xmath18@xmath19we write @xmath20 maji , biswas and roy defined the intersection of two soft sets as : [ intersection_def_of_maji]@xcite intersection of two soft sets @xmath0 and @xmath8 over @xmath1 is a soft set @xmath16 where @xmath21 and @xmath22 , ( as both are same set ) , and is written as @xmath23 pei and miao pointed out that generally @xmath24 and @xmath25 may not be identical and thus revised the above definition as : @xcite[intersection_def_of_pei ] let @xmath0 and @xmath8 be two soft sets over @xmath1 . intersection ( also called bi - intersction by feng et.al .
@xcite ) of two soft sets @xmath0 and @xmath8 is a soft set @xmath16 where @xmath21 and @xmath26 .
we write @xmath23 we further point out that in definition [ intersection_def_of_pei ] , @xmath27 must be nonempty to avoid the degenerate case . hence the definition [ intersection_def_of_pei ] is improved as : [ intersection_def_ours]let @xmath0 and @xmath8 be two soft sets over @xmath28 with @xmath29 .
intersection of two soft sets @xmath0 and @xmath8 is a soft set @xmath16 where @xmath21 and @xmath30 . we write @xmath23
first we define : [ df_sftclass]let @xmath1 be a universe and @xmath31 a set of attributes .
then the collection of all soft sets over @xmath1 with attributes from @xmath31 is called a soft class and is denoted as @xmath32 .
let @xmath32 and @xmath33 be soft classes .
let @xmath34 and @xmath35 be mappings . then a mapping @xmath36 is defined as : for a soft set @xmath37 in @xmath32 , @xmath38 @xmath39 is a soft set in @xmath33 given by@xmath40for @xmath41 @xmath42 is called a soft image of a soft set @xmath43 if @xmath44 then we shall write @xmath45 as @xmath46 let @xmath36 be a mapping from a soft class @xmath32 to another soft class @xmath47 and @xmath48 a soft set in soft class @xmath49 where @xmath50 let @xmath34 and @xmath35 be mappings .
then @xmath51 is a soft set in the soft class @xmath52 defined as : @xmath53for @xmath54 @xmath55 is called a soft inverse image of @xmath56 hereafter we shall write @xmath57 as @xmath58 [ ex_illustrating_def_of_fn]let @xmath59 @xmath60 and @xmath52 @xmath49 soft classes .
define @xmath34 and @xmath35 as : @xmath61choose two soft sets over @xmath1 and @xmath62 respectively as:@xmath63then the mapping @xmath64 is given as : for a soft set @xmath37 in @xmath65 @xmath66 where @xmath67 is a soft set in @xmath33 obtained as follows:@xmath68hence@xmath69next for the soft images , we have @xmath70where @xmath71 hence , we have @xmath72 [ df_union_of_images]let @xmath73 @xmath74 be a mapping and @xmath37 , @xmath75 soft sets in @xmath32
. then for @xmath76 soft union and intersection of soft images of ( f , a ) and ( g , b ) in ( x , e ) are defined as : @xmath77 [ df_union_of_inv_images]let @xmath73 @xmath74 be a mapping and @xmath37 , @xmath75 soft sets in.@xmath78 then for @xmath79 soft union and intersection of soft inverse images of soft sets @xmath80 are defined as:@xmath81 note that the null ( resp .
absolute ) soft set as defined by maji _ et.al .
@xcite , is not unique in a soft space _ @xmath52 _ _ rather it depends upon _
_ @xmath82 therefore , we denote it by @xmath83 ( resp .
@xmath84)@xmath85 if @xmath86 then we denote it simply by @xmath87 ( resp .
@xmath88)@xmath89 which is unique null ( resp .
absolute ) soft set , called full null ( resp .
full absolute ) soft set .
[ th_fn_properties]let @xmath73 @xmath74 , @xmath90 @xmath91 and @xmath92 be mappings .
then for soft sets @xmath93 @xmath94 and a family of soft sets @xmath95 in the soft class @xmath52 we have:@xmath96 @xmath97@xmath98 @xmath99 @xmath100 in general @xmath101@xmath102 @xmath103 in general @xmath104@xmath105 if @xmath106 then @xmath107 we only prove @xmath108@xmath109 for @xmath110 we show that @xmath111 consider@xmath112we consider the case , when @xmath113 as otherwise it is trivial . then@xmath114next , for the non - trivial case , using definition [ df_union_of_images ] and for @xmath115 , we have@xmath116from ( [ eq3 - 1 ] ) and ( [ eq3 - 2 ] ) , we have @xmath117@xmath102 for @xmath110 we show that @xmath118 consider@xmath119where @xmath120 we consider the case when @xmath121 as otherwise it is trivial .
thus@xmath122or @xmath123on the other hand , using definition [ df_union_of_images ] , we have@xmath124ignoring the trivial case , we get@xmath125 \supseteq u\left ( \dbigcup\limits_{\alpha \in p^{-1}\left ( \beta \right ) \cap \left ( a\cap b\right ) } \left ( f\left ( \alpha \right ) \cap g\left ( \alpha \right ) \right ) \right ) = f\left ( \left ( f , a\right ) ~\widetilde{\cap } ~\left ( g , b\right ) \right ) \beta \text { } \end{gathered}\]]@xmath126@xmath127for @xmath115@xmath128we consider the case when @xmath129 as otherwise it is trivial .
then@xmath130this gives @xmath105 . in theorem [ th_fn_properties ] ,
inequalities @xmath98 and @xmath131 can not be reversed , in general , as is shown in the following : the soft classes @xmath132 and mapping @xmath133 @xmath134 are as defined in example [ ex_illustrating_def_of_fn ] .
then@xmath135this shows that the reversal of inequality @xmath98 is not true.to show that the reversal of @xmath102 does not hold , choose soft sets in @xmath32 as:@xmath136then calculations show that@xmath137 [ th_fn_inv_properties]let @xmath73 @xmath74 , @xmath90 @xmath91 and @xmath92 be mappings .
then for soft sets @xmath93 @xmath94 and a family of soft sets @xmath95 in the soft class @xmath49 we have:@xmath96 @xmath138@xmath98 @xmath139@xmath109 @xmath140 in general @xmath141@xmath102 @xmath142 in general @xmath143@xmath105 if @xmath144 then @xmath145 we only prove @xmath108@xmath109 for @xmath146@xmath147next , using definition [ df_union_of_inv_images ] , we have@xmath148where @xmath149 from ( [ finv - eq3 - 1 ] ) and ( [ finv - eq3 - 2 ] ) , we obtain @xmath109.@xmath102 for @xmath146@xmath150this proves @xmath102.@xmath127for @xmath151 consider@xmath152this gives @xmath153
an important task of a medical expert system is to transform a patient s complaints / symptoms into a set of possible causes and their respective importance from the view point of a medical specialist .
a patient s case may easily be encoded into a soft set .
suppose following is the narration by the patient : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ i have three main complaints viz .
burning in stomach , headache and sleeplessness .
whatever sleep i get , is unrefreshing and semi - conscious .
i mean , i am always aware what is transpiring in the room when asleep .
i have some pain in joints and backbone .
to a less degree i also suffer from depression and anxiety .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ this can be written as the following soft set:@xmath154the medical knowledge may be encoded in the form of a look - up tables .
look - up tables are the computer representation of the notion of mapping in mathematics .
suppose our medical experts have provided us with following knowledge:@xmath155and@xmath156for the sake of ease in mathematical manipulation we denote the symptoms and gradations by symbols as follows:@xmath157and@xmath158 thus we have two soft classes @xmath32 and @xmath159 with @xmath160 , @xmath161 and @xmath162 @xmath163 @xmath32 is the soft class of symptoms and their importance for the patient , and @xmath33 represents causes and medical preference for treatment .
the soft set of patient s narration may be given as:@xmath164as a first task of the medical expert system , stored medical knowledge is to be applied on the given case .
this knowledge , in the language of computer programming , is given as look - up tables .
mappings @xmath34 and @xmath35 are defined as:@xmath165calculations give:@xmath166 a soft set , being a collection of information granules , is the mathematical formulation of approximate reasoning about information systems . in this paper , we define the notion of mapping on soft classes .
several properties of soft images and soft inverse images have been established and supported by examples and counterexamples .
finally , these notions have been applied to the problem of medical diagnosis .
it is hoped that these notions will be useful for the researchers to further promote and advance this research in soft set theory .
z. xiao , l. chen , b. zhong , s. ye , _ recognition for soft information based on the theory of soft sets , _
intl . conf .
services systems and services management 2005 , proc of vol2 issue 13 - 15 , ( june 2005)1104 - 1106 . | in this paper , we define the notion of a mapping on soft classes and study several properties of images and inverse images of soft sets supported by examples and counterexamples .
finally , these notions have been applied to the problem of medical diagnosis in medical expert systems.*keywords :* soft set ; soft class ; mapping on soft classes ; image of soft set ; inverse image of a soft set ; medical diagnosis in medical expert systems . _ _ |
there has been considerable recent interest in the study of the endpoint spectra of charmless semileptonic and radiative @xmath0 meson decays .
this effort is motivated by the fact that in order to remove overwhelming backgrounds due to decays to charmed states , stringent cuts must be applied to the data , which restrict the experimental analysis to within a few hundred mev of the kinematic endpoint .
hence it is crucial to understand in as much detail as possible the theoretical shape of the lepton or photon energy spectrum in the endpoint region , if one is to use these processes to extract reliably short - distance couplings such as kobayashi - maskawa matrix elements .
the current theoretical analysis makes use of the operator product expansion ( ope ) and the heavy quark effective theory ( hqet ) @xcite . within this context
one may compute a variety of corrections to the simple free quark decay picture , both perturbative @xmath2 and nonperturbative @xmath3 in origin .
an important result of this analysis is that the leading nonperturbative power corrections in the endpoint region may be resummed into a universal shape function , which describes the distribution of the light - cone momentum of the @xmath4 quark inside the @xmath0 meson @xcite .
since the same nonperturbative matrix elements describe the endpoints of both the lepton spectrum in charmless semileptonic decays and the photon spectrum in radiative decays , it is possible in principle to use a measurement of the photon spectrum to predict the shape of the lepton spectrum and thereby allow for a model - independent measurement of @xmath1 @xcite .
this relation is useful , of course , only if the dominant contribution to the shape of the lepton endpoint spectrum actually comes from the nonperturbative power corrections .
one possible source of trouble is radiative corrections , which near the endpoint suppress the theoretical cross - section by a factor @xmath5 $ ] , where @xmath6 is the scaled lepton energy . for @xmath7 sufficiently close to 1
, this sudakov suppression dominates the theoretical spectrum ; whether this is true over the entire experimentally allowed window is less clear @xcite .
there has been recent progress toward resumming the leading and subleading sudakov logarithms , which would reduce considerably the uncertainty due to this effect @xcite .
another potential source of large corrections near @xmath8 is instanton effects .
chay and rey @xcite have recently computed the one - instanton contribution to inclusive @xmath0 decays , in the dilute gas approximation ( dga ) .
their conclusion was that for charmless semileptonic decays this contribution diverges severely at @xmath8 , while it is small and under control for radiative decays .
unfortunately , their suggestion that one regulate this divergence by considering the energy spectrum only in the region @xmath9 , where @xmath10 , is not necessarily practical , given that the experimental analysis is restricted kinematically to the region @xmath11 . in the region of experimental interest , the effect of instantons is potentially large and dangerous .
unfortunately , it is also the region in which the dga begins to break down and multi - instanton processes become important . in this paper
we will investigate whether instantons spoil the relationship between the radiative and semileptonic endpoint spectra in a way that necessarily destroys its phenomenological usefulness .
we will adopt an approach used in similar situations in the study of baryon number violation in high energy collisions @xcite , in which we use the one - instanton result as a guide to an ansatz for the multi - instanton contribution .
this ansatz contains a small number of physical parameters , and we will investigate the size of instanton effects as a function of these parameters .
we will consider both the overall magnitude of the instanton contribution and the order - by - order behaviour of its moments , as compared to the nonperturbative corrections which arrive from higher order terms in the ope .
the limitations of such an approach are clear .
we will be dealing not with the true multi - instanton cross - section , which has not been computed , but with an ansatz which has been inspired by a one - instanton calculation which is valid in a different region .
nonetheless , we will come to conclusions which we believe are robust , and which indicate that large instanton corrections to the shape of the endpoint spectrum may be difficult to avoid .
we begin by summarizing the calculation of chay and rey @xcite of the effect of a single instanton on the lepton and photon energy spectrum . in the context of the ope
, the decay width is determined by the correlator of two quark bilinears .
for example , for the process @xmath12 , the differential decay rate is given by @xmath13 where @xmath14 is the product of the lepton matrix elements with a lepton phase space measure , and @xmath15 describes the interactions of the quarks @xcite .
the correlator is developed in a simultaneous expansion in @xmath16 and the off shell momentum of the @xmath17 quark , which is of order @xmath18 everywhere but at the boundaries of phase space . in terms of the scaled variables @xmath19 ( @xmath20 in the @xmath0 rest frame ) and @xmath21 , these boundaries are at @xmath22 and @xmath8 .
the calculation of the correlator ( [ wmunu ] ) in the dilute instanton background gives the instanton contribution to the decay width @xcite .
the instanton contribution enters as a contribution to the coefficients of the operators which appear in the ope .
the computation involves an integration over the instanton size @xmath23 , which diverges in the infrared .
chay and rey deal with this divergence by expanding the integrand in @xmath24 and interpreting the finite number of divergent terms as contributions to the matrix elements of operators in the ope .
this is appropriate insofar as the divergent contribution of large instantons is presumably regulated physically by the infrared growth of the qcd self - coupling .
the terms which are infrared convergent and hence calculable are interpreted as contributions to the coefficient functions in the ope . with some mild additional approximations , chay and rey
derive an expression for the leading one - instanton contribution to the doubly differential decay width , @xmath25 where @xmath26 .
the constant @xmath27 depends on the quark masses , the qcd scale @xmath28 and the number of light flavors , and is estimated by chay and rey to be @xmath29 for @xmath30mev and @xmath31gev .
note that the expression ( [ inst1 ] ) scales navely as @xmath32 and has strong divergences as @xmath33 and @xmath34 .
for the radiative decay @xmath35 , under the same approximations , chay and rey find the one - instanton contribution to the photon energy spectrum , @xmath36 where @xmath37 is the scaled photon energy in the @xmath0 rest frame , and @xmath38 is the lowest order free quark radiative decay width .
again , the instanton contribution diverges strongly as @xmath39 .
however , only integrals of eq .
( [ inst2 ] ) are actually meaningful , as we discuss below .
the divergent behaviour of the instanton contribution at the edges of phase space has a straightforward origin .
once the infrared divergences have been subtracted , the contribution of instantons to the coefficient functions comes from small instantons of typical size @xmath40 , where the scale @xmath41 is determined by the momentum of the final quark propagating in the instanton background .
since the instanton contribution contains the suppression factor @xmath42 , it is important only when @xmath43 is of order one , that is , when the invariant mass of the hadronic final state is of order the qcd scale rather than the bottom mass .
this occurs at the boundaries of phase space , where the final state light quark is driven to its mass shell .
hence this is the region where instanton effects become not only significant , but divergent
. an alternative approach to this calculation would be to cut off all instantons with @xmath44 , which would suppress the one - instanton contribution by @xmath45 , independent of @xmath7 .
such a cutoff might be natural for a diagram in which all propagators are in the instanton background , such as the calculation of the polarization operator in @xmath46 annihilation .
however , for semileptonic and rare @xmath0 decays the instanton insertions are only in one propagator , and the choice of cutoff should be governed solely by the kinematics of this light quark , not by the total energy released in the decay .
effectively , we cut off the integral over @xmath23 at @xmath47 , leading to a differential rate which is unsuppressed at the lepton energy endpoint . essentially , the question is whether the integral over @xmath23 should be performed before or after the integration over loop momenta .
if one performs the loop integration first , then the only remaining _ external _ momentum scale is set by @xmath18 , and a cut on @xmath44 would seem natural . if , however , we perform the integral over @xmath23 first , then a loop momentum - dependent cutoff prescription such as we use becomes possible , and in this case , for the reasons discussed above , is preferred . in the case of @xmath48 decays , it is possible to regulate the divergence in eq .
( [ inst2 ] ) in such a way that the contribution of instantons to the _ total _ decay rate is finite and , in fact , negligibly small @xcite .
the situation is similar to that of the contributions of sudakov double logarithms to this process @xcite ; so long as the cut @xmath49 on the photon momentum is not too stringent ( @xmath50 will do ) , it is possible to analytically continue the phase space integral away from any resonance region .
once this has been done , the one - instanton contribution is small , and calculable , everywhere .
hence , the one - instanton result gives a reliable estimate of instanton contributions to integrated quantities such as the total decay width .
unfortunately , this procedure will not work for @xmath12 decays . in this case one must perform an integration in the variable @xmath51 , and near the boundary of phase space
the contour can not be deformed away from the resonance region @xcite .
( this happens both because the endpoint of the integration is a function of @xmath7 and @xmath52 and can not be adjusted by hand , and because in certain regions of phase space the contour is pinched between two cuts . ) as a result , in this regime we do expect a large instanton contribution . of course , in this region the ope itself breaks down , as corrections from operators of higher twist become important .
however , the one - instanton correction has such severe divergences in this region that instantons become important even when @xmath53 , so it is not unreasonable to expect that the ope analysis gives the correct order of magnitude of instanton effects . even in the absence of nonuniversal
higher twist contributions , instantons contribute differently to semileptonic and rare @xmath0 decays .
the reason is essentially the dependence , in the semileptonic case , of the boundary point of the contour of integration on the kinematics of the leptons .
such a dependence does not affect universality in the absence of instanton effects , because the ope is an expansion in inverse powers of @xmath54 and therefore has a pole behaviour , which is insensitve to the shape of the contour of integration .
instanton zero modes , by contrast , give @xmath55 contributions to coefficient functions , so instanton corrections depend explicitly on the position of the boundary point of the contour .
the calculation discussed in the previous section shows that the one - instanton contribution to the semileptonic differential decay rate becomes of the same size as the lowest order result in the region @xmath56 , and diverges in the limit @xmath33 . in this region , then , we can not trust any more the one - instanton result , and we must include multi - instanton corrections as well . unfortunately , there is not at present a technology for performing such calculations .
the situation is similar to one obtaining in the study of high energy collisions , where the one - instanton correction to the total cross section grows exponentially with energy and violates unitarity in the multi - tev region .
possible ways of treating this problem have been discussed widely in the literature ( for a review , see ref .
most likely , multi - instanton corrections stop this dangerous growth and unitarize the amplitude at high energies .
while the detailed behaviour of the multi - instanton contribution is , of course , not known , one makes a hypothesis as to its qualitative form .
it is assumed that the instanton - mediated cross section has a threshold behaviour : it is dominated by the one - instanton contribution below the threshold , and hence exponentially suppressed ; it reaches the unitarity bound in the threshold region ; and it stays almost constant above threshold , in the multi - instanton regime .
these properties lead to an ansatz for the full instanton contribution , which is a step function with support above the threshold @xcite .
this ansatz has two parameters : the width of the step , corresponding to the position of the threshold , and the height , corresponding to the strength of the unitarized amplitude .
while such an ansatz is obviously extremely crude , it incorporates the one useful piece of information which may be extracted from the one - instanton calculation : the energy at which the instanton contributions become large . because of the rapid rise of the one - instanton contribution with energy ,
this threshold is actually predicted fairly reliably .
we will follow an analogous procedure in our discussion of the instanton correction to the semileptonic spectrum .
we will arrive at an ansatz for the multi - instanton contribution which is equally crude , but which we hope will again incorporate correctly the information provided by the one - instanton calculation . to try to account as honestly as possible for the large and uncontrolled uncertainty in our ansatz
, we will vary the parameters which define it over ranges which , in our opinion , are quite generous .
we note briefly the claim in the literature @xcite that the one - instanton result actually unitarizes `` prematurely '' , with its growth stabilized when it is still exponentially small .
this proposal is still somewhat controversial @xcite ; if correct , it will result in a strong suppression of all instanton effects .
we will not address this issue further , except to note that in our ansatz we allow for a significant variation in the overall normalization .
this normalization can be taken in principle to include the effect of premature unitarization , although if the phenomenon is real then it may lead to a stronger suppression than we consider below .
we begin with the expression ( [ inst1 ] ) for the one - instanton contribution to the doubly differential decay rate .
this expression is severely divergent as @xmath57 and @xmath33 , and particularly so when both limits are taken simultaneously .
let us suppose , then , that we believe the one - instanton calculation only over the region @xmath58 , where @xmath59 . outside of @xmath60 ,
multi - instanton configurations probably regularize the otherwise divergent decay rate , and the largest instanton contributions to the rate actually come from this region at the boundary of phase space .
hence we replace the one - instanton result by a step function ansatz , as shown in fig .
[ phasespace ] . in region 1 of fig .
[ phasespace ] , we take as our ansatz eq .
( [ inst1 ] ) , with @xmath61 ; in region 2 , we take eq .
( [ inst1 ] ) with @xmath62 ; and in region 3 , we take eq .
( [ inst1 ] ) with @xmath62 and @xmath63 .
elsewhere , the instanton - mediated decay rate is dominated by the small one - instanton contribution , and is taken to vanish .
next , we integrate this ansatz over @xmath52 to get a preliminary ansatz for @xmath64 . to within a factor of two for @xmath65
, we find the simple result @xmath66 since the function falls steeply for @xmath9 , we make the further simplification of setting the function to zero in this region . the final form for our ansatz , then , is a step function which takes the form @xmath67 the height @xmath68 of the step is given by the second case of eq .
( [ ansatzprelim ] ) .
we stress that this ansatz for the multi - instanton contribution to the width is extremely crude .
however , it does contain certain useful information gleaned from the one - instanton result .
because of the extremely strong dependence of @xmath23 on @xmath69 , we actually find a reasonable constraint on the width .
we assume that multi - instanton contributions come in and regularize the width when @xmath23 gets larger than some value @xmath70 , and let us take the very generous range @xmath71
. then we find , from the expression ( [ ansatzprelim ] ) , that @xmath72 .
this is the range of @xmath69 which we will allow in eq .
( [ ansatz ] ) .
it is somewhat more difficult to set a reasonable range for the area @xmath73 .
we would like to allow for the significant uncertainty in the derivation of eq .
( [ ansatzprelim ] ) , without losing entirely the strong dependence of @xmath74 on @xmath69 , which is physical .
our prescription will be to introduce an _ ad hoc _ multiplicative factor in the normalization of the ansatz , and to consider a wide variation in its value .
hence we will consider the functions @xmath75 where @xmath76 , @xmath77 , and @xmath78 .
these three functions cover a variation of two orders of magnitude in the true size of the multi - instanton contribution , compared with our nave ansatz ( [ ansatz ] ) .
combined with the restriction @xmath72 obtained above , we find @xmath79 . while the upper limit is not to be taken seriously , we believe that @xmath80 represents a reasonable lower limit on the size of multi - instanton effects .
given this ansatz for the multi - instanton contribution , what are the implications for the measurement of @xmath1 ? when instanton effects are included , do they dominate the shape of the endpoint spectrum , or is this shape still determined by the nonperturbative power corrections ? in order to frame this question properly , we must consider both the overall size of the instanton contribution and its effect on the moments of the spectrum .
the leading power corrections to the lepton energy spectrum may be expanded in a series of singular functions near @xmath8 , @xmath81 where @xmath82 .
the singular parts of this expression may be resummed into a `` shape function '' @xmath83 of width @xmath84 , so eq . ( [ exp1 ] ) takes the form @xcite @xmath85\,,\ ] ] where @xmath86 is a smooth function of @xmath7 .
it is convenient to define moments of the shape function , @xmath87 for @xmath88 .
the moments scale inversely with the bottom mass , @xmath89 .
if the spectrum is smeared with a weighting function of width @xmath90 near @xmath8 , for example @xmath91 $ ] , then the result may be written as a sum of these moments , @xmath92 it is this sum which is universal , in the sense that it appears both in the expression for the semileptonic endpoint spectrum and that for the photon energy spectrum in @xmath48 transitions .
if the smearing region @xmath90 is of order @xmath84 , then all terms in the sum ( [ momentsum ] ) are of the same order in the @xmath93 expansion @xcite .
instanton corrections appear as an additional term in the shape function , @xmath94 the new contribution to the right hand side of eq .
( [ momentsum ] ) takes the form @xmath95 where @xmath96 in analogy with eq .
( [ momdef ] ) .
universality will continue to hold if this new nonuniversal term is subleading , in some sense , compared to the universal series generated by the power corrections .
in fact , there are two criteria which we must impose .
the simplest is the condition that the total instanton contribution be small compared to the parton model rate .
how small this ought to be is somewhat a matter of taste . as an illustration , let us take @xmath97 to be a gaussian weighting function of width @xmath90 .
then the parton model rate is given by @xmath98 while the contribution of the instantons depends on the ansatz parameter @xmath69 , @xmath99 if we require the inequality @xmath100 for some @xmath101 , then we obtain lower limits on the smearing region @xmath90 as a function of @xmath69 .
these limits depend also on the normalization @xmath102 of the multi - instanton ansatz . in fig .
[ sigmacurves ] we show @xmath103 for @xmath76 , @xmath77 , and @xmath78 , and for @xmath104 ( a somewhat strict condition ) and @xmath105 ( a much looser one ) . the second criterion is somewhat more subtle , and concerns the behaviour of the moments at large values of @xmath106 . that the moments @xmath107 scale as @xmath108 follows immediately from the form of the operator product expansion @xcite . the instanton contribution , by contrast , is not an expansion in the inverse power of a momentum , and there is no reason for the instanton moments @xmath109 to show such a behaviour .
indeed , since the one - instanton result grows so steeply as @xmath33 , its moments at large @xmath106 may be large compared to the parton model result , even if they are suppressed at small @xmath106 .
the same is true of our multi - instanton ansatz ; a simple calculation yields @xmath110 the instanton moments are only subleading compared to the power corrections at large @xmath106 if @xmath111 .
since the natural width of the theoretical shape function @xmath83 is @xmath84 , this is just the condition that the instantons are concentrated in a region closer to the endpoint than the smearing given by the initial @xmath4 quark motion in the @xmath0 meson .
if so , then we can neglect not only the total instanton contribution but also the contribution of the instantons to the shape of the endpoint spectrum .
the estimate of @xmath69 which we obtained previously , @xmath72 , does not always satisfy this condition , except for quite large values of the qcd scale , @xmath112gev .
we can resolve this problem only by adjusting the width @xmath90 of the smearing function @xmath97 .
let us suppose that the dependence of @xmath90 on the bottom mass is given by @xmath113 where @xmath114 . as discussed in refs .
@xcite , if @xmath115 then all of the terms in the series ( [ momentsum ] ) are of the same order in @xmath84 , since @xmath116 and @xmath117 . for @xmath118 ,
all terms with @xmath119 are suppressed by at least @xmath84 and may be neglected , since they are of the same order as other terms which were dropped earlier .
hence @xmath115 provides a lower , but not an upper , limit on the size of the smearing region , as a function of @xmath18 .
now suppose that the condition @xmath111 is not satisfied
. then the higher moments @xmath107 are power - suppressed compared to @xmath120 , and there is in general an integer @xmath121 such that for @xmath122 we have @xmath123 .
if we insist on a smearing region size corresponding to @xmath115 , for which moments at all @xmath106 contribute equally to the sum ( [ momentsum ] ) , then this situation will obviously lead to trouble .
since for large @xmath106 the sum of moments is dominated by the instanton terms rather than those from the parton level , there will be no useful relation between the smeared spectra in radiative and rare semileptonic decays .
certainly , it will still be possible to use the observed photon spectrum in @xmath48 to predict certain contributions to the lepton energy spectrum in @xmath12 , but these contributions will not be the dominant ones . instead , the uncertainty in the shape of the lepton energy spectrum will be dominated by the uncertainty in the contributions of multi - instanton processes , which , as we have seen , is very large indeed . instead , we must smear over a larger region about @xmath8 , corresponding to an exponent @xmath118 in eq .
( [ epsdef ] ) . if we do so , then all moments with @xmath119 are subleading and may be ignored .
hence if we choose @xmath124 , then by the time @xmath125 , both @xmath109 and @xmath107 may be neglected .
in order to suppress the high moments of the instanton contribution , then , we are led to require a smearing width @xmath90 which might be significantly larger than previously expected .
the critical smearing depends on the size of the multi - instanton ansatz , and is parameterized by a function @xmath126 .
we can estimate @xmath127 by comparing our multi - instanton ansatz to the result which is obtained in the accmm model @xcite .
the moments @xmath107 in this model have been calculated by neubert @xcite , with the result @xmath128 for @xmath106 even .
a best fit to the spectrum yields @xmath129mev , and we take @xmath130gev .
( because of the symmetries of the model , @xmath107 vanishes for @xmath106 odd .
the model also has the curious feature that the moments exhibit an @xmath131 growth , which alters somewhat the condition on @xmath69 .
however , we remind the reader that the model is being used only as a somewhat crude comparison to an equally crude ansatz . ) we obtain a value of @xmath132 by comparing the moments in eqs .
( [ instmoms ] ) and ( [ accmmmoms ] ) , and from this the critical smearing exponent @xmath126 .
the result in the accmm model , however , is that even under the worst assumptions , @xmath133 , so this effect turns out to be relatively unimportant . in fact
, this is not so surprising : in our multi - instanton ansatz , we explicitly have cut off the strong divergence of the instanton contribution near the endpoint , so the effect of instantons on the shape of the endpoint spectrum is less important than their total contribution to the decay rate .
do the limits @xmath103 , summarized in fig .
[ sigmacurves ] , indicate that instantons constitute an important effect on the endpoint spectrum ?
the answer to such a question depends crucially on precisely how it is posed .
it is clear that since we are working with an extremely crude ansatz for the multi - instanton contribution , no numerical prediction which results is to be believed .
however , the real hope of this analysis was to show that instanton effects are sufficiently _ negligible _ that the proposal to measure the shape of the photon energy spectrum in radiative @xmath0 decays and use it to predict the shape of the endpoint spectrum in charmless semileptonic @xmath0 decays is left unaffected . for example , given the experimental constraints , the smearing region must satisfy @xmath134 . if we now focus on the nave normalization of the ansatz , @xmath76 , and on the `` loose '' criterion @xmath105 , we see that only for @xmath135 is @xmath136 . for @xmath137 ,
the instanton effect dominates the nonperturbative shape function over the experimental smearing region . but recall from section iii that the one - instanton result indicates that the entire region @xmath72 is likely to be allowed .
note that our ansatz hardly has to be pushed to its extremes for the relationship between the shapes of the radiative and charmless semileptonic spectra to be spoiled .
if we apply the `` strict criterion '' @xmath104 , then only the curve with the suppression factor @xmath78 is acceptable over the entire range @xmath72 .
we conclude , then , that _ there is no reason to believe that it is safe to neglect instantons in the analysis of the lepton energy endpoint spectrum . _ hence , we would have no particular confidence in the result of the proposed program to measure @xmath1 by comparing endpoint spectra , were it ever to be performed .
we stress that our result is interesting and important only in the _ negative _ sense . by no means
do we claim to have calculated the effect of multi - instanton configurations , or even to have estimated them with any particular accuracy .
what we have done , instead , is to analyze a well - motivated multi - instanton ansatz honestly and conservatively . within this ansatz
, we have not found instanton effects uniformly to be negligible , from which we have concluded that neither are they _ necessarily _ negligible in the real world .
we note that we would have reached a different conclusion , had considerable variation in the ansatz parameters yielded consistently negligible results . of course
, one may take the point of view that instanton effects could as easily be tiny as large , and that in the absence of better evidence one should proceed on this more hopeful assumption .
however , this nave approach , if applied to the extraction of @xmath1 , would lead to a systematic uncertainty which is unknown and probably unknowable . on the other hand , perhaps our ansatz can be refined , or improved , or replaced by something closer to the truth . eventually , perhaps , effects such as we have considered here may even be calculable .
we certainly hope that such advances will one day prove instantons to be unimportant to the semileptonic endpoint spectrum , and the proposed experimental analysis to be unaffected .
but unless good reasons arise for such confidence to be restored , we will consider the measurement of @xmath1 by the detailed analysis of the lepton energy endpoint spectrum to be intrinsically uncertain and untrustworthy .
this work was supported by the national science foundation under grant no .
a.f . also acknowledges the national science foundation for national young investigator award no .
phy-9457916 , and the department of energy for outstanding junior investigator award no .
de - fg02 - 094er40869 .
bigi , m. shifman , n.g .
uraltsev and a.i .
vainshtein , phys.rev .
lett . , 496 ( 1993 )
bigi , n.g .
uraltsev and a.i .
vainshtein , phys.lett .
* b293 * , 430 ( 1992 ) ; erratum , phys .
* b297 * , 477 ( 1993 ) ; b. blok , l. koyrakh , m. shifman and a.i .
vainshtein , phys . rev .
, 3356 ( 1994 ) . | a recent calculation by chay and rey has shown that instantons may make a significant contribution to the lepton energy spectrum near its endpoint . using an ansatz borrowed from the study of high energy baryon number violating processes
, we investigate whether these corrections could spoil the relation between the nonperturbative contributions to this spectrum and to the photon energy spectrum in radiative @xmath0 decays .
we find , in general , that this universality may well fail unless the spectrum is smeared over a region which is considerably larger than had previously been thought necessary .
this result affects the possibility of performing a reliable measurement of @xmath1 using inclusive decays . |
majorana fermions are ubiquitous for superconductors and fermionic superfluids .
the bogoliubov- de gennes equation for fermionic bogoliubov - nambu quasiparticles can be brought to a real form by unitary transformation .
this implies the linear relation between the particle and antiparticle field operators , which is the hallmark of a majorana fermion .
the fermionic statistics and cooper pair correlations give rise to majorana fermions , irrespective of geometry , dimensionality , symmetry and topology @xcite .
the role of topology is to protect gapless majorana fermions , which play the major role at low temperature , when the gapped degrees of freedom are frozen out . for some combinations of geometry , dimensionality and symmetry
these majorana fermions behave as emergent massless relativistic particles .
this suggests that majorana fermions may serve as building blocks for construction of the weyl particles of standard model @xcite .
here we consider the gapless majorana fermions , which appear as andreev bound states on the surfaces of superfluids and on topological objects in superfluids : quantized vortices , solitons and domain walls . in all cases the bound states are formed due to the subsequent andreev reflections of particles and holes .
the key factor for the formation of abs on the small defect with the size of the order of coherence length is a non - trivial phase difference of the order parameter at the opposite ends of particle trajectory . in general
it depends on the structure of the order parameter in real and momentum space which can be rater complicated .
the possibilities for the formation of abs are rather diverse , several of them are shown in in fig.[fig : andreevreflection ] .
the particularly interesting are the case when abs are topologically stable , which means that they have stable zero - energy majorana modes which can not be eliminated by the small perturbation of the system parameters .
schematically shown formation of andreev bound states localized ( a ) on domain wall , ( b ) on the edge , ( c ) inside the vortex core . in all cases
the mechanism is the subsequent particle - hole conversions via andreev reflections at the opposite ends of the trajectory @xmath2 .
the reflected particle ( hole ) picks up the phase of the order parameter @xmath3 and flips the group velocity direction @xmath4 as shown in the panel ( a ) . in general
the wave vectors of particle and hole in the bulk are slightly different @xmath5 where @xmath6 and @xmath7 are fermi momentum and velocity , @xmath8 is the energy . in case
if the order parameter phase difference is @xmath9 the closed loop can be formed even for @xmath10 , that is for the zero energy @xmath11 . in cases ( b
, c ) the phase difference appears due to the momentum dependence of the gap function and the phase winding around the vortex core correspondingly . ]
general properties of fermionic spectrum in condensed matter and particle physics are determined by topology of the ground state ( vacuum ) .
the classification schemes based on topology @xcite suggest the classes of topological insulators , fully gapped topological superfluids / superconductors and gapless topological media . in refs .
@xcite the classification is based on topological properties of matrix green s function , while the other schemes explore the properties of single particle hamiltonian and thus are applicable only to systems of free ( non - interacting ) fermions . among the fully gapped topological superfluids
there is the time - reversal invariant superfluid @xmath1he - b , thin films of chiral superfluid @xmath1he - a and thin films of the time - reversal invariant planar phase of superfluid @xmath1he .
the main signature of topologically nontrivial vacua with the energy gap in bulk is the existence of zero - energy edge states on the boundary , at the interface between topologically distinct domains @xcite and in the vortex cores @xcite . for superfluids and superconductors
these are andreev - majorana bound states ( ambs ) .
these are mainly propagating fermionic quasiparticles , which have relativistic spectrum at low energy @xcite .
however , for special geometries and dimensions , the andreev - majorana bound state represents the isolated non - propagating midgap state , called the majorana zero mode ( or majorino @xcite ) .
it is not a fermion , since it obeys a non - abelian exchange statistics @xcite .
this in particular occurs for the ambs in the vortex core of chiral @xmath0-wave superfluid - superconductor in 2 + 1 dimensions @xcite .
the gapless ambs takes place also on the surfaces , interfaces and in the vortex cores of the gapless topological media . among them
there are chiral superfluid @xmath1he - a with weyl points , the time - reversal invariant planar phase with dirac points and the time - reversal invariant polar phase with line of zeroes . the spectrum of andreev - majorana bound states there is non - relativistic and exotic : the zeroes of ambs spectrum form fermi arcs @xcite and flat bands @xcite .
the @xmath0-wave superfluid @xmath1he has been discovered in 1972 . however till now there is little understanding of superfluid @xmath1he films .
the information on recent experiments in confined geometry can be found in review @xcite . in thin films
the competition is expected between the chiral superfluid @xmath1he - a and time - reversal invariant planar phase , both acquiring the gap in the spectrum in quasi - two - dimensional case due to transverse quantization .
the fermionic spectra in both the 2d a phase and the planar phase have non - trivial topological properties .
these topological states provide the examples of systems featuring generic topological phenomena .
in particular the analog of integer quantum hall effect exist in the 2d a phase where the internal orbital momentum of cooper pairs plays the role of time reversal symmetry breaking magnetic field . in the time reversal invariant planar phase
the quantum spin hall effect can be realized . in close analogy with the 2d electronic systems
the topological invariant is determined by the number of fermionic edge modes with zero energy . in the superfluid systems
the edge zero modes are the abs localized at the superfluid / vacuum boundary or at the interfaces and domain walls separating superfluid states with different topological properties .
below we discuss in detail the topological properties and abs for the 2d a phase and the planar phase .
the order parameter in spatially homogeneous time reversal symmetry breaking @xmath1he - a phase is given @xmath12 where @xmath13 is spin pauli matrix and the @xmath14 are momentum projections to the anisotropy plane .
such order parameter describes the triplet cooper pairs with zero spin @xmath15 and non - zero oribital momentum @xmath16 projections onto the anisotropy axis .
the non - zero @xmath17 plays the role of internal magnetic field breaking the time - reversal symmetry of the systems . confined in @xmath18 plane
the 2d state of a phase is a fully gapped system . by the analogy with 2d electronic gase in quantized magnetic field the gapped ground states ( vacua ) in 2 + 1 or quasi 2 + 1 thin films of @xmath1he -
a are characterized by the following topological invariant @xcite : @xmath19 . \label{2
+ 1invariant}\ ] ] here @xmath20 is the green s function matrix , which depends on matsubara frequency @xmath21 ; the integration is over the whole ( 2 + 1)-dimensional momentum - frequency space @xmath22 , or over the brillouin zone and @xmath21 in crystals .
the expression ( [ 2 + 1invariant ] ) is the extension of the tknn invariant invented by thouless _
_ to describe the topological quantization of hall conductance@xcite
. schematic picture of the interface between two films of chiral @xmath23 superfluid with values @xmath24 and @xmath25 of topological invariant ( [ 2 + 1invariant ] ) .
the interface contains chiral ambs with spectrum @xmath26 , which move with group velocity @xmath27 . in general the algebraic sum of branches (
the number of left moving @xmath28 the number of right moving fermions ) is @xmath29 . on the lower panel the chiral branch of spinless ambs
is shown given by eq.([2dspinpolarized ] ) , when @xmath30 and @xmath31 .
for the spinful case in eq.([3he - a_film ] ) there are two anomalous branches of spectrum of edge states @xmath32 which are degenerate over spin .
the chiral branches produce an equilibrium mass current flowing along the interface . ]
the advantage of the topological approach is that one can choose for consideration the simplest form of green s function , which has the same topological properties and can be obtained from the complicated one by continuous deformation . for a single layer of @xmath1he - a film
one can choose @xmath33 where @xmath34 .
the pauli matrices @xmath35 and @xmath36 correspond to the bogoliubov - nambu spin and ordinary spin of @xmath1he atom respectively ; the parameter @xmath37 characterizes the amplitude of the superconducting order parameter .
the weak coupling bcs limit corresponds to @xmath38 . in this limit
one has @xmath39 , where @xmath40 is the gap in the spectrum and @xmath41 is fermi momentum , @xmath42 .
it is also instructive to consider the simplified case when there is only a single spin component , which corresponds to the fully spin polarized @xmath23 superfluid : @xmath43 we call this case as spinless fermions .
the topological invariant ( [ 2 + 1invariant ] ) for the state in eq.([2dspinpolarized ] ) with @xmath44 is @xmath45 , while for the state with @xmath46 one has @xmath47 .
according to the bulk - surface correspondence , at the interface between these two phases there must be the branch of the andreev - majorana edge states , which crosses zero energy level @xcite , see fig .
[ fig : edgestates ] . in the spinful case of eq.([3he - a_film ] ) , both spin components equally contribute to the topological invariant , and one has @xmath48 for @xmath44 and @xmath47 for @xmath46 . therefore there must be two branches of the andreev - majorana edge states , which cross zero energy level .
in general case the algebraic sum of anomalous branches ( the number of left moving minus the number of right moving fermions ) satisfies the index theorem , @xmath49 .
alternative to the 2d chiral a phase in thin films of superfluid @xmath1he the time - reversal invariant planar phase @xcite can become stable . while this phase has not been identified experimentally yet , in recent experiments @xcite strong suppression of the transverse gap has been observed . the order parameter which describes the spatially homogeneous time reversal invariant planar phase has the form @xmath50 . in this phase , the order parameters is anisotropic and vanishes for the @xmath51 direction , transverse to the film .
nevertheless , confined in 2d when @xmath52 this system is gapful .
being time - reversal invariant the planar phase has zero topological invariant of the type given by eq.([2 + 1invariant ] ) .
however it has an extra discrete symmetry , namely a combination of a @xmath53 spin rotation around @xmath54-axis followed by a @xmath55 phase rotation .
this modifies the topological classification , adding extra @xmath56 topological invariant obtained by volovik and yakovenko in ref . .
this invariant gives rise to the intrinsic spin - hall effect illustrated in fig .
[ fig : spinchargehall ] . an extra motivation to study this particular case of the planar phase
is that it can be considered as a corner stone for a dimensional reduction scheme which can be applied to general class - diii topological superconductors . in the next section
we discuss that the topological properties of a 3d system and an embedded ( 2 + 1)d system , which exist in any time - reversal invariant cross section of the momentum space , are connected . as an application of such a reduction we derive a generalized index theorem for 3d topological superconductors , which provides an example of the bulk - boundary correspondence in odd spatial dimensions .
an illustration of the intrinsic spin - current quantum hall effect due to andreev - majorana edge states in the stripe of the planar phase film with topological invariant @xmath57 in eq .
( [ n2 + 1prime ] ) . as distinct from @xmath1he -
[ fig : edgestates ] , the anomalous branches with different spin projections have opposite slopes .
this gives rise to the quantized spin hall effect without magnetic field , instead of the quantized hall effect in @xmath1he - a film @xcite . ] in the single layer case , the simplest expression for the planar phase green s function @xmath58 is determined by @xmath59 this phase is symmetric under time reversal .
the two spin components have opposite chiralities , as can be seen from identity @xmath60 that is why the contributions of the two spin components to the topological invariant ( [ 2 + 1invariant ] ) cancel each other , @xmath47 .
but the planar phase is still topologically non - trivial because of the discrete @xmath61-symmetry between the two spin components in eq.([planar_components ] ) . due to this symmetry
the matrix @xmath62 commutes with the green s function , which allows us to introduce the symmetry protected topological invariant@xcite : @xmath63 . \label{n2
+ 1prime}\ ] ] this invariant is robust to deformations , if the deformations are @xmath64-symmetric . for the state ( [ planar_state ] ) with @xmath44
one has @xmath57 . for the general case of the quasi 2d film with multiple layers of the planar phase
, the invariant @xmath65 belongs to the group @xmath66 .
the magnetic solid state analog of the planar phase is the 2d time reversal invariant topological insulator , which experiences the quantum spin hall effect without external magnetic field @xcite .
[ fig : spinchargehall ] demonstrates andreev - majorana edge states on two boundaries of the stripe of the single layer of planar phase film . as distinct from @xmath1he -
[ fig : edgestates ] , the anomalous branches with different spin projections are not degenerate : they have opposite slopes , which corresponds to the zero value of the invariant @xmath47 in eq.([2 + 1invariant ] ) . in case of superconductor with planar phase symmetry ,
the invariant @xmath65 determines quantization of spin hall effect . in applied voltage @xmath67
the spectra on two boundaries shift in opposite directions , changing the population of branches .
this produces the imbalance in the spin currents carried by edge states on two boundaries , giving rise to the non - zero total spin - current @xmath68 ( current of @xmath54-projection of spin along @xmath69-axis ) .
this is in the origin of quantized spin hall effect in the absence of magnetic field @xcite : @xmath70 in this time reversal invariant system the electric current qhe is absent .
the topological charge @xmath71 in eq.([2 + 1invariant ] ) , which determines quantization of the hall conductance in the absence of magnetic field @xcite , is @xmath47 , and the currents of different spin populations cancel each other .
the mass and spin currents carried by am edge state in @xmath0-wave superfluids have been considered in refs .
fully gapped @xmath72 fermionic systems - topological insulators and topological superconductors - are now under extensive investigation .
the interest to such systems is revived after identification of topological insulators in several compounds @xcite .
these systems are characterized by the gapless fermionic states on the boundary of the bulk insulator or at the interface between different states of the insulator .
historically , the topological insulators with fermionic zero modes at the interface have been introduced in works @xcite .
an example of the fully gapped topological superfluids is the b phase of superfluid @xmath1he .
much attention has been devoted to the investigation of bound fermion states on surface of @xmath1he - b .
the presence of andreev - majorana surface states in @xmath1he - b can be probed through anomalous transverse sound attenuation @xcite , and surface specific heat measurements @xcite .
these am bound states are supported by the non - zero value of the topological invariant in @xmath1he - b @xcite and have two - dimensional relativistic massless dirac spectrum @xcite .
phase diagram of topological states of triplet superfluid of @xmath1he - b type in equation in the plane @xmath73 .
states on the line @xmath74 correspond to the dirac vacua , which hamiltonian is non - compact .
topological charge of the dirac fermions is intermediate between charges of compact @xmath1he - b states .
the line @xmath75 marks topological quantum phase transition , which occurs between the weak coupling @xmath1he - b ( with @xmath44 , @xmath76 and topological charge @xmath57 ) and the strong coupling @xmath1he - b ( with @xmath46 , @xmath76 and @xmath77 ) .
this transition is topologically equivalent to quantum phase transition between dirac vacua with opposite mass parameter @xmath78 .
the gap in the spectrum becomes zero at this transition .
the line @xmath74 separates the states with different asymptotic behavior of the hamiltonian at infinity : @xmath79 .
the transition across this line occurs without closing the gap . ]
the topological superfluid / superconductor of the @xmath1he - b type are described by topological invariant @xmath65 , which is protected by symmetry : @xmath80\ , .
\label{3dtopinvariant_tau}\ ] ] here @xmath81 is the hamiltonian , or in case of interacting system the inverse green s function at zero frequency @xmath82 , and @xmath64 is matrix which commutes or anti - commutes with @xmath81 .
the proper model hamiltonian which has the same topological properties as superfluids / superconductors of the @xmath1he - b class is the following : @xmath83 where @xmath84 and @xmath85 are again the pauli matrices of bogolyubov - nambu spin and nuclear spin correspondingly . the symmetry @xmath64 , which enters the topological invariant @xmath65 in , is represented by the @xmath86 matrix , which anti - commutes with the hamiltonian : it is the combination of time reversal and particle - hole symmetries of @xmath1he - b . in the limit @xmath74 , eq.([eq : hamiltonianheb ] )
transforms to the dirac hamiltonian , where the parameter @xmath37 serves as the speed of light , while @xmath1he - b lives in the opposite limit @xmath87 . the topological phase diagram in the plane of parameters
@xmath88 is in fig .
[ qpt ] .
the mechanism of andreev - majorana bound states formation at the edge of @xmath1he - b is clear from the hamiltonian ( [ eq : hamiltonianheb ] ) .
let us consider the boundary plane at @xmath89 as shown schematically in the fig.([fig : reductionbtoplanar ] ) . then upon the normal reflection of particles and holes from the boundary some components of the gap function in ( [ eq : hamiltonianheb ] ) change sign .
therefore one obtains a non - zero phase of the gap along the effective trajectory as shown in fig.([fig : andreevreflection ] ) .
in particular for the trajectories normal to the boundary @xmath90 the overall gap function changes the sign leading to the formation of the zero - energy state localized at the boundary .
however this is not the whole story . indeed
if one formally assumes that the hamiltonian may have either negative effective mass , @xmath91 , or negative chemical potential , @xmath46 , the exact solution of the spectral problem yields no zero - energy states as will be discussed below .
the hint to the topological origin of the ambs in @xmath1he - b can be obtained from the topological phase diagram in fig .
[ qpt ] , which demonstrates that the system undergoes a topological quantum phase transitions as one changes the sign of chemical potential @xmath92 or effective mass @xmath93 .
the domain wall , which separates the states with different values of @xmath65 , should contain the zero energy states
the andreev - majorana zero modes .
( color online ) dimensional reduction of the surface - states spectral problem in 3d to that in the time - reversal invariant cross section of momentum space @xmath94 .
reduction from the @xmath95 topological superfluid @xmath1he - b results in the @xmath95 topological planar phase at @xmath94 . ]
to prove the existence of the andreev - majorana bound states on the surface of @xmath1he - b or at the interface one can use a dimensional reduction . let us assume that the boundary plane is at @xmath89 , so that the conserved longitudinal momentum projections are @xmath96 . to find the complete spectrum of bound states @xmath97 it is enough to consider a set of 2d spectral problems for the cross sections of momentum space @xmath98 where @xmath99 .
an example of such a dimensional reduction to the plane @xmath94 is shown in fig.[fig : reductionbtoplanar ] .
the @xmath100 hamiltonian in this cross section reduced from the @xmath72 phase exactly coincides with the hamiltonian of the planar phase .
therefore it is classified by the integer - valued topological invariant @xmath65 in eq.([n2 + 1prime ] ) , which can be shown to coincide with the topological invariant @xmath65 of the parent 3d @xmath1he - b phase in eq.([3dtopinvariant_tau ] ) .
the topologically protected andreev - majorana states in @xmath1he - b are thus related to the topologically protected edge states in the @xmath100 planar phase , see details in ref .
@xcite .
let us consider the spectrum of andreev - majorana fermions using the simplest model of the interface between the superfluid and the vacuum , in which the hamiltonian ( [ eq : hamiltonianheb ] ) changes abruptly at the boundary , with the boundary condition @xmath101 .
spectrum of andreev - majorana fermions , localized states on the surface of topological superfluid / superconductor of the @xmath1he - b class ( red solid lines ) for ( a ) @xmath102 and ( b ) @xmath103 .
the spectrum of bound states terminates when it merges with continuous spectrum in bulk ( yellow color ) , whose border is shown by blue dashed line .
the am bound states exist for @xmath104 . ] at low energies @xmath105 their spectrum is helical spectrum , being described by the hamiltonian @xmath106 @xcite .
interestingly an exact solution of the spectral problem demonstrates @xcite that the linear spectrum of ambs exist up to the merging point with the continuous spectrum of delocalized states .
for @xmath76 the exact spectrum of andreev - majorana fermions @xmath107 is shown by the red solid line in fig .
[ fig : spectrum ] for @xmath108 .
the bound states are confined to the region @xmath109 .
they disappear when their spectrum merges with the continuous spectrum in bulk .
the edge of continuous spectrum is shown by blue dashed line in fig . [
fig : spectrum ] .
if @xmath110 the minimum of the bulk energy spectrum increases monotonically with momentum @xmath111 , therefore the bulk gap is @xmath112 if @xmath113 the minimum of the bulk energy is non - monotonic function of @xmath111 having the minimum at @xmath114 where the bulk gap is @xmath115 the line @xmath116 marks the non - topological quantum phase transition the momentum space analog of the higgs transition @xcite , when the mexican hat potential as function of @xmath111 emerges for @xmath117 .
let us first consider the behavior of the spectrum of majorana fermions at the topological transition at which @xmath93 crosses zero .
when @xmath93 approaches zero , @xmath118 , the region of momenta where bound states exist shrinks and finally for @xmath91 , i.e. in the topologically trivial superfluid , no bound states exist any more . simultaneously the gap in bulk , which at small @xmath93 is @xmath119 according to eq .
, decreases with decreasing @xmath93 and nullifies at @xmath120 .
this corresponds to the conventional scenario of the topological quantum phase transition , when at the phase boundary between the two gapped states with different topological numbers the gap is closed .
the same happens at the tqpt occurring when @xmath92 crosses zero ( see phase diagram in fig .
[ qpt ] ) .
now let us consider what happens with bound states in the case if the tqpt occurs in the opposite limit , when @xmath93 changes sign via infinity , i.e. when @xmath121 crosses zero .
this topological transition occurs without closing of the gap . in this case
the bound states formally exist for all @xmath122 even in the limit @xmath123 .
however , in this limit the ultraviolet divergence takes place : the characteristic length scale of the wave function of the bound state @xmath124 .
so , if the tqpt from topologically non - trivial to the trivial insulator ( or superconductor ) occurs without closing the gap , the gapless spectrum of surface states disappears by escaping via ultraviolet .
this limit corresponds to formation of zero of the green s function , @xmath125 .
such scenario is impossible in the models with the bounded hamiltonian @xcite , which takes place in approximation of finite number of crystal bands . on
the other hand green s function zeroes can occur due to the particle interactions . as was found in ref .
@xcite , classifications of interacting and non - interacting fermionic systems do not necessarily coincide .
this is related to zeroes of the green s function , which according to ref .
@xcite contribute to topology alongside with the poles .
due to zeroes the integer topological charge of the interacting system can be changed without closing the energy gap , and it is suggested that this may lead to the occurrence of topological insulators with no fermion zero modes on the interface @xcite .
that is why we expect that the same scenario with escape to the ultraviolet takes place for the interacting systems : if due to zeroes in the green s function the tqpt in bulk occurs without closing the gap , the spectrum of edge states will nevertheless change at the tqpt , and this change occurs via the ultraviolet . finally let us mention , that the magnetic field violates time reversal symmetry , which generically leads to the finite gap ( mass ) in the spectrum of am fermions on the surface . at particular orientation of the magnetic field
there is still the @xmath61 dsicrete symmetry , which supports gapless ambs @xcite .
this symmetry is spontaneously broken at some critical value of magnetic feild , above which the am fermions become massive .
the surface of @xmath1he - b with massive am bound states represents the @xmath100 topological `` insulator '' : it is described by the topological invariant in eq.([2 + 1invariant ] ) .
the line on the surface , which separates the surface domains with different values of this topological invariant , contains @xmath126 gapless am fermions @xcite .
now we move to the am bound states which appear as edge and vortex states in the gapless topological systems . here
the zeroes in bulk lead to the extended zeroes on the surfaces , interfaces and vortex cores .
we start with point zeroes weyl points in chiral superfluids , which produce the lines of zeroes ( fermi arc ) on the surface , and the flat band in the vortex core . the topological origin of am bound states in @xmath72 chiral superfluids can be viewed by extension of the topology of the @xmath100 chiral system in sec .
[ 2dgapped ] to the @xmath72 case .
let us consider for simplicity the spinless fermions , or which is the same , the fermions with a given spin polarization .
then the green s function in eq.([3he - a_film ] ) extended to @xmath72 case is : @xmath127 where @xmath128 .
let us consider @xmath129 as parameter of the @xmath100 system .
then for each @xmath129 , except for @xmath130 , this green s function describes the fully gapped @xmath100 system the `` insulator '' , which is characterized by topological invariant in eq.([2 + 1invariant ] ) : @xmath131\ , .
\end{split } \label{n2 + 1}\ ] ] this insulator is topological for @xmath132 , where @xmath133 , and is topologically trivial for @xmath134 , where @xmath135 . at @xmath130 , the invariant ( [ n2 + 1 ] )
is not determined , since the corresponding @xmath100 system is gapless .
the bulk @xmath72 superfluid @xmath1he - a has two points in the spectrum @xmath136 , where energy is zero , see fig .
( [ arcboundary ] ) .
these nodes in the spectrum are topologically protected , since they represent the monopoles in the berry phase in momentum space and are characterized by the topological invariant in eq.([2 + 1invariant ] ) , where the integration now is over the 3d sphere around the weyl point in the @xmath72 space @xmath137 @xcite . in the vicinity of these points
the fermionic quasiparticles behave as chiral ( left - handed and right - handed ) weyl fermions in particle physics .
that is why such nodes are called the weyl points .
arrows in fig .
( [ arcboundary ] ) show the direction of the effective spin of the weyl fermion .
this spin is parallel to @xmath138 in the vicinity of @xmath139 , which means that the fermions living there are right - handed .
for the left - handed fermions near @xmath140 , their effective spin is anti - parallel to @xmath141 . ) .
however , as distinct from conventional fermi surfaces , this fermi surface has end points , and thus is called the the end points of the fermi arc are determined by projections of the bulk weyl points to the surface . ] . according to the bulk - surface correspondence , at each @xmath129 ,
for which @xmath142 , there should be one branch of andreev - majorana edge states which cross zero energy level , see fig .
[ fig : edgestates ] .
as a result one has the line of zero energy states in the range @xmath143 .
this line represents the fermi surface ( fermi line ) in the two - dimensional momentum space of bound states .
as the conventional fermi surface , it separates the positive and negative energy levels but as distinct from the conventional fermi surface , this fermi surface is not closed .
it has two end points , and this is the reason why this line is called the fermi arc .
the end points of the fermi arc coincide with the projection of the weyl points to the surface .
this is the consequence of bulk surface correspondence in the weyl systems @xcite . for the arbitrary direction of the surface with the angle @xmath144 between the normal to the surface and the axis @xmath54 , the fermi arc is concentrated in the range of momenta @xmath145 .
note that in @xmath1he - a the boundary conditions require @xmath146 . in crystals
, the weyl points can be moved to the boundaries of the brillouin zone , where they annihilate each other . as a result
one obtains the chiral @xmath72 topological insulator or the fully gapped chiral topological superconductor . since @xmath142 for all @xmath129 the topological fermi arc on the boundaries , transforms to the closed topological fermi surface . .
thick arrows show directions of fermi velocity at these fermi arcs . at @xmath52
the velocity is in the same direction , @xmath147 , which demonstrates that both fermi arcs have the same topological charge @xmath148 , which together satisfy the index theorem @xmath149 , in agreement with momentum space topology of weyl points in bulk @xmath1he - a on two sides of the soliton in fig .
[ fermiarcsolitonfig ] .
this leads to discontinuity in the spectrum of bound states at @xmath150 , where the spectrum merges with the bulk spectrum . ]
the similar fermi arcs appear on the domain walls or solitons separating the chiral phases with opposite chiralities .
one has @xmath151 on one side of the soliton / wall and @xmath152 on the other side .
according to the index theorem @xcite , the difference between these two values determines the number of the zero modes at the interface between the 2 + 1 topological insulators for each @xmath132 . as a result
the domain wall and soliton contain two fermi arcs instead of single fermi arc on boundary , see fig .
[ fermiarcsolitonfig ] .
fermi arc on the domain walls in @xmath1he - a @xcite has been considered in refs . .
[ arc ] includes also the bound states with non - zero energy and demonstrates that the fermi arc does represent the piece of the fermi surface , which separates the positive and negative energy levels .
the zero - dimensional point nodes in bulk ( the weyl points ) give rise to the one - dimensional nodes ( lines ) in the spectrum of ambs . in the same manner ,
the 1d nodal lines in bulk give rise to the 2d manifolds of am bound states with zero energy , see fig .
[ polarphase.eps ] .
let us consider the topological origin of such dispersionless spectrum the flat band on example of the polar phase of triplet superfluid / superconductor @xcite .
the hamiltonian for the polar phase is @xmath153 this superconductor obeys the time reversal and space inversion symmetry , and it has a line of zeroes in the form of a ring . have zero energy . the spectrum is shown for @xmath150 . ] for simplicity we consider the spinless fermions , or which is the same , the fully spin polarized fermions , whose hamiltonain is @xmath154 the spectrum of such fermions has the nodal line the ring @xmath155 , @xmath52 .
the stability of this nodal line is determined by the topological invariant protected by symmetry @xmath156 here the integral is along the loop @xmath157 around the nodal line in momentum space , see fig .
[ fig : spiral ] ; and the matrix @xmath158 anticommutes with the hamiltonian .
the winding number around the element of the nodal line is @xmath159 .
now let us consider the momentum @xmath160 as a parameter of the @xmath161 system , then for @xmath162 the system represents the fully gapped state the @xmath161 insulators .
this insulator can be described by the same invariant as in eq.([invariantforline ] ) with the contour of integration chosen parallel to @xmath129 . since at @xmath163 the hamiltonian tends to the same limit , points @xmath164 are equivalent , and the line @xmath165 forms the closed loop .
that is why the integral @xmath166 is integer - valued .
topological invariant @xmath167 in ( [ invariantforline2 ] ) determines the property of the surface bound states of the @xmath161 system at each @xmath160 . due to the bulk - edge correspondence
, the topological 1d insulator must have the surface state with exactly zero energy . since such states exist for any parameter within the circle @xmath168 , one obtains the flat band of am modes in fig .
[ fig : spiral ] _ left _ the continuum of the self conjugate bound states with exactly zero energy , @xmath169 , which are protected by topology .
such modes do not exist for parameters @xmath170 , for which the @xmath161 superfluid is non - topological . in a spinful polar phase with the hamiltonian ( [ eq : hamiltonianpolar ] ) the nodal ring in bulk gives rise to two surface flat bands with opposite chiralities for two directions of spin .
the tiny spin - orbit interaction leads to a small splitting of the andreev - majorana modes . .
if for a given @xmath171 the energy @xmath172 is nonzero for any @xmath129 , then the green s function @xmath173 describes the 1d fully gapped system
`` insulator '' . at each @xmath171 inside the projection of the line to the surface ,
this insulator is topological , since it is described by non - zero topological invariant ( [ invariantforline2 ] ) .
thus for such @xmath171 there is the gapless edge state on the surface .
the manifold of these zero - energy edge state inside the projection forms the flat band . ] in the multilayered graphene , when the number of the graphene layers tends to infinity , and if some small matrix elements are neglected , the formed @xmath72 system has line of zeroes , which also obeys the invariant similar to that in eq.([invariantforline ] ) .
this nodal line has a shape of a spiral @xcite , see fig .
[ fig : spiral ] .
let us again consider the momentum @xmath160 as a parameter of the @xmath161 system , then for @xmath174 , where @xmath175 is the dominating hopping element , the system represents the fully gapped system the @xmath161 insulators .
this insulator can be described by the same invariant as in eq.([invariantforline ] ) with the contour of integration chosen parallel to @xmath129 , i.e. along the 1d brillouin zone at fixed @xmath160 . due to periodic boundary conditions , the points @xmath176 , where @xmath177 is the distance between the layers , are equivalent and the contour of integrations forms the closed loop . as a result
one obtains the integer valued invariant @xmath178 for @xmath179 the @xmath161 insulator is topological , since @xmath180 .
this gives rise to the surface flat band .
since there are no cooper pair correlations , the fermionic bound states within the flat band are not the majorana modes .
the low - energy fermions bound to the vortex core play the main role in the thermodynamics and dynamics of the vortex state in superconductors and fermi - superfluids .
the spectrum of the low - energy bound states in the core of the axisymmetric vortex with winding number @xmath181 was obtained by caroli , de gennes and matricon for the isotropic model of @xmath2-wave superconductor in the weak coupling limit , @xmath182 : @xcite : @xmath183 this spectrum is two - fold degenerate due to spin degrees of freedom .
the integral number @xmath184 is the quantum number related to the angular momentum of the bound state fermions .
the minigap the level spacing @xmath185 corresponds to the angular velocity of the fermionic quasiparticle orbiting about the vortex axis .
the direction of rotation is determined by the sign of the winding number @xmath186 of the vortex .
the level spacing is typically small compared to the energy gap of the quasiparticles outside the core , @xmath187 .
so , in many physical cases the discreteness of @xmath184 can be neglected . in such cases
the spectrum crosses zero energy as a function of continuous angular momentum @xmath17 , and one may consider this as spectrum of quasi zero modes .
the fermions in this 1d `` fermi liquid '' are chiral : the positive energy fermions have a definite sign of the angular momentum @xmath17 .
the number of the branches crossing zero energy as function of continuous @xmath17 obeys the index theorem @xcite .
here we are interested in the fine structure of the spectrum , when its discrete nature is important .
this takes place for example in the ultracold fermionic gases near the feshbach resonance , when @xmath40 is not small .
_ left _ : schematic illustration of spectrum of the fermionic bound states in the core of @xmath188 vortex in @xmath2-wave superconductor . in the weak coupling limit
the lowest branches are equidistant : @xmath189 .
there are no zero energy states
. the spectrum is doubly degenerate over spin .
_ right _ : the spectrum of bound states in the most symmetric vortices in the @xmath0-wave superfluids : the ciral weyl superfluid @xmath1he - a and the time reversal invariant planar phase .
the spectrum is @xmath190 .
the branch with @xmath191 forms the flat band of andreev - majorana modes . ]
let us first consider the @xmath100 space - time and start with the weak coupling limit .
the majorana nature of the bogoliubov particles requires that the spectrum must be symmetric with respect to zero energy , i.e. for each level with energy @xmath8 there must be the level with the energy @xmath192 . for fermions on vortices
such condition is satisfied for two classes of systems . in the systems on the first class
the spectrum of andreev bound states @xmath193 .
vortices in @xmath2-wave superconductors belong to this class .
vortices of the second class have @xmath194 .
they contain the andreev - majorana mode with exactly zero energy level at @xmath191 . in the @xmath100 system
this mode is not propagating and is self - conjugated .
that is why it is called the majorana mode instead of the majorana particle ( see ref .
@xcite ) . let us for simplicity consider the spinless ( or fully spin polarized ) chiral @xmath23 superfluid in @xmath100 space - time , which is described by eq.([2dspinpolarized ] ) . as it was shown in ref .
@xcite the vortices with winding number @xmath188 or @xmath195 belong to the second class : @xmath196 and thus contain a single majorana mode at @xmath191 .
this mode is robust to perturbations , since it is self - conjugated and thus must obey the condition @xmath197 , see also @xcite . for the spinful fermions in eq.([3he - a_film ] ) there are two am modes corresponding to the two spin projections .
the even number of majorana modes is not robust to perturbations .
for example , the spin - orbit interaction splits two modes with @xmath198 .
the splitting is absent , if there is some discrete symmetry between the am modes , such as the mirror symmetry in ref .
@xcite . in the spinful @xmath23
there is a topological object , which carries a single majorana mode .
it is the half - quantum vortex @xcite . in a simple model ,
the half - quantum vortex is the vortex with @xmath199 in one spin component , while the other spin component has zero vorticity @xmath200 . as a result
such vortex contains single majorana mode , which is robust to perturbations
. however , the perturbations should not be too large . in the limit
when @xmath92 is negative and large , the bcs is transformed to the bec of molecules , where the majorana mode is absent .
the majorana mode disappears , when the chemical potential @xmath92 crosses zero . at @xmath75
there is a topological quantum phase transition ( tqpt ) , at which the topological invariant in eq.([2 + 1invariant ] ) changes from @xmath45 to @xmath47 .
the topological transition can not occur adiabatically , and in the intermediate state with @xmath75 , the spectrum in bulk becomes gapless . at @xmath75
the majorana mode merges with the continuous spectrum of bulk quasiparticles and disappears at @xmath46 .
this demonstrates the topological origin of the am mode , which exists inside the vortex only if the vacuum in bulk is topologically nontrivial .
one can easily extend the consideration in sec . [ vortexchiral ] to the @xmath72 case in the weak coupling limit .
the levels at @xmath201 remain equidistant according to the caroli - de gennes - matricon solution and they must be symmetric with respect to @xmath11 . this dictates the following modification of eq.([caroli ] ) for the most symmetric vortices in @xmath1he - a and in the planar phase : @xmath202 this equation suggests the flat band in the vortex core for @xmath191 , see fig .
[ carolifigure ] _
right_. now we show how such flat band emerges purely from the topological considerations , which do not use the weak coupling approximation .
topology of bound states on vortices in @xmath72 chiral superfluids can be obtained by dimensional extension of the topology in the @xmath100 case .
the am mode in the point vortex of fully gapped @xmath100 chiral superfluid transforms to the flat band of am modes inside the vortex line in @xmath72 chiral superfluids with weyl points in bulk .
let us consider again the @xmath23 state in eq.([3he - a_bulk ] ) , and choose for a moment the direction of the vortex line along the axis @xmath54 . in this case
@xmath129 is the quantum number of the bound states in the vortex core . for each @xmath129 in the range @xmath203
the green s function ( [ 3he - a_bulk ] ) describes the @xmath100 chiral superfluid with topological invariant @xmath133 in eq.([n2 + 1 ] ) , and this superfluid contains the point vortex .
the point vortex in the @xmath100 topologically nontrivial chiral superfluid contains the am mode with zero energy .
the continuum of the andreev - majorana modes in the range @xmath203 forms the flat band .
-axis ) determine the boundaries of the flat band in the vortex core .
weyl point in 3d systems represents the hedgehog ( berry phase monopole ) in momentum space @xcite . for each plane
@xmath204 one has the effective 2d system with the fully gapped energy spectrum @xmath205 , except for the planes with @xmath206 , where the energy @xmath205 has a node due to the presence of the hedgehogs in these planes .
topological invariant @xmath207 in is non - zero for latexmath:[$|p_z|
< p_f @xmath129 in this interval the system behaves as a 2d topological insulator or 2d fully gapped topological superfluid .
point vortex in such 2d superfluids has fermionic state with exactly zero energy . for the vortex line in the original 3d system with fermi points
this corresponds to the dispersionless spectrum of fermion zero modes in the whole interval @xmath209 ( thick line ) .
, width=302 ] this is demonstrated in fig .
( [ fig : vortex ] ) , in which the vortex axis is rotated by angle @xmath144 with respect to the direction to the wey points . in this case the invariant @xmath210 such flat band of am modes has been predicted by kopnin and salomaa in ref . for the @xmath188 vortex in @xmath1he - a .
this flat band is doubly degenerate over spin and thus it may split , for example , due to spin - orbit interaction ( the non - degenerate flat band of andreev - majorana fermions takes place in the core of half - quantum vortex ) . in superfluid @xmath1he the spin - orbit interaction is very small and can be neglected
. however , there can be another source of splitting : the symmetry of the vortex core can be spontaneously broken ( see @xcite ) .
the same doubly degenerate flat band should exist in the @xmath188 vortex in the @xmath72 planar phase , where the green s function is @xmath211 here now @xmath212 . for the @xmath72 planar phase
, the topological invariant @xmath65 in ( [ n2 + 1prime ] ) is extended to : @xmath213 , \end{split } \label{nk2 + 1}\ ] ] giving @xmath214 .
both flat bands , in the a - phase and in the planar phase , appear only for @xmath44 , when @xmath215 . for @xmath216
the superfluids are topologically trivial , @xmath217 , and the flat band does not exist .
dimensional extension of the @xmath100 planar phase allows to understand the topological properties of the vortex spectrum in @xmath1he - b .
the hamiltonian ( [ eq : hamiltonianheb ] ) for fermions in the bulk @xmath1he - b represents at @xmath52 the @xmath100 planar phase .
that is why at @xmath52 the @xmath188 vortex in @xmath1he - b contains two am bound states with zero energy , if the tiny spin - orbit interaction is neglected and the core symmetry is not spontaneously broken . for @xmath201
the zero energy modes are not supported by topology .
so the two branches of am modes split , and one may expect the behavior of the spectrum of am bound states in the most symmetric vortex as illustrated in fig .
[ o - vortex]a , b .
( _ a _ ) schematic illustration of spectrum of the fermionic bound states in the core of the of the most symmetric vortex ( @xmath218-vortex ) in @xmath1he - b .
two am states with zero energy exist at @xmath52 .
( _ b _ ) : the same vortex but in the topologically trivial state of the liquid , @xmath77 , does not have fermion zero modes .
the spectrum of bound states is fully gapped .
fermion zero modes disappear at the topological quantum phase transition , which occurs in bulk liquid at @xmath75 .
similar situation may take place for strings in color superconductors in quark matter @xcite . ] for @xmath1he - b , which lives in the range of parameters where @xmath219 in fig .
[ o - vortex]a , the gapless fermions in the core of the most symmetric vortex ( the so - called @xmath218-vortex @xcite ) have been found in ref .
@xcite . on the other hand ,
in the bec limit , when @xmath92 is negative and the bose condensate of molecules takes place , there are no gapless fermions , see fig .
[ o - vortex]b .
thus in the bcs - bec crossover region the spectrum of fermions localized on vortices must be reconstructed .
the topological reconstruction of the fermionic spectrum in the vortex core can not occur adiabatically .
it should occur only during the topological quantum phase transition in bulk , when the bulk gapless state is crossed .
such topological transition occurs at @xmath75 , see fig .
[ qpt ] . at @xmath46
the topological charge @xmath220 nullifies and simultaneously the gap in the spectrum of core fermions arises , see fig .
[ o - vortex]b .
this again demonstrates that the existence of fermion zero modes is closely related to the topological properties of the vacuum state .
the reconstruction of the spectrum of fermion zero modes at the tqpt in bulk can be also seen for vortices in relativistic superconductors @xcite .
spectrum of ambs in the axisymmetric @xmath221-vortex with spontaneously broken discrete symmetry in @xmath1he - b .
the am states with zero energy at @xmath52 , which were present in the most symmetric @xmath218-vortex in fig .
[ o - vortex ] , do not exist any more .
they split due to matrix element between the spin components , which appears due to symmetry breaking and move far away .
there are many non - topological branches of spectrum , which cross zero energy as function of @xmath129 and form the one - dimensional fermi surfaces .
the number of such branches @xmath222 . ] the spectrum in fig .
[ o - vortex ] _ left _ is valid only for the vortex state , which respects all the possible symmetries of the vortex core .
these symmetries are the space parity @xmath223 and the discrete symmetry @xmath224 .
the latter is the symmetry under the time reversal @xmath225 when it is accompanied by the @xmath53-rotation @xmath226 about the axis perpendicular to the vortex axis . in the cores of the experimentally observed vortices in @xmath1he - b both discrete symmetries
are spontaneously broken , while the combined symmetry @xmath227 is preserved @xcite .
such vortex is called the @xmath221-vortex .
the broken parity in the @xmath221-vortex leads to mixing between the two spin components in the core , as a result the two am modes at @xmath52 split .
this leads to the following spectrum in fig .
[ v - vortex ] @xcite . in the weak coupling regime , @xmath38
, there appear the large number ( on the order of @xmath228 ) of branches , which cross zero energy .
each crossing point corresponds to the one - dimensional fermi - surface .
this demonstrates that the topology in bulk determines the spectrum of the fermion zero modes on the b - phase vortices only if the symmetry of the vortex core is not violated .
this is the consequence of the mod 2 rule for majorana modes : the topological zero energy state survives the symmetry breaking only in case of the odd number of majorana modes .
so , for the realistic vortices the am mode can exist only in half - quantum vortices .
for the other vortices , such as in @xmath1he - b , the large number of the energy levels is nvolved .
that is why for the consideration it is more appropriate to use the quasiclassical approximation .
the latter leads to the other types of topological invariants describing the fermion zero modes on vortices , see e.g. refs . .
we considered the andreev - majorana bound states with zero energy on surfaces , interfaces and vortices in different phases of the @xmath0-wave superfluids : @xmath1he - a , @xmath1he - b , planar and polar phases .
these states are determined by topology in bulk , and they disappear at the quantum phase transition from the topological to non - topological state of the superfluid ( see example in fig . [ o - vortex ] ) .
this topology demonstrates the interplay of dimensions .
in particular , the 0d weyl point ( the berry phase monopole in momentum space ) gives rise to the 1d fermi arc on the surface ( sec .
[ fermiarcboundary ] ) .
the 1d nodal line in bulk produces the dispersional 2d band of am modes on the surface ( sec .
[ polarflatband ] ) .
the interplay of dimensions also connects the am states in superfluids in different dimensions .
for example , the property of the spectrum of bound states in the 3d @xmath1he - b is connected to the property of the spectrum in the 2d planar phase ( see sec .
[ ambsgapped ] for edge states and sec .
[ planarbphasevortex ] for bound states on vortices ) . the 0d am mode on a point vortex in 2d chiral superfluid ( sec .
[ vortexchiral ] ) gives rise to the 1d flat band of am modes on a vortex in the 3d chiral superfluid ( sec .
[ vortexflatbandsec ] ) .
the most robust zero energy edge states take place on the boundary of @xmath1he - a , or in general on boundaries and interfaces of chiral superfluids with the topological invariant @xmath71 in eq.([2 + 1invariant ] ) . in the other phases ,
the existence zero energy edge states is supported by symmetry , i.e. by the symmetry protected topological invariants @xmath65 in eqs .
( [ n2 + 1prime ] ) and ( [ invariantforline ] ) .
when the symmetry is violated in bulk or on the boundary / interface , the am bound states acquire gap . concerning the am states on vortices , only the states on half quantum vortices are fully robust to perturbations . in the singly quantized vortices
the fate of zero energy states depend on symmetry and its possible violation in bulk or spontaneous breaking inside the vortex core .
this the consequence of the @xmath61 classification of the am modes on vortices . on the other hand
, the spontaneously broken symmetry inside the vortex core may give rise to many non - topological branches of ambs , which cross zero energy as function of @xmath129 .
this is demonstrated in sec .
[ bbroken ] .
let us also mention the application to the relativistic theories .
the fermion zero modes obtained in the dirac systems , such as the modes localized on strings in ref .
@xcite , are not properly supported by topology .
the reason for that is that the dirac vacuum is marginal , and the topological invariants depend on the regularization of the green s function in the ultraviolet @xcite .
for example , in fig.[qpt ] the dirac vacuum is on the border between the trivial vacuum with @xmath77 and the topological vacuum with @xmath57 .
that is why the existence of the modes with exactly zero energy depends on the behavior of the green s function at infinity .
ms and gev acknowledge financial support by the academy of finland through its ltq coe grant ( project @xmath229250280 ) .
volovik , quantum phase transitions from topology in momentum space , in : `` quantum analogues : from phase transitions to black holes and cosmology '' , eds .
w.g . unruh and r. schtzhold ,
springer lecture notes in physics * 718 * ( 2007 ) , pp . 3173 ; cond - mat/0601372 .
heikkil and g.e .
volovik , dimensional crossover in topological matter : evolution of the multiple dirac point in the layered system to the flat band on the surface , pisma zhetf * 93 * , 6368 ( 2011 ) ; jetp lett .
* 93 * , 5965 ( 2011 ) ; arxiv:1011.4185 .
volovik , flat band in the core of topological defects : bulk - vortex correspondence in topological superfluids with fermi points , pisma zhetf * 93 * , 6972 ( 2011 ) ; jetp lett . * 93 * , 6669 ( 2011 ) ; arxiv:1011.4665 .
l. v. levitin , r. g. bennett , a. casey , b. cowan , j. saunders , d. drung , th .
schurig , j. m. parpia , b. ilic , n. zhelev , study of superfluid @xmath1he under nanoscale confinement : a new approach to the investigation of superfluid @xmath1he films , j. low temp . phys . *
175 * , 667680 ( 2014 ) .
volkov , a.a .
gorbatsevich , yu.v .
kopaev and v.v .
tugushev , macroscopic current states in crystals , jetp * 54 * , 391397 ( 1981 ) ; b.a .
volkov and o.a .
pankratov , two - dimensional massless electrons in an inverted contact , jetp lett . * 42 * , 178181 ( 1985 ) .
s. murakawa , y. tamura , y. wada , m. wasai , m. saitoh , y. aoki , r. nomura , y. okuda , y. nagato , m. yamamoto , s. higashitani , and k. nagai , new anomaly in the transverse acoustic impedance of superfluid @xmath1he - b with a wall coated by several layers of @xmath233he , phys .
lett . * 103 * , 155301 ( 2009 ) .
y. aoki , y. wada , m. saitoh , r. nomura , y. okuda , y. nagato , m. yamamoto , s. higashitani , k. nagai , observation of surface andreev bound states of superfluid @xmath1he by transverse acoustic impedance measurements , phys .
lett . * 95 * , 075301 ( 2005 ) .
makhlin , m. silaev , and g.e .
volovik , topology of the planar phase of superfluid @xmath1he and bulk - boundary correspondence for three - dimensional topological superconductors , phys .
b * 89 * 174502 ( 2014 ) ; arxiv:1312.2677 .
l. fidkowski and a. kitaev , effects of interactions on the topological classification of free fermion systems .
b * 81 * , 134509 ( 2010 ) ; topological phases of fermions in one dimension , phys . rev .
b * 83 * , 075103 ( 2011 ) . | we consider andreev - majorana ( am ) bound states with zero energy on surfaces , interfaces and vortices in different phases of the @xmath0-wave superfluids . we discuss the chiral superfluid @xmath1he - a , and time reversal invariant phases : superfluid @xmath1he - b , planar and polar phases .
the am zero modes are determined by topology in bulk , and they disappear at the quantum phase transition from the topological to non - topological state of the superfluid . the topology demonstrates the interplay of dimensions .
in particular , the zero - dimensional weyl points in chiral superfluids ( the berry phase monopoles in momentum space ) give rise to the one - dimensional fermi arc of am bound states on the surface and to the one - dimensional flat band of am modes in the vortex core .
the one - dimensional nodal line in the polar phase produces the two - dimensional flat band of am modes on the surface .
the interplay of dimensions also connects the am states in superfluids with different dimensions .
for example , the topological properties of the spectrum of bound states in the three - dimensional @xmath1he - b is connected to the properties of the spectrum in the two - dimensional planar phase ( thin film ) . |
the idea that instead of originating from a big bang singularity , the universe has emerged from a cosmological bounce has a long history ( see @xcite for a review with an extensive list of references to the original literature ) . the group of professor novello has made a lot of important contributions to the research on this topic .
however , it was realized only fairly recently @xcite that a bouncing cosmology with a matter - dominated phase of contraction during which scales which are probed today in cosmological observations exit the hubble radius can provide an alternative to the current inflationary universe paradigm of cosmological structure formation . in this review article
, we provide an overview of this matter bounce " scenario of structure formation , and we discuss some recent efforts at obtaining a non - singular bouncing cosmological background ( see also @xcite for reviews comparing the matter bounce with inflation and other alternatives to inflation ) .
inflationary cosmology @xcite ( see also @xcite ) has become the paradigm of early universe cosmology not only because it addresses some of the conceptual problems of standard big bang cosmology such as the horizon , flatness and entropy problems , but because it provided the first causal mechanism for generating the primordial fluctuations which could have developed into the structures we see today on large scales @xcite ( see also @xcite ) .
more specifically , it predicted a roughly scale - invariant spectrum of cosmological fluctuations which in simple models of inflation have a slight red tilt and which are gaussian and nearly adiabatic .
these predictions were spectacularly confirmed in recent precision observations of cosmic microwave background ( cmb ) anisotropies @xcite . on the other hand , it was known since long before the development of inflationary cosmology that any model which produces a roughly scale - invariant and almost adiabatic spectrum of cosmological perturbations will be a good match to observations @xcite .
inflationary cosmology is simply the first model based on fundamental physics which yielded such a spectrum . in the mean time , other models have been developed which predict this kind of spectrum , e.g. the ekpyrotic universe @xcite , string gas cosmology " @xcite , the varying speed of light " proposal @xcite , the conformal universe " @xcite , and the matter bounce scenario . in the following section , i shall show how quantum vacuum fluctuations originating on sub - hubble scales and exiting the hubble radius in a matter - dominated phase of contracting develop into a scale - invariant spectrum of curvature perturbations . as is well known since the pioneering work of hawking and penrose ( see e.g. @xcite for a textbook description ) ) , a cosmological singularity is unavoidable if space - time is described in terms of general relativity ( gr ) and if matter obeys certain energy conditions .
thus , in order to obtain a bouncing cosmology it is necessary to either go beyond einstein gravity , or else to introduce new forms of matter which violate the key energy conditions . in section 3 of this review
i will describe some concrete and recent models which yield a bouncing cosmology ( the reader is referred to @xcite for an overview of early work on obtaining bouncing universes ) .
the space - time background cosmology which we have in mind has time @xmath0 running from @xmath1 to @xmath2 .
the bounce point can be taken to be @xmath3 . for negative times
the universe is contracting . in the absence of entropy production at the bounce point
it is logical to assume that the contracting phase is the mirror inverse of the expanding cosmology of standard big bang cosmology , i.e. at very early times the universe is dominated by pressure - less matter , and at a time @xmath4 ( where @xmath5 is the time of equal matter and radiation in the expanding phase ) there is a transition to a radiation - dominated phase . if there is entropy production at the bounce , then the transition from matter to radiation domination will occur closer to the bounce point than @xmath6 . in figure 1
we show a space - time sketch of the bouncing cosmology background .
the horizontal axis is the comoving spatial coordinate , the vertical is time . the vertical line indicates the wavelength of a fluctuation model .
it should be compared with the comoving hubble radius @xmath7 ( labelled as such in the figure ) . as will be reviewed in the following subsection
, the hubble radius is the maximal scale on which causal and local microphysics can generate fluctuations . in order to have a causal generation mechanism , for fluctuations ,
it is thus important that wavelengths which are observed today originate on sub - hubble scales .
the first important point to take away from the figure is that fixed comoving scales start out with a wavelength smaller than the hubble radius and that hence a causal generation mechanism for fluctuations is possible in a bouncing cosmology , like it is in inflationary cosmology . in inflationary cosmology
it is the exponential decrease of the comoving hubble radius during the inflationary phase which allows for a causal generation mechanism .
we see that in a contracting universe dominated by regular matter a similar decrease in the comoving hubble radius occurs . in this section
we discuss how a scale - invariant spectrum of adiabatic curvature fluctuations emerges from initially vacuum perturbations .
whereas in inflationary cosmology there is a symmetry argument ( time translation invariance during the inflationary phase ) which underlies the scale - invariance of the cosmological fluctuations , in a contracting universe there is no such symmetry . as we will see below , it is only vacuum fluctuations which exit the hubble radius in a phase of matter domination which end up with a scale - invariant spectrum . before studying the generation and evolution of fluctuations in the matter bounce model ,
it is important to discuss to which extent this bouncing cosmology can address the conceptual problems of standard big bang cosmology which the inflationary scenario successfully addresses , namely the horizon " , flatness " , size " and entropy " problems . in the context of standard big bang cosmology
there is no causal explanation for the observed overall isotropy of the cosmic microwave background ( cmb ) since at the time of last scattering regions in the sky separated by more than a couple of degrees were causally disconnected @xcite .
this is the horizon " problem .
inflation solves this problem by providing a mechanism to exponentially increase the causal horizon in the early universe . in a bouncing cosmology
the particle horizon is infinite since time extends to @xmath1 .
hence , the entire spatial hypersurface at the time of recombination which we observe today in the cmb is within the causal horizon and there is no horizon problem . in standard big bang
cosmology spatial flatness is an unstable fixed point under dynamical evolution .
thus , the observed degree of spatial flatness requires extreme fine tuning of the initial conditions . during an inflationary phase spatial flatness becomes an attractor , thus mitigating the flatness problem .
a bouncing cosmology is neutral " with respect to the flatness problem : if we postulate a similar degree of spatial flatness an equal amount of time prior to the bounce point as is observed today a certain time interval after the bounce point , then the observations can be explained .
the size and entropy problems of the standard big bang model are based on the idea that the early universe should be described by initial conditions containing no hierarchy of scales : at the planck temperature the size of space should be planck size and it should contain a planckian value of the entropy . with these initial conditions inflation
is required to generated the large entropy and size observed today . in a bouncing cosmology
the universe starts large and cool .
thus , there are no size and entropy problems .
there is another aspect of the flatness problem which is not addressed in most bouncing cosmologies , namely the anisotropy problem : having a nearly homogeneous and isotropic bounce requires fine tuning of the initial anisotropies . in inflationary cosmology the anisotropies decrease , and thus the anisotropy problem is addressed .
similarly , in string gas cosmology there is no anisotropy problem , as shown in @xcite . in the ekpyrotic bouncing cosmology
@xcite anisotropies decrease in importance in the contracting phase .
a way of realizing the matter bounce in a framework which is safe from the anisotropy problem is to merge ideas from simple matter bounce models with ideas from ekpyrotic cosmology @xcite . the fluctuations which we are interested in are inhomogeneities in the curvature of space - time which are induced by perturbations in matter .
fluctuations observed today on large cosmological scales are small in relative amplitude .
hence , we can describe them by linearizing the einstein equations about the cosmological background solution .
the corresponding theory is called the theory of cosmological perturbations " ( see @xcite for an in - depth survey and @xcite for an overview ) . in the following
we summarize the essentials .
fluctuations in the space - time metric in cosmology can be classified according to how they transform under spatial rotations there are scalar , vector and tensor fluctuations . in total
there are ten modes . however , four of the modes correspond to coordinate transformation and are hence not physical .
there are only two scalar , two vector and two tensor modes which are physical .
the tensor modes correspond to gravitational waves and will not concern us here .
vector modes do not couple at linear order to matter perturbations and hence will not be discussed here . in an expanding universe , vector modes decay , and
this yields an additional reason to neglect them .
however , in a contracting universe they grow @xcite , and thus neglecting them is only justifiable at linear order.the modes of interest to us are the scalar modes , the ones which at linear order couple of matter inhomogeneities . for simple forms of matter such as scalar fields and perfect fluids
there is at linear order no anisotropic stress , and this eliminates one of the scalar modes , leaving one degree of freedom which describes curvature fluctuations . in longitudinal gauge , the metric including the scalar fluctuation mode @xmath8 can be written as ds^2 = a^2 ( ) , where @xmath9 is the cosmological scale factor and we have made use of conformal time @xmath10 related to physical time @xmath0 via @xmath11 . by expanding the full action ( einstein action plus action for matter ) to second order about the classical background cosmology , one can obtain the action for cosmological perturbations which yields the linearized equations of motion .
this action can be canonically quantized . by the logic of the previous paragraph
, the resulting action can be written in terms of a single dynamical variable which , in addition , can be canonically normalized .
the form of the canonical fluctuation variable @xmath12 was derived by mukhanov @xcite and sasaki @xcite ( see also @xcite ) .
the action for @xmath12 takes the following form [ pertact ] s^(2 ) = 1 2 d^4x , where a prime denotes the derivative with respect to conformal time , the subscript @xmath13 stands for the derivative with respect to the inth spatial comoving coordinate , and the function @xmath14 is a function of the cosmological background .
if the equation of state of matter is time - independent , then @xmath14 is proportional to @xmath9 .
for scalar field matter @xmath15 the canonical variable is given by @xmath16 \
, , \ ] ] where @xmath17 is the fluctuation of the matter field , @xmath18 , and @xmath19 the canonical variable @xmath12 is simply related to the variable @xmath20 which describes the curvature fluctuations in the comoving coordinate system @xcite and which we are interested in computing at late times : @xmath21 the equation of motion which follows from the action ( [ pertact ] ) is ( in momentum space ) @xmath22 where @xmath23 is the kth fourier mode of @xmath12 .
we see that each fourier mode satisfies a harmonic oscillator equation of motion with a time - dependent mass , the time dependence being given by the background cosmology .
the mass term in the above equation is in general given by the hubble expansion rate .
thus , we see that the hubble radius plays a key role in the evolution of fluctuations . the mode @xmath24 whose wavelength at time @xmath0 is equal to the hubble radius will be denoted by @xmath25 .
it follows that on small length scales , i.e. for @xmath26 , the solutions for @xmath23 are constant amplitude oscillations .
these oscillations freeze out at hubble radius crossing , i.e. when @xmath27 . on longer scales ( @xmath28 )
, there is a mode of @xmath23 which scales as @xmath29 .
this mode is the dominant one in an expanding universe , but not in a contracting one .
canonical quantization of the action for cosmological perturbations corresponds to imposing canonical commutation relations for each fourier mode @xmath23 .
if we impose vacuum initial conditions at some time @xmath30 , this implies @xmath31 before applying the above formalism to initial vacuum perturbations in the matter bounce scenario , we will review how a scale - invariant spectrum emerges in inflationary cosmology .
the definition of scale - invariance of the curvature power spectrum is p_r k^3 |v_k|^2 ~ k^n_s - 1 ~ const , i.e. @xmath32 , where @xmath33 is called the spectral index of scalar metric fluctuations .
note that the vacuum spectrum , i.e. the spectrum obtained with the values ( [ incond ] ) is not scale invariant .
rather , it is blue with @xmath34 ( more power on short wavelengths ) . in inflationary cosmology
the hubble radius is constant while the wavelength of comoving modes expands exponentially .
thus , provided the period of inflation is sufficiently long , all modes which are currently observed originate on sub - hubble scales during inflation . a mode with wavenumber @xmath24 exits the hubble radius at a time @xmath35 given by @xmath36 in inflationary cosmology , any classical fluctuations existing at the beginning of the period of inflation are red - shifted and ( in the absence of a trans - planckian mechanism which might re - populate the ultraviolet modes - see @xcite for discussions of this point ) leave behind a vacuum state .
thus , it makes sense to start fluctuations on sub - hubble scales in their vacuum .
the fluctuations will oscillate with constant amplitude while on sub - hubble scales .
however , on super - hubble scales @xmath23 will increase as @xmath9 . since long wavelengths spend more time outside the hubble radius they experience a bigger growth .
thus , the final spectrum will be less blue .
when measured at a fixed final time @xmath10 , the increase in the amplitude of @xmath23 will be proportional to @xmath37 , which from ( [ hubble3 ] ) is proportional to @xmath38 . hence , the slope of the power spectrum changes by @xmath39 , converting the vacuum spectrum into a scale - invariant one .
we will now see that a similar mechanism converts a vacuum spectrum into a scale - invariant one in the matter bounce scenario . in this section
we will be following modes which originate as quantum vacuum fluctuations on sub - hubble scales at early times and which cross the hubble radius during the phase of matter dominated contraction .
the vacuum spectrum has index @xmath34 . to convert it into a scale - invariant spectrum we require a mechanism which boosts long wavelengths compared to short wavelengths .
since @xmath23 grows on super - hubble scales , such a mechanism naturally arises in a contracting universe .
as we see below , in a matter - dominated phase of contraction the boost factor is exactly the right one to turn the vacuum spectrum into a scale - invariant one @xcite . from the equation of motion ( [ perteom2 ] )
it follows that @xmath23 will oscillate with constant amplitude until the scale exits the hubble radius .
while the equation of state is constant in time , we have @xmath40 . in a matter - dominated phase we have @xmath41 since @xmath42 .
hence , it follows by solving ( [ perteom2 ] ) in the limit @xmath43 that the dominant mode of @xmath23 scales as @xmath44 .
thus , the power spectrum of curvature fluctuations at some late time @xmath10 when the modes of interest to us are outside the hubble radius becomes p_r(k , ) & ~ & k^3 |v_k()|^2 a^-2 ( ) + & ~ & k^3 |v_k(_h(k))|^2 ( ) ^2 + & ~ & k^3 - 1 - 2 ~ const , where in the first line we have used the relation between @xmath20 and @xmath12 , in the second the growth rate of @xmath23 on super - hubble scales , and in the third line we have made use of the vacuum amplitude of @xmath23 at the hubble crossing time @xmath35 and the hubble radius crossing condition _
h(k ) ~ k^-1 .
thus , we have shown that vacuum perturbations which exit the hubble radius during the matter - dominated phase of contraction acquire a scale - invariant spectrum .
after the transition to a radiation - dominated phase of contraction ( if such a phase exists ) all modes are boosted by the same factor , and hence the scale - invariance of the spectrum persists at least until the bounce phase begins .
so far we have shown that before the bounce the spectrum of fluctuations of the curvature variable @xmath12 is scale - invariant .
however , we need to determine the curvature fluctuations in the expanding phase , i.e. after the bounce .
since the curvature fluctuations ( in the case of adiabatic perturbations ) are constant in time on super - hubble scales in an expanding universe , it is sufficient to compute the spectrum immediately after the bounce at the onset of the period of standard cosmology evolution .
the transfer of curvature fluctuations through a non - singular bouncing phase is a non - trivial topic . in models with only adiabatic fluctuations ,
the bottom line of many detailed studies is that on length scales which are longer than the time duration of the bounce the spectral index of the power spectrum of @xmath12 is not changed @xcite ( however , as stressed in @xcite , this is not a completely general result ) .
on the other hand , in the presence of entropy fluctuations changes are to be expected @xcite .
the reason why the analysis is non - trivial is related to the fact that new physics is required in order to obtain a non - singular bounce , and when evolving the fluctuations through the bounce phase the modes associated with the new physics must be taken into account . the are two ways of following fluctuations through the bouncing phase .
the first is my explicit numerical integration . in any given realization of the matter
bounce we know what the equations for the fluctuations are , and we can integrate them . however , to obtain a good understanding of the results , it is important to have an analytical method .
this method uses matching conditions at the transition surface from one phase to the next .
matching conditions were introduced by israel @xcite in the context of matching two solutions of einstein s equations across a time - like surface .
they were then generalized to the problem of matching across a space - like boundary hypersurface in cosmology by hwang and vishniac @xcite and by deruelle and mukhanov @xcite .
the matching conditions state that at the boundary between one phase and another both the induced metric on the boundary hypersurface and the extrinsic curvature must be continuous
. applied to the case of cosmological perturbations , one must first make sure that the background satisfies the matching conditions , a point emphasized in @xcite ( see also the caveats mentioned in @xcite ) .
once the matching of the background is successfully achieved , the matching conditions imply that the metric fluctuation variable @xmath45 and the extrinsic curvature fluctuation must be continuous across the bounce . if the matching surface is a constant energy density hypersurface , then the continuity of the extrinsic curvature implies the continuity of @xmath12 @xcite . to model a non - singular bounce , we can merge three phases , first the contracting phase obeying the equations of general relativity , next the bounce phase during which we can use the parametrization h ( ) = , where @xmath46 is some constant .
we have normalized the time axis such that the bounce point corresponds to @xmath47 .
we must now introduce two matching surfaces @xcite , first the transition surface between the standard contraction phase and the bounce phase , and second between the bounce phase and the standard expanding phase . at
the first transition surface the background is contracting on both sides , and hence the background matching conditions can easily be satisfied . similarly , at the second transition surface the background is expanding on both sides , and hence the background matching conditions are again satisfied . if both transition surfaces are constant energy density surfaces ( which will automatically be the case if the fluctuations are adiabatic ) , then @xmath12 is continuous across the surface . on length scales larger than the duration of the bounce
phase the non - trivial evolution during the bounce phase is independent of @xmath24 ( since the @xmath48 term in the equation of motion is negligible compared to the term which generalizes the @xmath49 term in ( [ perteom2 ] ) ) and hence , although it can change the amplitude of the spectrum , it will not change the spectral index @xmath33 .
we hence conclude that the scale - invariant pre - bounce spectrum survives the bounce . in figure 2
we present an example of the numerically obtained evolution of cosmological perturbations across a non - singular bounce , in the context of the model of @xcite which will be discussed below .
( vertical axis ) as functions of cosmic time ( horizontal axis ) .
these modes are distinguished by the comoving wave number @xmath24 , which are @xmath50 ( in red ) , @xmath51 ( in orange ) , @xmath52 ( in green ) , and @xmath53 ( in violet ) , respectively .
the inner insert shows the detailed evolution of @xmath23 during the bounce phase .
the initial conditions of the background field and background parameters are the same as those chosen in fig .
3 . all numerical values are in planck units @xmath54 .
the initial conditions for perturbation modes were chosen to be vacuum initial conditions .
the amplitude of @xmath23 grows during the bounce phase , but by the same factor for all long wavelength modes . ]
there is also the possibility of realizing the matter bounce scenario in the context of a singular bounce .
one example is obtained in the context of the string theory backgrounds with gravitomagnetic fluxes studied in @xcite . in this case
, there are cosmological solutions which correspond to a contracting radiation phase glued to an expanding radiation phase via an s - brane configuration , a configuration which has vanishing energy density and negative pressure and thus allows for the violation of the null energy condition which in turn allows for a cosmological bounce .
this model will be discussed further below . in this context
, great care must be taken when matching the cosmological fluctuations across the bounce since the extrinsic curvature of the background changes sign across the s - brane .
this issue has been discussed in @xcite based on the general analysis of @xcite .
other examples of singular bouncing cosmologies are the pre - big - bang scenario of @xcite and the ekpyrotic scenario @xcite .
as stressed in @xcite , in these cases the application of the matching conditions is only well defined after careful specification of the matching surface .
this resolved the disagreement between various results obtained in the literature for the final spectrum of curvature fluctuations in the ekpyrotic scenario , some of which obtained a scale - invariant spectrum @xcite , and others which did not @xcite .
calculations done in the higher - dimensional framework of the ekpyrotic scenario show that modes which from the four space - time dimensional point of view are entropy modes can mediate a successful transition of the spectrum across the ekpyrotic bounce @xcite ( see also @xcite ) . after these general remarks on matching conditions for cosmological fluctuations we will return the the main development , and we turn to a discussion of implementations of the matter bounce scenario .
as is well known , in order to obtain a bouncing cosmology it is necessary to either abandon general relativity as the theory which describes space and time , or else we must introduce matter which violates the null energy condition ( nec ) .
both options have been used to obtain realizations of a non - singular bouncing cosmology .
there is good reason to consider gravitational actions which differ from that of general relativity at high energy scales : since general relativity is not a renormalizable quantum theory , we know that in any approach to quantum gravity there will be terms in the action which contain higher derivatives . in this context
the singularity theorems of general relativity are no longer applicable and it is possible that nonsingular cosmological solutions emerge . on the other hand , since higher derivative terms in the action lead to a larger number of modes , the danger arises that the singularities will in fact be worse .
there are existence proofs of higher derivative actions which have non - singular cosmological solutions .
one example is the non - singular universe " construction of @xcite in which the action is constructed in order that space - time approached de sitter space at a certain limiting curvature .
this allows for bouncing cosmological solutions .
another example is the ( non - local ) action of @xcite which was constructed to be ghost - free about minkowski space - time .
once again , this action leads to bouncing cosmological solutions .
below , i will discuss a recent example of a bouncing cosmology arising from a modified gravitational action , namely the hoava - lifshitz bounce @xcite .
more numerous are examples of bouncing cosmologies which arise from modifying the matter sector .
below , i will discuss a few recent examples .
hoava - lifshitz ( hl ) gravity @xcite is a proposal for a power - counting renormalizable quantum theory of gravity in four space - time dimensions .
the approach is conservative in that it uses only the usual metric variables , but it is radical in the sense that it discards space - time diffeomorphism and local lorentz invariance as underlying symmetries of the theory .
hl gravity postulates the existence of a preferred time direction .
it maintains spatial diffeomorphism invariance , but space - dependent time reparametrizations are no longer a symmetry of the theory .
instead , hl gravity postulates a scaling symmetry which treats space and time anisotropically .
if @xmath55 , then there is a scaling symmetry if @xmath56 .
the gravitational action in hl theory is obtained by adding to the einstein action all terms which are consistent with the residual symmetries and which are power - counting renormalizable with respect to the above - mentioned scaling symmetry .
this leads to the possibility of adding higher space - derivative terms to the gravitational action - up to four extra space derivative terms are allowed .
these higher space - derivative terms can be moved to the right - hand ( matter ) side of the einstein equations .
as was realized in @xcite , in the presence of spatial curvature these terms act as matter with negative effective energy density , thus violating the nec and allowing for a bouncing cosmology .
specifically , if we insert the ansatz for a spatially homogeneous and isotropic metric into the hoava - lifshitz action , we obtain the following generalization of the friedmann - robertson - walker ( frw ) equations : h^2 & = & - - - + , [ heq ] + & = & - - + + , [ acceq ] in the above , @xmath57 is the usual hubble expansion rate , @xmath0 is physical time , an overdot indicates the time derivative with respect to @xmath0 , @xmath58 is the spatial curvature constant , @xmath59 is the cosmological constant , @xmath60 and @xmath61 are energy and pressure densities of matter , respectively , @xmath62 is the coefficient of the standard kinetic term in the gravitational action , and @xmath63 and @xmath10 are coefficients of the higher spatial derivative terms in the potential part of the gravitational action .
considering the first frw equation , we recognize all the terms except the third term on the right hand side .
this term is due to the higher space derivative terms in the gravitational action .
we see that it scales in time like radiation . if the kinetic term in the gravitational action is close to the form it has in general relativity , then @xmath64 and thus the new term acts like ghost radiation .
note that this term is absent if the spatial sections are flat , but that the coefficient does not depend on whether the spatial curvature is positive or negative .
the bouncing cosmology now arises as follows .
we begin with a matter - dominated phase of contraction . at low curvatures
the ghost radiation term is negligible .
however , as @xmath65 decreases , the importance of ghost radiation relative to matter increases . eventually there will be a time when the magnitude of the ghost radiation term catches up with that of regular matter ( for small values of @xmath65 the first two terms on the right hand side of the frw equations are negligible ) . at this time we have @xmath66 . from the second frw equation it follows that @xmath67 at this time .
thus , a cosmological bounce occurs . introducing quintom matter @xcite yields a way of obtaining a non - singular bouncing cosmology , as discussed in @xcite .
quintom matter is a set of two matter fields , one of them regular matter ( obeying the weak energy condition ) , the second a field with opposite sign kinetic term , a field which violates the energy conditions .
we can @xcite model both matter components with scalar fields .
let us denote the mass of the regular one ( @xmath68 ) by @xmath69 , and by @xmath70 that of the field @xmath71 with wrong sign kinetic term .
we assume that early in the contracting phase both fields are oscillating , but that the amplitude @xmath72 of @xmath68 greatly exceeds the corresponding amplitude @xmath73 of @xmath71 such that the energy density is dominated by @xmath68 .
both fields will initially be oscillating during the contracting phase , and both amplitudes grow at the same rate . at some point , @xmath72 will become so large that the oscillations of @xmath68 freeze out . then , @xmath72 will grow only slowly , whereas @xmath73 will continue to increase .
thus , the ( negative ) energy density in @xmath71 will grow in absolute values relative to that of @xmath68 .
the total energy density will decrease towards @xmath74 . at that point , @xmath66 by the friedmann equations .
[ heq2 ] h^2 = , where @xmath75 is the potential of the two fields . from the second frw equation [ dotheq ] h = - 4 g ( ^2 - ^2 )
it follows that since @xmath71 is oscillating while @xmath68 is only slowly rolling that @xmath76 when @xmath66 .
hence , a non - singular bounce occurs .
the higgs sector of the lee - wick model @xcite provides a concrete realization of the quintom bounce model , as studied in @xcite .
quintom models like all other models with negative sign kinetic terms suffer from an instability problem @xcite in the matter sector and are hence problematic .
in addition , they are unstable against the addition of radiation ( see e.g. @xcite ) and anisotropic stress . an improved way of obtaining a non - singular bouncing cosmology using modified matter is by using a ghost condensate field @xcite ( see also @xcite for related work ) . the ghost condensation mechanism is the analog of the higgs mechanism in the kinetic sector of the theory . in the higgs mechanism we take a field @xmath68 whose mass when evaluated at @xmath77 is tachyonic , add higher powers of @xmath78 to the potential term in the lagrangian such that there is a stable fixed point @xmath79 , and thus when expanded about @xmath80 the mass term has the safe " non - tachyonic sign . in the ghost condensate construction
we take a field @xmath68 whose kinetic term x - g^ _ _ appears with the wrong sign in the lagrangian .
then , we add higher powers of @xmath81 to the kinetic lagrangian such that there is a stable fixed point @xmath82 and such that when expanded about @xmath82 the fluctuations have the regular sign of the kinetic term : = m^4 ( x - c^2 ) ^2 - v ( ) , where @xmath83 is a usual potential function , @xmath70 is a characteristic mass scale and the dimensions of @xmath68 are chosen such that @xmath81 is dimensionless . in the context of cosmology ,
the ghost condensate is = c t and breaks local lorentz invariance .
now let us expand the homogeneous component of @xmath68 about the ghost condensate : ( t ) = c t + ( t ) . if @xmath84 then the gravitational energy density is negative , and a non - singular bounce is possible .
thus , in @xcite we constructed a model in which the ghost condensate field starts at negative values and the potential @xmath83 is negligible . as @xmath68 approaches @xmath77 it encounters a positive potential which slows it down , leading to @xmath84 and hence to negative gravitational energy density .
thus , a non - singular bounce can occur .
we take the potential to be of the form v ( ) ~ ^- for @xmath85 , where @xmath70 is the mass scale above which the higher derivative kinetic terms are important . for sufficiently large values of @xmath46 , namely 6 , the energy density in
the ghost condensate increases faster than that of radiation and anisotropic stress at the universe contracts .
hence , this bouncing cosmology is stable against the addition of radiation and anisotropic stress .
however , the model is still unstable to the presence of anisotropic stress during the long periods of matter and radiation domination which precede the bounce phase .
we now turn to a model which also solves this problem .
one of the key advantages of the ekpyrotic bouncing model @xcite is that anisotropies decay relative to the dominant component of the energy density during the period of contraction @xcite .
the ekpyrotic scenario is obtained by introducing a scalar field with negative exponential potential . while the scalar field is rolling down the potential towards @xmath77 , the equation of state parameter @xmath86 satisfied w 1 , since there is a partial cancellation between the positive kinetic energy density and the negative potential . from the continuity equation
it hence follows that the energy density @xmath87 increases faster than that in anisotropies .
the spectrum of adiabatic fluctuations in the ekpyrotic scenario is blue @xcite , and hence additional inputs ( e.g. entropy fluctuations ) are required to obtain a scale - invariant spectrum of curvature fluctuations .
the ekpyrotic bounce model of @xcite combines the advantages of the
matter bounce " scenario in terms of obtaining a scale - invariant spectrum of adiabatic fluctuations with the advantages of the ekpyrotic model in solving the anisotropy problem of a contracting universe .
the model is obtained by introducing a scalar field @xmath68 which combines the ghost condensate or galileon mechanisms for obtaining a non - singular bounce discussed in the previous subsection with a negative exponential potential which renders @xmath87 to be dominant over anisotropies at later stages during the contracting phase .
the model of @xcite is based on the most general form of single scalar field lagrangian giving rise to second - order field equations in four - dimensional spacetime @xcite @xmath88 where @xmath89 and @xmath90 are functions of a dimensionless scalar field @xmath68 and its canonical kinetic term @xmath81 , and where we have not written down a number of higher order operators which are not important at lower energy scales ( we shall construct our bounce to be consistent with the condition that these terms remain negligible ) .
following the idea behind the ghost condensate bounce of @xcite we assume that the term @xmath89 is given by @xmath91x + \beta x^2 - v(\varphi)~,\end{aligned}\ ] ] where we introduce a positive - definite parameter @xmath92 so that the kinetic term is bounded from below at high energy scales . for
@xmath93 a ghost condensate ground state with @xmath94 can arise .
this is the first key principle of our model .
note that the first term of @xmath89 involves @xmath95 since in the present paper we adopt the convention that the scalar field @xmath68 is dimensionless .
the second key principle of our model is to introduce a non - trivial potential @xmath75 for @xmath68 which is chosen such that ekpyrotic contraction is possible . in the specific model which we will discuss in the following
, the scalar field evolves monotonically from a negative large values to a positive large value .
the function @xmath96 is chosen such that a phase of ghost condensation only occurs during a short time when @xmath97 approaches @xmath98 .
this requires the dimensionless function @xmath99 to be smaller than unity when @xmath100 but larger than unity when @xmath97 approaches the origin .
we choose @xmath90 to be a simple function of only @xmath81 : @xmath101 where @xmath102 is a positive number .
this term is introduced to stabilize gradients .
the potential is chosen such that during the approach to @xmath77 from negative values we have ekpyrotic contraction : @xmath103 where @xmath104 is a constant .
based on the previous discussion , the function @xmath105 is chosen in the following form @xmath106 where @xmath107 is a positive constant greater than @xmath108 .
the background equations of motion which follow from the lagrangian ( [ lagrangian ] ) with the choices of the various free functions discussed above can be solved analytically using approximate methods .
they can also be solved numerically .
the results of the numerical integration are shown in figure 3 .
( vertical axis ) as a function of cosmic time ( horizontal axis ) .
the main plot shows that a non - singular bounce occurs , and that the time scale of the bounce is short .
the inner insert shows a blowup of the hubble parameter @xmath57 during the bounce phase .
all numerical values are in planck units @xmath54 .
the parameters were chosen to be @xmath109 . ] as is evident , a successful non - singular bounce occurs .
we have also studied @xcite the evolution of cosmological fluctuations through the bounce phase and verified that on large scales the spectral index does not change ( see figure 2 ) .
all the models with modified matter we have considered so far are toy models in that they do not arise from a theory which is well behaved in the ultraviolet .
we conclude this section by mentioning a model which originates from a theory which is well behaved in the ultraviolet , namely superstring theory . it has long been realized that string theory has the potential of solving the singularity problem of standard big bang cosmology .
as is well known @xcite , there is a maximal temperature of a gas of weakly coupled strings in thermal equilibrium , the hagedorn temperature @xmath110 . in the context of a toroidal compactification of heterotic string theory ,
it was realized @xcite that the temperature remains smaller than @xmath110 throughout . however , there are potential instabilities of string theory near the hagedorn temperature .
recently @xcite , ingredients to resolve both the hagedorn instability as well as the initial curvature singularity of string cosmology were proposed .
these ingredients are realized in a class of @xmath111 type ii superstring models where the presence of non - trivial gravito - magnetic " fluxes modifies the thermal vacuum by adding non - trivial momentum and winding charges , lifting the hagedorn instabilities . if @xmath112 denotes the radius of the euclidean time circle , the string partition function @xmath113 has a thermal duality symmetry about a critical value @xmath114 , which is of the order of the string scale : [ dual ] z(r_0 ) = z(r_c^2 / r_0 ) .
the inverse temperature @xmath115 measured in the string frame is then proportional to @xmath112 or its dual value in the asymptotic regimes : [ beta(r ) ] ~2r_0 r_0r_c , ~2 r_c^2r_0 r_0r_c .
thus the temperature acquires a maximal critical value . in the adiabatic approximation ,
cosmological dynamics very different from the usual particle physics induced dynamics can be obtained .
one possibility is a solution in which @xmath112 runs from @xmath74 to @xmath2 . at sub - critical values of @xmath112
, the universe can be taken to be in a contracting phase . at the critical value of @xmath112 ,
states characterized by non - trivial winding and momentum charges along the euclidean time circle become massless and give rise to a euclidean gauge theory . at the level of the four - dimensional low energy effective field
theory of dilaton - gravity , these massless states lead to extra terms in the effective action which correspond @xcite to an s - brane configuration ( see e.g. @xcite for early references on s - branes ) , and the effective action is that of dilaton - gravity coupled to a thermal gas of strings plus the s - brane configuration . an s - brane acts like a topological defect in space - like direction .
its energy density vanishes in the same way that the transverse pressure of a regular defect vanishes .
the pressure is negative .
hence , the s - brane provides the violation of the null energy condition ( nec ) which leads to a smooth transition from a contracting phase to an expanding phase ( when cosmological evolution is viewed in the einstein frame ) .
thus , for @xmath116 the universe expands .
thus , the setup of @xcite provides a realization of a bouncing cosmology .
the metric in this setup is continuous , but the derivative of the metric diverges . at physical temperatures
much lower than the critical value , the partition function of the string gas has the equation of state of radiation as long as supersymmetry is unbroken .
however , at low temperatures we expect supersymmetry to be broken , thus generating masses and leading to the emergence of a phase of matter domination at low temperature . in @xcite ,
the evolution of cosmological fluctuations in this bouncing cosmology has been studied in detail .
the background cosmology starts with a matter - dominated phase of contraction . at a certain temperature @xmath117 supersymmetry
is restored , and a transition to a phase of radiation domination will occur .
once the temperature rises to the dual value , the s - brane will be reached .
the s - brane mediates the transition to an expanding radiation phase .
the temperature will decrease , and below @xmath117 supersymmetry will break , allowing for the generation of a component of pressureless matter which dominates the late - time dynamics .
this background thus provides the requirements to implement the matter bounce scenario : we begin with quantum vacuum perturbations on sub - hubble scales in the contracting phase .
scales which exit the hubble radius before the onset of the radiation phase acquire a scale - invariant spectrum . in the absence of entropy fluctuations the perturbations are matched across the s - brane with no change in the spectral index .
hence , at late times in the expanding phase a scale - invariant spectrum of curvature fluctuations results .
the matter bounce " is an alternative to cosmological inflation as a mechanism for generating an almost scale - invariant spectrum of cosmological fluctuations . as in the case of inflationary cosmology
, the spectrum has a slight red tilt since smaller scales exit the hubble radius when the radiation component of matter is more important and the vacuum slope of the spectrum ( which corresponds to @xmath34 ) is rearing its head .
in fact , there is a transition to an @xmath34 spectrum on small scales which exit the hubble radius in the radiation phase of contraction .
the current observational constraints on a spectrum with such a transition from @xmath32 on large scales to @xmath34 on small scales were studied in @xcite , with the result that the power spectrum of quasar absorption lines already mildly constrains a model with no entropy production during the bounce .
if there is entropy production during the bounce , then the radiation phase of contraction is shorter than the radiation phase of expansion and the observational constraints recede .
the spectrum of gravitational radiation produced in a matter bounce is also scale - invariant . at first glance
, there will be less enhancement of the amplitude during the bounce phase than for scalar metric fluctuations , and hence the predicted tensor to scalar ratio @xmath118 will be consistent with current constraints .
however , a detailed study of this issue is needed .
since the curvature fluctuation @xmath20 increases on super - hubble scales during the contracting period , terms in the general expression for the bispectrum ( three point function of the curvature fluctuation variable ) derived in @xcite will be important , leading to a shape of the bispectrum which is different from the one obtained in simple inflationary models @xcite .
since the analog of the inflationary slow - roll parameter is of order unity in the case of the matter bounce , the predicted amplitude of the bispectrum will be of order unity @xcite .
thus , amplitude and shape of the bispectrum provide a way to differentiate between the matter bounce and simple inflationary models . possibly the most serious problem which the matter bounce scenario faces is the anisotropy problem ( see e.g. @xcite ) .
the only solution of this problem which we know of now is to postulate a phase of ekpyrotic contraction following the initial matter contraction phase . in the model of @xcite
this solves the anisotropy problem , whereas in the original new ekpyrotic scenario the anisotropy problem may re - appear during the bounce phase @xcite .
i have focused on single bounce models .
cyclic bouncing background cosmologies face several challenges .
in addition to the heat death " problem ( entropy increasing from cycle to cycle ) , there is the problem that the dynamics of cosmological fluctuations breaks the cyclicity of the model since fluctuations grow on super - hubble scales during the periods of contraction and lead to a jump in the index @xmath33 of the power spectrum by @xmath39 from cycle to cycle @xcite .
thus , the model loses predictability .
i wish to thank my collaborators on whose work i have drawn .
in particular , i thank dr .
yifu cai for collaboration on many of these topics and for allowing me to make use of the figures .
special thanks to professor xinmin zhang for collaboration over a period of many years and for wonderfully hosting me repeatedly at ihep in beijing .
this work was supported in part by an nserc discovery grant and by funds from the canada research chair program .
m. novello and s. e. p. bergliaffa , `` bouncing cosmologies , '' phys .
rept . * 463 * , 127 ( 2008 ) [ arxiv:0802.1634 [ astro - ph ] ] .
d. wands , `` duality invariance of cosmological perturbation spectra , '' phys .
d * 60 * , 023507 ( 1999 ) [ arxiv : gr - qc/9809062 ] .
f. finelli and r. brandenberger , `` on the generation of a scale - invariant spectrum of adiabatic fluctuations in cosmological models with a contracting phase , '' phys .
d * 65 * , 103522 ( 2002 ) [ arxiv : hep - th/0112249 ] .
r. h. brandenberger , `` alternatives to cosmological inflation , '' arxiv:0902.4731 [ hep - th ] ; + r. h. brandenberger , `` cosmology of the very early universe , '' aip conf . proc .
* 1268 * , 3 - 70 ( 2010 ) .
[ arxiv:1003.1745 [ hep - th ] ] ; + r. h. brandenberger , `` introduction to early universe cosmology , '' pos icfi * 2010 * , 001 ( 2010 ) [ arxiv:1103.2271 [ astro-ph.co ] ] .
a. guth , `` the inflationary universe : a possible solution to the horizon and flatness problems , '' phys .
d * 23 * , 347 ( 1981 ) .
r. brout , f. englert and e. gunzig , `` the creation of the universe as a quantum phenomenon , '' annals phys .
* 115 * , 78 ( 1978 ) .
k. sato , `` first order phase transition of a vacuum and expansion of the universe , '' mon . not .
soc . * 195 * , 467 ( 1981 ) .
a. a. starobinsky , `` a new type of isotropic cosmological models without singularity , '' phys .
b * 91 * , 99 ( 1980 ) . v. mukhanov and g. chibisov , `` quantum fluctuation and nonsingular universe .
( in russian ) , '' jetp lett . *
33 * , 532 ( 1981 ) [ pisma zh .
fiz . * 33 * , 549 ( 1981 ) ] . w. press , `` spontaneous production of the zeldovich spectrum of cosmological fluctuations '' , phys .
* 21 * , 702 ( 1980 ) .
a. a. starobinsky , `` spectrum of relict gravitational radiation and the early state of the universe , '' jetp lett .
* 30 * , 682 ( 1979 ) [ pisma zh .
fiz . * 30 * , 719 ( 1979 ) ] .
c. l. bennett _ et al .
_ , `` first year wilkinson microwave anisotropy probe ( wmap ) observations : preliminary maps and basic results , '' astrophys
. j. suppl . *
148 * , 1 ( 2003 ) [ arxiv : astro - ph/0302207 ] .
p. j. e. peebles , j. t. yu , `` primeval adiabatic perturbation in an expanding universe , '' astrophys .
j. * 162 * , 815 - 836 ( 1970 ) .
j. khoury , b. a. ovrut , p. j. steinhardt and n. turok , `` the ekpyrotic universe : colliding branes and the origin of the hot big bang , '' phys .
d * 64 * , 123522 ( 2001 ) [ arxiv : hep - th/0103239 ] .
r. h. brandenberger and c. vafa , `` superstrings in the early universe , '' nucl .
b * 316 * , 391 ( 1989 ) ; + a. nayeri , r. h. brandenberger and c. vafa , `` producing a scale - invariant spectrum of perturbations in a hagedorn phase of string cosmology , '' phys .
lett . * 97 * , 021302 ( 2006 ) [ arxiv : hep - th/0511140 ] ; + r. h. brandenberger , `` string gas cosmology , '' arxiv:0808.0746 [ hep - th ] . j. w. moffat , `` superluminary universe : a possible solution to the initial value problem in cosmology , '' int . j. mod .
d * 2 * , 351 ( 1993 ) [ gr - qc/9211020 ] ; + a. albrecht and j. magueijo , `` a time varying speed of light as a solution to cosmological puzzles , '' phys .
d * 59 * , 043516 ( 1999 ) [ astro - ph/9811018 ] .
v. a. rubakov , `` harrison - zeldovich spectrum from conformal invariance , '' jcap * 0909 * , 030 ( 2009 ) [ arxiv:0906.3693 [ hep - th ] ] ; + k. hinterbichler and j. khoury , `` the pseudo - conformal universe : scale invariance from spontaneous breaking of conformal symmetry , '' jcap * 1204 * , 023 ( 2012 ) [ arxiv:1106.1428 [ hep - th ] ] .
s. w. hawking and g. f. r. ellis , `` the large scale structure of space - time , '' _ cambridge university press , cambridge , 1973 _
s. watson and r. h. brandenberger , `` isotropization in brane gas cosmology , '' phys .
d * 67 * , 043510 ( 2003 ) [ hep - th/0207168 ] .
y. cai , d. easson and r. brandenberger , towards a nonsingular bouncing cosmology " , arxiv:1206.2382 [ hep - th ] .
v. f. mukhanov , h. a. feldman and r. h. brandenberger , `` theory of cosmological perturbations .
classical perturbations .
quantum theory of perturbations .
extensions , '' phys .
rept . * 215 * , 203 ( 1992 ) .
r. h. brandenberger , `` lectures on the theory of cosmological perturbations , '' lect .
notes phys .
* 646 * , 127 ( 2004 ) [ arxiv : hep - th/0306071 ]
. t. j. battefeld and r. brandenberger , `` vector perturbations in a contracting universe , '' phys .
d * 70 * , 121302 ( 2004 ) [ hep - th/0406180 ] .
v. f. mukhanov , `` quantum theory of gauge invariant cosmological perturbations , '' sov .
jetp * 67 * , 1297 ( 1988 ) [ zh
. eksp .
fiz . * 94n7 * , 1 ( 1988 ) ] . m. sasaki , `` large scale quantum fluctuations in the inflationary universe , '' prog
. theor .
phys . * 76 * , 1036 ( 1986 ) .
v. n. lukash , `` production of phonons in an isotropic universe , '' sov .
jetp * 52 * , 807 ( 1980 ) [ zh . eksp . teor .
fiz . * 79 * , ( 1980 ) ] . j. m. bardeen , `` gauge invariant cosmological perturbations , '' phys .
d * 22 * , 1882 ( 1980 ) .
j. m. bardeen , p. j. steinhardt and m. s. turner , `` spontaneous creation of almost scale - free density perturbations in an inflationary universe , '' phys .
d * 28 * , 679 ( 1983 ) .
r. h. brandenberger and r. kahn , `` cosmological perturbations in inflationary universe models , '' phys .
d * 29 * , 2172 ( 1984 ) .
d. h. lyth , `` large scale energy density perturbations and inflation , '' phys .
d * 31 * , 1792 ( 1985 ) .
r. h. brandenberger , `` inflationary cosmology : progress and problems , '' arxiv : hep - ph/9910410 .
r. h. brandenberger and j. martin , `` the robustness of inflation to changes in superplanck scale physics , '' mod .
lett . a * 16 * , 999 ( 2001 ) , [ arxiv : astro - ph/0005432 ] ; + j. martin and r. h. brandenberger , `` the transplanckian problem of inflationary cosmology , '' phys .
d * 63 * , 123501 ( 2001 ) , [ arxiv : hep - th/0005209 ] .
y. f. cai , t. qiu , r. brandenberger , y. s. piao and x. zhang , `` on perturbations of quintom bounce , '' jcap * 0803 * , 013 ( 2008 ) [ arxiv:0711.2187 [ hep - th ] ] ; + y. f. cai and x. zhang , `` evolution of metric perturbations in quintom bounce model , '' jcap * 0906 * , 003 ( 2009 ) [ arxiv:0808.2551 [ astro - ph ] ] .
y. f. cai , t. qiu , r. brandenberger and x. zhang , `` a nonsingular cosmology with a scale - invariant spectrum of cosmological perturbations from lee - wick theory , '' phys .
d * 80 * , 023511 ( 2009 ) [ arxiv:0810.4677 [ hep - th ] ] .
r. brandenberger , h. firouzjahi and o. saremi , `` cosmological perturbations on a bouncing brane , '' jcap * 0711 * , 028 ( 2007 ) [ arxiv:0707.4181 [ hep - th ] ] .
x. gao , y. wang , w. xue and r. brandenberger , `` fluctuations in a hoava - lifshitz bouncing cosmology , '' jcap * 1002 * , 020 ( 2010 ) [ arxiv:0911.3196 [ hep - th ] ] ; + x. gao , y. wang , r. brandenberger and a. riotto , `` cosmological perturbations in hoava - lifshitz gravity , '' phys . rev .
d * 81 * , 083508 ( 2010 ) [ arxiv:0905.3821 [ hep - th ] ] .
s. alexander , t. biswas and r. h. brandenberger , `` on the transfer of adiabatic fluctuations through a nonsingular cosmological bounce , '' arxiv:0707.4679 [ hep - th ] .
j. martin and p. peter , `` on the causality argument in bouncing cosmologies , '' phys .
lett . * 92 * , 061301 ( 2004 ) [ astro - ph/0312488 ] .
r. durrer and f. vernizzi , `` adiabatic perturbations in pre big bang models : matching conditions and scale invariance , '' phys .
d * 66 * , 083503 ( 2002 ) [ arxiv : hep - ph/0203275 ] .
israel , `` singular hypersurfaces and thin shells in general relativity , '' nuovo cim .
b * 44s10 * , 1 ( 1966 ) [ erratum - ibid .
b * 48 * , 463 ( 1967 ) ] [ nuovo cim .
b * 44 * , 1 ( 1966 ) ] .
j. c. hwang and e. t. vishniac , `` gauge - invariant joining conditions for cosmological perturbations , '' astrophys .
j. * 382 * , 363 ( 1991 ) .
n. deruelle and v. f. mukhanov , `` on matching conditions for cosmological perturbations , '' phys .
d * 52 * , 5549 ( 1995 ) [ arxiv : gr - qc/9503050 ] .
j. martin , p. peter , n. pinto - neto and d. j. schwarz , `` comment ondensity perturbations in the ekpyrotic scenario , '' phys .
d * 67 * , 028301 ( 2003 ) [ hep - th/0204222 ] .
i. florakis , c. kounnas , h. partouche and n. toumbas , `` non - singular string cosmology in a 2d hybrid model , '' nucl .
b * 844 * , 89 ( 2011 ) [ arxiv:1008.5129 [ hep - th ] ] ; + c. kounnas , h. partouche and n. toumbas , `` thermal duality and non - singular cosmology in d - dimensional superstrings , '' nucl .
b * 855 * , 280 ( 2012 ) [ arxiv:1106.0946 [ hep - th ] ] ; + c. kounnas , h. partouche and n. toumbas , `` s - brane to thermal non - singular string cosmology , '' arxiv:1111.5816 [ hep - th ] .
r. brandenberger , c. kounnas , h. partouche , s. patil and n. toumbas , fluctuations in non - singular bouncing cosmologies from type ii superstrings " , to be submitted .
m. gasperini and g. veneziano , `` pre - big bang in string cosmology , '' astropart .
* 1 * , 317 ( 1993 ) [ arxiv : hep - th/9211021 ]
. j. khoury , b. a. ovrut , p. j. steinhardt and n. turok , `` density perturbations in the ekpyrotic scenario , '' phys .
d * 66 * , 046005 ( 2002 ) [ hep - th/0109050 ] .
a. j. tolley , n. turok and p. j. steinhardt , `` cosmological perturbations in a big crunch / big bang space - time , '' phys .
d * 69 * , 106005 ( 2004 ) [ hep - th/0306109 ] .
p. peter and n. pinto - neto , `` primordial perturbations in a non singular bouncing universe model , '' phys .
d * 66 * , 063509 ( 2002 ) [ hep - th/0203013 ] .
d. h. lyth , `` the failure of cosmological perturbation theory in the new ekpyrotic scenario , '' phys
. lett .
b * 526 * , 173 ( 2002 ) [ hep - ph/0110007 ] ; + d. h. lyth , `` the primordial curvature perturbation in the ekpyrotic universe , '' phys . lett .
b * 524 * , 1 ( 2002 ) [ hep - ph/0106153 ] .
r. brandenberger and f. finelli , `` on the spectrum of fluctuations in an effective field theory of the ekpyrotic universe , '' jhep * 0111 * , 056 ( 2001 ) [ hep - th/0109004 ] . j. -c .
hwang , `` cosmological structure problem in the ekpyrotic scenario , '' phys .
d * 65 * , 063514 ( 2002 ) [ astro - ph/0109045 ] .
j. martin , p. peter , n. pinto neto and d. j. schwarz , `` passing through the bounce in the ekpyrotic models , '' phys .
d * 65 * , 123513 ( 2002 ) [ hep - th/0112128 ] .
t. j. battefeld , s. p. patil and r. brandenberger , `` non - singular perturbations in a bouncing brane model , '' phys .
d * 70 * , 066006 ( 2004 ) [ hep - th/0401010 ] .
a. notari and a. riotto , `` isocurvature perturbations in the ekpyrotic universe , '' nucl .
b * 644 * , 371 ( 2002 ) [ hep - th/0205019 ] .
f. finelli , `` assisted contraction , '' phys .
b * 545 * , 1 ( 2002 ) [ hep - th/0206112 ] ; + f. finelli , `` study of a class of four - dimensional nonsingular cosmological bounces , '' jcap * 0310 * , 011 ( 2003 ) [ hep - th/0307068 ] .
p. creminelli and l. senatore , `` a smooth bouncing cosmology with scale invariant spectrum , '' jcap * 0711 * , 010 ( 2007 ) [ arxiv : hep - th/0702165 ] ; + e. i. buchbinder , j. khoury and b. a. ovrut , `` new ekpyrotic cosmology , '' phys .
d * 76 * , 123503 ( 2007 ) [ arxiv : hep - th/0702154 ] ; + j. -l .
lehners , p. mcfadden , n. turok and p. j. steinhardt , `` generating ekpyrotic curvature perturbations before the big bang , '' phys .
d * 76 * , 103501 ( 2007 ) [ hep - th/0702153 [ hep - th ] ] .
r. h. brandenberger , v. f. mukhanov and a. sornborger , `` a cosmological theory without singularities , '' phys .
d * 48 * , 1629 ( 1993 ) [ arxiv : gr - qc/9303001 ] .
t. biswas , a. mazumdar and w. siegel , `` bouncing universes in string - inspired gravity , '' jcap * 0603 * , 009 ( 2006 ) [ arxiv : hep - th/0508194 ] .
r. brandenberger , `` matter bounce in horava - lifshitz cosmology , '' phys .
d * 80 * , 043516 ( 2009 ) [ arxiv:0904.2835 [ hep - th ] ] .
p. horava , `` quantum gravity at a lifshitz point , '' phys .
d * 79 * , 084008 ( 2009 ) [ arxiv:0901.3775 [ hep - th ] ] .
b. feng , x. l. wang and x. m. zhang , `` dark energy constraints from the cosmic age and supernova , '' phys .
b * 607 * , 35 ( 2005 ) [ arxiv : astro - ph/0404224 ] ; + b. feng , m. li , y. s. piao and x. zhang , `` oscillating quintom and the recurrent universe , '' phys .
b * 634 * , 101 ( 2006 ) [ arxiv : astro - ph/0407432 ] ; + y. f. cai , t. qiu , y. s. piao , m. li and x. zhang , `` bouncing universe with quintom matter , '' jhep * 0710 * , 071 ( 2007 ) [ arxiv:0704.1090 [ gr - qc ] ] ; + y. f. cai , t. t. qiu , j. q. xia and x. zhang , `` a model of inflationary cosmology without singularity , '' phys .
d * 79 * , 021303 ( 2009 ) [ arxiv:0808.0819 [ astro - ph ] ] .
t. d. lee and g. c. wick , `` negative metric and the unitarity of the s matrix , '' nucl .
b * 9 * , 209 ( 1969 ) ; + t. d. lee and g. c. wick , `` finite theory of quantum electrodynamics , '' phys .
d * 2 * , 1033 ( 1970 ) .
j. m. cline , s. jeon and g. d. moore , `` the phantom menaced : constraints on low - energy effective ghosts , '' phys .
d * 70 * , 043543 ( 2004 ) [ arxiv : hep - ph/0311312 ] .
j. karouby and r. brandenberger , `` a radiation bounce from the lee - wick construction ?
, '' phys .
d * 82 * , 063532 ( 2010 ) [ arxiv:1004.4947 [ hep - th ] ] ; + j. karouby , t. qiu and r. brandenberger , `` on the instability of the lee - wick bounce , '' phys .
d * 84 * , 043505 ( 2011 ) [ arxiv:1104.3193 [ hep - th ] ] . c. lin
, r. h. brandenberger and l. p. levasseur , `` a matter bounce by means of ghost condensation , '' jcap * 1104 * , 019 ( 2011 ) [ arxiv:1007.2654 [ hep - th ] ] .
t. qiu , j. evslin , y. f. cai , m. li and x. zhang , `` bouncing galileon cosmologies , '' jcap * 1110 * , 036 ( 2011 ) [ arxiv:1108.0593 [ hep - th ] ] ; + d. a. easson , i. sawicki and a. vikman , `` g - bounce , '' jcap * 1111 * , 021 ( 2011 ) [ arxiv:1109.1047 [ hep - th ] ] . j. k. erickson , d. h. wesley , p. j. steinhardt and n. turok , `` kasner and mixmaster behavior in universes with equation of state w @xmath119 1 , '' phys .
d * 69 * , 063514 ( 2004 ) [ hep - th/0312009 ] .
g. w. horndeski , int .
. phys . * 10 * , 363 ( 1974 ) .
c. deffayet , x. gao , d. a. steer and g. zahariade , `` from k - essence to generalised galileons , '' phys .
d * 84 * , 064039 ( 2011 ) [ arxiv:1103.3260 [ hep - th ] ] .
r. hagedorn , `` statistical thermodynamics of strong interactions at high - energies , '' nuovo cim .
suppl . * 3 * , 147 ( 1965 ) .
m. gutperle , a. strominger , `` space - like branes , '' jhep * 0204 * , 018 ( 2002 ) .
[ hep - th/0202210 ] .
h. li , j. q. xia , r. brandenberger and x. zhang , `` constraints on models with a break in the primordial power spectrum , '' phys .
b * 690 * , 451 ( 2010 ) [ arxiv:0903.3725 [ astro-ph.co ] ] .
j. m. maldacena , `` non - gaussian features of primordial fluctuations in single field inflationary models , '' jhep * 0305 * , 013 ( 2003 ) [ arxiv : astro - ph/0210603 ] . y. -f .
cai , w. xue , r. brandenberger and x. zhang , `` non - gaussianity in a matter bounce , '' jcap * 0905 * , 011 ( 2009 ) [ arxiv:0903.0631 [ astro-ph.co ] ] .
b. xue and p. j. steinhardt , `` evolution of curvature and anisotropy near a nonsingular bounce , '' phys .
d * 84 * , 083520 ( 2011 ) [ arxiv:1106.1416 [ hep - th ] ] ; + b. xue and p. j. steinhardt , `` unstable growth of curvature perturbation in non - singular bouncing cosmologies , '' phys .
lett . * 105 * , 261301 ( 2010 ) [ arxiv:1007.2875 [ hep - th ] ] .
r. h. brandenberger , `` processing of cosmological perturbations in a cyclic cosmology , '' phys .
d * 80 * , 023535 ( 2009 ) [ arxiv:0905.1514 [ hep - th ] ] .
p. j. steinhardt and n. turok , `` cosmic evolution in a cyclic universe , '' phys .
d * 65 * , 126003 ( 2002 ) [ hep - th/0111098 ] . | a bouncing cosmology with an initial matter - dominated phase of contraction during which scales which are currently probed with cosmological observations exit the hubble radius provides a mechanism alternative to inflation for producing a nearly scale - invariant spectrum of cosmological perturbations . in this review article
i first discuss the evolution of cosmological fluctuations in the matter bounce scenario , and then discuss various attempts at realizing such a scenario .
observational signatures which will allow the matter bounce to be distinguished from the inflationary paradigm are also discussed . |
the central idea in brane - world scenarios is that matter and its interactions , as we know them , must be localized on the brane .
domain wall spacetimes , providing regularizations of the randall - sundrum brane - world @xcite preserve the property of produce effective 4d gravity on the brane @xcite .
in addition , while it has been shown that fermion fields may be localized on the brane _ via _ their interaction with the scalar field from which the wall is made of @xcite , gauge field localization has been somewhat more elusive @xcite .
a lot of work has been done on the topic of localization in brane - worlds in which the field theoretic domain walls considered arise in abelian @xmath1-symmetric theories , although it is also possible to consider ( besides the @xmath1 symmetry ) continuos internal symmetry walls with interesting localization properties .
there are well known examples of domain walls in flat space - time that arise in a @xmath0 symmetric theory @xcite in which the full @xmath0 symmetry is not restored at the core of the wall @xcite .
this makes non - abelian domain walls of this sort ( rather its extensions to the gravitating case ) very interesting within the context of brane - worlds @xcite .
it has been shown that non - abelian branes in flat space that break @xmath3 to @xmath4 may produce , _ via _ the dvali - shifman ( ds ) mechanism @xcite , an @xmath2 effective theory on the brane @xcite .
on the other hand , self - gravitating @xmath0 domain walls , supporting 4d massless fermion excitations in its world volume , may also localize 4d massless non - abelian excitations via the ds mechanism , depending of the symmetry breaking pattern @xcite .
while domain walls in abelian @xmath1-symmetric theories are topologically stable , there is no global stability criterium for the non - abelian ones .
this lead us to resort to perturbative analysis , after a solution is found , to establish at least their local stability @xcite .
furthermore , the inclusion of gravity makes the perturbative analysis somewhat more intricate than the analogous one in flat spacetime due to the mixing between the field excitations of the non - abelian kink @xmath5 and the scalar components of the metric @xmath6 , in a theory invariant under diffeomorphisms acting both on the @xmath5 and on @xmath6 . in this work
we make a perturbative analysis for the 5d self - gravitating @xmath0 domain walls of @xcite in terms of diffeomorphism - invariant quantities .
this analysis not only shows that these walls are perturbatively stable , it also permits to study the gravitationally trapped content from the point of view of 4d observers localized on the brane .
we find , as expected , that the tensor and vector sectors of the fluctuations behave as its counterparts in the more familiar @xmath1-symmetric domain walls . on the other hand ,
the scalar sector of the fluctuations in the @xmath0 case shows a rather different behavior from the abelian one , being linked to the particular symmetry breaking pattern considered and with a spectrum that includes normalizable massless modes , i.e. , with massless scalar particles associated to the broken symmetries which are gravitationally trapped on the core of the wall .
in this section we briefly recall the results of @xcite regarding the properties of 5d selfgravitating domain walls formed in the two possible symmetry breaking patterns of @xmath0 .
let us consider the @xmath7-dimensional theory @xmath8,\ ] ] where @xmath9 is the scalar curvature , @xmath10 is the determinant of the metric , @xmath5 is a scalar field that transforms in the adjoint representation of @xmath2 and @xmath11 a potential such that ( [ theory ] ) is invariant under the transformations * @xmath12 , @xmath13 , * @xmath14 , @xmath15 where @xmath16 , are traceless hermitian generators of @xmath2 . ]
the einstein - scalar field equations for this system are @xmath17 where @xmath18 and @xmath19 where @xmath20 .
exact domain wall solutions of ( [ field_eq1],[field_eq2],[field_eq3 ] ) are available for a sixth - order higgs potential of the form @xmath21+h ( \text{tr}[\phi^2])^2+\lambda \text{tr}[\phi^4 ] \nonumber\\&&\!\!+ \alpha ( \text{tr}[\phi^2])^3 + \beta ( \text{tr } [ \phi^3])^2 + \gamma ( \text{tr } \phi^4)(\text{tr } \phi^2),\end{aligned}\ ] ] for special values of the couplings which yield integrable models . assuming that the geometry preserves 4d - poincare invariance , the metric _ ansatz _ is @xmath22 with @xmath23 and a domain wall solution @xmath5 of the form @xmath24 is proposed , where @xmath25 satisfy the boundary conditions @xmath26 with @xmath27 and @xmath28 appropriately chosen orthogonal diagonal generators of @xmath2 . imposing the integrability conditions @xmath29 @xmath30 for the special values of the couplings given by @xmath31 \
, \gamma , \nonumber \\
\beta & = & \frac{1}{6}\left [ \frac{\text{tr } ( \mathbf{m}^4 ) - \text{tr } ( \mathbf{p}^4 ) } { \text{tr } ( \mathbf{p}^2 \mathbf{m } ) [ 3 \text{tr } ( \mathbf{p}^2 \mathbf{m } ) - 2 \text{tr } ( \mathbf{m}^3 ) ] } \right ] \ , \gamma,\end{aligned}\ ] ] an exact domain wall solution is given by @xcite @xmath32 where @xmath33 , @xmath34 y @xmath35 can be written explicitly in terms of @xmath36 and @xmath37 , and @xmath38 is a numerical constant which depends on the choice of @xmath27 and @xmath28 .
on the other hand ,
@xmath39 \label{warp}\ ] ] and the space - time is asymptotically @xmath40 with cosmological constant @xmath41 . as discussed in @xcite ( see @xcite for the flat space case ) , the choice of @xmath27 and @xmath28 relies on the desired asymptotic values for @xmath5 at @xmath42 .
it is well known that in flat space - time there are two symmetry breaking patterns of @xmath2 by a field @xmath5 in the adjoint , involving the breaking to subgroups with the same rank as @xmath2 and occurring as minima of a fourth - order potential @xcite .
these are @xmath43 and @xmath44 the second type of symmetry breaking pattern yielding the largest residual symmetry . in the remainder of this section , we quote the results of @xcite concerning to the symmetry breakings ( [ first ] ) and ( [ second ] ) for the gravitating case with the sixth order potential and refer the reader to that work for a detailed discussion on these and the other issues related to localization of fermions and gauge fields within the context of brane - worlds . for this symmetry breaking pattern we have @xmath46 @xmath47 and @xmath48 . with @xmath49 and @xmath50 the unbroken symmetries at @xmath42 and @xmath51 , respectively , we have @xmath52 @xmath53 with the following embeddings @xmath54 in the sense that the cartan subalgebra de @xmath55 ( @xmath56 ) is a subspace of the cartan - subalgebra space of @xmath57 ( @xmath58 ) corresponding to the particular basis chosen for the lie algebra .
@xmath59 resembles the gauge group @xmath60 on which is based the higgs sector of the minimal 4d left - right symmetric theory .
this 5d non - abelian wall may be thought as the generalization to the gravitating case of the perturbatively stable flat space one for a quartic potential of the references @xcite . in this case , @xmath62 @xmath63 and @xmath64 . the unbroken symmetries are @xmath65 @xmath66 where @xmath67 is embedded in diferent manners in @xmath68 and @xmath69 . in this symmetry breaking pattern , the @xmath67 gauge bosons on the wall need to form massive glueballs of @xmath70 in order to escape into the bulk , so they can be localized on the wall _ via _ the dvali - shifman mechanism @xcite .
let @xmath6 and @xmath71 be the metric and the scalar field solutions of the field equations of the theory ( [ theory ] ) .
let @xmath72 and @xmath73 the metric and scalar field fluctuations , respectively , around the above background .
following @xcite we find that these fluctuations satisfy @xmath74 -\frac{2}{3}h_{ab}v(\phi)-\frac{2}{3}\!\!\left(\frac{\partial v(\phi)}{\partial \phi_q}\varphi_q \right ) \!\!g_{ab}\end{aligned}\ ] ] and @xmath75 where @xmath20 and @xmath76 for the theory ( [ theory ] ) with the sixth order potential ( [ potential1 ] ) we have @xmath77 + 6 \alpha(\text{tr}[\phi^2])^2 \right.\nonumber\\&&\left.+ 2\gamma\text{tr}[\phi^4]\right)\text{tr}[\phi \mathbf{t}^q]+ 6\beta\text{tr}[\phi^3]\text{tr}[\phi^2 \mathbf{t}^q ] \nonumber\\&&+ 4\left(\lambda + \gamma\text{tr}[\phi^2]\right)\text{tr}[\phi^3 \mathbf{t}^q]\end{aligned}\ ] ] and @xmath78 + 6\alpha(\text{tr}[\phi^2])^2 \right.\nonumber\\&&\left.+2\gamma\text{tr}[\phi^4]\right)\text{tr}[\mathbf{t}^q\mathbf{t}^p]+ 6\beta\left(\text{tr}[\phi^3](\text{tr}[\mathbf{t}^q\mathbf{t}^p\phi ] \right.\nonumber\\ & & \left.+ \text{tr}[\mathbf{t}^p\mathbf{t}^q\phi ] ) + 3\text{tr}[\mathbf{t}^q\phi^2]\text{tr}[\mathbf{t}^p\phi^2]\right)\nonumber\\ & & + 4(\lambda + \gamma\text{tr}[\phi^2])\left(\text{tr}[\mathbf{t}^q\mathbf{t}^p\phi^2 ] + \text{tr}[\mathbf{t}^p\mathbf{t}^q\phi^2 ] \right.\nonumber\\ & & \left.+\text{tr}[\mathbf{t}^q\phi\mathbf{t}^p\phi]\right ) + 8(h + 3\alpha\text{tr}[\phi^2])\text{tr}[\mathbf{t}^q\phi]\text{tr}[\mathbf{t}^p\phi]\nonumber\\ & & + 8\gamma\left ( \text{tr}[\mathbf{t}^q\phi^3]\text{tr}[\mathbf{t}^p\phi ] + \text{tr}[\mathbf{t}^p\phi^3]\text{tr}[\mathbf{t}^q\phi ] \right).\end{aligned}\ ] ] in the background @xmath79 provided by the @xmath0 kinks discussed in the previous section , ( [ delta_v ] ) reduces to @xmath80 and the right hand side of ( [ delta_v_2 ] ) turns out diagonal in @xmath81 .
however , although they get notably simplified , ( [ fluctuations1 ] ) and ( [ fluctuations2 ] ) are still extremely involved .
indeed , as is well known , @xmath72 and @xmath82 are not diffeomorphism invariant variables . under an infinitesimal diffeomorphism of the form
@xmath83 we have @xmath84 and @xmath85 where @xmath86 and @xmath87 describe the same physical perturbations .
hence we must take care of the general coordinate invariance of the theory ( [ theory ] ) . since the background @xmath79 provided by the @xmath0 kinks preserves 4d - poincare invariance ,
we decompose @xmath72 ( from the point of view of the 4-dimensional observers confined to the wall ) in scalar , vector and tensor sectors .
following @xcite , we set @xmath88 @xmath89 and @xmath90 where @xmath91 and @xmath92 from ( [ change1 ] ) , with @xmath93 where @xmath94 we have that under an infinitesimal diffeomorphism @xmath95 @xmath96 @xmath97 and @xmath98 and from ( [ change2 ] ) we have @xmath99 the @xmath100 sector is automatically diffeomorphism invariant , i.e. @xmath101 , while it is also possible to define an invariant divergenceless vector ( since we have one vector gauge function @xmath102 ) @xmath103 and two diffeomorphism invariant scalar fluctuations ( since we have two gauge functions @xmath104 y @xmath105 ) given by @xmath106 and @xmath107 such that @xmath108 , @xmath109 and @xmath110 . on the other hand for the diffeomorphism invariant non - abelian scalar field fluctuation
we find @xmath111 such that @xmath112 . in the ( generalized ) longitudinal gauge @xcite , @xmath113 and @xmath114 , the freedom of the coordinate transformations ( [ difeomorphism ] )
is completely fixed and we have @xmath115 , @xmath116 , @xmath117 and @xmath118 .
this leaves us with the tensor @xmath119 , vector @xmath120 and scalar @xmath121 sectors which decouple from each other at the linearized level .
the tensor and vector sectors of the fluctuations behave as their corresponding analogous in the standard abelian @xmath1 kinks , widely discussed in the literature .
it is the scalar sector of the theory under consideration that is quite different from the @xmath1 case . nevertheless ,
in order to render this work self - consistent , we will also discuss briefly the tensor and vector fluctuations . in the following
, the results will be expressed in the conformal coordinate @xmath122 such that @xmath123 in the tensor sector @xmath124 with @xmath125 , the modes @xmath126 , satisfy the schrdinger - like equation @xmath127 where @xmath128 and @xmath129 .
@xmath130 supports a massless bound state which can be identified with the graviton and a tower of non - normalizable massive states which propagate in the bulk .
as is well known , the above equation can be factorized as @xmath131 which implies the absence of modes with @xmath132 , ensuring the stability of the system under tensor perturbations . as in the abelian @xmath1 kink @xcite ,
the normalizable massless tensor mode is given by @xmath133 which reproduces 4d gravity on the core of the wall , while the continuum of massive modes gives small corrections to this behavior at short distances .
the vector sector @xmath134 satisfies the equations @xmath135 and @xmath136 whose solution is given by @xmath137 , with @xmath138 . on the other hand , from ( [ invariant_vector ] ) we see that @xmath139 must be an odd function of @xmath140 and hence @xmath141 . as in the abelian @xmath1 kink @xcite , there are no massless vector fluctuations localized on the wall . in the scalar sector
we have the set of diffeomorphism - invariant fluctuations @xmath142 , which are subject to two constraints .
the first one only involves scalars without charge under @xmath2 and is given by @xmath143 while the second one involves also the component @xmath144 of @xmath145 @xmath146 from ( [ constraint1 ] ) and ( [ constraint2 ] ) it follows that the fluctuations @xmath147 , @xmath148 and @xmath144 are not independent and correspond to a single physical scalar fluctuation . with @xmath149 , the modes @xmath150
satisfy @xmath151 where @xmath152 is given by @xmath153 @xmath152 does not support normalizable massless states .
notice that @xmath154 where @xmath155 implying the absence of modes with @xmath132 .
this scalar fluctuation has an exact analogous one in the abelian @xmath1 selfgravitating kink ( see @xcite ) .
now , let us consider the scalar fluctuation @xmath156 .
let @xmath157 with @xmath158 the modes @xmath159 satisfy schrdinger - like equations which depends on @xmath160 .
first , let us consider perturbations along the @xmath27 direction .
for @xmath161 , ( it should be recalled that the fluctuations @xmath147 , @xmath148 and @xmath144 are not independent ) @xmath162 satisfy @xmath163 where @xmath164 is given by @xmath165 in terms of @xmath166 , as defined in ( [ z ] ) , ( [ m_mode ] ) can be rewritten as @xmath167 and from ( [ super1 ] ) and ( [ super2 ] ) , it follows that the eigenvalues @xmath168 of the modes of @xmath169 and @xmath170 always come in pairs , except possibly for the massless ones .
hence there are no modes with @xmath132 for @xmath170 . on the other hand , the massless solution @xmath171 of ( [ m_mode ] )
is given by @xmath172 which is however no normalizable since @xmath171 is not bounded for @xmath173 .
this is the mode which is expected to be related , when gravity is switched off , to the translational zero mode of the flat space @xmath0 kink @xcite . as in the abelian @xmath1 kinks @xcite , once we include gravity , the massless mode is no normalizable and therefore no localized .
the behavior of the scalar fluctuation @xmath170 close parallels the one of the scalar fluctuation associated to the standard abelian @xmath1 kink @xcite .
now , let us consider perturbations along the other generators @xmath174 of @xmath2 different from @xmath27 . for @xmath175 , @xmath176satisfy @xmath177 where @xmath178 is given by @xmath179 with @xmath130 given by ( [ vq1 ] ) . for @xmath180 , we find @xmath181 for both symmetry breakings .
@xmath182 is everywhere positive and hence does not support bound states with @xmath183 . as expected , @xmath5 is stable to perturbations along the @xmath5 direction .
furthermore , these perturbations do not generate normalizable 4d massless modes independently of the symmetry breaking pattern .
for @xmath174 a generator of @xmath184 and @xmath185 , we find for the symmetry breaking * a * @xmath186 where @xmath187 is the function defined by @xmath188 with @xmath189 ( [ xi_zeta ] ) . these are perturbations along the six generators of @xmath190 , the @xmath191 signs correspond to the two different @xmath192 . on the other hand , for the symmetry
breaking * b * we find @xmath193 for the perturbations along the eight generators of @xmath194 . in both symmetry
breaking patterns , @xmath195 is everywhere positive and hence does not support bound states with @xmath196 . from the above results , it follows that perturbations along the generators of @xmath184 do not generate normalizable 4d massless modes .
next , let us consider perturbations along those generators of @xmath2 that are not generators of @xmath184 , which will be called broken generators . for a perturbation along @xmath174 such that @xmath174 is a generator of @xmath197 but not of @xmath198 ( plus signs ) or a generator of @xmath198 but not of @xmath197 ( minus signs )
, we find for the symmetry breaking * a * @xmath199 and for the symmetry breaking * b * @xmath200 these ones are perturbations along the @xmath201 generators that are broken only at one side of the wall , with @xmath202 for the symmetry breaking * a * and @xmath203 for the symmetry breaking * b * ( @xmath201 is the dimension of the coset @xmath204 ) . the schrdinger - like equation for these fluctuations can not be written in terms of a fake superpotential as we did for @xmath170 .
however , on general grounds and from the shape of @xmath178 for these perturbations ( see fig.[shape ] ) , besides a mild resonance behavior , trapped massless modes are expected . for the @xmath176 modes of the fluctuations within the class described by @xmath205 ( for the class described by @xmath206
, @xmath178 is the mirror image ) .
( b ) detail of the area shown in ( a).,title="fig:",scaledwidth=24.0% ] for the @xmath176 modes of the fluctuations within the class described by @xmath205 ( for the class described by @xmath206 , @xmath178 is the mirror image ) .
( b ) detail of the area shown in ( a).,title="fig:",scaledwidth=24.0% ] these gravitationally trapped massless modes correspond to rotations of @xmath5 within the class described by @xmath204 . in the non - gravitating setting the analogous zero modes behave as @xmath207 and are not normalizable @xcite .
hence , gravitationally trapped massless modes associated to these perturbations can be seen as the zero modes in flat space which turn out normalizable and get localized once we include gravity . for this type of perturbations
there are no bound states in the flat space case @xcite , we find no modes with @xmath132 neither in the gravitating case . finally , for @xmath160 such that @xmath174 is not a generator of @xmath197 nor of @xmath198 , we find that @xmath208 and the modes @xmath176 satisfy @xmath209 as stated above , @xmath130 supports a massless bound state , a tower of non - normalizable massive states with @xmath210 and no modes with @xmath132 .
these ones are perturbations along the @xmath211 generators that are broken everywhere , with @xmath212 for the symmetry breaking * a * and @xmath213 for the symmetry breaking * b*. the above results show that the 5d self - gravitating @xmath0 domain walls of @xcite are perturbatively stable .
additionally , on both symmetry breaking patterns , normalizable 4d massless modes of the kink fluctuations @xmath214 appear when @xmath174 is not a generator of @xmath184 .
the existence of normalizable massless scalar modes is reminiscent of the situation envisaged in @xcite , although in the latter these modes appears in abelian @xmath1-symmetric models with @xmath215 scalar fields ( @xmath216 ) and a potential generated by a fake superpotential .
dynamical localization of gauge fields _ via _ the ds - mechanism @xcite in non - abelian domain wall scenarios has been put forward in @xcite and discussed in @xcite for a flat space model based on a @xmath3 symmetry group and in @xcite for a gravitating @xmath0 model . besides the fact that the stability of non - abelian kinks is not guaranteed and should be proved at least perturbatively @xcite , it is clear that the 4d phenomenology on the brane may be influenced by the inclusion of gravity which couples metric fluctuations with the wall fluctuations .
we have thus carried out a diffeomorphism - invariant analysis of the fluctuations of the 5d self - gravitating @xmath0 domain walls of @xcite , in order to determine their perturbative stability as well as their localization properties from the point of view of 4d observers .
we find no modes with @xmath132 for the fluctuations of the self - gravitating @xmath0 kinks , implying that these are perturbatively stable as in the flat space - time case @xcite .
not surprisingly , the tensor and vector sectors of the fluctuations behave in the same way as the corresponding ones of the standard @xmath1 kinks .
there is a normalizable massless mode in the tensor sector which gives rise to 4d gravity on the brane and a tower of non - normalizable massive states which propagate in the bulk .
there is no localized vector fluctuation . on the other hand ,
the scalar sector of the fluctuations for the self - gravitating @xmath0 kinks greatly differs from its analogous for the abelian @xmath1 case .
we find no normalizable 4d massless modes associated to the unbroken subgroup @xmath184 on the core of the wall , independently of the symmetry breaking pattern considered .
however , there are as many normalizable 4d zero modes as there are broken generators , i.e. , we find gravitationally localized massless 4d scalar particles without charge under @xmath184 which can be identified as ( the 4d zero modes of ) the nambu - goldstone fields associated to the symmetry breaking @xmath217 .
this gravitational trapping of nambu - goldstone bosons is presumably shared with other non - abelian domain - walls , a subject that deserves a further separate investigation .
we wish to acknowledge fruitful discussions with adel khoudeir , alejandra melfo and aureliano skirzewski .
s. j. huber and q. shafi , phys .
d * 63 * ( 2001 ) 045010 [ arxiv : hep - ph/0005286 ] .
g. r. dvali , g. gabadadze and m. a. shifman , phys .
b * 497 * , 271 ( 2001 ) [ arxiv : hep - th/0010071 ] .
s. l. dubovsky , v. a. rubakov , int . j. mod
. phys .
* a16 * ( 2001 ) 4331 - 4350 .
[ hep - th/0105243 ] .
t. vachaspati , phys .
d * 67 * ( 2003 ) 125002 [ hep - th/0303137 ] .
a. davidson , b. f. toner , r. r. volkas and k. c. wali , phys .
d * 65 * ( 2002 ) 125013 [ hep - th/0202042 ] .
a. davidson , d. p. george , a. kobakhidze , r. r. volkas and k. c. wali , phys .
d * 77 * ( 2008 ) 085031 [ arxiv:0710.3432 [ hep - ph ] ] .
n. chen , t. a. ryttov and r. shrock , phys .
d * 82 * ( 2010 ) 116006 [ arxiv:1010.3736 [ hep - ph ] ] .
m. giovannini , phys .
d * 64 * ( 2001 ) 064023 [ hep - th/0106041 ] .
p. callin and f. ravndal , phys .
d * 70 * ( 2004 ) 104009 [ arxiv : hep - ph/0403302 ] .
m. shaposhnikov , p. tinyakov and k. zuleta , jhep * 0509 * ( 2005 ) 062 [ hep - th/0508102 ] . | the fluctuations of 5d self - gravitating non - abelian kinks which arise from the breaking of the @xmath0 symmetric theory are analyzed within the context of brane - worlds . while tensor and vector sectors of these fluctuations behave like its counterparts in the standard abelian @xmath1 kinks , the mixing between the field excitations of the non - abelian kink and the scalar components of the metric makes the pure scalar sector of the theory very interesting .
the spectrum of these scalar fluctuations , which includes gravitationally trapped massless modes on the core of the wall associated to the broken symmetries , is discussed for the two classes of kinks that break @xmath2 to its maximal subgroups . |
ensemble classifiers have become very popular for classification and regression tasks .
they offer the potential advantages of robustness via bootstrapping , feature prioritization , and good out - of - sample performance characteristics ( @xcite ) .
however , they suffer from lack of interpretability , and oftentimes features are reported as `` word bags '' - e.g. by feature importance ( @xcite ) .
generalized linear models , a venerable statistical toolchest , offer good predictive performance across a range of prediction and classification tasks , well - understood theory ( advantages and modes of failure ) and implementation considerations and , most importantly , excellent interpretability . until recently ,
there has been little progress in bringing together ensemble learning and glms , but some recent work in this area ( e.g. @xcite ) has resulted in publicly - available implementations of glm ensembles .
nevertheless , the resulting ensembles of glms remain difficult to interpret .
meantime , human understanding of models is pivotal in some fields - e.g. in translational medicine , where machine learning influences drug positioning , clinical trial design , treatment guidelines , and other outcomes that directly influence people s lives .
improvement in performance without interpretability can be useless in such context . to improve performance of maximum - likelihood models , @xcite proposed to learn multiple centroids of parameter space .
built bottom - up , such ensembles would have only a limited number of models , keeping the ensemble interpretable . in this paper
, we work from a model ensemble down .
we demonstrate that minimum description length - motivated ensemble summarization can dramatically improve interpretability of model ensembles with little if any loss of predictive power , and outline some key directions in which these approaches may evolve in the future .
the problem of ml estimators being drawn to dominant solutions is well understood .
likewise , an ensemble consensus can be drawn to the ( possibly infeasible ) mode , despite potentially capturing the relevant variability in the parameter space .
relevant observations on this issue are made in @xcite , who have proposed centroid estimators as a solution .
working from the ensemble backwards , we use this idea as the inspiration to compress ensembles to their constituent centroids . in order to frame the problem of ensemble summarization as that of mdl - driven compression , we consider which requirements
a glm ensemble must meet in order to be compressible , and what is required of the compression technique .
to wit , these are : 1 .
representation * the ensemble members needs to be representible as vectors in a cartesian space * the ensemble needs to be `` large enough '' with respect to its feature set * the ensemble needs to have a very non - uniform distribution over features 2 .
compression : the compression technique needs to * capture ensemble as a number of overlapping or non - overlapping clusters * provide a loss measure * formulate a `` description length '' measure it is easy to see that glm ensembles can satisfy the representation requirement very directly .
it is sufficient to view ensembles of _ regularized _ glms as low - dimensional vectors in a high - dimensional space .
the dimensionality of the overall space will somewhat depend on the cardinality of the ensemble , on the strictness of regularization used , on the amount of signal in the data , on the order of interactions investigated , and on other factors influencing the search space of the optimizer generating the ensemble of glms .
coordinates in this space can be alternately captured by ( ideally standardized ) coefficients or , perhaps more meaningfully , by some function of statistical significance of the terms . in this work
, we apply the latter . for representation , we choose a basis vector of subnetworks . in order to identify this basis vector ,
we have experimented with gaussian mixture decomposition ( gmm ) ( finding clusters of vectors in model space ) and hierarchical clustering . for performance reasons , we present results using the latter technique , despite its shortcomings : instability and inability to fit overlapping clusters ( this may lead to overfitting ) .
nevertheless , in practice we find that this latter technique performs reasonably well .
optionally , to summarize the clusters , centroids can be fit
_ de novo _ once these groups of models are identified , or medoids can be used , obviating the need for further fitting .
here we use the first method , refitting centroids from training data on just the terms occurring in the models in a given cluster .
lastly , bayesian information criterion ( _ bic _ ) satisfies the representation scoring requirement .
the likelihood term serves as the loss function and the penalty term captures `` description length '' ( @xcite ) .
the bic - regularized glm ensembles were fit for binary - outcome datasets used in @xcite and using the software from the same paper ( number of bags = = 100 , other settings left at defaults ) .
the result of this step was an ensemble @xmath0 which , ignoring the outcome variable and the intercepts , could be captured via a non - sparse matrix as follows : @xmath1 where @xmath2 , the ensemble dimensionality , refers to the number of fitted models and @xmath3 to the number of terms found in the whole fitted ensemble .
importantly , @xmath2 is always an arbitrary parameter - the fact that partially motivated our study . for each dataset , the fitted ensembles
were then compressed using the following procedure .
first of all , for each ensemble we created the significance matrix s : @xmath4 where @xmath5 , and the p - value is determined from the fit of the linear model @xmath6 of the glm ensemble ( s is the heatmap in figure [ figure1 ] ) .
each row of @xmath7 projects every model @xmath6 into a multivariate cartesian space where each axis corresponds to model terms observed in the whole ensemble and each coordinate corresponds to log - scaled term significance .
log - scaling is introduced to induce separability of models that share terms in the presence or absence of collinear covariates that would be expected to influence , but not obviate , shared terms significance .
note that , in this representation , if @xmath8 is missing , @xmath9 . as an example of what the @xmath0 and @xmath7 matrices would look like after this step , for the * compressed * ensemble in figure [ figure2 ] , @xmath0 = @xmath10 and @xmath7 = @xmath11 of course , in practice it is the full ensemble we are interested in representing in this manner en route to compressing it , so @xmath0 and @xmath7 will have their dimensions , @xmath2 and @xmath3 , of @xmath12 for a computational - biological application using a regularized glm ensemble construction approach .
having constructed the matrix s for each ensemble in this manner , we clustered its rows ( the models coefficients ) using ward s clustering criterion ( @xcite ) and euclidean distance metric .
we then traversed the resulting dendrogram from left to right , using each cutpoint to perform model assignment to clusters implied by the leaves of the resulting dendrogram ( figure [ figure1 ] ) .
we captured the assignment of models to k clusters produced by each step in this traversal as a column vector @xmath13 expressed as a categorical variable with @xmath14 unique values , such that @xmath15 .
then , for each model ensemble term @xmath16 , we extracted the vector @xmath17 - a column slice through the matrix s for term @xmath16 .
we next defined a linear model @xmath18 with @xmath19 parameters .
the success of ensemble compression via the @xmath14 clusters implied by @xmath13 could then be assessed by defining @xmath20@xmath21 being the log - likelihood of the model .
this evaluation of cost across clustering levels to find maximal average likelihood compression , @xmath22 , could be viewed as using the bayes factor as the loss function for optimization , and the process of describing the model ensemble by centroids ( or medoids ) of the clusters of models described by @xmath13 could be described as an mdl - driven compression of the ensemble , using the bic - penalized likelihood as the measure of optimal compression .
it is worth adding that the aforementioned gmm approach for cluster membership assignment , which can also be driven by bic(@xcite , @xcite ) , does not imply a specific nested membership of models in clusters , but generally results in cluster membership strongly correlated with that identified via the hierarchical model clustering technique .
while the gmm approach is more robust ( e.g. , it s not path - dependent , and does nt require specification of a linkage function ) , it scales worse when number of terms in the ensemble is large .
using the datasets described above , we performed 3-fold cross - validation repeated three times , and for each training fold and repeat fitted medoid- and centroid - compressed model ensembles .
for each held - out fold , we then computed out - of - sample auc for every method ( table [ table : table1 ] ) .
additionally , we performed paired one - tailed t - tests comparing medoid and centroid compression strategies to uncompressed ensembles across folds .
all aucs arising from repeats and folds were averaged prior to t - test to avoid pseudo - replication issues .
while medoids performed slightly worse , on average , than uncompressed ensembles , centroids performance was degraded only in the statistically `` suggestive '' sense ( 0.05 < p < 0.1 ) .
note that in our experience using this technique on real - world datasets , larger datasets and continuous outcomes result in even smaller , if any , degradation of performance , performance being especially degraded for logistic regression .
in other words , we believe that these results , reported for binary outcomes , are essentially a lower bound .
maximum - likelihood methods identify the most likely fit in the parameter space . however ,
unless the most likely fit is vastly superior to all others and is sharply defined , a rare scenario in practice , the total probability of this fit in the infinite model ensemble may be very small .
for that reason , model ensembles are thought to be superior to individual models .
ensemble construction gains power by sampling multiple models from the parameter space but , by so doing , loses interpretability by introducing alternative parameter configurations and values .
we build on the understanding that model parameter space centroids should be sufficient to capture predictive power of large ensembles ( @xcite ) , while observing that such centroids exhibit better interpretability by having fewer parameters among the alternative models . working top - down , we demonstrate and validate on several datasets a novel approach to summarizing ensembles of glms .
our data shows this approach can result in models nearly identical to full ensembles in performance and vastly superior in interpretability , owing to dramatically reduced ensemble sizes ( figure [ figure2 ] ) .
in addition , since this approach can operate on any models that can be shoehorned into a cartesian space , it shows promise for compressing and thus summarizing ensembles of other types - for instance , causal model ensembles with individual models represented as orderings ( @xcite ) .
we posit that our approach can alter applicability of ensemble methods in general , making their use possible for a wide range of applications where the bottleneck has been the interpretability of results .
future directions of research may include multivariate classification methods beyond gmm and hierarchical clustering , as well as extension of this methodology beyond ensembles of glms to other types of predictive ensembles .
presented at nips 2016 workshop on interpretable machine learning in complex systems
.aucs by method and dataset .
last column shows p - value of one - tailed paired t - test vs randomglm ( @xmath23 : the compressed ensemble performs as well as the full ensemble ) .
values are averaged over 3 folds and 3 repeats . because of multiple repeats , standard errors are not shown . [
cols=">,<,<,<,<,<,<",options="header " , ]
the authors would like to acknowledge leon furchtgott and fred gruber for their invaluable feedback on the manuscript , and fred gruber for his help with latex .
song , l. , langfelder , p. , horvath , s. ( 2013 ) random generalized linear model : a highly accurate and interpretable ensemble predictor . _ bmc bioinformatics _
14:5 pmid : 23323760 doi : 10.1186/1471 - 2105 - 14 - 5 teyssier , m. and koller , d ( 2005 ) .
ordering - based search : a simple and effective algorithm for learning bayesian networks . _ proceedings of the twenty - first conference on uncertainty in ai ( uai ) _ ( pp .
584 - 590 ) .
chitraa , v. , thanamani a.s .
( 2013 ) , review of ensemble classification . _
international journal of computer science and mobile computing a monthly journal of computer science and information technology _ , vol .
2 , issue . 5 , may 2013 , pg.307 312 liaw , a. , wiener , m. ( 2002 ) classification and regression by randomforest r news 1822 hansen , m.h . , and yu , b. model selection and the principle of minimum description length .
_ journal of the american statistical association _ 96.454 ( 2001 ) : 746 - 774 .
carvalho , l.e . ,
lawrence , c.e .
, centroid estimators for inference in high - dimensional discrete spaces _ pnas _ , 105 : 32093214 ( 2008 ) ward , j. h. , jr .
hierarchical grouping to optimize an objective function , _ journal of the american statistical association _ , 58 , 236244 ( 1963 ) fraley , c. , raftery , a.e . , murphy , t.b . , and scrucca , l. ( 2012 ) mclust version 4 for r : normal mixture modeling for model - based clustering , classification , and density estimation technical report no .
597 , department of statistics , university of washington fraley , c. and raftery , a.e .
( 2002 ) model - based clustering , discriminant analysis and density estimation journal of the american statistical association 97:611 - 631 | over the years , ensemble methods have become a staple of machine learning .
similarly , generalized linear models ( _ glms _ ) have become very popular for a wide variety of statistical inference tasks .
the former have been shown to enhance out - of - sample predictive power and the latter possess easy interpretability .
recently , ensembles of glms have been proposed as a possibility . on the downside
, this approach loses the interpretability that glms possess .
we show that minimum description length ( _ mdl_)-motivated compression of the inferred ensembles can be used to recover interpretability without much , if any , downside to performance and illustrate on a number of standard classification data sets . |
in computer aided geometric design and geometric modeling , we often use some common surface to construct a complex model .
these common surfaces are generally called basic modeling surfaces , and they should have some advantages such as simple and widely used .
the simple surfaces refer the ones with low degree , for instance , quadratic surfaces @xcite and cubic surfaces @xcite .
the widely used surface refer the ones common in industrial design , for instance , ruled surfaces @xcite , swept surfaces @xcite and translational surfaces @xcite . then a primary task is to have a well study for these basic modeling surfaces .
certainly , how to represent these surface is the first problem in studying surface .
as we know , two representation forms are usually used as parametric form and implicit form . for the parametric representation ,
it is a main popular geometric representation in cg and cad @xcite .
it is easy to render and is helpful for some geometric operations , such as the computation of curvature or bounds and the control of position or tangency .
however , it is difficult in positional relationship determination and collision detection .
another limitation is that the parametric surfaces have lower ability in topology comparing with the implicit surfaces .
this is also the reason that modeling technology of implicit surfaces is increasing in more applications .
however , it is difficult to classify and control the implicit surfaces .
hence , it is a problem to determining some basic modeling surfaces in the implicit form , and furthermore , find a parametric representation if it has . in this paper
, we prefer to study the translational surface which is commonly used in industrial design .
the translational surface is a simple solution to the problem of interpolating a surface passing through two curves .
hence , people generally give two space curves @xmath0 and @xmath1 with parametrizations @xmath2 and @xmath3 , respectively , and the two curves define a translational surface @xmath4 .
however , as we mentioned above , newly geometric modeling often begin with implicit equations @xcite . in this situation , for better control and design ,
it is necessary to find the basic modeling surface and compute their parametric representations if exist .
the main task of this paper is to parametrize a given algebraic translational surface . considering the parametric form of translational surface ,
the two variables are separated .
based on this fact we firstly find a space curve which can play the role @xmath0 .
then we compute a parametrization @xmath2 for it .
successively , we find @xmath1 and compute a parametrization @xmath3 for it . finally , we get a parametrization @xmath4 for the given algebraic translational surface .
the paper is organized as follows .
first , we introduce some preliminaries and previous results ( see section 2 ) . in section 3
, we show a necessary and sufficient condition for that an algebraic surface has a parametric representation of the form @xmath4 .
the proof is constructive , and then a method for computing @xmath5 is developed .
finally , we show how the computation of @xmath5 can be improved , computationally speaking .
more precisely , the final problem consist in deciding the existence of two space curves and to compute a rational parametrization of them ( see section 4 ) .
let @xmath6 be a surface over an algebraically closed field of characteristic zero @xmath7 , and let @xmath8,\,\,\ox:=(x_1 , x_2 , x_3)$ ] , be the irreducible polynomial defining implicitly @xmath6 . in the following ,
we analyze if @xmath6 is a _ translational surface _ ; i.e. whether @xmath6 admits a parametrization of the standard form @xmath9 where @xmath10@xmath11 and in the affirmative case , we compute it .
we denote by @xmath12 the space curve over @xmath7 defined by the rational parametrization @xmath13 , for @xmath14 . throughout this paper , @xmath6 is assumed not to be a cylindrical surface .
one can check that @xmath15 defines a cylindrical surface if and only if there exists a constant vector @xmath16 such that @xmath17 , where @xmath18 denotes the partial derivative of the polynomial @xmath19 w.r.t . the variable @xmath20 . for this case , it is not difficult to get a parametrization for @xmath6 ( see @xcite ) . as a special case , we get that @xmath6 is not a plane .
the plane case has a trivial solution since if @xmath6 is defined by the polynomial @xmath21 , a parametrization of the form given in eq .
( [ eq - form ] ) is @xmath22 , where @xmath23 . in the following , we present some properties concerning the parametrizations @xmath24 , and the space curves @xmath25 .
these results , will play an important role in sections 3 and 4 .
[ p - properties ] if @xmath6 is a translational surface , then the following properties hold : * if @xmath5 is proper , then @xmath26 and @xmath27 are both proper parametrizations . *
there exist @xmath26 and @xmath27 such that they are both proper parametrizations . *
it holds that @xmath28 .
* it holds that @xmath12 is not a line , for @xmath14 .
* * let us prove that @xmath26 is a proper parametrization .
similarly , one shows that @xmath27 is proper .
let us assume that @xmath26 is not proper .
then , there exists @xmath29 such that @xmath30 ( @xmath31 is the algebraic closure of @xmath32 , and @xmath33 is a new variable ) .
thus , @xmath34 ( @xmath35 is a new variable ) .
this implies that @xmath5 is not proper .
therefore , we conclude that @xmath26 is proper which is a contradiction . *
let us assume that @xmath2 or @xmath3 is not proper .
then , we may reparametrize each of them using for instance the results in @xcite .
that is , there exists @xmath36 proper , and @xmath37 such that @xmath38 . under these conditions , we have that @xmath39 satisfies that @xmath40 , and thus @xmath41 is a parametrization of @xmath6 of the form given in eq .
( [ eq - form ] ) . *
clearly @xmath28 since @xmath5 parametrizes a surface @xmath6 .
* let us assume that @xmath1 is a line ( we reason similarly , if @xmath0 is a line ) .
then , a proper parametrization of @xmath1 is given by @xmath42 .
then , since @xmath43 , we get that @xmath44 .
thus , @xmath5 parametrizes the surface @xmath6 defined by @xmath45 , and the surface defined by @xmath46 . since @xmath19 is irreducible , we get that @xmath47 which is impossible since @xmath6 is not a cylindrical surface . [ r - properties ] in the following , we compute @xmath48 being proper ( see statement @xmath49 in proposition [ p - properties ] ) .
however , we note that the properness of @xmath48 does not imply that @xmath5 is proper .
if @xmath6 is a translational surface , then @xmath6 admits a parametrization of the form given in eq .
( [ eq - form ] ) , and @xmath50 can be assumed to satisfied some additional properties . in particular , we claim that there exists a @xmath51 such that @xmath52 , and @xmath53 . without loss of generality , we prove the following lemma by setting @xmath54 .
[ l - prop1 ] let @xmath6 be a translational surface .
then , there exists a proper parametrization @xmath50 such that @xmath55 , and @xmath56 .
* let @xmath57 be a parametrization of @xmath6 such that @xmath58 are proper ( see statement @xmath49 in proposition [ p - properties ] ) .
we consider a change of variable on the variable @xmath59 of the form @xmath60 and such that @xmath61 is defined at @xmath62 .
that is , @xmath63 let @xmath64 , where @xmath65 .
note that @xmath66 ( @xmath61 is defined at @xmath67 ) , and @xmath68 is proper since @xmath65 and @xmath69 is invertible and @xmath70 is proper .
furthermore , @xmath71 is a parametrization of @xmath6 ( @xmath72 is a parametrization of @xmath6 , and @xmath73 ) .
in addition , @xmath74 ( since @xmath61 is defined at @xmath67 , then @xmath75 is defined at @xmath67 ) .
we consider @xmath76 such that @xmath77 observe that the rational function @xmath78 satisfying ( i ) , ( ii ) and ( iii ) exists since these three conditions determine a non empty open subset @xmath79 . indeed : clearly , @xmath80 is an open subset of @xmath7 .
in addition , if @xmath81 , then @xmath82 which implies that @xmath83 .
this is impossible since @xmath6 is a surface parametrized by @xmath57 and then @xmath84 .
finally , we consider @xmath4 , where @xmath85 , and @xmath86 .
note that @xmath50 is such that @xmath55 , @xmath87 ( by ( i ) , we have that @xmath88 ) .
furthermore @xmath27 is proper since @xmath68 is proper .
[ t20 ] from the proof of lemma [ l - prop1 ] , one deduces that we can have different @xmath51 such that @xmath52 and @xmath53 since the solution space @xmath80 is a non empty open subset of @xmath7 .
in addition , for every @xmath51 satisfying the above conditions , we obtain different proper parametrizations @xmath26 and @xmath27 solving the problem . in section 4 , we will take a deeper look at these @xmath89 to simplify our computation ( see theorem [ t - mp1 ] ) .
in this section , we provide a necessary and sufficient condition for that an algebraic surface @xmath6 is translational ; i.e. @xmath6 has a parametric representation of the form @xmath4 .
the proof is constructive and then , a method for computing @xmath5 is developed ( see theorem [ t - p1 ] , corollary [ c - p1 ] , and statement 2 in theorem [ t - necsuf ] ) . for this purpose , we assume that we are in the conditions stated in section 1 and first , we prove theorem [ t - p1 ] where we show a necessary condition to that @xmath6 admits a parametrization of the form given in eq .
( [ eq - form ] ) .
this condition provides a method for computing the parametrization @xmath90 .
[ t - p1 ] let @xmath6 be a translational surface .
there exists @xmath91 such that @xmath92 parametrizes properly a space curve @xmath93 defined by the equations @xmath94 , where @xmath95 .
* proof . * since @xmath6 admits a parametrization of the form given in eq .
( [ eq - form ] ) , we have that @xmath96 , and in particular @xmath97 ( see lemma [ l - prop1 ] ) .
in addition , from @xmath96 , we also get that @xmath98 thus , it holds that @xmath99 where @xmath100 ( see lemma [ l - prop1 ] ) . finally ,
we prove that the equations @xmath94 define a space curve . indeed : since @xmath6 is not a cylindrical surface , we get that @xmath101 for every @xmath91 ( note that @xmath102 and thus , if @xmath103 then @xmath104 which would imply that @xmath105 and @xmath6 is a cylindrical surface ) . in addition , since @xmath106 and @xmath19 is irreducible , we get that @xmath107 are linearly independent . from theorem
[ t - p1 ] , one easily gets the following corollary where it is shown that in fact , there exists _ infinitely many
_ vectors @xmath108 that can be used to compute a proper parametrization @xmath26 ( as many as vectors given by @xmath109 ) .
corollary [ c - p1 ] generalizes theorem [ t - p1 ] in the sense that for each vector @xmath110 we obtain a rational space curve @xmath0 ( this space curve will depend on the vector @xmath108 considered ) such that a proper parametrization of @xmath0 is given by @xmath111 .
thus , @xmath112 is a parametrization of @xmath6 , where @xmath113 and @xmath114 .
[ c - p1 ] let @xmath6 be a translational surface . for every @xmath115 it holds that @xmath116 parametrizes properly a space curve defined by the equations @xmath94 , where @xmath95 .
in addition , @xmath112 is a parametrization of @xmath6 , where @xmath113 and @xmath114 .
[ pset - paramspace ] * we prove theorem [ t - p1 ] and corollary [ c - p1 ] in a constructive way , but we should remind that there may exists suitable @xmath108 not only from @xmath117 ( see example [ e-5 ] ) .
we prefer to give more discussions for @xmath108 from @xmath117 since it is enough to lead a parametrization algorithm . * in order to compute a rational proper parametrization of a space curve @xmath118 , one may apply for instance the results in @xcite or in @xcite ( we remind that any space curve can be birationally projected onto a plane curve ) . in the following , we assume that the parametrization @xmath26 is computed ( see theorem [ t - p1 ] and corollary [ c - p1 ] ) , and we show how to compute the parametrization @xmath27 . for this purpose ,
we consider @xmath119 where @xmath120 $ ] , @xmath121 , and @xmath122 $ ] .
we denote by @xmath123 the variety generated by @xmath124 . in theorem
[ t - necsuf ] , we show that every rational space curve @xmath125 provides a parametrization @xmath50 such that @xmath126 parametrizes @xmath6 .
in fact , theorem [ t - necsuf ] provides a necessary and sufficient condition for that algebraic surface has a parametrization of the form given in eq .
( [ eq - form ] ) , and the proof is constructive ( see statement @xmath127 ) . more precisely , once a parametrization @xmath26 is computed ( see corollary [ c - p1 ] ) , we apply theorem [ t - necsuf ] to compute a parametrization @xmath27 ( if it exists ) .
if @xmath27 does not exist , we can conclude that @xmath6 is not a translational surface . in order to show theorem [ t - necsuf ] , we first need to show the following lemma where some properties of @xmath128 and @xmath129 are proved . [ prop2 ] the following statements hold : * @xmath130 ( and then @xmath131 ) .
in addition , @xmath132 for every @xmath133 .
* @xmath134\setminus { \bbb k}[\ox]$ ] . * @xmath134\setminus { \bbb k}[t_1]$ ] . * proof .
* * let us assume that @xmath135 .
then , @xmath136 for every @xmath133 where @xmath26 is defined . by applying the change of variable @xmath137
, one gets that @xmath138 which contradicts our assumption of @xmath6 not being a cylindrical surface .
+ in addition , we also have that @xmath132 for every @xmath133 .
otherwise , @xmath139 which implies that @xmath136 which leads us to a contradiction reasoning as before .
* let us assume that @xmath140 $ ] .
that is , @xmath141 .
let @xmath133 such that @xmath142 .
then , we have that @xmath143 ( note that from statement 1 above , we have that @xmath132 for every @xmath133 ) . deriving w.r.t @xmath144
, we get that @xmath145 since @xmath146 , we obtain that @xmath147 which implies that @xmath148 .
thus , @xmath149 and @xmath150 by applying the change of variable @xmath137 , one gets that @xmath151 and in particular , @xmath152 , for every @xmath133 where @xmath153 exists .
this contradicts our assumption of @xmath6 not being a cylindrical surface .
therefore , we conclude that @xmath134\setminus { \bbb k}[\ox]$ ] . * in order to prove that @xmath134\setminus { \bbb k}[t_1]$ ] , one reasons similarly as in statement 2 .
[ t - necsuf ] the following statements are equivalent : * @xmath6 is a translational surface .
* there exists a proper parametrization @xmath50 such that @xmath126 parametrizes @xmath6 .
* it holds that @xmath154 .
in addition , every rational space curve @xmath125 provides a parametrization @xmath50 such that @xmath126 parametrizes @xmath6 .
* * clearly , @xmath6 admits parametrization of the form given in eq .
( [ eq - form ] ) ( i.e. @xmath6 is a translational surface ) if and only if there exists a parametrization @xmath50 such that @xmath126 parametrizes @xmath6 . * if @xmath6 admits a parametrization of the form given in eq .
( [ eq - form ] ) , theorem [ t - necsuf ] holds , and then @xmath155 .
thus , @xmath156 . indeed : let us assume that @xmath157 and then @xmath158 let us prove that this is impossible . for this purpose , we first consider @xmath133 such that @xmath142 . now , taking into account that @xmath159 we apply the change of variable @xmath137 , and we obtain that @xmath160 then , for @xmath161 we get that @xmath162 . since @xmath45 is irreducible and @xmath163 is not a constant ( note that @xmath164 and thus , if @xmath165 then @xmath166 ; hence , @xmath167 which is impossible by statement 1 in lemma [ prop2 ] ) , we get that @xmath168 and then , @xmath169 . substituting in ( i )
, we obtain that @xmath170 since @xmath142 , one gets that @xmath171 with @xmath172 . deriving w.r.t @xmath144 in ( ii ) , we have that @xmath173 but @xmath174 , hence @xmath175 which implies that @xmath176 and @xmath177 thus , from this equality we get that @xmath178 by applying the change of variable @xmath137 , we have that @xmath151 which can only happen for @xmath6 being a cylindrical surface , which contradicts our assumption .
therefore , we conclude that @xmath156 .
+ finally , since @xmath156 and @xmath179 does not depend on @xmath144 , we get that @xmath180 which implies that @xmath154 .
* let @xmath181 be a parametrization of a rational space curve @xmath125 .
thus , @xmath182 . then , @xmath183 and hence @xmath184 where @xmath185 .
furthermore , according to lemma [ prop2 ] ( see statements 2 and 3 ) , @xmath123 defines an algebraic set independent to @xmath144 , which means that @xmath186 defines a surface . in the following ,
we illustrate theorems [ t - p1 ] and [ t - necsuf ] with an example .
[ e-1 ] consider the surface @xmath6 defined implicitly by the polynomial @xmath187.$ ] one may check that @xmath6 satisfy the assumptions introduced in section 2 .
thus , we first determine a rational proper parametrization of the space curve @xmath0 defined by the polynomials @xmath188 ( see theorem [ t - p1 ] ) . for this purpose , taking into account statement 2 in remark [ pset - paramspace ] , we compute a proper parametrization of the plane curve defined by @xmath189 we get that @xmath190 is a proper parametrization of the plane curve . thus , a proper parametrization of @xmath0 is given by @xmath191 where @xmath192 , and @xmath193 .
we get @xmath194 now , we compute a rational proper parametrization of the space curve @xmath1 . for this purpose , we determine @xmath195 , where @xmath196 we get that + @xmath197 + thus , we compute a proper parametrization of the space curve defined by @xmath198 ( see theorem [ t - necsuf ] ) . reasoning as above , we get that @xmath199 is a rational proper parametrization of @xmath1 . + finally , we obtain that a parametrization of @xmath6 is given by @xmath200 @xmath201 @xmath202 in example [ e-1 ] , we obtain a parametrization @xmath126 of the surface @xmath6 that depends on the vector @xmath203 introduced in theorem [ t - p1 ] .
observe that for a particular value of this vector ( such that the polynomials @xmath107 define a space curve ) , we obtain a particular parametrization of the form given in eq .
( [ eq - form ] ) . in section 4
, we deal with this question and we show how the computation of @xmath5 can be ( computationally speaking ) improved by choosing any particular value for the vector @xmath203 ( compare with corollary [ c - p1 ] ) . in particular , in section 4 , two theorems are proved . the first one , theorem [ t - p2 ] , shows how the computation of @xmath27 can be improved by considering a new ( simpler ) variety that generates the same space curves than @xmath204 .
the second theorem , theorem [ t - mp1 ] , allow us to consider particular values for the parameters @xmath205 which makes easier the computation of the parametrizations @xmath48 .
in theorem [ t - necsuf ] ( section 3 ) , we show how once the parametrization @xmath26 is computed ( see theorem [ t - p1 ] and corollary [ c - p1 ] ) , one may determine the parametrization @xmath27 ( and then , a parametrization @xmath206 of the given surface @xmath6 ) by computing a proper parametrization of a space curve @xmath207 , where @xmath119 @xmath208 , and @xmath121 .
in fact , theorem [ t - necsuf ] provides a necessary and sufficient condition for that algebraic surface having a parametrization of the form given in eq .
( [ eq - form ] ) . more precisely ,
once a parametrization @xmath26 is computed ( see corollary [ c - p1 ] ) , we apply theorem [ t - necsuf ] to compute a parametrization @xmath27 ( if it exists ) by applying statement @xmath127 in theorem [ t - necsuf ] .
if @xmath27 does not exist , we can conclude that @xmath6 is not a translational surface . in the following theorem ,
we show that the computation of @xmath27 can be improved in the sense that we do not need to compute explicitly @xmath204 but a simpler variety that generates the same space curves than @xmath204 .
more precisely , we prove that for _ `` almost all '' _ pair of values @xmath209 , it holds that any rational space curve @xmath207 can be defined by the polynomials @xmath210,\,i=1,2.$ ] [ t - p2 ] let @xmath118 be a rational space curve such that @xmath207 .
there exists a non empty open subset @xmath211 such that for every @xmath212 , the space curve @xmath213 is defined by the polynomials @xmath214 $ ] , where @xmath215 for @xmath216 * proof .
* let @xmath217 be a proper parametrization of @xmath118 , and let @xmath218 , where @xmath219 are new variables .
since @xmath156 for every @xmath144 , we get that @xmath220 for every @xmath219 , where @xmath221 .
let @xmath222 where @xmath223 $ ] denotes the leader coefficient of @xmath224 w.r.t @xmath225 . clearly , @xmath226 is a non empty open subset of @xmath227 .
now , we apply the properties of specialization of resultants ( see lemma 4.3.1 , p. 96 in @xcite ) , and we get that @xmath228 therefore , since @xmath229 , and @xmath230 are linearly independent ( note that @xmath231 ) , we deduce that @xmath232 for every @xmath212 , and in particular @xmath213 is defined by the polynomials @xmath210,\,i=1,2 $ ] .
[ r - p2 ] let @xmath233 .
since @xmath234 , it holds that , up to constants in @xmath235 , @xmath236 where @xmath237 , and @xmath238 is such that @xmath26 is defined ( note that by statement 1 in by lemma [ prop2 ] , we have that @xmath239 ) .
therefore , for every @xmath240 , where @xmath241 the space curve @xmath213 is defined by the polynomials @xmath242 note that @xmath80 is again a non empty open subset of @xmath227 . in the following example , we consider the surface @xmath6 introduced in example [ e-1 ] , and we show how the computation of @xmath27 can be improved by applying theorem [ t - p2 ] or remark [ r - p2 ] . in addition , we motivate the result that we will prove in theorem [ t - mp1 ] , and we do not consider a generic vector @xmath203 ( see example [ e-1 ] ) . instead , we take a particular value for this vector ( @xmath243 ) , and we show how a parametrization of the form @xmath244 is provided for the input surface @xmath6 . first , we compute a parametrization of the space curve @xmath0 defined by the polynomials @xmath188 ( see theorem [ t - p1 ] ) .
we take the particular values @xmath243 , and we get that @xmath245 is a proper parametrization of @xmath0 .
now , we apply remark [ r - p2 ] , and we compute a parametrization of the space curve @xmath1 defined by the polynomials @xmath246 and @xmath247 ( note that @xmath248 ) .
we get that @xmath249 is a rational proper parametrization of @xmath1 . in example
[ e-2 ] , we compute @xmath26 and @xmath27 for a particular value @xmath251 ( compare with example [ e-1 ] ) .
in fact , we may take any value for this vector except @xmath252 . in this case ,
@xmath253 and thus , @xmath254 do not define a space curve .
this remark gives us the idea that the computation of @xmath26 can be simplified since for particular values of the vector @xmath203 , we can obtain a parametrization @xmath26 . in order to show this property , we fist prove a technical lemma .
* according to the construction , @xmath261 is a rational conical surface , and then the implicit polynomial defining @xmath255 , @xmath262 , is homogenous w.r.t @xmath263 . then without loss of generality , setting @xmath264
, we get @xmath265 is homogenous in @xmath266 and @xmath225 . then , if @xmath267 is such that @xmath268 , we get that @xmath269 ( if @xmath270 ) or @xmath271 ( if @xmath272 ) .
thus , @xmath273 or @xmath274 .
similar discussions in the cases of setting @xmath275 or @xmath276 . in the following , we consider the surface @xmath255 defined by the parametrization @xmath277 , where @xmath278 .
observe that @xmath279 parametrizes a surface since its jacobian has rank 2 ; otherwise , @xmath280 which implies that @xmath281 and then @xmath282 for @xmath283 , @xmath284 and @xmath285 .
thus , @xmath286 which is impossible since @xmath27 is not a line ( see statement 4 in proposition [ p - properties ] ) . under these conditions ,
theorem [ t - mp1 ] shows that we can improve the computation of @xmath26 by taking a particular value for the vector @xmath203 introduced in theorem [ t - p1 ] ( compare examples [ e-1 ] and [ e-2 ] ) .
we illustrate this result with examples [ e-3 ] and [ e-4 ] .
[ t - mp1 ] let @xmath6 be a translational surface . for every @xmath257 it holds that there exists @xmath287 such that @xmath116 parametrizes properly a space curve defined by the equations @xmath94 , where @xmath95 .
in addition , @xmath112 is a parametrization of @xmath6 , where @xmath113 and @xmath114 .
* there exists @xmath289 such that @xmath290 .
we assume w.l.o.g . that @xmath291 ( otherwise , we consider the reparametrization @xmath292 ) . thus , we have that @xmath293 .
now , we apply corollary [ c - p1 ] and we conclude that @xmath116 parametrizes properly a space curve defined by the equations @xmath94 , where @xmath294 .
in addition , @xmath112 is a parametrization of @xmath6 , where @xmath113 and @xmath114 .
* there does not exist @xmath289 such that @xmath290 .
then , we consider a reparametrization @xmath295 such that @xmath296 , where @xmath297 , and @xmath298 .
note that @xmath299 is again a parametrization of the surface @xmath255 . under these conditions , since @xmath300 , we get that @xmath301 then , @xmath302 , and thus @xmath303 hence , @xmath304 , and in particular @xmath305 therefore , @xmath306 parametrizes properly a space curve defined by the equations @xmath94 , where @xmath294 .
in addition , @xmath112 is a parametrization of @xmath6 , where @xmath307 and @xmath308 .
* theorem [ t - mp1 ] provides a necessary ( but not sufficient ) condition for the computation of @xmath26 . *
taking into account lemma [ l - mp1 ] , one may apply theorem [ t - mp1 ] for some vector of the form @xmath91 with @xmath258 and @xmath259 for some @xmath260 .
[ e-3 ] we consider the surface @xmath6 over @xmath309 implicitly defined by the polynomial + @xmath310 + now , let @xmath0 be the space defined by the polynomials @xmath45 and @xmath311 + @xmath312 + ( see theorem [ t - mp1 ] and statement @xmath49 in remark [ r - mp1 ] ) . by applying statement 2 in remark [ pset - paramspace ]
( see also example [ e-1 ] ) , we compute a proper rational parametrization of @xmath0 .
we get @xmath313 now , let @xmath1 be the space curve defined by the polynomials @xmath246 and @xmath247 ( note that @xmath248 ; see theorem [ t - p2 ] , and remark [ r - p2 ] ) .
we get that a rational proper parametrization of @xmath1 is given by @xmath314 [ e-4 ] let @xmath6 be the surface over @xmath309 implicitly defined by the polynomial @xmath316 according to statement @xmath49 in remark [ r - mp1 ] , one can select @xmath317 . reasoning similarly as in example
[ e-3 ] , we get a proper parametrization of @xmath0 given by @xmath318 and a proper parametrization of @xmath1 given by @xmath319 then , @xmath4 is a proper parametrization of the surface @xmath6 .
observe that @xmath320 and @xmath321 furthermore , @xmath55 and @xmath322 ( see lemma [ l - prop1 ] ) .
+ now , we consider a different vector @xmath323 , and we get the space curve defined by the polynomials @xmath45 and @xmath324 . here ,
reasoning as above , we get the proper parametrizations @xmath325 thus , @xmath4 is a proper parametrization of @xmath6 .
however , @xmath326 and obviously , there does not exist @xmath327 such that @xmath328 note that @xmath329 parametrizes the surface @xmath255 defined by the polynomial @xmath330 , and @xmath331 observe that if we set @xmath332 , we get that @xmath317 , and in this case , we get the parametrizations given in eq .
( [ e-4eq1 ] ) and eq .
( [ e-4eq2 ] ) .
in addition , taking into account theorem [ t - mp1 ] , we also get parametrizations by computing @xmath333 and @xmath334 now @xmath335 is a parametrization of @xmath6 , and @xmath336 and @xmath337 , which support lemma [ l - prop1 ] and theorem [ t - p1 ] . finally , one should note that eq .
( [ e-4eq1 ] ) and eq .
( [ e-4eq3 ] ) define the same space curve @xmath0 .
similarly , eq . ( [ e-4eq2 ] ) and eq . ( [ e-4eq4 ] ) are proper parametrizations of the same space curve @xmath1 . finally , we give an example to illustrate statement 1 in remark [ pset - paramspace ] .
more precisely , theorem [ t - p1 ] and corollary [ c - p1 ] are proved in a constructive way , taking @xmath338 , for any @xmath287 .
however , there may exist suitable @xmath108 not only from @xmath117 .
[ e-5 ] let @xmath6 be the surface of example [ e-4 ] implictly defined by the polynomial + @xmath339 if we set @xmath340 , we get the proper parametrizations of @xmath0 and @xmath1 given by @xmath341 one can check that @xmath4 is a parametrization of the surface @xmath6 .
however , @xmath326 parametrizes the surface @xmath255 defined by the polynomial @xmath330 , and @xmath342 .
bricault , i. , monga , o. ( 1995 ) . _ modeling 3d objects with quadratic surface patches : application to the recognition and locating of anatomic structures .
_ computer vision , virtual reality and robotics in medicine , lncs905 , springer berlin heidelberg , pp .
460 - 464 .
glymph , j. , shelden , d. , ceccato , c. , mussel , j. , schober , h. ( 2004 ) . _ a parametric strategy for free - form glass structures using quadrilateral planar facets .
_ automation in construction .
13(2 ) , pp . 187 - 202 .
schroeder , w. j. , lorensen , w. e. , linthicum , s. ( 1994 ) .
_ implicit modeling of swept surfaces and volumes_. in proceedings of the conference on visualization94 ( pp .
40 - 45 ) .
ieee computer society press .
weinert , k. , du , s. , damm , p. , stautner , m. ( 2004 ) .
_ swept volume generation for the simulation of machining processes .
_ international journal of machine tools and manufacture .
44(6 ) , pp . 617 - 628 . | the algebraic translational surface is a typical modeling surface in computer aided design and architecture industry . in this paper
, we give a necessary and sufficient condition for that algebraic surface having a standard parametric representation and our proof is constructive .
if the given algebraic surface is translational , then we can compute a standard parametric representation for the surface .
* keywords : * translational surface , rational parametrization , reparametrization |
orbital angular momentum ( am ) is one of the most fundamental quantities in classical and quantum mechanics . in condensed matter physics ,
the bulk orbital am in chiral superconductors ( scs ) has been intensively studied @xcite .
a chiral sc is an exotic sc whose cooper pairs carry nonzero orbital am @xmath6 , and hence the time - reversal symmetry is spontaneously broken .
@xmath7he - a is the only material experimentally established as a chiral superfluid , whose pairing symmetry is chiral @xmath8 wave @xcite .
also , sr@xmath9ruo@xmath10 is widely believed to be a chiral @xmath8-wave sc @xcite .
therefore , most of the previous studies have focused on chiral @xmath8 wave and at least theoretically concluded @xmath11 in an ideal situation , where @xmath4 and @xmath5 are the bulk orbital am and the total number of particles at zero temperature , respectively @xcite . in general ,
chiral scs with @xmath6 are expected to have @xmath12 without an edge or a vortex @xcite , which we call the intrinsic value .
recently , it was found that the bulk orbital am in two - dimensional chiral scs with @xmath13 is remarkably reduced as @xmath14 @xcite . in two dimensions , chiral scs
are gapped and are topological scs characterized by the chern number @xmath15 and the presence of the majorana edge states . when @xmath16 , only one edge state is particle - hole symmetric ( phs ) by itself and
is called the phs edge state @xcite . on the other hand ,
when @xmath13 is even , each edge state is phs with another but not phs by itself .
such edge states are called the non - phs edge states @xcite .
when @xmath13 is odd , there are the phs and non - phs edge states .
the non - phs edge states give rise to the nonzero spectral asymmetry and reduce the bulk orbital am from the intrinsic value @xmath12 .
although chiral @xmath17 or @xmath3 wave has been proposed in upt@xmath18 @xcite , uru@xmath9si@xmath9 @xcite , and srptas @xcite , the theory can not be applied to these three - dimensional materials with nodes directly .
apparently , the above reduction of the bulk orbital am requires not @xmath13 but @xmath19 .
realization of two - dimensional time - reversal broken topological scs with @xmath19 is classified into two types .
the first type is intrinsic , namely , without impurities , and can be realized in a quantum anomalous hall insulator in proximity to an @xmath0-wave sc @xcite , a heterostructure of a hole - doped semiconductor thin film , a magnetic insulator , and an @xmath0-wave sc @xcite , and the systems with the help of @xmath8-wave scs @xcite .
the second type relies on impurities and can be realized in a lattice of magnetic impurities on the surface of an @xmath0-wave sc with the rashba spin - orbit interaction ( soi ) @xcite and that of nonmagnetic impurities in a chiral @xmath8-wave sc @xcite .
in these proposals , arbitrarily high chern numbers are available in principle . in return for using a conventional @xmath0-wave sc , sois play an important role in generating the effective chirality of cooper pairs . among these proposals , a hole - doped topological sc proposed in ref .
is a natural extension of an electron - doped topological sc composed of an electron - doped semiconductor thin film , a magnetic insulator , and an @xmath0-wave sc @xcite and offers an experimentally feasible system to verify the theory of the bulk orbital am .
we note that hole - doped semiconductors accompanied by ferromagnetism can be realized by mn doping and are called dilute magnetic semiconductors @xcite .
however , it is not trivial how the bulk orbital am is reduced in this system because the orbital am is not conserved in the presence of the sois . in this paper , we investigate the bulk orbital am in a hole - doped topological sc @xcite comparing with that in an electron - doped topological sc @xcite .
we calculate the bulk orbital am both by the green s function formula @xcite and on a circular disk . as mentioned above
, the non - phs edge states dramatically reduce the bulk orbital am in the case of @xmath19 .
in addition , tsutumi and machida pointed out that @xmath11 in chiral @xmath8-wave scs consists of @xmath20 from the majorana edge state and @xmath21 from the continuum states @xcite .
the authors already found that @xmath2 in an electron - doped topological sc is continuous at the topological phase transition and is nonzero even in the trivial phase which does not support the majorana edge state @xcite .
therefore it is important to reveal the difference of the contributions to the bulk orbital am from the phs edge , non - phs edge , and continuum states .
we also discuss the effects of the sois which are indispensable for topological scs .
we find that the conserved quantity is modified by sois . when the chern number is higher than @xmath22 , it can be nonzero owing to the presence of the non - phs edge states and the nonzero spectral asymmetry , which results in the reduction of the bulk orbital am as shown in refs . .
since the bulk orbital am calculated in the reciprocal space is continuous at the topological phase transition , that calculated in the real space shows a jump . differently from the existing literature ,
this reduction is not universal and can be tuned by sois in this system .
first , we review a hole - doped topological sc composed of a hole - doped semiconductor thin film , a magnetic insulator , and an @xmath0-wave sc @xcite , @xmath23 \tau_z - 2 h j_z \notag \\ & + [ ( \delta_{\rm h } - \delta_{\rm l})/2 + ( \delta_{\rm h } + \delta_{\rm l } ) \gamma_5/2 ] \tau_x , \label{eq : luttingertsc - hk}\end{aligned}\ ] ] where we choose the nambu basis , @xmath24^{\rm t}$ ] .
@xmath25 is a set of the pauli matrices for the nambu space , and @xmath26 is the total am of @xmath27 holes but is called spin to avoid confusion below .
we use the standard representation , @xmath28 [ eq : j3/2 ] and construct the @xmath29 matrices satisfying the anticommutation relations @xmath30
@xcite , @xmath31 [ eq : gamma3/2 ] the first , second , and third terms in eq .
are the two - dimensional luttinger hamiltonian , which describes a hole - doped semiconductor thin film @xcite .
we define @xmath32 and @xmath33 , in which @xmath34 $ ] is a two - dimensional wavevector , @xmath35 and @xmath36 are the luttinger parameters , and @xmath37 is the chemical potential . in a hole - doped semiconductor , confinement into a quantum well yields quantization of momentum @xmath38 and @xmath39 and opens up the gap @xmath40 between the heavy - hole bands with @xmath41 and the light - hole bands with @xmath42 .
this point is a crucial difference from an electron - doped semiconductor , in which confinement just shifts the chemical potential .
the fourth and fifth terms are the rashba soi and the zeeman interaction , respectively , and the sixth and seventh terms are the @xmath0-wave pairing potentials induced by the proximity effect .
note that we neglect some terms allowed by the theory of invariants @xcite .
the topological phase transitions occur when the gap closes at @xmath43 .
we can analytically obtain eight eigenvalues , @xmath44 and @xmath45 , and two phase transition lines , @xmath46 and @xmath47 , where the chern number changes by @xmath48 and @xmath49 , respectively .
the phase diagrams for different @xmath50 and @xmath51 are shown in fig .
[ fig : luttingertsc - phase ] .
@xmath1 is realized when the chemical potential lies in the lowest heavy - hole band alone , which is described by the cubic rashba model with the zeeman interaction in the large-@xmath52 limit @xcite .
we should note that the gap may accidentally close at @xmath53 , which does not change the chern number .
below we put @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 focusing on fig .
[ fig : luttingertsc - phase](a ) , leading to @xmath60 and @xmath61 .
once we choose energy and wavenumber units , the remaining important parameters are @xmath62 , @xmath63 , @xmath64 , @xmath65 , @xmath66 , and @xmath67 . among these , @xmath66 and @xmath67 drastically change the phase diagram as shown in fig .
[ fig : luttingertsc - phase ] . here
we arbitrarily choose @xmath68 since we can not estimate their experimental values .
@xmath62 takes @xmath69-@xmath70 for typical iii - v semiconductors such as gaas @xcite , and hence our choice of @xmath71 is realistic .
@xmath63 can be tuned experimentally by doping carriers or applying an external field .
note that @xmath52 is about @xmath72 for a gaas quantum well with @xmath73 thickness .
we calculate the bulk orbital am in two ways .
one is based on the green s function formula in the reciprocal space @xcite .
since there is no edge state in the periodic boundary condition , we do not discuss the effects of the non - phs edge states but interpret the resulting bulk orbital am as the intrinsic value .
here we calculate the hall viscosity @xmath74 , which is known to be equal to the bulk orbital am @xmath4 in two - dimensional gapped systems at zero temperature @xcite . by using the matsubara green s function @xmath75 , the chern number @xcite and the hall viscosity @xcite
can be evaluated by @xmath76 [ eq : luttingertsc - k ] here we introduce @xmath77 , and @xmath78 is the totally antisymmetric tensor with @xmath79 .
although the integrals are carried out over the @xmath80 space , the prefactor @xmath81 in eq .
is in two dimensions .
the other way is a straightforward calculation in the real space .
we solve the bogoliubov - de gennes ( bdg ) equation @xmath82 on a disk with its radius @xmath83 .
since the system preserves the rotational symmetry , the bdg wave function can be expressed as @xmath84^{\rm t } , \label{eq : luttingertsc - wfr}\ ] ] and the radial bdg equation is written by @xmath85 u_{\ell i 0}(r ) - \sqrt{3 } \alpha { \cal r}_{\ell + 1}(r ) u_{\ell i 1}(r ) + ( \sqrt{3 } \gamma_2/2 m ) { \cal s}_{\ell + 2}(r ) u_{\ell i 2}(r ) \notag \\ & + \delta_{\rm h } v_{\ell i 0}(r ) = e_{\ell i } u_{\ell i 0}(r ) , \label{eq : luttingertsc - bdg0 } \\ &
\sqrt{3 } \alpha { \cal r}_{-\ell}(r ) u_{\ell i 0}(r ) + [ -(\gamma_1 - \gamma_2 ) { \cal l}_{\ell + 1}(r)/2 m - \mu + w - h ] u_{\ell i 1}(r ) - 2 \alpha { \cal r}_{\ell + 2}(r ) u_{\ell i 2}(r ) + ( \sqrt{3 } \gamma_2/2
m ) { \cal s}_{\ell + 3}(r ) u_{\ell i 3}(r ) \notag \\ & - \delta_{\rm l } v_{\ell i 1}(r ) = e_{\ell i } u_{\ell i 1}(r ) , \label{eq : luttingertsc - bdg1 } \\ &
( \sqrt{3 } \gamma_2/2 m ) { \cal s}_{-\ell}(r ) u_{\ell i 0}(r ) + 2 \alpha { \cal r}_{-(\ell + 1)}(r ) u_{\ell i 1}(r ) + [ -(\gamma_1 - \gamma_2 ) { \cal l}_{\ell + 2}(r)/2 m - \mu + w + h ] u_{\ell i 2}(r ) - \sqrt{3 } \alpha { \cal r}_{\ell + 3}(r ) u_{\ell i 3}(r ) \notag \\ & - \delta_{\rm
l } v_{\ell i 2}(r ) = e_{\ell i } u_{\ell i 2}(r ) , \label{eq : luttingertsc - bdg2 } \\ & ( \sqrt{3 } \gamma_2/2 m ) { \cal s}_{-(\ell + 1)}(r ) u_{\ell i 1}(r ) + \sqrt{3 } \alpha { \cal r}_{-(\ell + 2)}(r ) u_{\ell i 2}(r ) + [ -(\gamma_1 + \gamma_2 ) { \cal l}_{\ell + 3}(r)/2 m - \mu - w + 3 h ] u_{\ell i 3}(r ) \notag \\ & + \delta_{\rm h } v_{\ell i 3}(r ) = e_{\ell i } u_{\ell i 3}(r ) , \label{eq : luttingertsc - bdg3 } \\ &
-[-(\gamma_1 + \gamma_2 ) { \cal l}_{\ell}(r)/2 m - \mu - w + 3 h ] v_{\ell i 0}(r ) + \sqrt{3 } \alpha { \cal r}_{\ell + 1}(r ) v_{\ell i 1}(r )
- ( \sqrt{3 } \gamma_2/2 m ) { \cal s}_{\ell + 2}(r ) u_{\ell i 2}(r ) \notag \\ & + \delta_{\rm h } u_{\ell i 0}(r ) = e_{\ell i } v_{\ell i 0}(r ) , \label{eq : luttingertsc - bdg4 } \\ &
-\sqrt{3 } \alpha { \cal r}_{-\ell}(r ) v_{\ell i 0}(r ) - [ -(\gamma_1 - \gamma_2 ) { \cal l}_{\ell + 1}(r)/2 m - \mu + w + h ] v_{\ell i 1}(r ) + 2 \alpha { \cal r}_{\ell + 2}(r ) v_{\ell i 2}(r ) - ( \sqrt{3 } \gamma_2/2 m ) { \cal s}_{\ell + 3}(r ) v_{\ell i 3}(r ) \notag \\ & - \delta_{\rm l } u_{\ell i 1}(r ) = e_{\ell i } v_{\ell i 1}(r ) , \label{eq : luttingertsc - bdg5 } \\ &
-(\sqrt{3 } \gamma_2/2 m ) { \cal s}_{-\ell}(r ) v_{\ell i 0}(r ) - 2 \alpha { \cal r}_{-(\ell + 1)}(r ) v_{\ell i 1}(r ) - [ -(\gamma_1 - \gamma_2 ) { \cal l}_{\ell + 2}(r)/2 m - \mu + w - h ] v_{\ell i 2}(r ) + \sqrt{3 } \alpha { \cal r}_{\ell + 3}(r ) v_{\ell i 3}(r ) \notag \\ & - \delta_{\rm l } u_{\ell i 2}(r ) = e_{\ell i } v_{\ell i 2}(r ) , \label{eq : luttingertsc - bdg6 } \\ &
-(\sqrt{3 } \gamma_2/2 m ) { \cal s}_{-(\ell + 1)}(r ) v_{\ell i 1}(r ) - \sqrt{3 } \alpha { \cal r}_{-(\ell + 2)}(r ) v_{\ell i 2}(r ) - [ -(\gamma_1 + \gamma_2 ) { \cal l}_{\ell + 3}(r)/2 m - \mu - w - 3 h ]
v_{\ell i 3}(r ) \notag \\ & + \delta_{\rm h } u_{\ell i 3}(r ) = e_{\ell i } v_{\ell i 3}(r ) , \label{eq : luttingertsc - bdg7}\end{aligned}\ ] ] [ eq : luttingertsc - bdg ] here we introduce @xmath86 , @xmath87 , and @xmath88 \partial_r + \ell ( \ell + 2)/r^2 $ ] .
we impose the boundary conditions @xmath89 ( @xmath90 ) and the normalization condition , @xmath91 = 1 .
\label{eq : rashbatsc - norm}\ ] ] owing to the phs @xmath92 with @xmath93 , when the bdg wave function eq .
has the eigenvalue @xmath94 , @xmath95^{\rm t } , \label{eq : luttingertsc - wfrc}\ ] ] has the eigenvalue @xmath96 .
therefore we only have to solve the radial bdg equation eq . for @xmath97 .
the total number of particles @xmath5 , the total spin @xmath98 , and the bulk orbital am @xmath4 can be calculated by @xmath99 , \label{eq : luttingertsc - nr } \\
j_z = & \sum_{\ell = -1}^{\ell_{\rm max } } \sum_i \int_0^r r { \rm d } r [ { \bm u}_{\ell i}^{\rm t}(r ) j_z
{ \bm u}_{\ell i}(r ) f(e_{\ell i } ) - { \bm v}_{\ell i}^{\rm t}(r ) j_z { \bm v}_{\ell i}(r ) f(-e_{\ell i } ) ] , \label{eq : luttingertsc - jzr } \\
l_z = & \sum_{\ell = -1}^{\ell_{\rm max } } ( \ell + 3/2 ) \sum_i \int_0^r r { \rm d } r [ { \bm u}_{\ell i}^2(r ) f(e_{\ell i } ) - { \bm v}_{\ell i}^2(r ) f(-e_{\ell i } ) ] - j_z , \label{eq : luttinegrtsc - lzr}\end{aligned}\ ] ] [ eq : luttingertsc - r ] respectively , in which @xmath100 is the fermi - dirac distribution function at @xmath101 .
we use the central difference method with @xmath102 grid points , which reduces a set of the differential equations eq . to the @xmath103 eigenvalue problem .
we put the disk radius @xmath104 , the number of grid points @xmath105 , and the am cutoff @xmath106 . figure [ fig : luttingertsc - eigen ] shows the low - energy spectra for @xmath107 .
there are one phs edge mode and one pair of the non - phs edge modes in fig .
[ fig : luttingertsc - eigen](b ) , which results from @xmath1 . on the other hand , there are two pairs of the non - phs edge modes in fig .
[ fig : luttingertsc - eigen](c ) , which results from @xmath108 .
figure [ fig : luttingertsc - am ] shows the @xmath109 and @xmath110 dependences of the total spin @xmath111 and the bulk orbital am @xmath2 obtained by the real and reciprocal spaces .
we find that @xmath111 obtained by the real space always coincides with that by the reciprocal space , while @xmath2 obtained by the real space is reduced compared with that by the reciprocal space for @xmath112 .
therefore these two are denoted by @xmath113 and @xmath114 , respectively , below .
this reduction of the bulk orbital am is expected from the previous studies @xcite but not strong when @xmath109 is small or @xmath110 is large .
this mechanism and quantitative criterion are discussed later .
we also find @xmath115 as in an electron - doped topological sc @xcite .
owing to the presence of the sois , not the orbital am @xmath116 but the total am @xmath117 is conserved . in the reciprocal space , we can easily check @xmath118 = & -[j_z , { \cal h}_{\vec k } ] = 2 i [ \sqrt{3 } \xi_{2k } ( \gamma_3 \cos 2 \phi - \gamma_4 \sin 2 \phi ) \notag \\ & + \alpha k ( j_x \cos
\phi + j_y \sin \phi ) ] \tau_z , \label{eq : luttingertsc - comm}\end{aligned}\ ] ] and @xmath119 = 0 $ ] for each @xmath120 .
therefore the eigenstates are the simultaneous eigenstates of @xmath121 . by summing up the eigenvalues of @xmath121 over the negative - energy eigenstates
, we obtain the bulk total am @xmath122 .
such discussion holds only when the negative - energy eigenstates are well defined , namely , the system is gapped . in the real space ,
eq . is rewritten as @xmath123 \notag \\ & + \frac{1}{2 } \sum_{\ell = -1}^{\ell_{\rm max } } ( \ell + 3/2 ) \sum_i \int_0^r r { \rm d } r [ { \bm u}_{\ell i}^2(r ) - { \bm v}_{\ell i}^2(r ) ] \notag \\ = & -\frac{1}{2 } \sum_{\ell = -1}^{\ell_{\rm max } } ( \ell + 3/2 ) \eta_{\ell } , \label{eq : rashbatsc - lzr2}\end{aligned}\ ] ] by using the spectral asymmetry , @xmath124 .
\label{eq : spec - asym}\ ] ] this is similar to the previous results on chiral scs @xcite but is modified by the presence of the sois @xcite . for @xmath1 as shown in fig .
[ fig : luttingertsc - eigen](b ) , the non - phs edge states cross @xmath125 at @xmath126 with @xmath127 being called the fermi am @xcite , which leads to the nonzero spectral asymmetry @xmath128 for @xmath129 and reduces the bulk orbital am by @xmath130 .
differently from chiral scs , the reduction of the bulk orbital am depends on @xmath109 and @xmath110 through the fermi am . to see this , we show the @xmath109 and @xmath110 dependences of the fermi am normalized by @xmath131 in fig .
[ fig : luttingertsc - fermi ] since we are interested not in @xmath132 but in @xmath133 .
we find that the fermi am is large enough to reduce the bulk orbital am almost to zero for @xmath134 but is small in the opposite case , leading to @xmath135 . here
@xmath136 is the effective strength of the rashba soi for the light - hole bands .
the factor @xmath137 comes from the matrix elements in @xmath138 and @xmath139 .
this is because the sharp decrease of the fermi am around @xmath140 is accompanied by the lifshitz transition in the normal state .
the lowest particle band is always one of the heavy - hole bands with @xmath141 , but the second lowest particle band changes from the other heavy - hole band with @xmath142 to one of the light - hole bands with @xmath143 at @xmath144 . in order to realize the topological phase with @xmath1 , @xmath43 of the second lowest particle band should be above from the chemical potential
however , when the strength of the rashba soi overcomes the zeeman interaction , this light - hole band has its minimum at @xmath53 , and hence the lifshitz transition can occur as shown in fig .
[ fig : luttingertsc - e0k ] .
the non - phs edge states connect the heavy - hole band with the antiparticle counterpart of the light - hole band before the lifshitz transition .
once the lifshitz transition occurs , the non - phs edge states connect the heavy - hole band with the inner part of the antiparticle counterpart of the light - hole band , which leads to the sharp decrease of the fermi am .
the @xmath109 and @xmath110 dependences of the bulk orbital am in the reciprocal space @xmath114 in a hole - doped topological sc can be understood in the same way as those in an electron - doped topological sc @xcite .
we translate the previous results for electrons to heavy holes . to do this
, we define the effective strength of the rashba soi for heavy holes by @xmath145 . for the small sois @xmath146 in the trivial phase , the @xmath0-wave
pairing potential is dominant , and hence we obtain @xmath147 . for @xmath148 in the topological phase with @xmath1 , we expect @xmath149 from the @xmath150- wave pairing potential . in between
, @xmath114 rapidly increases even in the trivial phase since the @xmath0-wave pairing potential is no longer dominant due to the pauli depairing effect , and the @xmath151-wave pairing potentials coexist . for the large sois @xmath152 , spins are tilted in the @xmath153 plane , and @xmath114 is suppressed through the conservation law @xmath154 . note that @xmath114 shown in fig .
[ fig : luttingertsc - am](b ) does not reach @xmath155 even for @xmath148 because the zeeman interaction does not overcome the other soi @xmath55 . to summarize these results , the @xmath109 dependence of the bulk orbital am is intriguing for the small sois as shown in fig .
[ fig : luttingertsc - am](a ) .
first , the system is trivial due to the @xmath0-wave pairing potential , and @xmath156 is almost zero .
second , the system is still trivial due to the coexistence of the @xmath151-wave pairing potentials , and @xmath156 rapidly increases .
finally , the system turns into the topological phases with @xmath157 , and only @xmath113 is strongly reduced by the non - phs edge states .
thus the @xmath109 dependence of @xmath113 in a hole - doped topological sc is nonmonotonic in contrast to that in an electron - doped topological sc . on the other hand , for @xmath158 ,
the system starts from the coexisting region . in the topological phases ,
the reduction of @xmath113 does not work due to the lifshitz transition in the second lowest particle band and the sharp decrease of the fermi am .
therefore the @xmath109 dependence of @xmath159 is almost monotonic and gradual since @xmath160 is suppressed by the large sois @xmath152 .
so far , we have concentrated on the bulk orbital am altogether . here
we discuss how the phs edge , non - phs edge , and continuum states contribute to the bulk orbital am separately .
the first and second lowest positive - energy eigenstates are continuum states in the trivial phase and become one of the non - phs edge states and the phs edge state , respectively , in the topological phase with @xmath1 .
we can reveal the effects of the phs and non - phs edge states more clearly by decomposing the bulk orbital am with respect to the index of eigenstates around the topological phase transition .
equation is decomposed as @xmath161 .
\label{eq : luttingertsc - lzr2}\end{aligned}\ ] ] remember that we have @xmath162 eigenstates labeled by @xmath163 , @xmath164 are dubbed the phs edge , non - phs edge , and continuum contributions , respectively , in the case of @xmath1 , as shown in fig .
[ fig : luttingertsc - eigen](b ) . in the case of @xmath165 as shown in fig .
[ fig : luttingertsc - eigen](a ) , all of them originate from the continuum states . in the context of chiral @xmath8-wave scs , tsutsumi and
machida previously showed that the bulk orbital am @xmath166 consists of @xmath167 from the phs edge state and @xmath168 from the continuum states and pointed out that these contributions can be experimentally distinguished by controlling the edge roughness @xcite .
thus the phs edge state is expected to play an important role for the bulk orbital am . as shown in ref . and fig .
[ fig : luttingertsc - am](a ) , the bulk orbital am increases in the trivial phase and does not jump at the topological phase transition apart from the reduction by the non - phs edge states , which indicates that the continuum contribution may change its sign from negative to positive at the topological phase transition . first , we carry out a similar analysis for an electron - doped topological sc , which is designed to be a chiral @xmath8-wave sc with use of the rashba soi @xcite , in order to compare with a chiral @xmath8-wave sc and a hole - doped topological sc of our interest .
we set @xmath169 to focus on its chiral @xmath8-wave behavior and change @xmath109 around the phase transition point @xmath170 .
see ref . for the hamiltonian and the other parameters we set .
in fig . [
fig : rashbatsc - radial ] , we show the radial distributions @xmath171 and @xmath172 . for @xmath173 , the phs edge contribution @xmath171 in fig . [
fig : rashbatsc - radial](b ) and the continuum contribution @xmath172 in fig .
[ fig : rashbatsc - radial](c ) are negative and positive , respectively , which is consistent with the previous result obtained by the quasiclassical eilenberger theory @xcite .
however , we can not find any significant difference between before and after the topological phase transition .
the @xmath174 eigenstate , which goes down toward @xmath125 and turns into the phs edge state for @xmath175 , is already localized at the edge even for @xmath176 and has the negative orbital am .
therefore , as @xmath177 from the phs edge state in a chiral @xmath8-wave sc is expected to be fragile against the edge roughness @xcite , not only the orbital am from the phs edge state in the topological phase but also that from one of the continuum states in the trivial phase is expected to be fragile .
now let us move on to a hole - doped topological sc .
figure [ fig : luttingertsc - radial ] shows the radial distributions @xmath164 .
we set @xmath178 to avoid the accidental gap closing at @xmath53 and to focus on its chiral @xmath3-wave behavior . in fig .
[ fig : luttingertsc - radial](b ) , the orbital am of the @xmath179 eigenstate , which turns into the phs edge state for @xmath112 , shows no significant difference between before and after the topological phase transition as in an electron - doped topological sc . on the other hand , the radial distribution from the @xmath174 eigenstate , which goes across @xmath125 and turns into the non - phs edge state for @xmath112 , discontinuously decreases at the topological phase transition .
this decrease is associated with the discontinuous reduction of the bulk orbital am @xmath180 .
note that the radial distribution is not positive because this eigenstate has the negative orbital am even in the trivial phase similar to the @xmath179 eigenstate .
thus the effect of the non - phs edge states can be found not in the radial distribution itself but in its change at the topological phase transition .
in conclusion , we have investigated the bulk orbital am in a two - dimensional topological sc which is composed of a hole - doped semiconductor , a ferromagnet , and an @xmath0-wave sc and is characterized by @xmath1 .
we have obtained @xmath114 by using the green s function formula of the hall viscosity in the reciprocal space and @xmath113 by solving the bdg equation on a circular disk .
our findings are as follows : ( 1 ) before the topological phase transition , @xmath156 starts to increase as in an electron - doped topological sc @xcite .
( 2 ) for @xmath181 , @xmath113 is reduced from @xmath114 almost to zero owing to the non - phs edge states with the large fermi am , which is expected from the recent studies on chiral scs with @xmath13 @xcite . as a consequence ,
@xmath113 is nonmonotonic as a function of the zeeman interaction .
( 3 ) in the opposite case , the fermi am of the non - phs edge states becomes small associated with the lifshitz transition in the normal state , and the reduction of the bulk orbital am does not work .
( 4 ) this effect competes with the other effect to tilt spins and suppress the bulk orbital am . thus a hole - doped topological sc offers an experimentally feasible stage to investigate the bulk orbital am in chiral @xmath3-wave scs and the interesting deviations due to the sois .
finally , let us comment on other models of two - dimensional topological scs previously proposed .
sois are the most important ingredients throughout this paper , which make the system topological and modify the conserved quantity .
the models without use of sois @xcite are out of our scope .
the explicit form of the conserved quantity depends on the details of the system as studied in ref . , but it is given by @xmath182 at least for the models in refs .
. therefore we expect that our findings ( 1 ) , ( 2 ) , and ( 4 ) above are generic for these models with @xmath19 . on the other hand ,
our finding ( 3 ) is quite specific to a hole - doped topological sc which makes use of @xmath27 holes .
we also note that the fermi am and the reduction of the bulk orbital am can be controlled by a confinement potential @xcite .
52ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1143/ptp.57.1836 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.43.596 [ * * , ( ) ] link:\doibase 10.1103/physrevb.21.980 [ * * , ( ) ] link:\doibase 10.1143/ptp.63.1083 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1143/jpsj.67.216 [ * * , ( ) ] link:\doibase 10.1016/s0375 - 9601(98)00438 - 1 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.184511 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.150409 [ * * , ( ) ] link:\doibase 10.1103/physreva.81.053605 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.214509 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.100506 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.134510 [ * * , ( ) ] link:\doibase 10.1103/physrev.123.1911 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.47.331 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1007/bf01141607 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.195301 [ * * , ( ) ] link:\doibase 10.1134/s0021364014230155 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.224519 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.75.657 [ * * , ( ) ] link:\doibase 10.1143/jpsj.81.011009 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.74.235 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.116402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.017004 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1143/jpsj.81.023704 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.100507 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.180503 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.020509 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.184516 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.157003 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.094504 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.184513 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.064507 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.236803 [ * * , ( ) ] link:\doibase 10.1103/physrevb.93.035134 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.160401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.020401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.040502 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.78.809 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.86.187 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.024502 [ * * , ( ) ] link:\doibase 10.1103/physrevb.69.235206 [ * * , ( ) ] link:\doibase 10.1103/physrev.102.1030 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.79.045308 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.085316 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.245309 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.93.174505 [ * * , ( ) ] | we investigate the bulk orbital angular momentum ( am ) in a two - dimensional hole - doped topological superconductor ( sc ) which is composed of a hole - doped semiconductor thin film , a magnetic insulator , and an @xmath0-wave sc and is characterized by the chern number @xmath1 . in the topological phase , @xmath2 is strongly reduced from the intrinsic value by the non - particle - hole - symmetric edge states as in the corresponding chiral @xmath3-wave scs when the spin - orbit interactions ( sois ) are small , while this reduction of @xmath2 does not work for the large sois . here
@xmath4 and @xmath5 are the bulk orbital am and the total number of particles at zero temperature , respectively . as a result ,
@xmath2 is discontinuous or continuous at the topological phase transition depending on the strengths of the sois .
we also discuss the effects of the edge states by calculating the radial distributions of the orbital am . |
anomalous normal - state properties of superconducting cuprates @xcite have stimulated intense theoretical investigations of models of strongly correlated electrons describing the interplay between antiferromagnetic ( afm ) ordering of reference ( undoped ) insulating substances and the itinerant character of charge carriers introduced by doping .
for the understanding of superconductivity the most challenging regime is that of intermediate ( optimum ) doping .
however , even the apparently simplest region of weak doping is not fully understood theoretically .
recently , the attention in experimental and theoretical investigations of cuprates has been given to characterization and understanding of different doping regimes @xcite . in a simple picture
, weak doping should correspond to the regime where properties vary linearly with the concentration of holes , i.e. one can deal with a semiconductor - like model where charge carriers ( holes ) are independent and well defined quasiparticles .
this requires a nonsingular variation of thermodynamic quantities with doping .
however , this scenario has been questioned near the metal
insulator transition based also on numerical solutions for some model systems @xcite , e.g. the hubbard model .
alternative possibilities include phase separation @xcite , quantum critical behavior @xcite or other instabilities at low doping .
still , singular behavior in a planar ( 2d ) system is expected only at @xmath2 , while @xmath3 should lead to a regular variation with doping . among the least understood properties of charge carriers in cuprates and correlated systems in general
are those related to the coupling of their orbital motion to an external magnetic field . evidently anomalous and not understood
is the hall constant in cuprates which reveals unusual temperature and doping dependence @xcite .
another quantity is the diamagnetic ( orbital ) susceptibility @xmath4 , which for noninteracting electrons corresponds to landau diamagnetism @xcite and seems to be connected to the hall response @xcite .
anomalous paramagnetic - like variation with magnetic field has been noticed within the ground state of the @xmath0-@xmath1 model @xcite at low doping .
recent @xmath3 studies of a single hole within the @xmath0-@xmath1 model @xcite confirm the existence of a paramagnetic regime at intermediate @xmath5 , though the systems studied were quite small .
conclusive experimental results on diamagnetic susceptibility are lacking @xcite , since the orbital part appears quite hidden by other contributions , although it could be distinguished via the anisotropy .
the aim of this paper is to study the thermodynamic properties and orbital response of correlated electrons at finite temperature in the low - doping regime .
most numerical studies of the @xmath0-@xmath1 model have so far focused on the ground - state properties @xcite , employing exact diagonalization of small systems , projector monte carlo , and density matrix renormalization group @xcite ( dmrg ) .
recently , the finite - temperature lanczos method ( ftlm ) has been introduced , which allows insight into the statics and dynamics at @xmath3 . in previous applications
certain thermodynamic quantities have also been investigated as a function of doping . in this paper
we focus on the low doping regime , where the method can be compared with the alternative approach , a novel adaptation of the worldline quantum monte carlo ( qmc ) cluster method @xcite which allows for the study of much larger systems at least for temperatures @xmath6 below which the minus - sign problem sets in .
large systems are particularly important for the study of diamagnetic response which appears to be quite sensitive to finite size effects . in both cases ,
new ways of dealing with the magnetic field are introduced .
related qmc methods have been used to study nonmagnetic properties of the @xmath0-@xmath1 model , in an exploratory calculation for doped chains and for ladders with 1 and 2 holes @xcite , in two dimensions at @xmath7 with 1 or 2 holes @xcite , and for chains at finite @xmath1 in a background of no holes @xcite . in the following ,
the planar @xmath0-@xmath1 model as a representative model for strongly correlated electrons and electronic properties of cuprates is studied , @xmath8 where @xmath9 , @xmath10 are fermionic operators , projecting out sites with double occupancy .
to approach the regime of strong correlations close to the real situation in cuprates , @xmath11 is used in most numerical calculations .
we also use @xmath12 .
the paper is organized as follows .
section ii of the paper is devoted to a brief introduction of both numerical techniques employed , qmc and ftlm . in sec .
iii results for several thermodynamic properties in the low - doping regime are presented and discussed .
iv is devoted to the discussion of the orbital susceptibility of the system .
results are obtained independently by the worldline qmc method and the ftlm . wherever possible ,
results of both methods for doped systems are compared and presented relative to the undoped heisenberg afm . for large enough systems we expect to reach a typical behavior in the low doping regime .
the loop cluster algorithm ( lca ) for the world - line qmc has been introduced by one of the present authors @xcite and recently adapted also to the @xmath0-@xmath1 model @xcite .
we briefly describe the worldline representation of the quantum qmc .
the hamiltonian , eq . [ eq : model ] , on a 2d square lattice can be split within the standard trotter - suzuki decomposition @xcite into four parts @xmath13 consisting of mutually commuting terms . this is equivalent to the well known checkerboard decomposition of hamiltonians in 1d .
the partition function is @xmath14^m= \nonumber\\ & = & \tr[e^{-\tbeta h_1}e^{-\tbeta h_2}e^{-\tbeta h_3}e^{-\tbeta h_4}]^m+ o(\tbeta^2)\approx \nonumber\\ & \approx&\sum_{\phi_1\ldots \phi_{4m}}\bra{\phi_{4m}}e^{-\tbeta h_1}\ket{\phi_1 } \bra{\phi_1}e^{-\tbeta h_2}\ket{\phi_2}\cdots \nonumber\\ & \cdots&\bra{\phi_{4m-1}}e^{-\tbeta h_4}\ket{\phi_{4 m } } , \label{eq : trotter}\end{aligned}\ ] ] where @xmath15 and @xmath16 .
the summation is taken over the complete orthonormal set of states @xmath17 . within each imaginary time step @xmath18
the time evolution operator is applied .
since the hamiltonian is total spin conserving , we can track time evolution of a particular spin along its so called _ worldline _ ( wl ) . because of the cyclic property of the trace , the wls are periodic in the imaginary time interval @xmath19 $ ] .
the time evolution operator acts only on @xmath20 plaquettes and the weight of the configuration @xmath21 factorizes into a product of plaquette weights .
the partition function @xmath22 is formally that of a ( @xmath23)-dimensional classical system . the thermal average of an observable @xmath24 can be obtained by @xmath25 such thermal expectation values are calculated by means of monte carlo ( mc ) importance sampling , where a sequence of configurations @xmath26 ( markov chain ) is constructed , which obeys detailed balance and reproduces the correct boltzmann distribution @xmath27 .
thermal expectation values now become simple averages @xmath28 in practice , monte carlo runs are finite , @xmath29 , leading to statistical errors which can be calculated from the standard deviation of partial data sets @xcite . in standard local algorithms
an update from one configuration @xmath30 to another @xmath31 in the markov chain represents a small local change of the wls .
therefore , consecutive configurations are highly correlated , which drastically increases the necessary number of monte carlo steps .
such difficulties are overcome in the lca @xcite which introduces global ( nonlocal ) stochastic updates that effectively reduce the correlations . in the lca formulation
also the continuous time limit @xmath32 can be taken @xcite avoiding the second order systematic error of eq .
[ eq : trotter ] . for certain observables improved estimators can be easily constructed allowing a potential reduction of statistical errors .
for more details we refer to the introductory paper @xcite .
the lca has recently been adapted to the @xmath0-@xmath1 model .
the update procedure is split into three substeps , allowing the application of the standard lca for the @xmath33 antiferromagnetic heisenberg model or for free fermions in all three cases . within each substep , only updates between two of the possible three states ( @xmath34 , @xmath35 , and hole @xmath36 ) are performed . for the weights of particular plaquettes and other technical details we refer to @xcite . in case of negative weights @xmath37 , their magnitude @xmath38 is taken for construction of the mc procedure , since the negative @xmath21 can not be taken as a probability .
( [ eq : avg ] ) becomes @xmath39 where @xmath40 denotes the expectation value with respect to the absolute value of the weight . in systems with such a `` sign problem '' , the average sign @xmath41 often becomes exponentially small with increasing system size and decreasing temperature @xmath5 , leading to a blow up of statistical errors @xcite .
let us briefly comment on the origin of negative signs in the wl formulation of the @xmath0-@xmath1 model . in the system with no doped holes the only source of negative weights are plaquettes , where two opposite spins exchange their positions representing a spin flip .
because of the periodicity of wls in time direction and the absence of holes spin flips always occur in even numbers , producing no net negative sign . in the pure heisenberg model , this sign
can also be transformed away by rotation of spins on one sublattice , resulting in all - positive plaquette weights @xcite .
for one hole doped into the afm one would naively not expect a sign problem , e.g. , in this case there is no sign in the exact diagonalization approach .
examining the particle wls surrounding the hole wl one finds , however , that an exchange of two fermions can occur when @xmath42 and @xmath43 , producing an odd number of spin flips , i.e. a negative sign , as can be seen schematically in a small @xmath44 system , @xmath45 where a loop motion of the hole @xmath46 around the system and a consecutive spin flip @xmath47 reproduce the original configuration with two fermions @xmath34 and @xmath48 interchanged .
measuring the sign here reduces to spin flip counting .
for higher concentration of holes a more general expression for the sign of the configuration can be obtained .
it links fermion wl ( @xmath49 ) and hole wl permutation ( @xmath50 ) @xmath51 so that for low doping it is preferable to measure @xmath50 rather than @xmath49 .
the sign problem also complicates the use of the improved estimators since for every observable a separate algorithm must be devised . to follow the emergence of the sign problem as well as the development of diamagnetic properties it is convenient to generalize the isotropic spin interaction term of the model eq .
( [ eq : model ] ) to an anisotropic one with general anisotropy parameter @xmath52 , @xmath53.\ ] ] this modifies the pure spin substep ( @xmath34 , @xmath35 ) of the @xmath0-@xmath1 lca so that otherwise independent loops are `` frozen '' together @xcite into clusters and updated stohasticaly . the results for @xmath54 as a function of inverse temperature @xmath55 are presented in fig .
[ fig : sign ] .
note that for a single hole in the system , @xmath56 , the relevant temperature scale in the anisotropic case is @xmath57 , i.e. there is no sign problem for @xmath58 .
at @xmath59 the sign starts to deteriorate rapidly , as can be seen in fig .
[ fig : sign ] , preventing the investigation of low temperature properties .
as expected @xmath54 decreases by adding additional holes @xmath60 . for this reason , within the doped @xmath0-@xmath1 model only chains and coupled chains @xcite have been investigated by lca so far , and in two dimensions the limit @xmath61 with @xmath62 @xcite .
recently , though , a way around the sign problem has been proposed by calculating fermion propagators for a background of no holes @xcite . for
a fixed number of holes , the average sign will converge as the system size increases .
this convergence could be taken as a another criterion that the limit of a dilute system has been reached . in the analysis of the @xmath0-@xmath1 model
the exact diagonalization of small systems using the lanczos algorithm has been extensively employed @xcite , predominantly in the investigation of the static and dynamic properties of the ground state .
more recently a ftlm combining the lanczos procedure and random sampling was introduced @xcite , allowing the calculation of @xmath3 static and dynamic properties of correlated systems .
the application is particularly simple for an arbitrary function of conserved quantities , e.g. @xmath63 , @xmath64 where @xmath65 , @xmath66 are ( approximate ) eigenfunctions and energies , respectively , obtained by diagonalization within the reduced orthonormal set , generated from the initial functions @xmath67 in @xmath68 lanczos steps .
@xmath69 is the dimension of the complete basis .
@xmath70 initial functions @xmath67 are chosen at random but with good quantum number @xmath71 .
usually it is enough to choose @xmath72 . for a more detailed discussion of the method and results
we refer to @xcite .
it is expected that @xmath3 reduces the finite - size effects of the measured quantities .
it is however important to realize that for a particular system , finite - size effects start to be pronounced at @xmath73 where , e.g. , some characteristic length - scale becomes larger than the system size . in our case of low doping , @xmath74 and @xmath75 ,
we find @xmath76 .
all our results are presented for @xmath77 where @xmath78 . in the present study @xmath79 so that at least 30 states are sampled in the thermal averages @xcite .
it should be stressed that the ftlm gives also the correct ground state within the chosen small system .
thermodynamic properties @xmath80 depend on the hole concentration @xmath81 . for the weak doping limit
, one would expect a linear dependence for most quantities . for a finite system size ,
the relevant parameter is thus the number of holes @xmath82 doped into the afm . in the low - doping regime
it makes sense to represent the results as a difference @xmath83 to distinguish the change in a particular quantity with doping , this notation is used in the following . if , e.g. , @xmath84 behaves quantitatively as @xmath85 one can conclude that the quantity changes linearly with the number of added holes @xmath82 , i.e.the holes behave as independent entities , and the system sizes are large enough so that the low - doping regime has indeed been reached .
such behavior is however not the only possibility at low doping , since one can expect , e.g. , even - odd effects in the case of pairing of holes .
the internal energy , defined by @xmath86 is calculated within ftlm as @xmath87 in eq .
( [ eq : ftlm_avg ] ) and in qmc as an expectation value of the corresponding operator , @xmath88 . results of both methods are presented in fig .
[ fig : eng ] . from fig .
[ fig : eng ] we first conclude that for @xmath89 the results at @xmath90 ( which overlap also with the results at @xmath91 ) obtained via the ftlm are essentially equivalent with qmc results for much larger lattices , at least for @xmath92 reached by the qmc method .
we note also a close agreement between qmc @xmath93 and @xmath89 , confirming the assumption of the low - doping regime and holes as independent quasiparticles in the @xmath5 window presented . in ftlm results , on the other hand , the difference between @xmath93 and @xmath89 is already visible , since @xmath94 here means already an appreciable doping @xmath95 . for @xmath2 the difference @xmath96 equals to the binding energy @xcite and in the continuum corresponds to the second derivative of the ground state energy with respect to the doping . in the chosen parameter regime ( @xmath97 )
the binding energy is negative and thus pointing to the attractive interaction between the holes . with the increase of the temperature
the bound state disintegrates and the difference @xmath96 approaches zero . in the case of a small system
the difference can even become positive but vanishes with increasing system size .
it is also evident from fig .
[ fig : eng ] that @xmath98 is not a monotonous function .
the ground state of a single hole introduced into the afm is quite well understood via analytical approaches @xcite and numerical calculations @xcite . for @xmath11 ,
the zero temperature result @xmath99 can be explained well by the interplay between the gain of the kinetic energy represented by the hopping term @xmath100 and the loss of local afm correlation energy around the hole .
@xmath98 has not been considered so far .
an interpretation of its behavior can be given as follows .
introducing a single hole into an afm destroys the local afm spin order and thus increases the exchange energy .
the increase is however expected to disappear at @xmath101 where the spin system becomes disordered . on the other hand
the ground - state kinetic energy in a disordered spin system is quite similar to the one in an afm , hence the decrease of the internal energy @xmath102 for @xmath101 .
this remains valid for @xmath103 where also higher hopping - related states become populated and finally @xmath104 , explaining turn back up for @xmath105 .
the specific heat defined by @xmath106\ ] ] is obtained as @xmath107 $ ] within the ftlm and the qmc method .
the results are presented in fig .
[ fig : heat ] .
the main effect of introducing holes into the afm insulator on @xmath108 is to decrease the peak at @xmath109 .
this appears in @xmath110 as a pronounced dip which slightly weakens and shifts its energy scale @xmath1 to lower values with doping , as can be seen from the line - shape in fig .
[ fig : heat ] . the entropy is @xmath111 we reconstruct it from the specific heat @xmath112 by numerical integration from high temperatures @xmath113 @xmath114 the high - temperature integration constants are chosen so that @xmath115 .
in contrast to @xmath102 and @xmath110 discussed previously , @xmath116 is not linear in @xmath82 in the low doping limit . in analogy with low - concentration systems , like the dilute classical gas , entropy
is expected to scale as @xmath117 , i.e. @xmath118 , so the change still depends explicitly on the system size .
this is also realized in fig.[fig : ent ] , where @xmath119 still varies with @xmath120 , yet curves for different @xmath120 appear parallel down to the lowest reachable temperatures .
it is particularly remarkable how large the entropy increase @xmath121 is even at the lowest @xmath122 .
this is indeed consistent with @xmath123 measured in cuprates @xcite . while this could be attributed partly to the logarithmic dependence on @xmath124 , at the same it is evident that the behavior is much closer to a system of classical particles than to a degenerate electron gas .
the uniform spin susceptibility can be evaluated as a thermodynamic quantity from @xmath125 where @xmath126 is the conserved total spin . in the ftlm
then eq .
( [ eq : ftlm_avg ] ) can be applied , while within the qmc method @xmath127 is related to the number of spin up and down wls .
it is instructive to present results both for @xmath128 with respect to the undoped afm , fig .
[ fig : schi ] , as well as for the effective curie constant ( difference of the square moment ) per hole @xmath129 in fig .
[ fig : moment ] . the results in figs .
[ fig : schi ] , [ fig : moment ] are easy to interpret for high @xmath130 .
each hole introduced into the system reduces the effective curie constant by one spin , i.e. , @xmath131 . on the other hand , at low @xmath132
the situation is reversed since @xmath133 .
this increase can be attributed to the relaxation of the afm order by the hole doping .
note that in an afm , @xmath127 achieves a maximum at @xmath109 while below that temperature it is reduced due to the longer range afm order .
it is interesting to note that at the lowest reachable temperature @xmath134 each hole effectively adds just one spin , i.e.@xmath135 .
in order to investigate the orbital response of the system , a homogeneous magnetic field @xmath136 perpendicular to the plane has to be introduced .
when we discuss the orbital magnetization and susceptibility , @xmath136 enters only in the kinetic term of eq .
( [ eq : model ] ) , via the peierls construction @xmath137 where the phases are given within landau gauge as @xmath138 with @xmath139 .
the relevant parameter for the strength of @xmath136 is the dimensionless flux per plaquette @xmath140 , where @xmath141 is the lattice spacing and @xmath142 is the unit quantum flux .
the dc orbital susceptibility of the system in the external magnetic field is @xmath143 , \label{eq : chidef}\end{aligned}\ ] ] where @xmath144 .
so far @xmath4 has been investigated only for a single hole @xmath56 by high - temperature expansion at @xmath145 , and by the ftlm for @xmath146 @xcite . in the latter study
it was realized that results are quite sensitive to finite - size effects , so it is desirable to get corresponding results also via the qmc method , where much larger lattices can be studied .
let us first derive the expression for the orbital susceptibility within the qmc method . as seen from eq .
( [ eq : peierls ] ) , the magnetic field affects only the hopping of the electrons . in the wl representation
this concerns matrix elements @xmath147 in eq .
( [ eq : trotter ] ) . for the plaquette representing the hopping between sites @xmath148 and @xmath149
the matrix elements become @xmath150 written in the @xmath151 , @xmath152 base .
the imaginary time propagator in the same base is @xmath153 thus the plaquette weights along the hole wl obtain an additional phase factor @xmath154 .
the weight of the whole configuration is a product , eq .
( [ eq : z ] ) , @xmath155 where the phases sum up @xmath156= \exp\big[i\oint\theta(\vc{r})\text{d}\vc{r}\big]= \nonumber\\ = \exp\big[i\frac{e}{\hbar}\oint\vc{a}(\vc{r})\text{d}\vc{r } \big]=e^{i\alpha{\cal s}}. \label{eq : phases}\end{aligned}\ ] ] the integral runs along the hole wl . here
@xmath157 is defined as the oriented area of the hole wl projected onto the plane in units of the lattice plaquette area @xmath158 . for more holes
, @xmath157 generalizes similarly to the sum of all hole wl areas .
now we can write the partition function in the magnetic field as @xmath159 for a given configuration @xmath160 there always exist @xmath161 ( imaginary time inversion ) with the same weight but @xmath162 , therefore the exponential in eq .
( [ eq : zz ] ) can be replaced by a @xmath163 function .
for @xmath164 we have @xmath165 and obtain the zero - field susceptibility from eq .
( [ eq : chidef ] ) , @xmath166 @xmath4 can be thus measured without the presence of a magnetic field .
this is just another consequence of the more general fluctuation dissipation theorem . even though @xmath167 is strictly positive , the thermal average in eq .
( [ eq : chi ] ) can become negative because of correlations between the sign of the weight and the area @xmath157 of the hole wl . from eq .
( [ eq : abs_avg ] ) we can deduce that @xmath168 when the configurations with negative sign tend to have larger @xmath169 than configurations with positive ones . the hole wl can obtain nonzero spatial winding number due to the periodic boundary conditions and small system size .
e.g. , the wl can run along the imaginary time , cross the system boundary , and complete time periodicity reconnecting with its spatially periodic image . in that case the area @xmath157 as defined by eq .
( [ eq : phases ] ) has no physical meaning .
therefore we restrict our simulation so that only the zero spatial winding loop updates are generated ( a discussion on fixing the winding numbers can be found in @xcite ) .
the hole is allowed to cross the system boundary as long as it does not increase the winding number .
the effect of the restriction is analogous to the movement of the hole doped into an infinite periodic spin background with a unit cell equal to the size of the system .
the results for thermodynamic quantities presented in the previous section agree within the errorbars with the unrestricted case . this restriction is weaker than closed boundary conditions , resulting in smaller finite size effects .
the introduction of finite @xmath170 into the model , eq . ( [ eq : peierls ] ) , reduces the translational symmetry and thus for a given system size increases the required minimal base set used in ftlm . in the present study a few mobile holes on a system of tilted squares with @xmath120 up
to 20 sites and periodic boundary conditions are considered .
it is nontrivial to incorporate phases @xmath171 corresponding to a homogeneous @xmath136 , being at the same time compatible with periodic boundary conditions @xcite .
this is possible only for quantized magnetic fields @xmath172 , where @xmath173 .
@xmath4 from eq .
( [ eq : chi ] ) can be calculated in ftlm only by taking a numerical derivative of the free energy @xmath174 , with @xmath175 from eq .
( [ eq : ftlm_z ] ) .
finite systems provide @xmath176 only for discrete values of @xmath177 .
@xmath4 can be obtained by fitting the @xmath176 dependence to the parabolic form @xmath178 in two points @xmath179 and @xmath180 .
the corresponding results for the susceptibility are denoted by @xmath181 .
in small systems ( @xmath182 ) the introduction of @xmath170 can lift some zero - field degeneracies .
the @xmath183 values are therefore systematically affected by larger finite - size effects . for the system with @xmath90 both values @xmath183 and @xmath184 agree quite reasonably with the qmc data .
let us first comment on the validity of results for @xmath4 . since @xmath4 deals with an orbital current represented by loop motion of charge carriers ( holes ) , it is much more sensitive to finite - size effects @xcite than most other correlation functions .
this was also the main motivation to employ the qmc method , where much larger systems can be reached . in fig .
[ fig : scale ] , finite - size scaling is performed for the qmc data for @xmath185 for the case of @xmath56 .
we can see that the different @xmath5 points do not cross upon changing the system size @xmath120 .
thus , at least qualitatively , results do not depend on the system size .
therefore , the sizes of choice for the qmc systems will be @xmath186 and @xmath187 . in fig .
[ fig : dchi ] , @xmath4 obtained via both methods is presented . for @xmath188 ,
the response is diamagnetic and proportional to @xmath189 as well as essentially @xmath1-independent @xcite .
the most striking effect is that the orbital response below some temperature @xmath190 turns from diamagnetic to paramagnetic , consistent with the preliminary results obtained via the ftlm @xcite . in order to locate the origin of this phenomenon ,
results for different @xmath1 and anisotropies @xmath52 are shown in fig .
[ fig : gammaj ] .
it appears that @xmath190 scales with @xmath191 , i.e. at @xmath145 the response is clearly diamagnetic at all @xmath5 , and for @xmath192 no crossing is observed with either method . at lower temperatures
@xmath193 , the diamagnetic behavior is expected to be restored .
this follows from the argument that at @xmath194 a hole in an afm should behave as a quasiparticle with a finite effective mass , exhibiting a cyclotron motion in @xmath195 .
the latter behavior should lead to @xmath196 @xcite .
numerically it is easiest to test this conjecture for a single hole and @xmath58 ( also true for @xmath145 ) .
namely , the qmc has no sign problem at @xmath58 , so that error bars are only due to the finite mc sampling .
results in fig .
[ fig : gammaj ] are quite interesting even for @xmath58 . at @xmath145 a monotonous increase of @xmath197
is observed , diverging as @xmath198 @xcite .
it can be explained as a gradual transition from a hole in a random spin background to a well defined quasiparticle , i.e. the ferromagnetic polaron , at @xmath198 .
the situation is more complicated for @xmath146 .
it is plausible that the difference to the @xmath145 case shows up at @xmath132 , where the afm short - range correlations appear .
relative to @xmath145 , a spin ordered state blocks the loop motion of holes , necessary for finite diamagnetic @xmath4 .
the effect is thus first a decrease of @xmath199 with decreasing @xmath5 , as seen in fig .
[ fig : gammaj ]
. turnover to the diverging diamagnetic @xmath4 should happen only when the coherent quasiparticle is formed at @xmath200 .
@xmath201 should scale with the inverse of the quasiparticle effective mass @xmath202 .
it is known that @xmath203 can become very large for the extreme @xmath58 case , in particular for larger @xmath1 .
this explains why we can not reach the coherent regime for @xmath204 even at @xmath205 ( fig .
[ fig : dchi ] ) , while the downturn is indeed observed for @xmath206 , @xmath207 , @xmath58 ( fig .
[ fig : gammaj ] ) . at @xmath208 ,
the results are qualitatively different .
the most pronounced effect is the change into a paramagnetic @xmath4 for @xmath209 .
the width of this @xmath5 window is quite large .
in fact within the ftlm and qmc data we are unable to locate the reentrance temperature @xmath201 into the diamagnetic response , although the latter is expected @xcite .
an argument for the low value of @xmath201 can be given in terms of a very shallow energy minimum which defines the quasiparticle dispersion near the ground - state of a hole in an afm within the @xmath0-@xmath1 model @xcite , hence the quasiparticle looses its character already at very low excitation energies .
still the paramagnetic response in the window @xmath210 remains to be explained .
let us finally discuss also results for @xmath4 for finite doping @xmath211 , as presented in fig .
[ fig : chich ] .
the easiest regime to interpret is that of a nearly empty band , i.e. @xmath212 , where @xmath4 is diamagnetic and nearly independent of @xmath5 . in this regime
the electron system is dilute and strong correlations are unimportant , hence landau diamagnetism is expected . at moderate temperatures @xmath101 and for an intermediate - doping regime , @xmath213 , @xmath4 is dominated by a paramagnetic response with a peak at approximately @xmath214 . as consistent with results at low doping , there is a weak diamagnetism at @xmath215 and @xmath216 , while the paramagnetic regime extends to @xmath217 for @xmath209 . for low temperatures @xmath218 quite pronounced oscillations in @xmath219 appear and
can be partly attributed to finite - system effects .
certain aspects of the above results for @xmath219 can be understood using the hte .
one is the asymmetry between @xmath220 and @xmath221 at higher @xmath130 . in the lowest order of the hte , only hopping of the electrons around a basic plaquette loop
has to be considered .
the signal on the @xmath222 side of @xmath219 is thus reduced by a factor of @xmath223 against the @xmath224 side since in the first case only the plaquettes with ferromagnetically aligned spins contribute .
let us first compare both numerical methods used in the analysis of the @xmath0-@xmath1 model at low doping .
the qmc method allows the studies of larger systems and the loop algorithm solves some serious drawbacks of the mc methods .
it is indeed very efficient in cases where there is no sign problem , e.g. the anisotropic model @xmath58 at @xmath56 . within the wl approach it is also very easy to formulate and measure certain responses like the orbital susceptibility @xmath4 .
still the method suffers from a sign problem ( for @xmath225 ) even for a single hole @xmath56 in an afm ( though not in a background of no holes @xcite ) .
results are thus in practice limited to @xmath226 for the isotropic case @xmath227 . on the other hand
the ftlm has no minus - sign problem but rather limitations due to small systems which can be studied .
these are even more pronounced in cases with @xmath170 where the translational symmetry is lost .
it is an interesting observation that within the ftlm , the limiting temperature @xmath228 for the @xmath0-@xmath1 model is in most cases quite close to the lowest @xmath5 reached by the qmc method .
we have presented several results for thermodynamic quantities , i.e.energy @xmath229 , specific heat @xmath108 , entropy @xmath230 and spin susceptibility @xmath127 , as a function of @xmath5 at low doping . at @xmath231
all results are consistent with the picture where holes introduced into the afm behave as independent ( nondegenerate ) particles .
the perturbation introduced into the afm by holes is quite large even at lowest @xmath132 , in particular visible from @xmath110 and @xmath123 , consistent with experiments in cuprates @xcite .
results for the orbital susceptibility @xmath4 , now obtained also for much larger systems using the qmc method , confirm the preliminary ftlm results @xcite , indicating an anomalous paramagnetic response at low doping in an intermediate window of temperatures @xmath232 ( for the isotropic model @xmath227 ) .
in fact within our numerical studies it is hard to reach the lower end of this window , meaning that @xmath233 .
still , the reentrance into a diamagnetic behavior is expected from theoretical arguments on the existence of a well defined quasiparticle at @xmath198 , as well as from more reliable qmc results for the @xmath58 case @xcite .
the paramagnetic response at intermediate @xmath5 can be viewed also as an extension of a more pronounced @xmath234 regime observed at finite doping @xmath235 at all @xmath5 .
the explanation can thus go in the direction proposed by laughlin @xcite , that at low doping @xmath220 we are dealing with quasiparticles ( with a diamagnetic response ) , being a bound composite of charge ( holon ) and spin ( spinon ) elementary excitations .
the binding appears however to be quite weak and thus easily destroyed by finite @xmath5 or @xmath124 , enabling the independent response of constituents , which apparently is paramagnetic . for a review ,
see m. imada , a. fujimori , and y. tokura , rev .
* 70 * , 1039 ( 1998 ) .
b. batlogg _ et al .
_ , physica c * 235 - 240 * , 130 ( 1994 ) .
emery and s.a .
kivelson , physica c * 209 * , 597 ( 1993 ) .
a. sokol and d. pines , phys .
lett . * 71 * , 2813 ( 1993 ) . for
a review see e.g. n.p .
ong , in _ physical properties of high temperature superconductors _
, ed . by d.m .
ginsberg ( world scientific , singapore , 1990 ) , vol
l. landau , z. phys . * 64 * , 629 ( 1930 ) .
rojo , g. kotliar , and g.s .
canright , phys . rev .
b * 47 * , 9140 ( 1993 ) .
p. bran , phys .
b * 54 * , 1391 ( 1996 ) ; p. bran , d. poilblanc , and r.b .
laughlin , nucl .
b * b473 * , 707 ( 1996 ) .
d. veberi , p. prelovek , and i. sega , phys .
b * 57 * , 6413 ( 1998 ) .
walstedt , r.f .
bell , l.f .
schneemeyer , j.v .
waszczak , and g.p .
espinosa , phys .
b * 45 * , 8074 ( 1992 ) ; m. miljak , v. zlati ' c , i. kos , j.d .
thompson , p.c .
canfield , and z. fisk , sol .
commun . * 85 * , 519 ( 1993 ) . for a review ,
see e. dagotto , rev .
* 66 * , 763 ( 1994 ) .
white and d.j .
scalapino , cond - mat/9907375 , and references therein .
evertz , g. lana , and m. marcu , phys .
* 70 * , 875 ( 1993 ) ; for a review , see h.g .
evertz , cond - mat/9707221v2 ( january 2000 ) , to be published in _ numerical methods for lattice quantum many - body problems _ , ed .
scalapino ( perseus books , frontiers in physics ) .
m. brunner , f.f .
assaad , and a. muramatsu , cond - mat/9904150 .
b. ammon , h.g .
evertz , n. kawashima , m. troyer , and b. frischmuth , phys .
b * 58 * , 4304 ( 1998 ) .
m. brunner and a. muramatsu , phys .
b * 58 * , r10100 ( 1998 ) .
trotter , proc . am .
* 10 * , 545 ( 1959 ) ; m. suzuki , prog .
phys . * 56 * , 1454 ( 1976 ) . , edited by m. suzuki ( world scientific , singapore , 1994 ) , p. 65 .
beard and u .- j .
wiese , phys .
lett . * 77 * , 5130 ( 1996 ) .
n. kawashima , j.e .
gubernatis , phys .
b * 50 * , 136 ( 1994 ) .
j. jakli and p. prelovek , phys .
b * 49 * , 5065 ( 1994 ) . for a review , see j. jakli and p. prelovek , adv .
phys . * 49 * , 1 ( 2000 ) , cond - mat/9803331 .
s. schmitt - rink , c.m .
varma , and a.e .
ruckenstein , phys .
lett . * 60 * , 2793 ( 1988 ) .
loram , k.a .
mirza , j.r .
cooper , and w.y .
liang , phys .
lett . * 71 * , 1740 ( 1993 ) .
p. henelius , s.m .
girvin , and a.w .
sandvik , phys .
b * 57 * , 13382 ( 1998 ) .
e. fradkin , _ field theories of condensed matter systems _ , ( addison - wesley , redwood city , 1991 ) , frontiers in physics vol . 82 , p. 252 .
laughlin , j. low temp .
phys . * 99 * , 443 ( 1995 ) . | finite - temperature properties of weakly doped antiferromagnets as modeled by the two - dimensional @xmath0-@xmath1 model and relevant to underdoped cuprates are investigated by numerical studies of small model systems at low doping .
two numerical methods are used : the worldline quantum monte carlo method with a loop cluster algorithm and the finite - temperature lanczos method , yielding consistent results .
thermodynamic quantities : specific heat , entropy and spin susceptibility reveal a sizeable perturbation induced by holes introduced into a magnetic insulator , as well as a pronounced temperature dependence .
the diamagnetic susceptibility introduced by coupling of the magnetic field to the orbital current reveals an anomalous temperature dependence , changing character from diamagnetic to paramagnetic at intermediate temperatures . |
regular expressions ( res ) , because of their succinctness and clear syntax , are the common choice to represent regular languages .
equivalent deterministic finite automata ( dfa ) would be the preferred choice for pattern matching or word recognition as these problems can be solved efficiently by dfas .
however , minimal dfascan be exponentially bigger than res .
nondeterministic finite automata ( nfa ) obtained from rescan have the number of states linear with respect to ( w.r.t ) the size of the res . because nfaminimization is a pspace - complete problem other methods
must be used in order to obtain small nfas usable for practical purposes .
conversion methods from resto equivalent nfascan produce nfaswithout or with transitions labelled with the empty word ( @xmath1-nfa ) .
here we consider several constructions of small @xmath0-free nfasthat were recently developed or improved @xcite , and that are related with the one of glushkov and mcnaughton - yamada @xcite .
the nfasize can be reduced by merging equivalent states @xcite .
another solution is to simplify the resbefore the conversion @xcite .
gruber and gulan @xcite showed that resin reduced star normal form ( ) achieve some conversion lower bounds .
our experimental results corroborate that resmust be converted to reduced . in this paper
we present the implementation within the * fado*system @xcite of several algorithms for constructing small @xmath0-free nfasfrom res , and a comparison of regular expression measures and nfasizes based on experimental results obtained from uniform random generated res .
we consider nonredundant resand resin reduced in particular .
let @xmath2 be an _ alphabet _ ( set of _ letters _ ) .
@xmath3 over @xmath2 is any finite sequence of letters .
the _ empty word _ is denoted by @xmath1 .
let @xmath4 be the set of all words over @xmath2 .
a _ language _ over @xmath2 is a subset of @xmath4 .
the set @xmath5 of _ regular expressions _ ( re ) over @xmath2 is defined by : @xmath6 where the operator @xmath7 ( concatenation ) is often omitted .
the language @xmath8 associated to @xmath9 is inductively defined as follows : @xmath10 , @xmath11 , @xmath12 for @xmath13 , @xmath14 , @xmath15 , and @xmath16 .
two regular expressions @xmath17 and @xmath18 are _ equivalent _ if @xmath19 , and we write @xmath20 .
the algebraic structure @xmath21 constitutes an idempotent semiring , and with the unary operator @xmath22 , a kleene algebra .
there are several ways to measure the size of a regular expression .
the _ size _ ( or _ ordinary length _ ) @xmath23 of @xmath24 is the number of symbols in @xmath17 , including parentheses ( but not the operator @xmath7 ) ; the _ alphabetic size _
@xmath25 ( or @xmath26 ) is its number of letters ( multiplicities included ) ; and the _ reverse polish notation size _
@xmath27 is the number of nodes in its syntactic tree .
the _ alphabetic size _ is considered in the literature @xcite the most useful measure , and will be the one we consider here for several remeasure comparisons .
moreover all these measures are identical up a constant factor if the regular expression is reduced ( * ? ? ?
3 ) . let @xmath28 be @xmath1 if @xmath29 , and @xmath30 otherwise .
a regular expression @xmath17 is _ reduced _ if it is normalised w.r.t the following equivalences ( rules ) : @xmath31 a recan be transformed into an equivalent reduced rein linear time .
a _ nondeterministic automaton _ ( nfa ) @xmath32 is a quintuple @xmath33 , where @xmath34 is a finite set of states , @xmath2 is the alphabet , @xmath35 the transition relation , @xmath36 the initial state , and @xmath37 the set of final states .
the _ size _ of an nfais @xmath38 . for @xmath39 and @xmath40 , we denote by @xmath41 , and we can extend this notation to @xmath42 , and to @xmath43 .
the _ language _ accepted by @xmath32 is @xmath44 .
two nfasare _ equivalent _ , if they accept the same language .
if two nfas@xmath45 and @xmath46 are isomorphic , and we write @xmath47 .
an nfais _ deterministic _ ( dfa ) if for each pair @xmath48 there exists at most one @xmath49 such that @xmath50 .
a dfais _ minimal _ if there is no equivalent dfawith fewer states .
minimal dfaare unique up to isomorphism . given an equivalence relation @xmath51 on @xmath34 , for @xmath39 let @xmath52_{e}$ ]
be the class of @xmath53 w.r.t @xmath51 , and for @xmath54 let @xmath55_{e}\mid q\in t\}$ ] .
the equivalence relation @xmath51 is _ right invariant _ w.r.t an nfa@xmath32 if @xmath56 and for any @xmath57 , @xmath40 if @xmath58 , then @xmath59 .
the quotient automaton @xmath60_e , f/_e)$ ] , where @xmath61_e,{\sigma},[q]_e ) \mid ( p,{\sigma},q ) \in \delta\}$ ] , satisfies @xmath62 . given two equivalence relations over a set @xmath34 , @xmath63 and @xmath64
, we say that @xmath63 is _ finer _ than @xmath64 ( and @xmath64 _ coarser _ than @xmath63 ) if and only if @xmath65 .
we consider three methods for constructing small nfas@xmath32 from a regular expression @xmath17 such that @xmath66 , i.e. , they are _
the position automaton construction was independently proposed by glushkov , and mcnaughton and yamada @xcite .
let @xmath67 for @xmath68 , and let @xmath69 .
we consider the expression @xmath70 obtained by marking each letter @xmath71 with its position @xmath72 in @xmath17 , @xmath73 . the same notation is used to remove the markings , i.e. , @xmath74 . for
@xmath9 and @xmath75 , let @xmath76 , @xmath77 , and @xmath78 .
let @xmath79 .
the _ position automaton _ for @xmath9 is @xmath80 , with @xmath81 and @xmath82 if @xmath83 , and @xmath84 , otherwise .
we note that the number of states of @xmath85 is exactly @xmath86 .
other interesting property is that @xmath87 is _ homogeneous _ , i.e.
, all transitions arriving at a given state are labelled by the same letter .
brggemann - klein @xcite showed that the construction of @xmath87 can be obtained in @xmath88 ( @xmath89 ) if the regular expression @xmath17 is in the so called _ star normal form _ ( ) , i.e. , if for each subexpression @xmath90 of @xmath17 , @xmath91 and @xmath92 .
for every @xmath68 there is an equivalent rein star normal form @xmath93 that can be computed in linear time and such that @xmath94 .
ilie and yu @xcite introduced the construction of the follow automaton from a re .
their initial algorithm begins by converting @xmath9 into an equivalent @xmath1-nfafrom which the follow automaton @xmath95 is obtained . for efficiency reasons we implemented that method in the * fado*library . the _ follow automaton _ is a quotient of the position automaton w.r.t the right - invariant equivalence given by the _ follow relation _ @xmath96 that is defined by : lcr @xmath97 & if & [ cols= " < " , ] @xmath98 .
let @xmath99 be a set of regular expressions .
then @xmath100 if @xmath101 and @xmath102 . for @xmath9 and @xmath40 , the set @xmath103 of _ partial derivatives _ of @xmath17 w.r.t .
@xmath71 is defined inductively as follows : @xmath104 this definition can be extended to sets of regular expressions , words , and languages . given @xmath9 and @xmath40 , @xmath105 for @xmath106 , @xmath107 , @xmath108 for @xmath42 , and @xmath109 for @xmath110 .
the _ set of partial derivatives _ of @xmath17
is denoted by @xmath111 given a regular expression @xmath17 , the partial derivative automaton @xmath112 , introduced by mirkin and antimirov @xcite , is defined by @xmath113 where @xmath114 , for all @xmath115 and @xmath40 .
@xmath116 champarnaud and ziadi @xcite showed that the partial derivative automaton is also a quotient of the position automaton .
et al . _
@xcite proved that for rereduced and in star normal form the size of its partial derivative automaton @xmath117 is always smaller than the one of its follow automaton @xmath118 .
the automata here presented @xmath87 , @xmath118 and @xmath117 can in worst - case be constructed in time and space @xmath119 , and have , in worst - case , size @xmath119 , where @xmath120 is the size of the re .
recently , nicaud @xcite showed that on the average - case the size of the @xmath87 automata is linear .
the best worst case construction of @xmath0-free nfasfrom reis the one presented by hromkovic _
et al . _
@xcite that can be constructed and have size @xmath121
. however this construction is not considered here .
it is possible to obtain in time @xmath122 a ( unique ) minimal dfaequivalent to a given one .
however nfastate minimization is pspace - complete and , in general , minimal nfasare not unique .
considering the exponential succinctness of nfasw.r.t dfas , it is important to have methods to obtain small nfas .
any right - invariant equivalence relation over @xmath34 w.r.t @xmath32 can be used to diminish the size of @xmath32 ( by computing the quotient automaton ) .
coarsest right - invariant equivalence _
@xmath123 can be computed by an algorithm similar to the one used to minimize dfas @xcite .
this coincides with the notion of ( auto)-bisimulation , widely applied to transition systems and which can be computed efficiently ( in almost linear time ) by the paige and tarjan algorithm @xcite .
left - invariant _ equivalence relation on @xmath34 w.r.t @xmath32 is any right - invariant equivalence relation on the reversed automaton of @xmath32 , @xmath124 , where @xmath125 if @xmath126 ( and we allow multiple initial states ) .
coarsest left - invariant equivalence _ on @xmath34 w.r.t @xmath32 , @xmath127 , is @xmath123 of @xmath128 .
* fado * @xcite is an ongoing project that aims to provide a set of tools for symbolic manipulation of formal languages . to allow high - level programming with complex data structures , easy prototyping of algorithms , and portability are its main features .
it is mainly developed in the ` python`programming language . in * fado * , regular expressions and finite automata are implemented as ` python`classes .
figure [ fig : regexpclass ] presents the classes for resand the main methods described in this paper . the `
regexp ` class is the base class for all resand the class ` position ` is the base class for marked res . the methods ` first ( ) ` , ` last ( ) ` and ` followmap ( ) ` ( where @xmath129 ) are coded for each subclass .
the method ` nfaposition ( ) ` implements a construction of the @xmath87 automaton without reduction to .
brggemann - klein algorithm is implemented by the ` nfapsnf ( ) ` method .
the methods ` nfafollowepsilon ( ) ` and ` nfafollow ( ) ` implement the construction of the @xmath118 via an @xmath0-nfa . the exact text of all these algorithms is too long to present here .
the method ` nfapd ( ) ` computes the @xmath117 and uses the method ` linearform ( ) ` .
this method implements the function @xmath130 defined by antimirov @xcite to compute the partial derivatives of a rew.r.t all letters .
algorithm [ alg : nfapd ] presents the computation of the @xmath117 .
@xmath131 @xmath132 @xmath133 stack @xmath134 @xmath135tail@xmath136 push(stack , pd ) @xmath137pd , head@xmath138pd , head@xmath139tail@xmath136 @xmath140pd@xmath136 figure [ fig : faclass ] presents the classes for finite automata . `
fa ` is the abstract class for finite automata .
the class ` nfar ` includes the inverse of the transition relation , that is not included in the ` nfa ` class for efficiency reasons . in the ` nfa `
class the method ` autobisimulation ( ) ` implements a nave version for compute @xmath123 , as presented in algorithm [ alg : equiv ] . given an equivalence relation the method ` equivreduced ( )
` builds the quotient automaton .
given an nfa**a * * , * a*.`requiv ( ) ` corresponds to @xmath141 , * a*.`lequiv ( ) ` to @xmath142 and * a*.`lrequiv ( ) ` to @xmath143 .
we refer the reader to gouveia @xcite and to * fado*webpage @xcite for more implementation details .
@xmath144 @xmath145 @xmath146 @xmath147 * return * @xmath148
uniform random generators are essential to obtain reliable experimental results that can provide information about the average - case analysis of both computational and descriptional complexity . for general regular expressions , the task is somehow simplified because they can be described by small unambiguous context - free grammars from which it is possible to build uniform random generators @xcite . in the * fado*system we implemented the method described by mairson @xcite for the generation of context - free languages .
the method accepts as input a context - free grammar and the size of the words to be uniformly random generated .
the random samples need to be consistent and large enough to ensure statistically significant results . to have these samples readily available , the * fado*system includes a dataset of random re , that can be accessed online .
the current dataset was obtained using a grammar for resgiven by lee and shallit @xcite , and that is presented in figure [ fig : grammar ] .
this grammar generates resnormalized by rules that define reduced res , except for certain cases of the rule : @xmath149 , where @xmath83 .
the database makes available random samples of reswith different sizes between @xmath150 and @xmath151 and with alphabet sizes between @xmath152 and @xmath153 .
in order to experiment with several properties of resand nfaswe developed a generic program to ease to add / remove the methods to be applied and to specify the data , from the database , to be used .
here we are interested in the comparison of several resdescriptional measures with measures of the nfasobtained using the methods earlier described . for reswe considered the following properties : the alphabetic size ( ) ; the size ( ) ; test
if it is in ( ) ; if not in , compute the and its measures ( , ) ; test if it is reduced ; if not reduced , reduce it and compute its measures ( , ) ; the number of states ( ) and number of transitions ( ) of the equivalent minimal dfa . for each nfa(@xmath87 , @xmath118 , and @xmath117 ) we considered the following properties : the number of states ( @xmath154 ) ; the number of transitions ( @xmath155 ) ; if it is deterministic ( ) ; and if it is homogeneous ( ) . all these properties were also considered for the case where the resare in , and for the nfasobtained after applying the invariant equivalences @xmath123 , @xmath127 , and their composition .
all tests were performed on samples of @xmath156 uniformly random generated res .
each sample contains resof size @xmath153 , @xmath157 , @xmath158 and @xmath159 , respectively .
table [ tab : resp ] shows some results concerning res .
the ratio of alphabetic size to size is almost constant for all samples .
almost all resare in , so we do not presented the measures after transforming into .
this fact is relevant as the reswere generated only _ almost reduced_. the column contains the percentage of resfor which their are reduced .
it is interesting to note that the average number of states of the minimal dfa ( ) is near ( i.e. , near the number of states of @xmath87 ) .
the standard deviation is here very high . for the sample of size @xmath159 , however , @xmath160 of the reshave @xmath161 .
more theoretical work is needed for a deeper understanding of these results .
table [ tab : nfasp ] and table [ tab : ratios ] show some results concerning the nfasobtained from res . in table
[ tab : nfasp ] the values not in percentage are average values . if @xmath87 is deterministic then the resis unambiguous ( and strong unambiguous , if in ) @xcite .
the results obtained suggest that perhaps @xmath162 of the reduced resare strong unambiguous .
note that if @xmath87 is not deterministic , almost certainly , neither @xmath117 nor @xmath118 are . for reasonable sized res ,
although @xmath87 are homogeneous it is unlikely that either @xmath117 or @xmath118 will be so .
it is not significant the difference between @xmath163 and @xmath164 . on average
@xmath165 seems linear in the size of the re , and that fact was recently proved by nicaud @xcite .
reductions by @xmath123 and @xmath127 ( or @xmath166 ) decrease by less than @xmath167 the size of the considered nfas(@xmath87 , @xmath118 , and @xmath117 ) .
in particular the quotient automata of @xmath87 are less than @xmath168 smaller than @xmath117 . in general , we can hypothesize that reductions by the coarsest invariant equivalences are not significant when resare reduced ( and/or are in ) .
we presented a set of tools within the * fado*system to uniformly random generate res , to convert resinto @xmath0-free nfasand to simplify both resand nfas .
these tools can be used to obtain experimental results about the relative descriptional complexity of regular language representations on the average case .
our experimental data corroborate some previous experimental and theoretical results , and suggest some new hypotheses to be theoretically proved .
we highlight the two following conjectures .
reduced reshave high probability of being in .
and the @xmath117 obtained from resin reduced seems to almost coincide with quotient automata of @xmath87 by @xmath166 .
a. almeida , m. almeida , j. alves , n. moreira , and r. reis . and guitar : tools for automata manipulation and visualization . in s.
maneth , editor , _
14th ciaa09 _ , volume 5642 of _ lncs _ , pages 6574 .
springer , 2009 .
l. ilie , r. solis - oba , and s. yu . reducing the size of nfas by using equivalences and preorders . in a.
apostolico , m. crochemore , and k. park , editors , _ 16th cpm 2005 _ , volume 3537 of _ lncs _ , pages 310321 .
springer , 2005 .
j. lee and j. shallit .
enumerating regular expressions and their languages . in m.
domaratzki , a. okhotin , k. salomaa , and s. yu , editors , _
9th ciaa 2004 _ , volume 3314 of _ lncs _ , pages 222 .
springer , 2005 . | regular expressions ( res ) , because of their succinctness and clear syntax , are the common choice to represent regular languages . however , efficient pattern matching or word recognition depend on the size of the equivalent nondeterministic finite automata ( nfa ) .
we present the implementation of several algorithms for constructing small @xmath0-free nfasfrom reswithin the * fado*system , and a comparison of regular expression measures and nfasizes based on experimental results obtained from uniform random generated res . for this analysis ,
nonredundant resand reduced resin star normal form were considered . |
the pulsar system psr b1259@xmath263 was discovered at parkes in 1989 and comprises a 47.76 ms radio pulsar orbiting a massive star ( ls 2883 ) in a highly elliptical ( @xmath8 ) orbit with a period of @xmath9 years @xcite .
recent optical spectroscopy @xcite yields an updated distance estimate to this source of @xmath10 kpc , in reasonable agreement with the dispersion measure ( dm ) derived distance of 2.7 kpc using the ne2001 model @xcite , so we adopt @xmath11 kpc .
the companion shows evidence for an equatorial disk in its optical spectrum , and has generally been classified as a be star @xcite .
the pulsar comes within @xmath12 au of its companion star at periastron , which is roughly the size of the equatorial disk @xcite .
the orbital plane of the pulsar is believed to be highly inclined with respect to this disk and so the pulsar crosses the disk plane twice each orbit , just before and just after periastron @xcite .
shock interaction between the relativistic pulsar wind and the wind and photon field of the be star is believed to give rise to the variable unpulsed x - ray emission observed throughout the orbit @xcite and the unpulsed radio and tev @xmath1 rays observed within a few months of periastron @xcite . at energies around 1 gev ,
the energetic gamma - ray experiment telescope ( _ egret _ ) provided only an upper limit for the 1994 periastron passage ( @xmath13 @xmath14 s@xmath15 for e @xmath0 300 mev , 95% confidence , @xcite ) . in tev
-rays the system was detected during the 2004 and 2007 periastron passages and flux variations on daily timescales were seen for energies @xmath16 0.38 tev in 2004 @xcite . for the 2010/2011 passage the time of periastron
@xmath4 was on 2010 december 15 .
by comparison to previous passages , the unpulsed radio and x - ray emission was expected to start rising in mid 2010 november peaking around @xmath17d in the pre - periastron phase and reaching another peak around @xmath18d in the post - periastron phase . by 2011 april
these emissions are expected to go back to their levels when the pulsar is far from periastron .
@xcite reported the first discovery of gev -ray emission from this system which was detected during the first disk passage .
a flaring gev -ray activity during the second disk passage was reported in @xcite and in @xcite .
recently @xcite reported with further details the gev -ray activity from this system .
we have assembled a multiwavelength campaign to monitor the system in radio , optical , x - rays , gev , and tev -rays during the 2010/2011 periastron passage .
here we describe the _ fermi_-lat detection of psr b1259@xmath263 in the @xmath19 mev range .
we also present a preliminary analysis of a portion of the radio and x - ray data to determine if there was any anomalous multiwavelength behavior compared to previous periastron passages .
we have analyzed lat data over the entire time period from the beginning of the _ fermi _ mission ( 2008 august 4 ; at which time the pulsar was nearing apastron ) through periastron up until 2011 april 22 which is after the passage of the pulsar through the dense equatorial wind of the massive star .
full analyses and interpretation of the multiwavelength data are deferred to subsequent papers .
analysis of the _ fermi _ lat data was performed using the _ fermi _ science tools 09 - 21 - 00 release .
the high - quality `` diffuse '' event class was used together with the p6@xmath20v3@xmath20diffuse instrument response functions . to reject atmospheric -rays from the earth s limb , we selected events with zenith angle @xmath21 .
we performed standard binned maximum likelihood analysis using events in the range 0.1100gev extracted from a 20@xmath22 region centered on the location of psr b1259@xmath263 .
the model includes diffuse emission components as well as @xmath1-ray sources within 20@xmath23 of the source ( based on an internal catalog created from 18 months of lat survey data ) .
the galactic diffuse emission was modeled using the ` gll\_iem\_v02 ` model and the isotropic component using ` isotropic\_iem\_v02 ` . to better constrain the diffuse model components and the nearby sources , we first generated a model using two years of data between 2008 august 4 and 2010 august 4 , a period during which the pulsar was far away from periastron .
we fixed spectral parameters of all the sources between 5@xmath23 and 15@xmath23 from the source , and left free the normalization factor of all the sources within 5@xmath23 that were flagged as variable source in the 1fgl catalog @xcite .
normalizations for the diffuse components were left free as well . for this time period ,
the source was not detected with the lat and we place a 95% upper limit on the photon flux above 100 mev f@xmath24 assuming a power law spectrum with a photon index @xmath25 .
the results of this fit were used to constrain the background source model for analyses on shorter timescales starting in november 2010 . in the source model ,
the normalization of the isotropic component was fixed to the 2-year value , while the normalization for the galactic diffuse component and three variable sources were left free .
we searched for -ray emission from this source on daily and weekly time scales during the first disk passage ( mid november to mid december 2010 ) .
no detection at the level of 5@xmath26 was observed from the source on these time scales .
integrating from @xmath27d ( the typical start of enhanced x - ray and unpulsed radio flux ) to periastron yielded a clear detection of excess -ray flux from the source with a test statistic ( ts ) of @xmath28 which corresponds to a detection significance of @xmath295@xmath26 @xcite . to estimate the duration of this enhanced emission and to get the best fit for the spectrum we looked at the cumulative ts as a function of time for integrations starting at @xmath27d ( figure [ fig : ts ] ) .
inspection of this plot reveals that the ts drops monotonically for integrations ending after @xmath30d , so we use this as the end of the integration for this initial period of detected @xmath1-ray flux . for the rest of the paper we will refer to this period as the `` brightening '' .
+ during the brightening , the detected -ray signal was in the energy range 0.11 gev and no significant emission was detected above 1 gev .
the spectrum in the energy range 0.1 - 1 gev is best described by a simple power law with photon index @xmath31 with average photon and energy fluxes of f = @xmath32 and f = @xmath33 respectively . because of the low signal - to - noise ratio during this period spectral fits to an exponentially cutoff power law were not constraining .
this period is shown as the shaded region on the top panel of figure [ fig : lat_lc ] , which shows the -ray flux in weekly time bins in the period @xmath34d to @xmath35d .
at about @xmath36d the source brightened rapidly , reaching a flux that was @xmath37 times higher than the integrated flux measured during the first disk passage ( figure [ fig : lat_lc ] ) .
this flare lasted for about 7 weeks .
the spectrum during this period ( @xmath36d to @xmath38d ) is best described by a power law with exponential cutoff .
the best - fit result is obtained for a photon index @xmath39 = ( 1.4 @xmath40 ) and a cutoff energy e@xmath41 = ( 0.3 @xmath42 ) gev .
the average photon and energy fluxes above 100 mev during this period are f@xmath43 ( @xmath44 10@xmath45 and g@xmath46 respectively .
+ in addition to the significant difference in flux and spectral shape between the brightening and the flare periods , the weekly time bins show clear evidence for a change in the photon index for a power - law fit on these time scales ( bottom panel of figure [ fig : lat_lc ] ) .
the phonon index softens from 22.5 during the brightening to a value of 3.5 around the peak of the flare .
after that , the index hardens over the rest of the flare period to its values during the brightening .
+ the top panel of figure [ fig : lat_lc_flare ] shows -ray flux as a function of time in 1-day time bins during the flare .
the daily source flux during this period is variabile at the 99.9% confidence level according to the method outlined in @xcite . during this strong - variability flaring period the source flux varied by a factor of 23 on daily time scales .
the previous upper limits from _
egret _ are fully consistent with the flux observed by during the same orbital phase @xcite . at the time of the bright emission seen by _
fermi _ , which was well above the sensitivity level of _ egret _ , _ egret _ was pointed elsewhere .
to search for evidence of -ray pulsations from psr b1259@xmath263 , we have constructed a radio ephemeris using observations from the 64-m parkes telescope .
using tempo2 @xcite , we fitted a timing model to 45 toas covering the range 2007 oct 6 through 2011 january 16 . using this ephemeris , we folded lat photons in the interval 2008 august 4 through 2010 august 4 , a total of 24 months of observation , all well away from the flaring region near periastron where non - pulsed -ray emission is observed .
we found no statistically significant indication of a pulsed -ray signal from this source .
the pulsed flux upper limit depends on the unknown pulse shape and assumed spectrum .
we therefore use the continuum upper limit on the energy flux above 100 mev of g@xmath47 erg @xmath14 s@xmath15 in our comparisons with the rest of the -ray pulsar population . comparing this pulsar to the rest of the lat - detected pulsars
, we find that most detectability metrics predict that this should be a -ray pulsar .
although the characteristic age of 333 kyr is fairly large , the spin period is short for a middle - aged pulsar and thus at a distance of 2.3 kpc the @xmath48 of 8.2@xmath49 erg s@xmath15 and magnetic field at the light cylinder of 2.9 @xmath50 g are well within the range where -ray pulsations are typically detected @xcite .
if we assume that the beaming factor @xmath51 is 1 , then the -ray efficiency of the pulsar is less than 0.7% , which is considerably lower than the 10% efficiencies that are typical in this @xmath48 range .
determining whether this pulsar is intrinsically under - luminous in the -ray band or if the low -ray luminosity is simply a geometric effect will require detailed modeling that includes geometrical information from radio polarization measurements .
we have also searched for -ray pulsations during the brightening and flare periods where continuum -ray emission was detected .
no pulsations were detected in these intervals .
this is consistent with this -ray emission originating from the intrabinary shock , which is well outside the light cylinder of the pulsar and thus is not expected to be modulated at the spin period .
pulsed emission was monitored at parkes to look for changes in the dm and rotation measure ( rm ) and determine the duration of eclipse of the pulsed signal .
pulses disappeared on @xmath52d and reappeared on @xmath53 . in the @xmath54 weeks leading up to the disappearance of the pulse , significant changes in the dm were observed .
the system was monitored at frequencies between 1.1 and 10 ghz using the atca array .
twelve observations spanning @xmath55d to @xmath56d were collected .
unpulsed transient radio emission was detected throughout the periastron passage with a behavior similar to that seen in previous observations @xcite , as shown in figure [ fig : radio - xray - flux ] .
x - ray observations of during this passage demonstrated the repeatability of the 1 - 10 kev light curve as shown in figure [ fig : radio - xray - flux ] . as with the periastron passage of 2004 , observed a rapid x - ray brightening starting at @xmath57d .
these observations confirmed the spectral hardening preceding the pre - periastron flux rise .
similarly to previous periastron passages ( see for example @xcite and references therein ) , observations with , and showed a rise of x - ray flux after periastron .
emission from the system is produced in the interaction of the pulsar wind with the stellar wind of the companion star .
observations in radio , x - ray and tev @xmath1-ray bands @xcite revealed a characteristic variability of this emission during the periods of periastron passage .
detection of the 0.1 - 10 gev band @xmath1-ray emission around periastron was not unexpected .
however , _ fermi _ observations reveal puzzling behavior of the source , which was not predicted in any model of @xmath1-ray emission from this system .
an unexpected strong flare , visible only in the gev band was observed some 30 days after the periastron passage and after the neutron star passage of the dense equatorial wind of the massive star . during this flare
the source was characterized by an extremely high efficiency of conversion of pulsar spin - down power into @xmath1-rays .
the highest day - average flux was f@xmath58 @xmath59 s@xmath60 with a spectral index of @xmath61 .
this corresponds to an isotropic @xmath1-ray luminosity of @xmath62 erg s@xmath15 , nearly equalling the estimated total pulsar spin down luminosity @xmath63 erg s@xmath15 @xcite .
this is illustrated in figure [ fig : synch ] where the horizontal red line shows the flux which would be produced when 100% of the spin - down power is converted into radiation emitted within one decade of energy , not taking into account possible beaming effects .
broadband spectra of emission around periastron are shown in figure [ fig : synch ] .
strong increases in gev flux and changes in -ray spectrum during the flare were not accompanied by noticeable spectral variations in the x - ray band . several possible mechanisms of production of 0.1 - 10 gev @xmath1-ray emission from the system were previously discussed : synchrotron , inverse compton ( ic ) , bremsstrahlung , or pion decay emission @xcite .
electrons with energies @xmath64 tev produce synchrotron emission in the energy range @xmath65\left[e_e/10^{14}\mbox { ev}\right]^2 $ ] ev .
alternatively , electrons with energies @xmath66 gev could produce @xmath1 quanta with energies @xmath67 ^ 2 $ ] ev via ic scattering of be star photons .
bremsstrahlung emission in the gev band could be produced by the gev electrons .
finally , the dense equatorial stellar wind could provide a sufficiently dense target for proton - proton interactions followed by decays of neutral pions into photons .
figure [ fig : synch ] shows example model fits to the persistent emission data .
the model shown in the upper panel assumes that high - energy particles escape with the speed of the stellar wind , as in the model of @xcite and @xcite .
slow escape of the high - energy particles leaves enough time for the efficient cooling of electrons via ic and/or bremsstrahlung / coulomb loss mechanisms . in the lower panel ,
high - energy particles are assumed to escape with the speed @xmath68 cm s@xmath15 , as in the model of @xcite . in this case
only synchrotron cooling is efficient .
the code used for the calculations is described in @xcite . in general , the flare could be explained either by anisotropy of the @xmath1-ray emission , or by an abrupt change of physical conditions in the emission region or by the appearance of a new emission component .
several possible sources of anisotropy are present in the system : relativistic beaming of the @xmath1-ray emitting outflow from the system @xcite , anisotropy of pulsar wind , or anisotropy of radiation field of the massive star .
the model for the flare spectrum ( cyan data points ) shown in the lower panel of fig .
[ fig : synch ] assumes a particular type of anisotropy which could appear at the high - energy end of the synchrotron spectrum , at the energies at which the synchrotron cooling distance is comparable to the gyroradius . in such a situation
the electron distribution could not be isotropized within the synchrotron cooling time scale .
the assumpiton that highest energy electrons with anisotropic initial velocity distribution cool before being isotropized , results in the increase of apparent luminosity by a factor @xmath69 where @xmath70 is the solid angle into which most of the highest energy synchrotron power is emitted .
another possibility to explain the flare is considered in the model shown in the upper panel of fig . [ fig : synch ] .
it iassumes a local increase of the density of stellar wind by a factor of @xmath71 which results in the increase of the bremsstrahlung component of emission spectrum .
clarification of the physical mechanism of the puzzling flare discovered by _
fermi _ requires more complete view of the properties of the flare , including information on system behaviour in optical and tev bands and on orbit - to - orbit variations of the gev flaring pattern .
d ( see text ) .
the shaded area marks the 10 days during which the lat was in a modified rocking mode giving @xmath72 higher exposure on the source .
the arrow marks the time we adopt as the end of the first emission period . ]
d to @xmath73d .
_ upper panel _
: @xmath0 100 mev flux , 2@xmath26 upper limits are drawn for points with ts @xmath74 5 .
_ bottom panel _ : variations of spectral index of a power law spectrum .
the shaded area shows the brightening period .
dashed line marks the time of periastron .
dashed - dotted lines marks the orbital phase during which _ egret _ observed this source in 1994 @xcite . ] the @xmath75 lat collaboration acknowledges support from a number of agencies and institutes for both development and the operation of the lat as well as scientific data analysis .
these include nasa and doe in the united states , cea / irfu and in2p3/cnrs in france , asi and infn in italy , mext , kek , and jaxa in japan , and the k. a. wallenberg foundation , the swedish research council and the national space board in sweden . additional support from inaf in italy and cnes in france for science analysis during the operations phase is also gratefully acknowledged .
the parkes radio telescope is part of the australia telescope which is funded by the commonwealth government for operation as a national facility managed by csiro .
we thank our colleagues for their assistance with the radio timing observations .
this work was supported in part by a nasa guest investigator program . | we report on the discovery of @xmath0 100 mev @xmath1 rays from the binary system psr b1259@xmath263/ls 2883 using the large area telescope ( lat ) on board _
fermi_. the system comprises a radio pulsar in orbit around a be star .
we report on lat observations from near apastron to @xmath3 days after the time of periastron , @xmath4 , on 2010 december 15 .
no -ray emission was detected from this source when it was far from periastron .
faint @xmath1-ray emission appeared as the pulsar approached periastron . at @xmath5d , the @xmath0 100
mev @xmath1-ray flux increased over a period of a few days to a peak flux 2030 times that seen during the pre - periastron period , but with a softer spectrum . for the following month
, it was seen to be variable on daily time scales , but remained at @xmath6 before starting to fade at @xmath7d .
the total @xmath1-ray luminosity observed during this period is comparable to the spin - down power of the pulsar .
simultaneous radio and x - ray observations of the source showed no corresponding dramatic changes in radio and x - ray flux between the pre - periastron and post - periastron flares . we discuss possible explanations for the observed -ray - only flaring of the source . |
one of the most remarkable results in the 3-yr data release from the wilkinson microwave anisotropy probe ( wmap ) experiment @xcite was the detection of large - scale e - mode polarization at millimeter wavelengths .
this was interpreted as the theoretically predicted signature of reionization , and allowed the wmap team to set new and tighter constraints on the optical depth of reionization @xmath2 . in turn ,
the well - known degeneracy between @xmath2 and the spectral index of primordial scalar perturbations @xmath3 was broken .
the final outcome was a claimed detection of @xmath4 at a statistical significance of almost @xmath5 @xcite .
one should bear in mind , however , the great potential for systematics effects in both the temperature and polarization measurements .
for instance , the precise level of power contribution from unresolved point sources affects @xmath3 directly .
an independent analysis of this particular issue by @xcite showed that the initial point source amplitude quoted by the wmap team was indeed too high , which biased @xmath3 to low values .
similarly , on large scales the likelihood approximation used by the wmap team was biased high @xcite , which also biased @xmath6 low .
after these corrections , the statistical significance of @xmath7 dropped to @xmath8 . for polarization the situation may be even more serious due to the strong sensitivity of @xmath2 and @xmath6 on the low-@xmath1 ee spectrum , combined with the low signal - to - noise ratio of the wmap data .
systematic effects , both from the instrument itself @xcite and from non - cosmological foregrounds @xcite , are much more likely to affect the results , and we are also much less likely to detect them .
it is therefore imperative to carefully check both the data and the analysis methods , in order to build up confidence in the final cosmological results . in this letter
, we start this task by computing the low-@xmath1 ee , eb , bb and foreground template amplitude posterior distributions from the wmap data .
we use a previously introduced gibbs sampling framework ( jewell et al . 2004 ,
wandelt et al .
2004 , eriksen et al . 2004 ,
larson et al.2007 ; hereafter jwel ) to estimate the posterior distributions . for full details on the method , we refer the interested reader to the quoted papers , and only summarize the principles here .
first we define our notation .
the desired distribution is denoted @xmath9 , where @xmath10 is the cmb signal , @xmath11 is the cmb power spectrum , @xmath12 is a set of foreground template amplitudes , and @xmath13 are the data .
the gibbs sampler is a markov chain monte carlo method , and , as such , maps out the full posterior by drawing samples from it .
while direct evaluation or sampling from the posterior @xmath14 requires inversion of a prohibitively large matrix , the gibbs sampling scheme @xcite uses the conditional densities of the joint posterior @xmath15 which is computationally feasible to sample from .
the algorithm may thus be described by the following sampling steps , @xmath16 here the symbol @xmath17 indicates sampling from the conditional distribution on the right hand side , which can be accomplished without inverting the signal - plus - noise covariance matrix ( see jwel for details ) . for the foreground template amplitude distribution
, we note that the required algorithm is identical to that employed for sampling monopole and dipole amplitudes @xcite .
we consider only the low - resolution foreground - corrected 3-yr wmap polarization data in this letter , as provided on lambda . these come in the form of three healpix sky maps , pixelized at @xmath18 , each having 3072 pixels in both
stoke s q and u. the wmap p06 sky cut is imposed on the data , leaving only 2267 pixels for the analysis .
two frequency bands are included , namely q - band ( 41 ghz ) and v - band ( 61 ghz ) .
in addition , we analyze the co - added map ( denoted qv ) , and also the two frequency maps jointly but not co - added ( denoted q+v ) . all maps are provided with a full noise covariance matrix @xcite , appropriately corrected for the p06 sky cut and removal of foreground templates .
the units used in this paper are thermodynamic @xmath19 . for foreground marginalization
, we adopt two individual templates .
first , we use the k
ka difference map , smoothed to @xmath20 fwhm resolution to reduce noise contributions , as a tracer of synchrotron emission .
second , for dust emission we adopt the low - noise template developed by the wmap team for their foreground correction procedure @xcite .
note that the specific shape of these templates are of minor importance ; if the provided sky maps are free of foregrounds , they should not correlate significantly with any non - cmb map .
we compare our results to the official wmap likelihood code , also available from lambda . to map out the appropriate posteriors , we have written a special - purpose mcmc wrapper around this likelihood code .
the first case considered is that adopted by the wmap likelihood code , namely the co - added qv map . for this analysis , we fix the eb and bb spectra to zero , and map out the corresponding marginalized ee posteriors @xmath1-by-@xmath1 , both with the gibbs sampler and by the wmap - based mcmc code . the results from this exercise are shown in figure [ fig : ee_comparison ] .
the agreement between the two approaches is very good , and this is an important validation of the wmap data processing method : first , we analyze the data at their native @xmath18 resolution , while the wmap team analyze maps downgraded to @xmath21 .
second , they marginalize over a single total foreground template , while we marginalize over the k
ka difference map and a dust template .
third , we use a gibbs sampler for the numerical work , while the wmap team uses a brute - force likelihood evaluator .
none of these differences affects the low-@xmath1 ee spectrum peak visibly , as will be quantified more precisely in the next section .
lcccc wmap & @xmath22 & @xmath23 & & + qv - band & @xmath24 & @xmath25 & & + q+v - band & @xmath26 & @xmath27 & & + q - band & @xmath28 & @xmath29 & & + v - band & @xmath30 & @xmath31 & & + wmap & @xmath32 & @xmath33 & @xmath34 & @xmath35 + qv - band & @xmath26 & @xmath36 & @xmath37 & @xmath38 + q+v - band & @xmath26 & @xmath39 & @xmath40 & @xmath41 + q - band & @xmath42 & @xmath43 & @xmath44 & @xmath45 + v - band & @xmath46 & @xmath47 & @xmath48 & @xmath49 + qv - band & @xmath50 & @xmath51 & @xmath52 & @xmath53 + q+v - band & @xmath54 & @xmath55 & @xmath56 & @xmath53 + q - band & @xmath57 & @xmath58 & @xmath59 & @xmath60 + v - band & @xmath61 & @xmath62 & @xmath63 & @xmath64 having validated the numerics and data processing steps in both procedures , we now expand the analysis and allow for non - zero eb and/or bb spectra .
we also compute the posteriors for each frequency band separately and jointly .
for presentational reasons , we report only band powers in ee and bb between @xmath656 and @xmath6520 , respectively .
cases of special interest are treated separately in subsequent sections . in order to achieve good convergence ,
@xmath66 samples were generated for the non - zero eb , @xmath6520 cases .
for all other cases , @xmath67 samples were generated .
the cpu time to generate one sample was @xmath68 seconds .
the results from these computations are summarized in table [ tab : band_power ] .
the ee posteriors with fixed eb = bb = 0 are shown in figure [ fig : ee_bin_power ] . again ,
note the excellent agreement between the wmap results and the qv and q+v cases in the two top sections .
however , even though the joint qv analyses agree well , the picture is considerably less clear when it comes to single bands and relaxed eb constraints . on the one hand
, there appears to be more ee power in the v - band data than in the q - band data . on the other hand
, there may seem to be small traces of bb power in the q - band data .
the bb posteriors develop a peak away from @xmath69 in the eb@xmath70 case .
this is not surprising .
since the signal covariance matrix is positive definite , one must have @xmath71 .
therefore , when marginalizing over @xmath72 , a non - zero peak emerges in both the ee and bb spectra individually . in figure
[ fig : foregrounds ] we show the foreground template amplitude posteriors for the q- , v- and qv - band data , for the case with fixed eb = bb = 0 . although no signs of significant residual foregrounds are observed in the co - added qv band , apparently confirming the fits made by the wmap team , the same is not true for each band individually . on the contrary ,
non - zero correlations are seen in both the q- and v - band data individually . for the q - band data , the marginal best - fit k
ka amplitude is @xmath73 , different from zero at @xmath74 .
the best - fit dust amplitude is @xmath75 . for the v - band data , the best - fit dust amplitude is @xmath76 , 2.3@xmath77 away from zero .
ka amplitude is @xmath78 .
these results may be compared to table 4 of @xcite .
the main difference between the two analyses is that while we apply the conservative p06 mask to the data , @xcite apply the much more aggressive processing mask described by @xcite .
the two analyses remove 26.4% and 5.7% of the sky , respectively .
considering that all cosmological analyses are carried out with the p06 mask , and that variations in the synchrotron spectral index are observed between the galactic plane and the high latitudes @xcite , it is not immediately clear to us why the more aggressive mask was chosen for this task by the wmap team .
the improvement in raw @xmath79 after further correcting the `` cleaned '' wmap maps for these residuals is -5.4 for the q - band and -3.3 for the v - band , respectively .
two other , minor differences are that we take into account both the full noise and cmb covariance matrices , while the wmap team includes only a white noise covariance approximation .
note that this carries no extra cost within the gibbs sampling framework .
we find evidence for a non - zero eb correlation in the v - band sky map .
this spectrum is shown in figure [ fig : eb_correlation ] for the q- and v - band sky maps individually , together with the pseudo-@xmath81 eb qv - band spectrum computed by the wmap team .
note the consistently negative correlation observed in the v - band .
although the significance of the negative correlation is not more than a few tenths per multipole , and the joint significance is not more than 1 to 2@xmath77 depending on binning scheme , it is observed consistently in every multipole up to @xmath82 . a similar , though weaker , trend is seen in the pseudo - spectrum computed by the wmap team from the qv combination .
we have performed a bayesian analysis of the low - resolution foreground - corrected 3-yr wmap polarization data using a previously described methodology based on the gibbs sampler . by doing so
, we validated the numerical implementation of the official wmap likelihood code , as well as their procedure for degradation of map resolution .
however , when generalizing the analysis to allow for non - zero eb and bb power spectrum components , and also considering individual frequency bands , we found several issues that may be important for the cosmological interpretation of these data .
first and foremost , when relaxing the constraints on eb and bb , noticeable differences between the q- and v - band posteriors are observed .
specifically , we find generally more ee power in the v - band data than in the q - band data , but also perhaps some hints of bb power in the q - band data . at the same time
, we have also found a negative eb correlation in the v - band map , as well as residual foregrounds in both maps .
if these tentative findings are confirmed by future experiments or additional years of wmap observations , significant shifts in cosmological parameters could be the result .
for example , if the v - band data alone were used for the wmap3 analysis instead of the qv combination , the amplitude of the @xmath656 ee detection would increase by 2050% , depending on whether bb is allowed to vary or not . consequently
, @xmath2 could increase from 0.09 to @xmath83 , and @xmath84 by a percent or two , comparable to its current nominal statistical uncertainty of 0.015 , from @xmath85 to @xmath86 . as discussed in the introduction , systematics are a serious concern for both the temperature and polarization data for both @xmath2 and @xmath6 .
it is important to bear in mind that the currently quoted uncertainties on these quantities often found in the literature are statistical only .
the unknown systematic uncertainties may turn out to be non - negligible for the currently available data sets , and , in particular , we believe it is too early to draw any firm conclusions concerning the precise value of @xmath6 . fortunately , planck will clarify these issues in the near future .
we wish to thank david larson for his contributions during the early phases of this project .
we acknowledge use of the legacy archive for microwave background data analysis ( lambda ) .
we acknowledge use of the healpix software @xcite and analysis package for deriving the results in this paper .
this work was partially performed at the jet propulsion laboratory , california institute of technology , under a contract with the national aeronautics and space administration .
hke acknowledges financial support from the research council of norway .
bdw acknowledges support by nsf grant no .
0507676 , the friedrich wilhelm bessel research prize by the alexander v. humboldt foundation and nasa subcontract no . | we apply a previously developed gibbs sampling framework to the foreground corrected 3-yr wmap polarization data and compute the power spectrum and residual foreground template amplitude posterior distributions .
we first analyze the co - added q- and v - band data , and compare our results to the likelihood code published by the wmap team .
we find good agreement , and thus verify the numerics and data processing steps of both approaches .
however , we also analyze the q- and v - band separately , allowing for non - zero eb cross - correlations and including two individual foreground template amplitudes tracing synchrotron and dust emission . in these analyses ,
we find tentative evidence of systematics : the foreground tracers correlate with each of the q- and v - band sky maps individually , although not with the co - added qv map ; there is a noticeable negative eb cross - correlation at @xmath0 in the v - band map ; and finally , when relaxing the constraints on eb and bb , noticeable differences are observed between the marginalized band powers in the q- and v - bands . further studies of these features are imperative , given the importance of the low-@xmath1 ee spectrum on the optical depth of reionization @xmath2 and the spectral index of scalar perturbations @xmath3 . |
when studying vector configurations or central hyperplane configurations , point sets on a sphere or great hypersphere arrangements , vector spaces or their duals , points on grassmannians , polytopes and their corresponding cellular decompositions in projective space , etc .
, an abstraction of an important equivalence class of matrices often plays a central role : an oriented matroid .
the theory of oriented matroids ( see @xcite ) provides us with a multitude of definitions for an oriented matroid that can be viewed as reflecting the variety of objects that a matrix can represent .
these definitions via circuit or cocircuit axioms ( @xcite , p. 103 ) , sphere systems ( @xcite , p. 227 ) , gramann plcker relations ( chirotope axioms ) ( @xcite , p. 126 , p. 138 , @xcite ) , hull systems ( @xcite ) , to mention just a few of them , differ a lot with respect to their motivational aspects , their algorithmical efficiency or their relation to the actual application .
each definition provides in general an additional insight for the motivating problem . in the research monograph on oriented matroids , @xcite ,
three chapters are devoted to axiomatics concerning oriented matroids and to the topological representation theorem for oriented matroids .
this central theorem in the theory of oriented matroids due to lawrence shows the equivalence of oriented matroids defined via sphere system axioms with oriented matroids defined via , say , the circuit axioms .
this remarkable result asserts that each oriented matroid has a topological representation as an oriented pseudosphere arrangement , even a piecewise - linear one , cf .
edmonds and mandel @xcite . other authors ( @xcite , @xcite )
have later simplified or complemented the original proof , but all use fundamentally the same approach : the face lattice ( tope ) formalism for oriented matroids and a shelling order to carry through the construction . finding a reasonably direct proof in rank @xmath0 , one that would rely on the structural simplicity of the planar case , has been posed as an open problem in the research monograph @xcite ( exercise 6.3 ) . in @xcite
such a proof in the rank 3 case was given .
unlike the previous ones , this was based on _ hyperline sequences _ , an equivalent axiomatization for oriented matroids which is particularly natural in rank @xmath0 . in this article
we generalize this proof to the arbitrary rank case .
the proof is inductive , direct and uses only one advanced result from topology , the generalized schnflies theorem .
the motivation for introducing hyperline sequences in rank 3 was presented in detail in our previous paper @xcite .
here we use the generalization to arbitrary rank , formally defined in section [ sec : hls ] .
the motivation is similar .
here is a brief account of it .
we use an index set @xmath1 with its standard total order .
we consider a labeled finite point set @xmath2 @xmath3 , ordered according to their labels .
we assume that the point set @xmath4 affinely spans @xmath5 .
we endow the affine euclidean space @xmath5 with an orientation , i.e. , for any @xmath6element subset of affinely independent points , we can tell whether the ordering of their indices defines a positive , or a negative , oriented simplex , respectively .
every @xmath7element subset of affine independent points of @xmath4 defines via the ordering of their indices an oriented hyperplane , i.e. , an affine hull of codimension @xmath8 together with an orientation .
furthermore , every @xmath9element subset of affine independent points of @xmath4 defines via the ordering of their indices an oriented hyperline , i.e. , an affine hull of codimension 2 together with an orientation of it .
although this affine model can be described further and the concept we are going to describe is useful for affine applications as well , in what follows we prefer to use the notational advantage that occurs when embedding this concept in the usual way in projective space .
first consider a bijective map between our former oriented @xmath7dimensional euclidean space @xmath5 and a non - central oriented hyperplane of the oriented space @xmath10 .
the former points @xmath11 then become non - zero vectors @xmath12 in @xmath10 , @xmath13 , the former oriented hyperplanes become oriented central hyperplanes in @xmath10 , and the former oriented hyperlines , defined via an index set @xmath14 with @xmath15 , become oriented central hyperlines ( subspaces of codimension 2 in @xmath10 together with an orientation ) . for
what follows we can also assume that we have started with an ordered set of vectors @xmath16 in our oriented space @xmath10 that has not necessarily an affine pre - image .
the oriented central hyperline with index set @xmath17 is an @xmath18 , and the restriction of @xmath19 to all elements in that hyperline yields an arrangement @xmath20 of non - zero vectors in @xmath21 .
we rotate an additional central oriented hyperplane in @xmath10 around one of our oriented central hyperlines with index set @xmath17 in the well defined positive sense , i.e. , the ordered set of vectors with index set @xmath17 together with two adjacent consecutive outer normal vectors of the hyperplane form a positive basis of our oriented space @xmath10 .
an orthogonal projection @xmath22 along the oriented hyperline onto its corresponding two - dimensional orthogonal oriented space @xmath23 ( that forms together with the oriented hyperline a positive orientation of the whole space ) shows the image of the additional central oriented hyperplane as a rotating ( mathematically positive ) central oriented line in the plane .
precisely all vectors @xmath12 that are not in the hyperline defined by @xmath17 appear after the projection as non - zero vectors @xmath24 their indices form a periodic circular sequence around the origin when we pursue the incidences of the vectors with the rotating line .
we store the cyclic order of incidences in a cyclic order of sets , where we write an index @xmath25 when the vector @xmath12 is consistent with the orientation of the oriented line , and @xmath26 otherwise .
we refer to this notation as * signed indices*. when an incidence position , say position @xmath27 , of the rotating oriented line with a vector occurs , we write the signed indices of all vectors that are incident with the oriented line in a set @xmath28 .
note that the period of the cyclic sequence is even , since there are two positions of the rotating oriented line in which it is incident with the vector @xmath29 , corresponding to the signed indices @xmath25 and @xmath26 .
we obtain a well defined oriented circular sequence @xmath30 ( * oriented cycle * ) of sets of signed indices , where after half the period there is a sign reversal for each element .
the combinatorics of the hyperline is encoded in the arrangement @xmath20 with index set @xmath17 together with the oriented circular sequence of incidences with the rotating line , @xmath31 , to which we also refer as a * hyperline*. we obtain a * hyperline sequence * when we write down all hyperlines arising in the vector arrangement .
this model can be generalized in two ways .
the combinatorial abstraction leads on the one hand to oriented matroids characterized as hyperline sequences and on the other hand it leads to topological equivalence classes of arrangements of pseudospheres .
we give a new proof of a one - to - one correspondence of oriented matroids and classes of arrangements of pseudospheres .
we further generalize arrangements of pseudospheres , and we show that one can still read off a hyperline sequence .
the paper is organised as follows .
we define hyperline sequences as a combinatorial abstraction of vector configurations in section [ sec : hls ] .
we recall the chirotope concept in section [ sec : chirotope ] and we show the equivalence of these two concepts in theorem [ thm : hls - chi ] .
furthermore , we discuss the concepts of deletion and contraction for both settings . we introduce the topological representation of oriented matroids via arrangements of oriented pseudospheres in section [ sec : psa ] with the corresponding concepts of deletion and contraction . in the proof of theorem [ thm : cells ] we use the generalized schnflies theorem to show the cell structure of an arrangement of oriented pseudospheres . in theorem
[ thm : psa ] we replace our two axioms for an arrangement of oriented pseudospheres with a single one .
section [ sec : x(a ) ] deals with the easier part of the main representation theorem : we obtain chirotopes and hyperline sequences from the topological representation .
we complete the proof of the one - to - one correspondence of hyperline sequences with their topological representation by induction .
section [ sec : flbase ] is devoted to the base case and section [ sec : flgeneral ] contains the essential part .
we finally discuss in section [ sec : wild ] the wild arrangement case as an easy supplement of our approach .
the whole proof is based heavily on the ideas that have been worked out already in @xcite by the first and the last two authors in the rank @xmath0 case .
but the experience with respect to topological arguments of the second author was decisive to arrive at our final version of the proof in the general rank case .
the last section and many improvements of the proof compared with the rank 3 version are due to him .
for instance , the uniform and non - uniform cases form no longer separate parts within the proof .
our aim in this section is to introduce the notion of hyperline sequences .
the geometric motivation of our definition comes from vector arrangements in euclidian space as explained in the introduction .
let @xmath32 be a finite totally ordered set .
let @xmath33 be a copy of @xmath34 .
the set @xmath35 of * signed indices * is defined as the disjoint union of @xmath34 and @xmath36 . by extending the map @xmath37 to @xmath38 for @xmath39
, we get an involution on @xmath35 .
we define @xmath40 . for @xmath41 , define @xmath42 and @xmath43 .
an * oriented @xmath44simplex * in @xmath34 is a @xmath45tuple @xmath46 $ ] of elements of @xmath35 , so that @xmath47 are pairwise distinct .
let an equivalence relation @xmath48 on oriented @xmath44simplices in @xmath34 be generated by @xmath49\sim [ x_1,\dots , x_{i-1},\overline{x_{i+1}},x_i , x_{i+2},\dots , x_{d+1}]$ ] , for @xmath50 . as usual , any oriented @xmath44simplex is equivalent to one of the form @xmath51 $ ] or @xmath52 $ ] , with elements @xmath53 of @xmath34 .
define @xmath54 = [ x_1,\dots , x_d,\overline{x_{d+1}}]$ ] .
if @xmath55 is a map with @xmath56 for all @xmath57 then @xmath58 are the vertices of a simplex in @xmath59 .
note that this simplex might be degenerate . in the following iterative definition of hyperline sequences we denote with @xmath60 the cyclic group of order @xmath27 .
a * hyperline sequence * @xmath4 over @xmath61 of rank @xmath8 is a non - empty subset @xmath62 so that @xmath63 and @xmath64 .
the oriented simplex @xmath65 $ ] is by definition a * positively oriented base * of @xmath4 for @xmath66 .
we define @xmath67 .
let @xmath68 , @xmath69 . a * hyperline sequence * @xmath4 of rank @xmath70 over @xmath61 is a map from @xmath71 to hyperline sequences of rank one , @xmath72 , so that @xmath73 for all @xmath74 , and @xmath75 is a disjoint union of @xmath76 .
we refer to @xmath76 as the * atoms * of @xmath4 and to @xmath77 as the * period length * of @xmath4 .
we say that @xmath78 is * incident * to an atom @xmath79 of @xmath4 if @xmath80 .
let @xmath81 so that @xmath82 and @xmath83 are not incident to a single atom of @xmath4 , and @xmath4 induces the cyclic order @xmath84 .
then , the oriented simplex @xmath85 $ ] is by definition a * positively oriented base * of @xmath4 .
we define the hyperline sequence @xmath86 over @xmath87 of rank @xmath70 as the map @xmath88 for @xmath74 .
a hyperline sequence @xmath4 of rank @xmath70 is determined by the sequence @xmath89 of atoms .
we define that two hyperline sequences @xmath90 and @xmath91 of rank @xmath70 are equal , @xmath92 , if @xmath93 , the number @xmath77 of atoms coincides , and @xmath90 is obtained from @xmath91 by a shift , i.e. , there is an @xmath94 with @xmath95 for all @xmath74 .
we prepare the axioms for hyperline sequences of rank @xmath96 with the following definitions .
let @xmath4 be a set of pairs @xmath97 , where @xmath20 is a hyperline sequence of rank @xmath98 and @xmath99 is a hyperline sequence of rank @xmath70 .
a * positively oriented base * of @xmath4 in @xmath100 is an oriented simplex @xmath101 $ ] in @xmath102 , where @xmath103 $ ] is a positively oriented base of @xmath20 and @xmath104 $ ] is a positively oriented base of @xmath99 .
we define @xmath105 .
the elements of @xmath4 are called * hyperlines*. an * atom * of @xmath4 in a hyperline @xmath100 is the pair @xmath106 , where @xmath107 is an atom of @xmath99 .
let @xmath108 be a set whose elements are pairs @xmath97 , where @xmath20 is a hyperline sequence of rank @xmath98 and @xmath99 is a hyperline sequence of rank @xmath70 .
the set @xmath4 is a * hyperline sequence * of rank @xmath96 over @xmath61 if it satisfies the following axioms .
* @xmath102 is a disjoint union of @xmath109 and @xmath110 , for all @xmath100 .
* let @xmath111 and let @xmath112 $ ] be a positively oriented base of @xmath113 . if @xmath114 then @xmath115 or @xmath116 .
* for all positively oriented bases @xmath117 $ ] and @xmath118 $ ] of @xmath4 , there is some @xmath119 so that @xmath120 $ ] or @xmath121 $ ] is a positively oriented base of @xmath4 . * for any positively oriented base @xmath117 $ ] of @xmath4 , @xmath122 $ ] is a positively oriented base of @xmath4 .
we call an oriented simplex @xmath123 a negatively oriented base of a hyperline sequence @xmath4 if @xmath124 is a positively oriented base of @xmath4 .
we connect these axioms to the geometric motivation exposed in the introduction .
let @xmath125 be non - zero vectors spanning @xmath10 .
let @xmath126 be a hyperline , i.e. a subspace of dimension @xmath98 spanned by elements of @xmath19 .
let @xmath127 be the orthogonal complement of @xmath128 in @xmath10 .
then @xmath129 is a set of non - zero vectors in @xmath21 , corresponding to the hyperline sequence @xmath20 in a hyperline @xmath97 .
the image of @xmath130 under the orthogonal projection onto @xmath127 yields a set of non - zero vectors in @xmath131 corresponding to the term @xmath99 in @xmath97 .
axiom ( h1 ) simply means that @xmath19 is a disjoint union of @xmath129 and @xmath130 .
axiom ( h2 ) corresponds to the fact that @xmath128 is uniquely determined by any @xmath9tuple of elements of @xmath129 in general position .
axiom ( h3 ) is the steinitz
mclane exchange lemma , stating that one can replace any vector in a base by some vector of any other base .
axiom ( h4 ) ensures that if @xmath132 $ ] is a positively oriented base of @xmath4 than so is any oriented simplex that is equivalent to @xmath117 $ ] ; this is a part of theorem [ thm : hls - chi ] below .
axiom ( h4 ) is related to the `` consistent abstract sign of determinant '' in @xcite .
it means that if @xmath6 points span an @xmath7simplex , then any subset of @xmath98 points spans a hyperline , and the orientation of the @xmath7simplex does not depend on the hyperline on which we consider the @xmath6 points .
we recall in this section the chirotope axioms for oriented matroids ( see @xcite , p. 126,p . 138 , @xcite ) .
let @xmath32 be as in the preceding section .
we denote by @xmath133 the set of all oriented @xmath44simplices in @xmath34 .
a * chirotope * @xmath134 of rank @xmath6 over @xmath34 is a map @xmath135 , so that the following holds .
* for any @xmath136 , there are @xmath137 with @xmath138)\not=0 $ ] . * for any @xmath139 holds @xmath140 , and if @xmath141 then @xmath142 . *
if @xmath143)\not=0 $ ] and @xmath144)\not=0 $ ] then there is some @xmath145 with @xmath146)\not=0 $ ] . * if @xmath147 so that @xmath148)\cdot \chi([x_1,\dots , x_{r-1},y_2 ] ) & \ge & 0\text { and}\\ \chi([x_1,\dots , x_{r-2},y_2,x_r ] ) \cdot \chi([x_1,\dots , x_{r-1},\overline{y_1}])&\ge & 0 , \end{aligned}\ ] ] then @xmath143)\cdot\chi([x_1,\dots , x_{r-2},y_1,y_2])\ge 0 $ ] .
a simplex @xmath149 is a * positively oriented base * of @xmath134 if @xmath150 . by axiom ( c2 ) , @xmath134 is completely described by the set of its positively oriented bases . in order to give axioms ( c1)(c4 ) a geometric meaning , we show how to construct a chirotope of rank @xmath6 from a vector configuration in @xmath151 .
let @xmath152 be a set of non - zero vectors spanning @xmath10 .
let @xmath153 be defined by @xmath154 and @xmath155 for all @xmath156 .
we get a map @xmath157 by setting * @xmath143)=+1 $ ] if the determinant @xmath158 $ ] is positive + ( i.e. if @xmath159 is a positive base of @xmath10 ) , * @xmath143 ) = -1 $ ] if @xmath158 $ ] is negative , and * @xmath143 ) = 0 $ ] if @xmath159 are linearly dependent .
we show that @xmath134 is a chirotope .
axiom ( c1 ) holds since @xmath19 spans @xmath10 .
axiom ( c2 ) follows from the symmetry of determinants .
axiom ( c3 ) is the steinitz
mclane exchange lemma , as axiom ( h3 ) above .
axiom ( c4 ) sais that the signs of oriented bases as defined by @xmath134 do not contradict the three summand gramann
plcker relation .
note that we do distinguish between @xmath134 and @xmath160 , although this is not usual in the literature .
our reason comes from geometry .
we believe if one deals with oriented objects ( vectors , oriented hyperspheres , etc . ) in @xmath59 or @xmath161 , then it is consequent to distinguish not only the two different orientations of any object , but also the two different orientations of @xmath59 or @xmath161 . by the following theorem ,
the notions of chirotopes and hyperline sequences are equivalent .
this connects our concept of hyperline sequences with other ways to look at oriented matroids .
the cyclic structure of a hyperline captures many instances of the @xmath0term gramann plcker relations at once .
therefore the proof of the theorem becomes rather long and tedious .
but we believe that this price is worth to pay .
namely , it is easier to deal with a few cyclic structures than with a multitude of gramann plcker relations , and therefore it is algorithmically more convenient and efficient to encode the structure of oriented matroids in hyperline sequences rather than in chirotopes .
[ thm : hls - chi ] the set of positively oriented bases of a hyperline sequence of rank @xmath6 over @xmath34 is the set of positively oriented bases of a chirotope of rank @xmath6 over @xmath34 , and vice versa .
let @xmath4 be a hyperline sequence of rank @xmath6 over @xmath34 , and let @xmath117 $ ] be a positively oriented base of @xmath4 . since the cyclic orders @xmath162 and @xmath163 are equal up to a shift , @xmath164 $ ] is a positively oriented base of @xmath4 as well .
this together with axiom ( h4 ) and an induction on the rank implies that all oriented simplices equivalent to @xmath117 $ ] are positively oriented bases of @xmath4 , hence , yields axiom ( c2 ) .
axiom ( c1 ) follows from axiom ( h1 ) , axiom ( c2 ) and induction on the rank .
the axioms ( h3 ) and ( c3 ) are equivalent . in the next paragraphs
we deduce axiom ( c4 ) from the cyclic order of hyperlines .
we use axiom ( c2 ) , that is already proven , but we do not mention any application explicitly . let @xmath147 satisfy the first two inequalities in axiom ( c4 ) .
the third inequality is to prove .
we can assume that @xmath165 defines some hyperline @xmath100 with period length @xmath77 , since otherwise axiom ( c4 ) is trivial .
for simplicity , we write @xmath166 $ ] in place of @xmath167)$ ] . if @xmath168=[x_{r-1},y_2]=-1 $ ] then we replace @xmath169 with @xmath170 , @xmath170 with @xmath169 , @xmath171 with @xmath172 and @xmath173 with @xmath174 .
this changes the signs of the factors in the first inequality , whereas the other inequalities remain unchanged .
similarly , we can assume without loss of generality that both factors of the second inequality and @xmath175 $ ] are non - negative .
let @xmath176 so that @xmath177 by shifting the hyperline , we assume that @xmath178
. then we have @xmath179 if @xmath175=0 $ ] then the third inequality is satisfied , and axiom ( c4 ) is proven .
if @xmath175=1 $ ] then @xmath180 .
assume that @xmath181=-1 $ ] , thus @xmath182 . with the natural order on @xmath183 we find @xmath184 .
this is a contradiction to @xmath185 .
hence @xmath104\ge 0 $ ] , which finishes the proof of axiom ( c4 ) . \2 .
conversely , let @xmath134 be a chirotope of rank @xmath6 over @xmath34 .
we wish to construct a hyperline sequence @xmath4 with the same positively oriented bases .
note that in our proof we do not mention all applications of axiom ( c2 ) explicitly .
if @xmath134 is of rank @xmath8 , then we define @xmath4 as the set of all @xmath186 with @xmath187 ) = + 1 $ ] .
since @xmath187 ) = -\chi([\overline x])$ ] and since @xmath188 by axiom ( c1 ) , @xmath4 is a hyperline sequence of rank @xmath8 , and it has the desired positively oriented bases by construction .
let @xmath134 be of rank @xmath70 .
fix some element @xmath39 .
we iteratively define a sequence @xmath189 of subsets of @xmath35 as follows .
we start with @xmath190)=1 & \text{\hspace{-1em } , and for all } y\in \mathbf e\\ & & \text{\hspace{-1em}with } \chi([e , y ] ) = 1\text { holds } \chi([x , y ] ) \ge 0\big\}. \end{aligned}\ ] ] for @xmath191 , pick some @xmath192 .
define @xmath193)=1 & \text{\hspace{-1em } , and for all } y\in \mathbf e\\ & & \text{\hspace{-1em}with } \chi([x^{(a)},y ] ) = 1\text { holds } \chi([x , y ] ) \ge 0\big\}. \end{aligned}\ ] ] we prove that @xmath194 ( and similarly @xmath195 ) is not empty . under the assumption @xmath196 , we inductively define elements @xmath197 as follows . by axiom ( c1 ) , there is some @xmath198 with @xmath199)=1 $ ] . for @xmath200 , since @xmath201 there is some @xmath202 with @xmath203)=1 $ ] and @xmath204)=-1 $ ] . since @xmath35 is finite , we find an index @xmath205 so that @xmath206 with @xmath207 ( note that @xmath208)=1 $ ] ) .
we choose @xmath25 minimal , which implies @xmath209)\ge 0 $ ] .
it follows @xmath210)\cdot \chi([x_i , x_{i-1 } ] ) & = & 1 \\ \chi([x^{(a)},x_i])\cdot \chi([x_{i-1},x_k ] ) & \ge&0 \\
\chi([x^{(a)},x_{i-1}])\cdot \chi([x_i , x_k ] ) & = & -1 , \end{aligned}\ ] ] which contradicts axiom ( c4 ) .
hence @xmath211 .
since @xmath212)=1 $ ] for @xmath213 , we have @xmath214)=-1 $ ] by axiom ( c2 ) , thus @xmath215 . in conclusion , @xmath194 is a hyperline sequence of rank @xmath8 .
we show that for @xmath216 and for all @xmath217 and all @xmath218 we have @xmath219 ) = \chi([y_2,x_2])$ ] .
let @xmath220 be the element that appears in the definition of @xmath79 . by definition @xmath221)=\chi([y_1,y_2])=1 $ ] .
hence we have @xmath222)\ge 0 $ ] and @xmath223)\ge 0 $ ] , thus @xmath222)=0 $ ] .
assume that @xmath219 ) \not= \chi([y_2,x_2])$ ] .
without loss of generality , we assume @xmath219)=-1 $ ] and @xmath224)\ge 0 $ ] . we obtain @xmath225)\cdot\chi([x_1,y_2 ] ) & = & 0\\ \chi([y_1,x_1])\cdot\chi([y_2,x_2 ] ) & \ge & 0\\ \chi([y_1,y_2])\cdot\chi([x_1,x_2 ] ) & = & -1 .
\end{aligned}\ ] ] this is a contradiction to axiom ( c4 ) , hence @xmath219 ) = \chi([y_2,x_2])$ ] .
this implies that @xmath194 does not depend on the choice of @xmath192 .
let @xmath226 be the least number so that @xmath227 . since @xmath194 is independent of the choice of @xmath192 , it follows @xmath228 for all @xmath229 . moreover , if @xmath230 is the least number with @xmath231 , we have @xmath232 for all @xmath229 . in order to prove that @xmath233 yields a hyperline sequence of rank 2 over @xmath34 with period length @xmath234 , it remains to show that @xmath35 is a disjoint union of @xmath235 . by choosing @xmath236 minimal and by the independence of @xmath194 from the choice of @xmath237
, it follows that @xmath238 are disjoint .
let @xmath239 .
there is some @xmath218 with @xmath219)=1 $ ] , by axiom ( c1 ) . since @xmath240 ) = 1 $ ] by definition , axiom ( c3 ) implies that @xmath241)$ ] or @xmath242)$ ] does not vanish . therefore and by axiom ( c2 )
, there is a least index @xmath25 so that @xmath243)=1 $ ] and @xmath244)\le 0 $ ] .
let @xmath245 with @xmath246)=1 $ ] . by definition of @xmath247 , we have @xmath248)\cdot \chi([x^{(i+1)},x_1 ] ) & \ge&0 , \\
\chi([y , x^{(i+1)}])\cdot \chi([x_1,x^{(i ) } ] ) & \ge&0 \end{aligned}\ ] ] and @xmath249 ) = 1 $ ] , and therefore @xmath250)\ge 0 $ ] by axiom ( c4 ) .
but this means @xmath251 . in conclusion ,
hence @xmath233 yields a hyperline sequence of rank 2 over @xmath34 , and by construction it has the same positively oriented bases than @xmath134 . finally , we come to the case of rank @xmath253 .
let @xmath254 be so that there are two elements @xmath255 with @xmath256)=1 $ ] .
such an @xmath9tuple exists by axiom ( c1 ) .
it is easy to verify that @xmath257\in \delta_1(e ) |\ ; \chi([x_1,\dots , x_{r-2 } , y_{r-1 } , y_r])=1\}\ ] ] is the set of positively oriented bases of a chirotope of rank 2 over some subset of @xmath34 , thus , of a hyperline sequence @xmath258 of rank 2 . by application of axiom ( c2 )
, one can also show that @xmath259\in \delta_{r-3}(e ) & | & \chi([y_1,\dots , y_r])=1\\ & & \text { for all } [ y_{r-1},y_r]\in c(x_1,\dots , x_{r-2})\ } \end{aligned}\ ] ] is the set of positively oriented bases of a chirotope of rank @xmath98 over some subset of @xmath34 , with @xmath260\in b(x_1,\dots , x_{r-2})$ ] . by induction , it is equivalent to a hyperline sequence @xmath261 of rank @xmath98 .
we collect all pairs @xmath262 to form a set ( not a multi - set ) @xmath4 , where @xmath263 .
it has the same positively oriented bases as @xmath134 , by construction .
it remains to show that @xmath4 is a hyperline sequence of rank @xmath6 over @xmath34 .
the most difficult to prove is axiom ( h1 ) .
let @xmath264 and @xmath265 . by definition , in an oriented simplex over @xmath34
any element of @xmath266 occurs at most once .
this implies that @xmath267 .
let @xmath268 .
it remains to show that @xmath269 , hence @xmath270 .
let @xmath260 $ ] and @xmath104 $ ] be positively oriented bases of @xmath20 and @xmath99 , thus @xmath271)=1 $ ] . by multiple application of axioms ( c2 ) and ( c3 ) , there is an index @xmath25 so that @xmath272)\not= 0 $ ] .
we have @xmath273 since @xmath274 .
we claim that for all @xmath175\in c(x_1,\dots , x_{r-2})$ ] holds @xmath275 ) = \chi([x_1,\dots , x_{i-1},e , x_{i+1},\dots , x_r]).\ ] ] by replacing @xmath276 with @xmath277 if necessary , we can assume that @xmath278)=1 $ ] . by axiom ( c2 ) , it suffices to consider the case @xmath279 . for simplicity , we denote @xmath166 $ ] for @xmath280)$ ] .
we have @xmath281=[x_i , y_2]=[e , x_r]=1 $ ] , and @xmath282=0 $ ] since @xmath274 .
if @xmath283=0 $ ] , then @xmath284\cdot [ e , x_i ] & = & 0\\ { } [ x_r , x_i]\cdot [ e , y_2 ] & = & 0\\ { } [ x_r , e]\cdot [ x_i , y_2 ] & = & -1 , \end{aligned}\ ] ] in contradiction to axiom ( c4 ) . hence @xmath283\not=0 $ ] .
since @xmath285\cdot [ e , x_r ] & = & 1\\ { } [ x_i , e]\cdot [ x_r , y_2 ] & = & 0\\ \end{aligned}\ ] ] and @xmath281=1 $ ] , it follows @xmath283=1 $ ] from axiom ( c4 ) .
this proves our claim .
the claim implies that @xmath269 , which finishes the proof of axiom ( h1 ) .
we check the remaining axioms .
since @xmath143 ) = \chi([x_1,\dots , x_{r-3},\overline{x_{r-2 } } , x_r , x_{r-1}])$ ] by axiom ( c2 ) , we have @xmath286 and this implies axiom ( h2 ) . axiom ( h3 ) is axiom ( c3 ) .
finally , axiom ( h4 ) is a special case of axiom ( c2 ) .
thus , @xmath4 is a hyperline sequence .
our proof of the topological representation theorem in sections [ sec : flbase ] and [ sec : flgeneral ] is by induction on the rank and the number of elements of a chirotope or hyperline sequence . in the rest of this section
, we expose the basic techniques for this induction .
let @xmath134 be a chirotope of rank @xmath6 over @xmath34 , and let @xmath287 .
we define the map @xmath288 as the restriction of @xmath134 to @xmath289 .
we call @xmath290 the * deletion * of @xmath291 in @xmath134 . in general , @xmath290 is not a chirotope .
however if the rank of @xmath134 is smaller than @xmath292 , then we find by the following lemma an element that can be deleted such that @xmath293 is a chirotope of the same rank .
[ lem : fullrank ] let @xmath134 be a chirotope of rank @xmath6 over @xmath34 .
if @xmath294 then there is an @xmath295 so that @xmath296 is a chirotope of rank @xmath6 over @xmath297 . by axioms ( c1 ) and
( c2 ) , there are @xmath298 with @xmath299)\not=0 $ ] .
since @xmath294 , we can pick some @xmath300 .
it is obvious that @xmath296 satisfies axioms ( c2 ) , ( c3 ) and ( c4 ) .
it remains to show that @xmath296 satisfies axiom ( c1 ) .
let @xmath301 .
by axiom ( c1 ) , there are @xmath302 with @xmath138)\not=0 $ ] .
if @xmath303 then by axiom ( c2 ) we can assume that @xmath304 .
since @xmath299)\not=0 $ ] from the preceding paragraph , we find some @xmath305 so that @xmath306 ) \not= 0 $ ] , by axiom ( c3 ) . since @xmath307 \in \delta_{r-1}(e\setminus \{i\})$ ] , the map @xmath308 satisfies axiom ( c1 ) and is therefore a chirotope .
let @xmath309 , with @xmath310 and @xmath311 .
we define @xmath312 as the set of all @xmath313 for which there exist @xmath314 so that @xmath315)\not= 0 $ ] .
the map @xmath316 is then defined by @xmath317 ) = \chi ( [ e_1,\dots , e_r])$ ] , for all @xmath318 \in \delta_{r-1}(e / r)$ ] .
it is obvious that @xmath319 satisfies axioms ( c2 ) , ( c3 ) and ( c4 ) .
it satisfies axiom ( c1 ) by definition of @xmath312 .
hence , @xmath319 is a chirotope of rank @xmath320 over @xmath312 , that is called the * contraction * of @xmath134 on @xmath291 .
finally , let @xmath4 be a hyperline sequence of rank @xmath6 over @xmath34 .
let @xmath134 be the chirotope that corresponds to @xmath4 .
for any @xmath287 so that @xmath290 ( resp .
@xmath319 ) is a chirotope , we define the * deletion * @xmath321 of @xmath291 in @xmath4 ( resp .
the * contraction * @xmath322 of @xmath4 on @xmath291 ) as the hyperline sequence associated to @xmath290 ( resp .
@xmath319 ) .
in the preceding sections , we extracted combinatorial data from vector arrangements and turned properties of these data into axioms , yielding hyperline sequences and chirotopes . in this section , we will generalize vector arrangements in a geometric way .
the main aim of this paper is to prove the equivalence of these geometric structures with hyperline sequences .
the idea is as follows .
let @xmath323 be the unit sphere , and let @xmath324 be a finite multiset of non - zero vectors .
any @xmath325 yields a pair @xmath326 of points in @xmath327 .
this is dual to a hypersphere @xmath328 , namely the intersection of @xmath327 with the hyperplane @xmath329 that is perpendicular to @xmath330 .
we put an orientation on @xmath331 , so that a positive base of @xmath331 together with @xmath330 is a positive base of @xmath151 .
this induces an orientation on @xmath332 . in conclusion , @xmath19 is dual to an arrangement of oriented hyperspheres in @xmath327 .
the geometric idea is now to consider arrangements of oriented embedded spheres of codimension @xmath8 in @xmath327 , intersecting each other similarly to hyperspheres , though not being hyperspheres in general .
we formalize this idea .
let @xmath161 denote the @xmath44dimensional oriented sphere @xmath333 and let @xmath334 denote the closed @xmath44dimensional ball .
a submanifold @xmath335 of codimension @xmath27 in a @xmath44dimensional manifold @xmath336 is * tame * if any @xmath337 has an open neighbourhood @xmath338 such that there is a homeomorphism @xmath339 sending @xmath340 to @xmath341 . an * oriented pseudosphere * @xmath342 is a tame embedded @xmath343dimensional sphere with a choice of an orientation .
obviously any oriented hypersphere is an oriented pseudosphere .
let @xmath344 be an embedding with image @xmath345 , inducing the desired orientation of @xmath345 . by a result of m. brown @xcite ,
the image of @xmath346 is tame if and only if @xmath346 can be extended to an orientation preserving embedding @xmath347 \to { \mathbb s}^d \hbox { with } \psi(\cdot)=\tilde\psi(\cdot,0).\ ] ] the image of an oriented pseudosphere @xmath345 under a homeomorphism @xmath348 is again an oriented pseudosphere , since the defining embedding @xmath349 can be extended to @xmath350 \to { \mathbb s}^d$ ] . by the generalized schnflies theorem , that was also proven by m. brown @xcite , @xmath351 is a disjoint union of two open balls whose closures are closed balls .
we call the connected component of @xmath351 containing @xmath352 ( resp .
@xmath353 ) the * positive side * @xmath354(resp .
* negative side * @xmath355 ) of @xmath345 .
the following definition of arrangements of oriented pseudospheres is similar to @xcite , p. 227 .
recall @xmath356 .
let @xmath357 be not necessarily distinct oriented pseudospheres , ordered according to their indices .
assume that the following conditions hold : * @xmath358 is empty or homeomorphic to a sphere , for all @xmath359 .
* let @xmath359 and @xmath360 with @xmath361 .
then @xmath362 is a pseudosphere in @xmath363 , and @xmath364 and @xmath365 are both non - empty .
then the ordered multiset @xmath366 is an * arrangement of oriented pseudospheres * over @xmath367 .
the arrangement is called * of full rank * if the intersection of its members is empty .
we omitted axiom ( a3 ) from @xcite , p. 227
, since it follows from the other two axioms .
we remark that in @xcite , arrangements of full rank are called _
essential_. obviously if @xmath368 are oriented hyperspheres then @xmath366 is an arrangement of oriented pseudospheres .
at the end of this section , we will characterize arrangements of oriented pseudospheres by a single axiom .
two ordered multisets @xmath369 and @xmath370 of oriented pseudospheres in @xmath161 are * equivalent * if there is an orientation preserving homeomorphism @xmath371 sending @xmath372 to @xmath373 and @xmath374 to @xmath375 , simultaneously for all @xmath360 .
we do not allow renumbering of the pseudospheres .
since the image of a pseudosphere under a homeomorphism @xmath376 is a pseudosphere , it is easy to observe that if @xmath377 is an arrangement of oriented pseudospheres , then so is @xmath378 .
we construct non - equivalent arrangements @xmath379 of @xmath380 oriented pseudospheres @xmath381 of full rank as follows . for @xmath382 ,
set @xmath383 in @xmath384 , define @xmath385 . in @xmath386 , define @xmath387 . in both @xmath384 and@xmath386 , define @xmath388 , for @xmath389 .
@xmath384 and @xmath390 are of full rank .
if @xmath348 is any homeomorphism that fixes @xmath391 setwise and maps @xmath392 to @xmath393 , then @xmath394 is orientation reversing .
thus , @xmath384 is not equivalent to @xmath386 . in the preceding section ,
we have defined deletion and contraction of chirotopes and hyperline sequences .
there is a similar notion for arrangements of oriented pseudospheres , as follows .
fix an arrangement @xmath395 of oriented pseudospheres in @xmath161 . for any @xmath396
, we obtain an arrangement @xmath397 of oriented pseudospheres over @xmath398 in @xmath161 ( axioms ( a1 ) and ( a2 ) are easy to verify ) .
we denote this arrangement by @xmath399 and call it the * deletion * of @xmath291 in @xmath400 . in general , @xmath399 is not of full rank , even if @xmath400 is .
it is intuitively clear from axiom ( a2 ) , that for @xmath359 one gets an arrangement of oriented pseudospheres on @xmath363 , induced by @xmath400 . in the following iterative definition of this induced arrangement
, the orientation of @xmath363 requires some care .
let @xmath401 , and let @xmath402 be a tame embedding defining @xmath403 with the correct orientation .
denote @xmath404 , for @xmath360 .
we obtain an ordered multiset @xmath405 of oriented pseudospheres in @xmath406 , where @xmath407 .
axioms ( a1 ) and ( a2 ) are easy to verify , thus , @xmath408 is an arrangement of oriented pseudospheres over a subset of @xmath409
. this construction can be iterated .
let @xmath359 so that @xmath410 .
list the elements of @xmath291 in ascending order , @xmath411 .
the * contraction * of @xmath400 in @xmath291 is then the arrangement of oriented pseudospheres @xmath412 note that @xmath413 and @xmath414 are related by an orientation _ reversing _
homeomorphism @xmath415 .
it is easy to see that @xmath416 is of full rank if and only if @xmath400 is of full rank .
an arrangement @xmath417 of oriented pseudospheres in @xmath161 yields a cellular decomposition of @xmath161 , as follows . for any subset @xmath418
, we consider the parts of the intersection of the pseudospheres with label in @xmath419 that are not contained in pseudospheres with other labels , @xmath420 by the next theorem , the connected components of @xmath421 are topological cells if @xmath400 is of full rank .
hence , it makes sense to refer to a connected component of @xmath421 as a * cell of @xmath400 with label @xmath419*. the * @xmath44skeleton * @xmath422 of @xmath400 is the union of its cells of dimension @xmath423 . by axiom ( a1 ) , @xmath424 is empty or a sphere , for @xmath359 .
thus , for any @xmath425-dimensional cell of @xmath400 exists exactly one other cell of @xmath400 with the same label , corresponding to the two points in @xmath426 .
the statement that the cells of an arrangement of oriented pseudospheres of full rank are in fact topological cells is essential for the proof of the topological representation theorem in sections [ sec : flbase ] and [ sec : flgeneral ] .
[ thm : cells ] let @xmath427 be an arrangement of oriented pseudospheres over @xmath367 .
any connected component of @xmath428 is a @xmath44dimensional cell whose closure is a closed ball .
if @xmath400 is of full rank , then for any @xmath418 , any connected component of @xmath421 is an open cell whose closure is a closed ball .
first , we prove that the theorem holds for @xmath428 , by induction on the number @xmath429 of elements of @xmath400 .
since @xmath430 , we have @xmath431 .
the base case @xmath432 is the generalized schnflies theorem @xcite , stating that an embedded tame @xmath343sphere in @xmath161 is the image of a hypersphere under a homeomorphism @xmath371 . in particular , the complement of a pseudosphere is a disjoint union of two @xmath44dimensional cells whose closures are balls . let @xmath433 , and let @xmath434 be the closure of a connected component of @xmath435 . by induction on @xmath429 , it is a cell of dimension @xmath44 .
the connected components of @xmath436 correspond to the closures of connected components of @xmath437 . by induction hypothesis ,
applied to the arrangement @xmath438 of @xmath439 pseudospheres in @xmath406 , these are closed balls of dimension @xmath440 whose boundary lies in @xmath441 . since @xmath442 is tame , @xmath436 is tame in @xmath434
hence , it follows from the generalized schnflies theorem @xcite that the closure of any connected component of @xmath443 is a ball of dimension @xmath44 . in conclusion ,
the connected components of @xmath428 are cells of dimension @xmath44 whose closures are balls .
it remains to prove the theorem in the full rank case with @xmath444 .
we can assume @xmath445 .
since @xmath400 is of full rank , @xmath446 , hence @xmath447 . by axiom ( a1 ) , @xmath448 is homeomorphic to some sphere @xmath449 .
there is an @xmath450 with @xmath451 and @xmath452 .
the set @xmath421 is mapped to @xmath453 by the restriction of a homeomorphism @xmath454 .
it has already been proven that the closure of any connected component of @xmath453 is a ball .
thus , the closure of any connected component of @xmath421 is a ball , as well .
we remark that the preceding theorem becomes wrong by dropping the hypothesis that pseudospheres are tame .
in fact , there are wild @xmath70-spheres in @xmath455 ( e.g. the famous horned sphere of alexander ) , whose complement is not a union of cells .
we will consider arrangements of oriented embedded spheres ( not necessarily tame ) in section [ sec : wild ] and prove that these wild arrangements have the same combinatorics than tame arrangements .
let @xmath456 be an ordered multiset of oriented pseudospheres in @xmath161 .
for @xmath359 , denote @xmath457 . in the remainder of this section , we show that one can replace axioms ( a1 ) and ( a2 ) by the following single * axiom ( a)*. * let @xmath359 so that @xmath458 for any proper subset @xmath459 of @xmath291
. then , @xmath460 is equivalent to an arrangement of @xmath461 oriented hyperspheres in @xmath161 .
[ thm : psa ] an ordered multiset @xmath462 of oriented pseudospheres in @xmath161 satisfies axiom ( a ) if and only if it is an arrangement of oriented pseudospheres .
first , we assume that @xmath400 satisfies axiom ( a ) and prove that @xmath400 is an arrangement of oriented pseudospheres .
let @xmath463 , and assume that @xmath464 .
replacing @xmath291 by a subset if necessary , we can assume that @xmath458 for any proper subset @xmath459 of @xmath291 .
by axiom ( a ) , @xmath465 is equivalent to an arrangement of oriented hyperspheres
. thus , @xmath363 is homeomorphic to a sphere , and axiom ( a1 ) holds .
let @xmath360 with @xmath466 ; we wish to prove ( a2 ) .
there is some @xmath467 with @xmath468 , so that @xmath469 for any proper subset @xmath459 of @xmath470 . since by axiom ( a ) , @xmath471 is equivalent to an arrangement of oriented hyperspheres , it follows @xmath472 since both @xmath363 and @xmath473 are spheres by axiom ( a1 ) ( that has already been proven ) and since @xmath466 by hypothesis , it follows @xmath474 thus , @xmath475 is a pair of speres of the same dimension @xmath476 , hence , @xmath477 . since @xmath471 is equivalent to an arrangement of oriented hyperspheres , axiom ( a2 )
holds for @xmath478 and @xmath479 . in conclusion
, @xmath400 is an arrangement of oriented pseudospheres .
secondly , we assume that @xmath400 is an arrangement of oriented pseudospheres and prove by induction on @xmath480 that ( a ) holds . if @xmath481 , then ( a ) is nothing but the generalized schnflies theorem @xcite . in the general case ,
let @xmath463 so that @xmath458 for any proper subset @xmath459 of @xmath291 , and let @xmath482 .
it follows easily that @xmath483 for any proper subset @xmath459 of @xmath484 .
thus , by axiom ( a ) we can assume that @xmath485 is an arrangement of oriented hyperspheres .
note that the connected components of @xmath486 are topological cells , for any proper subset @xmath419 of @xmath484 .
let @xmath487 be a hypersphere that does not contain @xmath488 .
our aim is to transform @xmath479 into @xmath128 , fixing @xmath485 cellwise , which implies axiom ( a ) for @xmath465 . by axiom ( a2 ) and by the generalized schnflies theorem ,
@xmath489 can be mapped to @xmath490 by some orientation preserving homeomorphism @xmath491 , the homeomorphism can be extended to a homeomorphism @xmath371 fixing all cells of @xmath485 , by the cone construction @xcite . we now proceed with transforming @xmath479 in cells of @xmath492 of higher dimension .
let @xmath459 be a proper subset of @xmath484 .
by induction on @xmath493 , we can assume that @xmath494 , for all @xmath495 .
let @xmath496 be the closure of a cell of @xmath497 with @xmath498 .
it follows from axiom ( a2 ) that @xmath499 is a tame ball in @xmath500 . as a consequence of the generalized schnflies theorem and since @xmath501 , we can map @xmath502 to @xmath503 by an orientation preserving homeomorphism @xmath504 that is the identity on @xmath505 .
it can be extended to a homeomorphism @xmath506 by the cone construction , fixing all cells of @xmath497 .
thus , we can transform @xmath507 into @xmath508 . finally , when we achieve @xmath509 , we have @xmath510 .
therefore , @xmath460 is equivalent to an arrangement of oriented hyperspheres .
this proves ( a ) , as claimed .
[ cor : n = d+1 ] if an arrangement @xmath400 of oriented pseudospheres in @xmath511 is of full rank and @xmath512 , then @xmath400 is equivalent to @xmath384 or @xmath386 .
let @xmath513 be minimal so that @xmath465 is of full rank . by the preceding theorem
, @xmath465 is equivalent to an arrangement of oriented hyperspheres of full rank .
such an arrangement consists of at least @xmath380 hyperspheres , hence @xmath514 and @xmath515 .
any arrangement of @xmath380 oriented hyperspheres of full rank is dual to an arrangement of @xmath380 unit vectors in @xmath516 that span @xmath516 .
since @xmath517 acts transitively on those vector arrangements , it follows that @xmath400 is equivalent to @xmath518 or @xmath386 , depending on its orientation .
the aim of our paper is to prove a topological representation theorem for hyperline sequences , i.e. , to establish a one - to - one correspondence between hyperline sequences and equivalence classes of arrangements of oriented pseudospheres . in this section ,
we settle one direction of this correspondence .
we associate to any arrangement of oriented pseudospheres a hyperline sequence , compatible with deletions and contractions .
we first expose the geometric idea . by a * cycle * of an arrangement @xmath400 of oriented pseudospheres in @xmath161 , we mean an embedded circle @xmath519 that is the intersection of some elements of @xmath400 .
let @xmath520 be a cycle of @xmath400 with a choice of an orientation .
it corresponds to a hyperline @xmath97 , as follows .
the positively oriented bases of @xmath20 correspond to @xmath343tuples of pseudospheres containing @xmath520 so that the system of positive normal vectors of the pseudospheres together with a positive tangential vector of @xmath520 forms a direct base of the oriented vector space @xmath59 .
the set of all elements of @xmath400 containing @xmath520 corresponds to @xmath109 .
the @xmath425-dimensional cells of @xmath400 on @xmath520 occur in a cyclic order , corresponding to the cyclic order of @xmath99 .
let @xmath521 .
if in a point of @xmath522 the cycle @xmath520 passes from the negative side of @xmath523 to the positive side , then we have an element @xmath524 in the corresponding atom of @xmath97 ( see figure [ fig : positivevertex ] ) .
further , in the second point of @xmath522 , the cycle @xmath520 passes from the positive to the negative side , yielding an element @xmath525 in the atom that is opposite to the first atom .
the meaning of axiom ( h3 ) in the setting of arrangements of oriented pseudospheres is that any two cycles of @xmath400 have non - empty intersection .
figure [ fig : g0 ] provides a visualization of axiom ( h4 ) in an arrangement of three oriented pseudospheres in @xmath526 : if we get from the pseudosphere @xmath527 the cyclic order @xmath528 , then we get the cyclic order @xmath529 from @xmath25 and the cyclic order @xmath530 from @xmath531 .
we formalize this idea in the rest of this section .
let @xmath532 be an arrangement of @xmath429 oriented pseudospheres of full rank in @xmath161 .
our aim is to associate to @xmath400 a hyperline sequence @xmath533 of rank @xmath380 over @xmath367 . by the equivalence of hyperline sequences and chirotopes ,
established in theorem [ thm : hls - chi ] , this also allows to associate a chirotope @xmath534 to @xmath400 , with the same positively oriented bases . if @xmath535 then define @xmath536 which is obviously a hyperline sequence over @xmath367 of rank 1 . in the case
@xmath537 , the orientation of @xmath538 yields a cyclic order @xmath539 on the points of @xmath540 . for @xmath541 , we define @xmath542 by 1 .
@xmath543 if @xmath544 and , along the cyclic orientation of @xmath538 , one passes in @xmath545 from @xmath546 to @xmath547 , and 2 .
@xmath548 if @xmath544 and one passes in @xmath545 from @xmath547 to @xmath546 .
it is easy to check that @xmath549 yields a hyperline sequence of rank 2 over @xmath367 . in the case
@xmath550 , let @xmath551 be an oriented cycle of @xmath400 .
let @xmath359 be the indices of oriented pseudospheres containing @xmath552 .
there is a hyperline sequence @xmath553 of rank 1 over @xmath291 , with @xmath554 ( resp .
@xmath555 ) if the orientation of @xmath523 coincides ( resp . does not coincide ) with the orientation of @xmath552 . as in the preceding paragraph
, we obtain a hyperline sequence @xmath556 of rank 2 over @xmath557 .
we collect the pairs @xmath558 to form a set @xmath559 , where @xmath552 runs over all oriented cycles of @xmath400
. axioms ( h1 ) and ( h2 ) are obvious for @xmath533 .
to prove axiom ( h3 ) , let @xmath560 $ ] and @xmath561 $ ] be two positively oriented bases of @xmath533 . by definition , the pseudospheres @xmath562 , @xmath563 ,
@xmath564 have no point in common .
in particular ,
one of them , say , @xmath562 , does not contain @xmath565 .
thus @xmath562 intersects @xmath566 transversely in @xmath567 , hence @xmath568 $ ] or @xmath569 $ ] is a positively oriented base of @xmath559 .
it remains to prove axiom ( h4 ) .
here we use the jordan schnflies theorem @xcite , stating that the complement of an embedded @xmath8sphere in @xmath526 is a disjoint union of two discs .
this holds even without the assumption of tameness . with this in mind , axiom ( h4 ) can be read off from figure [ fig : g0 ] . in the case
@xmath570 , let @xmath552 be an oriented cycle of @xmath571 .
let @xmath572 .
as in the preceding paragraph , the cyclic orientation of @xmath552 induces a cyclic order of the oriented points @xmath573 yielding a hyperline sequence @xmath556 of rank 2 over @xmath574 . since @xmath400 is of full rank ,
there are @xmath575 so that @xmath576 is a sphere of dimension @xmath577 disjoint to @xmath552 .
we may assume that @xmath556 yields the cyclic order @xmath578 , by changing the roles of @xmath25 and @xmath579 if necessary .
then @xmath580 is an arrangement of oriented pseudospheres in @xmath581 of full rank over @xmath582 , and by induction it corresponds to a hyperline sequence @xmath583 of rank @xmath440 over @xmath582 .
we show that @xmath583 does not depend on the choice of @xmath584 . by symmetry of @xmath25 and @xmath579
, it suffices to pick @xmath585 so that @xmath586 is a sphere of dimension @xmath577 disjoint to @xmath552 and @xmath556 yields the cyclic order @xmath587 , and to show that @xmath588 .
we will prove that the cyclic order of signed points on oriented cycles of @xmath589 coincides with those of @xmath590 .
it is clear that this implies @xmath588 .
let @xmath591 so that @xmath592 is a cycle of @xmath589 .
we consider the @xmath70sphere @xmath593 . both @xmath594 and @xmath595
are embedded @xmath8spheres in @xmath345 that are either equal or intersect in two points .
both @xmath596 and @xmath597 have the positive ( resp .
negative ) point of @xmath598 on their positive ( resp .
negative ) side .
thus , again using the jordan schnflies theorem , the situation is as in figure [ fig : parallel ] .
let @xmath500 be a connected component of @xmath599 disjoint from @xmath552 .
let @xmath600 .
since @xmath601 is equal to @xmath602 or @xmath603 intersects @xmath604 transversaly in @xmath605 , it follows that @xmath606 and @xmath607 are parallel arcs .
hence , the cyclic order of the four signed points of @xmath608 coincides with the order on @xmath609 .
this proves our claim @xmath588 .
since @xmath583 does not depend on the choice of @xmath25 and @xmath579 , we denote @xmath610 .
now , we define @xmath533 as the set of all pairs @xmath611 , where @xmath552 runs over all oriented cycles of @xmath400 , each cycle occuring in both orientations .
it remains to show that @xmath533 is a hyperline sequence of rank @xmath380 over @xmath367 .
first of all , @xmath533 is not empty .
it remains to check the four axioms .
axiom ( h1 ) is trivial .
any sub - arrangement of @xmath440 pseudospheres defines a cycle @xmath552 of @xmath400 , up to orientation .
hence axiom ( h2 ) follows from @xmath612 and @xmath613 .
axioms ( h3 ) and ( h4 ) follow from the corresponding axioms in the case @xmath550 , since we can argue by contraction onto a @xmath70sphere containing the two cycles of @xmath400 involved in axioms ( h3 ) and ( h4 ) . in conclusion
, @xmath533 is a hyperline sequence . by theorem [ thm : hls - chi ]
, we can define @xmath534 as the chirotope that is associated to the hyperline sequence @xmath559 with the same positively oriented bases .
let @xmath359 so that @xmath399 ( resp .
@xmath614 ) is an arrangement of oriented pseudospheres of full rank .
it is easy to see that @xmath615 ( resp .
@xmath616 ) .
[ thm : fl ] to any hyperline sequence @xmath4 of rank @xmath6 over @xmath367 , there is an arrangement @xmath617 of @xmath429 oriented pseudo hyperspheres in @xmath618 of full rank with @xmath619 .
the equivalence class of @xmath617 is unique .
we prove theorem [ thm : fl ] by induction on the number of elements and the rank of @xmath4 . in this section ,
we prove the base cases @xmath620 and @xmath621 .
the next section is devoted to the inductive step .
[ lem : flbasecase_r=1 ] theorem [ thm : fl ] holds for @xmath622 .
let @xmath4 be a hyperline sequence of rank @xmath8 over @xmath367 .
we will construct an arrangement @xmath617 of oriented pseudospheres in @xmath426 , with @xmath623 .
an oriented pseudosphere @xmath345 in @xmath624 is the empty set , together with the information whether @xmath625 or @xmath626 .
let @xmath627 be defined as follows . for @xmath360 , set @xmath628 if @xmath629 and @xmath630 otherwise .
it is obvious that @xmath623 and that @xmath617 is unique with this property .
[ lem : flbasecase_r=2 ] theorem [ thm : fl ] holds for @xmath631 .
let @xmath4 be a hyperline sequence of rank @xmath70 over @xmath367 .
we will construct an arrangement @xmath617 of oriented pseudospheres in @xmath538 with @xmath623 , and show that it is unique up to equivalence .
the hyperline sequence @xmath4 is a map from some cyclic group @xmath71 to non - empty subsets of @xmath632 .
we consider @xmath71 as a subgroup of @xmath538 .
an oriented pseudosphere corresponds to an embedding of @xmath633 into @xmath538 . for @xmath360 ,
let @xmath634 so that @xmath635 .
define @xmath636 and @xmath637 .
it follows easily that @xmath623 and that @xmath617 is essentially unique with this property .
[ lem : n = r ] there are exactly two chirotopes of rank @xmath292 over @xmath34 , namely one with @xmath638 $ ] as positively oriented base , and the other with @xmath639 $ ] as positively oriented base .
set @xmath640 . without loss of generality ,
let @xmath134 be a chirotope of rank @xmath6 over @xmath641 .
there are exactly two equivalence classes of oriented @xmath7simplices in @xmath642 , namely those equivalent to @xmath643 $ ] and those equivalent to @xmath644 $ ] .
thus , by axiom ( c2 ) , @xmath134 is completely determined by @xmath645)$ ] .
axiom ( c1 ) implies @xmath646)=\pm 1 $ ] .
hence there are at most two chirotopes of rank @xmath292 over @xmath34 .
the conditions in axioms ( c3 ) and ( c4 ) are empty for @xmath647 .
thus , there are two chirotopes of rank @xmath292 over @xmath34 .
[ lem : flbasecase_n = r ] theorem [ thm : fl ] holds for @xmath621 .
we prove the lemma for chirotopes rather than hyperline sequences .
let @xmath134 be a chirotope of rank @xmath6 over @xmath642 .
by lemma [ lem : n = r ] , @xmath134 is determined by whether @xmath648 $ ] is a positively oriented base of @xmath134 or not . in the former case ,
define @xmath649 , in the latter case define @xmath650 .
we have @xmath651 by construction , and the uniqueness of @xmath652 follows from corollary [ cor : n = d+1 ] .
this section is devoted to the inductive step in the proof of theorem [ thm : fl ] .
let @xmath653 .
suppose that theorem [ thm : fl ] holds for all hyperline sequences of rank @xmath6 with less than @xmath429 elements and for all hyperline sequences of rank less than @xmath6 .
thus , for any non - empty @xmath359 if the contraction @xmath322 ( resp .
the deletion @xmath321 ) is defined , then there is an essentially unique arrangement @xmath654 ( resp .
@xmath655 ) of oriented pseudospheres in @xmath656 ( resp . in @xmath327 ) of full rank with @xmath657 ( resp . with @xmath658 ) . by lemma [ lem : fullrank ]
, there is an element of @xmath4 , say , @xmath429 for simplicity , so that @xmath659 is a hyperline sequence of rank @xmath6 .
denote @xmath660 .
our aim is to construct an oriented pseudosphere @xmath661 as the image of a tame embedding @xmath662 , so that @xmath369 is an arrangement of oriented pseudospheres with @xmath663 .
we outline informally the idea of the construction of @xmath346 .
we start with the arrangement @xmath664 in @xmath665 .
we require that @xmath346 maps this arrangement `` consistently '' to the arrangement @xmath666 , in the sense that any cell in @xmath667 is mapped to a cell in @xmath668 in the correct orientation .
it turns out that this forces @xmath366 to be an arrangement of oriented pseudospheres .
moreover , we show that if @xmath442 intersects the cycles of @xmath669 in a way consistent with the rank 2 contractions of @xmath4 ( i.e. , the cyclic order on its hyperlines ) , then @xmath670 .
our construction of @xmath346 is iterative .
we start with defining @xmath346 on @xmath425dimensional cells of @xmath664 and show that if it is defined on @xmath44dimensional cells then it can be consistently extended to @xmath45dimensional cells .
it turns out that this is possible in an essentially unique way .
let us formalize this idea .
by induction hypothesis , the arrangement @xmath664 exists and is unique up to equivalence . for any element @xmath671 of @xmath672 ,
we denote by @xmath673 the oriented pseudosphere of @xmath674 that corresponds to @xmath25 . for any @xmath675 , set @xmath676 , and similarly @xmath677 for @xmath678 .
recall that the @xmath44dimensional skeleton @xmath422 of an arrangement @xmath400 of oriented pseudospheres is the union of its cells of dimension @xmath423 . for any @xmath359 ,
define @xmath679 .
let @xmath680 .
a * @xmath681admissible * embedding is an embedding @xmath682 so that for any @xmath678 with @xmath683 holds 1 .
@xmath684 or @xmath685 is a pseudosphere in @xmath363 , 2 . if @xmath686 , then any cycle of @xmath687 in @xmath363 is either contained in @xmath685 or meets both connected components of @xmath688 , and 3 . for
any @xmath689 holds @xmath690 and @xmath691 . by the following two lemmas , in our request for the pseudosphere
@xmath442 it suffices to study @xmath7admissible embeddings .
[ lem : admissibelarrangement ] let @xmath692 be a tame embedding that defines an oriented pseudosphere @xmath442 . if @xmath346 is @xmath7admissible then @xmath693 is an arrangement of oriented pseudospheres .
we prove that @xmath694 satisfies axioms ( a1 ) and ( a2 ) .
* let @xmath359 .
it is to show that @xmath363 is empty or homeomorphic to a sphere .
if @xmath695 then we are done since @xmath696 is an arrangement . if @xmath697 then we have @xmath698 , and @xmath699 is empty or homeomorphic to a sphere since @xmath664 is an arrangement .
* let @xmath359 and @xmath360 with @xmath700 . since @xmath346 is @xmath7admissible
, @xmath362 is a pseudosphere in @xmath363 .
it remains to show that @xmath701 and @xmath365 are both non - empty .
if @xmath702 then we are done , since @xmath696 is an arrangement of oriented pseudospheres .
+ if @xmath697 then @xmath703 and @xmath704 .
since @xmath705 , we have @xmath706 .
thus in this case , axiom ( a2 ) for @xmath400 follows from axiom ( a2 ) for @xmath664 applied to @xmath699 and @xmath673 , since @xmath707 and @xmath691 .
+ if @xmath708 , then @xmath709 since otherwise @xmath710 by definition of @xmath442 .
hence @xmath711 is a pseudosphere in @xmath363 .
since @xmath712 is of full rank , @xmath363 contains a cycle of @xmath712 . by the second property in the definition of @xmath7admissible embeddings ,
applied to the empty set , this circle meets both connected components of @xmath713 , which implies axiom ( a2 ) .
thus , @xmath400 is an arrangement of oriented pseudospheres .
[ lem : arrangementadmissibel ] let @xmath714 be a tame embedding that defines an oriented pseudosphere @xmath715 .
assume that @xmath716 is an arrangement of oriented pseudospheres .
if @xmath717 then there is an orientation preserving homeomorphism @xmath718 so that @xmath719 is @xmath7-admissible .
the arrangement @xmath720 of oriented pseudospheres is given by the pre - images of @xmath721 under @xmath722 . by induction hypothesis in the proof of theorem [ thm : fl ] , the equivalence class of @xmath723 is uniquely determined by the property @xmath724 . if @xmath717 , then @xmath725 .
hence , the arrangement @xmath726 is equivalent to @xmath727 .
let @xmath728 be the orientation preserving homeomorphism realizing this equivalence .
it easily follows that @xmath729 is @xmath7admissible . according to the preceding two lemmas
, we shall construct @xmath442 via a tame @xmath7admissible embedding .
moreover , we need to take into account the rank @xmath70 contractions of @xmath4 , as follows .
let @xmath730 be an @xmath8admissible embedding .
assume that for any contraction @xmath322 of rank @xmath70 with @xmath731 , the oriented @xmath425dimensional sphere @xmath732 extends the arrangement of oriented pseudospheres on the oriented cycle @xmath363 that is induced by @xmath733 to an arrangement equivalent to @xmath734 .
then we call @xmath730 * compatible with @xmath4*. in the next three lemmas , we prove that there is an essentially unique @xmath7admissible embedding whose restriction to the @xmath425skeleton of @xmath664 is compatible with @xmath4 .
[ lem:1admissibel ] there is a tame @xmath8admissible embedding @xmath730 that is compatible with @xmath4 .
it is unique up to composition with a homeomorphism @xmath735 that fixes @xmath736 .
let @xmath678 so that @xmath322 is of rank @xmath70 .
we first prove the uniqueness of @xmath730 .
if @xmath737 , then the cyclic order of the signed elements of @xmath322 uniquely determines in which cells of @xmath666 on @xmath363 the two points of @xmath738 must be mapped to , provided @xmath730 is compatible with @xmath4 .
if they are mapped to @xmath425dimensional cells of @xmath739 then their image is unique .
if they are mapped to @xmath8dimensional cells , then their image is unique up to a homeomorphism of these cells fixing the boundary .
if @xmath740 , then @xmath741 , and @xmath699 is a cycle of @xmath664 .
any @xmath425cell on @xmath699 is contained in some pseudosphere @xmath673 of @xmath664 that intersects @xmath699 transversely .
then @xmath742 , and if @xmath730 is @xmath8admissible then @xmath743
. moreover , if @xmath730 is @xmath8admissible then the intersection of @xmath744 with one side of a pseudosphere @xmath596 is mapped to the corresponding side of @xmath745 .
this uniquely determines the image of the two points of @xmath744 under @xmath730 .
we prove the existence of @xmath730 . according to the preceding two paragraphs , for any @xmath425dimensional cell @xmath236 of @xmath664 , a candidate for @xmath746
is given by the cyclic order of the rank @xmath70 contractions of @xmath4 .
we must ensure that the candidate does not depend on the choice of the contraction .
since any two cycles of @xmath666 are contained in some @xmath70sphere that is the intersection of pseudospheres in @xmath666 , we can assume by a contraction that @xmath4 is of rank @xmath0 .
then , the pseudospheres @xmath721 are cycles .
let @xmath747 . if @xmath748 then @xmath749 , hence , the cyclic order of signed points on @xmath479 is a copy of @xmath664 , possibly with opposite orientation . in this case ( and
similarly if @xmath750 ) it is easy to show that the candidates for @xmath730 imposed by @xmath25 and @xmath579 coincide .
it remains the case @xmath751 with @xmath752 .
the positive point @xmath236 of the oriented pseudosphere @xmath673 shall be mapped into the cell @xmath753 of @xmath666 on @xmath479 that corresponds to the atom of @xmath754 containing @xmath755 . if @xmath756 ( resp .
@xmath757 ) , then @xmath236 is the positive ( resp .
negative ) point of @xmath596 , thus shall be mapped into the cell @xmath758 on @xmath745 that corresponds to the atom of @xmath759 containing @xmath755 ( resp . containing @xmath429 ) . if @xmath760 then obviously @xmath761 .
otherwise , @xmath576 is a @xmath425sphere containing both @xmath753 and @xmath758 . up to symmetry
, we can assume that @xmath762 , which means that the atom of @xmath754 containing @xmath755 does also contain @xmath763 .
hence , @xmath753 corresponds to the atom of @xmath759 containing @xmath25 .
this atom also contains @xmath755 ( resp .
@xmath429 ) if and only if @xmath764 ( resp . ( x/\{j})/\{n } = - ( x/\{i})/\{n } ) .
thus @xmath765 by construction of @xmath753 and @xmath758 . in conclusion , the candidates for @xmath746 imposed by @xmath25 and @xmath579 coincide , which is enough to prove the existence of @xmath730 [ lem : extendtraces ] let @xmath766 , and let @xmath767 be a @xmath681admissible embedding . for any @xmath681dimensional cell @xmath434 of @xmath664
there is a cell @xmath768 of @xmath733 of dimension @xmath681 or @xmath769 so that @xmath770 .
let @xmath678 be maximal so that @xmath771 .
in particular , @xmath772 and @xmath773 .
for any cell @xmath774 of dimension @xmath775 in @xmath441 and any @xmath776 with @xmath777 , we have @xmath778 since @xmath767 is @xmath681admissible . by consequence , @xmath779 .
since @xmath434 is a cell of @xmath664 , @xmath780 or @xmath781 , for all @xmath782 . since @xmath767 is @xmath681admissible , @xmath783 or @xmath784 , for all @xmath785
thus , @xmath786 is contained in the closure of a connected component @xmath768 of @xmath787 which is a cell of @xmath739 of dimension @xmath788 .
[ thm : psi_exists ] there is an @xmath7admissible embedding @xmath346 whose restriction to @xmath789 is compatible with @xmath4 .
it is unique , up to a homeomorphism @xmath790 that fixes @xmath733 cellwise .
let @xmath767 denote the restriction of @xmath346 to @xmath791 , for @xmath792 .
we start with inductively proving the uniqueness of @xmath346 .
an @xmath8admissible embedding @xmath730 that is compatible with @xmath4 is essentially unique , by lemma [ lem:1admissibel ] . for @xmath793 ,
assume that @xmath767 is unique , up to a homeomorphism @xmath790 that fixes @xmath666 cellwise .
let @xmath434 be a @xmath681dimensional cell of @xmath664 , and let @xmath678 be maximal so that @xmath794 .
if @xmath346 is @xmath7admissible , then @xmath795 .
let @xmath796 be the cell of @xmath666 containing @xmath797 . since @xmath666 is of full rank , the cell @xmath768 is uniquely determined by @xmath798 . if @xmath799 then the @xmath7admissibility of @xmath346 imposes @xmath800
otherwise , it is a consequence of the generalized schnflies theorem @xcite that @xmath801 , being a tame ball of codimension @xmath8 , is unique up to a homeomorphism @xmath802 that fixes @xmath803 pointwise .
this can be extended to a homeomorphism @xmath790 fixing all cells of @xmath666 , by the so - called cone construction @xcite . in conclusion , @xmath804 is essentially uniquely determined by @xmath767 , which completes the proof of the uniqueness of @xmath346 .
secondly , we expose an iterative construction of @xmath346 .
an @xmath8admissible embedding @xmath730 that is compatible with @xmath4 exists , by lemma [ lem:1admissibel ] .
for @xmath793 , assume that @xmath767 is a @xmath681admissible embedding , and let @xmath434 be a @xmath681dimensional cell of @xmath664 . if there is a @xmath681dimensional cell @xmath768 of @xmath805 so that @xmath806 then we define @xmath807 . otherwise , by the preceding lemma there is a @xmath808dimensional cell @xmath768 of @xmath666 so that @xmath809 .
since @xmath786 is a tame union of tame cells of codimension one in @xmath803 , it follows from @xcite that @xmath786 is a pseudosphere in @xmath803 .
therefore @xmath786 bounds a tame cell @xmath810 .
we define @xmath811 .
we show that we can do this construction so that @xmath804 is an embedding .
let @xmath812 denote the image of @xmath767 , and let @xmath768 be a @xmath808dimensional cell of @xmath666 .
if @xmath767 is an embedding , then @xmath813 is a disjoint union of @xmath681dimensional spheres . if @xmath814 then these spheres bound a system of disjoint @xmath808dimensional cells in @xmath768 .
if @xmath815 then a priori the spheres might be linked , as depicted in figure [ fig : linked ] .
the thick dots indicate the images under @xmath346 of four @xmath425dimensional cells on a cycle of @xmath664 , whose cyclic order corresponds to the numbering .
it is shown in @xcite that this case does not occur . in conclusion
, our construction of @xmath804 produces an embedding .
it remains to show that @xmath804 is @xmath808admissible .
let @xmath678 with @xmath816 . 1 .
either @xmath817 is equal to @xmath363 or it is composed by tame cells of codimension one in @xmath363 . in the latter case , @xmath817 is a pseudosphere ( i.e. , tame ) by @xcite .
2 . let @xmath818 , and let @xmath819 be a cycle of @xmath666 that is not contained in @xmath817
. chose a subset @xmath820 with @xmath821 and @xmath822 .
the contraction @xmath823 is a hyperline sequence of rank @xmath70 .
since @xmath730 is compatible with @xmath4 , @xmath824 comprizes exactly two points @xmath825 , corresponding to the elements @xmath826 in @xmath823 .
+ there is some @xmath827 , so that @xmath823 induces the cyclic order @xmath828 .
let @xmath829 .
since @xmath767 is @xmath681admissible , it follows from lemma [ lem : admissibelarrangement ] that @xmath830 and @xmath817 induce on @xmath831 an arrangement of oriented pseudospheres .
if @xmath552 does not meet both connected components of @xmath832 , then the two points of @xmath833 are contained in a single component of @xmath834 , which is impossible for arrangements of oriented pseudospheres .
hence , @xmath804 satisfies the second property in the definition of @xmath808admissible embeddings .
we observe that an open cell is contained in the positive side of an oriented pseudosphere if and only if some point in its boundary is contained in the positive side of the oriented pseudosphere .
thus , if @xmath434 is a @xmath681dimensional cell of @xmath664 and @xmath835 , then @xmath836 .
since @xmath767 maps @xmath837 into @xmath372 , it follows @xmath838 and therefore @xmath839 .
similarly , if @xmath840 then @xmath841 . therefore @xmath804 is @xmath808admissible , which finishes the proof of theorem [ thm : psi_exists ] . in the remainder of this section , we finish the proof of theorem [ thm : fl ] .
let @xmath662 be a @xmath7admissible embedding whose restriction to the @xmath8skeleton of @xmath664 is compatible with @xmath4 . by theorem [ thm : psi_exists ] , @xmath346 exists .
let @xmath842 be the oriented pseudosphere defined by @xmath346 , and set @xmath694 .
by lemma [ lem : admissibelarrangement ] , @xmath400 is an arrangement of oriented pseudospheres .
we show that @xmath843 , hence , that the hyperlines @xmath844 coincide with those of @xmath533 . if @xmath845 then @xmath846 , since @xmath99 is a rank @xmath70 contraction of @xmath4 and @xmath730 is @xmath8-admissible and compatible with @xmath4 . if @xmath847 , then let @xmath848 .
since @xmath346 is @xmath7admissible , we have @xmath849 , thus , @xmath363 corresponds to the cycle @xmath850 of @xmath664 .
since @xmath851 , the cyclic order of points on this cycle coincides with @xmath99 .
if @xmath852 so that @xmath99 induces the cyclic order @xmath853 , then @xmath854 .
we have @xmath855 , since the topological representation theorem [ thm : fl ] holds in rank @xmath98 by induction hypothesis . therefore , @xmath856 .
it remains to prove the uniqueness of @xmath400 stated in the topological representation theorem [ thm : fl ] .
let @xmath857 be an arrangement of oriented pseudospheres with @xmath858 .
by induction hypothesis , we may use the uniqueness of @xmath666 and can assume that @xmath859 . let @xmath715 be the image of a tame embedding @xmath860 .
then , @xmath722 is @xmath7admissible by lemma [ lem : arrangementadmissibel ] , and it is obvious that its restriction to the @xmath8-skeleton of @xmath674 is compatible with @xmath4 . thus by theorem [ thm : psi_exists ] , @xmath861 coincides with @xmath346 up to a homeomorphism @xmath862 that fixes @xmath863 cellwise .
hence , @xmath400 is equivalent to @xmath864 .
this finishes the proof of the topological representation theorem [ thm : fl ] .
this section is an appendix to section [ sec : x(a ) ] .
we show here that one can get a hyperline sequence from an arrangement of oriented pseudospheres , even if one allows pseudospheres to be not tame .
let @xmath357 be embedded @xmath343dimensional spheres with a choice of an orientation .
we call the ordered multiset @xmath456 an * arrangement of oriented topological spheres * over @xmath367 if it satisfies axioms ( a1 ) and ( a2 ) , where the word `` pseudosphere '' is replaced by `` embedded sphere of codimension one '' .
an embedded sphere @xmath342 of codimension one is * wild * , if there is no homeomorphism @xmath865 mapping @xmath345 to @xmath866 . since there are infinitely many wild spheres in @xmath161 for all @xmath570 ( see @xcite ) , we are no longer allowed to use the generalized schnflies theorem @xcite .
however , @xmath351 has exactly two connected components , even if @xmath345 is wild . similarly to section [ sec : psa ] , we say that two arrangements @xmath366 and @xmath864 of oriented topological spheres in @xmath161 are equivalent if there is an orientation preserving homeomorphism @xmath867 with @xmath868 in the correct orientation , for @xmath869 . for @xmath359 , the definition of the contraction @xmath614 and the deletion @xmath399 is identic to the corresponding definition for arrangements of oriented _ _ pseudo__spheres . we wish to define a hyperline sequence @xmath533 associated to an arrangement @xmath400 of oriented topological spheres . by section [ sec : x(a ) ] , we know how to proceed if all spheres in @xmath400 are tame .
both in the construction of @xmath533 in section [ sec : x(a ) ] and in the proof that @xmath533 is indeed a hyperline sequence , we were using induction on the contractions of @xmath400 , based on contractions of rank @xmath8 and @xmath70 . the only topological argument in the induction step
was the use of the jordan schnflies theorem , that holds also without the assumption of tameness though , in the step from rank @xmath70 to rank @xmath0 .
it remains to remove the tameness condition in the base cases .
rank @xmath8 is trivial . for rank @xmath70 , observe that any two different points @xmath870 can be separated by small intervalls around @xmath871 and @xmath872 .
thus , any embedded sphere @xmath426 in @xmath538 is tame .
therefore , even if @xmath400 is not equivalent to an arrangement of pseudospheres , any rank @xmath70 contraction @xmath614 actually is an arrangement of pseudospheres , and we can read off a hyperline sequence @xmath873 . as in section [ sec : x(a ) ] , these contractions yield a hyperline sequence @xmath533 . in conclusion , although arrangements of oriented topological spheres are very complicated from a topological point of view , their combinatorics is simple enough to read off ordinary hyperline sequences .
nevertheless , there are more arrangements of oriented topological spheres than hyperline sequences , in the sense that there are non - equivalent arrangements @xmath874 and @xmath875 of oriented topological spheres ( for instance , a tame arrangement and a wild arrangement ) with @xmath876 .
hence , there is no analogue of the topological representation theorem [ thm : fl ] in the setting of arrangements of oriented topological spheres .
+ * acknowledgment : * we wish to thank gerhard burde , klaus johannson and martin scharlemann for their advises for topological references . | we present a new direct proof of a topological representation theorem for oriented matroids in the general rank case .
our proof is based on an earlier rank 3 version .
it uses hyperline sequences and the generalized schnflies theorem . as an application
, we show that one can read off oriented matroids from arrangements of embedded spheres of codimension one , even if wild spheres are involved .
* keywords : * oriented matroid , pseudosphere arrangement , chirotope , lawrence representation , hyperline sequences , generalized schnflies theorem . |
softly broken supersymmetry ( susy ) provides a very attractive framework for physics beyond the standard model ( bsm ) , in which the hierarchy problem is solved and the unification of gauge couplings can be realised @xcite .
despite these attractive features , the minimal supersymmetric standard model ( mssm ) suffers from the @xmath3 problem .
the superpotential of the mssm contains the bilinear term @xmath4 , where @xmath5 and @xmath6 are the higgs doublet superfields . in order to get the correct pattern of electroweak ( ew ) symmetry breaking the parameter @xmath3 is required to be in the tev region . at the same time
the incorporation of the mssm into supergravity ( sugra ) or grand unified theories ( gut ) implies that @xmath3 should be of the order of gut or planck scales .
an elegant solution to this problem arises within @xmath1 inspired susy models . at high energies
@xmath1 gut symmetry can be broken to the rank5 subgroup @xmath7 where in general @xmath8 and the two anomaly - free @xmath9 and @xmath10 symmetries originate from the breakings @xmath11 @xmath12 , @xmath13 @xmath14 @xmath10 . if @xmath15 or @xmath16 the extra @xmath17 gauge symmetry forbids the bilinear @xmath3 term but allows interaction @xmath18 in the superpotential . at low energies ( @xmath19 ) the scalar component of the sm singlet superfield @xmath20 acquires a non zero vacuum expectation value ( vev ) breaking @xmath17 and giving rise to an effective @xmath3 term .
within the class of rank5 @xmath1 inspired susy models with extra @xmath17 gauge symmetry , there is a unique choice of abelian gauge group that allows zero charges for right - handed neutrinos .
this is the @xmath21 gauge symmetry given by @xmath22 . only in this exceptional supersymmetric standard model ( e@xmath0ssm ) , which is based on the @xmath23 gauge group , right - handed neutrinos may be superheavy , shedding light on the origin of the mass hierarchy in the lepton sector @xcite-@xcite . to ensure anomaly cancellation the particle content of the e@xmath0ssm is extended to include three complete fundamental @xmath24 representations of @xmath1 .
these multiplets decompose under the @xmath25 subgroup of @xmath1 as follows : @xmath26 the first and second quantities in brackets are the @xmath27 representation and extra @xmath21 charge respectively , while @xmath28 is a family index that runs from 1 to 3 .
an ordinary sm family , which contains the doublets of left handed quarks @xmath29 and leptons @xmath30 , right - handed up and down quarks ( @xmath31 and @xmath32 ) as well as right handed charged leptons , is assigned to @xmath33 + @xmath34 .
right - handed neutrinos @xmath35 should be associated with the last term in eq .
( [ essm1 ] ) , @xmath36 .
the next - to - last term , @xmath37 , represents sm - singlet fields @xmath38 , which carry non - zero @xmath21 charges and therefore survive down to the ew scale .
the pair of @xmath39doublets ( @xmath40 and @xmath41 ) that are contained in @xmath42 and @xmath43 have the quantum numbers of higgs doublets .
they form either higgs or inert higgs @xmath39 multiplets .
other components of these @xmath27 multiplets form colour triplets of exotic quarks @xmath44 and @xmath45 with electric charges @xmath46 and @xmath47 , respectively .
these exotic quark states carry a @xmath48 charge @xmath49 , twice that of ordinary ones .
in addition to the complete @xmath50 multiplets the low energy matter content of the e@xmath0ssm can be supplemented by an @xmath39 doublet @xmath51 and anti - doublet @xmath52 from the extra @xmath53 and @xmath54 to preserve gauge coupling unification .
these components of the @xmath1 fundamental representation originate from @xmath55 of @xmath53 and @xmath56 of @xmath54 by construction .
anomaly cancellation is still guaranteed since @xmath51 and @xmath52 originate from the @xmath53 and @xmath54 supermultiplets .
the analysis performed in @xcite shows that the unification of gauge couplings in the e@xmath0ssm can be achieved for any phenomenologically acceptable value of @xmath57 consistent with the measured low energy central value .
the successful leptogenesis in the early epoch of the universe is the distinctive feature of the e@xmath0ssm .
indeed , the heavy majorana right - handed neutrinos may decay unequally into final states with lepton number @xmath58 , thereby creating a lepton asymmetry in the early universe . because in the e@xmath0ssm the yukawa couplings of the new exotic particles are not constrained by the neutrino oscillation data , substantial values of cp violating lepton asymmetries can be induced even for a relatively small mass of the lightest right handed neutrino ( @xmath59 ) so that successful thermal leptogenesis may be achieved without encountering gravitino problem @xcite . since sphalerons violate @xmath60 but conserve @xmath48 , this lepton asymmetry subsequently gets converted into the present observed baryon asymmetry of the universe through the ew phase transition @xcite .
as in the mssm the gauge symmetry in the e@xmath0ssm does not forbid lepton and baryon number violating operators that result in rapid proton decay .
moreover , exotic particles in @xmath1 inspired susy models give rise to new yukawa interactions that in general induce unacceptably large non diagonal flavour transitions .
to suppress these effects in the e@xmath0ssm an approximate @xmath61 symmetry is imposed . under this symmetry all superfields except one pair of @xmath40 and @xmath41 ( say @xmath62 and @xmath63 ) and one sm - type singlet field ( @xmath64 ) are odd .
the @xmath61 symmetry reduces the structure of the yukawa interactions to @xmath65 & + & h^u_{ij}(\hat{h}_{u } \hat{q}_i)\hat{u}^c_{j } + h^d_{ij}(\hat{h}_{d } \hat{q}_i)\hat{d}^c_j + h^e_{ij}(\hat{h}_{d } \hat{l}_i)\hat{e}^c_{j}+ h_{ij}^n(\hat{h}_{u } \hat{l}_i)\hat{n}_j^c\nonumber\\[2 mm ] & + & \dfrac{1}{2}m_{ij}\hat{n}^c_i\hat{n}^c_j+\mu'(\hat{l}_4\hat{\overline{l}}_4)+ h^{e}_{4j}(\hat{h}_d \hat{l}_4)\hat{e}^c_j + h_{4j}^n(\hat{h}_{u}\hat{l}_4)\hat{n}_j^c\ , , \label{essm2}\end{aligned}\ ] ] where @xmath66 and @xmath67 .
the @xmath39 doublets @xmath68 and @xmath69 and sm - type singlet field @xmath70 , that are even under the @xmath61 symmetry , play the role of higgs fields . at the physical vacuum they develop vacuum expectation values ( vevs ) @xmath71 generating the masses of the quarks and leptons . instead of @xmath72 and @xmath73 it is more convenient to use @xmath74 and @xmath75 .
the vev of the sm - type singlet field , @xmath76 , breaks the extra @xmath77 symmetry generating exotic fermion masses and also inducing that of the @xmath78 boson .
therefore the singlet field @xmath20 must acquire a large vev in order to avoid conflict with direct particle searches at present and past accelerators .
this also requires the yukawa couplings @xmath79 and @xmath80 to be reasonably large .
if @xmath79 or @xmath80 are large enough at the gut scale they affect the evolution of the soft scalar mass @xmath81 of the singlet field @xmath20 rather strongly resulting in a negative value of @xmath81 at low energies which triggers the breakdown of the @xmath21 symmetry .
although @xmath61 eliminates any problems related with baryon number violation and non - diagonal flavour transitions it also forbids all yukawa interactions that would allow the exotic quarks to decay . since models with stable charged exotic particles are ruled out by various experiments the @xmath61 symmetry must be broken . at the same time
, the breakdown of @xmath61 should not give rise to operators that would lead to rapid proton decay .
there are two ways to overcome this problem : the lagrangian must be invariant with respect to either a @xmath82 symmetry , under which all superfields except leptons are even ( model i ) , or a @xmath83 discrete symmetry , which implies that exotic quark and lepton superfields are odd whereas the others remain even ( model ii ) . if the lagrangian is invariant under the @xmath82 symmetry , then the terms in the superpotential which permit exotic quarks to decay and are allowed by the @xmath1 symmetry can be written in the form @xmath84 that implies that exotic quarks are diquarks .
if @xmath83 is imposed then the following couplings are allowed : @xmath85 in this case baryon number conservation requires the exotic quarks to be leptoquarks . in the e@xmath0ssm
some of the exotic quarks can be relatively light . then from fig .
[ fig1 ] one can see that the exotic quark production cross section at the lhc can be comparable with the cross section of @xmath86 production @xcite . in the e@xmath0ssm , the @xmath45 and @xmath44 fermions are susy particles with negative @xmath87parity so they must be pair produced and decay into quark squark ( if diquarks ) or quark slepton , squark lepton ( if leptoquarks ) , leading to final states containing missing energy from the lightest susy particle ( lsp ) . quarks with masses @xmath88 in comparison to top - quark pair production .
( right ) cross section at the 14 tev lhc for pair production of exotic @xmath89quarks as a function of their ( common ) mass @xmath90.,title="fig:",width=302]quarks with masses @xmath88 in comparison to top - quark pair production .
( right ) cross section at the 14 tev lhc for pair production of exotic @xmath89quarks as a function of their ( common ) mass @xmath90.,title="fig:",width=302 ] the lifetime and decay modes of the exotic coloured fermions are determined by the @xmath91 violating couplings .
assuming that @xmath45 and @xmath44 fermions couple most strongly to the third family ( s)quarks and ( s)leptons , the lightest exotic @xmath45 and @xmath44 fermions decay into @xmath92 , @xmath93 , @xmath94 , @xmath95 ( if they are diquarks ) or @xmath96 , @xmath97 , @xmath98 , @xmath99 ( if they are leptoquarks ) .
this can lead to a substantial enhancement of the cross section of either @xmath100 if exotic quarks are diquarks or @xmath101 if exotic quarks are leptoquarks . here
it is worth to point out that the sm production of @xmath102 is @xmath103 suppressed in comparison to the @xmath86 production cross section .
therefore light leptoquarks are expected to lead to the strong signal with low sm background at the lhc .
the results presented in fig . [ fig1 ] suggest that the observation of the d fermions might be possible if they have masses below about 1.5 - 2 tev @xcite .
similar considerations apply to the case of exotic @xmath104 and @xmath105 scalars except that they are non susy particles so they may be produced singly and decay into quark
quark ( if diquarks ) or quark lepton ( if leptoquarks ) without missing energy from the lsp .
it is possible to have relatively light exotic coloured scalars due to mixing effects .
the tevatron and lhc searches for dijet resonances ruled out the presence of light scalar diquarks .
however , scalar leptoquarks may be as light as @xmath106 gev since at hadron colliders they are pair produced through gluon fusion .
scalar leptoquarks decay into quark
lepton final states through small @xmath91 violating terms , for example @xmath107 , and pair production leads to an enhancement of @xmath108 ( without missing energy ) at the lhc .
other possible manifestations of the e@xmath0ssm at the lhc are related to the presence of @xmath78 boson .
the production of a tev scale @xmath78 will provide an unmistakable signal leading to enhanced production of @xmath109 pairs ( @xmath110 ) @xcite .
the differential distribution in invariant mass of the lepton pair @xmath109 in drell yan production is expected to be measurable at the cern collider with a high resolution and would enable one to not only confirm the existence of a @xmath78 state but also establish the possible presence of additional exotic matter , by fitting to the data the width of the @xmath78 resonance . at the lhc , the @xmath78 boson that appears in the @xmath1 inspired models can be discovered if it has a mass below @xmath111 @xcite@xcite .
the determination of its couplings should be possible if @xmath112 @xcite .
the new physics signals associated with the presence of @xmath78 boson and exotic particles predicted by the e@xmath0ssm were discussed in @xcite@xcite , @xcite@xcite .
recently the particle spectrum and collider signatures associated with it were studied within the constrained version of this model @xcite@xcite .
although the @xmath61 symmetry can only be an approximate one from here on we assume that @xmath61 symmetry violating couplings are small and can be neglected in our analysis .
this assumption can be justified if we take into account that the @xmath61 symmetry violating operators may give an appreciable contribution to the amplitude of @xmath113 oscillations and give rise to new muon decay channels like @xmath114 . in order to suppress processes with non diagonal flavour transitions the yukawa couplings of the exotic particles to the quarks and leptons of the first two generations
should be smaller than @xmath115 .
such small @xmath61 symmetry violating couplings can be ignored in the first approximation .
when @xmath61 symmetry violating couplings tend to zero only @xmath116 , @xmath117 and @xmath20 acquire non - zero vevs . the higgs effective potential can be written in the following form : @xmath118 v_f&=&\lambda^2|s|^2(|h_d|^2+|h_u|^2)+\lambda^2|(h_d h_u)|^2\,,\\ v_d&=&\dfrac{g_2 ^ 2}{8}\left(h_d^\dagger \sigma_a h_d+h_u^\dagger \sigma_a h_u\right)^2+\dfrac{{g'}^2}{8}\left(|h_d|^2-|h_u|^2\right)^2 + \dfrac{g^{'2}_1}{2}\left(\tilde{q}_1|h_d|^2+\tilde{q}_2|h_u|^2+\tilde{q}_s|s|^2\right)^2\,,\\[2 mm ] v_{soft}&=&m_{s}^2|s|^2+m_1 ^ 2|h_d|^2+m_2 ^ 2|h_u|^2+\biggl[\lambda a_{\lambda}s(h_u h_d)+h.c.\biggr]\ , , \end{array } \label{higgs1}\ ] ] where @xmath119 , @xmath120 and @xmath121 are the low energy @xmath39 , @xmath122 and @xmath21 gauge couplings while @xmath123 , @xmath124 and @xmath125 are the @xmath21 charges of @xmath117 , @xmath116 and @xmath20 .
the term @xmath126 represents the contribution from loop corrections to the higgs effective potential .
here @xmath127 , @xmath128 and @xmath129 .
the couplings @xmath119 and @xmath130 are known precisely . assuming gauge coupling unification
one can determine the value of extra @xmath77 gauge coupling .
it turns out that @xmath131 for any renormalization scale @xmath132 @xcite .
initially the ewsb sector involves ten degrees of freedom .
however four of them are massless goldstone modes which are swallowed by the @xmath133 , @xmath134 and @xmath78 gauge bosons that gain non - zero masses when higgs fields acquire vevs given by eq .
( [ essm21 ] ) . in the limit where @xmath135 the masses of the @xmath133 , @xmath134 and @xmath78 gauge bosons are @xmath136 where @xmath137 . when cp invariance is preserved the other degrees of freedom form two charged , one cp
odd and three cp - even higgs states .
the masses of the charged and cp - odd higgs bosons are @xmath138 where @xmath139 and @xmath140 are the loop corrections .
if all higgs states except the lightest one are considerably heavier than the ew scale the mass matrix of the cp even higgs sector can be diagonalised using the perturbation theory @xcite-@xcite
. then the masses of two heaviest cp even higgs states are set by @xmath141 and @xmath142 , i.e. @xmath143 the lightest cp even higgs state remains light , i.e. @xmath144 , even when @xmath142 and @xmath145 .
at least one cp even higgs boson is always heavy preventing the distinction between the e@xmath0ssm and mssm higgs sectors .
indeed , the mass of the singlet dominated higgs scalar particle @xmath146 is always close to the mass of the @xmath78 boson that has to be considerably heavier than @xmath147 .
when @xmath148 , vacuum stability requires @xmath142 to be considerably larger than @xmath141 and the ew scale so that the qualitative pattern of the higgs spectrum is rather similar to the one which arises in the pq symmetric nmssm @xcite-@xcite . in the considered limit
the heaviest cp even , cp odd and charged states are almost degenerate around @xmath142 and lie beyond the @xmath149 range @xcite .
if @xmath150 the charged , cp odd and second lightest cp even higgs states may have masses in the @xmath151 range .
however in this case we get a mssm type higgs spectrum with the lightest sm like higgs boson below @xmath152 and with the heaviest scalar above @xmath147 being singlet dominated and irrelevant .
susy models predict that the mass of the lightest higgs particle is limited from above .
the e@xmath0ssm is not an exception .
when the soft masses of the superpartners of the top - quark are equal , i.e. @xmath153 , the two - loop upper bound on the lightest cp - even higgs boson mass @xmath154 in the e@xmath0ssm can be written in the following form : @xmath155 \left(1-\dfrac{3h_t^2}{8\pi^2}l\right)\\[3 mm ] & + & \dfrac{3 h_t^4 v^2 \sin^4\beta}{8\pi^2}\left\ { \dfrac{x_t^2}{m_s^2}\biggl(1-\dfrac{1}{12}\dfrac{x_t^2}{m_s^2}\biggr)+l+ \dfrac{1}{16\pi^2}\biggl(\dfrac{3}{2}h_t^2 - 8g_3 ^ 2\biggr)\biggl ( 2\dfrac{x_t^2}{m_s^2}-\dfrac{1}{6}\dfrac{x_t^4}{m_s^4}+l\biggr)l\right\}\ , , \end{array } \label{higgs6}\ ] ] where @xmath156 $ ] and @xmath157 is the usual stop mixing parameter . at tree level
the upper limit on the mass of the lightest higgs particle is described by the sum of the three terms in the square brackets .
one - loop corrections from the top - quark and its superpartners increase the bound on the lightest cp - even higgs boson mass substantially while the inclusion of leading two - loop corrections reduces the upper limit on @xmath154 . in order to enhance the contribution of loop effects we assume maximal mixing in the stop sector ( i.e. @xmath158 ) .
we also adopt @xmath159 . then since @xmath160 is determined uniquely if we require the unification of gauge couplings the theoretical restriction on the lightest higgs mass ( [ higgs6 ] ) depends on @xmath161 and @xmath162 only .
the requirement of validity of perturbation theory up to the gut scale constrains the parameter space further setting a limit on the yukawa coupling @xmath161 for each value of @xmath162 .
relying on the results of the analysis of the renormalisation group ( rg ) flow in the e@xmath0ssm presented in @xcite one can obtain the maximum possible value of the lightest higgs boson mass for each particular choice of @xmath162 .
( right ) the dependence of the two loop upper bound on the lightest higgs boson mass on @xmath162 for @xmath163 , @xmath153 , @xmath164 and @xmath159 .
the solid , lower and upper dotted lines represent the theoretical restrictions on the mass of the lightest cp even higgs state in the mssm , nmssm and e@xmath0ssm respectively.,title="fig:",width=302 ] .
( right ) the dependence of the two loop upper bound on the lightest higgs boson mass on @xmath162 for @xmath163 , @xmath153 , @xmath164 and @xmath159 .
the solid , lower and upper dotted lines represent the theoretical restrictions on the mass of the lightest cp even higgs state in the mssm , nmssm and e@xmath0ssm respectively.,title="fig:",width=302 ] + the dependence of the tree level and two - loop upper bounds on the mass of the lightest higgs state on @xmath162 is examined in fig .
[ fig2 ] where these bounds are compared with the corresponding limits in the mssm and nmssm .
one can see that in the interval of @xmath162 from @xmath165 to @xmath166 the maximum value of the mass of the lightest higgs boson in the e@xmath0ssm is larger than the experimental lower limit on the sm like higgs boson even at tree level . at moderate values of @xmath162 ( @xmath167 )
the two loop upper limit on @xmath154 in the e@xmath0ssm is also considerably higher than in the mssm and nmssm .
it reaches the maximum value @xmath168 at @xmath169 . in the considered part of the parameter space
the theoretical restriction on the mass of the lightest cp - even higgs boson in the nmssm exceeds the corresponding limit in the mssm because of the extra contribution to @xmath170 induced by the additional @xmath171-term in the higgs scalar potential of the nmssm .
the size of this contribution , which is described by the first term in the square brackets of eq .
( [ higgs6 ] ) , is determined by the yukawa coupling @xmath161 .
the upper limit on the coupling @xmath161 caused by the validity of perturbation theory in the nmssm is more stringent than in the e@xmath0ssm due to the presence of exotic matter . as a result
the upper limit on @xmath154 in the nmssm is considerably less than in the e@xmath0ssm at moderate values of @xmath162 . at large @xmath172
the contribution of the @xmath171-term of the sm - type singlet field to @xmath173 vanishes . therefore with increasing @xmath162 the upper bound on the lightest higgs boson mass in the nmssm approaches the corresponding limit in the mssm . in the e@xmath0ssm
the theoretical restriction on the mass of the lightest higgs scalar also diminishes when @xmath162 rises .
but even at very large values of @xmath162 the upper limit on @xmath154 in the e@xmath0ssm is still @xmath174 larger than the ones in the mssm and nmssm because of the @xmath21 @xmath89-term contribution to @xmath175 ( last term in the square brackets of eq .
( [ higgs6 ] ) ) .
in the e@xmath0ssm the lightest susy particle tends to be the lightest inert neutralino .
the inert neutralino sector is formed by the neutral components of the inert higgsinos ( @xmath176 , @xmath177 , @xmath178 , @xmath179 ) and inert singlinos ( @xmath180 , @xmath181 ) . in the exact @xmath91 symmetry limit
the scalar components of the corresponding superfields do not acquire vevs and inert neutralino states do not mix with the ordinary neutralinos . in the field basis
@xmath182 the mass matrix of the inert neutralinos takes a form @xmath183 a_{12 } & a_{11 } \end{array } \right)\,,\qquad a_{\alpha\beta}=-\frac{1}{\sqrt{2 } } \left ( \begin{array}{ccc } 0 & \lambda_{\alpha\beta } s & \tilde{f}_{\beta\alpha } v \sin{\beta } \\[2 mm ] \lambda_{\beta\alpha } s & 0 & f_{\beta\alpha } v \cos{\beta } \\[2 mm ] \tilde{f}_{\alpha\beta } v \sin{\beta } & f_{\alpha\beta } v \cos{\beta } & 0 \end{array } \right)\ , , \label{icn1}\ ] ] so that @xmath184
. as before we choose the vev of the sm singlet field @xmath76 to be large enough ( @xmath185 ) so that the masses of all inert chargino states , which are formed by the charged components of the inert higgsinos @xmath186 , are larger than @xmath187 and @xmath78 boson is relatively heavy . in addition , we also require the validity of perturbation theory up to the gut scale .
the restrictions specified above set very strong limits on the masses of the lightest inert neutralinos . in particular , our numerical analysis indicates that the lightest and second lightest inert neutralinos ( @xmath188 and @xmath189 ) are typically lighter than @xmath190 @xcite@xcite .
therefore the lightest inert neutralino tends to be the lightest susy particle in the spectrum and can play the role of dark matter .
the neutralinos @xmath188 and @xmath189 are predominantly inert singlinos .
their couplings to the @xmath134boson can be rather small so that such inert neutralinos would remain undetected at lep . in order to clarify the results of our numerical analysis ,
it is useful to consider a simple scenario when @xmath191 , @xmath192 and @xmath193 . in the limit where off diagonal yukawa couplings vanish and @xmath194 the eigenvalues of the inert neutralino mass matrix can be easily calculated ( see @xcite ) . in particular the masses of two lightest inert neutralino states ( @xmath188 and @xmath189 )
are given by @xmath195 where @xmath196 are masses of the inert charginos . from eq .
( [ icn10 ] ) one can see that the masses of @xmath188 and @xmath189 are determined by the values of the yukawa couplings @xmath197 and @xmath198 .
they decrease with increasing @xmath162 and chargino masses . in this approximation
the part of the lagrangian , that describes interactions of @xmath134 with @xmath199 and @xmath200 , can be presented in the following form : @xmath201 r_{z\alpha\beta}=r_{z\alpha\alpha}\,\delta_{\alpha\beta}\,,\qquad r_{z\alpha\alpha}=\frac{v^2}{2 m_{\chi^{\pm}_{\alpha}}^2 } \biggl(f_{\alpha}^2\cos^2\beta-\tilde{f}_{\alpha}^2\sin^2\beta\biggr)\ , .
\label{icn14}\end{aligned}\ ] ] eqs .
( [ icn14 ] ) demonstrates that the couplings of @xmath199 and @xmath200 to the z - boson can be very strongly suppressed or even tend to zero .
this happens when @xmath202 .
although @xmath199 and @xmath200 might have extremely small couplings to @xmath134 , their couplings to the lightest cp even higgs boson @xmath203 can not be negligibly small if @xmath199 and @xmath200 have appreciable masses . when the susy breaking scale @xmath204 and the vev @xmath76 of the singlet field are considerably larger than the ew scale , the lightest cp even higgs state is the analogue of the sm higgs field and is responsible for all light fermion masses in the e@xmath0ssm .
therefore it is not so surprising that in the limit when @xmath194 the part of the lagrangian that describes the interactions of @xmath199 and @xmath200 with @xmath203 takes a form @xmath205 i.e. the couplings of @xmath203 to @xmath199 and @xmath200 are proportional to the mass / vev . in eq .
( [ icn12 ] ) @xmath206 is the set of inert neutralino eigenstates with positive eigenvalues , while @xmath207 equals 0 ( 1 ) if the eigenvalue corresponding to @xmath208 is positive ( negative ) . in our analysis
we require that the lightest inert neutralino account for all or some of the observed dark matter relic density .
this sets another stringent constraint on the masses and couplings of @xmath209 . indeed , because the lightest inert neutralino states are almost inert singlinos , their couplings to the gauge bosons , higgs states , quarks ( squarks ) and leptons ( sleptons ) are rather small resulting in a relatively small annihilation cross section of @xmath210 and the possibility of an unacceptably large dark matter density .
thus the bulk of the e@xmath0ssm parameter space , that leads to small masses of @xmath211 , is almost ruled out the masses of @xmath211 and @xmath212 tend to zero and inert singlino states essentially decouple from the rest of the spectrum . in this limit
the lightest non - decoupled neutralino may be rather stable and can play the role of dark matter @xcite .
the presence of very light neutral fermions in the particle spectrum might have interesting implications for the neutrino physics ( see , for example @xcite ) . ] .
a reasonable density of dark matter can be obtained for @xmath213 when the lightest inert neutralino states annihilate mainly through an @xmath76channel @xmath134-boson , via its inert higgsino doublet components which couple to the @xmath134boson .
if @xmath211 annihilation proceeds through the @xmath134boson resonance , i.e. @xmath214 , then an appropriate value of dark matter density can be achieved even for a relatively small coupling of @xmath211 to @xmath134 .
since the masses of @xmath199 and @xmath200 are much larger than the @xmath215quark mass and the decays of @xmath203 into these neutralinos are kinematically allowed , the sm like higgs boson decays predominantly into the lightest inert neutralino states and has very small branching ratios ( @xmath216 ) for decays into sm particles @xcite@xcite .
the lightest inert neutralino states can get appreciable masses @xmath217 only if at least one light inert chargino state and two inert neutralinos states , which are predominantly components of the @xmath39 doublet , have masses below @xmath218 .
the inert chargino and neutralinos states that are mainly inert higgsinos couple rather strongly to @xmath219 and @xmath134bosons and therefore can be efficiently produced at the lhc and then decay into the lsp and pairs of leptons and quarks giving rise to remarkable signatures which can be observed in the near future .
if the masses of @xmath199 and @xmath200 are very close then the decays of @xmath203 into @xmath220 will give rise to a large invisible branching ratio of the sm like higgs boson .
when the mass difference between the second lightest and the lightest inert neutralinos is larger than @xmath221 the invisible branching ratio remains dominant but some of the decay products of @xmath222 might be observed at the lhc . in particular , there is a chance that soft @xmath223 pairs may be detected .
since the branching ratios of @xmath203 into sm particles are extremely suppressed , the decays of the sm like higgs boson into @xmath224 could be important for higgs searches @xcite .
r.n . would like to thank ftpi , university of minnesota for its hospitality and m. a. shifman , k. a. olive , a. i. vainshtein , a. mustafayev , w. vinci , p. a. bolokhov , p. koroteev for fruitful discussions .
authors are grateful to x. r. tata , m. muehlleitner , l. clavelli , d. stockinger , d. j. miller , d. g. sutherland , j. p. kumar , d. marfatia , k. r. dienes , b. d. thomas for valuable comments and remarks .
the work of r.n . and s.p .
was supported by the u.s .
department of energy under contract de - fg02 - 04er41291 , and the work of m.s . was supported by the national science foundation phy-0755262 . the work of s.m .
is partially supported by the next institute .
s.f.k . and s.m .
acknowledge partial support from the stfc rolling grant st / g000557/1 .
d.j.m . acknowledges partial support from the stfc rolling grant st / g00059x/1 .
is thankful to the stfc for providing studentship funding . | we consider collider signatures of the exceptional supersymmetric ( susy ) standard model ( e@xmath0ssm ) .
this @xmath1 inspired susy model is based on the sm gauge group together with an extra @xmath2 gauge symmetry under which right handed neutrinos have zero charge .
to ensure anomaly cancellation and gauge coupling unification the low energy matter content of the e@xmath0ssm involve extra exotic matter beyond the mssm .
we discuss the collider signatures associated with the production of new particles predicted by the e@xmath0ssm and consider the implications of this model for dark matter and higgs phenomenology . since exotic quarks in the e@xmath0ssm can be either diquarks or leptoquarks they may provide spectacular new physics signals at the lhc . |
formation of a circumstellar disk is recognized as a natural step in the process of star formation and a vital step toward forming planets .
currently our knowledge of protoplanetary disk properties at anthropically interesting distances of 1 - 30 au is relatively limited .
the best spectroscopic constraints on disk temperature and density structure are dominated by material either within 0.5 au of the central source ( najita et al .
1996 ; carr , mathieu , & najita 2001 ) or at radii @xmath450 au ( dutrey , guilloteau , & simon 1994 ) .
this radial sampling results from the tracers utilized to date : dust emission and scattering and co ro - vibrational and rotational emission . using rotational lines as a tracer may allow study of disks at radii of 1 - 50 au .
recent work by thi et al .
( 1999 , 2001a , 2001b ) report @xmath0 @xmath5 and @xmath1 emission from pre - main - sequence and main - sequence stars with circumstellar disks . from infrared space observatory ( iso ) data , these authors derive gas temperatures of 100 - 200 k and masses up to 2@xmath610@xmath7 . in the case of disks around main - sequence stars ( thi et al .
2001a ) , previously thought to be mostly gas - free ( zuckerman , forveille , & kastner 1995 ) , the presence of a substantial reservoir of alters ideas on the formation of giant planets ( lissauer 2001 ) .
unfortunately , the large aperture and low spectral resolution of the observations leave open the question of whether the emission really comes from a disk .
three mid - ir pure rotational lines are readily available from high , dry sites if observed at moderately high spectral resolution : @xmath0 @xmath1 [ _ i.e. _ s(1 ) ] at 17 , @xmath2 [ _ i.e. _ s(2 ) ] at 12 , and @xmath8 [ _ i.e. _ s(4 ) ] at 8 .
the high spectral and spatial resolution possible with ground - based spectroscopy , together with large telescope apertures , can result in greater sensitivity to certain gas distributions than satellite observations .
in particular , ground - based spectroscopy is well suited for detecting point sources with narrow line emission , such as gas in keplerian orbit at a radius of 1 - 50 au from a solar - mass star .
we report here on observations of a small sample of young stars taken with the texas echelon cross echelle spectrograph ( texes ) in an effort to confirm the iso detections and explore the feasibility of ground - based observations of from circumstellar disks .
we used texes ( lacy et al .
2002 ) at the nasa 3 m infrared telescope facility ( irtf ) to observe the h@xmath9 @xmath1 transition [ @xmath10 or @xmath11 ] and the @xmath2 transition [ @xmath12 or @xmath13 ] .
all the observations were made with texes in high - resolution mode .
pertinent details regarding the observations may be found in table [ tab : obs ] .
all sources were observed while using the irtf offset guide camera as well as guiding on the dispersed continuum seen through the spectrograph .
the case of gg tau deserves special comment since it has weak continuum , is a strong case for iso detection ( thi et al .
1999 ; thi et al .
2001b ) , and has a unique geometry .
the source is a quadruple system composed of a pair of binaries .
gg tau a is a 0.25 binary with millimeter continuum emission ( guilloteau , dutrey , & simon 1999 ) and hst scattered light observations ( silber et al . 2000 ) showing a circumbinary ring extending roughly from 180 to 260 au ( 1.29 - 1.86 ) .
the ring is tipped at 37@xmath14 , resulting in an ellipse on the sky with the major axis running essentially e - w . when we observed gg tau a , we widened the slit to 3 and rotated the instrument to orient the slit so that the major axis of the projected ellipse would lie along our slit .
the weak continuum of gg tau meant we could not guide on continuum signal through the spectrograph .
therefore , we repeatedly checked our infrared boresight by observations of @xmath15 tau , a nearby , bright infrared source .
the maximum boresight offset found based on the @xmath15 tau observations was 0.5 , with a mean offset of 0.2 . by summing over @xmath164 along the slit
, we can confidently state that we observed essentially all of the gg tau a gap region .
correction for atmospheric transmission was done using bright infrared continuum objects , either stars or asteroids .
although most stars later than spectral type g show photospheric features at r@xmath17 , we can locate features using the kitt peak sunspot atlas ( wallace , livingston , & bernath 1994 ) and the atmos3 photospheric atlas ( geller 1992 ) and know that there are no features near the transitions .
asteroids have no features at our resolution . both the @xmath1 and @xmath2 lines are near telluric atmospheric lines , but the doppler shift of the source , the earth s motion , and the high spectral resolution available with texes helped to minimize atmospheric effects .
flux calibration was done using a blackbody and standard stars .
the data were reduced using the texes data pipeline ( lacy et al .
we extracted spectra from the data with several strategies : optimal extraction of the point source to look for emission coincident with the continuum , sums over the nodded region to look for diffuse emission covering half the slit length , and selected sums along the slit to look for isolated emission offset from the point source .
we saw no evidence for extended emission in any spectrum , although uniform emission on @xmath185 scales would not be recovered .
figure [ fig : s1 ] presents @xmath1 spectra for the three sources we observed that also have reported iso detections : gg tau , ab aur , and hd 163296 . to indicate the line flux reported by thi et al .
( 2001b ) , we overplot gaussians with an integrated line flux equal to the iso measurements . since iso was unable to resolve the line profiles , we have simply assumed gaussian fwhm matching our resolution ( 5 or 7.5 , depending on slit width ) and 30 , with the gaussians centered at the systemic v@xmath19 .
the figures clearly show a discrepancy between the reported line flux and our observations . of all our spectra , only the ab aur 12 spectrum shows a possible detection ( figure [ fig : s2 ] ) .
a gaussian fit to the ab aur data finds a feature centered at the systemic velocity with a fwhm=10.5 and a line flux of 2.0 ( @xmath201.0 ) @xmath21 .
there are no terrestrial lines and no photospheric absorption lines in the comparison star at this frequency that might artificially produce an emission feature .
we discuss the implications of these data , particularly in light of the non - detection of the @xmath1 line , below .
table [ tab : res ] summarizes the results for all our observations . in all cases , we measure continuum levels comparable to past measurements , although the noise for gg tau is such that many pixels must be combined to obtain a significant continuum measurement .
we established line flux errors by summing over the number of pixels corresponding to the assumed line widths at all parts of the spectrum with comparable atmospheric transmission .
we then use the line flux upper limit for the more conservative case of a 30 fwhm line to calculate a maximum mass of assuming temperatures of 150 k and 300 k. note that these masses assume no extinction and do not include gas in equilibrium with optically thick dust .
we observed six sources with reasonable evidence for disks and report no convincing detection of pure - rotational emission . our non - detections have bearing on the interpretation of previous iso reports and on the distribution of gas and dust in these sources . in all cases where we set limits , those limits are subject to uncertainties in the assumed gas temperature and the assumed line widths .
to provide an example of our sensitivity to these assumptions , we include two line widths and two gas temperatures in table [ tab : res ] . three reported iso detections of from circumstellar disks , out of three tested , are not seen with texes . for iso to detect emission that we would miss ,
the emission can either be extended or it can be spectrally broad . given that the iso - sws aperture at 17 covered 14 by 27 , compared with our effective aperture of 2 by 2 ( 280 au by 280 au at taurus : d=140 pc . ) , and that iso detected rotational lines from molecular clouds with even fairly modest incident uv fields ( li et al .
2002 ) , the presence of extended emission is possible .
thi et al . ( 2001b ) address this issue using @xmath22co 3 - 2 emission and conclude that gg tau shows no contaminating molecular material , that the bulk of molecular material for hd 163296 lies in the disk , but that ab aur may have a significant contribution from an extended envelope . for the case of broad lines ,
we believe we would have seen lines as broad as @xmath1650 ( see figure [ fig : s1 ] and table [ tab : res ] ) .
even for the case of gg tau , we would have seen a line as broad as 30 at the 2@xmath3 level , given the best fit line flux in thi et al .
( 2001b ) , although the uncertainties are large enough that these observations may not be in conflict , especially for a broad line .
recent @xmath23 @xmath1 observations at 2 show line widths @xmath2415 for gg tau ( j. bary , in preparation ) , while co ro - vibrational emission toward ab aur is unresolved at 12 ( g. blake , in preparation ) .
although the relationship between these tracers and the pure rotational emission depends in detail upon the excitation mechanism , these observations show that our assumption of 30 line widths when calculating mass upper limits is likely conservative . as a fiducial point , the projected keplerian velocity at a distance of 1 au around gg tau aa ( m = 0.8 m@xmath25 ; white et al .
1999 ) and assuming any circumstellar disk has the same orientation as the more extended circumbinary ring ( inclination @xmath26 ) , would be v=16 . for ab aur , taking m=2.4 m@xmath25 ( van den ancker et al .
1998 ) and inclination @xmath27 ( grady et al . 1999 ) , the projected keplerian velocity would be v@xmath28 . of course , emission lines broadened by keplerian rotation will generally have widths broader than their characteristic keplerian velocity .
although texes is a relatively new instrument , we feel reasonably confident in our results .
the successful detection of continuum emission from each source at levels consistent with past observations means we were pointed at the source and that our flux determinations are not dramatically in error .
we have detected rotational emission in sources such as uranus ( l. trafton , private communication ) and ngc 7027 ( h. dinerstein , in preparation ) .
the iso detections are the result of special data processing and are near the sensitivity limit ( thi et al .
2001b ) . because the iso reports of molecular gas reservoirs around main - sequence stars such as @xmath29 pic ( thi et al .
2001a ) have such important ramifications for our understanding of planetary formation , the current non - detections in pre - main - sequence disk sources strongly argue for follow - up observations of the main - sequence stars observed by iso , all of which are reported to be `` medium confidence '' detections .
follow - up observations are particularly needed for the well studied source @xmath29 pic , where fuse observations detect no absorption through the edge - on disk against a stellar o vi emission line ( lecavelier des etangs et al .
2001 ) and where na i emission shows a gas disk that seems to coexist with the well known dust disk ( olofsson , liseau , & brandeker 2001 ; heap et al .
2000 ) in the disk sources we observed , the absence of emission certainly does not mean there is no warm . where the dust in the disk is optically thick , an emission line would result only if there is spatial and/or temperature separation between the dust and , such as a gap in the dust disk or a hot layer above the optically thick disk ( dalessio et al .
1998 ; chiang & goldreich 1997 ) gaps are believed to be present in gg tau and gw ori ( guilloteau et al . 1999 ; mathieu , adams , & latham 1991 ) ; for these sources , the mass upper limits constrain the gas within the gap . in general , our observations argue , albeit weakly , for little spatial separation or temperature differentiation between the bulk of the gas and dust . as described in section
[ sec : res ] , the 12 spectrum of ab aur has a 2@xmath3 bump at the correct velocity for the @xmath2 line and with a reasonable width . when taken with the 3@xmath3 upper - limit on @xmath1 emission from this source , 3.3 @xmath21 , the @xmath2 line flux of 2.0 @xmath30 ( @xmath31 , @xmath32 ) implies t@xmath33 380 k ( 275 k , 500 k ) , assuming no differential extinction . this is warmer than the 140 k seen in co ro - vibrational lines ( g. blake , in preparation ) . for simple thermal emission
, we would expect the pure rotational lines to come from cooler regions than co ro - vibrational emission , although 140 k gas would not collisionally excite ro - vibrational co emission .
the 3@xmath3 upper - limit on @xmath34 line flux as observed with iso , 0.4 @xmath21 , ( thi et al .
2001b ) combined with the possible @xmath2 detection ( @xmath31 , @xmath32 ) suggest t@xmath24190 k ( 220 k , 170 k ) . given the errors and the resulting inconsistency , we consider the ab aur spectrum to give an upper limit to the @xmath2 integrated line flux of 4.0 @xmath6 10@xmath35 . as well as reobserving the @xmath1 and @xmath2 lines , ground - based observations of the @xmath8 line at 8 , which should be quite strong
if the gas is truly at 380 k , would help determine if the @xmath2 is actually as strong as it appears . while the next infrared satellites will have exceptional mid - infrared sensitivity
, they may not be able to detect low line - to - continuum ratio features such as may be present in the ab aur 12 data .
if ab aur were observed at r=600 , the maximum resolution available with the sirtf infrared spectrograph ( roellig et al . 1998 ) , a snr of @xmath161000 would give a 3@xmath3 detection of our derived line flux .
the recommended mid - infrared imager / spectrograph for ngst , with r=1500 ( mather & stockman 2000 ) , would require snr of 450 for a 3@xmath3 detection .
higher spectral resolution on ngst , still a design goal , would dramatically ease this type of observation .
future ground - based observations with 8 - 10 m class telescopes and high resolution spectrographs such as texes will improve on the sensitivity described here by an order of magnitude ( a factor of 100 in time ) . in conclusion ,
we have attempted to confirm several reports of pure - rotational emission from young circumstellar disks without success .
our observations generally set more stringent limits on gas within @xmath3650 au of the source and with widths @xmath3750 than obtained with iso .
we have found possible ( 2@xmath3 ) emission from one source , ab aur , although inconsistencies in derived gas temperatures lead us to consider the result as an upper limit .
the absence of emission suggests the gas and dust have little segregation in temperature and/or spatial extent , although more observations are necessary .
we have shown that ground - based , high - resolution spectroscopy is an effective way to search for narrow , pure rotational line emission such as that expected from circumstellar disks .
we are happy to acknowledge the day and night staff at the irtf ; without their hard work , none of this observing would be possible .
we appreciate the comments of ewine van dishoeck and an anonymous referee .
this work was supported by grant tarp 00365 - 0473 - 1999 from the texas advanced research program .
mjr acknowledges usra grant usra 8500 - 98 - 008 through the sofia program .
this research has made use of nasa s astrophysics data system and the simbad database operated at cds , strasbourg , france .
cccccc gg tau & @xmath1 & 40,000 & 3.0 & 11.5 & 7000 + ab aur & @xmath1 & 60,000 & 2.0 & 11.5 & 1800 + & @xmath2 & 83,000 & 1.4 & 7.5 & 2600 + hd 163296 & @xmath1 & 60,000 & 2.0 & 11.5 & 2400 + gw ori & @xmath1 & 60,000 & 2.0 & 11.5 & 1600 + & @xmath2 & 83,000 & 1.4 & 7.5 & 3800 + l1551 irs5 & @xmath1 & 60,000 & 2.0 & 11.5 & 2000
+ dg tau & @xmath2 & 83,000 & 1.4 & 7.5 & 4800 + ccccccccc gg tau & 1.1 & 2.8 ( 0.8 ) & 17 & 0.7 ( 1.4 ) & @xmath241.6 & @xmath243.9 & @xmath2447 & @xmath243.0 + ab aur & 24.4 & 30 ( 9 ) & 17 & 22.7 ( 2.3 ) & @xmath243.3 & @xmath247.2 & @xmath2491 & @xmath245.9 + & & & 12 & 18.7 ( 1.3 ) & @xmath244.0 & & @xmath241250 & @xmath248.7 + hd 163296 & 16.9 & 22 ( 6 ) & 17 & 15.3 ( 2.5 ) & @xmath242.0 & @xmath245.1 & @xmath2446 & @xmath243.0 + l1551 irs5 & & & 17 & 20.0 ( 1.7 ) & @xmath241.4 & @xmath244.3 & @xmath2451 & @xmath243.3 + gw ori & & & 17 & 6.8 ( 2.0 ) & @xmath241.6 & @xmath243.9 & @xmath24330 & @xmath2421 .
+ & & & 12 & 6.1 ( 0.5 ) & @xmath240.5 & @xmath242.2 & @xmath245300 & @xmath2437 . + dg tau & & & 12 & 5.9 ( 0.5 ) & @xmath240.6 & @xmath243.0 & @xmath24880 & @xmath246.2 + | we report on a limited search for pure - rotational molecular hydrogen emission associated with young , pre - main - sequence stars .
we looked for @xmath0 @xmath1 and @xmath2 emission in the mid - infrared using the texas echelon - cross - echelle spectrograph ( texes ) at nasa s 3 m infrared telescope facility .
the high spectral and spatial resolution of our observations lead to more stringent limits on narrow line emission close to the source than previously achieved .
one star , ab aur , shows a possible ( 2@xmath3 ) detection , but further observations are required to make a confident statement .
our non - detections suggest that a significant fraction , perhaps all , of previously reported emission towards these objects could be extended on scales of 5 or more . |
in this paper we investigate the origin and the consequences of the _ bi - stability jump _ of the stellar winds of early - type stars near spectral type b1 .
this bi - stability jump is observed as a steep decrease in the terminal velocity of the winds from @xmath8 for supergiants of types earlier than b1 to @xmath9 for supergiants of types later than b1 ( lamers et al .
we will show that this jump in the wind velocity is accompanied by a jump in the mass - loss rate with @xmath7 increasing by about a factor of five for supergiants with between 27 500 and 22 500 k. the theory of radiation driven winds predicts that the mass - loss rates and terminal velocities of the winds of early - type stars depend smoothly on the stellar parameters , with @xmath10 and @xmath11 ( castor et al .
1975 , abbott 1982 , pauldrach et al .
1986 , kudritzki et al .
this theory has not yet been applied to predict the observed jump in the ratio @xmath0 for supergiants near spectral type b1 .
the change from a fast to a slow wind is called the @xmath12-@xmath13 jump .
if the wind momentum @xmath14 were about constant across the bi - stability jump , it would imply that the mass - loss rate would _ increase _ steeply by about a factor of two from stars with spectral types earlier than b1 to later than b1 . unfortunately , there are no reliable mass - loss determinations from observations for stars later than spectral type b1 .
so far , a physical explanation of the nature of this bi - stability jump has been lacking . in this paper
, we attempt to provide such an explanation and we investigate the change in mass - loss rate that is accompanied by the change in @xmath15 .
the concept of a bi - stability jump was first described by pauldrach & puls ( 1990 ) in connection to their model calculations of the wind of the luminous blue variable ( lbv ) star p cygni (= 19.3 kk ) .
their models showed that small perturbations in the basic parameters of this star can either result in a wind with a relatively high mass loss , but low terminal velocity , _ or _ in a wind with relatively low , but high .
their suggestion was that the mechanism is related to the behaviour of the lyman continuum .
if the lyman continuum exceeds a certain optical depth , then as a consequence , the ionization of the metals shifts to a lower stage .
this causes a larger line acceleration @xmath16 and finally yields a jump in @xmath7 .
the models of pauldrach & puls ( 1990 ) for p cygni show that the wind momentum loss per second , @xmath14 , is about constant on both sides of the jump ( see lamers & pauldrach 1991 ) .
so lamers et al .
( 1995 ) put forward the idea that the mass - loss rate for normal stars could increase by about a factor of two , if @xmath15 decreases by a factor of two , so that @xmath14 is constant on both sides of the jump . whether this is indeed the case , is still unknown . to investigate the behaviour of the mass loss at the bi - stability jump
, we will derive mass - loss rates for a grid of wind models over a range in .
the main goal of the paper is to understand the processes that cause the bi - stability jump .
although our results are based on complex numerical simulations , we have attempted to provide a simple picture of the relevant physics .
we focus on the observed bi - stability jump for normal supergiants .
nevertheless , these results may also provide valuable insight into the possible bi - stable winds of lbvs .
it is worth mentioning that lamers & pauldrach ( 1991 ) and lamers et al .
( 1999 ) suggested that the bi - stability mechanism may be responsible for the outflowing disks around rapidly - rotating b[e ] stars .
therefore our results may also provide information about the formation of rotation induced bi - stable disks .
the paper is organized in the following way . in sect .
[ sec : simple ] we describe the basic stellar wind theory .
in particular we concentrate on the question : `` what determines and ? '' .
we show that is determined by the radiative acceleration in the _ subsonic _ region . in sect .
[ sec : method ] we explain the method that we use to calculate the radiative acceleration with a monte carlo technique and the mass - loss rates of a grid of stellar parameters . sect .
[ sec : isa ] describes the properties of the models for which we predict . in sect .
[ sec : predictions ] our predicted bi - stability jump in mass loss will be presented .
then , in sect . [ sec : origin ] we discuss the origin of this jump and show that it is due to a shift in the ionization balance of fe iv to fe iii .
then , we discuss the possible role of the bi - stability jump in on the variability of lbv stars in sect .
[ sec : lbv ] . finally , in sect .
[ sec : concl ] , the study will be summarized and discussed .
mass loss from early - type stars is due to radiation pressure in lines and in the continuum ( mainly by electron scattering ) .
since the radiative acceleration by line processes is the dominant contributor , the winds are `` line - driven '' , i.e. the momentum of the radiation is transferred to the ions by line scattering or line absorption .
line - scattering and line absorption occur at all distances in the wind , from the photosphere up to distances of tens of stellar radii .
so the radiative acceleration of the wind covers a large range in distance .
the equation of motion of a stationary stellar wind is @xmath17 where @xmath18 is the radiative acceleration .
together with the mass continuity equation @xmath19 and the expression for the gas pressure @xmath20 , where @xmath21 is the gas constant and @xmath22 is the mean mass per free particle in units of @xmath23 , we find the equation of motion @xmath24 where @xmath25 is the isothermal speed of sound .
for simplicity we have assumed that the atmosphere is isothermal . in this expression
the effective mass @xmath26 is corrected for the radiation pressure by electron scattering .
@xmath16 is the line acceleration .
the equation has a singularity at the point where @xmath27 , this critical point is the sonic point .
if the line acceleration @xmath28 is known as a function of @xmath29 , the equation can be solved numerically .
a smoothly accelerating wind solution requires that the numerator of eq .
[ eq : vdvdr1 ] reaches zero exactly at the sonic point where the denominator vanishes .
it should be stated that _ this _ critical point ( sonic point ) at @xmath30 and @xmath31 20 km s@xmath32 is _ not _ the same as the cak critical point .
the cak critical point is located much further out in the wind at @xmath33 and about @xmath34 . if the line acceleration @xmath16 in eq .
[ eq : vdvdr1 ] were to be rewritten as a function of velocity gradient instead of radius , then one would find the cak critical point .
pauldrach et al .
( 1986 ) showed that if the finite disk correction to the cak theory is applied , then the modified cak critical point moves inward and is located at @xmath35 and at @xmath36 km s@xmath32 .
this is much closer to the sonic point !
although the ( modified ) cak critical solution may well provide the correct mass - loss rate and terminal velocity , there is concern about its physical reality ( see e.g. lucy 1998 and lamers & cassinelli 1999 for a thorough discussion ) .
lucy ( 1998 ) has given arguments favouring the sonic point as the physical more meaningful critical point .
we will use the sonic point as the physically relevant critical point .
this is the point where the mass - loss rate is fixed . throughout the paper we will therefore refer to the _ subsonic _ part of the wind for the region close to the photosphere where the mass loss is determined , and to the _ supersonic _
part for the region beyond the sonic point where the mass - loss rate is already fixed , but the velocity has still to be determined
. the critical solution can be found by numerically integrating eq .
[ eq : vdvdr1 ] , starting from some lower boundary @xmath37 in the photosphere , with pre - specified values of @xmath38 and @xmath39 and with a trial value of @xmath40 . the value of @xmath40 that produces a velocity law that passes smoothly through the critical point is the correct value .
alternatively , for a non - isothermal wind with a pre - specified @xmath41-relation , one can integrate inwards from the critical point with an assumed location @xmath42 , and then adjust this value until the inward solution gives a density structure that reaches @xmath43 at the location where @xmath44 ( e.g. see pauldrach et al .
the critical solution specifies the values of @xmath45 , @xmath39 ( given by @xmath46 ) and @xmath40 at the lower boundary .
this fixes the value of via the mass continuity equation ( eq . [ eq : continuity ] ) .
note that is determined by the conditions in the _ subsonic _ region
! we will show below that an _ increase _ in @xmath28 in the subsonic region results in an _ increase _ in .
this can be understood because in the _ subsonic _ region , where the denominator of eq .
[ eq : vdvdr1 ] is negative , an increase in @xmath16 gives a smaller velocity gradient .
integrating from the sonic point inwards to the lower boundary with a smaller velocity gradient , implies that the velocity the lower boundary should be higher and hence the mass - loss rate , @xmath47 , must be higher . on the other hand ,
an increase in @xmath16 in the _ supersonic _ region , yields a larger velocity gradient and this would directly increase the terminal velocity .
another way to understand how an increase in is caused by an increase in @xmath16 below the sonic point , is based on the realization that the density structure of the subsonic region is approximately that of a static atmosphere .
this can be seen in eq .
[ eq : motion ] .
since the term @xmath48 is much smaller than the acceleration of gravity , it can approximately be set to zero in the subsonic region .
( this is not correct close to the sonic point . ) in an isothermal static atmosphere the density structure follows the pressure scaleheight . adding an extra outward force in the subsonic region results in an increase of the pressure - scaleheight and hence in a slower outward decrease in density .
this means that just below the sonic point , where @xmath49 , the density @xmath50 will be higher than without the extra force . applying the mass continuity equation ( eq . [ eq : continuity ] ) at the sonic point
then shows that the mass - loss rate will be higher than without the extra force in the subsonic region .
( see lamers & cassinelli 1999 for a thorough discussion ) .
a simple numerical experiment serves to demonstrate the dependence of on the radiative acceleration in the subsonic region .
we start with an isothermal model of the wind from a star of @xmath51 , @xmath52 , @xmath1 = 25 000 k , @xmath53@xmath54 k. we then specify the line acceleration @xmath28 in such a way that it produces a stellar wind with a mass - loss rate of @xmath55 and with a @xmath56-type wind velocity law @xmath57 where @xmath56 = 1 and @xmath58 km s@xmath32 .
( this @xmath28 is found by solving eq .
[ eq : vdvdr1 ] with this fixed velocity law ) .
this model is very similar to one of the models near the bi - stability jump that we will calculate in detail in sect .
[ sec : predictions ] . as a lower boundary
we choose the point where @xmath59 g @xmath60 at @xmath61 .
figure [ f_testbumps ] shows the resulting variation of @xmath28 . adopting this variation of @xmath28 and solving the momentum equation with the condition that the solution goes smoothly through the sonic point , we retrieve the input mass - loss rate and input velocity law , as one would expect .
the sonic point is located at @xmath62 where @xmath63 km s@xmath32 , and where @xmath64 cm s@xmath65 .
let us study what happens to and if we change the line acceleration in the subsonic region . to this purpose
we add a gaussian `` bump '' to @xmath28 .
this bump is characterized by @xmath66 where @xmath67 , @xmath68 describes the location of the peak at @xmath69 and @xmath70 gives the width of the bump ( @xmath71 ) . the line acceleration with the extra bumps
is shown in fig .
[ f_testbumps ] .
the solution of the momentum equation , with the condition that it passes smoothly through the sonic point , gives the velocity at the lower boundary and hence the mass - loss rate .
the upper panel of fig .
[ fig : models ] shows the resulting mass - loss rates as a function of the peak value of the bump in the line acceleration in the subsonic region .
we see that as the line acceleration in the subsonic region increases , increases .
once is fixed by the processes in the subsonic region , the radiative acceleration in the supersonic region then determines the terminal velocity that the wind will reach .
this can easily be seen in the following way . integrating the momentum equation
( eq . [ eq : motion ] ) in the supersonic region from the critical point @xmath42 to infinity , and ignoring the influence of the gas pressure , gives @xmath72 so @xmath73 here we have used the observed property that @xmath74 and that @xmath75 , so @xmath76 .
[ eq : vinfty ] says that is determined by the integral of @xmath28 in the _ supersonic _ region . the radiative acceleration in the _ supersonic _ part of
the wind will _ decrease _ as is forced to increase by an _ increase _ in the radiative acceleration in the _ subsonic _ part of the wind .
this is because the optical depth of the optically thick driving lines , which is proportional to the density in the wind , will increase .
thus an increase in results in an increase of the line optical depth .
this results in a decrease of @xmath16 in the supersonic region , which gives a lower terminal velocity .
we will estimate this effect below .
assume that the radiative acceleration by lines depends on the optical depth in the wind , as given by cak theory ( castor et al .
1975 ) .
@xmath77 where @xmath78 and @xmath79 are constants and @xmath80 is a reference value describing the acceleration due to electron scattering .
it is given by @xmath81 .
the optical depth parameter is @xmath82 where @xmath83 is the mean thermal velocity of the protons .
let us define @xmath84 as the radiative acceleration in the _ supersonic _ part of the initial wind model , i.e. without the increased mass - loss rate due to the bump in the subsonic region , and @xmath28 as the radiative acceleration of the model with the increased . from eqs .
[ eq : cak ] and [ eq : tcak ] with eq .
[ eq : continuity ] we find that @xmath85 where the superscript `` init '' refers to the initial model .
let us now compare the terminal velocities of the initial model without the bump , to that with the increased mass - loss rate due to the bump , in a simple but crude way , by solving the momentum equation in the supersonic part of the wind . if we neglect the terms due to the gas pressure and due to the gravity , the momentum equation in the supersonic part of the wind reduces to @xmath86 solving the equation for the initial model and the model with the increased results in the following expression @xmath87 so the ratio between the terminal velocities of the models with and without the increased mass - loss rate is @xmath88 where we adopted @xmath89 ( pauldrach et al .
1986 ) for the last expression .
we see that will decrease roughly as @xmath90 when the mass - loss rate increases .
the result is shown in the lower panel of fig .
[ fig : models ] . we realize that this numerical test is a drastic simplification of the real situation : ( a ) we have assumed an isothermal wind ; ( b ) we have taken the lower boundary at a fixed density ; ( c ) we have ignored possible changes in the ionization of the wind due to changes in and ( d ) we have ignored the role of the gas pressure and of gravity in estimating the change in .
however , this simple test serves the purpose of explaining _ qualitatively _ that the mass - loss rate depends on the radiative acceleration in the subsonic part of the wind only , and that an increase in the mass - loss rate due to an increase of @xmath16 in the subsonic region will also be accompanied by a decrease in . in the rest of the paper , we will _ quantitatively _ calculate radiative accelerations and mass - loss rates with a method which will be described in sect .
[ sec : method ] .
thus , an _ increase _ in the radiative acceleration in the _ subsonic _ region of the wind results in an _ increase _ of and a _ decrease _ in .
so , in order to understand the origin of the bi - stability jump of radiation driven winds , and to predict its effect on and , we should pay close attention to the calculated radiative acceleration in the _ subsonic _ part of the wind .
in order to understand the nature of the bi - stability jump , we calculate a series of radiation driven wind models for supergiants in the range of = 12 500 to 40 000 k. the calculation of the radiative acceleration of the winds requires the computation of the contributions of a very large number of spectral lines . to this end
, we first calculate the thermal , density and ionization structure of a wind model computed with the non - lte expanding atmosphere code isa - wind ( de koter et al .
1993)(for details , see sect .
[ sec : isa ] ) .
we then calculate the radiative acceleration by following the fate of a very large number of photons that are released from below the photosphere into the wind , by means of a monte carlo technique . in this section ,
we describe the basic physical properties of the adopted monte carlo ( mc ) technique which was first applied to the study of winds of early - type stars by abbott & lucy ( 1985 ) .
then , we describe the calculation of the radiative acceleration by lines with the mc method , and finally the method for calculating theoretical mass - loss rates .
the lines in the mc method are described in the sobolev approximation .
this approximation for the line acceleration is valid if the physical conditions over a sobolev length do not change significantly , i.e. @xmath91 where _ f _ is any physically relevant variable for the line driving , e.g. density , temperature or ionization fraction .
@xmath92 is a combination of thermal and turbulent velocities .
[ eq : sob ] shows that the validity range of the sobolev approximation is in practice somewhat arbitrary , since it depends on the value of the turbulent velocity which is poorly known .
nevertheless , the sobolev approximation is often used ( e.g. abbott & lucy 1985 ) and we will also adopt it in calculating the line acceleration and mass loss , mainly because of computational limitations .
we can not exclude that due to the use of the sobolev approximation we may predict quantitatively inaccurate line accelerations below the sonic point .
however , _ if _ an exact treatment would be followed , then this is expected to have a systematic effect on the line acceleration for _ all _ models .
therefore , we do not expect our conclusions regarding the origin of the bi - stability jump to be affected .
the sobolev approximation implies that for scatterings in the frame co - moving with the ions in the wind ( co - moving frame , cmf ) , the incident and emerging frequencies are both equal to the rest frequency of the line transition @xmath93 in the cmf . @xmath94 where @xmath95 and @xmath96 are the incident and emerging frequencies in the cmf .
in terms of quantities seen by an outside observer , these two cmf frequencies are given by : @xmath97 and @xmath98 where @xmath99 and @xmath100 are the incident and emergent frequencies for an outside observer ; @xmath101 and @xmath102 are the direction cosines with respect to the radial flow velocity of the photons at the scattering point and @xmath103 is the radial flow velocity of the scattering ion for an outside observer .
thermal motions of the scattering ions are assumed to be negligible compared to the motion of the outward flow .
note that we adopted the same velocity @xmath103 for the ion before and after the photon interaction ( eqs .
[ e_nuin ] and [ e_nuout ] ) .
this is justified since the change in velocity due to the transfer of momentum from a photon to an ion is very small , i.e. about @xmath104 cm s@xmath32 per scattering .. therefore , the change in @xmath99 and @xmath100 is mainly determined by a change in direction angle .
combining eqs .
[ e_nuin ] and [ e_nuout ] gives the conservation of co - moving frequency in a scattering event ( abbott & lucy 1985 ) .
@xmath105 because the energy and momentum of a photon are @xmath106 and @xmath107 , the equation can be rewritten in the following way : @xmath108 eq .
[ e_pe ] links the change in radial @xmath109 of a photon in an interaction with an ion with velocity @xmath103 to the @xmath110 loss of the photon . in order to determine the line acceleration @xmath16
we will need to derive the momentum transfer from the photons to the _ ions _ in the wind . for an outside observer ,
the conservation of radial momentum is : @xmath111 where @xmath112 is the mass of the moving ion and @xmath113 and @xmath114 are the radial velocities of the ion just before and after the scattering . for an outside observer : @xmath115 and @xmath116 again , the change in frequency is dominated by the change in direction angle .
so the change in radial velocity per scattering , @xmath117 , is small compared to @xmath103 and is given by @xmath118 since @xmath119 , eq .
[ eq : deltavlarge ] becomes @xmath120 this relation describes the velocity increase of the ion depending on the directions @xmath101 and @xmath102 of the photon . in case
@xmath101 = @xmath102 then @xmath121 = 0 , as one would expect .
the increase in the radial momentum @xmath122 of the scattering ion is now given by : @xmath123 where @xmath124 is the loss of _ radiative _ energy .
this equation shows that the increase in the momentum of the _ ions _ can be calculated from the loss of energy of the _ photons _ when these are followed in their path through the wind by means of the monte carlo method . multiplying both sides of eq .
[ e_gp ] by @xmath103 and using the fact that for each scattering @xmath125 so @xmath126 , gives : @xmath127 equation [ e_energy ] says that the gain of kinetic energy of the ions in the radial direction equals the energy loss of the photons .
the radiative acceleration as a function of distance is calculated by means of the mc technique by following the fate of the photons using the program mc - wind ( de koter et al .
1997 ) . in the calculation of the path of the photons we have properly taken into account the possibility that the photons can be scattered or absorbed & re - emitted due to true absorption or eliminated because they are scattered back into the star .
the radiative transfer in mc - wind is calculated in the sobolev approximation .
multiple line and continuum processes are included in the code .
the continuum processes included are electron scattering and thermal absorption and emission .
the line processes included are photon scattering and photon destruction by collisional de - excitation . in deciding
whether a continuum or a line event takes place , we have improved the code in the following way : the key point of the monte - carlo `` game '' is that _ line _ processes can only occur at specific points in each shell of the stellar wind , whereas _ continuum _ processes can occur at any point .
the correct way of treating the line and continuum processes is by comparing a random optical depth value to the _ combined _ optical depth for line and continuum processes along the photon s path .
first , this combined optical depth is compared to a random number to decide whether a continuum or a line process takes place .
this first part of the treatment is basically the same as described by mazzali & lucy ( 1993 ) for the case of line and electron scatterings only .
now , after it has been decided that the process will be a continuum process , a second random number is drawn to decide _ which _ continuum process will take place , an electron scattering or absorption .
the radiative acceleration of the wind was calculated by following the fate of the photons emitted from below the photosphere with the mc technique . to this purpose
the atmosphere is divided into a large number of concentric , thin shells with radius @xmath29 , thickness @xmath128 containing a mass @xmath129 .
the loss of photon energy due to all scatterings that occur within each shell are calculated to retrieve the total line acceleration @xmath28 per shell .
the total line acceleration per shell summed over all line scatterings in that shell equals @xmath130 where @xmath131 is the momentum of the ions in the shell .
the momentum gained by the ions in the shell is equal to the momentum lost by the photons due to interactions in that shell .
using the relationship between @xmath129 and @xmath128 for thin concentric shells , @xmath132 , and the derived relation between momentum and energy transfer of the photons @xmath133 ( eq . [ e_gp ] ) , @xmath28 can be rewritten as @xmath134 where @xmath135 is sum of the energy loss of all the photons that are scattered in the shell . now using mass continuity ( eq .
[ eq : continuity ] ) and the fact that the total energy transfer @xmath136 divided by the time interval @xmath137 equals the rate at which the radiation field loses energy , @xmath138 , i.e. @xmath139 , the expression for @xmath28 , which is valid for each shell , simply becomes ( abbott & lucy 1985 ) @xmath140 the line list that is used for the mc calculations consists of over @xmath141 of the strongest lines of the elements h - zn from a line list constructed by kurucz(1988 ) .
lines in the wavelength region between 50 and 7000 are included in the calculations with ionization stages up to stage vi . typically about @xmath142 photon packets , distributed over the spectrum at the lower boundary of the atmosphere
were followed for each model , i.e. for each adopted set of stellar and wind parameters . for several more detailed models we calculated the fate of @xmath143 photon packets .
the wind was divided in about 50 - 60 concentric shells , with many narrow shells in the subsonic region and wider shells in supersonic layers .
the division in shells is essentially made on the basis of a rosseland optical depth scale .
typical changes in the logarithm of this optical depth are about 0.13 .
we predict the mass - loss rates for a grid of model atmospheres to study the behaviour of near the bi - stability jump . for a given set of stellar parameters we calculate the mass loss in the following way : 1 . for fixed values of @xmath144 , @xmath1 , @xmath145 and
we adopt several values of the input mass loss @xmath146 ( within reasonable bounds predicted by cak theory ) .
2 . for each model
we adopt a wind with a terminal velocity of 1.3 , 2.0 or 2.6 times the effective escape velocity , given by @xmath147 a @xmath56-type velocity law with @xmath148 was adopted , appropriate for ob stars ( groenewegen & lamers 1989 ; puls et al . 1996 ) 3 . for each set of stellar and wind parameters we calculate a model atmosphere with isa - wind ( see sect . [
sec : isa ] ) .
this code gives the thermal structure , the ionization and excitation structure and the population of the energy levels of all relevant ions .
4 . for each model the radiative acceleration was calculated with the mc - wind program that uses the monte carlo method described above .
5 . for each set of stellar parameters and for
each adopted value of @xmath15 we check which one of the adopted mass - loss rates is consistent with the radiative acceleration .
this consistency was checked in the following way : + neglecting the term due to the gas pressure , one can write the equation of motion in the following way : + @xmath149 + using the expression for the line acceleration ( eq . [ eq : gl ] ) and integrating the equation of motion ( eq . [ eq : eqmotion ] ) from the stellar surface to infinity gives + @xmath150 + @xmath151 , is the total amount of radiative energy , summed over all the shells , that is lost in the process of line - interaction and is transfered into kinetic energy of the ions as given in eq .
[ e_energy ] .
equation [ eq : consistency ] states that the momentum transfered from the radiation into the wind is used to lift the mass loss out of the potential well and to accelerate the wind to . only _
one _ value of will satisfy this equation ( lucy & abbott 1993 ) .
this is the predicted mass - loss rate .
we note that eq .
[ eq : consistency ] only describes the``global '' consistency of the mass - loss rate with the radiative acceleration . for the set - up of the model atmosphere the velocity law @xmath152
is needed as input .
this means that although the calculation is globally consistent in terms of kinetic wind energy , the velocity is not necessarily locally consistent , since the equation of motion is not solved .
instead , we have used observed values for and @xmath56 for the velocity law . since the total amount of radiative energy in eq .
[ eq : consistency ] is mainly determined in the _ supersonic _ region , where the sobolev approximation is an excellent approximation , @xmath153 is accurately calculated .
this implies that if one adopts the correct values for the terminal velocity , one may predict accurate values for !
the calculation of the mass - loss rates by the method described in the previous section requires the input of a model atmosphere , before the radiative acceleration and can be calculated .
the model atmospheres used for this study are calculated with the most recent version of the non - lte unified improved sobolev approximation code isa - wind for stars with extended atmospheres . for a detailed description of this code
we refer to de koter et al .
( 1993 , 1997 ) . here
, we just make a few relevant remarks .
isa - wind treats the atmosphere in a unified manner , i.e. no artificial separation between photosphere and wind is assumed . this is distinct from the so - called `` core - halo '' approaches . in the photosphere
the density structure follows from a solution of the momentum equation taking into account gas and radiative pressure on electrons .
the velocity law follows from this density structure via the mass continuity equation . near the sonic point
, a smooth transition is made to a @xmath56-type velocity law for the supersonic part of the wind ( see eq .
[ eq : betalaw ] ) . the temperature structure in the wind
is computed under the assumption of radiative equilibrium in an extended grey lte atmosphere .
the temperature in the wind is not allowed to drop below a certain minimum value @xmath154 ( drew 1989 ) .
finally , the chemical species included explicitly in the non - lte calculations are h , he , c , n , o and si . the complexity of the model atoms is similar to that used by de koter et al .
( 1997 ) . for the iron - group elements , which are important for the radiative acceleration
, we calculate the ionization / excitation equilibrium in the _ modified nebular approximation _ ( see schmutz 1991 ) . in this representation
the ionization equilibrium is given by @xmath155 where @xmath156 and @xmath157 are
the electron density and temperature , @xmath158 and @xmath159 are the ion population numbers , @xmath160 = is the radiation temperature of ion @xmath161 at radial depth @xmath29 , and @xmath162 is a geometrical dilution factor as defined by schmutz et al .
the last factor of eq .
[ eq : nebular ] is the lte ionization ratio for a temperature .
the parameter @xmath163 , introduced by abbott & lucy ( 1985 ) , represents the fraction of recombinations going directly to the ground state .
the values of are obtained by inverting the above equation , using all 19 ionization ratios available from the isa - wind calculation .
the radiation temperature of an explicit ion is used that has its ionization potential closest ( but lower ) to that of the metal ion of interest .
for instance , the n ii / iii ratio is used to define the ionization equilibrium of fe iii / iv .
the excitation state of metastable levels is assumed to be in lte relative to the ground state .
for all other levels we adopt `` diluted '' lte populations , defined by @xmath164 where @xmath165 and @xmath166 are the excitation population numbers for the upper and lower levels .
clearly , the simplified treatment of the iron - group metals is prompted by the computationally intensive nature of the problem at hand .
it needs to be improved in the future , but we do not expect that our conclusions regarding the nature of the bi - stability jump would be affected .
( we return to this in the discussion in sect . [
sec : concl ] ) .
using the procedure as described in sect .
[ sec : determination ] , we calculated mass - loss rates for stars with a luminosity of @xmath167 and a mass of @xmath168 .
the models have effective temperatures between 12 500 and 40 000 k with a stepsize of 2500 k. these parameters are approximately those of ob supergiants , for which lamers et al .
( 1995 ) found the bi - stability in .
we calculated for wind models with a @xmath56-type velocity law with @xmath169 ( eq . [ eq : betalaw ] ) for three values of the @xmath0 = 2.6 , 2.0 and 1.3 .
lamers et al .
( 1995 ) found that @xmath170 for stars of types earlier than b1 , and @xmath171 for stars of types later than b2 .
for the determination of @xmath172 we used the effective mass @xmath173 , with @xmath174 .
the stellar parameters for the calculated grid are indicated in table [ t : parameters ] .
the models are calculated for solar metallicities .
the results are listed in table [ t : parameters ] .
this table gives the values of , @xmath145 , and for each temperature and for the three values of @xmath0 .
we also give the value of the wind efficiency factor @xmath175 , which describes the fraction of the _ momentum _ of the radiation that is transferred to the ions @xmath176 the fraction of the photon _ energy _ that is transferred into kinetic energy of the ions is also listed ( in column 8) .
the values for this @xmath110 efficiency number @xmath177 are a factor of about @xmath178 smaller than the wind _ momentum _ efficiency number @xmath175 , which is given in column ( 7 ) .
this is because a photon transfers a large fraction of its momentum during a scattering , but only a very small fraction ( of order @xmath179 ) of its energy .
the last column of table [ t : parameters ] marks three models that will be discussed in more detail in sect .
[ sec : origin ] .
ccrrrcccc @xmath0 & @xmath1 & @xmath180 & @xmath172 & @xmath15 & log & @xmath175 & @xmath177 & model + & ( k ) & ( @xmath181 & ( km s@xmath32 ) & ( km s@xmath32 ) & ( @xmath182/yr ) & & ( in 10@xmath183 ) & + 1.3 & 12 500 & 67.7 & 310 & 410 & - 6.32 & 0.095 & 0.103 & + & 15 000 & 47.0 & 380 & 490 & - 6.39 & 0.097 & 0.126 & + & 17 500 & 34.5 & 440 & 570 & - 6.28 & 0.146 & 0.221 & + & 20 000 & 26.4 & 500 & 650 & - 6.22 & 0.192 & 0.332 & + & 22 500 & 20.9 & 560 & 730 & - 6.15 & 0.254 & 0.493 & + & 25 000 & 16.9 & 630 & 810 & - 6.12 & 0.302 & 0.653 & c + & 27 500 & 14.0 & 690 & 900 & - 6.40 & 0.174 & 0.414 & + & 30 000 & 11.8 & 750 & 980 & - 6.58 & 0.126 & 0.326 & + & 32 500 & 10.0 & 810 & 1060 & - 6.58 & 0.136 & 0.382 & + & 35 000 & 8.6 & 880 & 1140 & - 6.43 & 0.207 & 0.626 & + & 37 500 & 7.6 & 940 & 1220 & - 6.37 & 0.255 & 0.826 & + & 40 000 & 6.6 & 1000 & 1300 & - 6.26 & 0.350 & 1.210 & + 2.0 & 12 500 & 67.7 & 310 & 630 & - 6.74 & 0.056 & 0.073 & + & 15 000 & 47.0 & 380 & 750 & - 6.62 & 0.088 & 0.138 & + & 17 500 & 34.5 & 440 & 880 & - 6.49 & 0.139 & 0.254 & + & 20 000 & 26.4 & 500 & 1000 & - 6.41 & 0.191 & 0.398 & + & 22 500 & 20.9 & 560 & 1130 & - 6.32 & 0.264 & 0.620 & + & 25 000 & 16.9 & 630 & 1250 & - 6.48 & 0.203 & 0.530 & + & 27 500 & 14.0 & 690 & 1380 & - 6.73 & 0.125 & 0.360 & + & 30 000 & 11.8 & 750 & 1500 & - 6.76 & 0.128 & 0.400 & + & 32 500 & 10.0 & 810 & 1630 & - 6.71 & 0.155 & 0.527 & + & 35 000 & 8.6 & 880 & 1750 & - 6.59 & 0.220 & 0.801 & + & 37 500 & 7.6 & 940 & 1880 & - 6.57 & 0.247 & 0.969 & + & 40 000 & 6.6 & 1000 & 2000 & - 6.48 & 0.325 & 1.356 & + 2.6 & 12 500 & 67.7 & 310 & 810 & - 6.95 & 0.045 & 0.070 & + & 15 000 & 47.0 & 380 & 980 & - 6.85 & 0.067 & 0.126 & + & 17 500 & 34.5 & 440 & 1140 & - 6.69 & 0.114 & 0.248 & + & 20 000 & 26.4 & 500 & 1300 & - 6.54 & 0.184 & 0.458 & + & 22 500 & 20.9 & 560 & 1460 & - 6.59 & 0.184 & 0.517 & + & 25 000 & 16.9 & 630 & 1630 & - 6.79 & 0.129 & 0.403 & b + & 27 500 & 14.0 & 690 & 1790 & - 6.95 & 0.098 & 0.337 & a + & 30 000 & 11.8 & 750 & 1950 & - 6.92 & 0.115 & 0.430 & + & 32 500 & 10.0 & 810 & 2120 & - 6.86 & 0.143 & 0.579 & + & 35 000 & 8.6 & 880 & 2280 & - 6.76 & 0.194 & 0.845 & + & 37 500 & 7.6 & 940 & 2440 & - 6.71 & 0.233 & 1.089 & + & 40 000 & 6.6 & 1000 & 2600 & - 6.68 & 0.266 & 1.327 & + the results are plotted in fig .
[ fig : bistability ] . for each of the three values of @xmath0
the value of is decreasing for decreasing @xmath1 between 40 000 and 30 000 k and also between 22 500 and 12 500 k. between about @xmath184 k and @xmath185 k ( slightly dependent on @xmath0 ) the mass loss _ increases _ with decreasing .
these increments in roughly coincide in @xmath1 with the observed bi - stability jump in @xmath0 near spectral type b1 , at about 21 000 k. for the ratio of @xmath0 = 2.6 , the increase in between model a and b equals 45 % . we know from the observations that @xmath0 jumps from 2.6 at the hot side of 21 000 k to 1.3 at the cool side of 21 000 k ( lamers et al .
1995 ) . including this observed jump in @xmath0 in the mass - loss predictions ,
provides an even steeper increase in from models a and b to the smaller value of @xmath0 = 1.3 , as is shown in the lower part of fig .
[ fig : bistability ] .
this figure shows an increase in of about a factor of five between @xmath184 and 20 000 k. this is our prediction for a bi - stability jump in @xmath7 .
the exact position of @xmath1 of the bi - stability jump in fig .
[ fig : bistability ] is somewhat ambiguous , since @xmath15 is adopted from observations , and does not directly follow from our models . for a discussion on the exact position of the jump in @xmath1 , see sect .
[ sec : concl ] . to test the sensitivity of our predictions of mass - loss rates for different shapes of the velocity law
, we calculated another series of models with @xmath56 = 1.5 .
since the differences are only about 10 % , we conclude that the predicted mass - loss rates are only marginally sensitive to the _ shape _ of the adopted velocity law .
another view at these results can be obtained by plotting the wind efficiency factor @xmath175 .
figure [ f_eta ] shows the behaviour of @xmath175 as a function of @xmath1 for the same grid of models as was presented for the mass - loss rates in the upper panel of fig .
[ fig : bistability ] . fig .
[ f_eta ] clearly shows that @xmath175 is not a constant function of @xmath1 .
the overall picture shows that for the three values of @xmath0 , @xmath175 decreases as @xmath1 decreases .
this is probably due to the fact that the maximum of the flux distribution shifts to longer wavelengths . at @xmath186 1800
there are significantly less lines than at @xmath187 1800 .
therefore , radiative acceleration becomes less effective at lower @xmath1 . in the ranges of @xmath188 and @xmath189 k
, @xmath175 is almost independent of the adopted value for @xmath0 .
this means that the behaviour of @xmath175 is intrinsically present in the model calculations and does not depend on the values adopted for @xmath0 . in the range of @xmath190
k , the situation is reversed .
@xmath175 now _ increases _ by a factor of 2 to 3 .
this means that the wind momentum loss , @xmath14 is _ not _ constant over the jump , but instead , jumps by a factor of 2 - 3 also .
since @xmath15 drops by a factor of about two , is expected to jump by a factor of about five , which was already shown in the lower panel of fig .
[ fig : bistability ] .
the behaviour of @xmath175 as a function of is not exactly the same for the three different series of models .
first , the size of the jump is different .
second , the jump occurs at somewhat different temperatures .
this is no surprise , since the ionization equilibrium does not only depend on @xmath191 , but on @xmath50 as well , a smaller value of the velocity @xmath15 , means a larger density @xmath50 in the wind .
hence , the jump is expected to start at a larger value of @xmath1 for a smaller value of @xmath0 .
this behaviour for the position of @xmath1 of the jump can be seen in in fig .
[ fig : bistability ] and in @xmath175 in fig .
[ f_eta ] .
in the previous section we have shown that the mass - loss rate increases around @xmath192 k. the next step is to investigate the physical process that causes the bi - stability jump .
therefore , we will look into the details of the line acceleration @xmath28 for three models around the bi - stability jump . for these models ( a , b and c in table [ t : parameters ] and fig . [ fig : bistability ] ) we made improved monte - carlo calculations , using @xmath193 packets of photons , to derive more details about the radiative acceleration .
first , we will investigate the line acceleration @xmath28 of the model at the hot side of the bi - stability jump .
this model a with @xmath184 k and @xmath0 = 2.6 , is our basic model .
then , we will compare model a to model b that has the same @xmath0 , but is situated on the cool side of the bistability jump , where @xmath1 = 25 000 k. by comparing models a and b , we can investigate the intrinsic increase in of 45 % in our model calculations due to the lower @xmath1 .
the next step is to compare @xmath28 of model b and model c which also has @xmath1 = 25 000 k , but a smaller ratio @xmath194 . by comparing model b and c ,
we can obtain information about the effects of a jump in @xmath15 .
finally , we check our approach for self - consistency by simultaneously calculating the mass - loss rate _ and _ terminal velocity .
model a has a mass - loss rate of log = -6.95 .
the behaviour of the line acceleration as a function of the distance from the stellar surface , @xmath28 is shown in fig .
[ f_modela ] .
the sonic point is reached at a distance of 1.025 @xmath145 .
it is clear that most of the line driving is produced far beyond the sonic point .
but , as was explained in sect . [ sec : simple ] the important region that determines the mass - loss rate is @xmath195 the sonic point
therefore , the part of the atmosphere around the sonic point is enlarged in fig .
[ f_modela](b ) . to investigate the origin of the jump ,
it is useful to know which elements are effective line drivers in which part of the stellar wind . therefore , extra monte - carlo calculations were performed .
the first extra monte - carlo simulation was performed with a line list containing only fe lines .
the second one was performed with a line list containing the lines of the elements c , n and o. figure [ f_modela](b ) shows that fe is the main line driver below the sonic point .
c , n and o , are important line drivers in the supersonic part of the wind , which can be seen in [ f_modela](a ) .
c , n and o contribute roughly 50 % of the line acceleration in the supersonic part of the wind . not indicated here , but relevant to mention is that si , cl , p and s are other important line drivers in the supersonic part of the wind .
ni was found @xmath196 to be an important line driver in any part of the stellar wind at all .
the mass - loss rate is determined by the radiative acceleration _ below _ the sonic point , and the terminal velocity is determined by the radiative acceleration in the _ supersonic _ part of the wind .
so our results show that the mass - loss rates of hot star winds are mainly determined by the radiation pressure due to fe !
the terminal velocities are mainly determined by the contributions of c , n and o. to understand the origin of the bi - stability jump in @xmath7 , we investigate the line acceleration due to fe . the ionization balance of fe for models a and b is plotted in fig .
[ f_ionization ] , top and bottom respectively .
the right hand figures show the enlargement of the ionization balance in the region near the sonic point . in model a ( = 27 500 k ) fe v has a maximum around @xmath197 , which can be seen in fig . [ f_ionization ] ( b )
then , due to the outward decreasing temperature , fe v decreases in favour of fe iv , which peaks around @xmath198 .
next , one may expect fe iv to decrease in favour of fe iii .
however , around @xmath199 fe iv re - ionizes due to a decrease of the density @xmath50 . in this region of the atmosphere , where @xmath200 is rapidly increasing , the effect of the decreasing @xmath50 is larger than the effect of the decreasing @xmath191 .
[ f_ionization ] ( b ) clearly shows that fe iv is the dominant ionization stage in the subsonic region of the stellar wind . in the region just below the sonic point ,
the ionization fraction of fe iv is 90 - 100 % whereas that of fe iii is less than 10 % .
however , this does not necessarily mean that fe iv is the main line driver . to investigate the contribution to the line acceleration @xmath16 of the different ionization stages of fe
some extra monte - carlo simulations were performed .
one simulation included only the lines of fe iii , another simulation included just the lines of fe iv .
the results for @xmath16 for fe iii and fe iv are plotted in fig .
[ f_afe ] .
it is surprising to note that , although fe iv is the dominant ionization stage throughout the wind , most of the driving is contributed by fe iii . below the sonic point fe
iii is clearly the most important iron line driver ( see fig .
[ f_afe](b ) ) . from the data shown in figures [ f_ionization ] and [ f_afe ] we conclude that the mass - loss rate of winds from stars with @xmath201 k is mainly determined by the radiative acceleration due to fe iii lines .
this suggests that the bi - stability jump is mainly due to changes in the ionization balance of fe .
we test this hypothesis in the next section . in the previous section
we have shown that the mass loss of model a is dominated by radiative acceleration due to fe iii lines . in this section
we investigate changes in the radiative acceleration due to fe as @xmath1 decreases .
this may explain the increase of near the bi - stability jump .
to this purpose we compare the ionization and @xmath16 of models a and b in detail .
the ionization balance of model b is shown in fig . [ f_ionization](c ) and ( d ) .
it shows that , due to a lower temperature , the decrease of the fe iv fraction drops to smaller values than for model a , which was shown in fig .
[ f_ionization](b ) .
the ionization fraction of fe iii below the sonic point in the case of model b is up to almost 40 % . to see whether this extra amount of fe iii can cause the increase in the line acceleration
, we must look at @xmath16 of fe for model b. since model a and b have different at the same @xmath202 , they have a different radiative surface flux .
the radiative acceleration will be proportional to this flux . in order to compare the values of @xmath16 of the two models , we scale the results to a flux of a @xmath1 = 25 000 k model .
so @xmath203 since @xmath204 for constant luminosity , this is also a scaling to the newtonian gravity of the models .
figure [ f_abfe ] shows the normalized @xmath16 of fe for the models a ( top ) and b ( bottom ) . the right hand figures show an enlargement of the region near the sonic point .
it shows that for model b @xmath16 of fe
iii around the sonic point is more than a factor two larger than for model a ( see figs .
[ f_abfe](b ) and ( d ) ) .
this extra amount of fe iii in model b causes an increase in the @xmath205 @xmath16 in the subsonic part of the wind also , as can be seen in fig .
[ f_abcabc](b ) .
_ we conclude that the increase in mass loss from model a to b is due to the larger radiative acceleration ( compared to the gravity ) of model b by a larger ionization fraction of fe iii below the sonic point .
_ now the effect of @xmath16 on @xmath15 will be examined .
therefore , model b is compared to model c. we remind that models b and c have the same , and hence the same radiative flux and gravity , but model c has a twice as small value of @xmath0 as model b. figure [ f_abcabc](a ) shows the normalized @xmath16 for models a , b and c. as expected , @xmath28 for model c is significantly smaller than @xmath28 for models a and b. this is obviously due to the smaller value of .
the integral @xmath206 in fig .
[ f_abcabc](a ) for model a and b is larger than for model c. the values of @xmath206 for the models are 2.34 @xmath207 10@xmath208 and 1.92 @xmath207 10@xmath208 @xmath209 s@xmath65 for models a and b respectively , and 6.12 @xmath207 10@xmath210 @xmath209 s@xmath65 for model c. using eq .
[ eq : vinfty ] and the values of from column ( 4 ) in table [ t : parameters ] , the output values for can be obtained from the values of the integral of @xmath16 .
the derived output values for for the models are = 2050 , 1860 and 920 km s@xmath32 respectively for the models a , b and c. these values are equal within 10 % to the input values for which were indicated in column ( 5 ) of table [ t : parameters ] .
we can conclude that a smaller value for is indeed consistent with a smaller value of the integral @xmath206 .
however , this is not an independent check , since the calculated line acceleration of optically thick lines ( in the sobolev approximation ) is inversely proportional to the sobolev optical depth which is proportional to @xmath211 .
hence , _ assuming _ a smaller terminal velocity will automatically result in a smaller calculated line acceleration .
cccccccccc @xmath1 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 & @xmath217 & @xmath218 & log @xmath219 & log @xmath220 + ( k ) & & & & & & & & ( @xmath182/yr ) & ( @xmath182/yr ) + 17 500 & 2.0 & 1.5 & 1.3 & 1.2 & 0.58 & 0.2065 & 1.2 & -6.21 & -6.27 + 30 000 & 2.0 & 2.5 & 2.4 & 2.4 & 0.85 & 0.0076 & 2.4 & -6.86 & -6.90 + in earlier sections we have demonstrated that the mass loss around the bi - stability jump increases . as we have used _ observed _ values for the ratio @xmath0 in our model calculations , we have not yet provided a self - consistent explanation of the observed bi - stability jump in @xmath0 . as a consistency test of our calculations and an attempt to explain the observed jump in the ratio @xmath0
, we proceeded to solve the momentum equation of line driven wind models around the bi - stability jump .
the approach we take is to combine predicted force multiplier parameters @xmath78 and @xmath79 ( see below ) from the monte carlo calculation with the analytical solution of line driven winds from cak .
we calculated the line acceleration @xmath16 for several models with different @xmath1 using the monte carlo method .
the values of @xmath16 were expressed in terms of the force multiplier @xmath221 ( eq .
[ eq : cak ] ) .
following cak we tried to express @xmath221 in terms of a power - law fit of the optical depth parameter @xmath222 ( eq . [ eq : tcak ] ) .
we found that in the range 20 000 @xmath223 27 500 , @xmath221 is not accurately fit by a power - law , since the ionization changes over this critical range in @xmath1 .
fortunately , _ just outside _ this temperature region , @xmath221 _ can _ be accurately represented in terms of @xmath78 and @xmath79 , i.e. @xmath224 therefore , we have calculated models with effective temperatures just below ( @xmath1 = 17 500 k ) and just above ( @xmath1 = 30 000 k ) this critical temperature range .
self - consistent values of @xmath15 and @xmath7 were thus found in the following way : 1 .
we started with an assumed ratio of @xmath0 = 2.0 ( see column ( 2 ) in table [ t : forcem ] ) .
the force multipliers @xmath225 were calculated and a power - law fit of the type eq .
[ eq : cakfit ] was derived .
the fit was found to be excellent in the important part of the wind between the sonic point and @xmath226 .
this yielded values of @xmath216 and @xmath217 .
next , the mass loss and terminal velocity were simultaneously calculated from these @xmath216 and @xmath217 parameters using the cak solution of the momentum equation .
note that the solution with the finite disk correction ( pauldrach et al .
1986 ) was _ not _ applied , since this is already properly taken into account in the values of @xmath216 and @xmath217 calculated in the monte carlo technique ( see sect .
[ sec : method ] ) .
the superscript , mc , to the force multiplier parameters was added to avoid confusion with @xmath78 and @xmath79 for a point - like source used by e.g. kudritzki et al .
the ratio @xmath0 can be derived from the simple cak formulation : + @xmath227 + the value for @xmath216 for the model of 30,000 k is significantly higher than values for @xmath79 that were calculated before ( e.g. pauldrach et al .
1986 ) , since the finite disk is already included in the @xmath228-parameter !
the new calculated terminal velocity ratio @xmath0 ( column ( 3 ) of table [ t : forcem ] ) was used in the next iteration .
new mass - loss rates were calculated from the mc approach using the procedure as explained in sect.[sec : determination ] .
the mass - loss rates are equal within 15 % to the mass - loss rates that can be calculated using the expression for of cak using @xmath216 and @xmath217 .
5 . the above procedure ( step 1 . through 4 . )
was repeated until convergence was reached . after four iterations ,
the ratio @xmath0 did not change anymore .
the intermediate values of @xmath0 are given in columns ( 3 ) , ( 4 ) and ( 5 ) of table [ t : forcem ] .
the final value for the ratio @xmath0 is given in column ( 8) .
for the hot model ( = 30 000 k ) the final ratio @xmath0 equals 2.4 ; for the cool model ( = 17 500 k ) @xmath0 = 1.2 .
these values are within 10 % of the observed values of @xmath0 , i.e. 2.6 and 1.3 respectively .
cak mass - loss rates were also calculated from the resulting _ final _ force multiplier parameters @xmath217 and @xmath216 ( given in columns ( 6 ) and ( 7 ) of table [ t : forcem ] and the _ final _ mass - loss rates are given in column ( 9 ) of this table .
note that the values of are only marginally different from the mass - loss rates that were calculated from the monte carlo approach ( column ( 10 ) of table [ t : forcem ] ) . in summary
; we have self - consistently calculated values of @xmath15 and @xmath7 of two models located at either side of the bi - stability jump .
we have found a jump in terminal velocity @xmath0 of a factor of two , similar as observed by lamers et al .
. moreover , the mass - loss rates calculated from the cak formulation are consistent with those obtained from our monte carlo approach .
_ this implies that the origin of the observed change in the ratio @xmath0 of a factor of two around spectral type b1 is identical to the predicted jump in mass - loss rate of a factor of five due to the recombination of fe iv to fe iii . _ from the results and figures presented above we conclude that the mass - loss rate of early - b supergiants near the bi - stability jump is mainly determined by the radiative acceleration by iron .
although fe iv is the dominant ionization stage in the atmosphere of stars near 25 000 k , it is fe iii that gives the largest contribution to the subsonic line acceleration .
this is due to the number of effective scattering lines and their distribution in wavelengths , compared to the energy distribution from the photosphere .
this implies that the mass - loss rates of b - supergiants are very sensitive to the ionization equilibrium of fe in the upper photosphere .
our models show that the ionization fraction of fe iii increases drastically between = 27 500 and 25 000 k. this causes an increase in the line acceleration below the sonic point and in turn increases the mass loss near the bi - stability jump .
luminous blue variables ( conti 1984 ) are massive stars undergoing a brief , but important stage of evolution . during this period
they suffer severe mass loss with values of up to @xmath229 .
lbvs are characterized by typical variations in the order of @xmath230 of 1 to 2 magnitudes .
nevertheless , the total bolometric luminosity of the star @xmath202 seems to be about constant .
the reason for the typical lbv variations is still unknown . for reviews
see nota & lamers ( 1997 ) .
leitherer et al .
( 1989 ) and de koter et al .
( 1996 ) have shown that it must be the actual radius of the star that increases during these typical variations .
therefore , @xmath1 decreases during the variations , if @xmath202 is about constant . in this paper , we have calculated the mass - loss behaviour for normal ob supergiants as a function of . despite many differences between ob supergiants and lbvs
, we can retrieve valuable information about the behaviour of during a typical lbv variation by investigating the behaviour of normal ob supergiants , since both types of stars are located in the same part of the hrd .
our calculations can be used as a tool to understand the mass loss changes of an lbv in terms of changes in during such a typical variation ( see also leitherer et al .
1989 ) . observations of lbvs show that for some lbvs that undergo typical variations is increasing from visual minimum to maximum , while for others it is the other way around : is decreasing .
this `` unpredictable '' behaviour of during an lbv variation is not a complete surprise , if one considers our values as a function of @xmath1 .
we have found that in the ranges @xmath231 and @xmath232 , decreases for a decreasing , whereas in the interval between @xmath233 , increases for a decreasing .
this shows that whether one expects an increasing or decreasing during an lbv variation depends on the specific range in between visual minimum and maximum .
this was already suggested by lamers ( 1997 ) , albeit a constant value of @xmath175 was anticipated .
our present calculations can not be used to model the observed lbv variations , because we have assumed solar metallicities , whereas the lbvs are known to have an enhanced he and n abundance ( e.g. smith et al . 1994 ) . moreover , since most lbvs have already suffered severe mass loss in the past , their @xmath234 ratio will be higher than for normal ob supergiants .
this means that lbvs are closer to their eddington limit , which one may expect to have an effect on also .
these combined effects explain the lack of a consistent behaviour of for lbv variations so far . especially since it is not sure that @xmath202 really remains constant during the variations ( see lamers 1995 ) .
we have investigated the nature of the observed jump in the ratio @xmath0 of the winds of supergiants near spectral type b1 .
calculations for wind models of ob supergiants show that around @xmath4 the mass - loss rate jumps due to an increase in the line acceleration of fe iii below the sonic point .
this jump in is found in three different series of models . in all cases ,
the wind efficiency number @xmath235 increases significantly , by about a factor of 2 to 3 , if decreases from about 27 500 k to about 22 500 k. observations show that the ratio @xmath0 drops by a factor of two around spectral type b1 . applying these values for @xmath0 , we predict a bi - stability jump in of about a factor of five . so is expected to increase by about this factor between 27 500 and 22 500 k. we have argued that the mass loss is determined by the radiative acceleration in the subsonic part of the wind , i.e. below @xmath236 .
we found that this radiative acceleration is dominated by the contribution of the fe iii lines .
therefore is very sensitive to both the metal abundance and to the ionization equilibrium of fe .
our models show that the ionization fraction of fe iii and the subsonic radiative acceleration increases steeply between = 27 500 and 25 000 k. this explains the calculated increase in in this narrow temperature range .
the exact temperature of the bi - stability jump is somewhat ambiguous .
observations indicate that the jump occurs around spectral type b1 , corresponding to @xmath1 @xmath237 21 000 k ( lamers et al .
1995 ) . if one would not completely trust the value of @xmath0 for the star hd 109867 ( number 91 in lamers et al .
( 1995 ) ) , because of its relatively large error bar , then @xmath1 of the observed jump can easily occur at a few kk higher .
in fact we can not expect the bi - stability jump to occur at one and the same temperature for all luminosity classes , because the jump is sensitive to the ionization balance ( mainly of fe iii ) in the subsonic region of the wind and hence to the gravity of the star .
our models predict that the jump will occur near @xmath1 @xmath237 25 000 k. however , this is sensitive to the assumptions of the models : the adopted masses and luminosities and to the assumption of the modified nebular approximation for the calculation of the ionization equilibrium of iron ( see sect . [
sec : isa ] ) .
a more consistent treatment of the ionization and excitation equilibrium of the fe - group elements may have two effects : _
i ) _ predicted from @xmath153 may alter , and _
ii ) _ @xmath1 at which the ionization ratio of e.g. fe iii / iv flips , may shift . nevertheless , in view of the very encouraging results using the modified nebular approximation in the modeling of uv metal - line forests ( de koter et al .
1998 ) , we expect the error in at which the dominant ionization of fe switches from iv to iii to be at most a few kk .
furthermore , if a more consistent treatment would yield a change in this would most likely produce a systematic shift .
since we are essentially interested in relative shifts in @xmath7 , we do not expect that our conclusions regarding the nature of the bi - stability jump would be affected .
it is relevant to mention that leitherer et al .
( 1989 ) calculated atmospheric models for lbvs and suggested that the recombination of iron group elements from doubly to singly ionized stages , which according to them , occurs around = 10 000 k , can explain increases when lbvs approach their maximum states .
we have found a fe iii / iv ionization / recombination effect around = 25 000 k for normal supergiants .
we also anticipate that somewhere , at a lower value of a similar ionization / recombination effect will occur for fe ii / iii , causing a _
second _ bi - stability jump .
lamers et al .
( 1995 ) already mentioned the possible existence of a second bi - stability jump around = 10 000 k from their determinations of , but the observational evidence for this second jump was meagre .
possibly , this second jump _ is _ real and we anticipate that this second jump could very well originate from a fe ii / iii ionization / recombination effect .
furthermore , we have shown that the elements c , n and o are important line drivers in the _ supersonic _ part of the wind , whereas the _ subsonic _ part of the wind is dominated by the line acceleration due to fe .
therefore , we do _ not _ expect cno - processing to have a large impact on , but it might have impact on the terminal velocities . finally , we would like to add that our calculations for around @xmath4 have only been performed for _ one _ value of @xmath238 , @xmath202 and h / he abundance .
is expected to depend on these stellar parameters , so calculating mass - loss rates for a wider range of stellar parameters will provide valuable information on the size of the bi - stability jump in and and will allow us to constrain its amplitude and location in the hrd .
we thank an anonymous referee for several useful suggestions .
jv acknowledges support from the netherlands foundation for research in astronomy ( nfra ) with financial aid from the netherlands organization for scientific research ( nwo ) .
adk acknowledges support from a nwo `` pionier '' grant to l.b.f.m .
waters ; from nwo spinoza grant 08 - 0 to e.p.j .
van den heuvel and from a grant from the nasa astrophysics data program nra 95-adp-09 to s.r | we study the origin of the bi - stability jump in the terminal velocity of the winds of supergiants near spectral type b1 .
observations show that here the ratio @xmath0 drops steeply from about 2.6 at types earlier than b1 to a value of @xmath0=1.3 at types later than b2 . to this purpose
, we have calculated wind models and mass - loss rates for early - type supergiants in a @xmath1 grid covering the range between @xmath2 and @xmath3 .
these models show the existence of a jump in mass loss around @xmath4 for normal supergiants , with increasing by about a factor five from @xmath5 to 22 500 k for constant luminosity .
the wind efficiency number @xmath6 also increases drastically by a factor of 2 - 3 near that temperature .
we argue that the jump in mass loss is accompanied by a decrease of the ratio @xmath0 , which is the observed bi - stability jump in terminal velocity . using self - consistent models for two values of @xmath1
, we have derived @xmath0 = 2.4 for = 30 000 k and @xmath0 = 1.2 for = 17 500 k. this is within 10 percent of the observed values around the jump . up to now , a theoretical explanation of the observed bi - stability jump was not yet provided by radiation driven wind theory .
to understand the origin of the bi - stability jump , we have investigated the line acceleration for models around the jump in detail .
these models demonstrate that _
@xmath7 increases around the bi - stability jump due to an increase in the line acceleration of fe iii _ below _ the sonic point_. this shows that the mass - loss rate of b - type supergiants is very sensitive to the abundance and the ionization balance of iron .
furthermore , we show that the elements c , n and o are important line drivers in the _ supersonic _ part of the wind .
the _ subsonic _ part of the wind is dominated by the line acceleration due to fe .
therefore , cno - processing is expected _ not _ to have a large impact on but it might have impact on the terminal velocities . finally , we discuss the possible role of the bi - stability jump on the mass loss during typical variations of luminous blue variable stars . |
epitaxially grown graphene on sic offers one of the most promising platforms for applications in high speed electronic devices that might replace silicon in future.@xcite however , the quality of the two - dimensional electron gas in epitaxial graphene on sic still falls behind the electronic quality of mechanically exfoliated graphene.@xcite while the maximum charge carrier mobility of epitaxial graphene on si(0001 ) is in the order of 1000 @xmath0/vs at room temperature,@xcite the mobility reaches two orders of magnitude higher values in exfoliated graphene.@xcite therefore a great deal of interest is focused on the understanding the differences between the electronic structure of exfoliated and epitaxial graphene and the consequences for potential applications .
since the crystallographic quality of both graphene layers seems to be equivalent,@xcite the interaction with the substrate remains the biggest unknown .
this is mainly because of the complicated structure and electronic properties of the carbon rich graphene / sic interface layer , which are still not fully understood.@xcite in this paper , we present a local study of electronic and vibrational properties of nanometer - sized areas of a graphene monolayer grown on sic(0001 ) and its @xmath1 interface layer by scanning tunneling microscopy ( stm ) . local scanning tunneling spectroscopy ( sts ) and inelastic electron tunneling spectroscopy ( iets ) measurements have revealed unexpected phenomena in epitaxial graphene that could not be observed in spatially averaged measurements , which are usually reported in literature .
localized states of the interface layer protruding through the first graphene layer have caused giant enhancement of inelastic tunneling of electrons from graphene particularly on the places with localized electron states of the graphene / sic interface layer .
the inelastic phonon contribution for the out of plane graphene acoustic phonon at 70 mv has reached a gigantic 50% of the total tunneling current .
our work reveals an unusual process of inelastic tunneling , which is principally different from previously reported phonon - mediated tunneling in mechanically cleaved graphene placed on sio@xmath2.@xcite
the growth of atomically thin graphene samples was carried out in situ in ultra - high vacuum ( uhv ) on n - type 6h - sic(0001 ) by thermal decomposition of si at elevated temperatures .
the growth process and have been done on a home - built electron - beam heater according to the preparation method described elsewhere.@xcite the sample temperature has been monitored by a pyrometer using emissivity 0.9 .
owing to inhomogeneous heating of the sample by the e - beam heater , a mixture containing very small atomically flat areas ( 10 - 20 nm ) of graphene mono- , bi- and interface layers has been produced as confirmed by low energy electron diffraction ( leed ) and stm .
scanning tunneling microscopy experiments were performed in an omicron gmbh lt - stm setup , working under uhv conditions ( 10@xmath3 mbar ) at 5 k. electrochemically etched w tips were used in the constant current stm mode .
scanning tunneling spectroscopy ( sts ) and inelastic electron tunneling spectroscopy ( iets ) have been obtained by using two lock - in amplifiers and superimposing an alternating voltage reference signal with a frequency 990 hz and amplitude 10 mv to the bias voltage applied to the sample .
figure 1 shows spatially averaged sts curves on 0 interface , 1st and 2nd graphene layers on sic , which are usually presented as local electronic structures of these layers.@xcite even though the sts measurements have been obtained on areas with very small sizes ( 10 - 20 nm ) that were surrounded by many large structural defects , they show comparable results to sts results reported by other groups on better quality graphene samples.@xcite however , averaging of sts curves is not appropriate in disordered systems such as the graphene / sic(0001 ) system is , because it mixes incorrectly the local density of states ( ldos ) at different locations .
we illustrate this in figure [ fig2 ] by a series of atomically resolved stm images of a first graphene layer taken at different bias voltages . at low bias voltages ( @xmath450 mv ) , the characteristic graphene atomic structure together with the larger @xmath1 superstructure
are visible , indicating a single graphene layer on sic(0001).@xcite however , when the bias voltage is increased , bright dots start to appear until they fully dominate the stm pictures at higher voltages . owing to these bias dependent topographic features , an average of sts spectra becomes bias dependent and therefore it does not reflect correctly an average of ldos .
spatially averaged sts curves on the 0 interface , 1st and 2nd graphene layers on sic(0001 ) .
scanning parameters : @xmath5 mv , @xmath6 pa for graphene monolayer and bilayer , and @xmath5 mv , @xmath7 pa for the 0 layer.,width=316 ] the observed disordered bright features do not originate from the graphene layer but from the underlying interface layer as has been previously discussed by rutter _
_ et al.__@xcite graphene thus shows transparency at higher bias voltages to bright features from the lower interface layer in stm .
although the bright features in the zero interface layer seem to be disordered on local scale , they manifest the @xmath1 reconstruction with respect to the sic crystal on larger scales as confirmed by large scale stm images and leed.@xcite interestingly , the positions of bright features are not the same in the filled and the empty states as symbolized by crosses and circles in figure 2c , d . circles and crosses point out the positions of the bright features in the filled states ( -200 mv ) and in the empty states ( 200 mv ) respectively .
local sts measurements on top of these features on a graphene monolayer have revealed clear localized electron states at -200 , -500 and 500 mv ( see figure 2e ) . on the other hand ,
sts spectra measured on regions with a graphene character ( no bright features are observed in stm ) have not shown any peaks in the ldos .
similar localized states as on the first graphene layer have also been measured on bright features in the zero interface layer by sts in figure 2f .
the carbon rich interface layer has semiconducting properties with a 400 mev gap pinned in between the @xmath4200 mv localized states in accordance with previous sts measurements.@xcite the spatial extension of these localized states is in the order of 0.5 nm .
( color online ) stm images of a @xmath8 nm@xmath9 area on single - layer graphene on sic(0001 ) taken with @xmath7 pa and @xmath10 mv ( a ) , 50 mv ( b ) , -200 mv ( c ) and 200 mv ( d ) .
circles point out the positions of the bright dots in the filled states and crosses in the empty states .
( e ) three local characteristic @xmath11 spectra on monolayer graphene on sic(0001 ) ( @xmath12 mv and @xmath13 pa ) .
( f ) three local characteristic @xmath11 spectra on the interface layer on sic(0001 ) ( @xmath5 mv and @xmath7 pa ) .
all sts curves have been averaged over 10 curves.,width=316 ] the origin of the localized states in the interface layer has been suggested to be either due to a different si - c bonding in the interface layer consisting of covalently bonded graphene layer to the sic(0001 ) surface@xcite or in the presence of si adatoms.@xcite both models propose correctly formation of localized states close to the fermi energy . however ,
the first model is supported by angle resolved photoelectron spectroscopy ( arpes ) studies on interface and graphene layers on sic(0001)@xcite and by transferring of monolayer graphene to bilayer graphene after decoupling of the interface layer by h@xmath2 intercalation.@xcite in arpes , the first graphene layer displayed well developed graphene @xmath14-bands extending up to the fermi level , the interface layer exhibited semiconducting properties with absent @xmath14-bands.@xcite two localized states at the binding energies 0.5 ev and 1.6 ev have been identified in the carbon rich interface layer with the @xmath1 reconstruction on the sic(0001).@xcite also the formation of empty electron states close to fermi energy with a localized character has been observed in the graphitization study of sic(0001 ) surface in inverse photoelectron spectroscopy.@xcite the localized states at @xmath4200 mev have not been identified in the photoemission experiments most probably because of their low intensities .
surprisingly , their energy coincides with a kink at 200 mev observed in the @xmath14-band dispersion near the @xmath15-point of monolayer graphene , whose origin has been suggested to be related to either electron - electron or electron - phonon interactions.@xcite sts spectra of graphene monolayers and bilayers display an unexpected gap - like feature at the fermi level ( see figure 1 or ref .
@xcite ) . from a thight - binding fit to photoemission data,@xcite however , one would not expect such a gap - like feature in sts because of the electron doping , which causes a shift of the dirac point ( the minimum in the graphene dos ) to -0.45 ev and -0.32 ev for monolayer and bilayer graphene layers respectively.@xcite also transport experiments suggest a higher electron density on a monolayer graphene on sic@xcite than on exfoliated graphene placed on sio@xmath2 , where the dirac point is in the vicinity of the fermi energy.@xcite recently the appearance of a gap - like feature at the fermi - level on exfoliated graphene supported on a silicon oxide surface has been explained by the inability to tunnel into the @xmath14-states due to a small tunneling probability at the fermi - level.@xcite this has been overcome by the assistance of a phonon at 63 mev coupled with @xmath16-states , which made the tunneling possible at energies higher than the phonon energy.@xcite the experimental findings of zhang _
_ have been supported by theoretical modeling of wehling _ _ et al.__@xcite in figure 3b , we show the observation of phonon contributions in iets on a nanometer - sized monolayer graphene on sic(0001 ) .
the inelastic tunneling features are observed as peaks ( or dips ) in the second derivative of current with respect to the voltage at the threshold where the electron energy associated with the bias voltage matches the oscillator energy .
the @xmath11 and @xmath17 spectra in figure 3 have been spatially averaged over 4096 points .
four inelastic peaks corresponding to out of plane acoustic graphene phonons at 16 and 70 mv can been identified in the @xmath17 spectrum on a graphene monolayer .
similar phonon modes at 16 and 58 mv have been found on graphite in iets before.@xcite phonon - induced inelastic tunneling in single molecules deposited on metal surfaces typically leads to conductivity changes in the order of only @xmath18,@xcite where the normalized change in differential conductance @xmath19 is obtained by normalizing the peak area in @xmath17 to conductance .
the inelastic peak intensities in monolayer graphene on sic are @xmath20 for both phonon contributions at 16 mv and 70 mv ( figure 4 ) .
surprisingly , the tunneling conductivity changed by a much larger factor 13 outside the gap - like feature on the exfoliated graphene.@xcite this has been explained by a different mechanism based on the phonon - mediated tunneling process which involves momentum - conserving virtual transitions between 2d electron bands in graphene .
averaged @xmath11 ( a ) and @xmath17 ( b ) spectra taken on graphene monolayer on sic(0001 ) and its tunnel current decay length @xmath21 as a function of bias voltage ( c ) .
an inelastic peak intensity @xmath19 is indicated at the each attributed phonon peak in @xmath17 .
scanning parameters : @xmath22 grid , @xmath23 mv and @xmath24 pa . the decay lengths have been determined from @xmath25 spectroscopy at fixed bias voltage by fitting it by an exponential function @xmath26 .
error bars represent standard deviations of the measurements.,width=316 ] the mechanism of the phonon - assisted tunneling in exfoliated graphene was supported by observation of bias dependent wavefunction spatial decay rates , where the tunnel decay length inside and outside the gap has been observed to be 0.25 @xmath27 and 0.45 @xmath27 respectively.@xcite bias dependent wavefunction spatial decay rates in monolayer graphene grown on sic are depicted in figure 4c .
the decay length @xmath21 has been determined from @xmath25 spectroscopy performed at fixed bias voltage @xmath28 by fitting it to an exponential function @xmath26 .
similarly like on exfoliated graphene , two different decay rates have been observed inside and outside the gap - like feature bounded in the @xmath4100 mv region , @xmath29 and @xmath30 . although the results measured on epitaxial graphene in figure 3 look similar to the data by zhang _
et al . _ measured on exfoliated graphene,@xcite the mechanism is different .
firstly , both out of plane acoustic phonon contributions at 16 mv and 70 mv have similar intensities but only the latter phonon can assist the virtual tunneling to @xmath16 electrons since it has the right momentum because it is centered at the @xmath31 points , whereas the other out - of - plane acoustic phonon at 16 mv can not play the same role because it is located at the @xmath32 point . secondly , the tunneling decay rates are observed to change exactly at the edge of the gap of the interface layer ( see figure 3 ) , whose states are known to have a large spatial extension since they are seen in stm even upon formation of two graphene layers above the sic interface . finally , the most important fact that disproves the phonon assisted tunneling in epitaxial graphene on sic is a spatially inhomogeneous character of the inelastic contribution .
( color online ) @xmath11 ( a , b ) and @xmath17 ( c , d ) maps at constant bias voltage indicated in the right top corner of a @xmath33 nm@xmath9 area on graphene monolayer on sic(0001 ) .
red regions indicate high intensity inelastic phonon contributions at -70 mv and high @xmath11 at -200 mv and blue regions mark out high intensity @xmath17 at 70 mv and @xmath11 at 200 mv . scanning parameters : @xmath23 mv and @xmath6 pa.,width=316 ] to illustrate the spatial dependence of inelastic tunneling intensity , we show simultaneously measured @xmath11 and @xmath17 maps on a graphene monolayer on sic(0001 ) in figure 4 .
the @xmath17 images depict intensities of the inelastic peak contribution of the phonon mode at @xmath470 mv and the @xmath11 maps portray intensities of the localized states at @xmath4200 mv .
the places of the high inelastic peak intensity coincide with the places where the @xmath4200 mv localized states are observed in the @xmath11 maps . for this reason ,
high intensity regions in figure 4 have been highlighted by red and blue color in negative and positive bias voltages respectively to highlight the correlation between @xmath34200 mv@xmath35 and @xmath3670 mv@xmath35 maps .
the iets peak intensities vary spatially by a large factor in @xmath17 maps , up to 50 among some places , as seen by the difference between the values of red / blue and gray regions .
the regions with high iets intensities are found at different locations in positive and negative bias voltage , similar to the bright features in figure 1 .
this inhomogeneous asymmetry can be also seen on three characteristic local @xmath11 and @xmath17 spectra depicted in figure 5 .
these spectra have been averaged only over 10 local measurements , therefore they exhibit a larger noise level in comparison to the spatially averaged iets spectra .
an iets curve measured on a position with a high iets intensity at -70 mv ( figure 5a ) shows a gigantic inelastic feature reaching @xmath37 in negative bias voltage , while the iets peak in positive voltage is half of this size .
such high iets signals have been observed predominantly at positions with high @xmath11 intensities at -200 mv .
these places most probably correspond to the localized states at -200 mv on the first and zero graphene layers .
moreover , a second harmonic phonon mode at -140 mv is observed in @xmath17 with an intensity approximately 5 times smaller than the intensity of the first harmonic mode .
similar results have been observed on places with a high inelastic peak at + 70 mv that are located at position with a high @xmath11 at 200 mv , implying a connection with localized states of the graphene monolayer in the empty states . in this case , an enormous first order inelastic peak together with the second harmonic contribution has been observed in the positive bias voltage . on the other hand ,
iets spectra obtained on locations free of localized states ( figure 5c ) have demonstrated relatively low intensity phonon contributions ( 10% ) for both 16 and 70 mv out - of plane phonons .
no second order phonon modes could be seen in these iets spectra .
important is to note that one should be careful in relating the high intensity @xmath11 regions at @xmath4200 mv with localized states since an increase in @xmath11 can also be caused by high intensity iets peaks at @xmath470 mv .
however , since the presence of localized states have also been independently proved by other stm groups on a monolayer graphene,@xcite it is highly probable that high intensity @xmath11 correlates with localized states at @xmath4200 mv originating in the graphene / sic(0001 ) interface layer .
@xmath11 and @xmath17 spectra on graphene monolayer on sic taken at @xmath23 mv and @xmath6 pa . the spectra represent typical individual @xmath11 and @xmath17 curves obtained in red reagions ( a ) , blue regions ( b ) and gray regions ( c ) in figure 4 .
inelastic peak intensities @xmath19 are indicated at the each attributed phonon peak in @xmath17.,width=316 ] typical inelastic tunneling experiments give rise to phonon fingerprints that reach only a few percent of the total tunneling current , however , we observe unexpectedly giant signals as high as 50% in the presence of localized states . this could indicate a strong electron - phonon ( e - ph ) coupling strength ( @xmath21 ) , but calculations by park _
_ et al.__@xcite have shown values for @xmath21 in the order of 0.05 at 200 mev binding energy .
this relative small value can not explain the large iets intensity .
the fact that the large iets intensities coincide with the positions of sharp localized electron states at @xmath4200 mv is supported by the dft calculation results of atta - fynn _ _ et al.__@xcite in this study , localized electron states stemming from defects or topological disorder exhibited an anomalously large e - ph coupling.@xcite hence , the observed localized states probably enhance the e - ph coupling , resulting in a larger iets intensity .
however , the presence of localized states might not be the only criterion of giant iets contributions because the iets have been measured on a very small graphene regions ( 10 - 20 nm ) confined among many structural defects .
therefore , there seem to be two conditions for the giant enhancement of the iets data : both the influence of the localized states at @xmath4200 mv and the presence of structural defects .
the structural defects are known to play a very important role in the scattering of electrons , which is an additional contribution for localization , thus causing together with localized states an anomalously large e - ph coupling .
in addition , in the @xmath11 spectra , higher harmonics are observed equidistantly spaced with the value of the vibration .
higher harmonics so called phonon ( vibrational ) side bands have been observed occasionally in scanning tunneling experiments in the resonant tunneling regime.@xcite the conditions for resonant tunneling are discussed in detail by galperin _
et al._:@xcite the higher order vibronic levels become visible if the tunneling electron stays relatively long on the molecule compared to the dephasing time and the localized electron state coincides with a vibration level .
this happens if the chemical interaction between electron state and molecule is relatively small , resulting in a narrow broadening of the vibronic level.@xcite as is shown , at @xmath4200 mv , localized electron states can couple with the out - of - plane phonons from graphene . because the @xmath11 data do not show periodic phonon peaks equidistantly around @xmath4200 mv , the resonant tunneling channel is not related to the localized states at @xmath4200 mv .
resonant tunneling through localized states at the fermi - level is difficult to determine because of the pseudogap , but it would be highly probable if the origin of the pseudogap is of many - body character , characterized by electron - electron and electron - phonon interactions.@xcite
in conclusion , a giant inelastic tunneling process has been observed in epitaxial graphene on sic(0001 ) in scanning tunneling experiments .
the inelastic tunneling channel reached half of the total tunneling current .
the mechanism of the giant tunneling is connected with the presence of sharp localized states originating in the interface with sic and strong electron - phonon coupling in graphene near a structural defect .
k. v. emtsev , a. bostwick , k. horn , j. jobst , g. l. kellogg , l. ley , j. l. mcchesney , t. ohta , s. a. reshanov , e. rotenberg , a. k. schmid , d. waldmann , h. b. weber , and t. seyller , nat . mat . 8 , 203 ( 2009 ) . | local electronic structures of nanometer - sized patches of epitaxial graphene and its interface layer with sic(0001 ) have been studied by atomically resolved scanning tunneling microscopy and spectroscopy .
localized states belonging to the interface layer of a graphene / sic system show to have an essential influence on the electronic structure of graphene .
giant enhancement of inelastic tunneling , reaching 50% of the total tunneling current , has been observed at the localized states on a nanometer - sized graphene monolayer surrounded by defects . |
the advent of quantum mechanics ( qm ) and the problems linked to its ontological nature changed our conception of physical reality in a radical way , problems that before had concerned just philosophers of science became central to the physics debate .
notions like observable , observer and measurement , which had not been problematic to the conception of physical theories , became fundamental and subject of numerous controversies .
furthermore , the classical realist conception of physical objects as independent bearers of properties , on which the ontology of classical theories was based , was also challenged ; becoming evident the interpretational difficulties of the theory , a problem which physics had never affronted before .
+ today , more than one hundred years after max planck formulated the quantum hypothesis , we still do not have a settled agreement about what quantum reality is or if there is something as a quantum reality at all .
nevertheless , the theory has been incredibly successful in its predictive role . for this reason
, many physicists think that probably there is no necessity of an interpretation of the theory that go further than the interpretation linked to its predictive nature .
but , despite this predictive success , it has not been possible to conciliate the theory in its instrumentalist form with general relativity ( gr ) , and it is becoming evident that a new formalism , which give us a completely new perspective of the theory , will be needed to solve this problem . +
numerous proposals have been advanced to solve the interpretational issues of qm , however , none of these have been able to transcend their heuristic argumentations with a solid mathematical machinery that captures and go further than the classical tools , settling the respective interpretation . our inability to do so
is probably a sign that behind the understanding of qm hides the necessity to transcend also the classical mathematical formalism that lies at the foundations of the tools used so far to conceive physical theories .
+ in recent years topos theory has captured the attention of people working on the foundations of qm as a possible route to reformulate the theory in a way flexible enough to include relativistic concepts , and where a definite interpretation could be finally settled .
the origins of this approach can be traced back to the work developed by takeuti in 1975 on boolean valued models of set theory .
takeuti proved that in a boolean valued model of set theory constructed over a complete boolean algebra of projector operators of a hilbert space , there is a correspondence between the self adjoint operators which spectral family is contained in the boolean algebra of projections and the real numbers of the boolean valued model @xcite@xcite . in those days
the importance of this result respect to its possible relation with interpretational issues of qm was discussed ( see @xcite ) but no conclusive results were obtained , and maybe due to the fact that the result used advanced tools of set theory and logic , it did not capture the attention of the physics community .
+ recently the work of c. isham , a. doering @xcite @xcite and others have brought the attention back to these methods and particularly to the idea that these tools can be used to obtain a new conception of the continuum useful to formulate quantum gravity theories and to obtain a new perspective of qm . even if not explicitly stated both , takeuti s and isham - doering s approaches , are a reformulation of the old idea proposed several times after the publication of the seminal paper in quantum logic ( ql ) by von neumann and birkhoff , regarding the necessity of a formalism founded over a quantum logic as a route to reformulate the theory in a way able to capture the essence of quantum reality .
however , it is still not clear that this new approach will give us a better understanding of the theory .
unfortunately , the intrinsic difficulties of cohen s forcing in the boolean formulation and the abstract categorical machinery of topos theory have obscured the potentiality of these methods to obtain a better picture of qm .
+ in 1995 x. caicedo introduced what can be considered so far the most user - friendly approach to kripke - joyal semantics @xcite ( the semantics intrinsic to a topos ) , giving a new perspective that avoids the technicalities linked to the categorical tools of topoi .
caicedo introduced a model theory of variable structures where it is possible to introduce a definition of genericity and a generic model theorem which unifies set theoretic forcing constructions and the classical theorems of model theory as completeness , compactness , omitting types etc .. in this context the approach to set theoretic forcing generalizes the scott - solovay boolean @xcite and the fitting intuitionistic @xcite formulations , simplifying the constructions and clarifying the essence of the proofs of classical independence results as the independence of the continuum hypothesis @xcite @xcite .
another remarkable fact is how interesting connections between classical logic and intuitionistic logic are revealed , showing that the logic of sheaves is not just intuitionistic but that constitutes a continuum of logics between classical and intuitionistic logic , where the independence results of set theory and the classical theorems of model theory can be conceived as a consequence of some limit process over this continuum .
+ in this work i apply these tools to qm , the idea in a few words is to show that the logic intrinsic to qm lies in this continuum of logics and then to show that the emergence of classicality can be conceived as a limit process over these logics . the logic used here will differ from the classical ql of von neumann and birkhoff ; its construction arises from the local character of truth of sheaf logic , which allows to introduce some contextual features of qm as those derived from the kochen - specker theorem and from the deutsch - everett multiversal interpretation of interference phenomena . over this logic a hierarchy of _ quantum variable sets _ is constructed that generalizes and simplifies the boolean approach of takeuti . in this model , two alternative proofs of takeuti s correspondence , between self adjoint operators and the real numbers of the model , are given .
this approach results to be more constructive , showing a direct relation with the gelfand representation theorem , and revealing also the importance of these results with respect to the interpretation of qm in close connection with the deutsch - everett multiversal interpretation of quantum theory .
finally it is shown how the collapse via generic models of this structure of quantum variable sets can help to explain the emergence of classicality also in close relation with the deutsch - everett perspective .
+ i have divided this work in two main sections , the first one is an introduction to sheaf logic as developed in caicedo s work , giving here an approach oriented to physicists . even if it is not possible to get a complete picture of these tools without some knowledge of model theory and set theory the guiding ideas are very natural and can be followed without being an expert in these fields .
the second part deals with the construction of the hierarchy of quantum variable sets and the results cited above .
hopefully , the ideas contained here will be also useful to people working with applications of topos theory in physics , particularly bringing attention to the fundamental role of logic that sometimes is forgotten giving prevalence to the geometric character of the theory .
as i argue in this work , the role of logic in these tools can be fundamental to obtain a new satisfactory picture of qm and probably the connections between geometry and logic , the most remarkable feature of topoi , will be fundamental in the construction of a future quantum gravity theory .
the notion of a sheaf of structures has its origins on the study of the continuation of analytic functions in the xix century , but the modern definition was introduced just few years after the end of the second world war , by leray , cartan and lazard , in the context of algebraic topology and algebraic geometry .
however , the idea of a sheaf of structures is naturally contained in the conception of spacetime which derives from the galilean relativity principle .
the galilean relativity principle asserts that the dynamical laws are the same when are referred to any frame in uniform motion .
this principle forces us to abandon the aristotelian picture of a fixed and absolute background space which constitutes a preferential frame where physical objects move .
thus , the idea of considering a point in space as the same point an instant later looses its meaning .
instead , galilean dynamics implies that there is not one fixed euclidean 3-dimensional space where the physical world is contained , but that to each instant corresponds a different 3-dimensional world @xcite , which is attached with the other worlds in a continuous way respect to the temporal order(see fig .
[ gspace ] ) .
each of these worlds is formed by objects which represent a snapshot of an extended object in time .
there also exist functions and relations defined over each world , which represent and describe the instantaneous physical attributes of these extended objects .
the continuous attachment refers to the possibility of seeing these extended objects as functions that at each time assign an object in the respective world in such a way that they form a continuous object as seeing from the perspective of the hole attachment . in the same way the functions and relations that give instantaneous descriptions attach also in a continuous way when seen as functions and relations of the attachment of the different instantaneous worlds .
+ we can summarize this description of galilean spacetime as follows .
we have a topological space @xmath0__temporal line _ _ , for each element @xmath1 there is an structure @xmath2 which is formed by a world @xmath3 which objects constitute a snapshot of extended objects in time , functions @xmath4 and relations @xmath5 which give the instantaneous properties of the extended objects at the instant @xmath6 .
even if at each instant @xmath6 we have different worlds @xmath3 , this worlds are of a same kind in the sense that the functions , relations and objects can be given analogous interpretations at each instant .
the different worlds @xmath3 attach in an extended universe @xmath7 in such a way that the attachment of the objects , functions and relations are continuous as seeing as extended objects , functions or relations defined in this extended world . in a few words at each instant @xmath6 we have an structure @xmath8 formed by a world of instantaneous snapshots of extended objects , functions and relations that attach in a continuous way .
all this features can be easily formalized and generalized using the notion of a sheaf of structures .
the ideas presented in this section follow @xcite , the proofs of the results presented here can be found there , however i will try to motivate the results and the main ideas can be understood without knowing all the technical details .
let @xmath9 be a topological space .
a sheaf over @xmath9 is a couple @xmath10 , where @xmath7 is a topological space and @xmath11 is a local homeomorphism , or in other words a continuous function such that for each @xmath12 there exists a neighbourhood @xmath13 of @xmath14 such that : 1 .
@xmath15 is open in @xmath9 2 .
@xmath16 is an homeomorphism .
given an open set @xmath17 in @xmath9 , a function @xmath18 such that @xmath19 is called a local section , if @xmath20 , @xmath21 is called a global section .
the set @xmath22 for @xmath1 is called the fibre over @xmath6 . using the sheaf
notion it becomes easy to define a generalization which captures the essence contained in the picture of galilean spacetime as described above . the formalism of the definition given below contains some technicalities of model theory , but the essence of the definition can be understood even if you are not used to the language of this field . given a fix type of structures @xmath23 a sheaf of @xmath24-structures @xmath25 over a topological space @xmath9 is given by : + a- ) a sheaf @xmath26 over @xmath9 .
+ b- ) for each @xmath1 , a @xmath24-structure @xmath27 , where @xmath28 ( the fiber that could be empty ) is the universe of the @xmath24-structure @xmath2 , and the following conditions are satisfied : + i. @xmath29 is open in @xmath30 seeing as subspace of @xmath31 , where @xmath32 is an n - ary relation symbol .
@xmath33 is a continuous function , where @xmath34 is an @xmath35-parameter function symbol .
@xmath36 such that @xmath37 , where @xmath38 is a constant symbol , is continuous .
the type of a structure mentioned in the definition above is a concept used in logic and model theory . in a few words
a type @xmath24 is a language with symbols of relations , functions and constants .
a @xmath24-structure @xmath39 is formed by a set of objects @xmath40 where the different symbols of relations , functions and constants find an interpretation as functions or relations over this set and the constants as elements of this set .
for example if @xmath41 , the structure which has as universe the rational numbers @xmath42 , where @xmath43 is interpreted as the multiplication of rational numbers and @xmath44 is interpreted as the number @xmath45 is a @xmath24-structure .
this can be rather confusing but in some sense is just telling us that we are attaching structures of the same kind . on the other hand the properties ( i ) , ( ii),(iii ) determine that the structures attach in a smooth or continuous way .
+ a sheaf of structures is a space extended over the base space @xmath9 of the sheaf as galilean spacetime extends over time .
the elements of this space will not be the points of @xmath7 but the sections of the sheaf conceived as extended objects .
the single values of these sections represent just punctual descriptions of the extended object .
+ let @xmath46 be an open subset of the base space , the set of sections defined over @xmath17 , @xmath47 can be seen also as the universe of an structure of the same type @xmath24 of the sheaf of structures .
this follows from the fact that the continuous attachment guarantees that if , for instance , we have sections @xmath48 defined over an open set @xmath13 and some relation @xmath49 holds at the node @xmath50 , taking @xmath51 which is an open set by the definition above , we have that for @xmath52 , @xmath53 holds
. then we can say that @xmath54 holds in @xmath17 if @xmath53 for all @xmath55 . in an analogous way
if @xmath56 for some section @xmath57 defined over @xmath13 and a @xmath58-parameter function symbol @xmath34 , by the definition above we have that @xmath59 is continuous , then there exists an open neighbourhood @xmath17 of @xmath6 such that @xmath60 , thus @xmath61 for all @xmath55 . using this
we can define a function @xmath62 such that @xmath63 if @xmath61 for all @xmath55 .
considering these kind of structures we have that for any open set @xmath64 the restriction of the sections define a natural homomorphism @xmath65 ( i.e a function which conserve the relations and commute with functions ) between @xmath66 and @xmath67 : @xmath68 sometimes it will be easier to define the structures @xmath66 this is something common in model theory to avoid some excess of formalism .
] than the complete sheaf of structures , as it will be the case of the hierarchy of quantum variable sets .
however given the structures @xmath66 for the open sets of @xmath9 and the morphisms @xmath69 , it is possible to reconstruct the sheaf of structures .
to show how to do this , we need two important definitions .
a presheaf of structures of type @xmath24 over @xmath9 is an assignation @xmath70 , such that to each open set @xmath71 is assigned a @xmath24-structure @xmath72 , and if @xmath64 it is assigned an homomorphism @xmath73 which satisfies @xmath74 and @xmath75 if @xmath76 .
it is clear that it is possible to define a presheaf of structures @xmath77 from a sheaf of structures assigning to each open set @xmath17 a @xmath24-structure which universes are the sets @xmath66 , and the homorphism are the @xmath69 defined above . given a presheaf of structures we can construct a sheaf of structures in the next way : let @xmath70 a presheaf of structures over @xmath9 .
let @xmath78 be the sheaf of structures over @xmath9 such that each fiber @xmath79 is defined by : @xmath80 where @xmath81 is the set of neighbourhoods of @xmath6 , and given @xmath82 and @xmath83 , @xmath84 let @xmath85_x$ ] the equivalence class of @xmath21 respect to @xmath86 .
we have that : @xmath87_x, ... ,[\sigma_n]_x)\in r^{x}\leftrightarrow \exists u\in\mathcal{v}(x)\text { such that } ( \sigma_1, ... ,\sigma_n)\in r^{\gamma(u)}\ ] ] @xmath88_x, ... ,[\sigma_n]_x)=[f^{\gamma(u)}(\sigma_1, ... ,\sigma_n)]_x.\ ] ] the space of the fibers @xmath89 is given the topology generated by the images of the sections @xmath90_x.\end{aligned}\ ] ] the sheaf @xmath78 is called the sheaf of germs of @xmath70 .
this definition can be understood using the analogy of galilean spacetime in the next way .
if to each time interval we assign objects such that their history or part of it develops in such interval , we can recover the instantaneous perspective identifying two objects at each instant if their histories coincide in an interval of time containing that instant .
in the same way a relation will hold from the instantaneous point of view if it holds in a time interval containing the respective instant .
this kind of contextuality will play a fundamental role in the logic that governs these models as we will see below .
+ the sheaf of germs @xmath91 associated to the presheaf of sections @xmath77 of a sheaf @xmath39 , is naturally isomorphic to the original sheaf .
indeed the function @xmath92_x\mapsto \sigma(x),\end{aligned}\ ] ] defines a natural isomorphism which sends in an isomorphic way each fiber @xmath93 to the fiber @xmath3 . on the other hand given a presheaf @xmath70 the presheaf @xmath94 associated to the sheaf of germs @xmath78 , results also isomorphic to the original presheaf just if the presheaf satisfies a further condition .
a presheaf of structures is said to be exact , if given @xmath95 and @xmath96 , such that if @xmath97 there exists an unique @xmath98 such that @xmath99 for all i. and the same holds for the relations i.e if we have some relations @xmath100 which are sent by the homomorphisms @xmath101 , @xmath102 to a same relation for all @xmath103 .
there exists an unique relation @xmath104 which is sent by the homomorphism @xmath105 to the relation @xmath106 for all @xmath107 .
we have then the next result .
if @xmath70 is an exact presheaf then it results isomorphic to the presheaf @xmath94 associated to the sheaf of germs @xmath78 , in the sense that @xmath108 as structures and the homomorphisms @xmath73 transform in the homomorphism @xmath109 .
we will define the quantum hierarchy of variable sets defining an exact presheaf , the above results allow to deal indistinctly with the presheaf of structures and the associated sheaf . the notion of truth " in classical physics is a contextual one . when we affirm that a property holds for a certain object at some instant , we are referring to a measurement realized in an extended interval of time containing that instant .
the mathematical models we use to describe such situations , even if based on an absolute notion of truth , permit to capture this contextual character using the notion of limit , which allows to define instantaneous properties in a coherent way . in some sense the notion of limit is what makes classical logic work in the conception of the continuum .
it is surprising then , that mathematical models which are based on an absolute notion of truth can capture and describe physical reality , where the notion of truth is contextual , in such an effective way .
nevertheless , it is probably this lack of contextuality what makes the classical formalism inappropriate to give a complete picture of quantum theory . in qm
the notion of contextuality recovers total new meanings related to the incompatibility of observables , and interference phenomena as we will see .
therefore , it can be fundamental to have a formalism that allows to include a notion of truth which can contain these contextual features .
this is the case of the sheaves of structures , which have a logic based on the next _ contextual - truth _ paradigm : _ if a property for an extended object holds in some point of its domain then it has to hold in a neighbourhood of that point . _ as the objects of a sheaf of structures are the sections of the sheaf , the logic which governs them should define when a property for an extended object holds in a point of its domain of definition .
the next definition contains a lot of logical language but it is very natural .
[ logic ] let @xmath110 be a first order language of type @xmath24 . given a sheaf of structures @xmath25 of type @xmath24 over @xmath9 , and a proposition @xmath111 we can define by induction @xmath112 ( which means @xmath25 forces @xmath113 in @xmath1 for the sections @xmath48 of @xmath25 defined in @xmath6 or in a more elementary fashion , the property @xmath114 holds at the node @xmath6 for the sections @xmath115 in the sheaf @xmath39 ) : + 1 . if @xmath114 is an atomic formula and @xmath116 are @xmath24-terms : @xmath117\leftrightarrow t_1^{\frak{a}_x}[\sigma_1(x), ...
,\sigma_n(x)]=t_2^{\frak{a}_x}[\sigma_1(x), ... ,\sigma_n(x)]\ ] ] @xmath118\leftrightarrow ( t_1^{\frak{a}_x}[\sigma_1(x), ... ,\sigma_n(x)], ... ,t_n^{\frak{a}_x}[\sigma_1(x), ... ,\sigma_n(x)])\in r^{x}\ ] ] 1. the equality between extended objects , or objects defined in function of extended objects holds at some node if and only if the equality holds for their punctual descriptions .
analogous for the relations .
@xmath119\leftrightarrow \frak{a}\vdash_x\varphi[\sigma_1, ... ,\sigma_n]$ ] and @xmath120 $ ] + 2. the conjunction of two properties holds for some extended objects at some node if and only if each property holds for those extended objects at the same node .
@xmath121\leftrightarrow \frak{a}\vdash_x\varphi[\sigma_1, ... ,\sigma_n]$ ] or @xmath122 $ ] .
+ 3. the disjunction of two properties holds for some extended objects at some node if and only if one of the properties hold for the extended objects in that node .
+ 4.@xmath123\leftrightarrow$ ] exists @xmath17 neighbourhood of @xmath6 such that for all @xmath55 , @xmath124 $ ] .
+ 4.the negation of a property for some extended objects hold at some node , if and only if there exists a neighbourhood of the node such that at each point in that neighbourhood the property does not hold for the extended objects .
@xmath125\leftrightarrow$ ] exists @xmath17 neighbourhood of @xmath6 such that for all @xmath55 if @xmath126 $ ] then @xmath127 $ ] .
@xmath128 exists @xmath21 defined in @xmath6 such that @xmath129 + 7 .
@xmath130 exists @xmath17 neighbourhood of @xmath6 such that for all @xmath55 and all @xmath21 define in @xmath131 , @xmath132 $ ] .
( in 4,5,7 @xmath17 has to satisfy @xmath133 ) numerals 1,2,3 , 4 clarify the content of the definition without the technicalities of the logical language ; 5,6 , 7 are more clearly understood using formal language .
it is clear from 4,5 , 7 how this is a contextual logic , however this contextuality is better expressed in the next result which tell us that a property is verified at some node if and only if it is verified in a neighbourhood of that node .
[ localtruth ] @xmath134 $ ] if and only if there exists a neighbourhood @xmath17 of @xmath6 , such that @xmath135 $ ] for all @xmath55 . using the above definition
it is possible to introduce a local semantics in a natural way , which will be more useful when dealing with presheaves of structures . given an open subset @xmath46 , and sections defined over @xmath17 , we say that a proposition about these sections holds in @xmath17 if it holds at each point in @xmath17 or in other words : @xmath136\leftrightarrow \forall x\in u , \mathfrak{a}\vdash_x \varphi[\sigma_1, ... ,\sigma_n]\ ] ] this definition is determined completely by the next result .
[ kripkejoyal ] @xmath137 $ ] is defined by : + 1 . if @xmath114 is an atomic formula : + @xmath138 .
+ @xmath139\leftrightarrow ( \sigma_1, ... ,\sigma_n)(u)\subseteq r^{\frak{a}}$ ] .
@xmath140\leftrightarrow \frak{a}\vdash_u\varphi[\sigma_1, ... ,\sigma_n]$ ] and @xmath141 $ ] .
@xmath142\leftrightarrow$ ] there exist open sets @xmath143 such that @xmath144 , @xmath145 $ ] and @xmath146 $ ] .
@xmath147\leftrightarrow$ ] for any open set @xmath148,@xmath149,@xmath150 $ ] .
@xmath151\leftrightarrow$ ] for any open set @xmath152,if @xmath153 $ ] then @xmath146 $ ] . + 6 .
@xmath154 there exists @xmath155 an open cover of @xmath17 and @xmath156 sections defined on @xmath157 such that @xmath158 $ ] for all @xmath107 .
@xmath159 for any open set @xmath152 and @xmath57 defined on @xmath160 , @xmath161 .
the logic just defined can be seen as a multivalued logic with truth values that variate over the heyting algebra of the open sets of the base space @xmath9 .
let @xmath162 sections of a sheaf @xmath39 defined over an open set @xmath17 , we define the _ truth value " _ of a proposition @xmath114 in @xmath17 as : @xmath163_{u}:=\{x\in u : \mathfrak{a}\vdash _ x \varphi [ \sigma_1, ... ,\sigma_n]\}\ ] ] from corollary [ localtruth ] we know that @xmath164_{u}$ ] is an open set , thus we can define a valuation as a topological valuation on formulas : @xmath165_{u}.\ ] ] the definition of the logic allows to define the value of the logic operators in terms of the operations of the algebra of open sets . for instance , the proposition @xmath166 is valid in a point if there exists a neighbourhood of that point where @xmath114 does not hold at each point of the neighbourhood , then @xmath166 holds in the interior of the complement of the set where @xmath114 holds i.e @xmath167_{u}= int ( u\setminus [ [ \varphi]]_{u}),$ ] .
reasoning in analogous way we have : 1 .
@xmath168_{u}=\{x\in u:\sigma_1(x)=\sigma_2(x)\}$ ] .
2 . @xmath169]_{u}=\{x\in
u : ( \sigma_1(x), ... ,\sigma_n(x))\in r^{\mathfrak{a}}\}$ ] 3 .
@xmath167_{u}= int ( ( u\setminus [ [ \varphi]]_{u}),$ ] 4 .
@xmath170_{u}=[[\varphi]]_{u}\cap[[\psi]]_{u},$ ] 5 . @xmath171_{u}=[[\varphi]]_{u}\cup
[ [ \psi]]_{u},$ ] 6 . @xmath172_{u}= int((u\setminus[[\varphi]]_{u})\cup [ [ \psi]]_{u}),$ ] 7 .
@xmath173_{u}=\bigcup_{\sigma\in \mathfrak{a}(w ) , w\subset u}[[\varphi(\sigma)]]_{w},$ ] 8 .
@xmath174_{u}=int(\bigcap_{\sigma\in \mathfrak{a}(w ) , w\subset u}[[\varphi(\sigma)]]_{w}).$ ] it follows then @xmath175\leftrightarrow [ [ \varphi[\sigma_1, ... ,\sigma_n]]]_{u}=u.\ ] ] therefore , since the open sets form a complete heyting algebra , and we have defined the logic operators in terms of the algebra operations , we have that the formulas that get the value 1 in a complete heyting algebras are forced in every node of the sheaf of structures .
this proves that the laws of the intuitionistic logic , which are those that assume the value 1 in a heyting algebra , are forced in each node of a sheaf of structures .
+ in this way each sheaf of structures is ruled by a logic intermediate between intuitionistic logic and classical logic .
intermediate because the geometry of the spaces which determine the sheaf will determine how close or far of each logic is the intrinsic logic of the sheaf ( see next section ) .
this connection between logic and geometry is a very interesting fact that can be important in a future theory of quantum gravity . in recent years
the use of topoi has been suggested to construct a kind of quantum geometry to unify gr and qm ( see for example @xcite ) , based on the conjectured inadequacy of the classical concept of manifold to construct a theory of quantum gravity .
the conjecture tacitly contained there is the necessity of an intuitionistic version of spacetime .
if an intuitionistic model can capture the essence of quantum mechanics , as i am proposing in this work , it will be natural to think that to unify gr and qm , we will need an intuitionistic version of spacetime .
i want to show how probably within general relativity , using the notion of sheaf of structures , we can also find strong motivations to think that this is the right route to follow .
+ the connections between logic and geometry included within the formalism i am presenting here ( and also in the topos formalism ) , were one of the main motivations to think that probably these methods can be useful in a future theory of quantum gravity .
as we will see below the logic structure of quantum mechanics in this context derives in its interpretation . on the other hand ,
these tools allow to obtain a new description of the continuum which arise from the logic structure of qm .
the continuum is what we use in geometry to measure , the notion of lorentzian manifold is an expression of the continuum to capture the essence of spacetime . if we can construct a definition of manifold which arise from the logic of qm , we will probably get automatically a quantum gravity theory . in other words , in some sense quantum mechanics represents logic , and general relativity is in essence geometry ; if we want to connect these two , the sheaf of structures formalism ( or the topoi formalism ) seems to be an appropriate route to follow .
+ to understand better what i mean by connections between logic and geometry lets consider a simple example .
consider @xmath176 and @xmath177 , where @xmath9 has the usual topology and @xmath7 has the usual topology of @xmath178 plus the open neighbourhoods of @xmath179 which are of the form @xmath180 where @xmath17 is a neighbourhood of @xmath181 . from @xmath7 and @xmath9 we can construct a natural sheaf with @xmath11 such that @xmath182
if @xmath183 and @xmath184 ( see fig [ hausdorff ] ) . consider the sections @xmath185 , @xmath186 such that @xmath187 if @xmath188 and @xmath189 and @xmath190 for all @xmath6 . even if @xmath191 and @xmath192 are two different sections , from the perspective of @xmath181 the sections are either different or equal .
in other words on the node @xmath181 we have : @xmath193 indeed it is clear that @xmath194 ( i.e @xmath195 does not hold ) because @xmath196 ( see 1 in the definition [ logic ] ) . on the other hand @xmath197 because each neighbourhood of @xmath181 contains points @xmath6 such that @xmath198
. however , note that if @xmath199 , @xmath200 .
the non validity of the excluded middle in @xmath181 depends on the fact that the hausdorff property is not valid on the fiber over @xmath181 .
it can be proved for an arbitrary sheaf the next result : @xmath201 thus , this example shows how the logic of the sheaf of structures is related with the geometric properties of the sheaf .
+ in general relativity , singularities in a spacetime @xmath202 can be added as boundary points in an extended manifold @xmath203 in such a way that the differential structure of the extended manifold captures the structure of the original manifold in some way . in some of these boundary constructions as those developed by geroch and schmidt ( see @xcite p.217 .
p.276 respectively ) some points , where the hausdorff property does not hold , appear as a result of the construction of the boundary .
furthermore this situation , in the schmidt construction , occurs in the schwarzschild and friedman solutions @xcite , which are our paradigms of strong curvature singularities .
these non - hausdorff features have been seen as defects of these boundary constructions , but within the context of the example exposed above , probably what these results are telling us is that the singularities are a result of the breakdown of the logic which governed these models .
the breakdown of the hausdorff property is probably the breakdown of the excluded middle in the singularity point in some appropriate sheaf description of spacetime .
the difficulty to construct a satisfactory boundary definition such that the extended manifold conserve all the desirable features of a classical spacetime , is probably due to the fact that this formalism is inadequate to capture these situations .
the breakdown of the hausdorff property is a sign of the breakdown of any possibility of classical description , non- hausdorff spacetimes affect the formulation of the cauchy problem , in non - hausdorff spacetime we can find bifurcating geodesics and classical determinism can loose all its sense .
a new conception of manifold is needed , but within this new conception , local determinism will probably have a new meaning based on the ontology of qm , this is why a theory of quantum gravity can not be constructed before settling what quantum reality is .
now we want to construct a mathematical universe which is founded over the logic intrinsic to qm .
as we saw above the logic of a sheaf of structures with base space @xmath9 can be seen as a multivalued logic which take its values over the algebra of open subsets of @xmath9 . therefore ,
if we construct sheaves over a topological space which algebra of open sets capture the logic of qm , we will find automatically models ruled by this logic .
+ the mathematical models that have been used so far in physics are founded over classical logic and classical set theory .
if we want an alternative mathematical universe founded over the logic of qm , it will be important to find one as a natural generalization of the classical universe .
thus it will be important for this model also to be a kind of model of set theory where all the classical constructions find a natural counterpart . +
the classical mathematical universe is the von neumann hierarchy , which is constructed inductively in the next way .
let @xmath204 @xmath205 @xmath206 where @xmath207 is the power set of @xmath208 ( i.e the set of all the subsets of @xmath208 ) .
the set @xmath209 where @xmath210 is the class of all ordinals , is the von neumann hierarchy of classical sets .
probably the notion of ordinal can be new for someone working in physics , but it is not fundamental to understand the model , just think ordinals as a generalization of natural numbers which cover all possible infinities . @xmath211 with the classical belonging relation , @xmath212 , between sets , is the universe of classical mathematics , a model ( in the sense of classical logic ) of the axioms of zfc ( zermelo - fraenkel+choice ) .
we will construct an analogous model which extend the notion of set , as a variable object over a sheaf , where the @xmath212 relation recover a new contextual meaning . to introduce the definition of the cumulative hierarchy we can use the interpretation that categorists use to introduce the objects of @xmath213 , the advantage here
is that the motivation translates literally in the definition ; we obtain a truly structure of variable of sets , where using the classical belonging relation between sets it is possible to define an analogous belonging relation between variable sets and not as arrows between objects .
the definition i present below was originally introduced in @xcite and i used it in @xcite,@xcite to get a new proof of the independence of the continuum hypothesis . in those articles
you can find a more detailed description , here i present a brief introduction enough to follow the main ideas contained below .
+ using the comprehension axiom of set theory , given a proposition @xmath214 , for any set @xmath40 we can construct a set @xmath215 such that @xmath216 if and only if @xmath217 and @xmath214 is truth `` for @xmath6 or in other words , @xmath218 as we have seen in the sheaves of structures the notion of truth is not absolute but it is based on a contextual truth paradigm . to see how this notion of truth translates
when we talk about sets lets consider a particular example .
let @xmath214 be the next proposition _ x is an even number greater or equal than 4 and x can be written as the sum of two prime numbers " _ . in this moment at my ubication on spacetime @xmath219__between the 21:00 and the 22:00 , of the 5 march 2011 , in some place in london " _
_ i can not assure that @xmath220 is even @xmath221 , i.e. that the goldbach conjecture is true .
however @xmath114 can be used to define a set at the node @xmath222 , @xmath223 the goldbach conjecture has been verified for a huge number of even numbers , and this set of numbers will keep growing each time we will find a new prime number .
probably ( maybe not ) some day the goldbach conjecture will be proved or disproved , but the important feature that follow from this example is that instead of conceiving sets as absolute entities , we can conceive them as variable structures which variate over our _ library of the states of knowledge_. it is natural then to conceive the set of nodes where our states of knowledge variates as nodes in a partial order or points in a topological space , that can represent , for instance , the causal structure of spacetime .
our ' ' states of knowledge " will be then structures that represent the sets as we see them in our nodes .
therefore , from each node we will see arise a cumulative hierarchy of variable sets , which structure will be conditioned by the perception of the variable structures in the other nodes that relate to it . or more precisely . [ hierarchy ]
let @xmath9 be an arbitrary topological space , the cumulative hierarchy of variable sets over @xmath9 is defined in the next way .
given @xmath224 denote the non empty open sets of @xmath9 ] we define inductively : @xmath226 the valuation @xmath13 over the open sets constitute an exact presheaf of structures , which sheaf of germs @xmath227 constitute the cumulative hierarchy of variable sets .
for each @xmath224 the set @xmath228 is a set of functions defined over @xmath229 which values for @xmath230 are functions over @xmath231 which values for @xmath232 are functions over @xmath233 and so on .
+ the next step is to define the @xmath212 relation . in a natural way
we define : which means that respect to the context @xmath17 , @xmath34 belongs to @xmath235 if and only if @xmath236 as classical sets relation on boolean valued models ] . at first sight
this construction seems to be odd but in the results presented below we will see that the constructions are extremely natural and extremely powerful .
the first result is that , as we expect , this model is a natural generalization of the universe of classical sets , a model where the axioms of set theory are satisfied respect to the logic of sheaves of structures @xcite@xcite@xcite . for any topological space @xmath9 , @xmath237 in other words for any @xmath1 ( or @xmath224 ) , at the node @xmath6 ( at the context @xmath17 ) the axioms of set theory are satisfied i.e @xmath238 ( @xmath239 ) . to see what this means lets see how the axiom of comprehension cited above is valid in this more general context .
+ * axiom of comprehension : * _ let @xmath240 be a formula . for every set @xmath241
there exists a set @xmath131 such that , @xmath242 if and only if @xmath243 and @xmath214 holds for @xmath6 _ following the logic of sheaf of structures , contained in theorem [ kripkejoyal ] , to prove the above result means to prove that for all @xmath245 and @xmath246 there exists an open cover @xmath247 of @xmath160 and @xmath248 such that for any @xmath249 and @xmath250 for @xmath251 , @xmath252 if and only if however we will omit these technicalities otherwise the notation will become very confusing . ]
@xmath253 and @xmath254 .
it may seem that proving a result in this new logic is very complicated but as we will see , this and the next results have elementary proofs .
let @xmath255 and @xmath256 @xmath257 note that we construct @xmath258 using the axiom comprehension in the classical universe .
it is clear that the variable set @xmath131 just defined satisfies the result described above taking the appropriate restrictions on the respective sets there cited .
note for example that as an open cover of @xmath245 we can take the cover formed by the set @xmath160 , in other results below we will need the definition of the existential using a non trivial open cover but here it is not the case .
however we need to prove that the @xmath131 just defined is an element of the structure of variable sets i.e we need to verify properties 1 - 3 in definition [ hierarchy ] .
let @xmath259 the minimum ordinal such that @xmath260 .
since @xmath261 then @xmath262 , therefore @xmath263 given @xmath264 and @xmath265 we have @xmath266 and @xmath267 . thus @xmath268 and @xmath269 , then @xmath270 .
property 3 in the definition follows in analogous way
. therefore @xmath271 .
even if generally @xmath228 and the classical von neumann hierarchy @xmath211 are not isomorphic for any open set @xmath17 there is an standard way to embed @xmath211 in @xmath228 .
given an arbitrary set @xmath272 let using this embedding it can be easily proved that the set forced as the empty set on each open set @xmath17 is precisely @xmath278 or in other words it holds @xmath279 in the same way it can be proved that the set forced as the set of natural numbers ( i.e. the minimum inductive set that contains the empty set ) is precisely @xmath280 @xcite . the sets forced as the integers and the rational numbers are also @xmath281 and @xmath282 respectively .
however the set forced as the real numbers generally not coincide with @xmath283 , for example as we will see below , this is the case of the set that is forced as the real numbers in the case related to a base space which captures the essence of quantum logic .
+ now we have a way to construct new mathematical universes over arbitrary topological spaces .
we need then a topological space able to capture the essence of quantum logic to obtain a mathematical quantum universe able to capture quantum reality .
as we saw above one important property of the logic of sheaves of structures is the possibility to define a contextual notion of truth .
this contextual truth paradigm is not just intrinsic to the notion of truth in classical physical reality but within the quantum realm it can recover new fundamental meanings .
+ one first aspect of contextuality arises from the kochen - specker theorem .
in a few words what the kochen - specker theorem says is that in qm does not exist a valuation function from the set of physical observables , intended as self adjoint operators on a hilbert space @xmath284 , if the dimension of @xmath284 is greater than two .
a valuation is a function @xmath285 from the set of self - adjoint operators @xmath286 on @xmath284 to the real numbers , @xmath287 , which satisfies : @xmath288 @xmath289 with @xmath290 a borel function . + on the other hand , locally or in a contextual sense , there exist valuations . if we consider an abelian von neumann subalgebra @xmath291 of the algebra of operators , the elements of the gelfand spectrum of @xmath291 ( i.e the positive linear functions @xmath292 of norm 1 such that @xmath293 for all @xmath294 ) when restricted to the self - adjoint operators in @xmath291 constitute a set of valuations which satisfy 1 and 2 above ( see @xcite ) .
what the kochen - specker theorem is telling us is that none of this local valuations can be extended globally . +
a history is characterized by the values that physical variables take on it ; therefore , we can see these valuations as histories , the kochen - specker theorem is telling us that the space of histories has a non trivial contextual structure .
thus , a first notion of context is given by the histories associated to an abelian von neumann subalgebra of the algebra of operators of a hilbert space .
as the elements of the gelfand spectrum of an algebra can be extended in different ways in two non compatible extensions of the original algebra , this contextual character is telling us that what is perceived as true by a history is conditioned by the context of similar histories , where similar is intended in this case as histories that can be expressed in function of the same observables .
+ a second notion of contextuality arises from the phenomenon of interference in quantum mechanics .
interference as intended in the deutsch - everett multiversal interpretation of quantum mechanics is the way as two histories can affect each other .
interference phenomenons are strong enough to be detected only between universes or histories that are very similar , thus what is perceived by a history depends on the context of histories close to it , where close is intended in the sense of being similar . to describe this second contextual feature we use projections operators on a hilbert space @xmath284 .
let @xmath291 be an abelian von neumann algebra of operators , @xmath295 the gelfand spectrum of @xmath291 and @xmath296 the set of projections of @xmath291 .
given @xmath297 and @xmath298 we have @xmath299 then @xmath300 .
consider a self - adjoint operator @xmath301 such that @xmath302 is the spectral representation of @xmath40 in @xmath291 .
let @xmath303 such that @xmath304 , then @xmath305 .
thus if @xmath40 represents a physical observable , we have that in all the histories @xmath285 such that @xmath305 the physical observable @xmath40 assumes the value @xmath306 .
therefore , given a proposition @xmath297 the set @xmath307 is a context of histories which are similar in the sense that some physical observables assume the same values or the values satisfy the same inequalities in each history .
+ the first notion of contextuality has played an important role in the development of the consistent histories program ( see @xcite ) and also it is the base of the approach based on topos theory developed by isham and others ( see @xcite and @xcite ) . the importance of the second notion has not been considered deeply but , as we will see below , this notion is fundamental and probably will explain how to relate classical and quantum realities .
i conjecture that these two notions of contextuality capture the essence of the logic of classical quantum mechanics . therefore , to develop a formalism able to capture the essence of quantum reality using the tools developed above ,
we need a base space @xmath9 which structure captures these two notions of contextuality . unfortunately to capture both notions
we will probably need something more general than a topological space and the respective generalization of the tools developed above , and it is here that tools of topos theory will be probably needed , this will be developed in a future work .
however , we can limit ourselves to a context @xmath308 where @xmath291 is an abelian von neumann algebra , and on @xmath9 we can consider the topology which open sets are the sets @xmath309 associated to a projector operator @xmath297 as defined above .
this will be enough to capture the importance of the second notion of contextuality .
the topology @xmath310 is a topology where all the open sets are clopen and where the algebra of open sets is a boolean algebra isomorphic to the boolean algebra of projector operators @xmath296 .
the cumulative hierarchy of variable sets constructed over the topological space @xmath311 where @xmath291 is an abelian von neumann algebra is what we will call _ the cumulative hierarchy of quantum variable sets .
_ the objects of this model will be the sections of the sheaf @xmath227 , which result to be extended objects that variate over the space of histories or universes @xmath312 , in a few words multiversal objects .
this characteristic will probably be ideal to describe quantum particles , to put it in d. deutsch words : _ thanks to the strong internal interference that it is continuously undergoing , a typical electron is an irreducibly multiversal object , and not a collection of parallel - universe or parallel - histories objects .
_ we will see that in this new mathematical universe the tools we use to understand quantum mechanics recover new fundamental meanings that can be fundamental to settle a definite interpretation of qm .
indeed , once we construct the continuum in this mathematical universe we find a fundamental result which shows the interpretative power of this model respect to quantum theory . as in the previous section consider @xmath291 an abelian von neumann subalgebra of the algebra of operators of a hilbert space @xmath284 , and @xmath308 the gelfand spectrum of @xmath291 with the topology given by the sets @xmath310 as defined above .
we want to construct the continuum , or in other words , the set which is forced as the real numbers on the hierarchy of variable sets over @xmath9 . to construct this set we will use the definition of the real numbers given by dedekind cuts , therefore at each open ( clopen ) set @xmath309 the object @xmath313 will be the set that is forced as the set of dedekind cuts i.e
since the set @xmath322 is the embedding of @xmath42 in @xmath323 , and the construction of the order relation does not introduce anything new , the order relation on @xmath322 is equivalent to the order relation in @xmath42 , i.e. @xmath324 using [ kripkejoyal ]
the above conditions mean : 1 .
\i ) there exist @xmath325 such that @xmath326 ( analogously for @xmath17 ) or + ii ) there exist an open cover @xmath327 of @xmath309 such that there exists @xmath328 and @xmath329 .
( analogously for @xmath17 ) 2 . for all the open sets @xmath330 if @xmath331 is open , @xmath332 , @xmath333 and @xmath334 then @xmath335 .
( analogously for @xmath17 ) .
3 . for all the open sets @xmath330 if @xmath336 then there exists an open cover @xmath337 of @xmath338 and @xmath339 such that @xmath340 and @xmath341 .
( analogously for @xmath17 ) .
4 . for all the open sets @xmath330 , given @xmath342 if @xmath331 , @xmath333 then there exists open sets @xmath343 , @xmath344 such that @xmath345 , @xmath346 and @xmath347 .
5 . for all the open sets
@xmath330 and @xmath348 not both @xmath349 and @xmath350 given a self adjoint operator @xmath301 , let @xmath351 be the spectral family of operators associated to @xmath40 ( see @xcite ) .
the family @xmath351 is a family of operators contained in @xmath296 which satisfy : where @xmath365 denotes the complement of @xmath366 in @xmath9 .
we want to show that @xmath367 and @xmath368 define a real number instead of @xmath369 , here again we have @xmath370 if @xmath330 , then we can use this notation without generating confusion ] , proving that the conditions 1 - 5 above hold : 1 .
a- ) @xmath371 is an open cover of @xmath309 , indeed , + @xmath372 + we have @xmath373 thus @xmath374 .
+ b- ) since @xmath375 , there exists @xmath376 such that @xmath377 , otherwise @xmath378 .
therefore , taking @xmath379 by definition we have @xmath380 .
a- ) let @xmath330 be an open set and @xmath381 .
suppose @xmath382 and @xmath383 , then @xmath384 .
since @xmath385 we have @xmath386 and @xmath387 .
thus @xmath388 which implies @xmath389 .
+ b- ) on the other hand if we suppose @xmath382 and @xmath390 , there exists @xmath391 , @xmath392 such that @xmath393 . since @xmath394
we can conclude that @xmath395 .
a- ) let @xmath330 be an open set and @xmath396 be such that @xmath383 .
since @xmath384 , @xmath397 is an open cover of @xmath338 ; indeed + @xmath398 + we have then @xmath399 , which implies @xmath400 .
+ b- ) if @xmath395 by definition there exists @xmath382 such that @xmath401 .
therefore given @xmath391 , such that @xmath402 we have @xmath403 .
4 . @xmath330 be an open set and @xmath404 such that @xmath333 .
suppose @xmath405 then @xmath406 .
thus for @xmath407 exists @xmath179 such that @xmath333 and @xmath406 , then @xmath390 . on the other hand if we suppose @xmath408 , then for all @xmath392 , @xmath409 .
particularly @xmath384 then @xmath383 . 5 .
let @xmath410 be an open set and @xmath348 .
suppose @xmath383 , then @xmath384 and for all @xmath382 , @xmath388 thus @xmath411 . in the same way
if we suppose @xmath395 , there exists @xmath382 such that @xmath401 , since @xmath412 we have @xmath406 .
thus @xmath413 .
thus each spectral resolution of the identity defines a real number in @xmath323 .
+ on the other hand if @xmath414 @xmath415 define a real number in @xmath323 we can find an spectral resolution of the identity in the next way .
given @xmath376 consider the set @xmath416 from ( 1 ) in the definition of a dedekind cut , for any @xmath362 , @xmath375 , there exists an open cover @xmath418 of @xmath309 and @xmath419 such that @xmath420 . on the other hand from ( 5 ) in the definition of a dedekind cut
we have that for any @xmath421 , @xmath422 .
summarising we have an open cover @xmath423 of @xmath309 and @xmath424 such that for every open set @xmath421 , @xmath425 .
therefore we have proved : @xmath426 for all @xmath362 .
if @xmath427 we have @xmath428 for all @xmath429 but this contradicts [ sf1 ] .
let @xmath431 and @xmath309 an open neighbourhood of @xmath21 . from ( 1 ) in the definition of a dedekind cut
there exists an open set @xmath432 , and @xmath433 such that @xmath434 and @xmath435 . then by property ( 2 ) of the definition of a dedekind cut for all @xmath436 we have @xmath437 .
therefore @xmath438 , since @xmath21 was arbitrary we have @xmath439 .
if @xmath441 , there exists a neighbourhood @xmath309 of @xmath21 such that @xmath442 for all @xmath443 .
let @xmath444 arbitrary , then @xmath442 for all @xmath445 , then @xmath446 for all @xmath444 thus @xmath447 . on the other hand
if @xmath448 , for all @xmath444 there exists a neighbourhood @xmath449 of @xmath21 such that @xmath450 for all @xmath445 .
let @xmath451 , then @xmath452 for all @xmath443 and @xmath453 .
then @xmath441 .
given a real number @xmath454 , define @xmath455 from the above results it is clear that the projections @xmath351 associated to the family of clopen subsets @xmath357 form a spectral resolution of the identity .
since there is a correspondence between self - adjoint operators in @xmath291 and spectral families in @xmath291 , we have a correspondence between the real numbers of the cumulative hierarchy of variable sets over @xmath312 and the self - adjoint operators in @xmath291 .
this is the correspondence proved by takeuti in the context of boolean valued models of set theory , the advantage here is that we get an explicit construction of the real numbers in function of the spectral representation of the respective self - adjoint operator .
this construction will be probably fundamental to understand the relation of quantum expressions and its classical counterparts .
+ there is an alternative proof of takeuti s correspondence in this context that even if it is less constructive than the proof presented above , it has a theoretical interest and it will probably be useful in more general contexts ; for the sake of completeness we include this proof here .
we start with a result which is proved in a more general context in @xcite ( theorem 2 , chapter vi ) , we give here the proof for the hierarchies of variable sets .
let @xmath9 be a topological space , and @xmath227 the hierarchy of variable sets constructed over @xmath9 .
let @xmath362 , then there is a correspondence between the objects forced as real numbers in @xmath323 and the real valued continuous functions over @xmath309 .
let @xmath362 , @xmath456 and @xmath414 @xmath415 such that define a real number in @xmath323 .
consider @xmath457 @xmath458 from the properties 1 - 5 above we have that the above sets defined a dedekind cut in the classical sense then a real number @xmath459 .
define the function @xmath460 such that @xmath461 . to show that the function @xmath462 is continuous consider an interval @xmath463 with @xmath464 and @xmath465 .
we have in particular that @xmath466 and @xmath467 since @xmath468 , then there exists @xmath338 and @xmath469 neighbourhoods of @xmath285 such that @xmath470 and @xmath471 .
thus , for every @xmath472 , we have @xmath473 and @xmath474 thus @xmath475 therefore @xmath476 .
since @xmath285 was arbitrary and the rational intervals form a base we have proved the continuity of @xmath462 .
+ consider now @xmath477 a continuous function and @xmath478 @xmath479 we want to prove that @xmath480 define a real number in @xmath323 i.e. that the conditions 1 - 5 above are satisfied .
let @xmath330 an open set : + 1 .
let @xmath481 , then @xmath482 form an open cover of @xmath338 , and we have @xmath483 and @xmath484 .
+ 2 . follows directly from the definitions of @xmath485 and @xmath486 .
let @xmath331 and @xmath487 , then for any @xmath488 we have @xmath489 . since @xmath34 is continuous there exists a neighbourhood of @xmath285 , @xmath490 and @xmath491 such that for any @xmath492 we have @xmath493 , then @xmath494 .
( analogously for @xmath495 ) .
let @xmath331 and @xmath496 such that @xmath333 .
consider @xmath497 and @xmath498 .
we have @xmath345 and @xmath499 , @xmath500 .
clear from the definition .
let @xmath291 an abelian von neumann subalgebra of operators of a hilbert space @xmath284 , and @xmath312 the gelfand spectrum of @xmath291 with the topology given by the sets @xmath310 as defined above .
there is a correspondence between the real numbers of the hierarchy @xmath227 and the self adjoint operators of @xmath291 .
let @xmath301 be a self adjoint operator and as before @xmath501 the associated spectral family .
define the function @xmath502 such that @xmath503 .
we want to show that @xmath504 is continuous .
consider @xmath505 such that @xmath333 , from the construction of the spectral family ( see @xcite theorem 5.2.2 ) we know @xmath506 which is an open set , then @xmath504 is a continuous function and by the above theorem it is associated with a real number in @xmath323 for each @xmath362 .
+ consider now @xmath507 such that @xmath414 @xmath415 define a real number for each @xmath362 , then from the above theorem we know there exists a continuous function @xmath508 associated to @xmath509 . we know that the sets @xmath310 are clopen sets also in the weak * topology over @xmath9 ( see again @xcite theorem 5.2.2 ) , then the function @xmath34 is also continuous in the weak * topology . therefore by the gelfand representation theorem ( @xcite theorem 4.4.3 ) to @xmath34 corresponds an operator @xmath510 such that @xmath511 for all @xmath512 . since @xmath513 for all @xmath512
, we have that @xmath510 is a self adjoint operator ( see @xcite theorem 4.3.8 ) .
the above results show how in this model deep results of operator theory recover a new interesting meaning .
the value of this new perspective is that it probably contains the way to obtain a new picture of qm with a definite interpretation .
lets see why .
+ given @xmath514 , @xmath515 is a constant dedekind cut over @xmath309 , given by : @xmath516 @xmath517 with @xmath330 open . on the other hand , consider the spectral family @xmath518 such that @xmath519 if @xmath520 and @xmath521 if @xmath522 , then the real number @xmath523 associated to this spectral family over an open set @xmath309 satisfies : for any open set @xmath330 since @xmath525 for any @xmath338 non empty . in the same way we have @xmath526 for any @xmath330 .
therefore @xmath515 is the real number associated to the operator defined by @xmath518 ( which is the operator @xmath527 ) . + as above consider again a self - adjoint operator @xmath528 , @xmath351 the spectral family associated to @xmath40 and @xmath529 the real number defined by @xmath40 . in @xmath530 , with @xmath514 , consider @xmath531 . since @xmath532 we have now remember that in the classical sense if we have two real numbers @xmath539 with @xmath540 and @xmath541 we define @xmath542 if and only if @xmath543 and @xmath544 if and only if @xmath545 . applying this definition to our model the results above translate as @xmath546 and @xmath547 respectively , where we have denoted the real number defined by the self - adjoint operator @xmath40 using the same symbol .
remember from section [ logic ] that @xmath548_x$ ] , @xmath549_x$ ] represent the set of elements of @xmath9 where the respective propositions hold , since @xmath550 we have shown that @xmath551_x\ ] ] @xmath552_x.\ ] ] the proof of this result was given in the context of boolean valued models in @xcite ( theorem 6.2 ) , comparing it with the proof here presented , it becomes clear that the sheaf based approach gives an important simplification . + as before we denote by @xmath556_x'$ ] the projector associated to the open set @xmath557_x$ ] . if @xmath558 is a state vector , we know from the classical formalism of qm that the probability of the observable @xmath40 to assume some value in the interval @xmath559 $ ] in the state @xmath560 is given by but within our model we see that @xmath560 is just measuring the set @xmath557_x$ ] of histories or universes where the proposition @xmath562 holds .
indeed , each quantum state @xmath558 defines a measure @xmath563 over @xmath9 given by : @xmath564 as we just saw these open sets are of the form @xmath565_x$ ] , where @xmath114 is a proposition about the quantum system , thus @xmath309 is the set of histories where the proposition @xmath114 is verified and @xmath566 is measuring the proportion of histories where any proposition is verified .
this is literally the interpretation given to quantum states in the deutsch - everett multiverse interpretation of quantum mechanics .
it is important to note that the type of propositions about the quantum system are not limited just to propositions referring to the values of physical variables as in the example above ; in this case the results are valid for any kind of propositions about the quantum system that can be expressed with the language of set theory , probably even propositions about emergent properties related to the conception of classical spacetime .
+ any new mathematical model that will improve our understanding of qm has to explain or give us a hint of how the classical reality of our everyday experience emerges from the deep quantum reality of the elementary particles that constitute everything .
surprisingly , in the models developed above there is a way to collapse the multiversal model of quantum variable sets to a classical model in such a way that what is perceived by the classical universe is conditioned by the structure of the multiversal model .
this can be done over any sheaf of structures , however , for the sake of clarity , we will explain how this process work on the hierarchy of variable sets .
the main idea in this section will be the notion of generic model and genericity .
this notion was originally introduced by paul cohen in his works on the independence of the continuum hypothesis and the axiom of choice ; becoming then fundamental in modern set theory and model theory . in the context of sheaves of structures it is possible to introduce the notion of generic model and genericity in a more general framework which unifies previous approaches , and where hypothesis about enumerability are not required ( see @xcite ) .
we will see then how to connect these ideas related to the foundations of mathematics to quantum mechanics via the hierarchy of quantum variable sets . + we start with an easy definition .
let @xmath9 be a topological space and @xmath567 the set of open sets of @xmath9 .
let @xmath568 , @xmath569 is called a filter of open sets of @xmath9 if : + i. @xmath570 .
+ ii . if @xmath571 then @xmath572 .
+ iii.given @xmath573 if @xmath64 then @xmath574 .
filters over the base space @xmath9 of a sheaf of structures will be the tools we will use to collapse the intuitionistic sheaf to a classical model . to do this we need an special kind of filters where the essence of being classical is captured in such a way that for each proposition the excluded middle holds in an open set of the filter , and for each existential proposition there exists an open set where the existential is verified in the classical sense .
in other words : let @xmath569 be a filter of open sets of @xmath9 , we say that @xmath569 is a generic filter of @xmath227 if : + i. given @xmath575 a first order formula ( a proposition ) and @xmath48 arbitrary sections of @xmath227 defined on @xmath576 , there exists @xmath577 such that @xmath578 \text { or } \vdash_q\neg\varphi[\sigma_1, ... ,\sigma_n].\ ] ] ii . given @xmath48 arbitrary sections of @xmath227 defined on @xmath576 , and @xmath579 a first order formula .
if @xmath580 then there exist @xmath577 and @xmath21 defined on @xmath338 such that @xmath581 .
we can associate a classical model to @xmath227 in the next way : + let @xmath582=\lim_{\rightarrow p\in\mathcal{f}}v(u)\ ] ] i.e @xmath582=\dot{\bigcup}_{p\in\mathcal{f}}v(p)/_{\sim_{\mathcal{f}}}\ ] ] where for @xmath583 and @xmath584 @xmath585 let @xmath85 $ ] be the class of @xmath21 .
we define relations and functions in the next way : @xmath87, ... ,[\sigma_n])\in \mathcal{r}^{\mathbb{v}^{x}[\mathcal{f } ] } \leftrightarrow\text { there exists } u\in\mathcal{f}:(\sigma_1, ...
,\sigma_n)\in r^{v(u)}\ ] ] @xmath586}([\sigma_1], ... ,[\sigma_n])= [ f^{v(u)}(\sigma_1, ... ,\sigma_n)]\ ] ] @xmath587 if @xmath569 is a generic filter over @xmath9 for @xmath227 we say that @xmath588 $ ] is a generic model .
let @xmath569 be a generic filter over @xmath9 for @xmath227 , then : @xmath589\models\varphi([\sigma_1], ... ,[\sigma_n])&\leftrightarrow&\text { there exists } p\in\mathcal{f}\text { such that } \vdash_p\varphi^g(\sigma_1, ...
,\sigma_n)\\ & \leftrightarrow&\{\lambda\in x:\vdash_{\lambda}\varphi^g(\sigma_1, ... ,\sigma_n)\}\in\mathcal{f},\end{aligned}\ ] ] where @xmath590 is the gdel translation of the formula @xmath114 .
it is not important here to know what is precisely the gdel translation , it is enough to know that @xmath591 is a form to reformulate @xmath114 in a classical equivalent way , for example , the gdel translation of the proposition @xmath592 is @xmath593 which in the sense of classical logic is equivalent .
the important fact to know is that a proposition can be deduced classically if and only if its gdel translation can be deduced intuitionistically . as an example we can choose the base space @xmath9 in such a way that in the collapsed model of @xmath227 the axiom of choice is valid but the continuum hypothesis does not hold ,
this is the intrinsic essence of paul cohen s proof ( see @xcite , @xcite @xcite ) .
+ to see how these results can be applied to quantum mechanics , lets consider again the cumulative hierarchy of quantum variable sets over the spectrum @xmath308 of an abelian von neumann sub algebra of the algebra @xmath594 of bounded operators of a hilbert space @xmath284 .
the important result here is that each history @xmath595 determines a generic filter .
remember that @xmath599 .
if @xmath600 , we have @xmath601 , then @xmath602 then @xmath603 . on the other hand if @xmath604 and @xmath605 is such that @xmath606 , this implies that @xmath607 , therefore @xmath608 then @xmath609 . finally as @xmath610 where @xmath611 is the identity operator , we have for any operator @xmath598 and any @xmath285 in @xmath9 @xmath612 then @xmath613 , thus @xmath614 .
we have shown that @xmath615 is a filter , to see that it is maximal suppose that @xmath616 is a filter such that @xmath617 .
let @xmath618 , then @xmath619 .
therefore @xmath620 then @xmath621 . so we conclude that @xmath622 and @xmath623 , then @xmath624 . thus @xmath625 is a maximal filter and by theorem [ generic ] a generic filter .
this last result tells us that to each history corresponds a classical universe , the collapsed universe @xmath626 $ ] obtained from the sheaf @xmath227 via the generic filters @xmath615 . in collapsed universes the propositions about the quantum system
assume definite truth values as in classical reality . and
these truth values depend on the structure of similar histories , which is represented by the filters @xmath615 which structures depend themselves on the structure of the hilbert space from which they arise . to understand how this process works
we will need to complete the reformulation of quantum mechanics within these models , but if this approach to understand emergence of classical reality is correct , it will reflect a new way to understand emergence in physics . in this case
we will not have the emergence as a weak classical limit where some parameters as velocity , scale or gravity tend to certain values but we will have a kind of emergence more close to the sense of emergence as intended in computation or biology , where a lower level structure , represented in this case by a sheaf of structures , determines what is perceived by a higher level structure represented in this case by the collapsed generic models .
+ any satisfactory reformulation of qm should have three fundamental characteristics ; first it has to settle a definite interpretation of the theory , second it has to explain the emergence of classical reality and finally it has to be flexible enough to be extended to general relativity . in this work
we saw that the model presented here is a good candidate to fulfil the first two conditions . and probably these kind of models are perfect to include gr from a totally new perspective ( see @xcite for example ) .
all these results show that probably we really need a formalism founded over a quantum logic to understand quantum reality and to construct quantum gravity theories .
x. caicedo , http://matematicas.uniandes.edu.co/archivos/publicaciones/logicadeloshacesdeestructura.pdf[_logica de los haces de estructuras ( the logic of sheaves of structures ) .
colombiana cienc .
( 1995 ) , no .
74 , 569586 . | based on the sheaf logic approach to set theoretic forcing , a hierarchy of quantum variable sets is constructed which generalizes and simplifies the analogous construction developed by takeuti on boolean valued models of set theory . over this model two alternative proofs of takeuti s correspondence , between self adjoint operators and the real numbers of the model , are given .
this approach results to be more constructive showing a direct relation with the gelfand representation theorem , revealing also the importance of these results with respect to the interpretation of quantum mechanics in close connection with the deutsch - everett multiversal interpretation .
finally , it is shown how in this context the notion of genericity and the corresponding generic model theorem can help to explain the emergence of classicality also in connection with the deutsch - everett perspective . |
sudan and guruswami s list decoding of reed - solomon codes @xcite has developed into algebraic soft - decision decoding by koetter and vardy @xcite . as reed - solomon codes are widely used in coding applications ,
algebraic soft - decision decoding is regarded as one of the most important developments for reed - solomon codes .
hence there have been many subsequent works to make the decoding method efficient and practical @xcite .
engineers have proposed fast electronic circuits implementing the algebraic soft - decision decoder @xcite .
one may say that the algebraic soft - decision decoding of reed - solomon codes is now in a mature state for deployment in applications @xcite .
reed - solomon codes are the simplest algebraic geometry codes @xcite .
therefore it is natural that the list decoding of reed - solomon codes was soon extended to algebraic geometry codes by shokrollahi and wasserman @xcite and guruswami and sudan @xcite .
however , it seems that no algebraic geometry codes other than reed - solomon codes have been considered for algebraic soft - decision decoding .
one reason for this unbalanced situation is presumably that the complexity of an algebraic soft - decision decoder for algebraic geometry codes would be prohibitively huge as the complexity for reed - solomon codes is already very large .
however , algebraic geometry codes have the advantage that they are longer than reed - solomon codes over the alphabet of the same size , promising better performance .
we may also expect that once we have an explicit formulation of algebraic soft - decision decoding for algebraic geometry codes , some clever ways to reduce the complexity to a practical level may be found , as has happened for reed - solomon codes @xcite . in this work
, we present an algebraic soft - decision decoder for hermitian codes .
hermitian codes are one of the best studied algebraic geometry codes , and they are often regarded as the first candidate among algebraic geometry codes that could compete with reed - solomon codes . to formulate an algebraic soft - decision decoder for hermitian codes
, we basically follow the path set out by koetter and vardy for reed - solomon codes .
thus there are three main steps of the decoding : the multiplicity assignment step , the interpolation step , and the root - finding step . for the multiplicity assignment step and the root - finding step , we may use algorithms in @xcite and @xcite , respectively . here
we focus on the interpolation step , the goal of which is to construct the @xmath0-polynomial whose roots give the candidate codewords . as for mathematical contents , this work is an extension of our previous @xcite and @xcite .
the core contribution of the present work is an algorithm constructing a set of generators of a certain module from which we extract the @xmath0-polynomial using the grbner conversion algorithm given in @xcite . in section 2 ,
we review the definitions of basic concepts and the properties of hermitian curves and codes .
we refer to @xcite and @xcite for the basic theory of algebraic curves and algebraic geometry codes , and @xcite and @xcite for grbner bases and commutative algebra . in section 3
, we formulate the algebraic soft - decision decoding of hermitian codes .
we present our interpolation algorithm in section 4 and a complexity analysis of the decoding algorithm in section 5 . in section 6 ,
we provide some simulation results of the algebraic soft - decision decoder .
as this work is an extension of @xcite , we omitted some proofs that can be found in that work but allowed some similar materials included here for exposition purposes .
let @xmath1 be a prime power , and let @xmath2 denote a finite field with @xmath3 elements .
the hermitian curve @xmath4 is the affine plane curve defined by the absolutely irreducible polynomial @xmath5 over @xmath2 .
the coordinate ring of @xmath6 is the integral domain @xmath7/{\langle{y^q+y - x^{q+1}}\rangle}={\mathbb{f}}[x , y],\ ] ] with @xmath8 and @xmath9 denoting the residue classes of @xmath10 and @xmath11 , respectively .
note that every element of @xmath12 can be written uniquely as a polynomial of @xmath8 and @xmath9 with @xmath9-degree less than @xmath1 , as we have @xmath13 .
so @xmath12 is also a free module of rank @xmath1 over @xmath14 $ ] .
the function field @xmath15 is the quotient field of @xmath12 . for each @xmath16 , there are exactly @xmath1 elements @xmath17 such that @xmath18
. therefore there are @xmath19 rational points @xmath20 of @xmath6 with @xmath21 , which can be grouped into @xmath3 classes of @xmath1 points with the same @xmath8-coordinates .
a rational point @xmath22 of @xmath6 is associated with a maximal ideal @xmath23 , and the local ring @xmath24 of @xmath6 at @xmath22 is the localization of @xmath12 at @xmath25 . for a nonzero @xmath26 ,
the valuation @xmath27 is the largest integer @xmath28 such that @xmath29 .
the projective closure of @xmath6 is a smooth curve with a unique rational point @xmath30 at infinity . the functions @xmath8 and @xmath9 on @xmath6 have poles at @xmath30 of orders
@xmath1 and @xmath31 , respectively , that is , @xmath32 and @xmath33 .
the genus of @xmath6 is given by @xmath34 .
it is well known that the number of rational points of the curve @xmath6 attains the maximum value possible for the genus and the size of the base field . for @xmath35 ,
the @xmath2-linear space @xmath36 has a basis consisting of @xmath37 for @xmath38 , @xmath39 , and @xmath40 .
therefore @xmath41 recall that the hamming space @xmath42 is an @xmath2-linear space with the hamming distance function @xmath43 . for @xmath44 ,
let @xmath45 .
the evaluation map @xmath46 defined by @xmath47 is a linear map over @xmath2 .
we now fix a positive integer @xmath48 .
the hermitian code @xmath49 is defined to be the image of @xmath50 by the evaluation map .
if @xmath51 , then @xmath52 is injective on @xmath50 , and the dimension of @xmath49 is equal to @xmath53 , which is @xmath54 for @xmath55 by the riemann - roch theorem .
note also that the minimum distance of @xmath49 is at least @xmath56 .
define @xmath57 for @xmath44 . for a vector @xmath58 ,
define @xmath59 we can easily prove that @xmath60 if @xmath61 , and @xmath62 otherwise . therefore @xmath63 for all @xmath64 .
let @xmath65 and @xmath66 .
we consider the hermitian curve @xmath6 defined by @xmath67 over @xmath68 .
there are @xmath69 rational points on @xmath6 , @xmath70 let @xmath71 .
the linear space @xmath72 is spanned by the basis @xmath73 .
hermitian code @xmath74 is a linear code over @xmath68 of length @xmath69 and dimension @xmath75 .
we use the following generator matrix for encoding @xmath76 note that the positions @xmath77 form an information set of @xmath78 .
our message is @xmath79 , which is encoded into the codeword @xmath80 the functions @xmath81 are as follows : @xmath82 we will continue this example throughout . the smooth surface @xmath83 has the coordinate ring @xmath84=r[z]$ ] .
the function field @xmath85 is the quotient field of @xmath86 .
a rational point @xmath87 of @xmath88 is a pair @xmath89 with @xmath44 and @xmath90 , and is associated with a maximal ideal @xmath91 .
the local ring @xmath92 of @xmath88 at @xmath87 is the localization of @xmath86 at @xmath93 .
a nonzero function @xmath94 defines a curve on the surface @xmath88 .
the multiplicity of @xmath95 at a rational point @xmath87 , denoted @xmath96 , is the largest integer @xmath28 such that @xmath97 .
we note the following properties of multiplicity on the surface @xmath88 .
let @xmath22 be a rational point of @xmath6 .
* if @xmath26 , then @xmath98 for every @xmath90 . * for @xmath26 , @xmath99 if @xmath100 , and @xmath62 otherwise . * for @xmath101 , @xmath102 for every rational point @xmath87 of @xmath88 .
suppose that some codeword of @xmath49 was sent through a noisy channel .
the output of the channel is some probabilistic information , for each location @xmath103 , of the plausibility of each @xmath90 .
the multiplicity assignment step translates the information to a doubly indexed list @xmath104\ ] ] of nonnegative integers , where we regard @xmath105 as assigned to the point @xmath106 .
the integer value @xmath105 would be chosen roughly proportional to the plausibility of the symbol @xmath107 according to the channel output .
we call @xmath108 the multiplicity matrix . corresponding to @xmath108 , define @xmath109\mid\text{$\operatorname{mult}_{(p_i,\gamma)}(f)\ge m_{i\gamma}$ for $ 1\le i\le n , \gamma\in{\mathbb{f}}$}\}},\ ] ] an ideal of @xmath110 $ ] .
we call @xmath111 the interpolation ideal
. note that by definition @xmath112 for a vector @xmath58 , the score of @xmath113 with respect to @xmath108 is defined as @xmath114 hence @xmath115 is also the sum of the multiplicities of the points through which the curve @xmath116 passes .
the task of the algebraic soft - decision decoder is to find the codeword that has the best score with respect to @xmath108 .
this codeword is presumed to be the most likely to have been sent , given the channel output .
[ djcqoe ] suppose that the codeword in the previous example is sent through a noisy channel and received data gives rise to the matrix of the plausibilities of symbols @xmath117 which is then translated to the multiplicity matrix @xmath118 where the rows are indexed by @xmath119 from top to bottom .
note that neither of the vectors @xmath120 and @xmath121 that have the best score with respect to @xmath108 is a codeword of @xmath74 .
[ lemabcd ] let @xmath122 $ ] be the multiplicity matrix
. then @xmath123/i_m=\sum_{i=1}^n\sum_{\gamma\in{\mathbb{f}}}\binom{m_{i\gamma}+1}{2}.\ ] ] since @xmath111 is a zero - dimensional ideal , @xmath124/i_m&=\sum_{i=1}^n\sum_{\gamma\in{\mathbb{f}}}\dim_{{\mathbb{f}}}{\mathcal{o}}_{(p_i,\gamma)}/i_m{\mathcal{o}}_{(p_i,\gamma)}\\ & = \sum_{i=1}^n\sum_{\gamma\in{\mathbb{f}}}\dim_{{\mathbb{f}}}\hat{\mathcal{o}}_{(p_i,\gamma)}/i_m\hat{\mathcal{o}}_{(p_i,\gamma ) } , \end{split}\ ] ] where @xmath125 denotes the completion of the local ring .
if @xmath126 is a uniformizing parameter of @xmath127 and @xmath128 , then @xmath129 is isomorphic to @xmath130 $ ] .
so @xmath131 is isomorphic to @xmath130/(s , t)^{m_{i\gamma}}$ ] .
the conclusion follows .
let @xmath132 with @xmath133 .
then @xmath134/(i_m+{\langle{z-\mu}\rangle})=\operatorname{score}(v).\ ] ] as in the previous proof , @xmath123/(i_m+{\langle{z-\mu}\rangle } ) = \sum_{i=1}^n\sum_{\gamma\in{\mathbb{f}}}\dim_{{\mathbb{f}}}\hat{\mathcal{o}}_{(p_i,\gamma)}/(i_m+{\langle{z-\mu}\rangle})\hat{\mathcal{o}}_{(p_i,\gamma)}.\ ] ] let @xmath126 be a uniformizing parameter of @xmath127 and @xmath128 again . we find that if @xmath135 , then @xmath136 is isomorphic to @xmath130/({\langle{s , t}\rangle}^{m_{i\gamma}}+{\langle{s - tu}\rangle})={\mathbb{f}}[[t]]/{\langle{t^{m_{i\gamma}}}\rangle}$ ] , but collapses to the zero ring otherwise . here
@xmath48 is some unit in @xmath129 .
the conclusion follows . for @xmath137 $ ] with @xmath138 ,
the @xmath48-weighted degree of @xmath95 is defined as @xmath139 for @xmath140 $ ] and @xmath141 , denote by @xmath142 the element in @xmath12 that is obtained by substituting @xmath143 with @xmath144 in @xmath95 . observe that if @xmath145 , then @xmath146 .
the algebraic soft - decision decoding of hermitian codes rests upon the following [ thmakwe ] suppose @xmath147 is nonzero . if a codeword @xmath148 of @xmath49 with @xmath149 satisfies @xmath150 then @xmath151 .
assume that @xmath152 is not zero in @xmath12 .
then @xmath153/{\langle{f , z-\mu}\rangle})\\ & \ge\dim_{\mathbb{f}}(r[z]/(i_m+{\langle{z-\mu}\rangle})=\operatorname{score}(c ) .
\end{split}\ ] ] for the first equality , see lemma 5 in @xcite .
this implies that if @xmath154 , we must have @xmath151 . in the interpolation step , the decoder picks a polynomial @xmath147 . then by proposition [ thmakwe ] , all codewords
whose score with respect to @xmath108 is big enough can be obtained from the roots of @xmath95 over @xmath12 .
thus the decoder can find among the candidates the codeword that has the best score with respect to @xmath108 .
it should be noted that for the best performance of algebraic soft - decision decoding , it is crucial for the decoder to find a polynomial in @xmath111 with the smallest @xmath48-weighted degree .
having the same weighted degree , the one with smaller degree in @xmath143 is preferred because this reduces the work of the root - finding step .
here the idea of grbner bases is relevant .
we call the elements in the set @xmath155 monomials of @xmath110 $ ] . recall that every element of @xmath110 $ ] can be written as a unique linear combination over @xmath2 of monomials of @xmath110 $ ]
. note that @xmath156 for two monomials @xmath157 , @xmath158 in @xmath159 , we declare @xmath160 if @xmath161 or @xmath162 when tied .
it is easy to verify that @xmath163 is a total order on @xmath159 .
notions such as the leading term and the leading coefficient of @xmath140 $ ] are defined in the usual way . for @xmath140 $ ] ,
the @xmath143-degree of @xmath95 , written @xmath164 , is the degree of @xmath95 as a polynomial in @xmath143 over @xmath12 .
now we define the @xmath0-polynomial of @xmath111 as the unique , up to a constant multiple , element in @xmath111 with the smallest leading term with respect to @xmath163 . by the definition
, the @xmath0-polynomial is an element of @xmath111 with the smallest @xmath48-weighted degree , and moreover it has the smallest @xmath143-degree among such elements .
therefore we may say that the @xmath0-polynomial is an optimal choice for the interpolation step .
the last step of algebraic soft - decision decoding is to compute roots of the @xmath0-polynomial over @xmath12 or the function field @xmath15 . only those roots that belong to @xmath50 yield candidate codewords .
if the list of the candidate codewords is empty , the decoder may declare decoding failure or resort to hard - decision decoding directly from the channel output .
if there are several codewords in the list , then the decoder chooses the codeword that has the best score , and outputs the received message by projecting the codeword on the information set .
[ cjops ] the @xmath0-polynomial of @xmath111 @xmath165 is obtained by the interpolation algorithm in the next section .
it turns out the @xmath0-polynomial has the factorization @xmath166 therefore a root - finding algorithm will output two roots .
the first root @xmath167 gives the codeword @xmath168 whose score is @xmath169 while the second root @xmath170 gives the codeword @xmath171 whose score is @xmath172 .
therefore the decoder chooses @xmath173 , and the received message is @xmath174 which is the correct sent message .
we will need upper bounds on the @xmath48-weighted degree and the @xmath143-degree of the @xmath0-polynomial of @xmath111 .
let @xmath0 denote the @xmath0-polynomial of @xmath111 .
[ propdkwd ] if @xmath175 is a finite set of monomials of @xmath110 $ ] such that @xmath176 then there is a set of coefficients @xmath177 such that @xmath178 .
lemma [ lemabcd ] implies that monomials in @xmath179 are linearly dependent over @xmath2 in @xmath110/i_m$ ] .
on the other hand , they are linearly independent over @xmath2 in @xmath110 $ ] . in a table
, we arrange monomials of @xmath110 $ ] such that the monomials in the same column have the same @xmath48-weighted degree and the monomials in the same row have the same @xmath143-degree .
let weighted degrees increase from left to right and @xmath143-degrees from bottom to top .
note that @xmath180 .
so we have the following table @xmath181 the symbol @xmath182 indicates that there is no monomial for the position .
the table of monomials of @xmath110 $ ] suggests the following formula .
let @xmath183 if @xmath184 is a weierstrass gap at @xmath30 , and @xmath185 otherwise .
note that @xmath186 for @xmath187 .
the number of monomials with @xmath48-weighted degree @xmath184 is @xmath188 let @xmath189 be the smallest integer such that @xmath190 let @xmath191 .
then the @xmath48-weighted degrees and the @xmath143-degrees of monomials up to the @xmath192th monomial are not greater than @xmath189 and @xmath193 , respectively .
now proposition [ propdkwd ] implies @xmath194 and @xmath195 .
@xmath196 , and @xmath186 for @xmath197 since @xmath198 .
so we have @xmath199 for our @xmath108 , @xmath200
. therefore @xmath201 , @xmath202 .
hence @xmath203 and @xmath204 .
let @xmath193 be a positive integer such that @xmath195 . define @xmath205_l={\{f\in r[z]\mid { \text{$z$-$\deg$}}(f)\le l\}}.\ ] ] note that @xmath110_l$ ] is a free module over @xmath12 of rank @xmath206 with a free basis @xmath207 .
define @xmath208_l.\ ] ] clearly @xmath209 is a submodule of @xmath110_l$ ] over @xmath12 .
recall that the ring @xmath210 $ ] is in turn a free module over @xmath14 $ ] of rank @xmath1 , with a free basis @xmath211 .
so we may view @xmath110_l$ ] as a free module of rank @xmath212 over @xmath14 $ ] with a free basis @xmath213 .
the elements of @xmath214_l$ ] will be called monomials of @xmath110_l$ ] .
it is clear that the total order @xmath163 is precisely a monomial order on the free module @xmath110_l$ ] over @xmath14 $ ] .
we also view @xmath209 as a submodule of the free module @xmath110_l$ ] over @xmath14 $ ] .
it is immediate that the @xmath0-polynomial of @xmath111 is also the element of @xmath209 with the smallest leading term with respect to @xmath163 . as a consequence of the definition of grbner bases
, @xmath0 occurs as the smallest element in any grbner basis of the module @xmath209 over @xmath14 $ ] with respect to @xmath163 .
we begin with let @xmath122 $ ] be a doubly indexed list of nonnegative integers . for each @xmath215 , let @xmath216 , and let @xmath217 $ ] . for each @xmath184 with @xmath218 , let @xmath219 be such that @xmath220 .
let @xmath221 $ ] where @xmath222 for @xmath223 and @xmath224 .
then as a module over @xmath12 , @xmath225 where @xmath226 and @xmath227 is such that @xmath228 . by the properties ( i ) ,
( ii ) , ( iii ) of local multiplicity , it is clear that @xmath229 . to show the reverse inclusion ,
let @xmath147 .
we can write @xmath230 for some @xmath231 $ ] and @xmath232 .
let @xmath233 .
if @xmath126 is a uniformizing parameter of @xmath127 , then @xmath126 and @xmath234 form a system of parameters of @xmath92 . recall that the completion @xmath235 is isomorphic to the power series ring @xmath236 $ ] . now
if @xmath237 , then in @xmath235 , @xmath238 for some unit @xmath48 in @xmath235 . since @xmath239 and @xmath240 are algebraically independent over @xmath2 , we see that @xmath241 .
then as this is true for all @xmath44 , it follows that @xmath242 . hence @xmath243 .
again by the properties of local multiplicity , @xmath244 .
thus we showed the reverse inclusion .
recall the multiplicity matrix @xmath122 $ ] .
let @xmath245 .
initially let @xmath246 and @xmath247 .
proceed inductively for @xmath248 .
choose @xmath219 such that @xmath249 if @xmath250 .
let @xmath251 such that @xmath252 .
let @xmath253 now let @xmath254 $ ] and @xmath255 $ ] .
observe @xmath256 for all @xmath257 , and therefore @xmath258 $ ] . by induction ,
we get for @xmath259 , @xmath260 as a module over @xmath12 . here
@xmath261 and @xmath262 for @xmath263 .
we may view the ideal @xmath226 as a module over @xmath14 $ ] .
indeed @xmath264 is a free module of rank @xmath1 over @xmath14 $ ] .
thus we obtain the input is an @xmath265 matrix @xmath122 $ ] of nonnegative integers .
the output is the generators @xmath266 of @xmath209 as a module over @xmath14 $ ] .
repeat steps b1 and b2 for @xmath267 .
* let @xmath268 for @xmath44 .
let @xmath269 . for each @xmath270 ,
let @xmath271 be such that @xmath272 .
set @xmath273 for @xmath274 , where @xmath275 is a set of generators of @xmath276 as a module over @xmath14 $ ] .
when @xmath277 is empty , @xmath278 so @xmath279 . *
set @xmath280 and for @xmath281 , set @xmath282 notice that if we compute @xmath283 by the method in the following subsection , @xmath284 has leading term @xmath285 with respect to lex order @xmath286 .
we continue from example [ djcqoe ] .
we show the first few steps to compute a set of generators of @xmath209 with @xmath202 using algorithm b. initially @xmath287 .
then @xmath288 as we compute in example [ jwqdd ] , @xmath289 as a module over @xmath14 $ ] .
so we set @xmath290 in step b2 , we compute ( setting @xmath291 arbitrarily ) @xmath292 now the matrix of @xmath105 is @xmath293 going on to @xmath294 , we have @xmath295 then @xmath296 as a module over @xmath14 $ ] .
hence @xmath297 now @xmath298 and the matrix of @xmath105 is @xmath299 proceeding this way until @xmath300 , we obtain a set of generators of the module @xmath209 .
we arrange the coefficients ( polynomials in @xmath8 ) of the generators in the following matrix @xmath301 where the rows are @xmath302 , @xmath303 , @xmath304 , @xmath305 ,
, @xmath306 , @xmath307 in this order , and the columns are coefficients of @xmath185 , @xmath9 , @xmath143 , @xmath308 , @xmath309 , @xmath310 , , @xmath311 , @xmath312 in this order .
we now tackle the task of computing a set of generators of @xmath264 as a module over @xmath14 $ ] .
for this , we switch to a different indexing of the rational points of @xmath6 by grouping the @xmath19 rational points into @xmath3 classes with the same @xmath8-coordinates .
thus the rational points are @xmath313 for @xmath314 and @xmath315 .
let @xmath316 if @xmath317 is the point @xmath127 .
also assume that for each @xmath314 , we have arranged the index @xmath318 such that @xmath319 are put in decreasing order , @xmath320 with the new notations , @xmath321 [ prop6 ] for @xmath322 , suppose that @xmath323 $ ] satisfy @xmath324 for all @xmath314 .
define for @xmath325 , @xmath326 then @xmath327 as a module over @xmath14 $ ] .
let @xmath328 .
then for @xmath314 and @xmath315 , @xmath329 therefore @xmath330 .
recall that we may view @xmath12 as a free module of rank @xmath1 over @xmath14 $ ] .
let @xmath331 be the submodule of @xmath12 generated by @xmath332 over @xmath14 $ ] .
then @xmath333 is isomorphic to @xmath334/{\langle{\prod_{1\le a\le q^2}(x-{\alpha}_a)^{\mu_{a , c}}}\rangle}.\ ] ] therefore @xmath335/{\langle{\prod_{1\le a\le q^2}(x-{\alpha}_a)^{\mu_{a , c}}}\rangle } = \sum_{1\le c\le q}\sum_{1\le a\le q^2}\mu_{a , c}.\ ] ] on the other hand , as @xmath336 by its definition , we have @xmath337 hence @xmath338 .
together with @xmath339 , this implies that @xmath340 .
[ jwqdd ] we compute generators @xmath341 , @xmath342 of @xmath343 .
we arrange the points as @xmath344 so that @xmath319 are in decreasing order , @xmath345 we will see in the next subsection that @xmath346 satisfies @xmath347 therefore @xmath348 generates @xmath343 as a module over @xmath14 $ ] . as @xmath350
, we see that @xmath351 in the completion of the local ring at @xmath352 . on the other hand ,
if @xmath353 is a rational point of @xmath6 , then @xmath354 , @xmath355 defines an automorphism of @xmath6 taking @xmath353 to @xmath352 .
hence at @xmath353 , we have @xmath356 now we consider the following problem .
suppose that @xmath357 are rational points on @xmath6 with distinct @xmath358 .
given some positive integers @xmath359 for @xmath360 .
we want to construct @xmath361 with @xmath362 $ ] such that @xmath363 for @xmath360 .
there are at least two ways to do this .
[ [ first - method ] ] first method + + + + + + + + + + + + for @xmath360 , let @xmath364 be the truncation of the series expansion of @xmath9 at @xmath365 modulo @xmath366 , and let @xmath367 $ ] be defined by @xmath368 then @xmath369 satisfies the required conditions by the chinese remainder theorem .
[ [ second - method ] ] second method + + + + + + + + + + + + + a somewhat more explicit way is as follows . if @xmath370 $ ] , then the condition @xmath371 is equivalent to the following linear conditions on the coefficients @xmath372 , @xmath373 for @xmath374 , where @xmath375 except @xmath376 , @xmath377 , @xmath378 for @xmath379 .
now let @xmath380 .
then the required @xmath95 can be determined by solving the linear system @xmath381 for the vector @xmath382 where @xmath383 is a certain vector of length @xmath192 and @xmath179 is a square matrix of size @xmath192 obtained by the horizontal join of @xmath384 matrices @xmath385_{0\le i\le{n-1},0\le j\le \mu_k-1}\ ] ] for @xmath386 .
the matrix @xmath179 is called a confluent vandermonde matrix in the literature , and is known to be invertible ( actually the determinant is @xmath387 @xcite ) .
therefore the linear system has a unique solution .
let us compute @xmath388 in the previous example by the second method . here
@xmath389 . if @xmath390 , then @xmath391 where @xmath392 and @xmath393 .
the solution of this linear system was given in the previous subsection .
for this task , we use the grbner conversion algorithm in @xcite that converts a set of generators of a submodule of @xmath14^n$ ] to a module grbner basis with respect to a special weighted monomial order .
we review the algorithm below .
let @xmath394 .
tuples in @xmath395 are ordered lexicographically such that @xmath352 is the first tuple in @xmath395 and the successor of @xmath396 is @xmath397 if @xmath398 or @xmath399 if @xmath400 .
thus @xmath401 is a basis for @xmath110_l$ ] as an @xmath14$]-module and the weight of the basis element @xmath402 is @xmath403 .
the index of @xmath140_l$ ] is defined to be the largest tuple @xmath396 such that the coefficient of @xmath402 is nonzero . in particular , if the leading term of @xmath140_l$ ] is @xmath404 with respect to @xmath163 , then @xmath405 .
note that @xmath406 for the generators @xmath407 of @xmath209 computed by algorithm b. the algorithm finds the element of @xmath209 with the smallest leading term .
initially set @xmath407 to be the initial set of generators of the module @xmath209 computed by algorithm b. let @xmath408 during the execution of the algorithm . for @xmath409 and @xmath410 in @xmath395 , we abbreviate @xmath411 . *
set @xmath412 . *
set @xmath28 to the successor of @xmath28 .
if @xmath413 , then proceed ; otherwise go to step i6 . *
set @xmath414 .
if @xmath415 , then go to step i2 . *
set @xmath416 and @xmath417 . *
if @xmath418 , then set @xmath419 if @xmath420 , then set , storing @xmath421 in a temporary variable , @xmath422 go back to step i3 .
* output @xmath407 with the smallest leading term , and the algorithm terminates .
algorithm i converts the initial basis given in to a grbner basis with respect to the order @xmath163 .
the computed grbner basis is @xmath423 the twelve rows represent the polynomials in the grbner basis of the module @xmath209 over @xmath14 $ ] . comparing the weights of the leading coefficients of the polynomials , which lie on the diagonal , we find that the polynomial represented by the eleventh row is the required @xmath0-polynomial of the ideal @xmath111 , given explicitly in example [ cjops ] equation .
elements of @xmath12 can be written uniquely as polynomials in @xmath9 of degree less than @xmath1 with coefficients in @xmath14 $ ] .
we assume that for computations in @xmath12 , we use this representation of elements of @xmath12 .
also we think of @xmath424 and @xmath425 for @xmath26 in this representation .
note that a straightforward way of multiplying two elements @xmath426 of @xmath12 takes @xmath427 multiplications on @xmath2 and that @xmath428 if @xmath429 and @xmath430 .
first we consider computing @xmath362 $ ] satisfying @xmath431 for @xmath360 as in section [ subsecc ] .
this computation takes @xmath432 multiplications on @xmath2 where @xmath380 , if we use gaussian elimination to solve the linear system . note also @xmath433 .
next we consider computing @xmath434 according to proposition [ prop6 ] in section [ subsecb ] .
the first product @xmath435 on the right side of has at most @xmath436 linear factors .
hence @xmath435 can be computed with @xmath437 multiplications on @xmath2 .
note @xmath438 . on the other hand , as @xmath439 , the second product @xmath440 can be computed with @xmath441 multiplications on @xmath2 .
note @xmath442 and @xmath443 .
then @xmath435 and @xmath440 can be multiplied with @xmath441 multiplications on @xmath2 .
hence , in total , computing @xmath434 takes @xmath441 multiplications on @xmath2 .
note @xmath444 and @xmath445 .
now we consider computations in steps b2 and b3 of algorithm b in section [ subseca ] .
fix @xmath446 .
computing @xmath283 ( @xmath447 ) , as shown above , takes @xmath448 multiplications on @xmath2 for each @xmath449 .
computing @xmath450 can be done with @xmath451 multiplications on @xmath2 .
note @xmath452 .
let @xmath453 denote the product of the right side in .
it is easy to verify @xmath454 and @xmath455 if @xmath456 .
so computing @xmath457 takes @xmath458 multiplications on @xmath2 .
note @xmath459 .
computing @xmath460 takes @xmath461 multiplications on @xmath2 .
summing up , an execution of algorithm b takes @xmath462 multiplications on @xmath2 .
lastly noting @xmath463 and using a result in @xcite , we see that an execution of algorithm i takes @xmath464 multiplications on @xmath2 .
therefore the algebraic soft - decision decoder of hermitian codes can be implemented in a way that takes @xmath465 multiplications on @xmath2 .
we implemented the algebraic soft - decision decoder ( sdd ) for hermitian codes in software . in this section ,
we present some simulation results that show the performance of the algebraic soft - decision decode for half - rate hermitian codes .
first we describe the general setup of our simulations .
we assume the awgn channel . for qpsk and qam modulations ,
the signal points correspond one - to - one with the symbols in the finite field over which the code is defined , and the posterior probabilities of the symbols are simply set to those of the corresponding signal points . for bpsk , each of the symbols correspond with a bit sequence , and the posterior probabilities of the symbols are set to the products of the posterior probabilities of the bits .
koetter and vardy s multiplicity assignment algorithm @xcite is used to translate the posterior probabilities to the values of the multiplicity matrix .
the multiplicity assignment algorithm accepts a parameter @xmath277 that limits the @xmath143-degree of the @xmath0-polynomial , thereby the list size of the candidate codewords to at most @xmath277 . from the multiplicity matrix ,
our interpolation algorithm finds the @xmath0-polynomial .
then wu and siegel s root - finding algorithm @xcite is used to compute the roots of the @xmath0-polynomial .
the list of candidate codewords is then formed from the roots . if the list is empty , then the decoder simply output the message part of the received vector determined by hard - decision directly from the posterior probabilities of the symbols
. if the list is not empty , the decoder outputs the message from the codeword that has the best score with respect to the multiplicity matrix .
the smallest field over which hermitian codes are defined is @xmath68 and the length of these codes is @xmath69 .
the length is extremely small , and it is perhaps unrealistic to expect the codes to be used in practice .
however the codes are amenable for simulations with somewhat larger @xmath277 .
figure [ fig1 ] show the performance of the half - rate @xmath466 $ ] hermitian code with qpsk .
the example used in previous sections was sampled from this simulation with snr @xmath467 and @xmath468 .
( 10,0 ) ; ( 2,6 )
( 10,6 ) ; ( 2,0 )
( 2,6 ) ; ( 10,0 )
( 10,6 ) ; in -1 , ... , -6 in 0.301 , 0.477 , 0.602 , 0.698 , 0.778 , 0.845 , 0.903 , 0.954,1 ( 2 , )
( 10 , ) ; in 2 , ... , 10 ( , 0 ) ( , 6 ) ; /in 2 , ... , 10 ( 0pt,2pt ) ( 0pt,0pt ) node[below ] @xmath469 ; /in 0 , ... , -6 ( 2pt,0pt ) ( 0pt,0pt ) node[left ] @xmath470 ; plot[yshift=6 cm ] fileawgn-qpsk-hd-uncoded-ber.table ; plot[mark = oplus , yshift=6 cm ] fileawgn - qpsk - sd - hermitian[8,4]-asd - l5-fer.table ; plot[mark = o , yshift=6 cm ] fileawgn - qpsk - sd - hermitian[8,4]-asd - l5-ber.table ; ( 9.8,5.8 ) node[legend , text width=15em ] plot coordinates ( -.5,0)(0,0 ) ; ber of uncoded qpsk + plot[mark = oplus , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; fer of hermitian [ 8,4 ] algebraic sdd for @xmath468 + plot[mark = o , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; ber of hermitian [ 8,4 ] algebraic sdd for @xmath468 + ; ( 6,0 ) [ yshift=-15pt ] node[below ] @xmath471 in db ; ( 2,3 ) [ xshift=-40pt ] node[rotate=90 ] error rate ; figures [ fig3 ] and [ fig4 ] show the performance of @xmath472 $ ] hermitian code with bpsk modulation .
for comparison , the figures also show the performance of the half - rate @xmath473 $ ] reed - solomon code .
observe that the performance curve of hermitian code more steeply decrease than that of reed - solomon code , and from around 5 db , the hermitian code outperforms the reed - solomon code . ( 0,0 )
( 10,0 ) ; ( 0,-7 )
( 10,-7 ) ; ( 0,0 )
( 0,-7 ) ; ( 10,0 )
( 10,-7 ) ; in -1 , ... , -7 in 0.301 , 0.477 , 0.602 , 0.698 , 0.778 , 0.845 , 0.903 , 0.954,1 ( 0 , ) ( 10 , ) ; in 0 , ... , 10 ( , 0 ) ( , -7 ) ; /in 0 , 1 , ... , 10 ( 0pt,2pt ) ( 0pt,0pt ) node[below ] @xmath469 ; /in 0 , ... , -7 ( 2pt,0pt )
( 0pt,0pt ) node[left ] @xmath470 ; plot fileawgn-bpsk-hd-uncoded.table ; plot[mark = triangle ] fileawgn - bpsk - sd - rs[16,8]-asd - l1-ber.table ; plot[mark = triangle ] fileawgn - bpsk - sd - rs[16,8]-asd - l2-ber.table ; plot[mark = triangle ] fileawgn - bpsk - sd - rs[16,8]-asd - l3-ber.table ; plot[mark = o ] fileawgn - bpsk - sd - hermitian[64,32]-asd - l1-ber.table ; plot[mark = o ] fileawgn - bpsk - sd - hermitian[64,32]-asd - l2-ber.table ; plot[mark = o ] fileawgn - bpsk - sd - hermitian[64,32]-asd - l3-ber.table ; ( .2,.2 ) node[legendsw , text width=14em ] plot coordinates ( -.5,0)(0,0 ) ; uncoded bpsk + plot[mark = triangle , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; rs [ 16,8 ] algebraic sdd for @xmath474 + plot[mark = o , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; hermitian [ 64,32 ] algebraic sdd for @xmath474 + ; ( 5,0 ) [ yshift=-15pt ] node[below ] @xmath471 in db ; ( 0,3.5 ) [ xshift=-40pt ] node[rotate=90 ] bit error rate ; ( 0,0 ) ( 10,0 ) ; ( 0,-7 )
( 10,-7 ) ; ( 0,0 ) ( 0,-7 ) ; ( 10,0 )
( 10,-7 ) ; in -1 , ... , -7 in 0.301 , 0.477 , 0.602 , 0.698 , 0.778 , 0.845 , 0.903 , 0.954,1 ( 0 , )
( 10 , ) ; in 0 , ... , 10 ( , 0 ) ( , -7 ) ; /in 0 , 1 , ... , 10 ( 0pt,2pt )
( 0pt,0pt ) node[below ] @xmath469 ; /in 0 , ... , -7 ( 2pt,0pt ) ( 0pt,0pt ) node[left ] @xmath470 ; plot[mark = triangle ] fileawgn - bpsk - sd - rs[16,8]-asd - l1-fer.table ; plot[mark = triangle ] fileawgn - bpsk - sd - rs[16,8]-asd - l2-fer.table ; plot[mark = triangle ] fileawgn - bpsk - sd - rs[16,8]-asd - l3-fer.table ; plot[mark = o ] fileawgn - bpsk - sd - hermitian[64,32]-asd - l1-fer.table ; plot[mark = o ] fileawgn - bpsk - sd - hermitian[64,32]-asd - l2-fer.table ; plot[mark = o ] fileawgn - bpsk - sd - hermitian[64,32]-asd - l3-fer.table ; ( .2,.2 ) node[legendsw , text width=14em ] plot[mark = triangle , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; rs [ 16,8 ] algebraic sdd for @xmath474 + plot[mark = o , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; hermitian [ 64,32 ] algebraic sdd for @xmath474 + ; ( 5,0 ) [ yshift=-15pt ] node[below ] @xmath471 in db ; ( 0,3.5 ) [ xshift=-40pt ] node[rotate=90 ] frame error rate ; figures [ fig5 ] and [ fig6 ] also show that the hermitian code outperforms the reed - solomon code with @xmath475-qam modulation , from around 8 db onward .
( 14,0 ) ; ( 2,-7 )
( 14,-7 ) ; ( 2,0 )
( 2,-7 ) ; ( 14,0 )
( 14,-7 ) ; in -1 , ... , -7 in 0.301 , 0.477 , 0.602 , 0.698 , 0.778 , 0.845 , 0.903 , 0.954,1 ( 2 , )
( 14 , ) ; in 2 , ... , 14 ( , 0 )
( , -7 ) ; /in 2 , ... , 14 ( 0pt,2pt ) ( 0pt,0pt ) node[below ] @xmath469 ; /in 0 , ... , -7 ( 2pt,0pt )
( 0pt,0pt ) node[left ] @xmath470 ; plot fileawgn-16qam-hd-uncoded-ber.table ; plot[mark = triangle ] fileawgn-16qam - sd - rs[16,8]-asd - l1-ber.table ; plot[mark = triangle ] fileawgn-16qam - sd - rs[16,8]-asd - l2-ber.table ; plot[mark = triangle ] fileawgn-16qam - sd - rs[16,8]-asd - l3-ber.table ; plot[mark = o ] fileawgn-16qam - sd - hermitian[64,32]-asd - l1-ber.table ; plot[mark = o ] fileawgn-16qam - sd - hermitian[64,32]-asd - l2-ber.table ; plot[mark = o ] fileawgn-16qam - sd - hermitian[64,32]-asd - l3-ber.table ; ( 2.2,.2 ) node[legendsw , text width=14em ] plot coordinates ( -.5,0)(0,0 ) ; uncoded @xmath475-qam + plot[mark = triangle , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; rs [ 16,8 ] algebraic sdd for @xmath474 + plot[mark = o , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; hermitian [ 64,32 ] algebraic sdd for @xmath474 + ; ( 8,0 ) [ yshift=-15pt ] node[below ] @xmath471 in db ; ( 2,3.5 ) [ xshift=-40pt ] node[rotate=90 ] bit error rate ; ( 2,0 ) ( 14,0 ) ; ( 2,-7 ) ( 14,-7 ) ; ( 2,0 ) ( 2,-7 ) ; ( 14,0 )
( 14,-7 ) ; in -1 , ... , -7 in 0.301 , 0.477 , 0.602 , 0.698 , 0.778 , 0.845 , 0.903 , 0.954,1 ( 2 , )
( 14 , ) ; in 2 , ... , 14 ( , 0 ) ( , -7 ) ; /in 2 , ... , 14 ( 0pt,2pt ) ( 0pt,0pt ) node[below ] @xmath469 ; /in 0 , ... , -7 ( 2pt,0pt ) ( 0pt,0pt ) node[left ] @xmath470 ; plot[mark = triangle ] fileawgn-16qam - sd - rs[16,8]-asd - l1-fer.table ; plot[mark = triangle ] fileawgn-16qam - sd - rs[16,8]-asd - l2-fer.table ; plot[mark = triangle ] fileawgn-16qam - sd - rs[16,8]-asd - l3-fer.table ; plot[mark = o ] fileawgn-16qam - sd - hermitian[64,32]-asd - l1-fer.table ; plot[mark = o ] fileawgn-16qam - sd - hermitian[64,32]-asd - l2-fer.table ; plot[mark = o ] fileawgn-16qam - sd - hermitian[64,32]-asd - l3-fer.table ; ( 2.2,.2 ) node[legendsw , text width=14em ] plot[mark = triangle , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; rs [ 16,8 ] algebraic sdd for @xmath474 + plot[mark = o , mark indices=2 ] coordinates ( -.5,0)(-.25,0)(0,0 ) ; hermitian [ 64,32 ] algebraic sdd for @xmath474 + ; ( 8,0 ) [ yshift=-15pt ] node[below ] @xmath471 in db ; ( 2,3.5 ) [ xshift=-40pt ] node[rotate=90 ] frame error rate ;
we presented an algebraic soft - decision decoder for hermitian codes .
software simulations show that hermitian codes perform better than reed - solomon codes for algebraic soft - decision decoding , as expected .
however , for the decoder to be really practical , reduction of the computational complexity remains an important problem .
one promising avenue is to generalize the idea of complexity reduction for reed - solomon codes in @xcite .
designing efficient electric circuits implementing the decoder is of course an issue to explore .
the extent of our computer simulations of the decoding algorithm was limited by our computing resources
. it would be good to have analytic results about the performance of the decoding algorithm .
there have been several analytic performance analyses for the algebraic soft - decision decoding of reed - solomon codes @xcite .
similar analyses may be done for hermitian codes .
our description of the decoding algorithm is interwoven with the particular structure of hermitian codes .
however , the underlying principle of the decoding algorithm seems to apply to a wider class of algebraic geometry codes . in particular , plane algebraic curves with one point at infinity are immediate candidates .
we leave an adequate treatment of this subject as a remaining work .
v. olshevsky and m. a. shokrollahi , `` a displacement approach to decoding algebraic codes , '' in _ fast algorithms for structured matrices : theory and applications _ ,
math.1em plus 0.5em minus 0.4em amer . math .
soc . , 2003 , vol .
265292 .
j. ma , a. vardy , and z. wang , `` efficient fast interpolation architecture for soft - decision decoding of reed - solomon codes , '' in _ proc .
ieee symp .
circuits and systems _ , kos , greece , may 2006 , pp . 48234826 .
j. ma , a. vardy , z. wang , and q. chen , `` direct root computation architecture for algebraic soft - decision decoding of reed - solomon codes , '' in _ proc .
ieee symp .
circuits and systems _
, new orleans , la , may 2007 , pp
. 14091412 . | an algebraic soft - decision decoder for hermitian codes is presented .
we apply koetter and vardy s soft - decision decoding framework , now well established for reed - solomon codes , to hermitian codes .
first we provide an algebraic foundation for soft - decision decoding
. then we present an interpolation algorithm finding the @xmath0-polynomial that plays a key role in the decoding . with some simulation results ,
we compare performances of the algebraic soft - decision decoders for hermitian codes and reed - solomon codes , favorable to the former .
hermitian codes , algebraic soft - decision decoding , interpolation algorithm , grbner bases . |
extended thermal x - ray emission from the hot ( @xmath0 k ) interstellar medium ( ism ) in giant elliptical galaxies is usually interpreted in terms of a ` cooling flow ' scenario [ see loewenstein for a recent review ] .
it implies that radiative cooling of optically thin hot plasma drives a subsonic inflow towards the centre of a potential well formed by the stellar component and a massive dark halo .
galactic cooling flows and cooling flows in clusters of galaxies are essentially different owing to a difference in the origin of the two media .
the intracluster medium is mostly primordial , cools from its high virial temperature and is accreted by the central galaxy , which provides a focus for the flow .
in contrast , the ism in elliptical galaxies is constantly replenished by mass loss from evolved stars .
this gas must be thermalized in the galaxy s gravity field and heated to x - ray temperatures from which it may cool down again .
thus , in hydrodynamic terms , galactic cooling flows are defined as systems where the mass flux is dominated by source terms as opposed to ` boundary terms ' as in cluster cooling flows @xcite .
several potential heat sources have been considered to counterbalance radiative cooling of the hot ism in ellipticals .
the energy supplied with normal stellar mass loss is limited by the value of the stellar velocity dispersion and may only provide temperatures a factor of @xmath1 lower than the ism temperatures @xcite . in contrast , gravitational energy liberated in case of gas _ inflow _ would provide too much heat to the central regions [ thomas and references therein ] .
heating by supernovae ( sn ) type ia remains a controversial issue . on the one hand ,
the sn rate is not very well constrained .
its value varies from @xmath2 snu , suggested by van den bergh & tammann , the uncertainty being a factor of the order of 1.5 , to @xmath3 snu derived by cappellaro , evans & turatto .
is the hubble constant in units of @xmath4 km s@xmath5mpc@xmath5 ; 1 snu = 1 supernova per century per @xmath6 solar bolometric luminosities . ] on the other hand , the low iron content of the diffuse hot gas in ellipticals estimated from high - quality broad band x - ray spectra provided by asca for a single - temperature plasma model @xcite suggests a very low efficiency of ism enrichment by type ia sne .
this implies either that the sn rate is lower than @xmath7 snu @xcite or that sn ejecta do not mix with the hot ism @xcite .
however , the multi - parametric analysis of x - ray spectra is a complex procedure based on iron l line diagnostic tools , and requires accurate atomic physics data @xcite .
the procedure is also model - dependent . using two - temperature multi - phase plasma models
buote obtained substantially better spectral fits for the same data sets with iron abundances of @xmath8 solar and relative element abundances fixed at their solar values .
his results are consistent with the type ia sn rate being up to a factor of @xmath1 lower than reported by cappellaro et al . .
clearly , better quality data are required to determine the type ia sn heating rate more precisely .
all of the above mentioned energy sources may be described as being continuously distributed within a galaxy and their local rates depending on the stellar mass density and velocity dispersion , the gas inflow velocity profile , and the shape of the gravitational potential .
there is a ` global problem ' @xcite of balancing heating and cooling both in the centre and in the outer regions , since the source terms depend in different ways on physical characteristics which vary with radius .
however , empirical scaling laws for ellipticals and the physics of radiative cooling and thermal instability imply certain restrictions on the radial dependences .
one possible solution to this problem is a hydrostatic hot gas recycling model @xcite for hot coronae of elliptical galaxies .
in the hydrostatic gas configuration all that remains of the gas dynamical equations are the algebraic source terms which balance to zero . in this
model two equilibrium conditions are simultaneously satisfied : ( 1 ) the stellar mass loss rate exactly balances the rate that mass cools locally from the flow ( dropout ) and ( 2 ) the rate that thermal energy is radiated away is exactly balanced by the energy input from stars and supernovae .
the recycling model can be used as a tool to distinguish between inflow and outflow regimes for a given set of galaxy parameters ( kritsuk , bhringer & mller 1998 ) .
periods of sporadic activity can drastically change the thermal state of the ism in the inner ` cooling flow ' region of a cluster or of an elliptical on a time scale of @xmath9 yr .
there are numerous observations of an interaction between radio sources and the hot gas both in central cluster galaxies and other ellipticals @xcite .
it is this complex central region that is most sensitive to any imbalance in the energy budget of the ism . since the ` thermal ' time scale is shorter there , the core of a few kpc in radius is the first to reach a new equilibrium state corresponding to the current status of heating .
high spatial and spectral resolution x - ray observations of these central regions are critical to distinguish among the various heat sources and flow regimes in the core .
the response of a cooling flow to energy injection by the central black hole that is fed by the cooling flow was simulated by binney & tabor ( 1995 ) . in their spherically symmetric numerical models
the black hole heats the central kiloparsec of the ism as soon as it begins to swallow gas from the ism .
the resulting expansion of the heated gas eliminates the cusped density profile and establishes a core that is roughly 2 kpc in radius .
after the central heat source has switched off , the core again cools catastrophically within 0.5 gyr . among several limitations of the model ( which include a peculiar feedback mechanism ) the authors
particularly note the restriction of spherical symmetry .
both the jet and the ensuing convective instabilities can not be properly modelled due to this symmetry restriction . in this paper
we analyze the hot gas flow regimes , which develop in elliptical galaxies in response to a small imbalance in the energy budget caused by a global instability of the equilibrium recycling model @xcite .
we use multi - dimensional numerical simulations to investigate compact cooling inflows and subsonic convection two alternative saturated nonlinear regimes for a slightly disturbed thermal equilibrium in the core .
here , we restrict ourselves to spherically symmetric ( on the average ) convective flows , which have not been studied before . the symmetry restriction will be relaxed in a subsequent paper which more closely addresses the astrophysical aspects of the problem .
the analysis presented here provides a basis for a better understanding of more complex convective flows and allows for a direct comparison with one - dimensional models .
a short description of our basic model is given in section 2 .
section 3 describes the numerical method .
the results of numerical experiments are presented in section 4 together with the detailed properties of the inner convective cores . in section 5
we summarize the results and discuss effects to be expected for the dynamical state of the hot gas in ellipticals from a larger imbalance between cooling and heating .
our numerical experiments are based on a physical model , which has been described in kritsuk et al . .
we refer the reader to that paper and references therein for a more detailed description than given below .
we assume that the local mass budget of the hot ism inside the galaxy is determined exclusively by mass supply from the old stellar population of the galaxy [ in the form of stellar winds , planetary nebulae , supernova explosions , etc . ; a detailed discussion can be found in mathews and mathews & brighenti ] , and by gas condensation to a warm phase due to thermal instabilities .
our motivation for including mass dropout is based on the fact that the assumed mass supply mechanisms inevitably cause nonlinear small scale perturbations in the hot phase , which can trigger thermal instabilities .
the rate of mass dropout is assumed to be a function of local conditions in the hot ism , and given by the expression @xmath10 , where @xmath11 is the density of the hot phase .
the spectrum of the initial small - scale perturbations in the gas and , in particular , its modification by heat conduction are described by a single parameter @xmath12 $ ] .
the heaviside function @xmath13 and the linear instability growth rate @xmath14 are defined as @xmath15 @xmath16 here @xmath17 is the hot gas temperature , @xmath18 is the nonequilibrium zero field radiative cooling function for solar composition and complete ionization of the plasma @xcite , and @xmath19 is the specific heat at constant pressure .
the first term in the _ rhs _ of equation ( [ growthrate ] ) coincides with field s instability criterion , and the second one describes stabilization due to stellar mass loss @xcite , @xmath20 is the specific stellar mass loss rate , and @xmath21 is the local stellar mass density . here
we assume @xmath22 @xcite .
since we are interested in relatively short evolutionary time scales ( @xmath23 gyr ) , we assume @xmath20 to be a constant .
the local energy budget of the hot ism is determined by radiative losses and energy supply from stars ( via sn type ia explosions and via thermalization of the kinetic energy of orbital motion for the gas shed by stars ) .
radiative cooling comprises two energy sinks due to ( i ) direct radiative cooling of the hot phase and ( ii ) due to radiative losses by the condensing material . finally , the hydrodynamic description for the hot phase is based upon the assumption that the sources are homogeneously distributed .
note that this assumption is fulfilled as long as the numerical resolution is not excessively high .
the set of conservation laws for mass , momentum , and energy of the hot phase can then be written as follows : @xmath24 @xmath25 @xmath26 = } \nonumber \\ & & \alpha\rho _ * e _ * -\dot\rho_{{\mathrm { ti}}}(e+p)/\rho-\rho^2\lambda + \rho \bmath{v\cdot\nabla}\phi . \label{energy}\end{aligned}\ ] ] here @xmath27 is the velocity vector , @xmath28 is the pressure , @xmath29 is the adiabatic index , @xmath30 is the specific internal energy , and @xmath31 is the characteristic specific energy of the heat source of temperature @xmath32 .
the energy density @xmath33 .
we define a spherically symmetric gravitational potential @xmath34 to be that of a two - component isothermal sphere , which includes stars and a massive dark halo @xcite .
the galaxy model is characterized by four parameters : the stellar velocity dispersion @xmath35 , the characteristic radius of the stellar mass distribution @xmath36 the ratio of the stellar and dark matter ( dm ) velocity dispersions @xmath37 and the ratio of the dark matter mass density and the stellar mass density at the centre of the galaxy @xmath38 for the prototype galaxy ngc 4472 , we set @xmath39 km s@xmath5 , @xmath40 kpc , @xmath41 , @xmath42 , which allows us to reproduce its surface brightness profiles both in the optical and in x - rays with a reasonable accuracy . in this case
sinks and sources of mass and energy for the hot phase are locally compensated everywhere in the galaxy if the parameters @xmath20 , @xmath43 , and @xmath44 are constant .
hydrostatic equilibrium requires an adjustment of the temperature of the hot gas @xmath45 and of the stellar velocity dispersion @xmath35 .
this is achieved by additional restrictions applied to the values of @xmath43 and @xmath32 coupled by @xmath46 where @xmath47 is the gas constant and the mean molecular weight @xmath48 [ see @xcite for more detail ] . for a specific stellar mass loss rate
@xmath49 s@xmath5 @xcite , hydrostatic equilibrium is achieved with a reasonable mass deposition efficiency of @xmath50 and a characteristic temperature of the heat source of @xmath51 k. this temperature is lower than the @xmath52 k corresponding to a sn type ia rate of @xmath53 snu @xcite , but it is higher than the @xmath54 k obtained for a supernova rate of @xmath55 snu @xcite . note , that we do not claim ngc 4472 to be a _
perfect _ galaxy with an isothermal hot ism , an isothermal stellar system and an isothermal dark halo .
rather we construct a simple model whose physical properties are based on observations of ngc 4472 .
the use of a restricted set of control parameters allows us to investigate in which way(s ) our model differs from the real object .
the evolutionary equations are solved by means of the amra code @xcite , which is based on the direct eulerian version of the ppm method of colella & woodward on an adaptively refined hierarchical system of completely nested grids @xcite . for the present 1d and 2d axisymmetric simulations we use spherical coordinates .
the computational domain extends from @xmath56pc ( @xmath57pc in 1-d simulations ) to @xmath58kpc in radius and covers ( in 2d simulations ) the wedge @xmath59 .
periodic boundary conditions are imposed in the angular direction . in the radial direction hydrostatic equilibrium
is assumed at @xmath60 and at @xmath61 , i.e. the radial ( and theta ) velocity is set to zero and the boundary pressure is defined in such a way that the pressure gradient at the boundary is equal to the hydrostatic one .
we allow for up to 4 levels of refinement with density and pressure jumps of 10 per cent used as refinement criteria . with refinement ratios of 4 and 2 in radius and angle , respectively , the resolution of our 2-d models is equivalent to that of a uniform grid run with @xmath62 zones ( @xmath63pc and @xmath64 ) .
the source terms are calculated explicitly @xcite using the operator split approach .
in the presence of heat conduction the equilibrium recycling model for the hot gas in giant ellipticals is meta - stable @xcite .
it means that , while arbitrary infinitesimal perturbations do not grow , perturbations of a finite amplitude can switch the system to a different equilibrium state or to a nonequilibrium catastrophic evolutionary regime .
since we are mostly interested in studying thermal effects and since the characteristic thermal time scale is usually longer than the dynamical time , it is quite natural to examine the response of the system to a class of perturbations , which do not disturb the hydrostatic equilibrium . in general
, the system can be perturbed in a number of different ways . however , the approximation of spherical symmetry , which allows us to reduce the problem to 1-d , is rather restrictive .
the simplest perturbation of this class is a global density perturbation @xmath65 which , at first sight , may seem to be unphysical because of its global nature
. however , due to stratification of the hot gas , the thermal time scale depends strongly on radius , i.e. for a time scale of a few @xmath66 yr the perturbation only affects the inner region because the rest of the hot gas has not enough time to react .
hence , for a sufficiently short time interval the global perturbation is equivalent to a more local one .
we have computed a series of spherically symmetric solutions with initial global density perturbation amplitudes @xmath67 and 0.2 , and compared the results obtained with different spatial resolution in the central cooling region .
left and right panels of fig . [ inflow ] illustrate our results for @xmath67 .
the left panel shows distributions of density temperature and velocity at @xmath68 , 50 , 100 , 150 , 200 , 250 , and 300 myr for our intermediate resolution run with a radial spacing of 12 pc .
the heavy lines indicate the initial conditions .
the quasi - steady regime is already established at @xmath69 myr after having passed through the linear growth regime which covers most of the earlier epochs .
then , for a very short time , a transient dynamic inflow occurs .
afterwards the flow remains quasi - steady .
the gas temperature in the cooling core drops below the equilibrium value @xmath70 by a factor of @xmath71 .
the density distribution shows a central cusp @xmath72 with a power index @xmath73 .
hence , this model predicts a compact central peak in the x - ray surface brightness .
this peak can be observed if the instrument has sufficient angular resolution , and if finite optical depth effects in the inflowing dense core will not render the approximation of an optically thin plasma invalid .
note , that heat conduction and/or magnetic pressure support can modify the stability of the initial equilibrium model and even suppress the formation of the cooling core @xcite .
the right panel shows the flow variables at @xmath74 myr soon after a compact cooling inflow has been established in the core region of @xmath75 kpc in radius .
the effects of the finite grid resolution are most evident in the velocity plot .
changes in the grid resolution effectively modify the position of the inner boundary thus determining the lowest velocity and temperature , and the highest density values attained on the grid . at the coarsest grid spacing of 48
pc the compact inflow region ( @xmath76 pc ) is covered by less than 10 grid zones and hence remains unresolved .
steep gradients of dependent variables introduce interpolation errors , and errors due to explicit calculation of the source terms .
these errors caused by the finite difference scheme are amplified by the physical thermal instability and , as a result , for @xmath77 the numerical solution displays high frequency noise in the temperature distribution for @xmath78myr .
using the same spatial resolution but doubling the size of the initial perturbation ( @xmath79 ) makes the inflow even more dynamic and the noise amplitude becomes so large that the numerical solution eventually switches from the inflow to the outflow regime ( see fig .
[ sketch ] ) .
grid resolution determines the amount of numerical dissipation of ppm .
although ppm does not utilize artificial viscosity explicitely , its interpolation scheme introduces some amount of dissipation , which depends on grid resolution . when increasing the resolution by a factor of 16 to @xmath80 pc the decreased numerical dissipation is found to be too small to prevent the development of small scale condensations due to the physical thermal instability inherent to flows with radiative cooling
[ see , e.g. , hunter , binney & tabor ] . as a result ,
a quasi - steady inflow solution is obtained ( solid lines in the right panel of fig . [ inflow ] ) , which becomes unstable soon after @xmath81 myr causing the breakup of the central density enhancement into multiple spherical condensation fronts . at an intermediate spatial resolution of 12 pc ( left panels and dashed lines in the right panels of fig . [ inflow ] ) the cooling inflow region
is well resolved by @xmath82 grid zones .
the numerical solution remains stable for at least 500 myr beyond the initial relaxation to a steady state inflow at @xmath83 myr .
note that at higher spatial resolution ( solid lines ; right panels ) the profiles are smoother than at lower grid resolution ( left panels ) .
the ripples present at lower resolution can be eliminated , for instance , by including a small phenomenological diffusion term in the equations , as in binney & tabor , or by taking into account thermal conductivity [ kritsuk et al . ] .
the numerical experiments discussed below are all performed with a minimum radial grid spacing of 12 pc . at a much coarser grid
important hydrodynamic scales remain unresolved , whereas a much finer grid is no longer consistent with our basic assumption that the sinks and sources of mass and energy are distributed continuously .
preliminary results obtained in two dimensions assuming axial symmetry indicate that cooling inflows are stable against random small velocity perturbations .
the stability is maintained due to advection of linear perturbations towards the inner boundary by the accelerating gas inflow . as an alternative to quasi - steady cooling inflows
we have studied a class of subsonic outflow solutions , which can be obtained through negative global perturbations added to the equilibrium density .
in such a case heating and mass supply from evolved stars in the galaxy overtake cooling and mass dropout , and instead of a compact cooling core a hot bubble develops in the centre ( fig .
[ sketch ] ) .
[ outflow ] shows an example of a subsonic outflow solution at @xmath68 , 100 , 200 , and 300 myr for @xmath84 .
in addition to the density , temperature , and velocity distributions we also show the entropy , which is defined here as @xmath85 numerical values for the entropy @xmath86 are given assuming @xmath87 .
as the gas in the hot bubble expands it compresses the ambient medium and thus creates a leading spherical quasi - isobaric condensation wave .
the wave front is clearly seen as a peak in density and as a sharp minimum in the temperature profiles .
the gas temperature in the bubble is a factor of @xmath1 higher than the equilibrium temperature @xmath70 .
it eventaully saturates at a value of about @xmath88 k which is fixed by the heating rate at the centre .
typical expansion velocities in the bubble are about 10 km s@xmath5 .
the maximum velocity continuously grows as the bubble expands while less and less dense outer layers of the hydrostatic corona start to feel the energy imbalance due to the initial global density perturbation .
the bubble will expand beyond 300 myr ( the state shown in fig .
[ outflow ] ) because of the special form of the initial perturbation .
thus , the solution will not tend towards a steady galactic wind . since on a timescale of a few gyrs the heating rate and stellar mass loss vary , such steady wind solutions are unrealistic , anyway .
negative velocities observed at @xmath89 kpc are related to a sound wave , which is triggered at the beginning of the simulation by small deviations of the initial model from an exact hydrostatic state .
the outflow solution is rayleigh - taylor unstable since the initially stable entropy stratification is modified by the expanding bubble . in particular
, one may expect that the unstable local entropy inversion associated with the forming condensation front will be smeared out by convective motion , if the restriction of spherical symmetry of the flow is relaxed , and , thus , such a sharp front will never form in reality . in order to initialize the flow in the axisymmetric case we impose besides a @xmath84 global density perturbation also a small random velocity perturbation with an amplitude of @xmath90 of the adiabatic sound speed .
this additional axisymmetric perturbation helps us to initiate two - dimensional flow . in the central region , @xmath91kpc ,
the grid resolution is @xmath92 zones .
the evolution has been followed for 300 myr .
convection starts to develop after @xmath93 myr from the beginning of the simulation in the form of small - scale vertical finger - like structures .
the fingers originate the central region of unstable entropy inversion .
high entropy material flows upwards and low entropy material
downwards , restoring a marginally stable stratification with an adiabatic temperature gradient @xmath94 as the fingers flowing in opposite directions accelerate , kelvin - helmholtz roll - up transforms them into mushroom - like structures typical for the developing convective pattern ( * ? ? ?
* section 6 ) . by @xmath95 myr
a convective core with a radius of @xmath75 kpc forms and large scale structures become dominant . at that time the radial velocities of the convective motion reach @xmath96 km s@xmath5 , and convection approaches a quasi - stationary regime . still later ,
when the outer gas layers start to feel the imbalance of mass and energy gains and losses due to the global initial density perturbation , these layers become involved into the convective motion , too . as the temperature in the hot bubble grows and the convective core expands , large scale convective structures with higher flow velocities appear .
at @xmath97 myr the linear size of the largest mushroom is about 3 kpc and typical flow velocities are @xmath98 km s@xmath5 .
a little bit earlier , at @xmath99 myr , the maximum mach number of 0.54 is recorded for a stream of upflowing hot gas .
thus , the convection driven by initially small ergs s@xmath5 .
a conservative estimate of @xmath100 ergs s@xmath5 for the typical power supplied by agns in the form of mechanical energy of jets is much larger than this value ( see , e.g. , owen , eilek & kassim ) .
the typical power required to sustain radio core sources in elliptical galaxies is about @xmath101 ergs s@xmath5 .
our simulations show that a smaller ( @xmath102 ) initial density perturbation develops a convective core with a radius of @xmath103 kpc by 300 myr . ]
( @xmath104 ) deviations from local thermodynamic equilibrium remains subsonic . in a realistic 3-d situation
one could expect even lower convective velocities due to the extra dimension available for a redistribution of the kinetic energy .
snapshots of the temperature , entropy and vorticity distributions for the inner 4 kpc are shown in figs [ temperature][vorticity ] .
the temperature maps clearly exhibit the expansion of the hot bubble and the distorted analogue of the spherically symmetric leading condensation front , which is formed by the uppermost parts of the mushroom caps .
the mushrooms are best seen in the entropy maps owing to the formation of an entropy ` plateau ' in the core .
the mean value of the entropy in the core steadily grows .
consequently , the convective pattern changes from blue via green and yellow to red . the vorticity maps highlight small scale structures in the velocity field and help in tracing the mushroom clouds as counter - rotating eddies forming the caps ( fig . [ vorticity ] ) .
the convective core as a whole has a well defined sharp outer boundary defined by the largest mushroom caps .
the global maximum of the hot gas density is typically slightly off centre .
for further details on the flow evolution we refer to a set of computer animations available at http://www.mpa-garching.mpg.de/hydro .
analysis of the angular averages of the hydrodynamic variables allows one to compare the properties of the new equilibrium , the flow evolves slowly on a time scale of hundreds of myrs which is longer than the typical life span of a convective element ( see sections [ fourier ] , [ integral ] ) . ]
established in the convective core with those of the hydrostatic recycling model , and those of the outflow solution .
[ meanflow ] shows angular averages at 300 myr for the central 4 kpc region .
convection has smeared out the condensation front , which forms in the spherically symmetric case . as a result
, the averaged convective solution falls in between the equilibrium model and the 1-d outflow .
the gas density near the very centre is by a factor of @xmath105 smaller than in the equilibrium , but still larger by a factor of @xmath106 than in the 1-d case . unlike in a compact inflow
, there will be no sharp peak in the x - ray brightness in the central 1 kpc in this case ( see also the emissivity distribution in the upper panel of fig .
[ meansrc ] ) . an observation with sub - arcsec resolution of a nearby elliptical galaxy , like ngc 4472 , would be a critical experiment to distinguish between the compact inflow and convective outflow realizations in the core .
the relatively small density fluctuations about the mean indicate that compressibility is not very important , which is consistent with a subsonic convection regime .
the average temperature grows by a factor of @xmath1 in the central 1 kpc and has a shallow minimum at the position of the condensation front .
local relative temperature variations in the convective flow are about @xmath107 , mainly at the interface between the convective core and the ambient hydrostatic gas .
the average pressure profile shows clear deviations from the hydrostatic stratification .
local pressure fluctuations are @xmath108 per cent , and hence imply a quasi - isobaric regime of convection .
the angular averaged entropy distribution is almost flat within a 2 kpc core and rises further outwards .
there is no entropy inversion as in the 1-d case , i.e. the system has settled to a new equilibrium state via nonlinear saturation of the convective instability .
both vertical and horizontal gas velocities are quite high compared to typical cooling flow values of about @xmath109 km s@xmath5 .
however , the mean radial velocity at @xmath97 myr lies in the range @xmath110kms@xmath5 , which is similar to the 1-d case , where @xmath111kms@xmath5 .
thus , there is no considerable mass flux in the radial direction . in order to assess the efficiency of radiative cooling in the turbulent core
, we define the volume emissivity @xmath112 as @xmath113 the two terms on the _ rhs _ in the above equation describe radiative losses from the cooling condensations and direct radiative cooling of the hot gas phase , respectively . the upper panel in fig .
[ meansrc ] shows the distribution of the emissivity with radius for the inner 4 kpc .
the peak central emissivity in the convective core is an order of magnitude lower than in hydrostatic equilibrium . for @xmath114kpc
the emissivity distribution roughly follows the one of the recycling model .
local variations of the emissivity by a factor of @xmath115 are detected in the condensation front region between 2 and 3.5 kpc .
the lower panel in fig.[meansrc ] shows the relative rates of mass supply from stellar winds and mass deposition due to local thermal instabilities .
notice that for @xmath116kpc the mass sink is switched off due to a stabilization by intensive sn - heating ( see equation [ growthrate ] ) .
mass dropout dominates on average at radii from 2 to 3 kpc .
the oscillations of the mass dropout rate are rather intense between 2 and 3.5 kpc , where the convective flow turns over .
all of the best x - ray observations of elliptical galaxy x - ray halos show positive temperature gradients in the central @xmath11710 kpc @xcite .
xmm-_newton _ observations of m87 and its x - ray halo report @xmath118 kev in the central circle of @xmath119 ( @xmath120 kpc ) in radius , a ` steep ' temperature rise to @xmath121 kev at @xmath122 ( @xmath123 kpc ) , and a shallow rise to @xmath124 kev in the range from @xmath125 to @xmath126 @xcite .
yet the measured temperature gradients do not give us much information about the temperature distribution within the central kpc .
only the latter would help us to verify the model assumptions concerning the thermal balance of the hot gas near the centre .
the measured gradients can rather point to a slight non - isothermality of the underlying gravitational potential , which traces the transition from the dark matter halo / cluster dominated outer region to the stellar core of m87 @xcite . at the same time , the flat ( on the average ) entropy distribution in the core region of our model is supported by the assumed spherical symmetry of the underlying gravitational potential , average gas density distribution , and the net heat source . as we mention in the introduction , this assumption is restrictive .
even a small asymmetry , e.g. , in the central heating efficiency would create a hybrid in / out - flow with a more stable ( on the average ) entropy stratification , i.e. with a mild positive average temperature gradient .
as soon as the asymmetric heating is sufficient and , still , centrally concentrated ( as in the case of agn heating ) , such a flow would not necessarily display a central density / x - ray brightness cusp .
observationally , one can expect to see convection in the form of uprising hot bubbles interacting with a background flow of cooling gas .
it is possible , that flows of this kind are observed in x - rays by _
chandra _ and in radio on vla in ngc 1275 @xcite and in m87 @xcite .
such flows seem to be largely subsonic , and the interface of a hot bubble with the background flow does not imply shock waves unlike in the model of heinz , reynolds & begelman .
rather , it would be physically similar to the interface of our convective core with the surrounding hydrostatic corona , displaying lower temperatures , larger mass deposition rates and a higher x - ray surface brightness
. we shall consider such flows in detail elsewhere .
a spectral decomposition of a flow helps to reveal information which is not apparent otherwise , like the distribution of energy over length scales and effects of the finite grid resolution . in order to reveal this information we have computed one - dimensional fourier transforms along the @xmath127-coordinate direction at various radii .
strong gravitational stratification of the gas density and the lack of periodicity in the radial direction would dominate the spectrum in the radial direction , and thus would complicate the analysis . for a scalar quantity of interest , @xmath128
, we define the horizontal fourier transform to be @xmath129 which is a function of time , radius , and the horizontal wavenumber @xmath130 . here
@xmath131 is the horizontal periodic length .
the power spectrum is the square of the modulus of its complex fourier transform , namely @xmath132 .
we have used simple periodograms [ see , e.g. , press , teukolsky , vetterling & flannery ] as the power spectra estimates for discrete data sets given on the finest level grid covering the inner convective region of the computational domain . for the spectral analysis we have used the velocity field .
when the density is constant , the power spectrum of the velocity field is similar to a kinetic energy spectrum . as the gas density varies only slightly about its mean value with angle at a given radius ( see @xmath133 and @xmath134 envelopes in fig .
[ meanflow ] ) , the density factor is always negligible .
however , since radial modes are ignored in our 1d analysis , the sum of the power spectra of the velocity components is not expected to be a measure of the total kinetic energy . since instantaneous power spectra are noisy and since they fluctuate considerably in time , we use temporal averages to estimate the typical power distribution with wavenumber .
the convective core expands as the source terms associated with the stellar system drive the hot bubble at the centre .
hence , we had to take averages over a period of time shorter than the expansion time scale ( @xmath9 yr ) but longer than the lifetime of the largest mushroom - like structures in our convective flow ( @xmath135 myr , see section [ integral ] ) . averaging over a time interval 25 myr sufficiently suppresses the oscillations in the mean spectra at small wavenumbers .
we use angle brackets @xmath136 in the following to denote the computed temporal averages .
[ power ] shows the power spectra of the two velocity components at radii of 0.71 , 1.32 , 1.93 , 2.54 , and 3.15 kpc , respectively .
obviously , the power is concentrated towards the large scales , as already mentioned in section [ flowevol ] when discussing the settling of the convective regime .
the spectra do not show any significant radial dependence .
the mean slope @xmath137 of the spectra @xmath138 is @xmath139 . following the work of porter & woodward , who investigated the numerical dissipation of the ppm scheme
, we estimate that in our simulations numerical dissipation starts to become important for @xmath140 , i.e. for structures smaller than about 16 angular zones .
batchelor ( 1969 ) in his paper on two - dimensional isotropic turbulence derived an analogue of the kolmogorov energy spectrum . for
2d flows the spectrum of vorticity is universal . if @xmath141 is the vorticity spectrum , then @xmath142 is the analogue of kolmogorov s power law . for the energy spectrum of 2d flows
one has @xmath143 . from our fourier analysis
we obtain as expected @xmath144 and @xmath145 in the interval @xmath146 $ ] .
this explains the transition from small scale to large scale mushrooms when convection sets in .
the characteristics of the convective flow in the inner region can also be examined by calculating the angular averaged energy and a set of energy fluxes , which we define below following hurlburt , toomre & massaguer .
the angular integrated total energy equation ( [ energy ] ) can be expressed as @xmath147 with @xmath148 where @xmath149 is the heat flux due to a spherically symmetric expansion or contraction of the model , @xmath150 is the ` enthalpy ' or convective heat flux , @xmath151 is the kinetic heat flux , and @xmath152 is the net energy source term .
these fluxes and the source term are defined as @xmath153 @xmath154 @xmath155 @xmath156 \:r^2\sin{\theta}\:d\theta.\ ] ] the above definitions involve spatially fluctuating variables @xmath157 , where @xmath158 is the angular mean value and @xmath159 is the fluctuation about that mean value .
they allow us to discuss correlations between the radial velocity and a number of fluctuating quantities .
such correlations provide useful insights into energy generation and energy transport mechanisms in the convective interior of the hot ism .
we further define a total energy flux @xmath160 . in a steady state with @xmath161 the total flux does not depend on radius @xmath162 .
if a convective flow achieves a statistically steady state , then for sufficiently long time averages the average total energy flux @xmath163 does not dependent on radius as well , while the individual contributing fluxes may vary considerably with @xmath162 . before turning to the fluctuations
let us consider the inner 4kpc region where convective motion occurs in a mean ` background ' flow after the onset of the instability .
fig.[initialfluxes ] shows the radial dependence of the total energy flux @xmath164 and net source @xmath152 at four representative moments of time before and shortly after the instability sets in . at these times
the total flux is directed outwards and increases from zero at @xmath165 to values around @xmath166erg / s at @xmath167kpc mainly due to a slow overall expansion of the gas .
this expansion is stimulated partly by an outgoing weak sound wave generated by the initially imperfect hydrostatic equilibrium , and also by the increasing source terms blowing a hot bubble at the centre . since the flow is spherically symmetric for @xmath168 myr , the major contribution to the total flux is due to gas expansion in an external gravity field , i.e. @xmath169 . the small wrinkles visible at @xmath170kpc in the flux diagram at @xmath171myr hint towards a contribution from fluctuations of the developing convective flow , but still @xmath172 .
by @xmath97 myr , when the convective flow already covers the whole inner region ( @xmath173 kpc ) , the global expansion driven by the action of the source terms remains the dominant contributor to the total energy flux ( fig .
[ fluxes ] ) .
note that the mean energy flux carried away from the inner region at @xmath174 myr amounts to about 7 per cent of the luminosity of the hot gas in the whole computational domain @xmath175 kpc ( see fig .
[ integral ] ) .
our analysis indicates that the fluctuations play only a secondary role for the energy transport .
the mean angular integrated convective flux , @xmath176 , is positive everywhere except for the region around @xmath177 kpc , where the mean source @xmath178 attains its minimum , yielding an upward convective transport of thermal energy . like in case of an incompressible fluid a positive flux value corresponds to motions where hot fluid parcels rise and cold ones sink .
the depression in @xmath176 at @xmath179 kpc spans the same range in radius as the condensation front smeared by convection . in this region mass dropout due to thermal instabilities dominates over mass supply from stellar winds , and the enthalpy flux , carried by fluctuations from the inner region , decreases .
the mean kinetic flux @xmath180 steadily grows in absolute value , as the convective core becomes more extended giving rise to larger convective structures with higher velocities .
the kinetic flux is positive ( i.e. directed outwards ) at @xmath181 kpc and weakly negative further out in the region where convective motions overshoot through the condensation front .
the sign of @xmath151 is related to differences in strength and filling factor between upward and downward directed motions . in our two - dimensional simulations upflows are considerably faster than downflows in the interior of the convective core and there is no apparent asymmetry in the filling factors ( see the angular velocity diagram in fig .
[ meanflow ] ) . in order to follow the evolution of global characteristics of the flow we computed the total hot gas luminosity @xmath182 and the net mass deposition rate @xmath183 as integrals over the whole computational domain ( fig.[integral ] ) .
the computed total luminosity is close to the x - ray luminosity @xmath184 erg s@xmath5 of ngc 4472 at a distance of 17 mpc .
there is a mild tendency for the luminosity to grow with time as the gas shed by stars in the galaxy slowly accumulates in the convective core .
low amplitude oscillations in the luminosity on a time scale of @xmath185myr correspond to a life cycle of the largest convective structures . assuming a linear scale of @xmath186kpc and @xmath187myr
, one can estimate a characteristic velocity @xmath188kms@xmath5 .
this value is typical for the convective core which developed in the simulations by @xmath189myr ( see fig .
[ meanflow ] ) .
a time scale of 25myr is also a reasonable choice for the time averaging interval to be used when computing mean power spectra . the net mass deposition rate @xmath183 , given by the volume integral @xmath190 is initially positive , since a negative global density perturbation @xmath191 breaks the balance between the mass supply to the hot phase from stellar winds and mass loss from the hot phase due to local thermal instabilities .
the mass supply from stars always dominates , and @xmath183 grows from 0.5 to 0.75m@xmath192yr@xmath5 ( fig.[integral ] ) since the system evolves further away from hydrostatic equilibrium . for our model galaxy with a stellar mass @xmath193m@xmath192 the rate of steady supply of gas from stars , @xmath194m@xmath192yr@xmath5 , implies that the gas condensation rate varies from 0.7 at @xmath68 to @xmath195m@xmath192 yr@xmath5 at 300myr . as one can see in fig .
[ integral ] , @xmath183 slightly oscillates in phase with the x - ray luminosity .
maxima of @xmath183 record moments when streams of rising hot material form new large scale mushroom - like structures .
in this paper we have studied hot gas flows originating in the central few kpc of hot galactic coronae in response to a slight energy supply imbalance in a model initially in equilibrium .
it is shown that a negative energy budget drives a quasi - steady compact cooling inflow , which is stable to small perturbations .
excessive heating creates an isentropic ( on average ) convective core in the vicinity of the galaxy centre .
provided the net energy gain in the core is not too high , the convection remains subsonic with typical flow velocities ( both , radial and angular ) of the order of @xmath196kms@xmath5 .
the characteristic convective pattern consists of a manifold of mushroom - like structures of different scales .
the power spectra of the velocity components are dominated by the largest - scale mushrooms .
the emission weighted gas temperature in the core is slightly rising towards the centre .
unlike in the case of a cooling inflow , the x - ray surface brightness of a convective core does not display a sharp maximum at the centre .
the inner parts of the convective core are dominated by mass supply from evolved stars , while in the colder outer shell mass drop out dominates the mass budget .
large - scale mushrooms ( mainly their caps ) can probably be marginally detected by high resolution x - ray images with good photon statistics , as their volume emissivity is about a factor of 2 higher than that of the surrounding gas .
we have considered an idealized case with a spherically symmetric heat source and small deviations from ` thermal ' equilibrium near the centre .
however , we can extrapolate our results to a more realistic hydrodynamic picture .
first , stronger heating in the vicinity of the centre can drive a supersonic convection with velocities in excess of 400kms@xmath5 and produce a much larger ( here 3 - 4kpc ) convective region .
second , a type ia sn heating rate which is lower all over the galaxy , and an excessive heating by an agn in the central region will create a hybrid flow with an inner convective core and an outer ` cooling flow ' , which will vanish in convection when the inflow velocity is very subsonic .
our preliminary results indicate that such a hybrid dynamic flow can be stable and can last as long as the central source is active .
third , a low type ia sn heating rate without extra energy supply near the centre will drive a more extended massive cooling inflow with an even more pronounced central surface brightness peak . it would be harder to turn such a dynamic inflow into an outflow by a strong perturbation ( e.g. , a burst of agn activity , or a merger ) than a weaker compact inflow .
finally , we expect that an asymmetric central heat source in excess of the equilibrium value would create bubble(s ) filled with a ( mean ) isentropic convective flow with properties similar to the ones described above .
_ chandra _ , _ newton _ and future x - ray missions will teach us more about the hydrodynamics of ` cooling flows ' .
ak was partly supported by the russian foundation for basic research ( project 98 - 02 - 19670 ) and by the federal targeted programme _ integration _ ( project a 0145 ) .
tp was partly supported by grant 2p03d.014.19 from the polish committee for scientific research .
the simulations were performed on the risc cluster of the mpa , and on the sgi power challenge at the interdisciplinary centre for computational modelling in warsaw . | we use hydrodynamic simulations with adaptive grid refinement to study the dependence of hot gas flows in x - ray luminous giant elliptical galaxies on the efficiency of heat supply to the gas .
we consider a number of potential heating mechanisms including type ia supernovae and sporadic nuclear activity of a central super - massive black hole . as a starting point for this research
we use an equilibrium hydrostatic recycling model @xcite .
we show that a compact cooling inflow develops , if the heating is slightly insufficient to counterbalance radiative cooling of the hot gas in the central few kiloparsecs .
an excessive heating in the centre , instead , drives a convectively unstable outflow .
we model the onset of the instability and a quasi - steady convective regime in the core of the galaxy in two - dimensions assuming axial symmetry . provided the power of net energy supply in the core
is not too high , the convection remains subsonic .
the convective pattern is dominated by buoyancy driven large - scale mushroom - like structures . unlike in the case of a cooling inflow , the x - ray surface brightness of an ( on average ) isentropic convective core does not display a sharp maximum at the centre . a hybrid model , which combines a subsonic peripheral cooling inflow with an inner convective core , appears to be stable .
we also discuss observational implications of these results .
[ firstpage ] hydrodynamics instabilities cooling flows galaxies : ism ism dark matter |
in standard @xmath0cdm cosmological models , cold dark matter halos form from the gravitational collapse of dark matter particles , and they assemble hierarchically , such that smaller halos merge to form larger and more massive halos . according to the current paradigm of galaxy formation , galaxies form within halos , due to the cooling of hot gas .
halos and galaxies evolve simultaneously , and the evolution of a galaxy is affected by its host halo .
if the halo is accreted by a larger halo , the galaxy will be affected by it as well , and may interact or merge with the galaxies within the new host halo .
such ` satellite ' galaxies in halo substructures no longer accrete hot gas , which instead is only accreted by the ` central ' galaxy in the halo .
the central galaxy consequently continues to grow , while other massive galaxies may merge into it , and therefore it is expected to be the most luminous and most massive galaxy in the halo . for these reasons , current analytic and semi - analytic models distinguish between central and satellite galaxies , which at a given time are at different stages of evolution , or may have evolved differently .
as galaxies evolve they transform from star - forming late - type galaxies into massive bulge - dominated galaxies with little or no ongoing star formation .
it is thought that central galaxies undergo such a transformation by experiencing a major merger followed by agn feedback preventing additional gas cooling and star formation .
satellite galaxies may have their star formation suppressed or ` quenched ' by a number of other processes , such as ram - pressure stripping of the cold gas reservoir , ` harassment ' by other satellites , and ` strangulation ' following the stripping of the hot gas reservoir , the latter of which appears to be the dominant process ( _ e.g. _ , weinmann et al .
2006 , van den bosch et al .
galaxies in relatively dense environments tend to reside in groups and clusters hosted by massive halos .
recent analyses with galaxy group catalogs have argued that many of these galaxies are very red with very low star formation rates , in contrast with galaxies in low - mass halos in less dense environments , many of which are still quite blue with significant star formation ( _ e.g. _ , weinmann et al .
2006 , berlind et al .
measurements of the environmental dependence of galaxy color have found trends that are qualitatively consistent with these claims ( _ e.g. _ , zehavi et al .
2005 , blanton et al .
2005a , tinker et al .
2007 , coil et al .
2008 ) . in order to better understand galaxy and halo evolution ,
more models are needed that can explain the environmental dependence of color , and more measurements of correlations between color and environment are needed to better constrain such models .
skibba & sheth ( 2008 ) have taken a step in this direction : they developed and tested a halo model of the color dependence of galaxy clustering in the sloan digital sky survey ( sdss ) .
their model successfully explains the correlation between color and environment , quantified by the color mark correlation function , while assuming that all environmental correlations are due to those of halo mass .
they distinguish between central and satellite galaxies , whose properties are assumed to be determined by host halo mass .
the purpose of this paper is to further investigate these central and satellite galaxy colors , and in particular to compare the predictions of the model with measurements from recent galaxy group catalogs ( yang et al .
2007 , berlind et al .
this paper is organized as follows . in the next two sections ,
we briefly introduce the color mark model and the galaxy group catalogs . in section [ groupcatcompare ] , we compare the satellite color - magnitude sequence of the model to that of the yang et al . catalog , and we compare the central and satellite colors of the model and both group catalogs as a function of group richness , which is a useful proxy for halo mass .
we summarize our results in section [ discuss ] .
our halo model of the color dependence of galaxy clustering is described in ( skibba & sheth 2008 ; hereafter ss08 ) , and we refer the reader to this paper for details . briefly , our model is based on the model of luminosity dependent clustering of skibba et al . ( 2006 ) , which explained the observed environmental dependence of luminosity by applying the luminosity - dependent halo occupation distribution ( hod ) that was constrained by the observed luminosity - dependent correlation functions and galaxy number densities in the sdss ( zehavi et al .
2005 , zheng et al .
the model of galaxy colors in ss08 added constraints from the bimodal distribution of @xmath1 colors of sdss galaxies as a function of @xmath2-band luminosity .
we made two assumptions : ( i ) that the bimodality of the color distribution at fixed luminosity is independent of halo mass , and ( ii ) that satellite galaxies tend to follow a particular sequence in the color - magnitude diagram , one that approaches the red sequence with increasing luminosity : @xmath3 these observational constraints and additional assumptions allowed ss08 to model the central and satellite galaxy color ` marks ' as a function of halo mass , @xmath4 and @xmath5 .
ss08 used the central and satellite galaxy marks to model color mark correlation functions , in which all correlations between color and environment are due to those between halo mass and environment .
the modeled mark correlation functions were in very good agreement with their measurements with volume - limited sdss catalogs , reproducing the observed correlations between galaxy color and environment on scales of @xmath6 .
the two - point mark correlation function is simply the ratio @xmath7 , where @xmath8 is the traditional two - point correlation function and @xmath9 is the same sum over galaxy pairs separated by @xmath2 , but with each member of the pair weighted by the ratio of its mark to the mean mark . in practice , we measure the _ projected _ clustering of galaxies , and so we use the following analogous statistic for both the measurements and the models , the marked projected correlation function : @xmath10 where the projected two - point correlation function is @xmath11 if mark correlations are consistent with unity , then the mark is not correlated with the environment at that scale ; if the mark correlations are above unity , which is the case for galaxy luminosity and color ( skibba et al . 2006 ; ss08 ) , then higher values of the mark tend to be located in denser environments at that scale . as an example , we show the @xmath1 color mark correlation function for @xmath12 in figure [ mcfexample ] , reproduced from ss08 ( their figure 6 ) .
the solid curve shows the halo model s prediction , which is in agreement with the measurements of the color mark correlations in the sdss , using petrosian colors .
had we instead used @xmath13 as the mark , the resulting mark correlations would be quantitatively stronger , but with larger uncertainties the sdss measurements would have similar statistical significance and constraining power as the @xmath1 mark measurements ( see skibba et al .
2006 ) .
color mark correlation function for @xmath12 .
points show sdss measurement for petrosian colors , and solid curves show the fiducial model , both taken from skibba & sheth ( 2008 ) .
dashed curves show the range of color mark correlations predicted by the model if color gradients in groups and clusters are included ( hansen et al .
2007 ) . ]
a few assumptions were made in the model that are worth discussing .
firstly , it was assumed that the central galaxies lie at the center of their host dark matter halos , as is commonly assumed in hod and conditional luminosity function ( clf ) studies .
however , central galaxies are often offset from the center of the potential well ( van den bosch et al .
2005 ) , and this offset appears to weakly dependent on galaxy color ( skibba et al . 2008b , in prep . ) .
secondly , it was assumed that satellite galaxies follow the dark matter profile , while satellite galaxy number density profiles have been found to be less concentrated ( _ e.g. _ , hansen et al . 2005 , yang et al .
these two effects are both too weak to significantly affect the marked galaxy clustering on scales of @xmath14 , however , and if the effects were stronger , then there would have been a discrepancy with the observed unmarked correlation function @xmath15 as well , which was not the case .
thirdly , we also assumed that there are not mark gradients within halos , that is , that the colors of galaxies are independent of their distance from the halo center .
in contrast , hansen et al . ( 2007 ) have shown that the @xmath1 colors of galaxies are @xmath16 redder in the inner regions of groups and clusters compared to the cluster outskirts ( see their figure 10 ) .
van den bosch et al .
( 2008b ) have shown a similar fractional increase in galaxy colors with decreasing halo - centric radius , down to halo masses of @xmath17 .
these color gradients appear to be stronger than luminosity gradients in groups and clusters ( hansen et al .
2005 , martnez & muriel 2006 , weinmann et al .
such a positional dependence of galaxy colors would only affect the color mark correlations while leaving the unmarked correlation function unchanged . in order to model position - dependent color marks in mark clustering statistics ,
the density profile of satellite galaxies must be replaced by a weighted profile .
the calculation is done in fourier space ( see sheth 2005 , and appendix b of ss08 ) , and we replace @xmath18 , which is the fourier transform of @xmath19 , by a weighted number density profile @xmath20 where @xmath21 and @xmath22 quantifies the dependence of galaxy color on halo - centric radius , which we assume to be independent of mass because it is independent of cluster richness ( hansen et al . 2007 ; _
cf_. , van den bosch et al .
the minimum and maximum amount of position - dependence of galaxy colors result in slightly stronger color mark correlations at small scales ( lower and upper dashed curves in lower panel of figure [ mcfexample ] ) . or in other words , color gradients in halos imply a slightly stronger environmental dependence of galaxy color in halo environments ( @xmath23 ) .
note that color gradients in clusters are much stronger at @xmath24 ( loh et al .
2008 ) and would have a stronger effect on color mark clustering at such redshifts . finally , the analytic halo - model description of the color dependence of galaxy clustering uses the mean of the halo occupation distribution and the mean colors of central and satellite galaxies .
we are assuming that at fixed halo mass , the halo occupation distribution and color distributions are approximately independent of the environment an assumption which is justified by some recent results ( blanton & berlind 2007 , van den bosch et al .
2008b , skibba & sheth 2008 ) . to explore the central and satellite galaxy color distributions predicted by the model , we construct mock galaxy catalogs .
the mock catalogs are observationally constrained by the halo occupation distribution as a function of luminosity , determined from galaxy clustering measurements in the sdss ( zheng et al .
2007 ) , and the bimodal color distribution as a function of luminosity in the sdss , fit as the sum of two gaussian distributions , which we refer to as the ` red sequence ' and the ` blue sequence ' . in effect , since central and satellite galaxies have a range of colors at fixed luminosity , in the model satellite colors are drawn from the red sequence with some luminosity dependent probability and are otherwise drawn from the blue sequence ; central galaxy colors are drawn with a similar procedure , but with a different luminosity - dependent probability .
the details of the algorithm are described in skibba et al .
( 2006 ) and ss08 .
ss08 showed that these mock catalogs , which include not only the means of the galaxy colors but their scatter as well , yield color mark correlation functions that are consistent with the analytic model and with sdss measurements .
the analysis in section [ groupcatcompare ] , in which we examine the colors of central and satellite galaxies , will constitute a test of this model , and in particular of the model s two assumptions described at the beginning of this section .
we will compare the model s predictions of central and satellite galaxy colors to those of galaxy group catalogs , described in the following section .
we will first compare our model to the yang et al .
( 2007 ; hereafter y07 ) group catalog , which was constructed by applying the halo - based group finder of yang et al .
( 2005a ) to the new york university value - added galaxy catalog ( nyu - vagc ; blanton et al .
2005b ) , which is based on the sloan digital sky survey ( sdss , york et al .
2000 ) data release 4 ( adelman - mccarthy et al .
2006 ) . the group finder uses halo properties as a function of the total luminosity or stellar mass of groups , and it also identifies ` groups ' with only a single member .
we used only those galaxies with spectroscopic redshifts from the sdss or with redshifts taken from other surveys ( their ` sample ii')that is , we excluded fiber - collided galaxies that were not assigned fibers . including such galaxies does not affect the mean colors of central and satellite galaxies ( shown in section [ groupcatcompare ] ) , although this is not the case for the mean central galaxy luminosities ( see appendix of skibba et al . 2007 ) .
y07 have used mock catalogs to account for the effects of the survey edges when estimating group halo masses ( see their paper for details ) .
we only compare to volume - limited catalogs constructed from the sample ii group catalog in this work , and we have accounted for groups that overlap the redshift limits . in each group , we identify the ` central ' galaxy as the most luminous in the @xmath2 band , and the remaining galaxies brighter than the luminosity threshold are the ` satellites ' . labeling the most massive galaxy in a group , rather than the brightest , as the central one does not affect our results
. we will also compare to the central and satellite galaxy colors of a volume - limited group catalog of berlind et al .
( 2006a ; hereafter b06 ) , which is drawn from the sdss large - scale structure sample ` sample14 ` from the nyu - vagc ; the sample is a subsample of sdss dr4 ( adelman - mccarthy et al .
fiber - collided galaxies are included , and each collided galaxy was given the redshift of its nearest neighbor , which was shown to be an adequate correction at least for groups with ten or more members . to account for the survey edges , b06 excluded all groups whose centers lie less than 500 kpc from an edge in the tangential direction or less than 500 km @xmath25 from an edge in the radial direction .
groups were identified using a halo - based friends - of - friends algorithm , which used halo occupation distribution models and the group multiplicity function .
the group - finding algorithm only used the galaxy positions in redshift - space , as opposed to the algorithm used by y07 . in the following analysis ,
the magnitudes and colors of galaxies in the two group catalogs are petrosian , and have been @xmath26-corrected and evolution corrected to @xmath27 .
the main purpose of this paper is to analyze the colors of central and satellite galaxies in groups and clusters .
we compare predictions of the color mark model of skibba & sheth ( 2008 ) to measurements from the yang et al .
( 2007 ) and berlind et al .
( 2006 ) galaxy group catalogs .
this analysis also constitutes a test of the model , which , as described in section [ model ] , made two assumptions in addition to those necessary to model the luminosity dependence of galaxy clustering : ( i ) that the bimodal color distribution at fixed luminosity is independent of halo mass , and ( ii ) that satellite galaxies tend to follow a particular color - magnitude sequence @xmath28 that approaches the red sequence with increasing luminosity ( equation [ csatseq ] ) .
we test the latter assumption first . in figure [ censatcmd ]
we show a color - magnitude diagram with the model s satellite galaxy color - magnitude sequence .
it is similar to figure 2 in ss08 , but the contours are for a volume - limited catalog constructed from the y07 group catalog , with limits @xmath29 , @xmath30 , consisting of 59,085 galaxies .
the figure shows the color - magnitude contours of central and satellite galaxies in the catalog ( red and blue contours , respectively ) .
central and satellite galaxies appear to have somewhat similar bimodal color distributions at faint luminosities .
the majority of bright blue galaxies and luminous red galaxies are centrals , however , and as we will show later , they tend to reside in less massive and more massive halos , respectively .
we compare the satellite galaxy color - magnitude sequence @xmath28 ( [ csatseq ] ) of the model to the mean satellite colors in the y07 group catalog in figure [ censatcmd ] .
they are in excellent agreement : the model s satellite color sequence is approximately only 0.01 magnitudes redder than that of the group catalog and within the poisson errors across most of the luminosity range .
the agreement between the model and the group catalog is encouraging , and in particular it supports the use of ( [ csatseq ] ) as the satellite galaxy color sequence , which could be used as a constraint for galaxy formation models that include physical processes that quench the star formation of satellites . volume - limited sdss catalog constructed from the y07 group catalog
. contours for central galaxies ( with brightest @xmath2-band luminosity ) and satellite galaxies are red and blue , respectively . the satellite galaxy color - magnitude sequence @xmath28 of the ss08 model ( [ csatseq ] ) is shown as the thick solid line ; the mean satellite galaxy colors at fixed luminosity of the group catalog are shown as the square points , with poisson errors .
the color - magnitude sequence of central galaxies in the model ( [ ccenseq ] ) and the group catalog are also shown , by the dashed line and triangle points , respectively .
the measured central and satellite colors ( squares and triangles ) are slightly offset , for clarity . ]
we also show the color - magnitude sequence of central galaxies in figure [ censatcmd ] . in the model
, this sequence is implied by the satellite sequence ( [ csatseq ] ) , the luminosity - dependent halo occupation distribution , and the observed color - magnitude constraints : @xmath31 . \label{ccenseq}\ ] ] ( see ss08 for details ) .
this sequence is consistent with the measurement from the y07 catalog , at least for @xmath32 . in both the model and the y07 catalog , at fixed luminosity ,
central galaxies tend to be _ bluer _ than satellite galaxies .
van den bosch et al . ( 2008a ) found the same result , at fixed stellar mass .
however , in a given halo , the central galaxy tends to be significantly brighter and more massive than its satellites , and it also tends to be redder , as we will show in section [ ccencsatn ] .
figure [ censatcmd ] showed the mean central and satellite colors as a function of luminosity , @xmath33 and @xmath28 .
we now compare the central and satellite color distributions , @xmath34 and @xmath35 , in figures [ ccenlcompare ] and [ csatlcompare ] .
the model predictions are determined from a mock galaxy catalog ( described in section [ model ] ) , and they are compared to measurements from the y07 group catalog .
color of central galaxies as a function of luminosity .
solid black histograms show the predictions of the model ( skibba & sheth 2008 ) ; dashed blue histograms show the measurements from the yang et al .
( 2007 ) group catalog . ] , but for satellite galaxies .
the brightest luminosity bin is excluded because of poor number statistics . ]
overall , the agreement between the model and the group catalog is quite good , especially for @xmath36 ( or @xmath37 ) . for central galaxies , the bimodality of the color distribution is stronger in the model than in the data at faint luminosities
this is partly due to the fact that the double - gaussian fit to the color distribution at fixed luminosity is not perfect , and slightly underpopulates the ` green valley ' between the red and blue sequences at faint luminosities .
this does not explain the discrepancy in the faintest bin , however .
faint central galaxies tend to reside in low mass halos or small groups ( _ e.g. _ , skibba et al .
2007 ) , so the discrepancy could be due to inaccuracies in the model or its constraints at low mass or to inaccuracies in the group catalog in poor groups .
an interesting result is that both the model and the group catalog agree that the red sequence and blue sequence are peaked at approximately the same colors at fixed luminosities for central and satellite galaxies , and that the widths of the red and blue sequences narrow similarly with increasing luminosity for centrals and satellites .
the difference between the central and satellite color distributions is simply that the blue fraction at a given luminosity is lower for satellites : the red sequence of satellites becomes significantly populated at fainter luminosity than does the red sequence of centrals .
we now investigate the colors of central and satellite galaxies as a function of group richness , which is strongly correlated with halo mass .
we constructed a volume - limited catalog from the y07 group catalog similar to the one above , but with the following limits : @xmath38 and @xmath39 .
we begin by examining the mean colors of centrals and satellites .
the mean color of central galaxies in groups containing @xmath40 galaxies was defined as the mean value @xmath41 of all central galaxies of such groups ( rather than @xmath42 or @xmath43 ) .
the mean satellite color in groups of @xmath40 galaxies was computed similarly ( _ i.e. _ , the mean of satellite @xmath1 ) .
we have also measured the mean colors of central and satellite galaxies in the similar volume - limited catalog of b06 .
this @xmath44 catalog consists of 21301 galaxies in 4119 groups having three or more members . in comparison ,
the corresponding y07 catalog contains 15234 galaxies in 2163 groups ( when fiber - collided galaxies are included ) , which are significantly smaller numbers considering that their catalog was drawn from a later sdss data release .
the berlind et al .
catalog has an overabundance of low-@xmath40 groups ( see their appendix and that of skibba et al . 2007 ) , which does not appear to be the case for yang et al .
however , berlind et al .
only claim to be complete for @xmath45 , and for such richnesses the two group catalogs are in better agreement .
we will compare the mean central and satellite colors from both group catalogs to those predicted by the halo occupation model of skibba & sheth ( 2008 ) .
the model colors are computed the same way as the central and satellite luminosities were computed in skibba et al .
( 2007 ) : @xmath46 and @xmath47 where @xmath48 is the halo mass function , @xmath49 is the halo occupation distribution , and @xmath50 is the group multiplicity function .
figure [ yangberlind ] shows the results . in general , at fixed group richness ,
satellite galaxies tend to be bluer than central galaxies , by up to 0.1 mags .
these trends are similar at fixed halo mass in the model ( not shown ) . because of the dependence of the mean satellite colors on the luminosity threshold ( @xmath51 in eq .
[ csatn ] ) , the difference between the colors of centrals and satellites would be even larger if fainter galaxies were included .
colors as a function of group richness , for @xmath44 .
solid curves show the model predictions ( using equations [ ccenn ] and [ csatn ] ) , and red triangles and blue squares show the measurements from the b06 and y07 galaxy group catalogs , respectively .
vertical error bars are the average of the poisson errors ( estimated from the number of groups in each bin ) and bootstrap errors ( estimated from the variance of 10 times as many pseudo - samples as the number of groups ) .
horizontal error bars show the @xmath40 bin widths . ] the model s central and satellite galaxy colors are in very good agreement with both group catalogs , from poor groups to rich clusters .
the error bars on the measurements are fairly large , but the halo - model predictions are also uncertain , due to uncertainties in the luminosity - dependent hod and in the color - magnitude constraints
. the central galaxy colors are systematically redder in the yang et al .
( 2007 ) catalog than in the berlind et al .
( 2006a ) catalog .
it could be due to the different cosmology assumed by yang et al .
( @xmath52 , @xmath53 ) , or it may be related to the different color - dependent clustering at fixed group mass measured from the catalogs ( berlind et al .
2006b , wang et al . 2007 ) .
it is of only weak statistical significance , however .
for the model , we have assumed that the bimodal color distribution at fixed luminosity is independent of halo mass .
therefore , the satellite colors are only weakly dependent on group richness or halo mass because the satellite luminosities are .
skibba et al .
( 2007 ) found that satellite galaxy luminosity is almost independent of group richness in their halo occupation model and in the group catalogs of yang et al .
( 2005a ) and berlind et al .
( 2006a ) .
the result in the upper panel of figure [ yangberlind ] shows that this is also the case for satellite galaxy colors , and lends justification to the model s assumption that the color distribution at given luminosity is independent of mass .
this result has an interesting consequence : the fact that both the luminosity and color of satellites is nearly independent of halo mass implies that the evolution of satellite galaxies is nearly independent of the environment .
we discuss this further in section [ discuss ] .
finally , we compare the central and satellite color distributions as a function of group richness , @xmath54 and @xmath55 , for @xmath44 , in figures [ ccenncompare ] and [ csatncompare ] .
the model predictions are determined from a mock galaxy catalog ( described in section [ model ] ) , and they are compared to measurements from the yang et al .
( 2007 ) and berlind et al .
( 2006a ) group catalogs .
color of central galaxies as a function of group richness , for @xmath44 .
solid black histograms show the predictions of the model ( skibba & sheth 2008 ) ; dashed blue histograms show the measurements from the yang et al .
( 2007 ) group catalog ; dotted red histograms show the measurements from the berlind et al .
( 2006a ) catalog . a very large @xmath40 bin
is not included , because of poor number statistics . ] , but for satellite galaxies .
the three large @xmath40 bins are chosen to be the same as the @xmath56-richest , @xmath57-richest , and richest bins in figure [ yangberlind ] . ]
overall , there is impressive agreement between the model and group catalogs , even for groups with few members . for central galaxies in poor groups , the b06 catalog has a slightly weaker red peak and a slightly more populated blue bump than in the y07 catalog , with the model between them . for satellite galaxies in poor groups ,
the red sequence is peaked at a slightly redder color in the group catalogs than in the model , and the blue bump is slightly more populated in the catalogs .
these differences are very small , however .
a comparison of the figures makes the following two conclusions evident .
firstly , for central galaxies , while a significant fraction of them populate the blue cloud in small groups or low - mass halos , the vast majority of them in large groups or massive halos ( @xmath58 ) have moved onto the red sequence .
secondly , for satellite galaxies , the color distribution has a significant blue bump even in massive groups and clusters .
while there are large numbers of red satellites in such systems , of course , there are also a significant number of blue satellites as well , presumably tending to be located in the cluster outskirts , implied by the color gradients discussed in section [ model ] .
moreover , the satellite color distribution is almost independent of group richness or halo mass .
this occurs in the model by construction : we have assumed that the color distribution at fixed luminosity is independent of mass , and satellite luminosity only weakly depends on richness .
the agreement with the group catalogs constitutes evidence in support of this assumption .
to summarize , we have compared the halo model of galaxy colors of skibba & sheth ( 2008 ) to the colors of galaxies in the yang et al .
( 2007 ) and berlind et al .
( 2006a ) group catalogs .
the model assumes that satellite galaxies tend to follow a particular sequence along the color - magnitude diagram ( [ csatseq ] ) , such that it approaches the red sequence at bright luminosities , and we have found that this satellite color sequence is in excellent agreement with measurements from the yang et al . group catalog .
this constitutes support for the skibba & sheth model , and for the satellite color sequence itself , which could be used as a constraint for galaxy formation models on the physical processes that quench the star formation of satellite galaxies .
the satellite color sequence as a function of luminosity can easily be converted into a sequence as a function of halo mass as well ( see section 2.2 of skibba & sheth ) .
the agreement between the model and the group catalog suggests that a significant fraction of faint satellites ( with @xmath59 ) , which reside in low - mass halos as well as cluster - sized halos ( skibba et al .
2007 ) , are still forming stars out of a dwindling gas supply and are in the process of being quenched and transformed into ` red and dead ' galaxies .
van den bosch et al .
( 2008a ) recently used the yang et al ( 2007 ) group catalog to explore the impact of various transformation mechanisms that are believed to operate on satellite galaxies .
based on the colors and concentrations of galaxies at fixed stellar mass , they found that the main mechanism that causes the transition of satellite galaxies from the blue to the red sequence is strangulation , in which the hot diffuse gas around newly accreted satellites is stripped , removing its fuel for future star formation .
they ruled out other mechanisms , such as ram - pressure stripping , which is efficient only in massive halos , and harassment , which alters galaxies morphologies .
kang & van den bosch ( 2008 ) used these results with their semi - analytic model to show that strangulation is inefficient and takes a few gyr to operate .
a similar conclusion was recently obtained by font et al .
( 2008 ) . who used a more sophisticated stripping model .
cattaneo et al .
( 2008 ) have recently used another semi - analytic model to show that the observed ` archaeological downsizing ' , in which stars in more massive galaxies tend to form earlier and over a shorter period , can be reproduced if strangulation shuts down star formation only above a critical halo mass @xmath60 .
gilbank & balogh ( 2008 ) came to the same conclusion when attempting to reproduce the red sequence dwarf - to - giant ratio in clusters and the field .
this critical mass is slightly larger than the minimum halo mass for @xmath12 , and at this mass scale the satellite color sequence of the skibba & sheth ( 2008 ) model is indeed approaching the red sequence .
we also analyzed the colors of central and satellite galaxies as a function of group richness , quantified by the number of galaxies in a group more luminous than a given threshold .
we showed that at fixed richness or halo mass , central galaxies tend to be redder than satellites , and the color difference increases with richness .
in contrast , at fixed luminosity or stellar mass , centrals tend to be bluer than satellites .
this is simply explained by the fact that in a given halo , the central galaxy is usually the brightest and most massive galaxy .
these results suggest that as central and satellite galaxies evolve , they may follow different paths along the color - magnitude diagram .
for example , central galaxies become more luminous as they evolve , whether secularly or with mergers , and may continue to form stars and remain blue even after forming a significant bulge component , and then move towards the red sequence after agn feedback has operated ; on the other hand , satellites may experience star formation quenching due to strangulation and approach the red sequence while they are still faint and disk - dominated .
these issues are investigated further in skibba et al .
( 2008a , in prep . ) , using morphology mark correlation functions with the sdss galaxy zoo catalog of visually classified morphologies .
for the colors of both central and satellite galaxies as a function of group richness , the model of skibba & sheth ( 2008 ) and the two group catalogs of yang et al .
( 2007 ) and berlind et al .
( 2006a ) are in very good agreement .
central galaxies tend to be the reddest galaxy in a halo and are much redder than the typical satellite in large groups .
satellite galaxy color , unlike that of centrals , is almost independent of group richness .
the distributions of central and satellite colors as a function of richness in the model and group catalogs are also in good agreement .
central galaxies have a bimodal color distribution in small groups , but the vast majority of them in larger groups have moved onto the red sequence .
in contrast , the satellite color distribution is almost independent of group richness .
this occurs in the model by construction : we assumed that the color distribution at fixed luminosity is independent of halo mass , and satellite luminosity only weakly depends on richness .
the agreement with the group catalogs supports this assumption .
it is worth emphasizing that , while some authors focus on mean galaxy properties ( _ e.g. _ , martnez & muriel 2006 , conroy & wechsler 2008 ) or on red or blue fractions ( _ e.g. _ , weinmann et al .
2006 , van den bosch et al .
2008a ) , our model has predicted the mean and _ distributions _ of central and satellite galaxy colors , as a function of luminosity and richness , in agreement with the sdss data for @xmath32 .
this is quite a feat , considering that our model is fairly simple , based on luminosity - dependent clustering and color - magnitude constraints , with very few assumptions .
it is interesting that satellite galaxy color appears to be almost independent of host halo mass , and that this is also the case for satellite galaxy luminosity ( skibba et al .
2007 ) . since galaxy color is tightly correlated with stellar mass - to - light ratio ( bell et al . 2003 ) , this implies that satellite galaxies of a given stellar mass can also be found in halos of a wide range of masses .
this suggests that what most determines a satellite galaxy s properties , and its evolution in general , is its stellar mass , not its host halo mass .
in addition , galaxy color and luminosity are the primary properties that are most predictive of a galaxy s environment ( blanton et al .
2005a ) , so the flat relations of satellite galaxy color and luminosity with halo mass suggest a dearth of environmental dependence for the transformation of satellite galaxies .
this point was recently made by van den bosch et al .
( 2008b ) , who used the yang et al .
( 2007 ) group catalog to show that the color and concentration of satellite galaxies are almost completely determined by their stellar mass , with only a very weak dependence on halo mass and halo - centric radius .
one consequence of this is that ` pre - processing ' in groups can not be the dominant process that differentiates the cluster galaxy population from that of the field .
finally , brown et al . (
2008 ) recently completed a study of the evolution of the luminosity and stellar mass of red central and satellite galaxies , using halo occupation models .
we can do a few simple comparisons between our results and theirs .
they find that the stellar masses of luminous red central galaxies scales with halo mass to the power of @xmath61 .
using our model s relationship between central galaxy color and halo mass ( lower panel of figure [ yangberlind ] ) with the relation between colors and stellar mass - to - light ratios ( bell et al .
2003 ) , we can estimate the relation between central galaxy stellar mass and halo mass for the model .
we find that the slope of this relation approaches @xmath62 , similar to their result .
brown et al . also show that approximately 50% of @xmath63 mass halos host central galaxies that are red , and this fraction increases with halo mass .
our model predicts a fraction of @xmath64 at this mass , although our separation of ` red ' and ` blue ' galaxies is in terms of @xmath1 color , while theirs is in terms of @xmath65 . finally , brown et al .
also conclude that the fraction of stellar mass within the satellite population increases with host halo mass , which is consistent with our results and skibba et al .
i would like to thank xi kang , frank van den bosch , and ravi sheth for valuable discussions , and i thank frank van den bosch for providing the galaxy group catalog of yang et al ( 2007 ) .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , cambridge university , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpa ) , the max - planck - institute for astrophysics ( mpia ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
adelman - mccarthy j. k. et al . , 2006 ,
apjs , 162 , 38 baldry i. k. , glazebrook k. , brinkmann j. , ivezi . , lupton r. h. , nichol r. c. , szalay a. s. , 2004 , apj , 600 , 681 beisbart c. , kerscher m. , 2000 , apj , 545 , 6 bell e. f. , mcintosh d. h. , katz n. , weinberg m. d. , 2003 , apjs , 149 , 289 berlind a. a. et al . , 2006a , apjs , 167 , 1 berlind a. a. , kazin e. , blanton m. r. , pueblas s. , scoccimarro r. , hogg d. w. , 2006b , astro - ph/0610524 , apj submitted blanton m. r. , eisenstein d. , hogg d. w. , schlegel d. j. , brinkmann j. , 2005a , apj , 629 , 143 blanton m. r. et al . , 2005b , aj , 129 , 2562 blanton m. r. , berlind a. a. , 2007 , apj , 664 , 791 brown m. j. i. , zheng z. , white m. , dey a. , jannuzi b. t. , benson a. j. , brand k. , brodwin m. , croton d. j. , 2008 , astro - ph/0804.2293 cattaneo a. , dekel a. , faber s. m. , guiderdoni b. , 2008 , astro - ph/0801.1673 coil a. et al .
2008 , apj , 672 , 153 conroy c. , wechsler r. h. , 2008 , astro - ph/0805.3346 cooray a. , sheth r. , 2002 , phys .
, 372 , 1 font a. s. , bower r. g. , mccarthy i. g. , benson a. j. , frenk c. s. , helly j. c. , lacey c. g. , baugh c. m. , cole s. , 2008 , astro - ph/0807.0001 gilbank d. g. , balogh m. l. , 2008 , mnras , 385 , l116 gunn j. e. et al . , 1998 , aj , 116 , 3040 gunn j. e. et al . , 2006 , aj , 131 , 2332 hansen s. m. , mckay t. a. , wechsler r. h. , annis j. , sheldon e. s. , kimball a. , 2005 , apj , 633 , 122 hansen s. m. , sheldon e. s. , wechsler r. h. , koester b. p. , 2007
, astro - ph/0710.3780 kang x. & van den bosch f. c. , 2008 , astro - ph/0801.1843 loh y .- s . , ellingson e. , yee h. k. c. , gilbank d. g. , gladders m. d. , barrientos l. f. , 2008 , astro - ph/0802.3726 martnez h. j. , muriel h. , 2006 , mnras , 370 , 1003 navarro j. f. , frenk c. s. , white s. d. m. , 1997 , apj , 490 , 493 sheth r. k. , tormen g. , 1999 , mnras , 308 , 119 sheth r. k. , 2005 , mnras , 364 , 796 skibba r. a. , sheth r. k. , connolly a. j. , scranton r. , 2006 , mnras , 369 , 68 skibba r. a. , sheth r. k. , martino m. c. , 2007 , mnras , 382 , 1940 skibba r. a. , sheth r. k. , 2008 , astro - ph/0805.0310 , mnras , accepted skibba r. a. , bamford s. p. , nichol r. c. , slosar a. et al .
, 2008a , in prep .
skibba r. a. , van den bosch f. c. , yang x. , mo h. j. , 2008b , in prep .
stoyan d. , stoyan h. , 1994 , fractals , random shapes , and point fields .
wiley , chichester van den bosch f. c. , weinmann s. m. , yang x. , mo h. j. , li c. , jing y. p. , 2005 ,
mnras , 361 , 1203 van den bosch f. c. , yang x. , mo h. j. , weinmann s. m. , macci a. v. , more s. , cacciato m. , skibba r. , kang x. , 2007 , mnras , 376 , 841 van den bosch f. c. , aquino d. , yang x. , mo h. j. , pasquali a. , mcintosh d. h. , weinmann s. m. , kang x. , 2008a , mnras , 387 , 79 van den bosch f. c. , pasquali a. , yang x. , mo h. j. , weinmann s. , mcintosh d. h. , aquino d. , 2008b , astro - ph/0805.0002 wang y. , yang x. , mo h. j. , van den bosch f. c. , weinmann s. m. , chu y. , 2007 , astro - ph/0711.4431 weinmann s. m. , van den bosch f. c. , yang x. , mo h. j. , 2006 , mnras , 366 , 2 yang x. , mo h. j. , van den bosch f. c. , 2003 , mnras , 339 , 1057 yang x. , mo h. j. , van den bosch f. c. , jing y. p. , 2005a ,
mnras , 356 , 1293 yang x. , mo h. j. , van den bosch f. c. , weinmann s. m. , li c. , jing y. p. , 2005b ,
mnras , 362 , 711 yang x. , mo h. j. , van den bosch f. c. , pasquali a. , li c. , barden m. , 2007 , apj , 671 , 153 yang x. , mo h. j. , van den bosch f. c. , 2008 , apj , 676 , 248 york d. g. et al . , 2000 , aj , 120 , 1579 zehavi i. et al . , 2005 ,
apj , 630 , 1 zheng z. , coil a. l. , zehavi i. , 2007 , apj , 667 , 760 | current analytic and semi - analytic dark matter halo models distinguish between the central galaxy in a halo and the satellite galaxies in halo substructures .
it is expected that galaxy properties are correlated with host halo mass , and that central galaxies tend to be the most luminous , massive , and reddest galaxies in halos while the satellites around them are fainter and bluer .
using a recent halo - model description of the color dependence of galaxy clustering ( skibba & sheth 2008 ) , we investigate the colors of central and satellite galaxies predicted by the model and compare them to those of two galaxy group catalogs constructed from the sloan digital sky survey ( yang et al .
2007 , berlind et al .
2006a ) . in the model , the environmental dependence of galaxy color is determined by that of halo mass , and the predicted color mark correlations were shown to be consistent with sdss measurements .
the model assumes that satellites tend to follow a color - magnitude sequence that approaches the red sequence at bright luminosities ; the model s success suggests that bright satellites tend to be ` red and dead ' while the star formation in fainter ones is in the process of being quenched . in both the model and the sdss group catalogs
, we find that at fixed luminosity or stellar mass , central galaxies tend to be bluer than satellites .
in contrast , at fixed group richness or halo mass , central galaxies tend to be redder than satellites , and galaxy colors become redder with increasing mass .
we also compare the central and satellite galaxy color distributions , as a function of luminosity and as a function of richness , in the model and in the two group catalogs . except for faint galaxies and small groups , the model and both group catalogs are in very good agreement .
[ firstpage ] methods : analytical - methods : statistical - galaxies : clusters : general - galaxies : formation - galaxies : evolution - galaxies : clustering - galaxies : halos - large scale structure of the universe |
the recent renewal of interest in multiferroic materials led to discovery of many novel compounds where electric polarization is induced by ordered magnetic states with broken inversion symmetry.@xcite the electric polarization in multiferroics is very susceptible to changes in spin ordering produced by an applied magnetic field , which gives rise to dramatic effects such as the magnetically - induced polarization flops and colossal magnetocapacitance.@xcite magnetoelectric interactions also couple spin waves to polar phonon modes and make possible to excite magnons by an oscillating electric field of light , which gives rise to the so - called electromagnon peaks in photoabsorption.@xcite electromagnons were recently observed in two groups of multiferroic orthorombic manganites , @xmath0mno@xmath1 ( @xmath0 = gd , tb , dy , eu@xmath2y@xmath3 ) and @xmath0mn@xmath4o@xmath5 ( @xmath0 = y , tb ) .
@xcite ferroelectricity in @xmath0mno@xmath1 appears in a non - collinear antiferromagnetic state with the cycloidal spiral ordering and the magnetoelectric coupling originates from the so - called inverse dzyaloshinskii - moriya mechanism .
@xcite in ref .
[ ] it was noted that the same mechanism can couple magnons to photons and that an oscillating electric field of light can excite rotations of the spiral plane .
however , the selection rule for the electromagnon polarization resulting from this coupling does not agree with recent experimental data@xcite and , moreover , the inverse dzyaloshinskii - moriya mechanism of relativistic nature is too weak to explain the strength of the electromagnon peaks in @xmath0mno@xmath1 .
these peaks seem to originate from the exchange striction , i.e. ionic shifts induced by changes in the heisenberg exchange energy when spins order or oscillate.@xcite this mechanism explains the experimentally observed polarization of electromagnons .
since the heisenberg exchange interaction is stronger than the dzyaloshinskii - moriya interaction , it can induce larger electric dipoles . in ref .
[ ] it was shown that the magnitude of the spectral weight of the giant electromagnon peak in the spiral state of rare earth manganites is in good agreement with the large spontaneous polarization in the e - type antiferromagnetic state,@xcite which has not been reliably measured yet but is expected to exceed the polarization in the spiral state by 1 - 2 orders of magnitude.@xcite from the fact that the mechanism that couples magnons to light in rare earth manganites is different from the coupling that induces the static polarization in these materials we can conclude that electromagnons can also be observed in non - multiferroic magnets . in this paper we focus on electromagnons in materials exhibiting a linear magnetoelectric effect , i.e. when an applied magnetic field , @xmath6 , induces an electric polarization , @xmath7 , proportional to the field , while an applied electric field , @xmath8 , induces a magnetization , @xmath9 .
this unusual coupling takes place in antiferromagnets where both time reversal and inversion symmetries are spontaneously broken.@xcite it is natural to expect that when an electric field applied to a magnetoelectric material oscillates , the induced magnetization will oscillate too .
such a dynamical magnetoelectric response , however , requires presence of excitations that are coupled both to electric and magnetic fields .
they appear when magnons , which can be excited by an oscillating magnetic field ( antiferromagnetic resonances ) , mix with polar phonons , which are coupled to an electric field .
thus in materials showing a linear magnetoelectric effect , for each electromagnon peak there is an antiferromagnetic resonance with the same frequency .
this reasoning does not apply to all magnetoelectrics and the dc magnetoelectric effect is not necessarily related to hybrid spin - lattice excitations .
as will be discussed below , in materials with collinear spin orders electromagnons either do not exist or have a relatively low spectral weight . in this paper
we argue that electromagnons should be present in non - collinear antiferromagnets showing strong static magnetoelectric response . as a simple example
, we consider a kagom lattice antiferromagnet with the 120@xmath10 spin ordering , shown in fig .
[ fig : model ] .
such an ordering has a nonzero magnetic monopole moment , which allows for a linear magnetoelectric effect with the magnetoelectric tensor @xmath11 for electric and magnetic fields applied in the plane of the kagom lattice.@xcite a relatively strong magnetoelectric response was recently predicted for kagom magnets with the kitpite crystal structure , in which magnetic ions are located inside oxygen bipyramids.@xcite in this structure the oxygen ions mediating the superexchange in basal planes are located outside the up - triangles forming the kagom lattice and inside the down - triangles or vice versa ( see fig .
[ fig : model ] ) , in which case magnetoelectric responses of all triangles add giving rise to a large magnetoelectric constant .
spin state . here ,
@xmath12 and @xmath13 denote , respectively , the nearest - neighbor and next - nearest - neighbor exchange constants , the solid arrows denote spins , while the empty arrows denote the shifts of the ligand ions . ] this paper is organized as follows . in sec .
[ sec : symmetry ] we analyze the symmetry of magnon modes and the magnetoelectric coupling in the kagom lattice magnet with the kitpite structure and show that the dc magnetoelectric effect in this system is related to presence of electromagnon modes .
the common origin of the dc and ac magnetoelectric responses implies existence of relations between static and dynamic properties of magnetoelectric materials , derived in sec .
[ sec : relations ] . in sec .
[ sec : softening ] we discuss softening of ( electro)magnons and the resulting divergence of the coupled magnetoelectric response . in sec .
[ sec : phasediagram ] we discuss the transition from a magnetoelectric to a multiferroic state at a strong spin - lattice coupling and plot the phase diagram of our model system . in section
[ sec : discussion ] we discuss the importance of non - collinearity of spins for dynamic magnetoelectric response and possible electromagnons in known magnetoelectric materials . in sec .
[ sec : conslusions ] we conclude .
the coupling of magnetic excitations in the kagom magnet to electric field , resulting from the heisenberg exchange striction or any other non - relativistic interaction , can be found using the method outlined in ref .
[ ] . to simplify notation , we consider a single up - triangle , which has the same point symmetry as the whole kagom lattice with the @xmath14 spin ordering shown in fig .
[ fig : model ] .
the form of the magnetoelectric coupling is constrained by the @xmath15 and @xmath16 symmetry operations:@xcite @xmath17 \right .
\nonumber \\ & + & \left .
\frac{e_{y}}{\sqrt{6 } } \left [ \left(\mathbf{s}_{1}\cdot\mathbf{s}_{3}\right ) + \left(\mathbf{s}_{2}\cdot\mathbf{s}_{3}\right ) - 2 \left(\mathbf{s}_{1}\cdot\mathbf{s}_{2}\right ) \right ] \right\}. \label{eq : hme1}\end{aligned}\ ] ] we then replace @xmath18 by @xmath19 , where the unit vectors @xmath20 describe the @xmath14 spin ordering in the @xmath21 plane and @xmath22 is the oscillating part , which is the superposition of the orthogonal magnon modes in the triangle ( the zero wave vector magnons for the kagom lattice ) : @xmath23 where @xmath24 labels the magnon , @xmath25 are the out - of - plane components of the magnons and @xmath26 are the in - plane components ( see fig .
[ fig : modes ] ) , @xmath27 the single - magnon excitation by an electric field is described by the terms linear in @xmath28 , while the terms quadratic in @xmath28 give rise to the photoexcitation of a two - magnon continuum . since spins order in plane ,
the polarization oscillations are induced by the in - plane oscillations of spins and the coupling of electric field to magnons , obtained from eq.([eq : hme1 ] ) , has the form : @xmath29 where @xmath30 .
this magnetoelectric coupling is only nonzero in the magnetically ordered state with broken time reversal symmetry , which is why the coupling constant is proportional to @xmath31 .
spin ordering .
the thin arrows indicate the directions of average spins , while the thick arrows show the in - plane components of the magnon modes . ]
we note that eq.([eq : hme2 ] ) can also be obtained by a conventional symmetry analysis of magnetic modes .
the vector @xmath32 and , hence , the corresponding amplitude , @xmath33 , forms a one - dimensional representation @xmath34 , while @xmath35 and @xmath36 form a two - dimensional representation @xmath37 ( see table [ tab : ireps ] ) . the direct product , @xmath38 , where @xmath39 is the symmetry of the spin ordering , transforms as the doublet of the in - plane components of the electric fields , @xmath40 , which leads to eq.([eq : hme2 ] ) .
this general symmetry analysis is , however , insensitive to the microscopic mechanism of the magnetoelectric coupling , whereas the derivation staring from eq.([eq : hme1 ] ) only applies to non - relativistic mechanisms .
.the transformation properties of several irreducible representations of the space group of the kagom lattice and time reversal operation @xmath41 . [ cols="^,^,^,^,^ " , ] table [ tab : ireps ] shows that the coupling to the in - plane components of magnetic field has the form , @xmath42 while the hamiltonian describing magnon modes with zero wave vector in the kagom layer is , @xmath43 if we consider , for example , the microscopic spin hamiltonian describing the antiferromagnetic nearest - neighbor and next - nearest - neighbor heisenberg exchange interactions with the exchange constants , respectively , @xmath12 and @xmath13 and the easy plane magnetic anisotropy @xmath44 , @xmath45 for which the @xmath14 spin ordering shown in fig .
[ fig : model ] is a classical ground state,@xcite we get @xmath46\langle s \rangle^2 , & m^{-1 } = \delta \langle s \rangle^2,\\ \kappa_0 = 0 , & \kappa = 3\left(j_1 + j_2\right ) .
\end{array}\ ] ] the linearized equations of motion for spins in applied electric and magnetic fields are obtained from eqs.([eq : hme2]),([eq : hm ] ) , and ( [ eq : hspin ] ) , if we impose the commutation relations for the amplitudes of the in - plane and out - of - plane parts of @xmath47 : @xmath48 = i \delta_{\alpha,\beta}.\ ] ] from these equations we find the frequencies of the three magnon modes with zero wave vector : @xmath49 and @xmath50 . minimizing the spin energy with respect to @xmath51 and @xmath52 in the static limit
, we obtain an effective magnetoelectric coupling , @xmath53 where the magnetoelectric coefficient @xmath54 .
furthermore , the @xmath55-mode can be excited by both electric and magnetic field oscillating with the frequency @xmath56 in the direction parallel to the @xmath57 axis , while the @xmath58-mode can be excited by both @xmath59 and @xmath60 , which shows that the static linear magnetoelectric effect in this non - collinear magnet is related to the presence of electromagnon and antiferromagnetic resonance peaks with equal frequencies in the optical absorption spectrum .
the common origin of the static and dynamic magnetoelectric response of non - collinear magnets leads to quantitative relations between dc susceptibilities and spectral weights of peaks in the optical absorption spectrum .
these relations simplify when the coupling of magnons to electric field is mediated by a single optical phonon .
the description of magnons in terms of conjugated coordinates and momenta is very convenient for derivation of these relations , since the coupled magnon and optical phonon are in this approach just a pair of coupled oscillators : @xmath61 where @xmath62 and @xmath63 are the coordinates and momenta of the optical phonons coupled to magnons , which also form a two - dimensional representation .
the magnetoelectric response of such a system is easy to calculate .
the result can be expressed in terms of observable quantities , such as the ` dressed ' magnon and phonon frequencies , @xmath64 and @xmath65 , and the spectral weights of the magnon and phonon peaks excited by an electric and magnetic field .
we denote the spectral weight of the electromagnon peak by @xmath66 where @xmath67 is the imaginary part of the dielectric ac susceptibility , while the spectral weight of the antiferromagnetic resonance is , @xmath68 where @xmath69 is the imaginary part of the magnetic ac susceptibility .
the integration in eqs.([eq : smage ] ) and eq.([eq : smagh ] ) goes over an interval of frequencies around @xmath64 .
the two spectral weights for the phonon @xmath70 and @xmath71 are defined in a similar way .
we assume that the magnon and phonon peaks are sufficiently narrow and can be separated from each other .
the four spectral weights satisfy a relation , @xmath72 following from the fact that an electric field only interacts with the ` bare ' phonon , while a magnetic field is only coupled to the ` bare ' magnon .
the relations between the dc and ac magnetoelectric responses of the coupled spin - lattice system have the form , @xmath73 where @xmath74(@xmath75 ) is the increase of the real part of the dielectric constant ( magnetic permeability ) at zero frequency resulting from the magnon and phonon peaks ( we use the gaussian units ) .
the first two equations are the kramers - kronig relations for the real and imaginary parts of dielectric and magnetic susceptibilities , while the last relation follows from equations of motion .
combining eqs.([eq : weights ] ) and ( [ eq : kk ] ) we can express the ratio of the spectral weights of the electromagnon and the antiferromagnetic resonance through the ratio of the static magnetoelectric constant @xmath76 and magnetic susceptibility @xmath77 : @xmath78 for @xmath79 , the ratio of the spectral weights is just the square of the ratio of the dc magnetoelectric constant and magnetic susceptibility . due to the spin - lattice coupling ,
the ` dressed ' magnon frequency , @xmath64 , is lower than its ` bare ' value , @xmath80 ( assuming that the ` bare ' magnon frequency is smaller than the ` bare ' phonon frequency ) . as the spin - lattice coupling increases , @xmath64 vanishes at a critical value of the coupling . according to eq.([eq : kk ] ) , this results in the simultaneous divergency of @xmath81 , @xmath82 and @xmath76 , indicating an instability towards a multiferroic phase , which is both ferroelectric and ferromagnetic .
another manifestation of this instability is the fact that as the spin - lattice coupling approaches the critical value , the magnetoelectric constant @xmath76 tends to its upper bound equal the geometric mean of the dielectric and magnetic susceptibilities , @xmath83 , imposed by the requirement of stability with respect to applied electric and magnetic fields.@xcite
to study the transition from the magnetoelectric state of the kagom magnet to the multiferroic state in more detail , we consider a simple microscopic model , in which positions of magnetic ions are fixed , while ligand ions are allowed to move .
the spin - lattice coupling originates from the dependence of the exchange constants on displacements of ligand ions mediating the superexchange in the direction perpendicular to the straight line connecting two neighboring spins .
we denote the positions of the three ligand ions outside up - triangles by @xmath84 , @xmath85 , and @xmath86 , while the position of a ligand ion inside a down - triangle is denoted by @xmath87 .
then , for example , the exchange constant for the spins @xmath88 and @xmath89 is @xmath90 while for the spins @xmath91 and @xmath92 it is @xmath93 ( see fig .
[ fig : model ] ) .
furthermore , we assume that phonons are dispersionless and the lattice energy for a pair of the up- and down - triangles is , @xmath94 where @xmath95 is the spring constant . for the @xmath14 structure shown in fig .
[ fig : conf](a ) , the magnetoelectric constant @xmath76 in eq.([eq : fme ] ) is given by @xmath96 where @xmath97 is the charge of the oxygen ion , @xmath98 is the average magnetic moment on each site , @xmath99 is the unit cell volume , and @xmath100 is the dimensionless spin - lattice coupling constant .
an estimate for the magnetoelectric constant , @xmath101 , for the model parameters appropriate for the kitpite structure ( @xmath102 , @xmath103mev , @xmath104 , @xmath105 , and @xmath106 ) agrees well with the result of ab initio calculations.@xcite furthermore , @xmath107 , so that @xmath108 .
thus , the electromagnon peak in kitpite should be much stronger than the antiferromagnetic resonance peak , which is also the case for rare earth manganites with a spiral ordering.@xcite at @xmath109 , the magnetoelectric constant diverges and so do the dielectric and magnetic susceptibilities : @xmath110 since in our model there are two polar phonons coupled to a magnon with a given polarization ( one in the up - triangle and another in the down - triangle ) , the magnetoelectric constant comes close to its upper bound but does not reach it at @xmath111 : @xmath112_{g = 1 } \approx 0.985.\ ] ] surprisingly , the softening of the @xmath33 magnon mode , which is not coupled to polar lattice distortions , occurs at a lower value @xmath113 . since @xmath114 ,
the softening in this case means vanishing velocity of the @xmath33-mode .
the velocity vanishes , because away from the @xmath115-point in the magnetic brillouin zone magnons with different symmetry become mixed and the @xmath33-mode is coupled to the electromagnon modes .
as the spin - lattice coupling grows and the electromagnon frequency decreases , the lowest - frequency magnon branch is pushed down , which ultimately reduces the velocity of the @xmath33-mode to zero .
-state with zero wave vector , ( b ) the incommensurate ferroelectric state , and ( c ) the collinear multiferroic state .
the short solid arrows show the spin directions , while the short empty arrows denote the shifts of the ligand ions in these states .
the long solid and empty arrows show the direction of , respectively , the spontaneous magnetization and polarization.,width=264 ] although kitpite , for which @xmath116 , is far away from the instabilities discussed in the previous section , it is interesting to study behavior of magnetoelectric materials when the spin - lattice coupling becomes strong , in particular , in view of the dramatic magnetoelectric effects recently observed in multiferroics . as the magnetoelectric constant becomes large close to the transition between magnetoelectric and multiferroic states , it is important to understand possible scenarios of such a transition . in this section
we present the analytical and numerical study of the phase diagram of the kitpite layer for strong spin - lattice couplings . in particular ,
we show that none of the continuous transitions involving magnon softening , discussed in sec .
[ sec : softening ] , actually takes place , as the strong spin - lattice coupling makes the frustrated kagom magnet unstable towards a first - order magnetoelastic transition that relieves the frustration .
this frustration - driven instability is similar to the one found in spinels , where a collinear ordering of spins appears together with a lattice deformation.@xcite we show that the transformation of a non - collinear magnetoelectric state into a collinear multiferroic state can involve two transitions and an intermediate phase , which is ferroelectric but not ferromagnetic . to understand the origin of magnetoelastic instabilities at strong spin - lattice coupling , we first consider a single up - triangle and integrate out the lattice degrees of freedom , @xmath117 ( @xmath118 ) .
then the total energy of the triangle takes the form of an effective spin hamiltonian with quadratic and bi - quadratic interactions : @xmath119.\ ] ] the bi - quadratic interactions favor collinear spins and for @xmath120 , the lowest - energy spin configuration is a collinear state of the @xmath121 type ( the spins lie in the lattice plane ) .
similarly , an effective spin hamiltonian for a down - triangle , where the exchange along all bonds is mediated by a single ligand ion located inside the triangle , has the form , @xmath122 \nonumber \\ & ~ & + \frac{\left(j'_{1}\right)^2}{2k } \sum_{i \neq j \neq k } \left(\mathbf{s}_i \cdot \mathbf{s}_j\right ) \left(\mathbf{s}_j \cdot \mathbf{s}_k\right).\end{aligned}\ ] ] in this case the critical coupling is lower : @xmath123 . due to the inequivalence of the up- and down - triangles in the kitpite structure the transition from the @xmath124-state shown in fig .
[ fig : conf]a to a fully collinear state shown in fig .
[ fig : conf]c goes in two steps via an intermediate state , where only the spins in the down - triangles are collinear while the spins in the up - triangles are still non - collinear ( see fig . [
fig : conf]b ) . .
for relatively small @xmath125 the magnetoelectric ( me ) phase with the 120@xmath10 spin ordering [ see fig .
[ fig : conf](a ) ] undergoes a first - order transition into the incommensurate ( ic ) spin state , which is ferroelectric ( fe ) [ see fig . [
fig : conf](b ) ] , as the coupling constant @xmath126 increases .
further increase of @xmath127 results in the transition into the collinear multiferroic ( mf ) phase [ see fig . [
fig : conf](c ) ] , with the parallel spontaneous polarization and magnetization , @xmath128 . for larger @xmath125 , the me state undergoes a direct transition into the fully collinear mf state .
also plotted are the dot - dash line , where the @xmath33-mode in the me state softens and the dotted line , at which the magnetoelectric response of the me state diverges ( both phenomena do not occur because of the intervening first - order transitions to the fe and mf states).,width=302 ] the transition to the intermediate state does not fully lift the frustration : the number of combinations of the down - triangles with collinear spins and up - triangles with the @xmath14-angle between spins grows exponentially with the system size . in our model
this degeneracy is removed by the next - nearest - neighbor interactions between spins , which select the state shown in fig .
[ fig : conf]b .
the next - nearest - neighbor interactions in the vertical direction induce a small spin canting , as a result of which spins in down - triangles are not strictly collinear and the angles between spins in up - triangles deviate from @xmath14 by the angle @xmath129 .
furthermore , the nearest - neighbor interactions in the remaining two directions are frustrated in the commensurate spin state with the wave vector @xmath130 , where @xmath131 ( @xmath132 being the lattice constant ) .
this frustration is lifted , when the spin ordering becomes incommensurate with the lattice : @xmath133 , where @xmath134 .
the incommensurate state has zero net magnetization , but its electric polarization is nonzero . for the state shown in fig . [
fig : conf](b ) the polarization vector is parallel to the @xmath135 axis .
this polarization originates not from the inverse dzyaloshinskii - moriya mechanism , which makes incommensurate spiral states in e.g. @xmath0mno@xmath1 ferroelectric , but from the fact that all bonds connecting parallel spins are parallel to each other , which forces the ligand ions in all down - triangles to shift in the same direction , since @xmath136 where @xmath137 and the labeling of spins is the same as in fig .
[ fig : model ] .
for the state shown in fig .
[ fig : conf](b ) , where bonds connecting ( nearly ) parallel spins are oriented along the @xmath57 axis , @xmath138 , while @xmath139 and hence the polarization @xmath140 is nonzero .
we note that the coupling of exchange interactions to strains , which gives rise to magnetoelastic transitions in frustrated spinels , @xcite also favors the ferroelectric state .
the absence of inversion symmetry in the kitpite layer allows for the piezoelectric coupling , @xmath141 where @xmath142 is the strain tensor , so that the ligand displacement , @xmath87 , the electric polarization , @xmath7 , and the strains are coupled to each other .
the zero - temperature phase diagram of the kagom layer with @xmath127 and @xmath125 along the horizontal and vertical axes is shown in fig . [
fig : pd ] . for small @xmath125 ,
the incommensurate ferroelectric ( ic fe ) state , discussed above , intervenes between the magnetoelectric ( me ) state with the 120@xmath10 spin ordering [ see fig . [
fig : conf](a ) ] and the fully collinear multiferroic ( mf ) state shown in fig .
[ fig : conf](c ) , in which the spontaneous polarization and magnetization are parallel to each other , @xmath143 . as the ratio @xmath125 grows , the interval of the coupling constant @xmath127 where the intermediate state is stabilized shrinks and above the tricritical point the me state undergoes a direct transition into the collinear mf state along the critical line @xmath144 .
also plotted are the dot - dash line , at which the @xmath33-mode would soften and the dotted line , at which @xmath76 , @xmath145 , and @xmath146 would diverge , if the me state would survive at strong spin - lattice couplings .
we showed that the static magnetoelectric response of non - collinear antiferromagnets can be related to hybrid magnon - phonon modes coupled to both electric and magnetic fields .
such magnetoelectric materials are analogs of displacive ferroelectrics the dielectric response of which is governed by optical phonon modes .
if spins in an ordered state are collinear , the exchange striction can not couple an electric field to a single magnon , as the expansion of scalar products of parallel or antiparallel spins begins with terms of second order in @xmath28 , which give rise to photoexcitation of a two - magnon continuum ( the so - called charged magnons"@xcite ) .
electromagnons in collinear magnets can still originate from mechanisms involving relativistic effects , such as the exchange striction induced by the antisymmetric dzyaloshinskii - moriya interaction , which is proportional to the vector product of two spins . in @xmath147 transition metal compounds
such couplings are weak compared to the exchange striction driven by the heisenberg exchange , so that the spectral weight of electromagnons in collinear magnets should be relatively low .
magnetoelectric materials with collinear spin orders may rather be analogs of ` order - disorder ' ferroelectrics with the static magnetoelectric response originating from thermal spin fluctuations .
cr@xmath4o@xmath1 seems to be an example of such a material : its magnetoelectric coefficient passes through a maximum below nel temperature and then strongly decreases when temperature goes to zero and spin fluctuations become suppressed.@xcite we note that the rotationally invariant coupling eq.([eq : hme1 ] ) may also originate from purely electronic mechanisms , such as the polarization of electronic orbitals induced by a magnetic ordering .
@xcite ab initio calculations suggest that in rare earth manganites the electronic mechanisms of magnetoelectric coupling are as important as the exchange striction.@xcite on the other hand , the increase of the spectral weight of the electromagnon peaks in @xmath0mno@xmath1 below the spiral ordering temperature occurs largely at the expense of the strength of the optical phonon peak at @xmath148@xmath149 , suggesting the dominant role of the spin - lattice coupling .
@xcite if electronic mechanisms dominate and an electromagnon gets its spectral weight from frequencies much higher than those of optical phonons , eq.([eq : kk ] ) should to be modified in an obvious way , while eq.([eq : emagafmr ] ) , where @xmath150 should be replaced by @xmath151 , is still valid .
we note that the non - collinearity of spins by itself does not guarantee strong magnetoelectric effect and electromagnon peaks the crystal structure is equally important .
thus , in the layered kagom antiferromagnet , the iron jarosite kfe@xmath1(oh)@xmath152(so@xmath153)@xmath4,@xcite which has the spin ordering shown in fig .
[ fig : model ] , the ligand ions are located outside of both up- and down - triangles , which cancels the magnetoelectric effect due to the heisenberg exchange striction .
the cancellation also occurs in triangular magnets with the @xmath14 spin ordering , as they contain three different spin triangles , such that spins in one triangle are rotated by @xmath154 with respect to spins in two other triangles@xcite ( more generally , the linear magnetoelectric effect can only be induced by a spin ordering with zero wave vector ) .
we note , however , that the lattice trimerization in hexagonal manganites@xcite makes the three types of spin triangles inequivalent and destroys the cancellation .
this can be also seen from the symmetry properties of the a@xmath155 and b@xmath155 phases of hexagonal manganites@xcite allowing for the magnetoelectric term @xmath156 in the a@xmath157-phase , which has a toroidal moment , and the term @xmath158 in the a@xmath4-phase , which has a magnetic monopole moment .
whether electromagnons in these phases can be observed , depends on the magnitude of the trimerization and remains to be explored.@xcite
in conclusion , we showed that magnets with non - collinear spin orders resulting in a linear magnetoelectric effect may also show electromagnon peaks in optical absorption spectrum .
while electromagnons should be present in many non - collinear magnets , the specific feature of magnetoelectric materials is that some magnon modes can be excited by both electric and magnetic fields , i.e. electromagnons are also antiferromagnetic resonances .
we derived a simple relation eq.([eq : emagafmr ] ) between the ratio of the spectral weights of the electromagnon and antiferromagnetic resonance peaks and the ratio of the static magnetoelectric constant and magnetic susceptibility , which can be used to estimate the strength of electromagnon peaks on the basis of dc measurements . to make our consideration more specific
, we considered a kagom lattice magnet with the kitpite structure , where the ligand ions are positioned in a way that gives rise to a relatively strong linear magnetoelectric effect.@xcite using the symmetry analysis we identified the magnon modes that are coupled to both electric and magnetic fields and give rise to the linear magnetoelectric effect .
we showed that the softening of these modes at a strong spin - lattice coupling results in the divergence of the magnetoelectric constant as well as of magnetic and dielectric susceptibilities , signaling the instability of the magnetoelectric state towards a multiferroic state with spontaneously generated @xmath7 and @xmath9 .
however , the detailed study of the phase diagram of this model revealed that the electromagnon softening does not actually take place , since the first - order transition to the collinear multiferroic state occurs at a lower value of the spin - lattice coupling .
in some region of model parameters a ferroelectric incommensurate - spiral phase intervenes between the magnetoelectric and multiferroic phases .
while in known spiral magnets , ferroelectricity is likely induced by the inverse dzyaloshinskii - moriya mechanism , in our model it results from the stronger exchange striction mechanism due to the collinearity of spins in half of the triangles .
these magnetoelastic instabilities are typical for frustrated magnets , where non - collinear spin orders usually occur .
we also discussed the possibility to observe electromagnons in known magnetoelectric materials .
we hope that our study will stimulate experimental work in this direction .
a review : s .- w . cheong and m. mostovoy ,
nature mater .
* 6 * , 13 ( 2007 ) .
a review : t. kimura , annu . rev . mater .
37 * , 387 ( 2007 ) .
a review : r. ramesh and n. spaldin , nature mater .
* 6 * , 21 ( 2007 ) .
t. kimura , t. goto , h. shintani , k. ishizaka , t. arima , and y. tokura , nature ( london ) * 426 * , 55 ( 2003 ) .
n. hur , s. park , p.a .
sharma , j.s .
ahn , s. guha , and s - w .
cheong , nature ( london ) * 429 * , 392 ( 2004 ) .
t. goto , t. kimura , g. lawes , a.p .
ramirez , and y. tokura , phys .
lett . * 92 * , 257201 ( 2004 ) .
g. smolenskii and i. chupis , sov .
usp . * 25 * , 475 ( 1982 ) .
a. pimenov , a.a .
mukhin , v.yu .
ivanov , v.d .
travkin , a.m. balbashov and a. loidl , nature physics * 2 * , 97 ( 2006 ) .
a. pimenov , t. rudolf , f. mayr , a. loidl , a.a .
mukhin , and a.m. balbashov , phys .
b * 74 * , 100403(r ) ( 2006 ) .
r. valds aguilar , a.b .
sushkov , c.l .
zhang , y.j .
choi , s .- w .
cheong , and h.d
. drew , phys .
b * 76 * , 060404(r ) ( 2007 ) .
sushkov , r.v .
aguilar , s .
park , s .- w .
cheong , and h.d . drew , phys .
. lett . * 98 * , 027202 ( 2007 ) .
h. katsura , n. nagaosa , and a.v .
balatsky , phys .
* 95 * , 057205 ( 2005 ) .
m. kenzelmann , a.b .
harris , s. jonas , c. broholm , j. schefer , s.b .
kim , c.l .
zhang , s .- w .
cheong , o.p .
vajk , and j.w .
lynn , phys .
lett . * 95 * , 087206 ( 2005 ) .
i.a . sergienko and e. dagotto , phys .
b * 73 * , 094434 ( 2006 ) .
m. mostovoy , phys .
96 * , 067601 ( 2006 ) .
a. malashevich and d. vanderbilt , phys .
* 101 * , 037210 ( 2008 ) .
h. katsura , a. v. balatsky , and n. nagaosa , phys .
lett . * 98 * , 027203 ( 2007 ) .
n. kida , y. ikebe , y. takahashi , j.p .
he , y. kaneko , y. yamasaki , r. shimano , t. arima , n. nagaosa , and y. tokura , phys .
b * 78 * , 104414 ( 2008 ) .
sushkov , m. mostovoy , r. valds aguilar , s .- w .
cheong and h. d. drew , j. phys . :
condens . matter * 20*,434210 ( 2008 ) .
y. takahashi , n. kida , y. yamasaki , j. fujioka , t. arima , r. shimano , s. miyahara , m. mochizuki , n. furukawa , and y. tokura , phys .
* 101 * , 187201 ( 2008 ) .
r. valds aguilar , m. mostovoy , a.b .
sushkov , c.l .
zhang , y.j .
choi , s - w .
cheong , and h.d
. drew , phys .
lett . * 102 * , 047203 ( 2009 ) .
b. lorenz , y .- q .
wang , and c .- w .
chu , phys .
b * 76 * , 104405 ( 2007 ) .
sergienko , c. sen , and e. dagotto , phys .
* 97 * , 227204 ( 2006 ) .
s. picozzi , k. yamauchi , b. sanyal , i.a .
sergienko , and e. dagotto , phys .
99 * , 227201 ( 2007 ) .
landau and e.m .
lifshitz , _ electrodynamics of contiuous media _ ( pergamon , oxford , 1984 ) . for a recent review
see , m. fiebig , j. appl .
d * 38 * , r123 ( 2005 ) .
spaldin , m. fiebig , and m. mostovoy , j. phys . : condens .
matter * 20 * , 434203 ( 2008 ) .
delaney , m. mostovoy , and n.a .
spaldin , phys .
* 102 * , 157203 ( 2009 ) .
bulaevskii , c.d .
batista , m.v .
mostovoy , and d.i .
khomskii , phys .
b * 78 * , 024402 ( 2008 ) .
a. b. harris , c. kallin and a. j. berlinsky , phys .
b * 45 * , 2899 ( 1992 ) .
m. elhajal , b. canals and c. lacroix , phys .
b * 66*,014422 ( 2002 ) .
brown , jr .
hornreich , and s. shtrikman , phys . rev . *
168 * , 574 ( 1968 ) .
lee , c. broholm , t. h. kim , w. ratcliff , and s .- w .
cheong , phys .
* 84 * , 3718 ( 2000 ) .
o. tchernyshyov , r. moessner , and s.l .
sondhi , phys .
88 * , 067203 ( 2002 ) .
o. tchernyshyov , r. moessner , and s. l. sondhi , phys .
b * 66 * , 064403 ( 2002 ) .
k. penc , n. shannon , and h. shiba , phys .
* 93 * , 197203 ( 2004 ) .
a. damascelli , d. van der marel , m. grninger , c. presura , t.t.m .
palstra , j. jegoudez and a. revcolevschi , phys .
lett * 81 * , 918 ( 1998 ) .
g. rado , phys . rev . * 128 * , 2546 ( 1962 ) . h. yatom and r. englman , phys . rev . *
188 * , 793 ( 1969 ) .
s. miyahara and n. furukawa , preprint arxive:0811.4082 .
d. grohol , k. matan , j .- h .
cho , s .- h .
lee , j.w .
lynn , d.g .
nocera , and y.s .
lee , nature mater . * 4 * , 323 ( 2005 ) .
see e.g. b.b .
van aken , t.t.m .
palstra , a. filippetti , and n.a .
spaldin , nature mater . * 3 * , 164 ( 2004 ) .
the phases of hexagonal manganites are described e.g. in m. fiebig , th .
lottermoser , and r. v. pisarev , j. appl .
93 * , 8194 ( 2003 ) .
a large trimerization below nel temperature was reported in s. lee , a. pirogov , m. kang , k .- h .
jang , m. yonemura , t. kamiyama , s .- w .
cheong , f. gozzo , n. shin , h. kimura , y. noda , and j .-
park , nature * 451 * , 805 ( 2207 ) . | we show that strong electromagnon peaks can be found in absorption spectra of non - collinear magnets exhibiting a linear magnetoelectric effect .
the frequencies of these peaks coincide with the frequencies of antiferromagnetic resonances and the ratio of the spectral weights of the electromagnon and antiferromagnetic resonance is related to the ratio of the static magnetoelectric constant and magnetic susceptibility . using a kagom lattice antiferromagnet as an example
, we show that frustration of spin ordering gives rise to magnetoelastic instabilities at strong spin - lattice coupling , which transform a non - collinear magnetoelectric spin state into a collinear multiferroic state with a spontaneous electric polarization and magnetization .
the kagom lattice antiferromagnet also shows a ferroelectric incommensurate - spiral phase , where polarization is induced by the exchange striction mechanism . |
the problem of particle production in electric field on de sitter geometry received a lot of attention in the literature @xcite . this problem can be studied using nonperturbative methods @xcite and the main results are related to the number density of produced particles .
more recent studies suggest that this effect could also appear as a perturbative phenomenon @xcite .
this is because in de sitter qed these processes are no longer forbidden by the simultaneous energy - momentum conservation as in flat space case .
the cosmological mechanism of particle creation based on nonperturbative methods use the approximation of the out field at late times to obtain the density number of produced particles . in the case of scalar pair production
it was shown that the results are valid only when the mass of the particle is larger than the expansion parameter @xmath0 @xcite , and this corresponds to weak gravitational fields .
the perturbative approach to the problem of particle production in de sitter geometry , use the exact solutions of the free field equations for computing the transition probabilities .
the results obtained so far using this approach show that these probabilities are nonvanishing only when the expansion factor is comparable ( or larger ) with the mass of the scalar particle , that corresponds to strong gravitational fields . in the limit @xmath0 , these probabilities are very small or even vanish since this limit is close to the flat space case .
so it is clear from the start that a direct comparation between this two different mechanisms of particle production needs to be considered with care . in this paper
we use the result obtained in @xcite , for computing the total probability for the process of scalar pair production in external field .
our results prove that the total probability is nonvanishing only when @xmath1 and vanish in the minkowski limit .
further we will make a series of observations related to a recent work @xcite , that compare the perturbative and nonperturbative methods for obtaining particle production in external field .
our analysis show that the results obtained in @xcite , can not be used for comparing the density number obtained in @xcite with the existing nonperturbative results .
the letter is organized as follows : in the second section we present the results obtained for pair production in electric field using the perturbative approach . in the third section
the total probability is computed and in section four we make a graphical analysis of the total probability
. section five is dedicated to the comparation between perturbative and nonperturbative approaches to the problem of particle production on de sitter geometry . in section six
we present our conclusions and in appendix we give the mathematical formulas that help us to establish the final results for total probability .
in this section we want to summarize the results obtained in ref.@xcite , where the production of the scalar particles in electric field on the de sitter expanding universe was studied . in ref.@xcite the calculations
were performed using a perturbative approach .
+ in order to obtain the expression of the transition amplitude we define the @xmath2 and @xmath3 fields as being the exact solutions of the klein - gordon equation in momentum basis @xcite : @xmath4 where @xmath5 is the hankel function of first kind , @xmath6 .
in addition we note : @xmath7 with @xmath8 .
we recall that in @xcite the external electric field @xmath9 was produced by the vector potential @xmath10 : @xmath11 where the contravariant component of the four vector potential was used , since the scalar qed is constructed with vectors .
the final result for the transition amplitude obtained in ref.@xcite , is expressed in terms of unit step function @xmath12 and gamma euler function @xmath13 : @xmath14,\end{aligned}\ ] ] where the function @xmath15 was defined as @xcite : @xmath16.\end{aligned}\ ] ] we must mention that @xmath17 is obtained when @xmath18 .
the result obtained in @xcite show that the momentum conservation law is broken in this process .
this is a direct consequence of the fact that the vector potential @xmath19 is a function of the spatial coordinate @xmath20 , resulting that the spatial integral gives a result which is not dependent of the delta dirac function .
the presence of the external field ( [ c1 ] ) in de sitter geometry leads to the breaking of the momentum conservation law as was shown in @xcite .
in this section we will present the main steps for calculating the total probability of scalar pair production in electric field on de sitter space . for presenting our arguments
we restrict only to the case when @xmath21 with the observation that we consider the ratio of the momenta close but not equal to unity such that @xmath22 . using equations
( [ p ] ) and ( [ ln ] ) from appendix , the @xmath23 functions that define the probability can be brought in this case to the form : @xmath24 the total probability is obtained by integrating after the final momenta @xmath25 the probability density .
since the particles are emitted in pairs we will study the situation when the momenta @xmath26 are emitted on the same direction with the direction of the electric field , which is given by the unit vector @xmath27 .
the total probability will also be computed in the case when the particles will be emitted in the directions which do not coincide with the direction of the electric field .
the nominator can be expressed in terms of the angle between the momenta vectors @xmath28 : @xmath29 using the equation ( [ re1 ] ) , the expression for probability density becomes ( @xmath21 ) : @xmath30 where @xmath31 are the angles between momenta vectors and the vector of electric field .
then the total probability can be obtained by integrating after the final momenta : @xmath32 we analyse only the case @xmath21 since for @xmath33 the calculations are similar .
the integration after momenta modulus is taken such that @xmath34 while @xmath35 , where @xmath36 is the particle momentum in the presence of the electric field @xmath37 .
this cutoff of the momentum is also used in @xcite for avoiding divergent results for the number of particles .
here we adopt the same cutoff since we study the case when @xmath21 .
a few important observations need to be made now . in the case of a perturbative calculation
the probability is a function dependent of the particle momenta ( see eq.([po ] ) ) and in this case one may also apply other regularization methods for obtaining the total probability . in the existing nonperturbative results this is no longer valid since the approximations made for the out state give number densities which are not dependent on the particle momenta and the only remaining choice is to make a cutoff for the upper limits of the momenta @xcite .
first we analyse the case when the momenta of the pair have the same orientation @xmath38 and are emitted on the direction of the electric field @xmath39 .
in other words we will study the total probability when the particles are emitted on given directions and for that reason it will be sufficiently to solve only the momenta modulus integrals .
using now equations ( [ fa ] ) and integrating after the final momenta @xmath25 the total probability in the case @xmath40 gives : @xmath41 further we will compute the total probability in two distinct cases when the particles are emitted in other directions which make various angles with the direction of the electric field given by @xmath27 .
the final results in these cases turns to be : @xmath42 the total probability depends on the parameter @xmath43 , and from our analytical expression one can observe that this function is highly convergent for large @xmath44 , since is proportional with a factor @xmath45 . from the above equation
we observe that the processes in which the particles are emitted on the direction of the electric field are favoured since in this case the probability is larger ( see eqs.([pt1]),([pt2 ] ) ) .
this result proves that there is a net distinction between the situations when the particles are emitted on the direction of electric field or when the particles are emitted in other directions . in @xcite
the density number obtained using nonperturbative method depend on the orientation of the particle momenta relatively to the field directions .
the same conclusion can be reached if we use perturbative methods as can be seen from our exact perturbative results obtained in eqs.([pt1])-([pt2 ] ) .
our results do not sustain the conclusion reached in @xcite about the perturbative result and prove that the probability depend on the direction and momenta orientation of the particles , relatively to the electric field direction .
moreover in the perturbative case one can study the interesting situation when the particles are emitted on the field direction but their momenta have opposite orientation .
this situation is not analysed in @xcite because there is no analytical expression for transition probability since the calculations were done numerically . a closer look at the perturbative result obtained in @xcite leave some questions .
let us recall the amplitude result , using the external field given in @xcite , @xmath46 , where @xmath47 is the conformal time and @xmath48 is the intensity of the electric field .
then the amplitude obtained in @xcite without solving the temporal integrals is proportional with : @xmath49 where the variable of integration is @xmath50 . from the delta dirac term @xmath51 it is immediate that @xmath52 that means that the momenta are equal in modulus @xmath53 .
the problem now is related to the temporal integral which contain in general the result when @xmath54 and a direct computation of this integral prove that the result can be expressed in terms of gauss hypergeometric functions that have the algebraic argument @xmath55 , which can not be equal with @xmath56 , since in these conditions these functions become divergent . by solving numerically the integral in the case @xmath53
one needs to cut the upper limit of the temporal integral .
this means that the upper limit of the variable @xmath57 is finite , which also lead to questions regarding the justification for this cutoff .
so an analytical study of these integrals is required for a correct and well justified result .
another observation about to the result obtained in @xcite is related to the justification of the result in two dimensions .
the interaction between photons and charged scalar particles it is known in physics as the scalar quantum electrodynamics .
this theory is constructed in four dimensions and for that we have a good reason : there is no electromagnetic field in one spatial dimension since the photons have polarization @xmath58 , which is perpendicular on the photon momenta @xmath59 @xcite(the field is transversal ) .
this is why it is impossible to calculate perturbative amplitudes in two dimensions ( one spatial dimension and one temporal dimension ) when one study interactions intermediate by photons , or to reduce the result obtained in four dimensions to two dimensions .
in one spatial dimension one could not construct the theory of electromagnetic field and further the reduction formalism and interaction theory .
all the qed models in two dimensions have to be considered with care since the they work with massless fields and the interaction lagrangean will contain additional terms , which will change the definition of the transition amplitudes .
all these aspects should be clear from the minkowski qed @xcite and the same observations apply for de sitter qed @xcite .
so the amplitude from @xcite computed in two dimensions should be considered with care since as we explained above there is no perturbative qed in two dimensions .
in fact the perturbative methods @xcite seems to be very good in describing what happens at large expansion @xmath60 , with the observation that in this case the nonperturbative approach becomes problematic in the case of a scalar field as was shown in @xcite .
further we plot the total probability as function of parameter @xmath43 , giving different numerical values to the parameter @xmath61 .
we present in our plots the situation when the particles are emitted on the direction of electric field and the situation when the pair is emitted in other directions which do not coincide with the direction of the electric field .
, @xmath62 .
the solid line is for @xmath40 , the dashed line for @xmath63 and the point line for @xmath64 , @xmath65 .
the solid line is for @xmath40 , the dashed line for @xmath63 and the point line for @xmath64 , @xmath66 .
the solid line is for @xmath40 , the dashed line for @xmath63 and the point line for @xmath64 , @xmath67 .
the solid line is for @xmath40 , the dashed line for @xmath63 and the point line for @xmath64 from figs.([fg1])-([fg4 ] ) , we observe that the total probability drops rapidly to zero as the parameter @xmath68 and this behaviour is also preserved when we increase the intensity of the electric field ( see figs.[fg3]-[fg4 ] ) . indeed the probability becomes larger as we increase the parameter @xmath61 , but have nonvanishing values only when @xmath69 .
the conclusion is that the perturbative result is nonvanishing only when the electric field is coupled with a strong gravitational field and vanish for @xmath0 .
this observation is important since the nonperturbative results for computing the beta bogoliubov coefficient are obtained approximating the klein - gordon equation for @xmath70 .
it is well known that the schwinger effect appear in the minkowski theory as a nonperturbative phenomenon and in de sitter space applying the same nonperturbative methods this effect could also be obtained @xcite .
returning now to the de sitter qed it was proved that any external field coupled with the strong gravity of the early universe will give nonvanishing probabilities for the qed processes in the first order of perturbation theory @xcite , which clearly must reduce to zero in the minkowski limit as the laws of energy - momentum conservation ask .
so as long as one use the perturbative methods to investigate pair production in electric fields on de sitter geometry , we do not expect to recover the schwinger effect in the minkowski limit . in @xcite ,
the graphycal results show ( using a numerically evaluation of the integral ) the number of particles in terms of @xmath71 for given values of @xmath72 , without taking into account the above observations and compare the results of the perturbative method and nonperturbative method , for the same values of the parameter @xmath72 . in our opinion
this is is not the right think to do since the two methods give good results for completely different rates of space expansion .
this is because the bogoliubov coefficients are computed approximating the solutions of the free field equation in the limit @xmath0 @xcite .
another problem worth to be mentioned is that the nonperturbative calculations are performed in two dimensions , since in four dimensions there are serious technical problems for completing the calculations . for finding some similarities between the results from @xcite and the result obtained in @xcite
one needs to take the potential given in eq.([c1 ] ) and solve the klein - gordon equation in four dimensions in the case @xmath0 , which is not at all an easy task .
then approximate the out state such that the bogoliubov coefficients could be determined .
the last step requires to obtain an analytical formula for the perturbative amplitude at @xmath0 , which is very small but not zero . in the perturbative case
this result can be obtained and the probability is proportional in this case with a factor @xmath73 . even so the results will be different since the methods are of completely different nature . to be more clear
let us consider the pair production amplitudes calculated with the field produced by a point charge ( or a distribution of point charges ) in minkowski qed , which are zero .
in contrast if we study the same problem in de sitter qed @xcite , the result is nonvanishing only at large expansion .
so we study the problem of pair production in weak electric fields coupled with the strong gravity ( large expansion @xmath74 ) of the early universe .
this problem have nothing in common with the schwinger effect which is obtained using a nonperturbative approach in minkowski space @xcite .
this result can be obtained using the same nonperturbative approach in de sitter space @xcite , when @xmath0 .
our results refers to the case when the electric fields are weak and gravity becomes strong and in the end we recover the minkowski limit which is zero as we expect . in the nonperturbative approach
the situation is inverse since the electric fields are strong and the equations are studied at @xmath0 , which clearly means weak gravity .
finally we conclude that the result obtained in @xcite do not take into consideration the above observations and in addition there are problems in considering qed calculations in two dimensions as we mentioned in the previous section .
in this section we will make a series of observations related to the perturbative and nonperturbative approach to the problem of pair production in external field on de sitter geometry .
first we want to comment some of the affirmations made in @xcite , about the result obtained in @xcite . in the introduction of @xcite
it was suggested that in ref.@xcite , was obtained the number of particles using a perturbative approach . in @xcite
only the probability density was computed and further integration after the final momenta needs to be done for obtaining the particle number .
in fact in the perturbative case the problem of defining the particle number is not well studied .
another observation is that in @xcite only the nonperturbative results about the problem of pair production are cited giving the impression that only ref.@xcite use perturbative methods .
in fact there is an extended literature that use perturbative methods @xcite , and it is only fair to mention these papers .
the problem of particle production due to the fields interactions was also studied in @xcite . in @xcite
it was proven that this kind of particle production could become important and must be taken into account along side with the cosmological particle production . in the case when one use external fields as those given in @xcite the spatial integrals and temporal integrals give completely different results .
these results show that the momentum is no longer conserved for pair production in external fields , which have a dependence of spatial coordinates .
moreover the perturbative probability depends on the ratio @xmath72 and on the particle momenta @xmath75 as one can see from equation ( [ po ] ) .
when the total probability is computed the integrand will depend on the momenta and in the case of divergent integrals one can apply regularization methods . in the nonperturbative case the approximations done for the out modes have as consequence the lose of the momenta dependence in the beta @xmath76 bogolibov coefficient .
then an integration over the final momenta will give a quantity which is at least linearly divergent depending on the number of dimensions used in the problem .
for that reason a cutoff of the upper limits of the momenta integration need to be done as we mentioned in the previous section . even so one can justify these cutoffs and important results
were obtained in @xcite .
but in the present paper we do not want to criticise the nonperturbative method which gives important results as we mentioned above .
instead we want to point out that both approaches to the problem of particle production must be used in order to obtain a clear picture about the phenomenon of particle production in gravitational fields .
one can not neglect the pair production processes that arise when the perturbative methods are used , which do not contradict the existing nonperturbative results @xcite .
the results obtained so far using perturbative methods prove that some physical consequences related to the conservation laws and the mechanism of separation between matter and antimatter could be studied and this method is more suitable for analysing the phenomenon of pair production in strong gravitational fields @xcite .
an interesting observation can be made related to the results obtained by this two methods . in the limit
@xmath0 the probabilities computed with perturbations is proportional with a term of the form @xmath73 .
the beta bogolibov coefficient is also proportional with a term of this type .
so for @xmath0 the perturbative results reproduce up to some factors ( which depend on the particle momenta ) , the nonperturbative results but this is not an argument for comparing two different mechanisms of particle production . the mechanism of cosmological particle production use the fact that the vacuum could become unstable and for that reason the solutions of the free field equations of two different local charts could be related using bogoliubov transformations .
these transformations mix the solutions which describe the particle with the solutions which describe the antiparticle . in this case
the solutions of the field equations are defined locally and depend on the local chart , which means that the `` in '' and `` out '' modes are not the same .
these effects could be observed using local particle detectors .
this is the nonperturbative approach to the problem of particle production @xcite .
if one use the perturbative method then the modes are globally defined on entirely manifold and do not depend on local coordinates providing that the vacuum state is stable and unique @xcite . in this case
the quantum states are measured using a global apparatus which consist of conserved operators .
then the transition probabilities can be calculated using the exact solutions of the free field equations . in the scalar qed on de sitter geometry we can study the effect of electromagnetic interaction on the particle production @xcite .
these effects are present only when the gravitational field is still strong ( @xmath1 ) and vanish in the limit @xmath0 .
so the two mechanisms that study the problem of particle production are of completely different nature and it is premature to draw definitive conclusions in the absence of experimental evidence or to criticize the existing perturbative results without having strong mathematical and physical arguments @xcite .
in this paper we study the total probability for massive scalar particle production in electric field on de sitter geometry .
our results prove that the probability depend on the parameter @xmath72 and vanish in the minkowski limit .
this result is expected since we consider here the field produced by a distribution of point charges in de sitter geometry .
our study proves that the probability is nonvanishing only for @xmath1 regardless of the intensity of the electric field and the probability varies with the directions in which particles are emitted .
these conclusions are confirmed also by our graphical analysis . when the electric field is strong
we obtain that the probability of particle emission is larger when the particles are emitted in the direction of the electric field .
our results also prove that a direct comparison between nonperturbative methods and perturbative methods in what concerns the phenomenon of particle production , must be considered with care since the two mechanisms are completely different .
for obtaining function @xmath15 , we use : @xmath77 and @xmath78 99 j. schwinger , phys .
rev . * 82 * , 664 ( 1951 ) .
e. schr " odinger , physica * 6 * , 899 ( 1939 ) .
c.w.misner , k.s.thorne and j.a.wheleer , _ gravitation _
( w.h.freeman and company new york , 1973 ) .
i. i. cotescu and c. crucean , _ phys .
d _ * 87 * , 044016 ( 2013 ) .
s. weinberg , _ the quantum theory of fields _ ( cambridge university press , cambridge , 1995 ) .
s.drell and j.d.bjorken , _ relativistic quantum fields _ ( mc graw - hill book co. , new york 1965 )
. m. a. bloi , mod .
a * 29 * , 1450138 , ( 2014 ) .
l.parker , phys.rev.lett .
* 21 * , 562 ( 1968 ) .
l.parker , phys.rev . * 183 * , 1057 ( 1969 ) .
l.parker , phys.rev.d * 3 * , 346 ( 1971 ) .
n. d. birrel , p. c. w. davies and l. h. ford , _ j. phys .
a _ * 13 * , 961 ( 1980 ) .
n. d. birrel and p. c. w. davies , _ quantum fields in curved space _ ( cambridge university press , cambridge 1982 ) .
n. nicolaevici , mod .
a * 30 * , 1550046 ( 2015 ) .
k.h.lotze , nucl.phys.b * 312 * , 673 ( 1989 ) .
k.h.lotze , class.quant.grav .
* 2 * , 351 ( 1988 ) .
k.h.lotze , class.quant.grav . * 5 * , 595 ( 1985 ) .
v.m.villalba , phys.rev.d * 52 * , 3742 ( 1995 ) .
j.garriga , phys.rev.d * 49 * , 6343 ( 1994 ) .
buchbinder , e.s .
fradkin and d.m .
gitman , fortschr .
* 29 * , 187 ( 1981 ) .
i. l. buchbinder and l. i. tsaregorodtsev , int . j. mod
. phys . a * 7 * , 2055 ( 1992 ) .
g. w. gibbons and s. w. hawking , phys .
d * 15 * , 2738 ( 1977 ) .
j.haro and e.elizalde , j.phys .
a * 41 * , 372003 ( 2008 ) .
i. i. cotescu and c. crucean , prog.t.phys * 124 * , 1051 ( 2010 ) . c. crucean , phys .
d * 85 * , 084036 ( 2012 ) .
i. i. cotescu , c. crucean , a. pop , int . j. mod . phys.a
* 23 * ( 2008 ) 2563 .
c. crucean and m. a. bloi , int .
a * 30 * , 1550088 ( 2015 ) . | in this letter we study the problem of scalar particle production in external electric field in de sitter geometry . the total probability is calculated using the previously obtained result in ref.@xcite for transition amplitude in external electric field on de sitter space .
then we make a graphical study of the total probability in terms of the ratio mass of the particle / expansion factor .
our results show that the probability depend of the direction in which the particles are emitted and that the probability becomes maxim when particles are emitted on the direction of the electric field . in the minkowski limit we obtain that the probability is vanishing . |
strong coulomb electron - electron ( e - e ) interactions can drive transitions from metallic to exotic insulating states , the most well known of which are the mott - hubbard semiconductor ( mhs ) and the wigner crystal ( wc ) .
the mhs is a characteristic of systems with carrier concentration per site @xmath5 and is driven by strong onsite e - e repulsion , the hubbard @xmath6 interaction . depending upon the lattice structure the critical @xmath6
at which the metal - insulator ( mi ) transition occurs , @xmath7 , can be 0@xmath8 or finite @xcite .
in contrast , the wc occurs in systems with @xmath9 , and is characterized by charge - ordering ( co ) , _ i.e. _ , a periodic arrangement of _ single _ charge carriers on the lattice .
the wc is driven by strong onsite as well as inter - site coulomb interactions @xcite .
although in principle the wc is likely at any arbitrary @xmath10 , it has been been studied most intensively for @xmath11 bipartite lattices , where the nearest neighbor ( n.n ) coulomb repulsion can drive the mi transition @xcite .
the combined effects of e - e and electron - phonon ( e - p ) interactions are also of interest , usually in one dimension ( 1d ) , where the mhs can further exhibit the spin - peierls ( sp ) transition .
importantly for our purpose here the above semiconducting states have been intensively studied over the past several decades , and are largely understood , although arguments regarding the magnitude of @xmath7 for formation of the mhs state in specific lattices or the detailed mechanism of the mi transition may continue to persist . in the present work , we discuss a new correlated - electron semiconductor , the paired - electron crystal ( pec ) , that occurs in @xmath11 systems in the presence of moderate to strong geometric lattice frustration @xcite .
we believe that our work has direct application to 2:1 cationic or 1:2 anionic charge - transfer solids ( cts ) that exhibit correlated insulator
sc transitions , and further applies to other inorganic strongly - correlated @xmath4-filled materials .
the combined effects of e - e interactions and geometric lattice frustration are of strong current interest @xcite .
the bulk of the work here is for @xmath5 , where the hubbard model in the limit @xmath12 reduces to the heisenberg spin hamiltonian .
interest in the consequences of lattice frustration stems from the seminal proposal by anderson that the ground state of the heisenberg antiferromagnet ( haf ) model on a triangular lattice is a quantum spin liquid ( qsl ) with no spin ordering even at zero temperature@xcite .
the type of wavefunction usually assumed to describe a qsl is often referred to as a resonating valence bond ( rvb ) state . whether or not rvb states appear in the square lattice for @xmath10 slightly different from 1 , and the relationship of such states to superconductivity ( sc ) in doped strongly correlated semiconductors remains contentious .
an extension of the rvb theory of dopant - induced sc in @xmath13 is the proposal that frustration - induced sc occurs in the anisotropic triangular lattice within the simple hubbard model even for @xmath10 exactly 1 , where a narrow superconducting phase is straddled on both sides by broader paramagnetic metallic ( pm ) and antiferromagnetic ( afm ) insulator phases @xcite .
it has been claimed that this transition explains the sc in the cts @xcite .
recent numerical work by us and others , however , have determined that sc is absent within the @xmath5 triangular lattice hubbard model @xcite and the earlier results are artifacts of mean - field approximations .
while the ground state of the haf on the isotropic triangular lattice is now known to be the ordered 120@xmath14 afm rather than the originally proposed qsl state @xcite , other frustrated lattices , most notably the kagom lattice @xcite , have been investigated in the search for qsl states .
proposed ground states here include various types of qsl states @xcite as well as valence - bond solid ( vbs ) states @xcite .
the literature on vbs states has a long history going back to the well known ghosh - majumdar model @xcite .
the common theme in works on vbs is the frustration - driven transition from the afm state to a total spin @xmath15 singlet state .
we have found a similar frustration - driven transition from the afm to a @xmath15 state in strongly correlated systems with @xmath11 , where reduction to a spin hamiltonian is not possible .
in contrast to the voluminous literature on correlated and frustrated systems at @xmath5 , the literature on frustrated @xmath13 is relatively sparse and new .
the discovery of sc@xcite in hydrated na@xmath16coo@xmath1 has spurred interest in correlated systems away from @xmath5 @xcite , although to the best of our knowledge only isotropic triangular lattices have been studied .
we will specifically focus on @xmath11 within the present work on the triangular lattice with varying anisotropy . in the square lattice limit at this density
, we show that spontaneous in - phase dimerization occurs in the presence of electron - phonon interactions modulating n.n hopping integrals , leading to an _ effective _ @xmath5 system with one electron per dimer and afm order .
this result is the origin of the so - called the dimer mott - hubbard model @xcite that is commonly used to describe the 2:1 cationic or 1:2 anionic organic cts .
very recently we have proposed that under the influence of lattice frustration this dimer mott - hubbard afm state gives way to a spin - paired state that we termed the pec @xcite .
the pec is different from any of the above more well known correlated semiconducting states in that it is a wc of heitler - london spin - singlets simultaneously charged - ordered and spin @xmath15 . alternately , the pec is the @xmath11 equivalent of the @xmath5 vbs .
a conceptually similar state was postulated for the electron gas many years back by moulopoulos and ashcroft @xcite .
there is a fundamental similarity between this earlier work and ours , in that in both cases the pairing is driven by the exchange interaction .
our previous work @xcite only considered a limited set of parameters , and the full phase diagram was not discussed .
here we give a more complete phase diagram , including the competition with the wc state that was ignored before . as with the vbs state at @xmath5 ,
the pec is a consequence of frustration - induced quantum effects .
its extraordinary stability at @xmath11 is a commensurability effect ( recall that mhs and wc formation also require commensurability ) . in section [ model ]
we introduce the model we consider .
we include e - p interactions to stabilize the lattice dimerization that gives the effective @xmath5 dimer lattice .
we present the physical mechanism behind the pec formation by briefly discussing simple molecular clusters , for which we show that spin - singlet formation in @xmath11 necessarily requires charge disproportionation .
the charge disproportionation in the infinite one - dimensional ( 1d ) chain and the so - called zigzag ladder leads to periodic co , viz . , the simplest pecs . following the discussions of these simple cases , we introduce the two - dimensional ( 2d ) lattice that will be the focus of this work , and discuss the different possible phases .
in section [ results ] we present out numerical results for the 2d system , covering a wide region of parameter space . in section [ discussion ] we discuss the relevance of our results for several classes of @xmath11 materials and the outlook for understanding unconventional sc .
we particularly emphasize the cases of the organic charge - transfer solids cts @xmath0-(bedt - ttf)@xmath1cu@xmath1(cn)@xmath2 and etme@xmath2sb[pd(dmit)@xmath1]@xmath1 which have been described as qsls within the effective @xmath5 scenario @xcite .
we believe frustration - induced charge disproportionation is an alternate possibility .
the hamiltonian we consider contains electron hopping , semi - classical inter - site and onsite e - p couplings , and onsite and n.n .
coulomb interactions : @xmath17 in eq .
[ ham ] , @xmath18 indexes the different bond directions in the lattice ; for example @xmath19 in 1d and @xmath20 in the 2d square lattice .
our actual calculations ( see below ) are for the anisotropic triangular lattice , @xmath21 .
@xmath22 is the electron hopping between sites @xmath23 and @xmath24 with electron creation ( annihilation ) operators @xmath25 ( @xmath26 ) .
@xmath27 is the inter - site e - p coupling constant , @xmath28 is the corresponding spring constant , and @xmath29 is the distortion of the bond between sites @xmath23 and @xmath24 .
@xmath30 is the intra - site phonon coordinate and @xmath31 is the intra - site e - p coupling with corresponding spring constant @xmath32 .
both @xmath29 and @xmath30 are determined self - consistently @xcite .
@xmath27 are in general taken close to the minimum value needed for the transition to occur , our goal being the replication of the same instability from finite cluster calculations that would occur in the infinite system for @xmath33 coupling . @xmath6 and @xmath34 are on - site and n.n .
coulomb interactions , respectively .
the physically relevant range of @xmath34 is @xmath35 based on comparison between @xmath5 and @xmath11 cts @xcite .
we first present a simple qualitative discussion of coupled spin - singlet and co formation at @xmath11 .
the ideas are quite general and we argue that the mechanism is _ independent _ of dimensionality .
the key requirement is that the density must be exactly @xmath11 , as the effect requires commensurability .
consider a single dimer of two sites with one electron .
the electron populations per site are 0.5 each , but the quantum mechanical wavefunction for the system is the superposition @xmath36 , where @xmath37 and @xmath38 are site charge densities .
if one now brings two of these dimers together , as in insert ( a ) of fig .
[ cofig ] , the composite wavefunction of the two - dimer system can be written as @xmath39 . if the two electrons are in a spin - singlet state then within the simple hubbard hamiltonian the configuration @xmath40 , in which singlet stabilization can occur from a single n.n .
hop that creates a virtual double occupancy , must dominate over the configurations @xmath41 and @xmath42 , in which singlet stabilization requires two and three hops , respectively .
thus as the singlet bond between the dimers gets stronger we expect a charge difference @xmath43 between sites belonging to the same dimer ( nominally between sites 1 and 2 , or between sites 3 and 4 in the linear chain of fig .
[ cofig ] ) .
while some charge disproportionation must occur in finite linear chains from end effects alone , we note that our proposed picture demands that similar charge disproportionation occurs between members of the same dimer even in the case of the periodic molecule shown in the insert ( b ) of fig .
[ cofig ] . in this case
the charges on the sites connected by the diagonal bond must be larger than 0.5 , while the charges on the two other sites must be smaller .
importantly , the modulation of charge density , bond orders , as well as spin - singlet pairing all occur cooperatively , and any of these observables may be used as an order parameter in the case of a real transition . in fig .
[ cofig ] we have plotted @xmath43 versus the hopping integral @xmath44 corresponding to the dotted bonds in the molecules shown in the insert , for the ground spin - singlet state . in both cases , as @xmath44 increased from zero @xmath43 becomes nonzero and increases with @xmath44 .
importantly , in the spin - triplet @xmath45 state the sign of @xmath43 is reversed in the cyclic molecule , indicating repulsive interaction among the electrons .
coupled co and singlet formation occur also in the sp state in 1d @xmath11 systems @xcite .
a key difference from the sp transition in @xmath5 is that for @xmath11 , a mi transition first occurs at a intermediate temperature , followed by the sp transition at low temperature . for @xmath11
, the mi leads to either a bond - dimerized or to a charge - ordered wc state with equal bond lengths @xcite .
provided the n.n .
coulomb interaction is not too strong , a sp transition occurs from either insulating state @xcite , resulting in a ground state with period-4 co @xmath460110@xmath46 .
this state is the simplest realization of the pec and may visualized as a second dimerization of dimer units of molecules ; the singlet bond giving the spin gap ( sg ) forms _ between _ adjacent dimer units .
the intermediate temperature bond - dimerized states , wc state , and pec state with sg are all found experimentally in quasi-1d cts @xcite .
zigzag ladder systems , coupled two - stack systems in which each site on one stack is coupled to two sites on the other stack , are a second realization of the pec state .
cts zigzag ladder materials that are @xmath11 have been found with spin - gap transition temperatures much larger than in 1d @xmath11 sp materials @xcite .
the insulating ground state in this case may be understood again as a pec state occurring in a zigzag ladder lattice , with singlet bonds oriented between the two chains @xcite . unlike the 1d case ,
bond orders are now modulated in several lattice directions , leading to a larger sg than for the 1d pec case @xcite .
interestingly , in both the linear chain and the zigzag ladder , the @xmath11 pec is obtained by simply removing alternate spin - singlet bonds from the corresponding @xmath5 vbs , and replacing them with pairs of vacancies . in this paper
we will focus on ground state solutions of eq .
[ ham ] in 2d in the presence of variable lattice frustration .
the lattice we choose is a 2d square lattice with a single frustrating bond , as shown in fig .
[ lattices](a ) .
thus @xmath21 within eq .
[ ham ] for this lattice . in most of the results we will present , @xmath47 ,
although we will also consider @xmath48 in some cases .
energies will be given in units of @xmath49 .
we will take the frustrating bond @xmath50 in the range @xmath51 , covering the wide region between the unfrustrated square lattice ( @xmath52 ) and the nearly isotropic triangular lattice ( @xmath53 ) .
for the inter - site e - p coupling , unless denoted otherwise we choose @xmath54 and @xmath55 , with similarly identical spring constants @xmath56 .
for all calculations we assume periodic boundary conditions .
our calculations are largely for @xmath57 , but variable @xmath58 . in reference
we presented limited numerical results for a select set of coulomb interaction parameters ( @xmath59 , @xmath60 , @xmath61 ) demonstrating transition from nel antiferromagnetism to the pec state in this lattice when @xmath44 exceeds a critical value @xmath62 . for completeness and for giving an introduction to the various competing states in 2d we briefly review these results here . for small @xmath44 , the self - consistent solution of eq .
[ ham ] gives spontaneous dimerization along the @xmath63 axis as shown in fig .
[ lattices](b ) . the dimerized lattice state is effectively @xmath64-filled with one carrier per dimer and has nel afm order between dimers for finite @xmath6 .
the nel order is very clearly observable from the spin - spin correlations calculated for @xmath65 clusters @xcite .
importantly , the charge density @xmath66 of all sites is exactly 0.5 .
as @xmath44 increases , frustration reduces the strength of the antiferromagnetic correlations . provided the n.n .
coulomb interaction @xmath34 is not too strong ( see below ) , at a critical @xmath67 antiferromagnetic correlations disappear and charge disproportionation develops , with the charge densities within each dimer becoming inequivalent .
this is shown in fig .
[ lattices](c ) . in this state , the charge densities follow the pattern @xmath461100@xmath46 along the @xmath63 and @xmath68 directions , and the pattern @xmath461010@xmath46 along the @xmath69 direction .
the bond distortion is period-4 along @xmath63 and period-2 along @xmath69 .
the strongest bond orders @xmath70 occur between adjacent charge - rich ` 11 ' sites in the @xmath68 direction .
importantly , the spin - spin correlations change dramatically for @xmath71 .
they are strongly negative between the bonded ` 11 ' sites along the @xmath68 direction ( see fig . [ lattices](c ) ) , and are nearly zero between either member of the pair and all other sites , indicating the formation of spin - singlet bonds @xcite .
any of the observables @xmath43 , bond order between the charge - rich sites , or z - z spin correlations , @xmath72 , may be used as order parameters for the pec state @xcite . in section [ results ]
we present further details of the pec phase and the full parameter dependence of eq .
[ ham ] , with the goal of demonstrating that _ ( i ) the transition to the pec that we are interested in is driven by quantum effects due to frustration only ; and ( ii ) the pec occurs over a broad region of parameter space . _ given the number of parameters in eq .
[ ham ] , it should be relatively easy to generate co driven by specific ( presumably artificial ) choices of @xmath34 .
such classical results would be uninteresting .
we therefore consider several distinct choices of coulomb interactions : ( i ) @xmath73 and all @xmath74 , ( ii ) @xmath73 , @xmath75 , and @xmath61 , ( iii ) @xmath73 and @xmath76 .
pec formation occurs in all of these parameter regions .
we also show that for sufficiently strong n.n .
coulomb interactions corresponding to parameter region ( ii ) , the wc phase with checkerboard co ( fig .
[ lattices](d)-(e ) ) is the ground state of eq .
we also will consider other modifications of the basic lattice , viz .
, sign of @xmath44 opposite site to @xmath49 , and @xmath77 .
finally , we will argue that our results are not consequences of finite size effects and are to be expected in the thermodynamic limit .
figs . [ bigplot](a)-(d ) show the charge disproportionation @xmath79 , bond order @xmath80 , n.n . z - z spin - spin correlation @xmath72 , and spin gap @xmath81 as a function of @xmath44
. @xmath81 is defined as the excitation energy from the ground @xmath15 state to the lowest @xmath45 state .
@xmath6 here is finite but all @xmath82 terms are zero .
sites @xmath23 and @xmath24 in figs . [
bigplot](b)-(c ) correspond to two ` 1 ' sites in the pec state connected by a @xmath44 bond ( filled circles connected by heavy lines in fig .
[ lattices](c ) ) . as with the results in reference which included the @xmath83 and @xmath84 interactions , for small @xmath44
the charge density is 0.5 on all sites , and antiferromagnetic order can be seen in the spin - spin correlations ( not shown here , see fig .
3(a ) in reference ) .
nonzero @xmath62 in fig .
[ bigplot ] is a consequence of the nature of the diagonal bonds in fig .
[ lattices](b ) ; the diagonal bonds inside each plaquette with two strong dimer bonds actually strengthen the afm , and only the inter - plaquette diagonal bonds have a frustrating effect . at small @xmath44 these two effects appear to cancel , and there is only a weak effect on the afm . for @xmath71 ,
@xmath43 becomes nonzero , with the charge pattern as shown in fig .
[ lattices](c ) .
similarly , the bond order between paired ` 11 ' sites increases abruptly , and the z - z spin - spin correlation between these sites becomes strongly negative , and nearly zero with all other lattice sites ( see reference ) , indicating formation of a singlet bond .
although @xmath81 is nonzero in all cases in a finite cluster , we nevertheless see a large jump in @xmath81 at @xmath62 , also indicating spin - singlet formation . the increase in bond order strength and strength of spin - spin correlation
clearly follow the same pattern as @xmath79 .
these results show that the n.n .
coulomb interaction is not essential for formation of the pec state . unlike the wc phase where co is driven by the n.n .
coulomb interaction , the pec state is a consequence of geometric lattice frustration .
increasing @xmath6 moves @xmath62 to larger @xmath44 : @xmath6 tends to strengthen the afm phase and therefore increasing @xmath6 is expected to make the afm order persist for stronger lattice frustration .
the transition also becomes more discontinuous to changes in @xmath44 as @xmath6 increases , suggesting it may be continuous for small @xmath6 and first order for large @xmath6 . in both the 1d and in the zigzag ladder lattice ,
the pec state occurs unconditionally even in the noninteracting limit ( @xmath85 ) for any finite e - p coupling . in these two cases ,
the unconditional occurrence of pecs is a consequence of simple nesting .
in contrast , in the isotropic @xmath11 2d band considered here , the lack of nesting forbids an unconditional peierls transition . in agreement with this
, we found that for @xmath86 , the pec phase did not occur .
instead , the self - consistent calculations converged to disordered states with no clear charge pattern .
this is an indication that in the thermodynamic limit , the preferred ground state is one of uniform charge .
this result is reminiscent of that in quantum spin systems : the simplest vbs transition , the sp transition , can be predicted from nesting behavior in the 1d xy model following jordan - wigner transformation @xcite .
this is , however , not true in 2d frustrated spin systems .
we next consider the effect of n.n .
coulomb interactions @xmath75 , but @xmath61 . figs .
[ bigplot](e)-(h ) show the same order parameters as in figs .
[ bigplot](a)-(d ) for @xmath88 . comparing the data with and without @xmath82 , the effect of moderate @xmath82 is to _ strengthen _ the pec state
the magnitude of all order parameters increase when @xmath89 for a fixed value of @xmath6 . for fixed @xmath6 , the afm
pec boundary @xmath62 also moves to smaller @xmath44 with increasing @xmath82 .
[ uv - phase ] shows the phase diagram in the @xmath44-@xmath6 plane for both the @xmath90 and @xmath89 cases . here
, for each value of @xmath6 , @xmath44 was varied until the afm
pec transition occurred , and the first @xmath44 where @xmath43 became nonzero was taken as the afm pec boundary .
one expects that when @xmath82 is above a critical value @xmath91 , the @xmath461100@xmath46 pec co will give way to the checkerboard wc state . in 1d
, this transition occurs exactly at @xmath92 in the limit @xmath12 , and at a larger @xmath91 for finite @xmath6 @xcite .
previous exact diagonalization for a 2d cluster@xcite ( without however e - p interactions or @xmath44 as considered here ) found @xmath93 for @xmath94 , and showed that @xmath91 increases when @xmath95 .
[ wcplot ] shows the evolution of @xmath96n and diagonal bond orders with @xmath82 for @xmath97 . while difficult to see in fig .
[ wcplot ] due to the choice of axis scales and parameters , in the pec region @xmath43 increases with increasing @xmath82 .
for the parameters of fig .
[ wcplot ] ( @xmath6=6 , @xmath98 , @xmath99 , @xmath100 , @xmath101 ) the charge order pattern changes from the pec ( fig . [ lattices](c ) ) to the checkerboard wc ( fig . [ lattices](d)-(e ) ) at @xmath102 .
in addition , @xmath79 increases sharply when entering the wc phase . at larger @xmath103
a slight cusp occurs in the @xmath43 versus @xmath82 plot . at the same time
, the pattern of bond orders changes : for @xmath104 in the wc phase the bond orders alternate strong - weak along the @xmath68 direction ( as shown in fig .
[ lattices](e ) ) , while for @xmath105 the bond orders along @xmath68 are uniform ( fig .
[ lattices](d ) ) . within the wc phase region
there are therefore two sub - phases : a phase with equal length bonds in the diagonal @xmath68 ( @xmath44 ) directions , and a phase in which these bonds become dimerized .
the added bond dimerization in the diagonal direction will result in a spin gap , and we denote this phase as the wigner crystal spin gap ( wc - sg ) phase .
a similar spin - gapped wc phase can be found in a small region of parameter space in the 1d model @xcite .
[ vt - phase](a ) shows the resulting phase diagram in the @xmath44-@xmath82 plane for @xmath6=6 .
the @xmath91 we find in the @xmath52 limit is slightly larger ( @xmath97 , @xmath106 ) than the results of reference ( @xmath94 , @xmath107 ) ; however , both the smaller @xmath6 here as well as the e - p coupling in eq .
[ ham ] would be expected to increase @xmath91 . unlike the afm , pec , and wc phases , the wc - sg phase is limited to a relatively narrow range of parameters .
larger lattice calculations are needed to confirm whether the wc - sg phases persists in the thermodynamic limit .
the @xmath109 interaction destabilizes the checkerboard - pattern wc , leading to a metallic phase in the absence of e - p interactions @xcite . here
we will consider parameters @xmath76 . for eq.[ham ] without e - p interactions , @xmath94 , and @xmath110 , exact diagonalization found@xcite that in this case a metallic phase exists for @xmath82 up to at least @xmath1115 .
charge fluctuations within the metallic phase adjacent to the wc were speculated to cause a co - to - sc transition @xcite . in our calculations , summarized in the phase diagram in fig .
[ vt - phase](b ) for @xmath97 , we also found that the wc phase does not occur , but rather than being metallic the system is insulating either afm at small frustration or pec at large frustration . in this case
pec transition can occur over a wide range of lattice frustration , @xmath112 . within the pec phase the co pattern remains the same for all @xmath82 , although @xmath43 increases with @xmath82 as in the @xmath61 case considered in the previous section . although our calculations are for one value of @xmath109 only , fig .
[ vt - phase](b ) suggests that the pec region is broadened relative to that in fig .
[ vt - phase](a ) for any @xmath113 .
as shown in the previous section , variation of coulomb interactions can cause a substantial variation in the extent of frustration needed to form the pec state .
next we show the effect of varying the one - electron parameters in eq .
[ ham ] , @xmath114 and @xmath27 . due to the lack of particle - hole symmetry in the anisotropic triangular lattice
, differences might be expected when @xmath44 is taken as negative .
however , as fig .
[ uv - phase ] shows , we found only a small variation in the afm
pec phase diagram when the sign of @xmath44 is changed .
this is consistent with the expected mechanism for spin frustration in an effectively @xmath64-filled band : the frustrating exchange interaction is proportional to @xmath115 , so reversing the sign of @xmath44 should only change the effective frustration at higher order . fig .
[ uv - phase ] also shows the effect of changing @xmath116 .
as expected for a cooperative transition , stronger e - p coupling increases the size of the pec region .
the effect of anisotropy of the inter - site e - p interaction is shown in fig .
[ txty - alpha - phase](a ) , where @xmath117 is fixed at 1.1 and @xmath118 is varied .
increasing either @xmath117 or @xmath118 separately strengthens the pec , as shown in fig .
[ txty - alpha - phase](a ) , where the @xmath119 ratio is varied .
pec phase diagram is relatively insensitive to the ratio of @xmath120 . in fig .
[ txty - alpha - phase](b ) , @xmath121 is fixed at 1 and the value of @xmath122 is varied the resulting @xmath62 does not depend sensitively on the choice of @xmath120 . summarizing sections [ results](a)-(d ) , we see the pec state in our calculations for a wide range of parameters . in every case , we first arrived at parameters that placed the system in the afm or wc phases in the @xmath52 limit , and then varied only @xmath44 .
thus the transition to pec is a consequence of frustration alone .
with variation of e - p coupling and coulomb interactions , the amount of frustration needed to drive the afm
pec transition can vary over a sizable range of frustration .
we will discuss the implications of this further in relationship to the cts materials further in section [ discussion ] .
we have performed several checks on our calculations that indicate that the pec state found in our numerical calculations is an intrinsic property of eq .
[ ham ] and _ not _ induced by finite size effects . _
( i ) noninteracting bandstructure : _ one common finite - size effect in numerical calculations are changes in the fermi - level degeneracy or level crossings of the noninteracting system . in the lattice of fig .
[ lattices](a ) the fermi level degeneracy does not change throughout the range @xmath123 , remaining 2-fold degenerate throughout this range .
this degeneracy is broken by @xmath63-axis dimerization , shown in fig .
[ lattices](b ) , giving a nondegenerate fermi level for @xmath123 . at @xmath53 the degeneracy at the fermi level increases to 6-fold ; in the presence of interactions the ground state can become triplet ( @xmath124=1 ) for @xmath125 .
hence we stop at @xmath126 where the ground state is @xmath15 . _
( ii ) interactions : _ as mentioned above , no transition occurs for @xmath85 .
this further indicates that the transition is not a feature of the single - particle bandstructure . _
( iii ) commensurability : _ we have verified that the pec state does not occur for electron densities different from @xmath11 ; for example , no transition to pec or any `` nonmetallic '' state occurs for 6 or 10 electrons on the @xmath127 lattice .
in the following we discuss how our theory applies to real materials , and may even give insight to the mechanism of correlated - electron superconductivity in @xmath11 materials .
the superconducting organic cts share many characteristics of other strongly - correlated superconductors , in particular the high - t@xmath128 cuprates , including reduced dimensionality and the presence of afm near sc . at the same time ,
sc in the cts occurs under pressure at a constant carrier density of @xmath11 rather than under the influence of doping . a variety of exotic insulating states _ in addition to afm _ , including co @xcite and spin - gapped states @xcite , as well as possible qsl states @xcite , are proximate to the superconducting state in the cts .
our work shows that only the afm phase is described by the dimer mott - hubbard model .
since in all cases the experimental systems are structurally related , with identical or near - identical molecular components , we believe that the same mechanism of sc should apply to them . in reference
we had pointed out how the pec concept can perhaps lead to such a unifying theory . here
we expand on this theme .
we briefly review here experimental evidence for pec formation in several 2d cts families . _
( i ) @xmath129-(et)@xmath1x : _ co corresponding to the pec and spin gap are found in the @xmath129-(et)@xmath1mm@xmath130(scn)@xmath131 family @xcite . in @xmath132=rbzn
, the co occurs below the mi transition at @xmath133190k , while the sg appears below 20k @xcite .
the charge order pattern in the co phase below the mi transition has been experimentally determined for @xmath132=rbzn and follows a horizontal stripe pattern ( see fig . 6 in reference ) .
the horizontal co is definitely not the wc .
rather , the co pattern is precisely as expected in the pec , with @xmath461100@xmath46 co along the two directions of largest hopping ( the @xmath134-directions in the @xmath129-(et)@xmath1x lattice ) , and @xmath461010@xmath46 order along the direction of weakest hopping ( @xmath135-direction in @xmath129-(et)@xmath1x ) .
experiments have revealed that with decreasing temperature , the @xmath135-axis lattice parameter decreases @xcite .
the decrease in the lattice parameter implies increased carrier hopping in this direction and therefore increased frustration within our theory , giving the transition to the singlet pec and observed spin gap . _ ( ii ) @xmath116-(et)@xmath1x : _ the crystal structure of @xmath116-(et)@xmath1x is quite similar to that of @xmath129-(et)@xmath1x .
the existence of a sg opening below 136k has been known for some time in @xmath116-(et)@xmath1i@xmath2 @xcite . below the 136k transition co
is found which has been confirmed to be of the same pattern as in @xmath129-(et)@xmath1x ; see for example fig . 2 in reference . _
( iii ) @xmath31-(meso - dmet)@xmath1pf@xmath136 : _ this 2d cts exhibits a pressure - induced transition from co to sc @xcite . while the charge order pattern in this cts is referred to as `` checkerboard '' by the authors , the checkerboard pattern refers to meso - dmet _ dimers _ as units . in terms of meso - dmet
_ monomer _ units , the co pattern is the same as the pec in fig .
[ lattices](c ) , with @xmath461100@xmath46 in two directions and @xmath461010@xmath46 along the third direction ( see fig . 2 in reference . ) _ ( iv ) @xmath137-x[pd(dmit)@xmath1]@xmath1 : _ in this family the materials dmit molecules are arranged in dimers . the frustration varies with the cation @xmath138 , with the least frustrated in the series showing afm order @xcite . among the materials with larger frustration
, @xmath139 has a sg transition at 25k to what has been described as a vbs state @xcite .
the experimentally determined bond and charge distortion patterns ( see fig .
3(b ) in reference ) are exactly as expected for the pec , with period 4 charge and bond distortions along what is the @xmath63 axis in fig .
[ lattices ] .
the intradimer charge disproportionation in particular argues against the sg state from being a simple vbs , which would require equal charge densities on all the molecules .
there has been much recent interest in these cts @xcite specifically because they present possible realizations of the long awaited qsl @xcite . in both cases
the materials have nearly isotropic triangular lattices of dimer unit cells ( corresponding to fig .
[ lattices](b ) with @xmath53 ) within an effective @xmath5 model . in @xmath0-(et)@xmath1cu@xmath1(cn)@xmath2 ( hereafter @xmath0-cn ) the estimate for the heisenberg exchange integral between n.n .
dimers is @xmath140220250 k @xcite . @xmath141h
nmr experiments find absence of long range magnetic order down to 32 mk @xcite .
very similar behavior is also seen in etme@xmath2sb[pd(dmit)@xmath1]@xmath1 ( hereafter dmit - sb ) @xcite .
the insulating ground state in @xmath0-cn is close to sc , transition to superconductivity occurring under moderate pressure @xcite .
sc is found in the dmit family as well @xcite .
recent experiments in both @xmath0-cn and dmit - sb have found peculiarities that appear to be unexpected within qsl theories .
below we list experiments that seem to indicate that apparent qsl behavior at low temperatures is giving way to a `` hidden order '' , and perhaps even charge disproportionation , which by itself would be against spin - only models .
\(i ) a second order phase transition is seen at 6 k in @xmath0-cn in measurements of heat - capacity c@xmath142 @xcite , @xmath143c nmr relaxation rate 1/t@xmath144 @xcite , and lattice expansion coefficients @xcite .
the last experiment finds strong lattice effects at the transition , indicating possible role of charge degrees of freedom @xcite .
a symmetry - breaking and/or topological ordering transition at t @xmath145 1 k has also been observed in dmit - sb @xcite .
\(ii ) the specific heat c@xmath142 in @xmath0-cn is linear in t for t between 0.75 and 2.5 k , indicating a gapless energy spectrum @xcite .
the sommerfeld coefficient @xmath146 is nonzero and large even at @xmath14775 mk @xcite .
equally perplexingly , c@xmath142 is independent of magnetic field up to 8 t , indicating absence of zeeman coupling of spins to the field @xcite .
in contrast , thermal conductivity measurements down to 80 mk indicate a spin gap @xcite .
\(iii ) the temperature dependence of the thermal conductivity of dmit - sb suggests gapless mobile excitations , but magnetic field dependence of the thermal conductivity again indicates a gap @xcite . taken together , ( ii ) and ( iii )
suggest gapless spin - singlet excitations but gapped spin - triplet excitations in @xmath0-cn and perhaps also dmit - sb .
it is conceivable that the gap in dmit - sb is nodal @xcite .
\(iv ) measurements of dielectric response in @xmath0-cn have shown increasing and frequency - dependent dielectric constant below 60 k , and possible antiferroelectric ordering of dipoles at t@xmath1486 k @xcite .
the latter requires unequal site charges on the molecules within the dimer unit cells @xcite .
it may be relevant in this context that @xmath143c - nmr experiments on both @xmath0-cn @xcite and dmit - sb @xcite find unusual line broadenings at low t that can not be ascribed to disorder @xcite , which might also indicate charge disproportionation .
similar line broadening at low t in etme@xmath2p[(dmit)@xmath1]@xmath1 occurs at the transition to the pec @xcite , which in turn gives way to superconductivity under pressure @xcite .
although it is as yet not entirely clear whether or not existing spin - liquid models @xcite with appropriate modifications can explain the above anomalies , it appears that for both @xmath0-cn and dmit - sb , spin degrees of freedom alone can not describe the low temperature ( below 6k in @xmath0-cn and below 1k in dmit - sb ) properties .
rather , any description of the ground state must involve charge as well as spin degrees of freedom .
assuming that the above experiments and their interpretations are correct , the proper theoretical model should have built - in significant electron - lattice coupling and should lead to charge disproportionation and excitation energy spectrum with gapless singlet excitations and gapped spin excitations .
we believe that the highly frustrated @xmath11 model satisfies all of the above criteria , in addition to providing the starting point for a theory of superconductivity in the cts ( see below ) .
t - linear specific heat but gapped magnetic susceptibility were noted in nonmetallic vanadium bronzes@xcite as far back as 1978 , and at the time was considered to be a distinctive proof for bipolarons .
the nonzero spin - singlet degeneracy within this model comes from tunneling motion of the bipolarons , causing them to `` flip flop '' between equivalent configurations @xcite .
we propose that a similar mechanism is at play in the present case at large @xmath44 . particularly in @xmath0-cn the lattice structure and
the orientations of molecules are such that the spin - singlet bonds , which occur between monomers belonging to two neighboring dimers can flip flop between the monomers ( see fig . 4 in reference .
) within our proposed picture for @xmath0-cn , the transition at 6k is to the spin - singlet charge - disproportionated state , with short range fluctuating order .
magnetic excitations even in this state , however , require breaking the spin - singlet bonds .
note that the observed difference between @xmath0-cn and dmit - sb is expected within the @xmath11 model , since the simple description as a triangular lattice of dimers is no longer enough and the detailed couplings between the monomers in the materials are indeed different because of their different crystal structures . as mentioned above , numerical results for @xmath149
are not useful due to the highly degenerate ground state becoming spin - triplet in our finite clusters .
we have suggested elsewhere that the occupied spin - singlet bonded ` 11 ' sites of the pec can be thought of as _ effective _ single sites doubly occupied by charge carriers , and similarly the pairs of ` 00 ' vacancies can be thought of as single vacant sites @xcite .
such a mapping would transform the pec to an effective checkerboard co with alternate sites ( in the square lattice representation ) occupied by double occupancies and vacancies . the effective hamiltonian that describes the checkerboard co in this case is a @xmath5 extended hubbard model with weak _ attractive _ @xmath6 whose origin is the exchange interaction that stabilizes the singlet bond within the original @xmath11 hamiltonian . the n.n .
interaction within the effective hamiltonian remains repulsive to simulate the checkerboard ordering of the effective doubly occupied sites .
while such a mapping is not rigorous , similar mapping of n.n .
spin - bonded sites to double occupancies has been routinely used in the literature on bipolaron models @xcite .
the difference between our work and traditional bipolaron models is that the spin - singlet bonding within our work is driven primarily by afm correlations , while within the bipolaron models it is a consequence of effective attraction due to the overscreening of the e - e repulsion by e - p interactions @xcite .
we have investigated the consequences of stronger frustration@xcite in the @xmath11 anisotropic triangular lattice within the effective @xmath5 extended hubbard model with attractive @xmath6 @xcite .
we found a co - to - sc transition within the effective model @xcite , suggesting a pec - to - sc transition within the @xmath11 model with repulsive interactions .
although this result is not a proof of transition to sc within the repulsive @xmath11 model , it nevertheless is instructive and provides the direction for future research .
demonstration of sc within the actual model will require further work .
what is interesting though is that the proposed scenario can give a unified approach to sc in all organic cts , irrespective of whether the insulating state proximate to sc is afm @xcite , co @xcite or vbs @xcite .
recall that within existing mean field theories the afm - to - sc transition is driven by spin fluctuations @xcite , while the co - to - sc transition is driven by charge fluctuations @xcite . even if we ignore that recent precise numerical calculations @xcite have demonstrated the absence of sc within the proposed spin - fluctuation models in this context ( thereby raising doubts also about mean - field theory of charge - fluctuation mediated superconductivity ) , different mechanisms of sc for structurally related materials with identical or near - identical molecular components appear to be unrealistic . within our proposed scenario ,
the charge - ordered or vbs systems that exhibit sc are pecs in the semiconducting state , while the afm systems under pressure transform to pecs first and then to sc
( although due to actual crystal structures the `` width '' of the pec region could be narrow to vanishing within the latter ) .
the sc phase within this scenario is a paired - electron liquid .
there is considerable overlap between these ideas and the one proposed by moulopoulos and ashcroft for the continuous electron gas @xcite .
the thrust of our work has been to understand within a unified theoretical approach the variety of exotic insulating states that are proximate to the superconducting state in 2:1 cationic or 1:2 anionic 2d cts .
we have shown here that the peculiarities of these materials originate from the unique behavior of @xmath11 in the presence of both strong e - e interactions and lattice frustration .
it is conceivable that the complex behavior of apparently unrelated inorganic families can be understood within the same broad theoretical approach .
we list two such classes of materials below to point out the hitherto unnoticed similarities between them and @xmath11 cts .
layered cobaltates na@xmath16coo@xmath1 have attracted wide attention because of their 2d structure , tunable carrier concentration , and the occurrence of sc @xcite .
the co ions form a 2d triangular lattice , are in their low - spin state , and their valence ranges from co@xmath150 at @xmath151 to co@xmath152 at @xmath153 .
charge carriers are @xmath154 holes on the co@xmath152 sites and the hole density @xmath155 .
trigonal distortion splits the occupied @xmath156 orbitals into @xmath157 and @xmath158 orbitals .
lda calculations have suggested that although for large @xmath63 the @xmath157 orbitals occur below the fermi level and an @xmath158-only description is valid , this description breaks down at small @xmath63 where @xmath157 orbitals can be nearly degenerate @xcite .
in contrast , correlated - electron calculations find that the @xmath158
@xmath157 energy separation is positive and relatively large for all @xmath63 @xcite , suggesting that low energy excitations can likely be described within @xmath158-only single - band models .
the temperature dependent magnetic susceptibility @xmath159 shows a peculiar @xmath10-dependence within the family , with small @xmath10 ( large @xmath63 ) exhibiting strongly correlated behavior and large @xmath10 ( small @xmath63 ) exhibiting weakly correlated behavior @xcite .
we have recently shown that @xmath10-dependent @xmath159 , exactly as seen in the cobaltates , is expected within the single - band extended hubbard model on a triangular lattice @xcite . equally interestingly ,
@xmath159 behavior in na@xmath16coo@xmath1 is very similar to that in the family of cts as a whole , where also @xmath159 shows a systematic @xmath10-dependence that is understood within the single - band extended hubbard model @xcite .
it is tempting to compare now the superconducting states in hydrated cobaltates and cts with this apparent similarity in mind .
although superconductivity in na@xmath16coo@xmath1 @xmath160 @xmath69h@xmath1o occurs at @xmath161 @xcite , it is now established that the co - ion valency here is determined not only by the na content , but also by h@xmath2o@xmath8 ions .
there have been several reports that sc here occurs over a very narrow range of hole density @xmath10 , and that maximum @xmath162 occurs at or very close to co - ion valency 3.5 + , corresponding to @xmath11 @xcite .
if this is confirmed from future experimental work , it would appear that like the cts , cobaltates are yet another example of a frustrated 2d @xmath11 superconductor , suggesting that the mechanism of sc in the two families is related .
the b sublattice in spinel compounds ab@xmath1x@xmath131 form
a frustrated three - dimensional ( 3d ) pyrochlore lattice and usually consist of transition metal cations that possess partially filled @xmath156 @xmath163-orbitals .
for integer occupancies of @xmath163-electrons per @xmath164 cation the geometrical degeneracy of the underlying lattice is often lifted by orbital ordering ( oo ) , leading to formation of spin - singlet dimers .
only four of the many spinel compounds are superconducting , of which three have effective carrier density @xmath11 : liti@xmath1o@xmath131 , curh@xmath1s@xmath131 , and curh@xmath1se@xmath131 . in liti@xmath1o@xmath131
there is one @xmath163-electron per two ti@xmath165 ions ; in curh@xmath1s@xmath131 and curh@xmath1se@xmath131 the rh@xmath165 ions have average @xmath163-hole occupancy of @xmath64 .
jahn - teller instability or oo can lead to occupancy of the same @xmath156 orbitals , making the filled bands exactly @xmath4-filled , as in the superconducting organic cts ( the apparently different spinel superconductor cuv@xmath1s@xmath131 , also has noninteger number of @xmath163-electrons per v@xmath165 ion . ) yet another similarity between the organics and the @xmath11 spinels is the proximity of the superconducting state to exotic semiconducting states . thus cuir@xmath1s@xmath131 and lirh@xmath1o@xmath131
, both isoelectronic with curh@xmath1s@xmath131 , undergo mi transitions that are accompanied by co. in cuir@xmath1s@xmath131 the ir ions are charge - ordered as ir@xmath150-ir@xmath150-ir@xmath152-ir@xmath152 along specific directions @xcite .
this would correspond exactly to the @xmath460011@xmath46 co in our notation here . as in the 2d pec , these inorganic 3d systems are spin - gapped due to the formation of ir@xmath152-ir@xmath152 singlet bonds .
it is interesting to recall that similar singlet bonds between ti@xmath150 ions had been proposed @xcite many years back within bipolaron theories of ti@xmath131o@xmath166 and liti@xmath1o@xmath131 .
as explicitly shown in our work here , the singlet bond formation is a natural consequence of frustration and @xmath4-filling .
work is currently in progress to extend the pec concept to the checkerboard lattice , thought to be 2d equivalents of the pyrochlore lattice , multiple orbitals per site @xcite .
in summary , there is an extraordinarily strong tendency to form spin - singlets in systems with charge carrier concentration precisely @xmath64 .
naturally , in @xmath11 this spin - singlet state is accompanied by co. the stability of the pec derives from the commensurability of the pec at @xmath11 . in the anisotropic triangular lattice , the pec consists of charge arrangements @xmath461100@xmath46 in two directions and @xmath461010@xmath46 in the third direction .
thus although @xmath11 in principle is incommensurate on the triangular lattice , `` separately commensurate '' periodic charge arrangements are nevertheless possible .
we thank m. abdel - jawad , h. fukuyama , k. kanoda , and r. kato for helpful discussions .
this work was supported by the us department of energy grant de - fg02 - 06er46315 .
rtc thanks the university of arizona for support while on sabbatical . | we present a study of the effects of simultaneous charge- and spin - frustration on the two - dimensional strongly correlated quarter - filled band on an anisotropic triangular lattice .
the broken - symmetry states that dominate in the weakly frustrated region near the rectangular lattice limit are the well known antiferromagnetic state with in - phase lattice dimerization along one direction , and the wigner crystal state with the checkerboard charge order . for moderate to strong frustration , however , the dominant phase is a novel spin - singlet paired - electron crystal ( pec ) , consisting of pairs of charge - rich sites separated by pairs of charge - poor sites .
the pec , with coexisting charge - order and spin - gap in two dimension , is the quarter - filled band equivalent of the valence bond solid ( vbs ) that can appear in the frustrated half - filled band within antiferromagnetic spin hamiltonians .
we discuss the phase diagram as a function of on - site and intersite coulomb interactions as well as electron - phonon coupling strength .
we speculate that the spin - bonded pairs of the pec can become mobile for even stronger frustration , giving rise to a paired - electron liquid .
we discuss the implications of the pec concept for understanding several classes of quarter - filled band materials that display unconventional superconductivity , focusing in particular on organic charge transfer solids .
our work points out the need to go beyond quantum spin liquid ( qsl ) concepts for highly frustrated organic charge - transfer solids such as @xmath0-(bedt - ttf)@xmath1cu@xmath1(cn)@xmath2 and etme@xmath2sb[pd(dmit)@xmath1]@xmath1 , which we believe show frustration - induced charge disproportionation at low temperatures .
we discuss possible application to layered cobaltates and @xmath4-filled band spinels . |
we show below through physical examples that low energy nuclear reactions have been induced by _ all _ of the four fundamental interactions : gravitational , strong , electromagnetic and weak .
* gravity * : gravitational interactions are well known to cause nuclear reactions and fusion in a star . were it not for nuclear activity
, a star would be dynamically unstable and undergo gravitational collapse@xcite .
in fact , standard theory predicts the collapse of a star when the nuclear fuel is exhausted as the star can no longer counteract the inward compression due to gravitation . *
nuclear * : a laboratory example of low energy strong interaction fusion is provided by a fast discharge in fine deuterated polymer fibers . in such fibers ,
deuterons are accelerated to speeds high enough to overcome the barrier due to mutual coulomb repulsion , giving rise to the production of @xmath0 neutrons through low energy reactions such as @xmath1 in the same set of experiments@xcite , also _
non deuterated _ fibers exhibit an `` anomalous '' production of neutrons , at a rate over 6 orders of magnitude larger than that expected through natural contamination of deuterons in a normal material .
such experimental evidence strongly suggests an explanation in terms of weak interaction induced nuclear reactions@xcite .
this will be discussed in a later sec.[nondeuterated ] .
* electromagnetic * : purely electromagnetically induced fusion has been observed through the coulomb explosion of large deuterium clusters -stripped of their electrons by lasers- and their subsequent blow up resulting in nuclear fusion@xcite . in more recent experiments ,
other charged nuclei have been exploded and the ensuing nuclear reactions have been observed . *
weak * : in the corona of the sun , magnetic fields emerge from one sunspot to dive into another .
the coronal magnetic field accelerates the electrons and protons enough to cause production of neutrons through weak interactions@xcite .
these neutrons then cause further nuclear transmutations of the available material accounting for the presence of anomalously formed nuclei on the surface of the sun and in the solar corona@xcite . once in a while ,
magnetic flux tubes explode leading to spectacular solar flares releasing extremely high energy particles some of which reaching the earth .
often , the resultant electromagnetic fields are large enough to disturb terrestrial satellite communications and secondary muons produced in the upper atmosphere in sufficient numbers to have been detected in the underground cern l3+c detector@xcite . in the following work
, we shall analyze in some detail , various mechanisms and phenomena found both in nature and in the laboratory related to low energy nuclear reactions .
we shall make a general classification of these diverse phenomena in order to delineate common themes ( such as collectivity , thermal or non - thermal nature of their induction ) as well as profound differences ( such as the time scale for induction by a particular interaction ) between nuclear reactions caused by different interactions under diverse circumstances .
we shall also illustrate the interplay between various interactions necessary for the genesis of a given physical process .
for the energy production in a star , all four interactions are essential .
in the standard model of star formation , collectivity plays a central role .
it is asserted that undamped density fluctuations cause an astronomically large number of interstellar gas ( essentially hydrogen ) to clump together , whereupon gravity begins to compress them further thereby increasing the protostar s temperature . to appreciate the subtle interplay existing between different interactions in causing nuclear reactions , the leitmotif of our paper , it is useful to recall the paradigm shifts in the view that earlier physicists had to make regarding the main source of a star s energy production and stability , as recounted by hans bethe@xcite , a leading architect of gravitational low energy nuclear theory ( gravitational lent ) . to estimate
the stability of a star , say our sun , helmholtz -around the 1850s- was the first to use the only tool available to him : newtonian gravity .
if one gram of matter falls on the sun s surface , it acquires a potential energy @xmath2 given by @xmath3 a similar energy must have been set free when the sun was assembled . through the virial theorem ,
he argued , one half of it must have been set free as kinetic energy [ which allows us to estimate the temperature ] and the other half must have been radiated away . at present , the outward flux of radiation from our sun is @xmath4 hence , it was estimated that if gravitation supplies the energy , then in @xmath5 the sun would have radiated away all its energy and would have undergone a gravitational collapse . but
darwin , other biologists and geologists argued with helmholtz ( and kelvin@xcite , who concurred with helmholtz ) that they needed a much longer life time for the sun and that his gravitational mechanism for the source of energy production in the sun must be in error .
of course , they were right .
in fact , as we will see , gravitation supplies the necessary force to hold and compress the particles collectively and raise their temperature for positively charged nuclei to overcome the coulomb barrier but it is the weak force which ignites the nuclear fire by providing neutrons without which no deuterons or heavier nuclei could be formed . in 1895 , radioactivity in the weak sector of the standard model of fundamental interactions would be discovered by henri becquerel , pierre curie and marie curie and through it the age of the earth and later that of meteorites and other geological objects would be determined .
one deduces the age of the sun from that of the meteorites to be 4.5 billion years in close agreement with the age 4.56 billions deduced from solar oscillations@xcite .
clearly , we need some interaction other than gravitation to fuel the sun .
since the sun has existed for at least 4.5 billion years , we may deduce that the gain in energy per gm to be provided by whatever agency must be at least @xmath6 if not gravitation , then which of the other three em , weak or strong can be the source ? if for simplicity we assume that initially we only had electrons and protons and we want to fuse the protons into heavier nuclei , then we need the intervention of neutrons .
electromagnetism can not do it by itself because it can not change the charge .
also , the direct strong interaction between two protons to change their charges is well nigh negligible since the temperature of the core is estimated to be about @xmath7 .
this corresponds to a mean thermal energy for the electrons and protons to be a mere @xmath8 .
hence , the only alternative left is to invoke the weak interaction .
the fermi theory does provide a possibility of neutron production through the inverse beta decay @xmath9 this reaction was first proposed by wick and later independently by yukawa and sakata for the observed phenomena of electron capture from the k and l shells from a large proton rich nucleus .
but notice that in an electron capture , the electron is bound near the nucleus , the proton is bound within the nucleus and the subsequent neutron remains bound to the nucleus .
all that escapes is the unobserved neutrino . of course , this is not quite exact .
as the closest to nucleus electron disappears , a higher shell electron jumps down to replace it , a third to displace the second , etc .
, thus leading to a cascade of emitted photons .
it is precisely through the emitted photons , chiefly in the x - ray region , that the process can be directly registered at all . alvarez@xcite
provided the first clinching experimental evidence that the @xmath10 nucleus captures one of its orbital electrons and converts to the stable nucleus @xmath11 .
the life time of k capture and hence the transformation of one atom to a different atom can be of the order of seconds to months depending upon the mass difference between the parent and the daughter atom .
it is well to remember that even in the vacuum , prior to the intervention of temperature , several orders of magnitude difference in the rates can occur due to differences in phase space . instead , for the reaction in eq.([g4 ] ) , both the electron and the proton have negligible kinetic energy on the nuclear @xmath12 scale . hence , the mean @xmath13 of the reaction is negative : @xmath14 .
one may consider the thermally induced process @xmath15 where the thermal radiation @xmath16 supplies kinetic energy to the initial system .
but the available thermal energy even near the core of the sun @xmath17 is miniscule in comparison with the threshold energy @xmath18 required for the reaction to proceed .
hence , the process can occur through thermal fluctuations albeit with an exceedingly small probability .
the dramatic difference between the time scales involved between this case , bethe s estimate for the core of the sun is @xmath19 , and the observed k captures , typically a few days , is due to the large exponential threshold barrier in the first case .
this will be of crucial interest for us in what follows , where the reaction in eq.([g4 ] ) would be invoked for specially designed condensed matter systems in which the electron kinetic energy can be boosted enough or in a relativistically covariant language , the electron mass can get renormalized by external electric fields radiated from currents , @xmath20 so as to overcome the threshold barrier @xmath13 value turning positive , as in the electron capture case , with substantial rates for neutron production . in eq.([g6 ] ) , anticipating effects due to electron collectivity , we have generalized the reaction in eq.([g4 ] ) to the case of @xmath21 initial electrons which end up with @xmath22 final electrons after the weak destruction of a single electron . as we shall see in sec.([betatron ] ) , the presence of a large number of electrons in dense matter systems can , through the darwin interaction , change the energy distribution of a single electron . in such cases , the rates for the reaction in eq.([g6 ] ) can be quite different from that of eq.([g5 ] ) or for electron capture . given the depressingly low rate due to the threshold barrier for the direct production of neutrons through eq.([g5 ] ) , von weizscker @xcite suggested the exothermic reaction @xmath23 with a gain of energy @xmath24 from a transmutation of two protons . for this reaction , we can estimate the energy gain / gm to be @xmath25 a value quite close to that needed in eq.([g3 ] ) in order to keep the sun warm .
the completion of the mechanism described through the above processes requires the reaction rates for this `` @xmath26 '' channel of energy production .
it was first computed by bethe and critchfield@xcite and it involves all three , em , weak and strong interactions of the standard model of particle physics . at low relative kinetic energy between the two protons ,
we have a strong coulomb repulsion inhibiting the entire process eq.([k ] ) .
this is a two step process : first we have the essentially `` zero '' range weak fermi process @xmath27 followed by the very short range strong interaction process @xmath28 , visualized in the feynman diagram in fig .
@xmath29 where @xmath30 is an intermediate `` virtual '' state neutron , which combines with the second proton to fuse into a deuteron . with an eye towards later usage for other processes ,
let us decompose the computation into ( i ) a repulsive coulomb gamow suppression factor @xmath31 and ( ii ) a weak - strong cross section for a generic process @xmath32 where @xmath33 refer to the charge of the nuclei @xmath34 and @xmath35 refers to some specified final state .
the effective cross section reads @xmath36 where @xmath37 in the above equation , @xmath38 denotes the relative velocity , @xmath39 is the reduced mass , @xmath40 is the fine structure constant and @xmath41 is the relative kinetic energy when the two nuclei are far apart .
the cross section @xmath42 is in principle computable from first principles of the known weak , em and strong interactions . for low velocities , @xmath43 and hence @xmath31 is approximately @xmath44 since , the reactions eq.([k3 ] ) of interest are exothermic for which the cross sections @xmath45 for small @xmath46 , it is convenient to write the effective cross section as @xmath47 where @xmath48 is a characteristic of the nuclear reaction under investigation .
bethe - critchfield@xcite obtain for deuteron production through the pp chain eq.([k ] ) @xmath49 which is exceedingly small in comparison to purely nuclear reactions , i.e , where @xmath35 does not contain leptons and so no weak processes are involved .
in fact , the reaction eq.([k ] ) has never been observed in any earth laboratory . by contrast , for non - leptonic initial and final states , one finds@xcite @xmath50 typically ( @xmath51 ) results when the transition is radiative , i.e. whenever there is a photon in the final state whereas for purely nuclear transitions we find numbers in the ( @xmath52 ) range .
let us pause to understand why the reaction eq.([k8 ] ) has a more than @xmath53 lower probability to occur than a typical strong reaction eq.([k9 ] ) . since the former is a weak process with a cross section proportional to @xmath54 , @xmath55 being the fermi coupling constant and @xmath56 , we expect a suppression factor of about @xmath57 with respect to the strong process .
the extra suppression by about @xmath58 arises due to the reduced probability [ @xmath59 , with @xmath60 the nucleon mass ] for a virtual neutron to combine with a proton to form a deuteron . of course , eq.([k7 ] ) is not the end of the story .
it would if it were a _ beam _ experiment , that is if the incident energy @xmath41 were fixed , but there is a maxwellian velocity distribution which has to be integrated upon . hence , bethe and later bahcall and others inserted it and the rate of nuclear reaction at a given temperature @xmath61 is proportional to @xmath62^{1/2 } \int de e \sigma_{eff}(e ) e^{-e/(k_bt ) } .\ ] ] this integration changes the final temperature or equivalently the mean energy dependence considerably in the rate of the reaction . one further point regarding @xmath63 needs be made .
further variations can and do occur if there are resonances .
this is discussed below . for room temperature non - thermal induction of neutron production such as electron capture in a plasmon polariton set up ,
there is no cross section only a capture rate . to calculate a cross section from a rate one requires two beams so that one may define a cross section as a rate divided by a flux . in the context of computing rates when the beams are in reality oscillating plasma oscillations ,
one is perhaps better off simply quoting the rates of processes . in the fermi theory of point interaction ,
the electron has to be on top of the nucleus for the reaction to go .
the method suggested@xcite employs the process of radiation induced electron capture by protons or deuterons to produce ultra low momentum neutrons and neutrinos . in the neighborhood of metallic hydride cathodes , protons or deuterons can do such inverse beta decay conversions through the collective plasma modes of the charged particles .
hence , its proper description is through eq.([g6 ] ) and not through a two body cross section since electrons and protons or deuterons oscillate back and forth at a frequency @xmath64 and have a root mean square lateral displacement @xmath65 and when a renormalized invariant mass @xmath66 with @xmath67 is reached for an initial electron , it can cause the fermi inverse beta reaction .
this is strikingly different from a beam situation .
failure to do so and analyze this reaction using two body cross section at a fixed velocity would lead to quite erroneous results .
the deuterons produced from the pp reaction can in turn produce @xmath69 and an @xmath68 : @xmath70 @xmath71 @xmath72 as discussed in the sec.[tokomak ] barring selection rules , @xmath73 for reactions eq.([a1 ] ) and eq.([a3 ] ) should be @xmath74 , whereas for eq.([a2 ] ) it should be @xmath75 . also , as mentioned in eq.([w1 ] ) , the neutron in eq.([a2 ] ) has an energy of @xmath0 which has been measured in exploding wire experiments@xcite as well as in the coulomb explosion experiments@xcite .
schwinger@xcite considered the radiative reaction eq.([a3 ] ) and observed that at low energies , if one includes just the @xmath73-wave , the dipole radiation would not contribute because of the positive intrinsic parities of the @xmath76 , @xmath77 and @xmath78 and negative parity of @xmath79 and hence it should be highly suppressed .
the same argument should also apply to the radiative decay eq.([a1 ] ) . on the other hand ,
all the intrinsic parities are positive in eq.([a2 ] ) and hence this process can proceed via @xmath73-wave .
it is thus encouraging that both in exploding wire and coulomb explosion experiments , reaction eq.([a2 ] ) has been observed and to our knowledge reaction eq.([a1 ] ) not .
it is suppressed at the level of @xmath80 in the vacuum .
proceeding on with the pp chain , it has become customary in the astrophysics literature to invoke the bethe - von weizscker weak process twice and write it as a four initial proton process @xmath81 needless to say that there is no possibility to ever measure this reaction directly on earth . on the other hand , for a partial verification of the pp chain in the sun ,
the detection of neutrinos arising from the fast part of the cycle was undertaken . as discussed in the next section , nuclei between helium and carbon
are unstable .
one such unstable process which concerns bound and free electron capture from a @xmath82 nucleus leading to @xmath83 nucleus and emission of an electron neutrino has been widely investigated@xcite .
the electron capture reaction @xmath84 half - life has been measured in the laboratory@xcite @xmath85 = \ ( 4.61 \pm\ 0.001 ) \times 10 ^ 6\ sec.,\ ] ] leading to a rate @xmath86 .
this implies , that after @xmath87 , we should find about @xmath88 atoms and the same number of neutrinos if we had an initial supply of say @xmath53 @xmath82 atoms .
historically , the inverse electron capture reaction @xmath89 was employed for the solar neutrino detection on earth by davis and others .
we mention the cno cycle employed in massive stars simply to direct attention to the role of resonances .
early workers @xcite had difficulty in bridging the gap between @xmath90 and @xmath91 in the stars since the elements in this gap are unstable . given the great abundance of helium , the process of fusing 3 alpha particles was invoked @xmath92 since it requires the simultaneous presence of three particles in the initial state , the probability should be extremely small except that this process is blessed with a double resonance .
salpeter@xcite pointed the first resonance which is the unstable nucleus @xmath93 with approximately the same mass as two alpha particles and hoyle@xcite noted the second i.e. , @xmath93 and @xmath94 have the same mass as an excited state of @xmath95 .
of course , the reaction eq.([c1 ] ) can not be observed in the laboratory , but the resonances can and have been observed .
bethe first emphasized the great importance of the above process which should be the main reaction producing stable carbon , active when the temperatures reach around @xmath96 .
carbon is of course considered crucial for the buildup of elements . around the same temperature ,
another capture of an alpha particle produces oxygen @xmath97 we shall not follow the astrophysical consequences any further , except to remind the reader of the intricate interplay between all the four interactions in producing various low energy reactions leading to very different rates .
we now turn our attention to laboratory examples where strong interaction fusion reactions , such as @xmath98 have been directly observed through clean signals of threshold @xmath99 neutrons . as discussed in the last section ,
the coulomb barrier is quite large at low energies , see eq.([k7 ] ) , and thus it is well known that in the vacuum this process would have a vanishingly small probability . in the core of the stars , when the temperature becomes sufficiently high to overcome the barrier , the process is supposed to occur . in reality , the results of such experiments do not yet have practical importance .
an alternative method using fast electrical discharge through thin deuterated polymer fibers has been very successful in causing fusion in many laboratories .
this method is not directly thermal but uses an electrical impulse to accelerate the deuterons enough to overcome the coulomb barrier .
we shall discuss it in the next subsection using an example of a particular experiment .
let us focus on the exploding wire experiment by stephanakis _
et al_@xcite .
their experimental setup is summarized below : fibers used : polyethylene and polypropylene density of deuterium @xmath100 high power discharge @xmath101 typical discharge currents @xmath102 peak voltage @xmath103 duration of the discharges @xmath104 length of the fibers @xmath105 diameter of the fibers @xmath106 mean kinetic energy of the electrons : ( @xmath107 mean electron speed @xmath108 through the equipartition theorem , it was deduced that when the plasma contained deuterium ions , it had also a temperature equivalent to kinetic energies in the range @xmath109 , sufficient to overcome the coulomb barrier and cause fusion .
reading from the graph of their integrated yield @xmath110 of @xmath99 neutrons , the yield seems to scale initially with the diameter @xmath111 according to @xmath112 and then it seems to saturate for higher diameters .
the authors state , without proof , that the variation in the yield with fiber diameter can be understood on the basis of pressure and energetics .
their integrated yield @xmath113 neutrons of @xmath99 for thin deuterated fibers of diameter @xmath114 .
it is indeed a hefty rate giving ample proof that fusion had occurred .
but they were puzzled by an anomalously high production of neutrons by natural i.e. , non deuterated fibers as discussed below . while the production of @xmath0 neutrons from deuterated fibers appeared to satisfy expectations from the standard strong interaction branch eq.([s1 ] ) , an apparent paradox was generated when yields from normal fibers were found to be considerably higher than that expected .
the isotopic abundance of deuterium in normal matter is about @xmath115 .
the fusion yield being proportional to the square of the density of deuterium , the expectations were that natural fiber yield should be about @xmath116 lower than the deuterated fiber yield , for the same fiber geometry .
but measurements proved otherwise .
the normal yields are about @xmath117 to @xmath118 orders of magnitude higher than that expected .
it is natural to conclude that some mechanism other than the strong interaction fusion is responsible for this effect .
there are also other natural phenomena where neutrons have been observed in normal matter with virtually nil deuteron contamination .
examples are lightning@xcite and thunderstorms@xcite . in both cases , an anomalous production of neutrons has been reported . for a normal fiber ,
if all we have are electrons and protons then the only way to produce neutrons is via weak interactions we shall discuss at some length the dynamics of weak neutron production in a later section .
here we shall just anticipate some results .
as discussed earlier , for an electron - proton system , we are not afflicted by the coulomb barrier , as a matter of fact the process is helped by coulomb attraction . but to produce a neutron , about @xmath119 needs to be put in the system to reach threshold for this reaction .
to thermally excite an electron to such an energy is not possible even at the core of most stars .
however , under special circumstances there are electromagnetic means e.g. , high electric and magnetic fields@xcite , or `` smart '' materials@xcite which can directly convert elastic or other mechanical energy into em signals , to supply such an energy . for a thin fiber
, it is useful to consider an initial state of a large number of ( @xmath120 ) interacting electrons undergoing a weak process with a proton @xmath121 we shall see later that the importance of having a large number of `` spectator '' electrons , even though only one is destroyed at a time , is due to the induction of a coherent darwin interaction@xcite between the electrons which leads to a high collective contribution to the reaction energy thereby enhancing the nuclear activity .
it produces a strong flux of neutrons in normal fibers , lightning and in thunderstorms .
in recent experiments , exploding molecular clusters of deuterium atoms have been produced .
first , a weak laser pulse hits the cluster internally ionizing the atoms within the cluster and it is then followed by a strong laser pulse photo - ejecting a large number of electrons completely out of molecular cluster .
this leaves the cluster with a large positive charge ( @xmath122 ) in a sphere of small radius @xmath123
. clusters with large @xmath124 in a small @xmath123 explode . ejected deuterons from different clusters then collide with sufficient kinetic energy to overcome the coulomb barrier and lead to observed fusion events .
such experiments have now been repeated in several laboratories around the world and results bear out their theoretical analysis .
let us estimate what magnitudes for the fields are required for a coulomb explosion .
let @xmath125 , @xmath41 and @xmath126 denote respectively the voltage , electric field and the stress on the system of charges .
@xmath127 the _ tensile strength _ of a material @xmath128
is defined as the maximum allowed stress before the material disintegrates .
@xmath129 typical explosion fields are of the order of @xmath130 we shall see later that similar fields would be required for inducing weak processes through surface plasmon polaritons on metallic hydride surfaces .
we can deduce a hot disintegration temperature @xmath131 for coulomb explosion devices , using eqs.([e1]-[e2 ] ) : @xmath132 of the order of the temperature near the core of the sun . through the collision between two deuterons from different clusters , zweiback _
et al _ @xcite and buersgens _ et al _ @xcite , observed the production of neutrons via the reaction eq .
( [ s1 ] ) : @xmath133 the peak of their signal showing up at the expected characteristic neutron energy of @xmath134 .
the cluster size can be measured through light scattering .
when the wave length @xmath135 of the incident light in a raleigh light scattering experiment is large on the scale of the cluster radius @xmath123 , then the elastic scattering amplitude is determined by the polarizability @xmath68 .
@xmath136 so the elastic differential cross - section reads @xmath137 and the integrated elastic cross - section reads @xmath138 the total cross - section is given by @xmath139 the radius @xmath123 of the clusters has been determined empirically via light scattering cross sections as a function of the temperature @xmath61 of the gas jet producing the clusters . for a gas pressure of @xmath140 , as the temperature was increased from @xmath141 to @xmath142 , the measured radii of the clusters decreased from about @xmath143 to about @xmath144 .
the temperature @xmath61 of the gas jet producing the clusters also determines the absorption cross section for the light .
thus , the cluster radius can be determined by the absorption of light .
such measurements have also been made .
in addition , the yield as a function of the radius of the clusters have been measured .
the experimental results are consistent with theoretical computations described above .
the success achieved with deuteron fusion led to a series of experiments using other molecular clusters@xcite through which detailed dynamical mechanisms for the explosions could be understood along with the structure of the molecules themselves .
it has become an imaging tool as well , see for instance@xcite
a detailed review of electro - weak interaction induced lent [ ew lent ] can be found@xcite , hence we shall be very brief here . for systems containing only electrons and protons , as amply discussed in the preceding sections , weak interactions between electrons and protons are essential in producing neutrons . once there is a supply of low energy neutrons , subsequent nuclear transmutations can occur at a high rate .
it is crucial for the production of initial neutrons from low energy @xmath145 in a condensed matter environment , that an input energy @xmath146 be supplied somehow to overcome the threshold barrier [ see eq.([g5 ] ) ] .
to accelerate the electrons thermally to mev s in a condensed matter system is impracticable and one is left to search for em means for the needed level of electron acceleration .
there are essentially three options for electron acceleration which have already produced good results \1 .
* fast lasers * : very fast femtosecond lasers can and have been employed successfully to create mono - energetic electron beams from @xmath147@xcite up to @xmath148@xcite and table top electron accelerators have been constructed . while such table top accelerators are extremely useful devices for all sorts of studies , they are not quite suitable for inducing electroweak lent . \2 . *
magnetic means * : as mentioned in the introduction , an excellent example of ew lent in nature is provided through the magnetic field acceleration of electrons and protons via the betatron mechanism in the solar corona .
we shall discuss it in the next section ( [ corona ] ) .
* electric means * : a laboratory example is an electric field acceleration of electrons when piezoelectric rocks are crushed .
this would be discussed in sec([smart ] ) devoted to lent induced by ` smart ' materials .
a very promising avenue for ew lent yet not completely demonstrated in the laboratory is through plasmon polaritons on the surface of metallic hydrides . under suitable conditions in particular , [ @xmath149 , where @xmath41 is the mean electric field , @xmath64 is the resonant frequency , @xmath150 is the electron mass and @xmath151 , weak production of neutrons can occur .
in fact , experimentally the naples group has observed neutrons generated through plasma in a mizuno type battery .
the detection of neutrons were made through nuclear cr-39 detectors and these results are repeatable@xcite .
same as above also holds for piezo electrics.the difference is that the resonant frequency for metallic hydrides is in the infrared whereas acoustic frequencies are in the microwave range .
thus , considerably higher electric fields are needed for lent to occur on metallic surfaces .
galileo first saw solar sunspots through his optical telescope and to this day the physical origin of these dark sunspots is unclear . early in the last century
, hale and others@xcite discovered that magnetic flux tubes exit out of one sunspot to then dive into another sunspot .
such an uninterrupted magnetic flux tube has a part of it submerged below the solar surface where the electronic density is high and a part which is out in the solar corona where the density of electrons is low . a schematic picture is shown in fig .
such a `` floating '' flux tube is held up by the magnetic buoyancy@xcite .
there are oppositely directed vortex currents of electrons and protons which loop around the outer magnetic flux tubes .
an intricate pattern of stable steady state magnetic flux tubes form a `` magnetic carpet '' and they store considerable magnetic energy .
this collective magnetic energy @xmath152 can be estimated to be high enough to cause inverse beta decay similar to that discussed previously @xmath153 let @xmath154 denote a small change in the current going around the vortex circumference .
it would lead to a small change in the magnetic field energy @xmath155 @xmath156 if @xmath157 denotes the length of the vortex circumference of the magnetic flux tube , then the destruction of one electron through the weak reaction eq.([s1 ] ) implies @xmath158 with @xmath46 denoting the component of the relative velocity between an electron and a proton which is tangential to the circumference . letting @xmath159 and @xmath160
, we have @xmath161 \left[\frac{v}{c}\right].\ ] ] for a cylindrical flux tube , @xmath162 = \frac{r}{2}.\ ] ] yielding @xmath163 \left[\frac{b}{kilogauss}\right ] \left[\frac{v}{c}\right].\end{aligned}\ ] ] employing the estimates @xmath164 we find @xmath165 that is quite large enough for an appreciable production of neutrons from the protons .
these neutrons thereby give rise to further nuclear transmutations into higher mass nuclei via neutron capture and subsequent beta decays .
let us discuss here a non - thermal acceleration process which leads to a generation of quite energetic particles in the solar corona .
as described in the last section , a mechanism is required to induce nuclear reactions well outside the solar photosphere . the central feature of our solar accelerator @xcite centers around the faraday s law and is closely analogous to a betatron@xcite .
conceptually , the betatron is a step up transformer whose secondary coil is a toroidal ring of accelerating charged particles circulating about a time varying magnetic flux tube .
the solar magnetic flux tube acts as a step up transformer mechanism in transferring particle kinetic energy outward from the inner photosphere to circulating charged particles located in the corona outside . circulating currents located deep in the photosphere
can be viewed conceptually as a net current @xmath166 circulating around a primary coil .
circulating currents found high in the corona can be viewed as a net current @xmath167 circulating around a secondary coil .
if @xmath168 and @xmath169 represent , respectively , the charged particle kinetic energies in the primary and secondary coils , then one finds the step up transformer power equation @xmath170 where @xmath171 represent the voltages across the primary , secondary coils respectively .
the total kinetic energy transfer @xmath172 the essence of the step up transformer mechanism is that the kinetic energy distributed among a very large number of charged particles in the photosphere can be transferred via the magnetic flux tube to a distributed kinetic energy shared among a distant much smaller number of charged particles located in the corona , i.e. a small accelerating voltage in the primary coil produces a large accelerating voltage in the secondary coil .
the resulting transfer of kinetic energy is collective from a large group of charged particles to a smaller group of charged particles .
the kinetic energy per charged particle of the dilute gas in the corona may then become much higher than the kinetic energy per charged particle of the more dense fluid in the photosphere . in terms of the connection between temperature and kinetic energy , the temperature of the dilute gas in the corona will be much higher than the temperature of the more dense fluid photosphere .
it is clear that for the betatron , or equivalently the transformer mechanism , to be efficient in sustaining large electrical fields , the coronal electrical conductivity must not be too large . for an estimation of the conductivity ,
we need to know the density of the electrons in the corona .
fortunately , we have a quite reliable experimental measurement of the electronic density through its effect on the gravitational bending of light in the radio wave band near the sun .
given its importance , we shall sketch the methodology of this measurement .
bending of an em beam near the sun may be looked upon as a change in the refractive index , from unity for the vacuum . for frequencies
@xmath173 , where @xmath174 is the plasma frequency , the effective refractive index @xmath175 at a position @xmath176 , where the charged particle density is @xmath177 , reads @xcite @xmath178 successful measurements of the gravitational bending of electromagnetic waves with frequencies in the visible and all the way down to the high end of the radio spectrum are legion .
these experiments provide a direct proof that any coronal conductivity disturbance on the expected gravitational bending of electromagnetic waves for frequencies down to @xmath179 gigahertz must be negligible .
deviations from the general relativistic effect have allowed quite accurate measurements of the electronic density in the solar corona .
the estimates from even lower frequency radio wave probes used for gravitational bending @xcite put the coronal conductivity in the megahertz range . for comparison
, we note that the typical conductivity of a good metal would be more than ten orders of magnitude higher@xcite .
hence , these data lead us to conclude that the solar corona is close to being an insulator and eons away from being a metal and there are no impedimenta toward sustaining electrical fields within it @xcite .
thus , our proposed transformer mechanism and its subsequent predictions for the corona remain intact . when sufficiently high kinetic energy is reached in the circulating currents around the flux tubes floating in the corona
, the flux tubes may explode violently into a solar flare .
the loss of magnetic energy during the flux tube explosion is rapidly converted into charged particle kinetic energy .
the relativistic high energy products of the explosion yield both nuclear and elementary particle interactions .
the extraordinary explosion which occurred on july 4 , 2000 was strong enough to have generated particles , for example an excess muon flux , some of which were observed experimentally at the cern l3 + c detector@xcite .
it offered a wonderful opportunity to infer that protons with an average energy of @xmath180 were produced in the solar corona .
independently , the baksan underground muon measurement@xcite provided experimental evidence for protons of energy in excess of @xmath181 in the solar flare of september 29 , 1989 .
we show below some estimates@xcite .
for a large solar flare exploding in a time ( @xmath182 ) , the break up of a magnetic tube would yield , through the faraday law , a mean acceleration voltage @xmath183 around the walls given by @xmath184 inserting as before , @xmath185 , where @xmath31 denotes the mean magnetic field before the explosion and @xmath186 the inner cross - sectional area , we have for the mean acceleration energy @xmath187 for a cylindrical geometry , we may again rewrite it in a useful form @xmath188 \left[\frac{\pi r^2}{\lambda \times kilometer}\right].\ ] ] for a coronal mass ejecting flux tube exploding in a time @xmath189 seconds , we use the estimated values to derive @xmath190 physically , it corresponds to a colliding beam of electrons and protons with a beam energy of @xmath191 .
experimental measurements by l3 and baksan confirm the very high energy acceleration of electrons and protons and our proposed mechanism is consistent with the data .
in @xcite we have also made predictions for the flux of the positron from such flares .
needless to say , that the measurements already made belie the notion that all nuclear reactions in the sun occur in the core of the sun .
similar conclusions can be drawn for any star with a high enough magnetic field .
it is often overlooked that in a d - d or a d - t system@xcite -driven by modest static electric fields- inertial electrostatic confinement has been achieved leading to fusion .
such devices known as fusors have existed since the 1960 s .
the basic idea is to use a hollow transparent ( perforated ) spherical cathode enclosed by a concentric spherical anode , and the device filled with deuterium or deuterium and tritium at low pressure .
ionization of the gas occurs via corona discharge
. a potential difference of around 80 kv in commercial devices accelerates deuterons radially , and despite collisional losses , substantial number of collisions take place with energies over 15 kev and fusion takes place with the release of neutrons . much lower voltages can be used .
in fact , the voltage threshold for d - t fusion ( the lowest ) is very modest , comparable to that used in many photocopiers , well below the 10 kv of old black and white television sets or furnace ignition transformers , and also well below the typical 15 kv for neon signs or the 25 kv used in old - style colour tv sets . for d - d fusion
, the threshold voltage is again quite modest around 20 kv , and corresponds to an energy that can and indeed has been reached by pyroelectric devices as discussed in section [ pyro ] .
we stress that this is not a speculative idea , but rather a mature technology , and tabletop devices can be purchased commercially@xcite which produce ( @xmath192 ) neutrons per second .
indeed , a quick internet search will readily reveal that fusors are popular science fair projects built by students .
amongst a large class of `` smart '' or more appropriately , specifically designed materials with special properties , pyro and piezo electric materials have been demonstrated both theoretically as well as experimentally to lead to low energy nuclear reactions .
a pyroelectric crystal develops an electric field due to ( adiabatic ) changes in its temperature and its opposite , that is an applied electric field causing an adiabatic heating or cooling of the system is called the electrocaloric effect . in a piezo electric crystal , mechanical stress is directly converted to the development of electric fields and vice versa . under suitable conditions , sufficient electric fields [ voltages ] have been generated in both cases to lead to neutron production .
we shall discuss each of them briefly . in @xcite , it was experimentally shown that through two pyroelectric crystals when heated or cooled produced nuclear @xmath193 fusion evidenced by the observation of the characteristic signal of @xmath0 neutrons .
the system was used to ionize the gas and accelerate the ions up to @xmath194 . as we have seen in the previous sections , e.g. eq.([s1 ] ) , this acceleration is much more than sufficient to cause @xmath193 fusion .
the measured yields are in good agreement with the calculated yield .
the formal definition of a pyroelectric crystal is that within a single domain , the electric induction @xmath195 .
when such a pyroelectric crystal is heated or cooled , it gets spontaneously polarized under equilibrium conditions .
the often quoted phenomenological expression presumably valid in the linear regime gives the effective electric field @xmath196 generated inside a pyroelectric crystal due to a temperature change @xmath197 @xmath198 where @xmath199 is the pyroelectric coefficient .
pyroelectric crystals are modeled as parallel plate capacitors with the accelerating voltage @xmath125 given by @xmath200 where @xmath201 is the area , @xmath202 is the dielectric constant and @xmath203 the thickness of the crystal .
hence , in this simple model , the energy imparted to an electron or an ion of charge @xmath204 reads @xmath205 crystals were employed since they have a particularly large value for this coefficient , @xmath206 at atmospheric pressure , the change in polarization charge is quickly masked by the accumulation of charges from the air .
hence , it is in the vacuum that we see the surface charge change and the resulting electric field . for , @xmath207 and @xmath208 for a two crystal set up@xcite , the increase in energy should be @xmath209 whereas the maximum electron energy gain measured is only about @xmath194 . for ions , it is even less .
the deuterium fill gas in a vacuum chamber was ionized by the strong electric field between two polarized crystals and the deuterons were accelerated into a deuterated polystyrene target at energies above the threshold for fusion . to obtain the maximum accelerating voltage ,
two crystals were employed to double the voltage and a catwhisker tip was employed to increase the electric field .
the crystals were heated up to @xmath210 and then cooled to room temperature .
very clear signals for neutrons with energies @xmath211 were observed by the usual time of flight method . over six decades
, considerable experimental evidence has been gathered for high energy particle production during the fracture of certain kinds of crystals @xcite .
fracture induced nuclear transmutations and the production of neutrons have been clearly observed@xcite .
the production of neutrons appears greatly enhanced if the solids being fractured are piezoelectric materials .
common examples of a piezoelctric are : hair , bone , quartz . in a recent paper@xcite , we have theoretically analyzed the problem and shown the manner in which the mechanical pressure in a piezoelectric stressed solid about to fracture can organize the energy so that neutrons can be produced . since some of the arguments may not be familiar to many readers , we shall repeat below some of the relevant theory .
the energy per unit volume of a piezoelectric material may be written as @xmath212 where @xmath213 is the entropy per unit volume ; @xmath214 and @xmath215 are the stress and strain tensors , @xmath216 and @xmath217 are the polarization and electric field vectors .
the adiabatic piezoelectric tensor is defined as@xcite @xmath218 the effective mechanical electric field interaction energy density , up to quadratic order , reads @xmath219 leading to the piezoelectric hamiltonain @xmath220 eq.([p4 ] ) provides a direct link between a large strain ( say a fracture caused by stress ) causing a large electric field .
if one uses this interaction twice , then to second order in @xmath221 , the em photon propagator would receive a contribution from the mechanical acoustic phonon propagator in the following way .
let @xmath222 and @xmath223 be the adiabatic electric susceptibilities at constant strain and at constant stress respectively @xmath224 these two are related through the elastic response tensor @xmath225 via the thermodynamic identity @xmath226 the above static equations are easily generalized to dynamical frequency dependent susceptibilities @xmath227 and @xmath228 and the phonon propagator @xmath229 .
phonon modes described by the phonon propagator affect the dynamical susceptibilities through the relations @xmath230 clearly then , the electric susceptibility of a piezoelectric material would be affected by the frequencies associated with the mechanical acoustic frequencies occurring in fractures of such materials .
the velocity of sound @xmath231 compared to the velocity of light @xmath232 obeys @xmath233 and hence for similar sized cavities @xmath234 these phonon modes will then appear in @xmath235 .
the nuclear reactions are supposed to proceed through the weak capture of high electric field accelerated electrons by protons as described in section([ewlent ] ) .
we need to calculate the mean energy of the electrons in condensed matter when accelerated by an electric field @xmath236 and the electron energy is estimated to be @xmath237 if @xmath238 denotes the mean squared electric field strength in a band width ( @xmath239 ) , then eqs.([p10],[p11 ] ) imply @xmath240 if @xmath64 is a dominant frequency , then the above equation may be written as @xmath241 let @xmath242 be the ( electrically ) renormalized mass . then , in terms of the vacuum electron mass @xmath150 , we may write @xmath243 for electron capture to occur , @xmath244 . for a rough estimate of the electric field , we may assume that most of the stress at a fracture is electric , in which case @xmath245 where we have used griffith s rule that @xmath246 , where @xmath247 is the stress necessary to break all the chemical bonds . for a relevant frequency in the microwave range
@xmath248 we have @xmath249
. we can convert this into the neutron production rate through the fermi interaction @xmath250 \sim\ \left[\frac{g_f m^2}{\hbar c}\right]^2 \left[\frac{mc^2}{\hbar}\right ] \gamma^2\\ \sim
0.6\ hz\ { \rm for}\ \gamma = 30.\end{aligned}\ ] ] the transition rate per unit time per unit area of micro - crack surface may be written as @xmath251 , \ ] ] where @xmath252 is the number of protons per unit area in the first few layers of the quartz granite .
typical values are @xmath253 a neutron production rate sufficiently large to be measured and to cause even other nuclear transformations . incidentally , in the piezoelectric case , the neutrons are not ultra cold .
traditional `` hot fusion '' programs have been able to generate neutrons at total rate of about @xmath254 in the ( @xmath255 ) and ( @xmath256 ) modes , unfortunately lasting only for about 2 seconds @xcite .
thus , great improvements are still required for a steady state running of the tokamaks along with the curtailment of the generated radioactivity . as emphasized by byrne@xcite
, fusion reactors would require a `` biological shield '' , as their outermost shell made of some heavy element such as lead , whose purpose is to remove the surviving @xmath257 neutrons and to attenuate the @xmath79 ray flux .
but , this would generate a substantial amount of undesirable radioactivity@xcite .
in fact , byrne concludes that `` _ it is not immediately obvious that the problems of radioactive waste disposal , which have plagued fission power programmes across the world , will be entirely absent in the thermonuclear power plants of the future . _ '' in view of the formidable scientific and technical problems with hot fusion programs yet to be resolved , it is pertinent to ask whether tokomaks , machines constructed at great public cost , may be put to a different but important practical use .
given the extremely large electromagnetic fields created with great ingenuity within these devices , we propose to use it as a proof of concept machine for electro - weak lent by running electrons and protons rather than just deuterons and tritons as has been done so far .
it may be difficult at the beginning to adapt a tokomak for this purpose but our proposal seems feasible .
the leitmotif of this paper has been to provide direct evidence that low energy nuclear transmutations [ lent ] can and have been induced by gravitational , weak , em and strong interactions both in nature as well as in the laboratory . also , that quite often three or four of the fundamental interactions are needed to lead to the final lent . due to this interplay of different interactions each with its own widely different coupling strength ,
one finds a great diversity in the rates of the reactions .
hence , we may conclude that the reality of lent is no longer a matter of debate because it is soundly based on the standard model.the focus must now be on the employment of technology to realize electroweak lent and thereby provide sorely needed practical applications of the standard model of particle physics .
extraordinary advances already made in realizing up to @xmath148 table top electron accelerators and high gain factors in generating electric fields in surface plasmon polariton resonances on nano particles bode well for the future of electroweak lent .
ys would like to thank his colleagues at the u. of rome la sapienza , pisa , perugia , naples , boston and at cern for inviting him for seminars on this and related topics .
special thanks are due in particular to professor n. cabibbo , who during his last years greatly encouraged ys to pursue ideas presented here and to , dr . r. chidambaram who has been an ever patient and sympathetic listener .
a. v. gurevich , v. p. antonova , a. p. chubenko , a , n. karashtin , g. g. mitko , m. o. ptitsyn , v. a. ryabov , a. l. sheptov , yu . v. shlyugaev , l. i. vildanova , and k. b. zybin , _ phys
lett . _ * 108 * , 125001 ( 2012 ) .
see , for example : r. p. ashley , g. l. kulcinski , j. f. santarius , s. krupakar murali , g. piefer ; ieee publication 99ch37050 , pg .
35 - 37 , 18th symposium on fusion engineering , albuquerque nm , 25 - 29 october 1999 .
v. a. klyuev , a. g. lipson , yu .
p. toporov , b. v. deryagin , v. j. lushchikov , a.v .
streikov , e. p. shabalin , _ sov .
* 12 * , 551 ( 1986 ) ; v. klyuev _ et al .
_ , _ kolloidn .
* 88 * , 1001 ( 1987 ) . | employing concrete examples from nuclear physics it is shown that low energy nuclear reactions can and have been induced by all of the four fundamental interactions ( i ) ( stellar ) gravitational , ( ii ) strong , ( iii ) electromagnetic and ( iv ) weak .
differences are highlighted through the great diversity in the rates and similarity through the nature of the nuclear reactions initiated by each . |
since the singularity confinement criterion was introduced as a discrete analogue of the painlev test @xcite , many discrete analogues of painlev equations have been proposed and extensively studied @xcite .
discrete painlev equations have been considered as 2-dimensional non - autonomous birational dynamical systems which satisfy this criterion and which have limiting procedures to the ( continuous ) painlev equations . in recent years
it was shown by sakai that all of these ( from the point of view of symmetries ) are obtained by studying rational surfaces in connection with extended affine weyl groups @xcite .
on the other hand , recently kajiwara _ et al _ ( kny ) @xcite have proposed a birational representation of the extended weyl groups @xmath9 on the field of rational functions @xmath10 , which is expected to provide higher order discrete painlev equations ( however , this representation is not always faithful , for example it is not faithful in the case where @xmath11 or @xmath12 equals @xmath13 and in the case of @xmath14 ) . in the case of @xmath1 and @xmath15 ,
the actions of the translations can be considered to be 2-dimensional non - autonomous discrete dynamical systems and therefore to correspond to discrete painlev equations .
special solutions and some properties of these equations have been studied by several authors @xcite . in the case of @xmath1 and @xmath2 ,
the action of the translation was thought to be a symmetric form of the @xmath6-discrete analogue of painlev v equation ( @xmath6-@xmath16 ) .
however , the symmetry @xmath17 does not coincides with any symmetry of discrete painlev equations in sakai s list , ( in the case of @xmath1 and @xmath18 , it coincides with an equation , which is associated with a family of @xmath3 surfaces and whose symmetry is @xmath19 , in sakai s list ) .
so it is natural to suspect that the symmetry might be a subgroup of a larger group associated with some family of rational surfaces . in this paper
we show that in the case of @xmath1 and @xmath2 the action of the translation can be lifted to an automorphism of a family of rational surfaces of the type @xmath3 , i.e. surfaces such that the type of the configuration of irreducible components of their anti - canonical divisors is @xmath3 , and therefore that the group of these automorphisms is @xmath4 ( hence it is not @xmath6-@xmath16 by sakai s classification ) .
the action can be decomposed into two mappings which are conjugate to the @xmath6-@xmath7 equation .
it is also shown that the subgroup of automorphisms which commute with the original translation is isomorphic to @xmath20 .
the birational representation of @xmath21 on @xmath22 proposed by kajiwara _
et al _ ( kny ) @xcite is an action on @xmath22 ( @xmath23 , @xmath24 and the indices @xmath25 are considered in modulo @xmath26 respectively ) defined as follows .
we write the elements of the weyl group corresponding to the simple roots as @xmath27 and the elements corresponding to the rotations of the dynkin diagrams as @xmath28 the action of these elements on @xmath22 are defined as & & r_i(x_ij)=x_i+1,j , r_i(x_i+1,j)=x_ij , r_k(x_ij)=x_ij , ( k i , i-1 ) , + & & s_j(x_ij)=x_i , j+1 , s_j(x_i , j+1)=x_ij , s_k(x_ij)=x_ij , ( k j , j-1 ) , + & & ( x_ij)=x_i+1,j , ( x_i , j)=x_i , j+1 , where p_ij=_a=0^n-1 ( _ k=0^a-1 x_i , j+k+1 _ k = a+1^n-1 x_i+1,j+k+1 ) , & & q_ij=_a=0^m-1 ( _ k=0^a-1 x_i+k+1,j _ k = a+1^m-1 x_i+k+1,j+1 ) .
for example in the @xmath29 case , @xmath30 and @xmath31 .
it was shown by kny that this action is a representation of @xmath21 as automorphisms of the field @xmath22 .
but it is still an open problem when this representation is faithful . in the case of @xmath1 and @xmath15 ,
one can see it is faithful by considering the actions on the root systems which we discuss later . in the case of @xmath29 ,
the variable transformation : ( ) ^1/2=a_j , & & ( ) ^1/2=f_j , reduces the actions of @xmath32 to & & r_0(a_i)=a_i , + & & r_0(f_i)= + & & ( a_i)=a_i , ( f_i)= + & & r_1=r_0 + & & s_i(a_j)=a_ja_i^-c_i , j , s_i(f_j)=f_j ( ) ^u_i , j + & & ( a_i)=a_i+1 , ( f_i)=f_i+1 where @xmath33 and @xmath34 are ( c_i , j)= ( rrrr 2 & -1 & 0 & -1 + -1 & 2 & -1 & 0 + 0 & -1 & 2 & -1 + -1 & 0 & -1 & 2 ) , & & ( u_i , j)= ( rrrr 0 & 1 & 0 & -1 + -1 & 0 & 1 & 0 + 0 & -1 & 0 & 1 + 1 & 0 & -1 & 0 ) . by the variable transformation we have @xmath35 ,
but if we remove this constraint and set @xmath36 , the actions also generate @xmath21 .
+ the element @xmath37 is a translation of @xmath38 and provides a discrete dynamical system : c |a_0=a_0 , |a_1=a_1 , |a_2=a_2 , |a_3=a_3 , + [ qp5 ] contrary to the case where these mappings are considered to be field operators , we define the composition of mappings as that of functions . for example , for @xmath39 and @xmath40 , we have @xmath41 .
+ by the change of variables a_3=1/(a_0 a_1 a_2 q ) , & & f_0=x , f_1=y , f_2=c / x , f_3=d / y this equation reduces to the following 2-dimensional non - autonomous discrete dynamical system : : ( x , y ) & & ( , ) , [ qp6 ] \ { + , . where the change in the parameters is given by & & ( , , , , , ) = ( a_0,a_1,a_2,q , , ) .
the notion of space of initial conditions ( values ) was first proposed by okamoto @xcite for the continuous painlev equations and was recently applied by sakai @xcite for the discrete painlev equations . in the discrete case
it is linked to automorphisms of certain families of rational surfaces .
the relations of surfaces and groups of these automorphisms were also studied by many authors from the algebraic - geometric view point @xcite . in this section , following sakai s method , we construct the space of initial conditions for @xmath42 .
let @xmath43 and @xmath44 be rational surfaces and let @xmath45 and @xmath46 be surfaces obtained by the successive blow - ups @xmath47 and @xmath48 .
a rational mapping @xmath49 is called _ lifted _ from a rational mapping @xmath50 , if @xmath51 holds for any point where @xmath52 and @xmath53 are defined . + ( space of initial conditions ) .
let @xmath54 s be rational surfaces and let @xmath55 be a sequence of birational mappings . a sequence of rational surfaces @xmath56 is ( or @xmath57 themselves are ) called the space of initial conditions for @xmath58 if each @xmath59 is lifted to an isomorphism , i.e. bi - holomorphic mapping , from @xmath57 to @xmath60 .
+ let us consider the mapping ( [ qp6 ] ) to be a birational mapping from @xmath61 to itself . blowing up successively at the indeterminate points of @xmath42 and @xmath62 ( as usual ) , we have the surface @xmath54 as described in fig.[a3 ] .
the total transforms of the points of successive blow - ups are @xmath63 where we use @xmath64 instead of @xmath6 .
the mapping @xmath65 ( [ qp6 ] ) can be lifted to the sequence of isomorphisms @xmath66 .
let @xmath43 be a rational surface which is given by blow - ups ( [ blowup ] ) from @xmath61 , where @xmath67 .
we denote the linear equivalence class of total transforms of @xmath68 ( @xmath69 ) on @xmath43 by @xmath70 ( resp .
@xmath71 ) .
the picard group of the surface @xmath43 , i.e. the group of linear equivalence classes of divisors on @xmath43 , is denoted as @xmath72 .
the mapping @xmath42 induces the push - forward mapping @xmath73 ( @xmath74 ) as h_0 2h_0+h_1-e_2-e_4-e_6-e_8 & + h_1 h_0 + 2h_1-e_2-e_4-e_6-e_8 & + e_1 h_0+h_1-e_2-e_4-e_8 & e_2 e_5 + e_3 h_0+h_1-e_2-e_4-e_6 & e_4 e_7 + e_5 h_0+h_1-e_4-e_6-e_8 & e_6 e_1 + e_7 h_0+h_1-e_2-e_6-e_8 & e_8 e_3 .
+ in this section we shall consider automorphisms of the family of @xmath43 with various parameters .
the picard group of @xmath43 is the @xmath75-module ( lattice ) @xmath76 the intersection number of any two divisors can be calculated by using the intersection form @xmath77 let @xmath78 be divisors as ll d_0=h_0-e_1-e_2 , & d_1=h_1-e_3-e_4 + d_2=h_0-e_5-e_6 , & d_3=h_1-e_7-e_8 .
the anti canonical divisor of @xmath43 , @xmath79 , is uniquely decomposed into prime divisors ( for generic parameters ) as @xmath80 the connections of the @xmath81 are expressed by the following diagram .
@xmath82 such rational surfaces are said to be of the @xmath3 type which is the type of those surfaces in sakai s list associated with the @xmath6-@xmath7 equation .
an automorphism @xmath83 of @xmath72 is called a _ cremona isometry _
@xcite if @xmath83 preserves i ) the intersection form on @xmath72 , ii ) the canonical divisor @xmath84 , iii ) the semigroup of effective classes of divisors .
+ let @xmath43 and @xmath45 be rational surfaces . in general , if a mapping @xmath42 is an isomorphism from @xmath43 to @xmath45 , its action on the picard group @xmath85 ( or @xmath86 ) ( @xmath87)is always a cremona isometry .
+ a cremona isometry preserves the set @xmath88 and its orthogonal ( with respect to the intersection form ) lattice , [ q ] & & q:= z_0 _ 1_2_3_4 _ 5 where the @xmath89 s are ll _ 0=e_5-e_6 , & _
1=e_1-e_2 + _ 2=h_1-e_1-e_5 , & _ 3=h_0-e_3-e_7 + _ 4=e_3-e_4 , & _ 5=e_7-e_8 .
+ we define the action @xmath90 on @xmath72 corresponding to @xmath89 as @xmath91 where @xmath92 .
the cartan matrix is as follows .
( c_i , j)&= & ( rrrrrr 2 & 0 & -1 & 0 & 0 & 0 + 0 & 2 & -1 & 0 & 0 & 0 + -1 & -1 & 2 & -1 & 0 & 0 + 0 & 0 & -1 & 2 & -1 & -1 + 0 & 0 & 0 & -1 & 2 & 0 + 0 & 0 & 0 & -1 & 0 & 2 ) .
( 200,30)(50,30 ) ( 100,20)(1,0)20 ( 125,20)(1,0)20 ( 150,20)(1,0)20 ( 122.5,22.5)(0,1)20 ( 147.5,22.5)(0,1)20 ( 97.5,20 ) ( 122.5,20 ) ( 147.5,20 ) ( 172.5,20 ) ( 122.5,45 ) ( 147.5,45 ) ( 97.5,10)(0,0)@xmath93 ( 122.5,10)(0,0)@xmath94 ( 147.5,10)(0,0)@xmath95 ( 172.5,10)(0,0)@xmath96 ( 135,50)(0,0)@xmath97 ( 160,50)(0,0)@xmath98 we also define the actions of the generators of automorphisms of the dynkin diagrams @xmath99 and @xmath100 as _ 102345 : & & ( h_0,h_1,e_1,e_2,e_3,e_4,e_5,e_6,e_7,e_8 ) + & & ( h_0,h_1,e_5,e_6,e_3,e_4,e_1,e_2,e_7,e_8 ) + _ 543210 : & & ( h_0,h_1,e_1,e_2,e_3,e_4,e_5,e_6,e_7,e_8 ) + & & ( h_1,h_0,e_3,e_4,e_1,e_2,e_7,e_8,e_5,e_6 ) , where the suffixes of @xmath101 mean corresponding permutations
. for convenience we also use the automorphism @xmath102 . for simplicity
we write @xmath99 and @xmath103 as @xmath104 and @xmath105 respectively . here , the group of automorphisms of the lattice @xmath106 which preserve the intersection form ( considered to be the inner product ) is @xmath107 ( see @xcite 5.10 ) .
since each element of @xmath108 does not preserve effectiveness of divisor classes , we have the following theorem ( it can be confirmed directly that @xmath4 does preserve effectiveness by realization as automorphisms of the family of surfaces ) .
[ cremona ] the group of cremona isometries of the surface @xmath43 is @xmath4 .
the mapping @xmath109 acts on the root basis as & & ^2:(_0,_1,_2,_3,_4,_5 ) + & & ( _ 0+k,_1+k,_2-k , _ 3-k,_4+k,_5+k ) where @xmath110 corresponds to the canonical central element and @xmath111 is written by generators as & = & _ 54_10w_5w_4 w_1w_0w_2w_3w_2 . since we have defined the composition of mappings as in sect . 2 ,
the composition of actions on the picard group should like a changes of basis , i.e. for @xmath112 and @xmath40 , we have @xmath113
. + each cremona isometry of th.[cremona ] can be realized as birational mapping on @xmath61 which is lifted to an automorphism of the family of surfaces @xmath43 ( cf .
@xcite ) .
the birational actions of generators are as follows : @xmath114 is mapped to & ( x , y ; a_0 , , a_2 , , c , ) & + & ( x , y ; a_0 , , a_2 , , c , ) & + & ( , ; a_0 , , a_2 , , c , ) & + & ( , ; , a_1 , , a_3 , , d ) & + & ( x , y ; , a_1 , , a_3 , , d ) & + & ( x , y ; , a_1 , , a_3 , , d ) & + & ( , ; , , , , , ) & + & ( , ; , , , , , ) & .
if one would prefer to relate these transformation to sakai s paper @xcite explicitly , it is sufficient to put @xmath115 and @xmath116 as @xmath117 and @xmath118 .
( this coincidence is up to the coefficients of @xmath119 and @xmath120 in each birational transformation , because the normalization is different in sakai s paper . while there are 8 parameters in that paper , there are only 6 parameters here .
if we consider surfaces whose picard group and effective classes are the same as those of the surfaces @xmath43 and normalize by automorphisms of @xmath61 , then only `` 6 '' parameters remain . )
we define the mapping @xmath121 as @xmath122 , then @xmath121 is expressed as : ( x , y ) & & ( , ) , \ { + . with the change of parameters
: & & ( , , , , , ) = ( , q a_1 , , q , d , c ) .
the mapping @xmath123 is the @xmath6-@xmath7 equation .
actually it acts on the root basis as & & ( _ 0,_1,_2,_3,_4,_5 ) ( _ 0,_1,_2+k,_3-k,_4,_5 ) .
if one would prefer to relate these transformation to sakai s @xmath124 @xcite , it is sufficient to put @xmath117 and @xmath125 and @xmath126 and @xmath127 ( the normalization is changed by introducing the new variable @xmath128 ) .
the mapping @xmath129 can be decomposed into conjugate mappings of @xmath6-@xmath130 as & & ^2=(_543210w_2 w_2 _ 543210 ) ( w_2 w_2 ) .
in this section we investigate the subgroup of @xmath4 whose elements commute with @xmath109 ( we denote it as @xmath131 ) .
we prove the following theorem .
@xmath132 holds , where the meaning of each symbol is : @xmath133 is @xmath134 ; @xmath135 is the extended weyl group defined by the root basis @xmath136 and the automorphisms of dynkin diagram @xmath137 and @xmath138 ; it coincides with the original @xmath139 except for the extension ; @xmath140 is the extended weyl group defined by the root basis @xmath141 and the automorphisms of dynkin diagram @xmath142 .
notice that the subset of the lattice @xmath143 ( [ q ] ) whose elements are preserved by @xmath129 is @xmath144 the lattice @xmath145 can be described as q_b&=&span_(_0,_1,_2,_3,_0 ) + & = & span_(_0,_1,_2,_0,_1 ) .
[ lemma1 ] the group @xmath131 preserves the lattice @xmath145 . let @xmath146 and let @xmath147 . since @xmath148 , @xmath149 is preserved by @xmath129 and therefore @xmath150 .
[ lemma2 ] the group whose actions preserve the lattice @xmath145 and the intersection form is @xmath151 .
first we consider @xmath152 .
@xmath152 is written as @xmath153 where @xmath154 . from the equation -2 & = & _
i_i= w(_i)w(_i ) + & = & -(b_i,0-b_i,1)^2-(b_i,1-b_i,2)^2-(b_i,2-b_i,3)^2- ( b_i,3-b_i,0)^2 - 2(c_i,0)^2 , we have @xmath155 or @xmath156 . in the former case we have that @xmath157 which is a contradiction .
hence @xmath158 preserves the lattice @xmath159 and therefore the action of @xmath158 on this lattice coincides with an element of the extended affine weyl group @xmath160 ( see @xcite 5.10 ) .
next we consider @xmath161 and @xmath162 .
@xmath163 is written as @xmath164 and we have @xmath155 or @xmath156 for each @xmath165 . since the rank of @xmath166 equals that of @xmath145 , @xmath167 can not be zero .
hence we have @xmath168 and therefore the action of @xmath158 on the lattice @xmath169 coincides with an element of the extended affine weyl group @xmath170 .
+ now we are ready to prove the theorem .
+ by lemma [ lemma1 ] and lemma [ lemma2 ] it follows that the action of @xmath171 on @xmath145 is given by @xmath172 .
one of the extensions of this action onto @xmath143 is given in appendix [ app2 ] .
we investigate the variety of extensions .
suppose the actions of @xmath173 on the lattice @xmath145 are the same .
we write @xmath174 as @xmath175 .
since @xmath176 are linearly independent , @xmath177 can be written as @xmath178 using the relations @xmath179 and @xmath180 , we have \
{ -2b_0+b_1+b_3&=&0 + a_0+b_0 - 2b_1+b_2&=&1 + b_1 - 2b_2+b_3&=&0 + -a_0+b_0+b_2 - 2b_3&=&-1 + -a_0 - 2c_0&=&-1 . + . and hence & = & _ 0+k+(2_0+_0 + 2_1+_2-_0 ) + & = & _ 0+k+(_2+_3 ) .
moreover from @xmath181 , we have & = & _ 0+k _ 0+_2+_3+k . since @xmath182 and @xmath183
, we have : if @xmath184 ( @xmath185 ) , then & & ( , , , , , ) + & = & ( _ 0 + z k , _ 1 + z k , _ 2 - z k , al_3 - z k , _ 4 + z k , _ 5 + z k ) and if @xmath186 ( @xmath185 ) , then & & ( , , , , , ) + & = & ( _ 0 + _ 2+_3 + z k , _ 1 + _ 2+_3 + z k , -_3 - z k , + & & -_2 - z k , _ 2+_3 + _ 4 + z k , _ 2+_3+_5 + z k ) . in the former case @xmath187 and in the later case @xmath188 . since @xmath189 does not commute with @xmath129 ( @xmath190 ) , the action of @xmath191 is extended by @xmath134
. + _ acknowledgment . _
the author would like to thank h. sakai , j. satsuma and r. willox for discussions and advice .
the author is also grateful to the referee for his useful comments and suggestions .
by studying the actions of the elements of original weyl group @xmath0 on the root basis , we can realize its embedding into the @xmath4 as follows r_0 & = & w_2w_3 w_2 + r_1 & = & w_0w_1 w_4 w_5
w_2 w_3 w_2 w_0w_1 w_4 w_5 + & = & w_0 w_1 w_4 w_5 _ 10 _ 54 + s_0 & = & w_5 w_3 w_5 + s_1 & = & w_1 w_2 w_1 + s_2 & = & w_3 w_4 w_3 + s_3 & = & w_0 w_2 w_0 + & = & _ 543210 _ 10 .
as mentioned in section [ comm ] the actions of @xmath172 on the lattice @xmath145 are not uniquely expanded to onto the lattice @xmath143 .
the variety of extensions is @xmath192 . here
, we give one of the embeddings of @xmath172 into @xmath4 . for @xmath193 or @xmath194
we define the action on the lattice @xmath143 as @xmath195 where @xmath196 . in order to find its description in terms of the fundamental elements of @xmath4
, we use the relation w_w_()=w_w_w _ , where @xmath197 and @xmath198 are real roots .
w__0 & = & w__3+_5=w_3w_5w_3 + w__1 & = & w__1+_2=w_1w_2w_1 + w__2 & = & w__3+_4=w_3w_4w_3 + w__3 & = & w__0+_2=w_0w_2w_0 + w__0 & = & w__0+_1+_2+_3= w_3w_0w_2w_1w_2w_0w_3 + w__1 & = & w__2+_3+_4+_5 = w_2w_5w_3w_4w_3w_5w_2 + finally , we define the action of automorphisms of the dynkin diagram of @xmath172 : @xmath199 as @xmath200 is mapped to _ ( 1032 ) : & & ( _ 0+_1+_2+_3+_4 , _ 0+_1+_2+_3+_5 , -_0-_1-_2 , + & & -_3-_4-_5 , _ 0+_2+_3+_4+_5 , _
1+_2+_3+_4+_5 ) + _ ( 3210 ) : & & ( _ 0+_1+_2+_3+_5 , _ 0+_1+_2+_3+_4 , -_0-_1-_2 , + & & -_3-_4-_5 , _ 1+_2+_3+_4+_5 , _
0+_2+_3+_4+_5 ) + _ ( 10 ) : & & ( _ 0+_2+_3+_4+_5 , _ 1+_2+_3+_4+_5 , -_3-_4-_5 , + & & -_0-_1-_2 , _ 0+_1+_2+_3+_4 , _ 0+_1+_2+_3+_5 ) . hence ,
these are written in termes of the fundamental elements of @xmath4 as : _ ( 1032)&=&_543210 w_5w_4 w_1w_0w_2w_3w_2 + _ ( 3210)&=&_543210_54_10 w_5w_4 w_1w_0w_2w_3w_2 + _ ( 10)&=&= _ 54_10 w_5w_4 w_1w_0w_2w_3w_2 . | we study the birational representation of @xmath0 proposed by kajiwara - noumi - yamada ( kny ) in the case of @xmath1 and @xmath2 . it is shown that the equation can be lifted to an automorphism of a family of @xmath3 surfaces and therefore the group of cremona isometries is @xmath4 ( @xmath5 ) .
the equation can be decomposed into two mappings which are conjugate to the @xmath6-@xmath7 equation .
it is also shown that the subgroup of cremona isometries which commute with the original translation is isomorphic to @xmath8 .
graduate school of mathematical sciences , university of tokyo , komaba 3 - 8 - 1 , meguro - ku , tokyo 153 - 8914 , japan + |
a thermal counterflow in he ii is internal convection of two fluids , namely the normal fluid and the superfluid .
when the counterflow velocity exceeds a critical value , a self - sustaining tangle of quantized vortices appears to form superfluid turbulence . in low aspect ratio channels ,
superfluid turbulence makes the mysterious transition .
the increase in the counterflow velocity is observed to change the laminar state to the first turbulent state ti , and next to the second turbulent state tii @xcite .
melotte and barenghi suggested that the transition from the ti to tii state is caused by the transition of the normal fluid from laminar to turbulent @xcite .
the recent developments of the visualization technique have enabled us to confirm the scenario .
guo @xmath0 have followed the motion of seeded metastable @xmath1 molecules by a laser - induced - fluoresence technique to observe that the normal fluid can be turbulent at relatively high velocities @xcite . to understand the mysterious transition of counterflow quantum turbulence , it is necessary to address the coupled dynamics of the two fluids ;
the superfluid is described by the vortex filament model , the normal fluid by the navier - stokes equation , and they are coupled through the mutual friction @xcite . however , it is difficult to solve fully the coupled dynamics . as the first essential step , we address the ti state in a square channel with prescribing the velocity field of the normal fluid to a poiseuille profile .
our simulation obtains a statically steady state of an inhomogeneous vortex tangle .
baggaley @xmath0 @xcite studied numerically a thermal counterflow between two plates .
they prescribed a poiseuille and turbulent profiles for the velocity field of the normal fluid .
an inhomogeneous vortex tangle was obtained , where vortices concentrated near the solid boundaries .
they suggested that their results supported the scenario proposed by melotte and barenghi .
the better understanding of the ti and tii states would be obtained by studying the flow in a low aspect ratio channel where all boundaries are solid except for the flow direction .
this is because the ti and tii states are actually observed in low aspect ratio channels and another turbulence tiii state is observed in high aspect ratio channels @xcite .
in a vortex filament model @xcite a quantized vortex is represented by a filament passing through a fluid and has a definite vorticity .
this approximation is very suitable in he ii , since the core size of a quantized vortex is much smaller than any other characteristic length scale . at zero temperature
the vortex filament moves with the superfluid velocity @xmath2 where @xmath3 is the velocity field produced by vortex filaments , @xmath4 by solid boundaries , and @xmath5 is the applied superfluid velocity .
the velocity field @xmath3 is given by the biot - savart law ; this work addresses the full biot - savart integral @xcite .
since @xmath5 represents the laminar flow of the superfluid , @xmath5 requires irrotational condition , which is supposed to be uniform .
the velocity field @xmath4 is obtained by a simple procedure ; it is just the field produced by an image vortex which is constructed by reflecting the filament into the surface and reversing its direction .
taking into account the mutual friction , the velocity of a point @xmath6 on the filament is given by @xmath7,\ ] ] where @xmath8 and @xmath9 are the temperature - dependent coefficients , and the prime denotes derivatives of @xmath10 with respect to the coordinate @xmath11 along the filament .
the velocity field of the normal fluid is prescribed to be a poiseuille profile . in a rectangular channel
the poiseuille profile is given by @xmath12 \frac { \cos(m \pi y / 2 a ) } { m^3 } , \ ] ] where @xmath13 and @xmath14 are coordinates normal to the flow direction @xmath15 , and @xmath16 and @xmath17 are halves of the channel width along the @xmath13- and @xmath14- axes @xcite .
in this study , all simulations are performed under the following conditions .
we study thermal counterflow of he ii at temperatures _
k , 1.6 k and 1.9 k. the computing box is @xmath18 .
periodic boundary conditions are used along the flow direction @xmath15 , while solid boundary conditions are applied to the channel walls at @xmath19 and @xmath20 .
all simulations start with eight randomly oriented vortex rings of radius @xmath21 .
the vortex line density ( vld ) is defined as @xmath22 , where the integral is performed along all vortices in the sample volume @xmath23 .
the vortex tangle reaches the statistically steady state .
figure 1 ( a ) shows the time evolution of vld .
fluctuations are larger than those in a uniform counterflow @xcite , which is attributable to the mechanism characteristic of this system discussed in section 4.1 . .
( b ) vortex line density averaged over the statistically steady state as a function of @xmath24 .
, title="fig:",scaledwidth=100.0% ] ( a ) .
( b ) vortex line density averaged over the statistically steady state as a function of @xmath24 .
, title="fig:",scaledwidth=100.0% ] ( b ) the statistically steady state is known to exhibit the characteristic relation @xmath25 @xcite with the parameters @xmath26 and @xmath27 .
we regard the counterflow velocity @xmath24 as the spatially averaged amplitude of @xmath28 .
figure 1 ( b ) shows the vld temporally averaged over the statistically steady state , which almost satisfies the relation .
table 1 shows the comparison of @xmath26 among the present work @xmath29 , the simulation @xmath30 under the periodic boundary condition @xcite and a typical experiment @xmath31 @xcite .
the values of @xmath29 are lower than the values of @xmath30 obtained under the uniform counterflow .
the difference of @xmath26 comes from the difference of the mechanism sustaining the vortex tangle .
the origin of the discrepancy between @xmath29 and @xmath31 is not clear , but this may be attributable to neglecting the effect of the vortex tangle on the poiseuille flow of the normal fluid through mutual friction . .
line density coefficients @xmath26 .
numerical results @xmath29 by this work , @xmath30 by adachi
_ @xcite and experimental results @xmath31 by childers and tough @xcite . [ cols="^,^,^,^ " , ] in order to estimate the inhomogeneity of the vortex tangle , we divide the computational box to @xmath32 sub - volumes and define the vld at a sub - volume as the local vld .
figure 2 ( a ) shows the spatially dependence of @xmath33 , which is obtained by averaging the local vld spatially over the flow direction and temporally over the statistically steady state .
we estimate the inhomogeneity of the vortex tangle by a spatial variance @xmath34 of @xmath35 .
figure 2 ( b ) shows the characteristic dependence of @xmath34 on @xmath24 and @xmath36 .
firstly , @xmath34 of 1.6k is the largest among three temperatures .
secondly , @xmath34 at 1.3k and 1.6k increases with @xmath24 , while @xmath34 at 1.9 k decreases with @xmath24 . the dependence on @xmath36 is understood as discussed in section 4.2 , but the dependence on @xmath24 is not known . ) .
one can see the concentration near the solid boundaries .
( b ) dependence of the spatially variance @xmath34 of @xmath35 on @xmath24 and @xmath36 .
, title="fig:",scaledwidth=100.0% ] ( a ) ) .
one can see the concentration near the solid boundaries .
( b ) dependence of the spatially variance @xmath34 of @xmath35 on @xmath24 and @xmath36 .
, title="fig:",scaledwidth=100.0% ] ( b )
section 4.1 addresses the mechanism for sustaining the inhomogeneous vortex tangle . in section 4.2
we discuss how the vortex tangle becomes inhomogeneous depending on temperature . in section 4.3
we consider how the aspect ratio of the channel cross section affects the two fluids . as shown in fig .
1 ( a ) , the vld shows the non - linear oscillation with large amplitude in the statistical steady state , which is much different from the case of the uniform counterflow @xcite .
the non - linear oscillation comes from the space - time oscillation of the vortex tangle through the mutual friction under the poiseuille flow .
figure 3 shows the space - time pattern of the vortex tangle at 1.9 k. the period of the non - linear oscillation is about 0.6 s , consisting of four stages ( a)-(d ) . in fig .
3 ( a ) corresponding to the minimum of the vld , vortices are dilute , where vortices remain only near the solid walls . then the vortices near the walls invade to the central region in fig . 3 ( b ) .
these vortices make lots of reconnections in the central region subject to the large counterflow in fig .
hence the vld increases significantly to the maximum . eventually in fig .
3 ( d ) the poiseuille flow excludes the vortex tangle from the central region toward the solid walls .
thus the vld around the central region decreases , and the vortices are absorbed by the solid boundaries .
then the vortex tangle returns to the stage of fig .
therefore the vortex tangle repeats the periodic motion , resulting in the non - linear oscillation of fig .
the vortex tangle sustained by this mechanism is more dilute than the case of the uniform counterflow for the same @xmath36 and @xmath24 .
1.9 k , @xmath37 ) , corresponding to the results of fig .
1 ( a ) . , title="fig:",scaledwidth=100.0% ] ( a ) depletion ( 2.00 s ) 1.9 k , @xmath37 ) , corresponding to the results of fig .
1 ( a ) . , title="fig:",scaledwidth=100.0% ] ( b ) invasion ( 2.14 s ) 1.9 k , @xmath37 ) , corresponding to the results of fig .
1 ( a ) . , title="fig:",scaledwidth=100.0% ] ( c ) maximum ( 2.30 s ) 1.9 k , @xmath37 ) , corresponding to the results of fig .
1 ( a ) . , title="fig:",scaledwidth=100.0% ] ( d ) exclusion ( 2.42 s ) temperature dependence of @xmath34 in fig .
2 ( b ) is caused by the temperature dependent parameters , namely the density ratio @xmath38 of the two fluids and the mutual friction coefficients @xmath8 and @xmath9 .
the increase in temperature causes two competitive effects as described in the following .
one effect is that the increase of @xmath39 decreases @xmath40 .
the thermal counterflow requires the conservation of mass , yielding @xmath41 where @xmath42 is spatially averaged @xmath43 over the cross section of the channel .
the ratio @xmath44 increases with temperature .
thus the counterflow velocity @xmath45 becomes more uniform at higher temperature , since @xmath5 is uniform and @xmath46 is non - uniform .
figure 4 shows the temperature dependence of the counterflow velocity profile .
the other effect is that the enhancement of the mutual friction increases @xmath47 .
the mutual friction makes the vortex tangle inhomogeneous under the non - uniform @xmath48 .
therefore the increase in temperature renders the profiles of @xmath48 uniform but enhances this action of the mutual friction .
these two competitive effects maximize the inhomogeneity @xmath47 at some @xmath36 .
( a ) @xmath36=1.3 k ( @xmath39=0.052 )
( b ) @xmath36=1.9 k ( @xmath39=0.738 ) the aspect ratio of the channel plays an important role in the density and the profile of the vortex tangle .
the increase of the aspect ratio from unity reduces the counterflow velocity gradient along the long side of the cross section .
thus the exclusion of the vortex tangle shown in fig .
3 ( d ) does not work so much along the long side of the cross section .
hence the vld should increase compared to the vld in a low aspect ratio channel . it would be meaningful to consider how the aspect ratio affects the normal fluid , though it is prescribed in this formulation .
the linear stability analysis of the navier - stokes equation shows that the critical reynolds number for turbulence transition of a viscous fluid increases significantly with decreasing the aspect ratio @xcite . according to this result
, the normal fluid in counterflow could remain laminar in a low aspect ratio channel even if the superfluid becomes turbulent and the vortex tangle disturbs the normal fluid , which may correspond to the ti state . in order to understand the ti state ,
therefore , the studies in a low aspect ratio channel like this work will be indispensable .
this work was supported by jsps kakenhi grant number 26400366 and mext kakenhi `` fluctuation & structure '' grant number 26103526 .
99 tough j t 1982 _ progress in low temperature physics _ , ed d f brewer ( amsterdam : north - holland ) , vol
. 8 melotte d l and barenghi d f 1998 _ phys .
lett . _ * 80 * 4181 guo w , cahn s b , nikkel j a , vinen w f , and mckinsey d n 2010 _ phys .
lett . _ * 105 * 045301 kivotides d 2007 _ phys .
_ b * 76 * 054503 baggaley a w and laizet s 2013 _ phys
. fluids _ * 25 * 115101 schwarz k w 1985 _ phys .
_ b * 31 * 5782 adachi h , fujiyama s and tsubota m 2010 _ phys . rev .
_ b * 81 * 104511 1998 _ the handbook of fluid dynamics _ , ed r w johnson ( boca raton : crc press ) vinen w f 1957 _ proc .
london _ ser .
a * 242 * 493 childers r k and tough j t 1976 _ phys .
_ b * 13 * 1040 tatsumi t and yoshimura t 1990 _ j. fluid mech .
_ * 212 * 437 | we perform a numerical analysis of superfluid turbulence produced by thermal counterflow in he ii by using the vortex filament model .
counterflow in a low aspect ratio channel is known to show the transition from laminar flow to the two turbulent states ti and tii .
the present understanding is that the ti has the turbulent superfluid and the laminar normal fluid but both fluids are turbulent in the tii state .
this work studies the vortex tangle in the ti state .
solid boundary condition is applied to walls of a square channel , and the velocity field of the normal fluid is prescribed to be a laminar poiseuille profile .
an inhomogeneous vortex tangle , which concentrates near the solid boundaries , is obtained as the statistically steady state .
it is sustained by its characteristic space - time oscillation .
the inhomogeneity of the vortex tangle shows the characteristic dependence on temperature , which is caused by two competitive effects , namely the profile of the counterflow velocity and the mutual friction . |
long - lived quantum coherence been observed in photosynthesis after laser excitation @xcite .
it has attracted much attention on how quantum coherence could be enhanced in complex biological environment and how it may play a key role in efficient exciton transports @xcite .
it is well known that the photon - to - charge conversion quantum efficiency of photosynthesis in plants , bacteria , and algae can be almost 100% under certain conditions .
while photosynthesis converts sunlight into chemical energy , solar cell converts sunlight into electric energy . according to shockley and queisser ,
the efficiency of photovoltaic energy conversion is limited to 33% , based on the energy band gap and solar spectrum , due to the radiative recombination of electron - hole pairs , thermalization , and unabsorbed photons @xcite .
various attempts have been made to improve the performance of photovoltaic devices @xcite .
mimicking photosynthesis presents a promising route by which to increase the efficiency of the current solar cell technology @xcite .
consequently , there has been a long - standing and increasing interest in the understanding of the physics describing the energy conversion within photosynthesis .
recently , quantum coherence has demonstrated its crucial role in the energy conversion during photosynthesis @xcite .
similarly , it has been shown that quantum coherence can be used to alter the conditions of the detailed balance and thereby enhance the quantum efficiency in photocell @xcite . in principle , the shockley - queisser model is a two - extended - level model . by incorporating more levels and tuning them carefully
, the conversion efficiency can be improved .
recently , creatore _
et al . _
@xcite have shown that the delocalized quantum state is capable of improving the photocurrent of a photocell by at least 35% in compared with a photocell with the localized quantum state when treating the photon - to - charge conversion as a continuous carnot - like cycle @xcite . within their model , the two delocalized states , called the bright and dark states , of the dipole - dipole interacting two donors play a key role in improving the efficiency of the pv cell .
due to the constructive interference , the optical transition rate between the ground and the bright states becomes two times stronger than the uncoupled donor case .
while it is blocked through the bright state due to the destructive interference , the electron transition from the excited donor to the acceptor is made only through the dark state and its rate is two times larger than the uncoupled donor case , due to the constructive interference .
consequently , the presence of quantum coherence of the delocalized donor states alters the conditions for the thermodynamic detailed balance ; that results in the enhancement of the efficiency of the photocell . in this paper , we show that a photocell with three suitably arranged electron donors coupled via dipole - dipole interactions can result in an enhancement of photocurrents by about 49.5% over a classical photocell . while inspired by creatore _
et al . _
@xcite , our three coupled donors , rather than the two coupled ones , makes another big improvement in the efficiency of a pv cell .
the origin of the photocurrent enhancement is explained by the key roles of the delocalized excited states of the donor system .
the dipole - dipole coupling between donors make three degenerate and localized one - exciton levels split into three delocalized levels : the bright , almost - dark , and dark states .
the photon absorption and emission rates between the ground and the bright excited state becomes about 2.91 times larger than that of the uncoupled donor case , which is due to the constructive interference of three donors . while the electron transferring from the donor to the acceptor through the almost - dark state
is enhanced by about 2.91 times compared to the uncoupled donor case , but is almost blocked through the bright state , which are also due to the constructive and destructive interferences of the delocalized donor states
. basically , essential physics of our triple - donor model is similar to that of creatore _
s two donor model , but more enhanced by collective properties .
while it seems challenging , our proposed model could be realized by nanotechnologies inspired by natural light - harvesting structures .
before introducing a photovoltaic cell model with three donors , we discuss in detail how a configuration of two dipoles moments of two donor affects the efficiency of a pv cell in creatore _
s model @xcite .
the excitation of a molecule is simply modeled as a two - level system with the ground state @xmath0 and excited state @xmath1 .
the optical transition between them is characterized by the optical dipole moment @xmath2 . for a molecular aggregate composed of electric neutral molecules ,
the intermolecular interaction is given by the electrostatic dipole - dipole coupling @xcite @xmath3 where dipole moment @xmath4 is located at @xmath5 , @xmath6 at @xmath7 , and @xmath8 is the radius vector from @xmath4 to @xmath6 .
typically , the strength of @xmath9 is much weaker than the excitation energy @xmath10 .
the exciton dynamics of the aggregate is described by hamiltonian @xcite @xmath11 where @xmath12 and @xmath13 are the pauli raising and lowering operators , respectively . according to eq .
( [ dipole - dipole ] ) the strength of @xmath9 depends on how dipole moments are aligned . in creatore _
s paper @xcite , the donor is a dimer where the dipole moment @xmath4 is always perpendicular to the radius vector @xmath14 so the second term in eq .
( [ dipole - dipole ] ) vanishes .
the dipole - dipole coupling is given by @xmath15 with angle @xmath16 between two dipole moments .
this gives rise to the simple angle - dependence energy gap @xmath17 between the symmetric and antisymmetric excited states .
the spontaneous decay rates are also proportional to @xmath18 . is optically bright and has the absorption and emission rate @xmath19 , but has no electron transition channel to the donor .
the antisymmetric state @xmath20 is dark but has the electron transfer path to the donor .
the electron transition rate @xmath21 between the bright state @xmath22 and dark state @xmath20 is caused by thermal phonons .
, scaledwidth=45.0% ] molecules in aggregates , however , are more probable to be aligned collectively , not independently .
we study how the h and j aggregate donor alignments affect the efficiency of pv cells . as illustrated in fig . 1
, we consider the two dipole moments tilted at the same angle @xmath16 with respect to the vertical axis .
the angle dependence of the dipole - dipole coupling of eq .
( [ dipole - dipole ] ) becomes @xmath23 this implies the angle dependence of the davidov energy splitting @xmath24 between the symmetric and antisymmetric states and explain the transition between the h - aggregate and the j - aggregate at the magic angle @xmath25 when the angle is measured from @xmath14 . here
the angle is measured with respect to the vertical axis so one has the magic angle @xmath26 as shown in fig .
[ mangle ] . between the bright and dark states and
( b ) the current enhancement as a function of the tilt angle @xmath16 with respect to the vertical axis . in ( a ) the alignment of two dipole moments is shown as the blue arrows .
the two parallel dipole moments are aligned in head to head manner ( h - aggregate ) at @xmath27 and in head to tail manner ( j - aggregate ) at @xmath28 . in
( b ) the black arrow points to the magic angle @xmath29.,scaledwidth=50.0% ] in contrast to creatore _ et al . _
s configuration , the symmetric state @xmath30 in our model is always an optically active state ( bright state ) @xcite . for angles less than @xmath31
, this level is higher than the antisymmetric ( dark ) state @xmath32 so the optical transition is shifted to the blue ( h - aggregate ) . on the other hand , for angles greater than @xmath31 , the antisymmetric state is higher so the optical transition is changed to the red ( j - aggregate ) .
note that classically the total dipole moment is always @xmath33 because the two dipole moments point to the same direction .
the dipole matrix element between the ground and bright states is @xmath34 so the optical transition rate @xmath19 , proportional to the square of the dipole matrix element , becomes doubled , i.e , @xmath35 , in compared with an uncoupled donor case .
we calculate how the current enhancement is dependent on the angle @xmath16 , as plotted in fig .
[ mangle ] .
the two energy levels @xmath36 for the symmetric and antisymmetric states move to @xmath37 as the angle @xmath16 increases .
this affects the bose - einstein distributions , @xmath38 of thermal photons , @xmath39 and @xmath40 of thermal phonon through the gaps @xmath41 , @xmath42 , and @xmath43 , respectively .
because of @xmath44 , the distribution @xmath45 is dependent little on the angle @xmath16 .
however , @xmath39 and @xmath40 are strongly affected by the angle @xmath16 so drastic changes in current enhancement . for h - aggregate case @xmath46 , @xmath39 increases but @xmath40 decreases as angle @xmath16 increases .
so the the current enhancement decreases as angle @xmath16 increases . for j - aggregate case @xmath47 ,
the bright state is lower than the dark state .
an electron in the bright state jumps to the dark state via only the absorption @xmath21 of thermal phonons ( in h - aggregate case , the transition from the bright to dark states can be done via the stimulated emission and spontaneous emission of thermal phonons , @xmath48 ) .
thus , the transition from the donor to acceptor is very low , and the current enhancement is negative as shown in fig .
[ mangle ] . in our model as well as
_ s model @xcite , two donors are coupled to the acceptor so the bright state has no electron transferring channel to the acceptor because of the destructive interference .
if a donor system is composed of many molecules ( for example a linear chain ) , it is likely that only some donor molecules ( or the molecules at the end site ) are coupled to the acceptor so the transition path of the bright state to the donor would not be blocked .
the photocell model , proposed here , is depicted in fig . [ model ] .
the picture of a classical cyclic engine is described as the following : @xmath49 , @xmath50 , and @xmath51 represent three identical and initially uncoupled donor molecules which are aligned around an acceptor molecule @xmath52 .
initially , the system starts in the ground state @xmath0 .
the cycle of electron transport begins with the absorption of solar photons populating the uncoupled donor excited states @xmath53 , @xmath54 , and @xmath55 .
then the excited electrons can be transferred to the acceptor molecule , the charge - separated state @xmath56 , with any excess energy radiated as a phonon .
the excited electron is then assumed to be used to perform work , leaving the charge - separated state @xmath56 decaying to the sub - stable state @xmath57 .
the recombination between the acceptor and the donor is also considered with a decay rate of @xmath58 , where @xmath59 is a dimensionless fraction .
this loss channel brings the system back into the ground state without producing a work current , which could be a significant source of inefficiency .
finally , the state @xmath57 decays back to the charge neutral ground state , closing the cycle . if considering the quantum effects resulting from the long - range dipole - dipole interaction , the new element of the system is the formation of new optically excitable states through strong exciton coupling among the donor molecules @xcite . ,
become excited by absorbing incident photons and their excited electrons are transferred to the acceptor @xmath52 .
the pink and blue shadowed regions surrounding the molecules represent the molecular orbitals representing the spatial distribution of electron density.,scaledwidth=40.0% ] for simplicity , we assume that three donors ( @xmath60 and @xmath51 ) are identical and degenerate , so the uncoupled excited states @xmath61 and @xmath55 of the three donors have the same excitation levels @xmath62 . furthermore , their dipole moments are aligned to the same direction , @xmath63 as depicted in fig . [ model ] .
we assume the dipole - dipole interaction between only nearest neighbors .
the dipole - dipole couplings between @xmath49 and @xmath50 , and @xmath50 and @xmath51 are denoted by @xmath64 , but there is no coupling between @xmath49 and @xmath51 .
the hamiltonian for the system of three interacting donors is written as @xmath65 it is straightforward to obtain the three single - excitation states of hamiltonian ( [ threeh ] ) : @xmath66 , @xmath67 , and @xmath68 .
the corresponding eigenvalues are obtained as @xmath69 , and @xmath70 .
the dipole moment between the state @xmath71 and the ground state @xmath72 is enhanced / weakened by constructive interference between the individual transition dipole matrix elements , @xmath73 , while the dipole moment of the state @xmath74 cancels due to destructive interference .
this means the state @xmath74 , comprised of the antisymmetric combination of the uncoupled @xmath53 and @xmath55 states , describes an optically forbidden dark state . on the contrary ,
the @xmath75 and @xmath76 states describe two optically active bright states with photon absorption and emission rates @xmath77 and @xmath78 , respectively , in compared with the uncoupled case , @xmath79 . in other words , @xmath75 is much brighter than @xmath76 , as the photon absorption and emission rate of @xmath75 is enhanced while that of @xmath76 is weakened .
obviously , the dark state @xmath74 has a resultant charge transfer matrix element equal to zero .
the intermolecular dipole interaction will also modify the transition rate between the donors and acceptor .
the electron transfer matrix elements leading to charge separation have been chosen to have the same magnitudes @xmath80 .
also , we assume that the acceptor molecule hosts an electron within its lowest unoccupied molecular orbital , which is characterized by the shape of the @xmath81-orbitals ( see fig .
[ model ] ) . we have also assumed that the donor molecules are located close to different lobes of the acceptor molecular orbital ; this leads to electron transfer matrix elements with the same magnitudes but different signs , i.e. , @xmath82 . due to effects of the dipole - dipole interactions ,
the eigenstates of the three optically excited donors are no longer uncoupled , but are coherent exciton states .
the bright states @xmath71 have matrix elements @xmath83 , giving decreased / enhanced electron transfer rates of @xmath84 , in compared with the uncoupled case @xmath85
. these modifications of electron transfer matrix elements play a crucial role in the enhancement of photocurrents within our photocell model .
the assumptions surrounding the electron transfer matrix elements is identical to that in ref . .
another crucial procedure in our model is phonon - mediated energy relaxation , which can be very effective between exciton states with strong pigment overlap @xcite .
these relaxations are included in our kinetic model via the relaxation rates @xmath86 . assuming that the new donor states are directly populated by the absorption of weak incoherent solar photons , the kinetics of the optically excited states obey the pauli master equation ( pme ) by treating the donor - light , electron transfer , and bright - dark relaxation coupling by second - order perturbations @xcite .
the pmes for the uncoupled case , describing the processes as shown in fig . [ states2 ] ( a ) , are written as @xmath87 + \gamma_c[n_c p_{\alpha } -(1+n_c)p_{1}],\\[5pt ] \dot{p}_{2 } & = \gamma_h[n_hp_b -(1+n_h)p_{2 } ] + \gamma_c[n_c p_{\alpha } - ( 1+n_c)p_{2}],\\[5pt ] \dot{p}_{3 } & = \gamma_h[n_hp_{b } -(1+n_h)p_{3 } ] + \gamma_c[n_c p_{\alpha } -(1+n_c)p_{3}],\\[5pt ] \dot{p}_{\alpha } & = \gamma_c(1+n_c)(p_{1 } + p_{2 } + p_{3 } ) -3\gamma_c n_cp_{\alpha } - \gamma(1+\chi)p_{\alpha},\\[5pt ] \dot{p}_{\beta } & = \gamma_c[n_c p_{b } - ( 1+n_c)p_{\beta } ] + \gamma p_{\alpha } \end{split}\ ] ] where we use the notation @xmath88 with indices @xmath89 running as @xmath90 .
similarly , the pmes for the dipole - dipole coupled case , whose processes are shown in fig .
[ states2 ] ( b ) , are given by @xmath91 + \gamma_{12}[n_{12}p_{2 } -(1+n_{12})p_{1 } ] + \gamma_{13}[n_{13}p_{3 } -(1+n_{13})p_{1 } ] + \gamma_{1c}[n_{1c}p_{\alpha } -(1+n_{1c})p_{1}],\\[5pt ] \dot{p}_{2 } & = \gamma_{12}[(1+n_{12 } ) p_{1 } -n_{12}p_{2 } ] + \gamma_{23}[n_{23}p_{3 } -(1+n_{23 } ) p_{2}],\\[5pt ] \dot{p}_{3 } & = \gamma_{3h}[n_{3h}p_{b } -(1+n_{3h } ) p_{3 } ] + \gamma_{23}[(1+n_{23 } ) p_{2 } -n_{23}p_{3 } ] + \gamma_{13}[(1+n_{13 } ) p_{1 } -n_{13}p_{3 } ] + \gamma_{3c}[n_{3c}p_{\alpha}- ( 1+n_{3c } ) p_{3}],\\[5pt ] \dot{p}_{\alpha } & = \gamma_{1c}[(1+n_{1c } ) p_{1 } -n_{1c } p_{\alpha } ] + \gamma_{3c}[(1+n_{3c } ) p_{3 } -n_{3c } p_{\alpha } ] -\gamma(1+\chi)p_{\alpha},\\[5pt ] \dot{p}_{\beta } & = \gamma p_{\alpha } + \gamma_c[n_c p_b - ( 1+n_c ) p_{\beta } ] , \end{split}\ ] ] where index @xmath89 of @xmath92 runs as @xmath93 . in eqs .
( [ e1 ] ) and ( [ e2 ] ) , the equation of motion for @xmath94 is determined by the conservation of the probability , @xmath95 . in eqs .
( [ e1 ] ) and ( [ e2 ] ) , @xmath45 and @xmath96 ( @xmath97 ) stand for the average numbers of photons with frequencies matching the transition energies from the ground state @xmath0 to the excited states @xmath98 and @xmath75 ( @xmath76 ) , respectively .
@xmath40 and @xmath99 ( @xmath100 ) are the thermal occupation numbers of ambient phonons at room temperature , @xmath101 , with energies @xmath102 in eq .
( [ e1 ] ) and @xmath103 ( @xmath104 ) in eq .
( [ e2 ] ) .
@xmath105 , @xmath106 , and @xmath107 represent the corresponding thermal occupations at @xmath108 with energies @xmath109 , @xmath110 , and @xmath111 , respectively .
@xmath112 is the corresponding thermal occupation at @xmath108 with the energy @xmath113 .
the rates in eqs .
( [ e1 ] ) and ( [ e2 ] ) obey local detailed balance and correctly lead to a boltzmann distribution for the level population if the thermal averages for the photon and phonon reservoirs are set to a common temperature , such as room temperature .
we consider the initial condition to be a fully occupied ground state , i.e. , @xmath114 . to calculate the population of each state , we use the following parameters : the energy levels are @xmath115 , @xmath116 , @xmath117 ; the transfer rates are @xmath118 , @xmath119 , @xmath120 , @xmath121 @xcite .
we assume that the superposition states are stable under the steady - state operation , so that @xmath122 have to satisfy the relationship : @xmath123 @xcite . here
, we choose as a limiting condition : @xmath124 .
we also employed this as a limiting condition for creator s model , to create an appropriate comparison with our model .
[ pop_dynamics ] ( a ) and ( b ) show the populations of each state in the absence and presence of coupling . due to the dipolar interaction among donors , the populations of the donors ground state , @xmath0 ,
is significantly decreased while the populations of the acceptors states , @xmath56 and @xmath57 , are notably increased in the presence of coherence when the system reaches the steady - state operation .
these changes are responsible for the enhanced photocurrents . of the levels from the numerical solution of the pauli master equation ( eq . [ e1 ] ) for uncoupled donors .
( b ) numerical solutions of the pauli master equation ( eq . [ e2 ] ) for coupled donors.,title="fig:",scaledwidth=45.0% ] of the levels from the numerical solution of the pauli master equation ( eq . [ e1 ] ) for uncoupled donors .
( b ) numerical solutions of the pauli master equation ( eq . [ e2 ] ) for coupled donors.,title="fig:",scaledwidth=45.0% ] as a function of the transition rate @xmath125 between the donors and acceptor . using the upper limit condition for @xmath126 , @xmath127 , and @xmath128
, we can get a current enhancement as high as 49.5% . on the other hand when @xmath129
, there is no current enhancement.,scaledwidth=45.0% ] as a function of the recombination rate @xmath59 using @xmath119 . @xmath130 and @xmath131 are the electric current in the excitonically coupled and uncoupled cases , respectively , when the system reaches steady - state operation .
the red line represents the current enhancement for the system with three donors ; the blue line represents the current enhancement for the system with two donors proposed by creator _
; the black line represents the current enhancement of our model comparing to creator s model in the presence of dipolar coupling.,scaledwidth=45.0% ] , at room temperature .
the blue lines represent the dimensionless current @xmath132 and power generated ( @xmath133 ) of the system with three uncoupled dipoles ( @xmath134 ) ; the red lines represent the dimensionless current @xmath135 and power generated ( @xmath136 ) of the system with three coupled dipoles ( @xmath137 ) ; the green lines represent the dimensionless current ( @xmath138 ) and power generated ( @xmath139 ) of the system with two coupled dipoles , which is proposed in ref .
.,scaledwidth=45.0% ] taking a modest recombination rate @xmath140 with @xmath141 , fig .
[ eff ] shows the current enhancement as a function of the transition rate , @xmath125 , using the other parameters listed before . under the upper limit condition , when @xmath129 , there is no current enhancement .
this means that the charge transfer via the channels @xmath142 , @xmath143 , and @xmath144 are as fast as the combined transfer through the independent channels @xmath145 , @xmath146 , and @xmath147 .
however , when @xmath148 , the coherent coupling leads to substantial current enhancements when compared to the configuration without coupling .
[ eff ] also shows that the current enhancement may reach as high as 49.5% , comparing this with 35% in creator s model .
this can be explained by two factors : ( i ) the optical transition rate between the ground and the bright state is enhanced from 2 times to 2.9 times .
( ii ) the electron transition from the almost - dark ( dark state in ref .
@xcite ) to acceptor is increased from 2 times to 2.9 times .
so a simple calculation shows the enhancement of pv model , @xmath149 . we have also explored the effect of the recombination rate , @xmath140 , on the current enhancement . in fig .
[ chi ] , we show not only the current enhancement for the system comprised of three donors , ( which is proposed here , ) but also the current enhancement for the system with two donors , ( that proposed in ref . , ) under similar electron transfer rate conditions .
the results show that although the overall current is lower for faster recombination , the relative enhancement of the photocurrent is actually slightly larger for strong recombination .
this behavior is analogous to that in the system with two donors @xcite . from fig .
[ chi ] , we also notice that the current enhancement in our three - donor system is much larger than that in creatore s two - donor system at any value for the recombination rates .
however , the current enhancement , on the order of @xmath150 , from the two - coupled - donor model to the three - coupled - donor model is very small . within this scheme ,
we assume there to be a _
load _ connecting the acceptor levels @xmath151 and @xmath152 . according to fermi - dirac statistics ,
the voltage , @xmath153 , across this load can be expressed as @xmath154 , where @xmath155 is the fundamental charge of the electron @xcite .
thus , we can assess the performance of our proposed photocell in terms of its photovoltaic properties .
the steady - state current - voltage @xmath156 characteristic and power generated are shown in fig . [ power ] .
the current and voltage are evaluated using the steady - state solutions of the pmes ; the calculations are performed at increasing rate @xmath157 as other parameters are fixed , from open circuit regime where @xmath158 ( @xmath159 ) to the short circuit regime where @xmath160 .
the power , @xmath161 , is evaluated by the formula @xmath162 . in fig .
[ power ] , the peak current enhancement is roughly 23.4% in uncoupled three - donor system ( @xmath134 ) relative to the coupled three - donor system ( @xmath137 ) . following the definition contained in ref .
@xcite , the peak delivered power enhancement is about 23.0% in the uncoupled three - donor system ( @xmath134 ) relative to the coupled three - donor system ( @xmath137 ) .
we also show in fig .
[ power ] that the current - voltage characteristic and power generated for the system with two coherent donors , proposed in ref .
@xcite . from these results
, we find that when compared to the two coherent donor system , the three coherent donor system has an enhancement of 6.3% in both peak current and peak delivered power .
the study of photosynthesis has inspired a new method by which we may harness quantum effects and coherent coupling amongst chromophores for the formation of coherent superposition to realize an artificial light - harvesting system at the molecular scale . in this paper , we propose a simple model to improve the performance of a theoretical photocell system . with suitably arranged electron donors , the photocurrents and power can be greatly enhanced through harnessing quantum effects .
the studied system is a photocell where the excitations are assumed resonant ; for solar cells the excitation is done by solar radiation which has broad spectrum .
however , the presented approach can be utilized in solar cells in different ways .
one approach is to extend the system into _
n_-dipole ( extended bands ) and use solar radiation for excitation .
another possibility is to host the dipole aggregates in solar cell materials close to the lumo to suppress recombination and hence increase the collected photogenerated carriers @xcite .
developing new concepts to harvest and utilize energy based on lessons learned from nature like those in photosynthesis is of great current interest .
examining the current - voltage characteristic and power generated for the system with three coherent dipoles , we have found an efficiency enhancement of about 6.3% compared with two coherent dipoles .
this encouraging trend suggest a promising novel design aspect of photosynthesis - mimicking photovoltaic devices .
engel , t.r .
calhoun , e.l .
read , t .- k .
ahn , t. manal , y .- c .
cheng , r.e .
blankenship , and g.r .
fleming , _ evidence for wavelike energy transfer through quantum coherence in photosynthetic systems _ ,
nature * 466 * , 782 ( 2007 ) .
calhoun , n.s .
ginsberg , g.s .
schlau - cohen , y.c .
cheng , m. ballottari , r. bassi , and g.r .
fleming , _ quantum coherence enabled determination of the energy landscape in light - harvesting complex ii _ , j. phys .
b , * 113 * , 16291 ( 2009 ) .
d. abramavicious , b. palmieri , and s. mukamel , _ extracting single and two - exciton couplings in photosynthetic complexes by coherent two - dimensional electronic spectra _
, chem . phys . * 357 * , 79 ( 2009 ) .
g. panitchayangkoon , d. hayes , k.a .
fransted , j.r .
caram , e. harel , j. wen , r.e .
blankenship , and g.s .
engel , _ long - lived quantum coherence in photosynthetic complexes at physiological temperature _ , pnas , * 107 * , 12766 , ( 2010 ) .
e. harel , a.f .
fidler , and g.s .
engel , _ real - time mapping of electronic structure with single - shot two - dimensional electronic spectroscopy _ proc .
usa , * 107 * , 16444 ( 2010 ) .
d. hayes , g.b .
griffin , and g.s .
engel , _ engineering coherence among excited states in synthetic heterodimer systems _ , science , * 340 * , 1431 ( 2013 ) .
e. romero , r. augulis , v.i .
novoderezhkin , m. ferretti , j. thieme , d. zigmantas , and r. van grondelle , _ quantum coherence in photosynthesis for efficient solar - energy conversion _
* 10 * , 676 ( 2014 ) .
m. mohseni , p. rebentrost , s. lloyd , and a. aspuru - guzik , _ environment - assisted quantum walks in photosynthetic energy transfer _ , j. chem .
* 129 * , 174106 ( 2008 ) .
plenio , and s.f .
huelga , _ dephasing - assisted transport : quantum networks and biomolecules _ , new j. phys . ,
* 10 * , 113019 ( 2008 ) .
p. rebentrost , m. mohseni , i. kassal , s. lloyd , and a. aspuru - guzik , _ environment - assisted quantum transport _ , new j. phys . * 11 * 033003 ( 2009 ) .
j. zhu and s. kais , _ modified scaled hierarchical equation of motion approach for the study of quantum coherence in photosynthetic complexes _ , j. phys . chem .
b , * 115 * , 1531 ( 2011 ) . s
yeh and s. kais , _ population and coherence dynamics in light harvesting complex ii ( lh2 ) _ , j. chem .
phys . * 137 * , 084110 ( 2012 ) .
w. shockley and h.j .
queisser , _ detailed balance limit of efficiency of p - n junction solar cells _
, j. appl .
* 32 * , 510 , ( 1961 ) .
p. wrfel , _ physics of solar cells _
( wiley - vch , berlin , 2009 ) . _ quantum efficiency in complex systems , part ii : from molecular aggregates to organic solar cells _ edited by u. wrfel , m. thorwart , and , e. r. weber , semiconductors and semimetals , vol .
85 ( academic press , san diego , 2011 ) .
miller , e. yablonovitch , and s.r .
kurtz , _ strong internal and external luminescence as solar cells approach the shockley - queisser limit _ , ieee j. photovoltaics , * 2 * , 303 ( 2012 ) .
alharbi , _ carrier multiplication applicability for photovoltaics ; a critical analysis _ , j. phys .
d * 46 * , 125102 ( 2013 ) .
f.h . alharbi and s. kais , _ theoretical limits of photovoltaics efficiency and possible improvements by intuitive approaches learned from photosynthesis and quantum coherence _ ,
energ . rev .
( in press ) ; arxiv:1402.1923 .
r.e . blankenship , d.m .
tiede , j. barber , g.w .
brudvig , g. fleming , m. ghirardi , m.r .
gunner , w. junge , d.m .
kramer , a. melis , t.a .
moore , c.c .
moser , d.g .
nocera , a.j .
nozik , d.r .
ort , w.w .
parson , r.c .
prince , and r.t .
sayre , _ comparing photosynthetic and photovoltaic efficiencies and recognizing the potential for improvement _ , science * 332 * 811 ( 2011 ) .
scully , _ quantum photocell : using quantum coherence to reduce radiative recombination and increase efficiency _ , , * 104 * , 207701 , ( 2010 ) .
scully , k.r .
chapin , k.e .
dorfman , m.b .
kim , and a. svidzinsky , _ quantum heat engine power can be increased by noise - induced coherence _ , proc .
usa , * 108 * , 15097 ( 2011 ) .
svidzinsky , k.e .
dorfman , and m.o .
scully , _ enhancing photovoltaic power by fano - induced coherence _ , , * 84 * , 053818 , ( 2011 ) . k.e .
dorfman , d.v .
voronine , s. mukamel , and m.o .
photosynthetic reaction center as a quantum heat engine
usa , * 110 * , 2746 , ( 2013 ) . c. creatore , m.a .
parker , s. emmott , and a.w .
chin , _ efficient biologically inspired photocell enhanced by delocalized quantum states _ , , * 111 * , 253601 , ( 2013 ) .
it should be noted that this efficiency may not be analogous to the shockley quisser limit @xcite .
the hamiltonian ( [ hamil_exciton ] ) is identical to an xy spin hamiltonian .
one exciton state corresponds to the one spin flipped state .
the single exciton eigenstates of hamiltonian ( [ hamil_exciton ] ) are given explicitly as @xmath163 .
for simplicity , these are written as @xmath164 .
h. van amerongen , l. valkunas , and r. van grondelle , _ photosynthetic excitons _ ( world scientific , singapore , 2000 ) .
a. shnirman , y. makhlin , and g. schn , _ noise and decoherence in quantum two - level systems _ ,
saikin , a. eisfeld , s. valleau , and a. aspuru - guzik , _
photonics meets excitonics : natural and artificial molecular aggregates _ ,
nanophotonics * 2 * 21 ( 2013 ) . | the high quantum efficiency of photosynthetic complexes has inspired researchers to explore new routes to utilize this process for photovoltaic devices .
quantum coherence has been demonstrated to play a crucial role within this process .
herein , we propose a three - dipole system as a model of a new photocell type which exploits the coherence among its three dipoles .
we have proved that the efficiency of such a photocell is greatly enhanced by quantum coherence .
we have also predicted that the photocurrents can be enhanced by about 49.5% in such a coherent coupled dipole system compared with the uncoupled dipoles .
these results suggest a promising novel design aspect of photosynthesis - mimicking photovoltaic devices . |
reality has the habit of intruding on the prodigies of purest thought and encumbering them with unpleasant embellishments .
so it is astonishing when the chthonian hammer of the engineer resonates precisely to the gossamer fluttering of theory .
such a moment may soon be at hand in the practice and theory of quantum computation .
the most compelling theoretical question , @xmath1localization , " is yielding an answer which points the way to a solution of quantum computing s ( qc ) most daunting engineering problem : reaching the accuracy threshold for fault tolerant computation .
after shor s discovery [ s1 ] of a polynomial time factoring algorithm in the quantum model qc , skeptics properly questioned whether a unitary evolution could ever be induced to process information fault tolerantly .
the most obvious tricks , such as making a backup copy , useful in a dissipative system ( e.g. pencil and paper ) are unavailable in quantum mechanics . to overcome these difficulties , a remarkable theoretical framework based on @xmath1stabilizer codes , `` @xmath1transversal gates , '' @xmath1cat - state - ancilli , " and nested concatenations of these
was erected [ s2 ] , [ s3 ] , [ a , b - o ] , [ k1 ] , and [ klz ] .
while the result is a consistent recipe for fault - tolerant quantum computation , the accuracy threshold which would allow this combinatorial behemoth to overcome its own overhead has been estimated as about @xmath2 , one i.i.d .
error per one million physical gate operations and requiring gates accurate also to one part in a million .
this places a formidable task before the engineer and physicist .
but within the year the beginnings of a new idea on fault tolerance had been generated by kitaev [ k2 ] . while the term is not yet present in that paper the idea is to construct ( first mathematically ) a @xmath1quantum medium `` and to store quantum states as topological structures within the medium and ( eventually ) manipulate these states , that is , apply gates to them , by topological transformations of the medium . for our purposes , we define a quantum medium as a collection of many finite level systems coupled together by a hamiltonian @xmath3 obeying a strong locality condition : the individual systems are located in a @xmath4dimensional lattice or a more irregular cellulation of a surface @xmath5 .
we postulate a constant @xmath6 so that @xmath7 and each @xmath8id , where the identity is on all tensor factors(= subsystem ) not located within some ball @xmath9 of diameter @xmath10 in the lattice .
for example , the heisenberg magnet with @xmath11 is a quantum medium of diameter @xmath12 .
( but engineer be warned ; localizing @xmath13 within balls of diameter @xmath14 implies @xmath15ary interaction for @xmath16 .
controlling effective @xmath15ary terms for @xmath17 will be tricky in the extreme and probably will require enforcing symmetries to cancel lower order terms . )
kitaev s @xmath1toric code '' [ k2 ] in which quantum states are stored as first homology of a torus , can be counted as having @xmath18 ; they require @xmath19ary interactions .
we study here a partial generalization of the toric code which also stores quantum information in a degenerate ground state @xmath20 of a quantum medium .
the medium is on a disk with point - like defects which we treat as punctures .
the dimension of @xmath20 , @xmath5 the punctured disk , grows exponentially with the number of punctures .
transformations of @xmath5 , that is braidings ( up to isotopy ) of the punctures in space - time , @xmath21 , operate unitarily on @xmath20 . other work ( [ k2 ] , [ p ] , and [ k , b ] )
also explores the realization of elements of computation by braiding anyonic @xmath1quasi - particles `` or @xmath1defects '' of a quantum medium .
the vision is that stability of computation , at least sufficient to reach the @xmath2 threshold for @xmath1software `` error correction , is to be realized by the discreteness of algebraic topology : two @xmath22homology cycles are never @xmath1close , '' two words in the braid group are equal or distinct .
more exactly , it is geometry not topology which will confer stability . working in a lattice model one
may calculate [ k2 ] that the perturbation hamiltonian @xmath23 must be raised to the length scale @xmath24 before nonzero terms , @xmath25 ground state @xmath26 , are encountered and so the splitting of the ground state is estimated to be proportional to @xmath27 .
the length scale in the previous two examples are : @xmath28 ( length of shortest essential cycle ) ; and in the anyonic context , the closest that two defects are allowed to come to each other during braiding . the @xmath1engineering goal `` is to construct a quantum medium on a material disk whose ground state admits many localized excitations ( @xmath1anyons '' ) whose braidings effect computationally universal unitary transformations of the ground state .
it is further hoped that actual @xmath1errors , `` the result of unwanted noisy excitations , are to be removed automatically by some relaxation process in which the system is coupled to a cold bath by another much weaker hamiltonian @xmath29 .
the mathematicians first cut at the engineering goal is to produce a quantum medium with these properties and this is accomplished by the theorem below .
this @xmath1first cut '' is not yet interesting to experimentalists since the hamiltonian contains summands which have as many as @xmath30 nontrivial indices , but it represents an exact existence theorem .
the question for physicist is whether this phase can also be represented perturbatively with a simple hamiltonian , perhaps a rvb model [ a ] , [ n , s ] .
this would be a major step toward physical realization .
consider a rectangle @xmath31 of euclidian square lattice consisting of @xmath32 boxes by @xmath33 boxes .
associate a @xmath4level spin system @xmath34 with each of the @xmath35 box edges in @xmath31 .
the disjoint union of these spin systems has hilbert space @xmath36 .
there is a time dependent local hamiltonian @xmath37 with fewer than @xmath38 terms and each @xmath39 having @xmath30 or fewer indices , supported in at most a @xmath40 rectangle of boxes - @xmath1diameter @xmath41 .
`` for @xmath42 , the ground states of @xmath43 form a sub - hilbert space @xmath44 , and geometrically determines @xmath45 exceptional points or @xmath1defects '' spaced out along the midline of @xmath31 . within @xmath46 there is a @xmath1computational `` sub - hilbert space @xmath47 , @xmath48 @xmath49 may be identified with the @xmath50witten - chern - simons modular functor at level @xmath51 of the @xmath52punctured disk with the fundamental representation of @xmath53 labeling each of the @xmath54 boundary components .
the braid group @xmath55 of the defects acts unitarily on @xmath46 according to the jones representation at level @xmath41 .
any quantum algorithm can be efficiently simulated on @xmath56 by restricting the action of @xmath55 to a @xmath1computational subspace . ''
the representation is implemented adiabatically by gradually deforming @xmath57 to @xmath58 and then to @xmath59 and so on .
the passage from @xmath57 to @xmath58 involves turning off an exceptional term @xmath60 which defines a defect site and turning on a new term @xmath61 , which determines an alternative , adjacent , site for the defect at time @xmath62 .
each braid generator can be implemented in @xmath63 times steps .
we believe , based on a conjectural energy gap , that the geometry confers stability to this implementation which increases exponentially , error = @xmath64 , under refinement of the lattice on @xmath31 by a factor of @xmath24 , while the number of time step needed for a computation increases only linearly in @xmath24 .
so far , we have only discussed the @xmath1engineering `` : the quest to specify @xmath3 ( which will be described in the proof ) .
let us take a brief digression from that sulfurous underworld of grinding gears to the elysian fields of abstract thought .
the witten - chern - simons theory descends from the signature (= pontryagin form ) in dimension 4 and every step of the desent to lower dimension leads to deeper abstraction until mathematical wit is well nigh exhausted as the point ( dimension @xmath0 ) is reached . to tell this story in its barest outline , we restrict to @xmath65 , and borrow from atiyah [ a ] , freed [ f ] , and walker [ w ] .
the signature of a closed @xmath19manifold is an integer as is the pontryagin class of an @xmath53 bundle over a closed @xmath19manifold .
an @xmath50bundle over a closed @xmath66manifold is topologically trivial but if endowed with a connection acquires a secondary @xmath1 chern - simons '' class in the circle @xmath67 .
quantizing [ wi ] at level @xmath68 , leads to the topological jones - witten - chern - simons invariant @xmath69 which is morally an average of the classical chern - simons invariant over all connections .
the invariant for a closed surface @xmath5 ( with some additional structure ) is a finite dimensional vector space @xmath56 ; and each @xmath66manifold bounding @xmath5 determines a vector @xmath70 . before dividing by gauge symmetry ,
the vector space @xmath71 is the infinite dimensional space of sections of the associated complex line bundle to a natural @xmath72bundle over the space of @xmath53 connections @xmath73 on @xmath53 bundles over @xmath5 .
a @xmath66manifold @xmath74 with connection , @xmath75 , on a bundle extending the bundle over the boundary , @xmath76 , determines a map @xmath77 by integrating the chern - simons form over @xmath78 .
the consistent choices for such functionals constitute the total space of this @xmath1natural `` @xmath72bundle . in general , a map @xmath79 is @xmath1consistent '' if it obeys the additivity properties of the chern - simons integral : @xmath80 .
symplectic reduction followed by quantization as explained in [ a ] produces a finite dimensional @xmath56 from @xmath71 with @xmath81 depending only on the topology of @xmath74 .
the definition of the witten - chern - simons invariant for a surface with boundary is a collection of vector spaces indexed by certain labelings . for a @xmath82manifold
the invariant seems to be a certain type of @xmath83category `` while the correct definition for a point is but dimly perceived and the object of current research .
several authors assert that it is unnecessary to finish the progression , that we can be content with a theory whose smallest building blocks are @xmath1pairs of pants '' ( three - punctured- spheres ) .
the invariant for these while technically a vector in a @xmath4vector space is easily understood in terms of sets of vector spaces parameterized by @xmath1labelings `` of the boundary circles so no unusual categorical abstractions need be mastered .
the reason for this assertion is that using a handle body decomposition all closed @xmath66manifold invariants can be calculated from gluing along surfaces with smooth boundary ; gluings along faces with corners on the boundary , which one would encounter computing from a cellulation , can be avoided .
but the freed - walker program rejects this advice on two grounds .
first localizing @xmath20 not merely to @xmath1pants , '' but to cells ( i.e. neighborhoods of points ) may give more natural consistency conditions , to replace the @xmath84 consistency equations of [ w ] ; which in turn could eventually lead to classification of modular functors and a conceptual understanding .
second , to paraphrase edmund hillary , we should localize to points @xmath1because they are there . "
the hyperbole of the first paragraph can now be made sound .
cs5 is a universal model for quantum computation and for the physicist / engineer to implement it , a local hamiltonian @xmath3 must be described . for the pure mathematician to be satisfied with his understanding of cs5
it must be localized to points .
the two objectives are certainly similar in spirit and possibly identical .
to clarify the connection , we introduce an intermediate concept , undoubtedly plebeian , but dear to a topologist .
we would like when possible to describe a vector in a modular functor as a linear combinations of @xmath1admissible `` up to @xmath1equivalence . ''
this , after all , is exactly how we understand homology : @xmath85 is an equivalence class of admissible pictures . to be admissible
the picture must be a closed @xmath82manifold , the equivalence relation is bordism . both @xmath86manifold - ness `` and @xmath1bordism '' can be defined by local conditions which are the combinatorial analogs of @xmath1closed `` and @xmath1co - closed '' familiar from de rham s theory of differential forms . in kitaev
s toric code these condition are imposed by vertex and face operators @xmath87 and @xmath88 respectively .
there is a subtle shift here from the usual way of thinking of homology as equivalence classes of cycles to the @xmath1harmonic " representative which is merely the equally weighted average of all cycles in the homology class . in this way
quotients and equivalence classes are never encountered and homology is located within cycles , within chains , just as a c.s.s .
code space is located within the fixed space of stabilizers built from products of @xmath89 s and further within the fixed space of stabilizers , @xmath90 s . to generalize from homology , we should think of a picture as ( linear combinations of ) anything we can draw on a surface @xmath5 . if helpful , we allow various colors and/or notational labels , framing fields , etc@xmath91 , and even additional dimensions bundled over @xmath5 .
but in the present case no such embellishments are required .
what is important that if we move the surface by a diffeomorphism , the picture should also move and move canonically .
thus if @xmath5 is a torus it would not suite our purposes to draw the picture of @xmath92 in a solid torus @xmath93 , @xmath94 : a meridial dehn twist on @xmath5 extends over @xmath93 , twisting the picture , but a longitudinal dehn twist does not have any obvious way to act on a picture drawn in @xmath93 .
( to anticipate , a modular functor will have an @xmath95matrix which can transform a picture in one ( call it the @xmath1inside `` ) solid torus to a picture in the dual ( @xmath1outside '' ) solid torus where longitudinal a dehn twist does act . but
resorting to the @xmath95matrix does not solve our problem since its input and output pictures are on a scale of the injectivity radius of the surfaces and hence nonlocal . )
we demand that @xmath1admissibility `` and @xmath1equivalence '' of pictures be locally determined , i.e. decided on the basis of restriction to small patches on @xmath5 . to make the connection with lattice models , we consider @xmath5 discretized as a cell complex ; the conditions must span only clumps of cells of constant combinatorial diameter .
as in the example of harmonic @xmath82cycles , @xmath1equivalence " is a slight misnomer : what we impose instead are invariance condition on the ( linear combinations of ) admissible pictures representing any fixed @xmath96 which ensure that the stabilized vectors are in fact equally weighted superpositions of all admissible pictures representing @xmath97 .
now consider the question , perhaps the first question a geometric topologist should ask about a modular functor @xmath98 ; can you draw a ( local ) picture of it on @xmath5 so that the mapping class group of @xmath5 acts on @xmath20 by the obvious induced action on pictures ?
we should not expect it to be easy to discover the local rules for the pictures associated to a given modular functor @xmath56 and in fact they may not exist in much generality . recall that a three manifold @xmath74 bounding @xmath5 , @xmath99 determines a vector @xmath100 so we might think of our proposed picture @xmath101 drawn on @xmath5 as some ghostly recollection of @xmath74 . the present understanding of modular functors is closely related to surgery formulas on links , but to think in this way we must choose a @xmath1base point " @xmath66manifold @xmath102 with @xmath103 to hold the links .
this choice seems to create an asymmetry which should not be present in @xmath104 .
thus for a pictorial representation of @xmath56 which is derived from surgery , we expect only part of the mapping group @xmath105 that part extending over @xmath106 will act locally . to localize v , this problem must be overcome .
let us propose a meta theorem or @xmath1principle `` that solving the @xmath1picture problem , '' which we call @xmath1combinatorial localization , `` should imply both the freed - walker program , which we call @xmath1algebraic localization '' and the design problem for the hamiltonian @xmath3 which we call @xmath1physical localization . "
.2 in = 4.9 in figure 1 .2 in the solid arrow is asserted with some confidence at least as a mathematical statement ; the dotted arrow is speculative . while the solid arrow seems unlikely to have a literal converse : ground states of even simple hamiltonians in dimension @xmath107 are too complicated to draw pictures of ; conceivably the dotted arrow might be an equivalence constituting a culmination of the freed - walker program .
we show how to represent cs5 ( and by extension all csr ) on a disk with marked points by local pictures .
since the representation of quantum computing within cs5 [ flw ] only used the braid group acting on a disk with marked points , this partial solution to the combinatorial localization problem will suffice to prove the theorem ( once we have explained the solid arrow in figure 1 ) . for any @xmath108
, cs@xmath109 has a combinatorial localization on any cellulated disk with marked labeled points , ( labels @xmath110 lie on the marked points and disk boundary ) provided the cellulation has bounded combinatorics and the marked points stay sufficiently far from each other and the boundary . for a concrete statement , let us take the cellulated disk to be a rectangle @xmath111 with a square euclidean cellulation .
we suppose that all marked points are at least @xmath109 lattice spacings from the boundary and @xmath112 from each other . the marked points and @xmath113 are all assigned the label @xmath114 ( the irreducible @xmath115 dimensional representation of @xmath116 ) . in this circumstances
it is easy to build a trivalent @xmath117collared rooted tree " @xmath118 for the disk with marked points as shown in figure 2 .
all straight segments of the tree are to be more than @xmath119 lattice bonds in length ; the root is on @xmath113 and the leaves are the marked points .
the @xmath120collard condition is that an @xmath121 relative regular neighborhood @xmath122 of lattice cells - the region within the dashed line - should be imbedded in @xmath31 .
the box counts in the statement of the theorem are designed to permit a ( discontinuous ) family of @xmath118 s to be found for at all times during braiding .
we say that the boxing of @xmath31 is relative to the location of the marked points if it has this property .
the key lemma 2.1 will show that for roomy boxing that two discrete pictures , which we regard as smoothly equivalent are in fact combinatorially equivalent .
more precisely , the infinity of smooth averaging operators acting on the space of combinatorial pictures has exactly the same joint fixed set as a finite subset of combinatorial operators .
let us begin with a geometric interpretation of cs5@xmath123 .
for a closed surfaces @xmath5 it is implicit in [ k , l ] .
let @xmath5 bound a handle body @xmath3 .
a general @xmath66manifold @xmath74 with boundary @xmath5 can now be represented as a @xmath1blackboard framed `` surgery diagram in @xmath3 .
the special cabling morphism @xmath124 of the temperley - lieb category ( see chapter 12 [ k , l ] or [ r , t ] ) when composed into the surgery diagram yields a linear combinations of @xmath82 manifolds , each labeled by @xmath125 '' .
we may write @xmath3 as a planar surface cross interval , @xmath126 , so that @xmath127 , where @xmath128 denotes @xmath129 with its orientation reversed . now projecting these @xmath82manifolds to @xmath129
, we see a linear combination of immersed @xmath82labeled @xmath82manifolds with overcrossings indicated at double points .
this pictures determines the vector @xmath130 .
the kauffman relations at a root of unity , in our case @xmath131 , allow extensive simplification of these pictures via the recoupling formalism .
in fact each @xmath132 can be encoded as a labeling of a fixed ( framed , imbedded , and vertex planar ) trivalent graph , which is a spine for a @xmath129 .
it is an important observation of walker s ( personal communication ) and gelca s [ g ] that this description can be extended to labeled surfaces with boundary .
( verification follows directly from the gluing axiom . ) in the case of a disk with @xmath133 marked points @xmath134 - treating marked points as crushed boundary components - the modular functor with @xmath135 labels @xmath136 has as its basis @xmath137admissible labelings with boundary condition on a fixed trivalent tree imbedded in @xmath138 , rooted on @xmath139 , with leaves on the marked points .
the boundary condition is that the label on the root is the label given on @xmath140 and each leaf has the label associated to its marked point . as in [ flw ] , we only need consider the case where all labels @xmath12 .
the ( framed ) braid group acts on the labeled tree @xmath93 via its imbedding in the disk . to see
the induced action on @xmath141 ( we drop labeling subscripts ) , perturb the imbedding of @xmath93(rel its endpoints ) by pushing it downward into a three ball @xmath142 $ ] , where we think of @xmath138 identified with @xmath143 .
now implement any desired braid @xmath144 as a diffeormorphism of @xmath145 $ ] where @xmath146 is small with respect to the previous push .
viewed from above , @xmath147 has overcrossings but the recoupling @xmath148 rules ( and isotopies ) allow @xmath147 to be described in the original basis of @xmath137admissible labelings on @xmath93 ( with root and leaves still carrying the label 1 ) .
for example the simplest kauffman relations , on strands of @xmath147 labeled by @xmath149 " read : @xmath150{drawingh.eps } = e^{\pi i/10 } ) ( e^{-\pi i/10 } \
, \includegraphics[width=.55cm , height=.35cm]{drawingf.eps}\tn { and } \
, \bigcirc = e^{\pi i/5 } + e^{-\pi i/5 } = : d.\ ] ] a detailed example : the effect of a single braid generator , is given immediately following the statement of lemma 2.1 to elucidate the recoupling of braids .
there is a topological observation inherent in inducing the braid action on @xmath151 . by capping off , any diffeomorphism of a planar surface
extends to the two sphere and can be extended further to a diffeomorphism of the @xmath66ball @xmath152 .
the action on @xmath56 comes from projecting this topological extension acting on labeled trivalent trees back into the original planar surface ( after crushing the inner boundary components to points ) .
in fact , it is the correspondence between @xmath66manifolds and diagrams which proves that we have correctly specified the action on the functor , for we have @xmath153 where @xmath154 . generally ,
when a surfaces @xmath5 has genus @xmath155 there will be no way of including it in the boundary of a @xmath66manifold @xmath156 so that all diffeomorphisms of @xmath5 extend over @xmath74 .
however it is a triviality that any diffeomorphism of @xmath5 extends over @xmath157 by product with id@xmath158 .
now let this extension act on the appropriate equivalence classes of framed @xmath137admissibly labeled trivalent graphs imbedded in @xmath157 projected back into @xmath5 to define the action on any @xmath159level @xmath160 modular functor @xmath56 .
thus the @xmath1doubled " functor @xmath161 has a combinatorial localization , i.e. is describable by local pictures .
this may have some relation to unpublished work of kitaev and kupperberg ( private communication ) on local descriptions for drinfeld doubles .
we set aside for later study the problem of devising combinatorial local rules for the necessary elementary equivalences of such trees @xmath93 : @xmath162moves , ribbon equivalence , vertex half - twist equivalence , and regular homotopy .
one would hope to define a quantum medium for cs5 of individual systems with levels to record labels @xmath163 ( and possible additional levels to store other information ) and terms @xmath164 with at most @xmath165 indices ( as in a @xmath162symbol ) corresponding to these elementary equivalences . while this count seems correct in the smooth setting ,
there the crude hilbert space is infinite dimensional which may create new difficulties .
we have not been able to find a discrete setting in which all the equivalences are expressed efficiently . for the purpose of this lecture ,
we stay with discrete models for quantum media built from @xmath4level systems , but to do this we accept terms @xmath164 with up to @xmath30 indices .
the fundamental @xmath4dimensional representation of @xmath53 generates @xmath53 s complex representation ring and as a result recoupling theory achieves a very simple result : an element @xmath166 is a linear combination of imbedded @xmath82manifolds each labeled by @xmath149 `` , i.e. the standard @xmath4dimensional representation and given the boundary condition : each @xmath82manifold of the linear combination meets each marked point ( and @xmath139 ) once .
thus @xmath1manifoldness '' and the @xmath1boundary condition " define admissibility for our picture .
this makes good sense combinatorically in the lattice of @xmath31 , as well as , smoothly .
we point out that our notion of @xmath82manifold is strict : at each vertex @xmath167 or @xmath115 edges ( not @xmath168 ) should be occupied .
it is time to define the local equivalence moves between pictures .
we are working within the temperley - lieb category modulo the relation that the @xmath169 jones - wenzl projector is trivial .
this is our most interesting relation . as a smooth equivalence relation
this has only one form but combinatorially , we need to impose two versions of it according to how the output endpoints are grouped .
we denote these by and .
the second picture stands for : in conventional projector notation ( [ k , l ] ) .
a second relation says that removing a circle which bounds a disk free from punctures multiples the diagram by the scalar @xmath170 , @xmath171 .
a third relation replaces the undercrossing that arise through braiding with legitimate morphisms if the category . in terms of smooth pictures ,
the relation replaces the @xmath1virtual " uncrossing in the middle diagram with a two term sum : the middle picture is @xmath1 virtual " ; it is not actually an admissible picture to be assigned a weight .
this relation requires a little care and lattice space to discretize since we do not want to permit the intermediate picture : which would represent the wrong boundary data at the indicated defect .
recall that each defect is labeled by @xmath114 representing the @xmath4dimensional irreducible representation of @xmath116 which is recorded by a single line leaving the defect .
finally , a fourth class of equivalence permits isotopy .
again the reader should note that enough neighboring sites should be observed by the appropriate @xmath164 to preserve imbeddedness .
for example , cases 1 and 2 are allowable , case 3 is not .
there will be isotopy relations for arc endpoints as well .
for example , in cases @xmath114 and @xmath115 of figure 3 imagine the open circle filled to become an end point and the shorter of the two line segments meeting it deleted .
morally , we should define operators @xmath172 which enforce the average of the initial @xmath173 and final @xmath174 configuration of cases @xmath114 and @xmath115 .
however there is a detail , to get the overall phase correct , and not settle for merely a projective representation , we must fix a base point direction : say the positive ray emanating from each endpoint at @xmath175 degrees and find positive semidefinite @xmath172 s which assign zero norm to @xmath176 and @xmath177 $ ] in cases @xmath114 and @xmath115 respectively .
these operators correspond to asserting equivalences : @xmath178 and @xmath179 .
the general rule is that a state obtained by clockwise ( counterclockwise ) isotopy through the base point direction must be adjusted by the phase @xmath180 before being averaged .
similarly there is an isotopy relation for the arc end point on the boundary circle of the disk @xmath138 .
here some point on the boundary is chosen an phase is adjusted by @xmath181 as this point is crossed clockwise ( counterclockwise ) .
let us return to the raltions @xmath182 .
combinatorically the first may be written out with the left hand side a @xmath183 lattice square foliated by parallel straight lines ( of label @xmath12 ) .
s [ we ] , recursion formula , yields an identity equating @xmath168 parallel lines with a linear combination of @xmath184 @xmath1smaller `` terms each containing @xmath1turn arounds . ''
the form of the relation is shown below : the coefficients @xmath185 are rational functors of @xmath10 which can be computed from the wenzl s recursion relation for projectors ( see pg . 18 [ k , l ] or [ we ] ) .
figure 6.0 is merely the lattice counterpart of the more familiar smooth relation , figure 6.0@xmath186 , which may be applied within any diagram ( at @xmath187 ) whenever four @xmath82 labeled lines are found running parallel .
obviously figure 6.1 also has a smooth counterpart .
the admissibility conditions and the above four classes of @xmath1equivalences `` must be rewritten as operators @xmath188 and @xmath189 respectively ; collectively denoted @xmath190 .
let @xmath191 denote the ground state of the soon - to - be - defined hamiltonian @xmath192 .
let @xmath56 denote the csr modular functor of the disk with @xmath45 marked points and all labels @xmath12 . via recoupling
, we may describe @xmath56 in the fashion of homology .
set @xmath193 [ admissible pictures ] and write : @xmath194 , where @xmath195 is the smooth - category equivalence relation corresponding to our four combinatorical equivalences : @xmath196 .
lemma 2.1 will prove that under the @xmath1roomy hypothesis '' @xmath195 and @xmath196 induce identical equivalence classes of admissible pictures ( which of course are combinatorial objects ) .
so we may also write @xmath197 .
our goal is to tailor @xmath3 so that the ground states @xmath198 correspond bijectively to linear functionals @xmath199 under the map @xmath200 .
this will identify @xmath191 with @xmath201 , but since @xmath56 has a canonical nonsingular hermitian inner product @xmath202[wi ] and [ k , l]@xmath203 this also gives an isomorphism @xmath204 .
the inner product @xmath205 is defined on pictures by imbedding the disk @xmath138 into the @xmath206plane , deforming @xmath207 upward rel endpoints and @xmath208 downward rel endpoints .
the union of the deformed pictures @xmath209 is a ( vertically framed ) link in @xmath210 and it kauffman bracket is @xmath205 . note that the vertical framing is singular where @xmath207 and @xmath208 share a common lattice bonds meeting @xmath211 and @xmath212 ; here the convention is to bend such bonds of @xmath208 slightly clockwise at the endpoints internal to @xmath138 and counterclockwise at an endpoint on @xmath139 .
the definition of the @xmath213 operators is quite obvious .
consider , a vertex @xmath97 in the interior of @xmath31 .
a hermitian @xmath214 with @xmath168 indices whose ground state is spanned by classical states of valence @xmath167 or @xmath115 at @xmath97 is said to enforce @xmath86manifoldness `` at @xmath97 .
clearly the ground state of @xmath214 has dimension @xmath215 .
to enforce , instead , a @xmath1defect '' or marked point labeled by the fundamental representation , @xmath11 " of @xmath53 , we would use instead a hermitian operator @xmath216 with ground state spanned by the four classical states of valence @xmath12 at @xmath97 .
there are @xmath218 nonboundary edges @xmath219 ( shown in bold ) .
if @xmath220 are the nonempty ( classical ) manifold configuration of these edges , i.e. valence @xmath221 at each of the four internal vertices , and iff @xmath222 and @xmath223 , set @xmath224 and let @xmath225 be the set of such vectors .
let @xmath226 be the hermitian operator with @xmath218 indices on @xmath227 whose ground state is orthogonal to span @xmath225 .
@xmath228 is the operator which @xmath1allows isotopy across @xmath144 . "
to remove circles which bound disks we need , in the presence of isotopy , only introduce operators which deletes a box .
this operator may be written as @xmath229 where @xmath230 is a unit vector proportional to @xmath231 .
we postpone the definition of the operator corresponding to figures 3 and 4 since this must involve the dynamics @xmath232 `` of @xmath57 .
some trick is needed to avoided adding new levels to our system to encode @xmath1crossings . ''
the projector corresponding to , figure 6.0 , requires a @xmath233index operator acting on a @xmath183 grid of edges or @xmath1box `` @xmath234 whose @xmath82 dimensional excited state is spanned by the vector obtained by putting all fourteen term in figure 6.0 on the left hand side of the equation .
similarly the projector corresponding to , figure 6.1 is a 30 index operator acting on the bonds of a region the shape of l.h.s . in figure 6.1 .
this @xmath1nobby box '' @xmath235 is a @xmath236 rectangle union an additional small box in the middle of one of the long sides .
now we turn to the dynamics .
almost all conditions @xmath164 that combine to yield @xmath3 are permanent , only the end point operators @xmath237 should change as we execute braiding . because of the technical problem illustrated in figure 4 ; any lattice resolution into a superposition of two @xmath82manifolds as in figure 3 may cause collision with other strands .
one way to deal with this problem is to locate the marked points on a second lattice @xmath238 consisting of the mid points of the edges in the original lattice @xmath239 of boxes in @xmath31 .
this means that we have to add additional @xmath4index @xmath73 operators holding equal the two classical states on both halves of the original edges , i.e. ground state @xmath240 , and that the end point operators @xmath241 actually occur ( with @xmath4dimensional ground states ) on the finer lattice @xmath238 , @xmath242 .
the dynamics consists of moving an endpoint diagonally on @xmath238 , i.e. translating one unit horizontally or vertically in the structure of @xmath239 . in figure 8 the endpoint @xmath124 is moved horizontally to @xmath243 by replacing : @xmath244 with @xmath245 .
if @xmath124 and @xmath243 are immediately adjacent in @xmath239 the operator swap will cause the end point to travel around a corner .
this operator swap can be performed gradually by slowly turning the appropriate terms on or off . if the adiabatic theory is applicable , and following the proof of lemma 2.1 we discuss the heuristics for an energy gap ( in the theromdynamic limit ) for the family @xmath57 , @xmath246 will be carried to a unique ground state @xmath246 of @xmath58 .
this ground as a functional on pictures is identical to @xmath246 provided pictures are identified according to the obvious isotopy rules ( and phase rules at endpoints ) .
if the lattice is refined by a linear factor @xmath24 , tunneling to an undesired orthogonal ground state @xmath247 should , by arguments analogous to those for the stability of homology classes [ k2 ] , have amplitudes scaling like @xmath248 .
the mathematical description for adiabatic evolution of the system is via the natural connection @xmath73 on the tautological bundle over the complex grassmannian @xmath249 : the time evolution of @xmath250ground states @xmath251 defines a path in @xmath249 and @xmath252transport covers this motion with a unitary ( i.e.isometric ) identification @xmath253 , for all @xmath254 . after a braiding @xmath144
is completed at time @xmath255 , the self - identification @xmath256 is the representation of the braid b. a ground state @xmath257 defines a functional @xmath258 on @xmath23 via orthogonal projection .
the @xmath259 have been chosen to correspond to @xmath196 precisely so that a unique extension @xmath260 exists : .1 in = .6 in .1 in and @xmath260 satisfies @xmath261 . conversely given a functional @xmath260 on @xmath262 the @xmath263 associated to @xmath260 by
the formula above lies in the null space of each @xmath217 , so in fact @xmath264 .
the important remaining point is to see that after braiding , when the marked points have been returned to there original sites set - wise , that the induced transformation on the ground state is precisely , up to error @xmath265 , the unitary cs5 representation originally introduced by jones [ j ] and studied in [ flw ] . but
this follows from the recoupling theory as presented in [ k , l ] provided we show that the combinatorial relations that we have imposed through hermitian operators @xmath259 in fact are sufficient to span all the relations implied by the infinitely many smooth relations between pictures , that is @xmath266 . for this
the following lemma suffices .
let @xmath267 be a linear relation between admissible combinatorial pictures in @xmath268 which holds under @xmath195 , the smooth recoupling theory associated to cs@xmath109 .
provided that the configuration @xmath269 is roomy in the rectangle @xmath31 , the same relation already holds under @xmath196 . before proving the lemma
let us carry out a simple calculation to get a feel for how the action of braiding is computed via recoupling theory .
if the reader wishes to try more complicated examples , the formulas on pages 93 - 100 of [ k , l ] are helpful . here
we compute the effect of a braid generator on a vector @xmath270cs5 ( @xmath66punctured disk ) where each boundary component has label@xmath12 ( the @xmath4dimensional representation of @xmath271 and to account for phase each boundary has a marked base point . @xmath272{drawinga3.eps},\ ] ] which as a labeled tree is : @xmath273{drawinga4.eps}\ ] ] let @xmath144 be the counterclockwise braided of the right most pair of punctures
. then @xmath274 is represented by : @xmath275{drawinga5.eps } = \underset{\lq\lq\tn{virtual picture"}}{\includegraphics[width=2.5cm , height=.85cm]{drawinga6.eps } } = \ ] ] @xmath276{drawinga7.eps } \ , \,+ \,\ ,
a^{-1}\,\ , \includegraphics[width= 2.5cm , height=.85cm]{drawinga8.eps } , \tn { where } a = e^{2 \pi i/10}.\ ] ] now @xmath277{drawingg.eps}}$ ] is our notation for the jones - wenzl idempotentent @xmath278{drawingi.eps}}= \f{1}{\sqrt{d^{2}-1}}\left(\includegraphics[width=.75cm , height=.25cm]{drawingf.eps}- \f{1}{d } \supset \subset
\right)$ ] where @xmath279 and , as usual , the open ends in the above diagrams can be interpreted as permitting arbitrary ( but constant ) extension to the outside .
note : the orthogonality relations * @xmath280 @xmath281
* @xmath282{drawingf.eps } -\f{1}{d } \supset \subset \right ) , \f{1}{\sqrt{d^{2}-1}}\left(\ , \includegraphics[width=.75cm , height=.25cm]{drawingf.eps } -\f{1}{d}\supset \subset \right ) > = \\ \f{1}{d^{2}-1}\left ( \ , \includegraphics[width=.55cm , height=.30cm]{drawinge.eps}\ , -\f{2}{d}\ , \includegraphics[width=.50cm , height=.33cm]{drawingd.eps}\ , + \f{1}{d^2}\,\includegraphics[width=.55cm , height=.32cm]{drawingc.eps } \right)= \f{1}{d^2 -1}\left(d^2 - \f{2}{d } d + \f{1}{d^2 } d^2\right)=1 $ ] , and * @xmath282{drawingf.eps } -\f{1}{d } \supset \subset \right),\f{1}{d } \supset\ \subset > = \f{1}{d\sqrt{d^{2}-1}}\left ( \ , \includegraphics[width=.50cm , height=.33cm]{drawingd.eps } -\f{1}{d } \ ,
\includegraphics[width=.55cm , height=.32cm]{drawingc.eps } \right ) = \\ \f{1}{d\sqrt{d^{2}-1}}\left ( d-\f{1}{d}d^2\right ) = 0 $ ] from the definition of @xmath277{drawingg.eps}}$ ] we have : @xmath285 @xmath277{drawingg.eps}}- \f{1}{d}\supset\subset$ ] .
so we use this to expand the two parallel lines in the second term of @xmath286 to get : @xmath287{drawinga9.eps}\right)\\ = & a \psi_\circ + \f{a^{-1}}{d } \left ( \sqrt{d^{2 } -1 } \quad \includegraphics[width=2.5cm , height=.85cm]{drawinga12.eps } + \ , \f{1}{d}\quad \includegraphics[width=2.5cm , height=.85cm]{drawinga11.eps } \right ) ( * * ) \\ = & a \psi_\circ + \f{\sqrt{d^{2}-1}}{d } \,a^{-1 } \psi_2 + \f{a^{-1}}{d } \psi_\circ\\ \tn{where } \psi_2 : = & \ , \includegraphics[width=2.5cm , height=.85cm]{drawinga12.eps } , \\ = &
\left(a + \f{a^{-1}}{d}\right)\psi_\circ + \f{\sqrt{d^{2}-1}}{d } \,a^{-1 } \psi_2\end{aligned}\ ] ] as a check on unitarity note that under the sequelinear pairing , @xmath288 * proof of lemma 2.1 . * the argument is based on the birkhoff curve shortening principle where by a family of imbedded arcs and circles can be @xmath1pulled tight `` to a shorter geodesic position without crossings developing .
we work combinatorically . by the @xmath1roomy hypothesis '' there is an @xmath120collared tree @xmath289 .
assign a positive weight @xmath290 to each bond @xmath291 ( or @xmath82cell ) of the cellutation of @xmath31 so that @xmath124 grows rapidly with distance to @xmath93 : as a good first approximation , we may take @xmath292 where @xmath293minimum number of bonds joining @xmath291 to @xmath93 . now
for any ( classical ) picture @xmath294 define its length @xmath295 .
permitting combinatorial isotopy ( rel the marked points ) and the removal of small circles , but not the undercrossing , , or relations , we may pull @xmath294 tight by local moves to equivalent pictures ( up to a scalar ) which steadily reduce @xmath296 until a local minimum is reached . call this step @xmath1pull tight `` .
because of our weight function @xmath124 , the new @xmath294 will try to lie mainly in a small neighborhood of @xmath93 , and only occupancy of the bonds close to @xmath93 will force parts of the picture to lie farther away . also the picture , seeking to occupy the bonds near @xmath93 efficiently will have its strands running parallel to @xmath93 in @xmath297 lattice blocks @xmath291 a thwart the middle @xmath120bonds of each of the distinguished length @xmath119 segments of @xmath93 . also at the trivalent vertices of @xmath93
near which sufficiently many strands pass , we would like to see copies of l.h.s . figure 6.1
. this will be true up to a small isotopy ( across a few boxes ) and can be made true on the nose by modifying the weight function @xmath124 by adding a small term proportional to the distance from each trivalent vertex out to a distance @xmath109 from that vertex .
now apply at some site in a @xmath234 or at some site @xmath235 if the opportunity presents .
this breaks @xmath294 into @xmath298 and for all @xmath299 , @xmath300 .
pull tight again to remove the slack created by the @xmath1turn arounds '' in figure 6.0 or 6.1 .
alternate pulling tight with applications of or until no further reductions in length can be made in this way .
call this cycle @xmath1pull and cut `` . with a slight abuse of notation let @xmath301 denote one of the terminal classical states of this process .
now allow a single @xmath1under crossing '' move ( figure 3 ) to further reduce @xmath302 if such a move is available .
now alternated the @xmath1pull and cut " cycle with single under crossing moves until no daughter picture ( still denote @xmath301 ) can have its length reduced by further iteration of this process .
manifestly , all the daughter pictures @xmath301 now lie in @xmath122 and pass through all boxes @xmath234 parallel to @xmath93 and with @xmath303 or few strands and pass through each @xmath235 in a standard way according to some admissible triple as explained below .
note that @xmath304 is not admissible . at this point
it is simple to formally reorganize the term of this sum @xmath305 as @xmath306 where @xmath307 is an admissible labeling of @xmath93 .
as explained in [ k , l ] , the leaves and root @xmath93 of @xmath308 are always labeled by @xmath114 ( this is our choice ) and the admissibility condition says that other edges ( i.e. components of the intrinsic @xmath82skeleton of @xmath93 ) are labeled by @xmath309 taken from @xmath310 so that at each trivalent vertex of @xmath93 the following relations hold on the triple of incident labels @xmath311 and @xmath312 : @xmath313 an admissible labeling @xmath307 is interpreted as a linear combination of pictures by replacing each edge with the jones - wenzl projector corresponding to its label .
the set of admissible labeled trees @xmath314 is an orthogonal basis for the modular functor @xmath315 , defined topologically using the smooth equivalence relation .
( the subscript @xmath316 is to emphasis that the smooth relations are used in this definition ; of course @xmath317 . ) because of the assumed @xmath318 , @xmath319 for all admissible @xmath320 . but
each @xmath301 is an imbedded arc pairing @xmath321 of the @xmath322leaves @xmath323 root@xmath324 in @xmath122 satisfying the additional admissibility restriction at each trivalent vertex of @xmath93 .
such pairings are an alternative ( though not orthogonal ) basis for the modular functor @xmath325 so collected in this basis we have for each pairing type @xmath321 , we have @xmath326 where @xmath327 but all @xmath301 of a fixed type are clearly combinatorally equivalent @xmath328 .
thus we have found a combinatorial path through applications of @xmath328 from @xmath329 to the empty picture , or more precisely to a sum of zero times various pictures .
@xmath330 unlike [ k2 ]
the individual summands of @xmath3 do not commute .
the ground states of @xmath3 has been computed topologically , however the spectrum spec@xmath26 is less accessible .
the most important question is the existence of an energy gap above the ground state which is constant under lattice refinement , @xmath331 , i.e. in the thermodynamic limit .
the following heuristics motivate the conjectured energy gap . in finite classical systems such as random walk on a graph diffusion time
is well known to scale inversely with the spectral gap of the laplacian .
similarly , in some simple quantum mechanical systems where exact calculation is possible , the energy gap scales inversely to the diffusion time between classical states . in [ k2 ] where direct calculation yields an energy gap above the ground state , the classical states are cycles and the @xmath1diffusion `` is through elementary bordisms . since we have set up our ground state to be analogous to homology : @xmath332 with pictures playing the role of cycles and our @xmath259 playing the role of bordisms we expect similar diffusion properties and hence an energy gap . in lemma 2.1
the proof shows that equivalent pictures @xmath207 and @xmath208 are connected by a @xmath1path '' @xmath333 of deformations ( @xmath1down " from @xmath207 to a neighborhood of @xmath118 and then back up to @xmath208 ) .
rapid diffusion corresponds to observing that there are a plethora of such paths and in fact the procedure for finding @xmath333 is highly under determined .
more difficult would be a rigorous implication between diffusion and spec@xmath26 . extending the analogy with [ k2 ] , in both cases when the lattice is refined by a factor of @xmath24 , a sequence of @xmath334 local operators
is required to transform between a pair of orthogonal ground states .
so given the existence of an energy gap , the hamiltonian @xmath3 will be stable to order @xmath335 in perturbation theory ; formally corresponds to tunneling amplitudes between orthogonal ground states which scale like @xmath336 .
there are several important open questions .
the first is a rigorous treatment of the energy gap , but this is probably too difficult in the present model .
another is how to deal with errors in the form of actual rather than @xmath1virtual " excitation which have already been discussed in the context of tunneling .
can a coupling to a could bath repair such errors or are more active measures required ?
for example , can broken endpoint pairs of a @xmath82manifold find each other and cancel through some imposed attraction ( as suggested by dan gottesman in conversations ) or merely through random walk ?
nearby error pairs may be more serious in cs5 than in the toric codes since isotopy class not just homology needs to be preserved ; the wrong reconnection pairing would result in an unrecoverable error . to make this unlikely , should additional terms be included into our hamiltonian @xmath3 which could force distinct strands to be widely separated ?
this would put more weight on the simpler pictures , which are the ones that the quantum medium can most easily correct if damaged .
kitaev s very general notion of quantum media with its several antecedents in the study of quantum statistical mechanics looks likely to become a central object of study shared between theoretical physics , solid state physics , and topology .
the main disappointment of the present investigation is the complexity of the local hamiltonian @xmath3 used to construct stable universal quantum computation .
one sees no easy road to radically simplifying it and still obtaining an exact description of cs5 .
however another path may be open . in our discussions ,
kitaev has suggested ( also see page 46 [ p ] ) that simpler lattice hamiltonians may renormalize in the scaling limit to topological modular functors .
perhaps the most interesting topological theories , such as cs5 , because of their simplicity will have large @xmath1basins of attraction " under renormalization and that identifiable universality classes of quantum media may not only exist mathematically but may even lie within the reach of engineers .
p. w. shor , algorithms for quantum computation : discrete logarithms and factoring .
_ 35th annual symposium on foundations of computer science ( santa fe , nm , 1994 ) , 124134 , ieee comput .
press , los alamitos , ca , 1994_. | the mathematical problem of localizing modular functors to neighborhoods of points is shown to be closely related to the physical problem of engineering a local hamiltonian for a computationally universal quantum medium . for genus
@xmath0 surfaces , such a local hamiltonian is mathematically defined .
braiding defects of this medium implements a representation associated to the jones polynomial and this representation is known to be universal for quantum computation . _
dedicated to my teachers and collaborators : alexei kitaev , greg kuperberg , kevin walker , and zhenghan wang .
their work has been the inspiration for this lecture . _ |
a paper of griffiths and li in this journal @xcite has triggered a lively discussion on the electrostatic equilibrium distribution of charge on a thin , straight , cylindrical conductor .
while suspecting that as the diameter - to - length ratio of the conductor approaches zero the charge distribution on it tends to uniformity at all points sufficiently far from the ends , they encountered certain puzzling aspects of the problem in their numerical studies .
in particular , they identified an apparent contradiction in the strict limit of zero diameter , posing the question that `` if the charge density _ were _ constant , on a truly one - dimensional needle , how could the force on an off - center point be zero ? '' contributions by good @xcite and andrews @xcite followed the work of griffiths and li , offering further evidence of the uniformity conjecture of the latter , as well as alternative methods of analyzing the problem .
however , it fell to jackson @xcite to provide a clear resolution of the equilibrium paradox quoted above by carefully considering the limit of vanishing diameter - to - length ratio and the crucial role of the slow , inverse - logarithmic approach of charge density to uniformity in that limit .
a subsequent paper by jackson @xcite reported on his discovery that the problem in question had already been considered by none other than james clerk maxwell ! maxwell , treating `` the electrical capacity of a long narrow cylinder '' more than 130 years ago @xcite , had derived the inverse - logarithmic approach to uniformity noted by andrews and highlighted by jackson , and had eloquently summarized the behavior of electrostatic charges on a thin , straight conductor .
maxwell s pioneering words on the subject turned out to be a very fitting last word as well .
the purpose of this contribution is to investigate the distribution of charge on a thin , _ curved _ conductor , which we will often refer to as a _
filament_. the intuition underlying this inquiry is the expectation that a sufficiently thin , smoothly curved filament , when viewed up close , would appear long and straight , so that the equilibrium behavior of charges on the filament should be nearly uniform just as in the case of straight conductors .
this is indeed what we find and describe in this paper .
we approach the problem of charge distribution on a physical filament in two steps .
the first step is the treatment of the fields of any sufficiently smooth , one - dimensional distribution of charge , which we will often refer to as an _ ideal filament _ , with no regard to its conductivity properties .
here we consider the behavior of the electric field components and the electrostatic potential in the limit of vanishing distance from an arbitrary point of the filament , and carry out an asymptotic analysis to calculate the divergent parts of these quantities explicitly .
we find that the divergent parts of the radial and azimuthal electric field components , as well as the electric potential , are determined by the linear charge density of the filament at the point in question , whereas the axial component of the electric field is determined by the spacial derivative of linear charge density , or equivalently , the electric dipole density , at the point .
the intuition underlying these results is discussed throughout , and the results are summarized in eqs .
( [ 19 ] ) and ( [ 20 ] ) . in the second step of our analysis , we relate the fields of the ideal filament to those of a physical one by casting the former as the symmetry axis of a curved , cylindrical conductor . here , we show that the equipotential surfaces of the physical filament approach those of the ideal one as its diameter vanishes , and use this to deduce the inverse - logarithmic approach of charge distribution to uniformity , the primary result of this paper , in eq .
( [ 24 ] ) .
the rest of this paper is organized as follows . in
iia we set up the geometry of the ideal filament and formulate the required regularity conditions . in
iib we present the derivation of the electric field components near the ideal filament , and in iic we deal with the electrostatic potential .
we then formulate the requirements of relating the fields of the physical filament to those of the ideal one in iiia . in
iiib we consider the consequences of enforcing those requirements , whence we derive the result expressing the approach of charge density to uniformity .
an alternative derivation of this result using energy minimization is developed in
examples of straight and circular filaments providing analytical illustrations of eqs .
( [ 19 ] ) and ( [ 20 ] ) are given in
ivb we present the results of a numerical study of an elliptical filament , including an interactive computer program for calculating them .
concluding remarks are presented in v.
consider a charged filament in the limit of zero diameter , i.e. , idealized as an open or closed curve .
we will assume this curve to be of class @xmath0 , or twice continuously differentiable , and parameterized by @xmath1 , @xmath2 , where @xmath3 is the arc length parameter @xcite .
these regularity conditions guarantee the existence and continuity of the principal normal vector and curvature , as well as the tangent vector , at every point .
thus @xmath4 is the unit vector tangent to the curve at point @xmath1 , with @xmath5 .
furthermore , the unit principal normal is given by @xmath6 , where @xmath7 is the curvature at point @xmath3 .
these two unit vectors , together with the unit _ binormal _ vector @xmath8 , constitute a local orthogonal triad at each point of the curve .
note that the regularity conditions stated above guarantee that @xmath7 is a continuous function , allowing us to define a radius of curvature @xmath9 which is continuous and has an infimum which we denote by @xmath10 .
note also that the same conditions imply that the curve is non - self - intersecting , i.e. , that @xmath11 , except for closed filaments for which @xmath12 . the minimum radius of curvature and the length of the filament are the two relevant length scales for points near the filament , and we will find it convenient to use the scale parameter @xmath13 in the following analysis .
consider a point @xmath14 of the filament corresponding to @xmath15 , where @xmath16 for closed filaments and @xmath17 in case of open filaments .
as defined above , @xmath18 , @xmath19 , and @xmath20 are the tangent , normal , and binormal unit vectors at point @xmath14 .
we will also consider a nearby point @xmath21 located at @xmath22 on the normal plane of the filament at point @xmath14 , where @xmath23 , @xmath24 , is a unit vector in the normal plane and @xmath25 is the distance between @xmath14 and @xmath21 . here and below
, @xmath26 is a dimensionless , positive number which represents the distance @xmath27 in units of @xmath28 . clearly , as @xmath29 , @xmath21 approaches @xmath14 . )
form the triad of tangent , normal , and binormal unit vectors at point @xmath14 of the solid curve .
the pair ( @xmath30 ) defines the transverse plane at point @xmath14 , with the dotted circle delimiting the cross section of the curved cylinder .
the azimuthal angle @xmath31 and the pair ( @xmath32 ) define a polar coordinate system on the transverse plane . ]
figure 1 illustrates the geometry of the filament and the coordinate system used to describe it .
it also suggests that as @xmath31 ranges from @xmath33 to @xmath34 , the line @xmath27 sweeps the surface of a normal disk of radius @xmath25 centered on point @xmath14 .
relative to this disk , @xmath35 is the `` radial '' unit vector at point @xmath21 and @xmath36 the azimuthal unit vector .
these unit vectors , together with @xmath18 , constitute an orthogonal triad at point @xmath21 similar to the three unit vectors of a cylindrical coordinate system .
note that as @xmath3 ranges from @xmath37 to @xmath38 and @xmath31 sweeps the interval @xmath39 $ ] , the set of points @xmath40 define the surface of a curved , circular cylinder of fixed diameter @xmath41 @xcite .
the electric field at any point @xmath42 not on the filament is given by @xmath43/{|\mathbf{r}-\mathbf{r}({s})|}^{-3 } , \label{1}\ ] ] where @xmath44 is the linear charge density on the filament , assumed to be continuously differentiable , and @xmath45 .
in particular , the electric field at point @xmath21 is given by @xmath46 $ ] . in the following
we will be concerned with the axial , radial , and azimuthal components of the electric field at point @xmath14 , which we respectively define as the limit of @xmath47 , @xmath48 , and @xmath49 as @xmath21 approaches @xmath14 .
we then find @xmath50 , \label{2 } \\ & & { e}_{\textrm{\textrm{ra}}}(\alpha,{s}_{\scriptscriptstyle 0})={\lim}_{\epsilon \rightarrow 0 } { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0})\cdot \mathbf{e}[\mathbf{r}({s}_{\scriptscriptstyle 0})+\epsilon l { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0 } ) ] , \label{3 } \\ & & { e}_{\textrm{\textrm{az}}}(\alpha,{s}_{\scriptscriptstyle 0})={\lim}_{\epsilon \rightarrow 0 } { \boldsymbol{\hat{\alpha}}}(\alpha,{s}_{\scriptscriptstyle 0})\cdot \mathbf{e}[\mathbf{r}({s}_{\scriptscriptstyle 0})+\epsilon l { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0})].\,\,\,\,\,\,\ , \label{4}\end{aligned}\ ] ] furthermore , we are only interested in the leading , divergent contributions to these limits , and will organize the following calculations accordingly .
it should also be noted here that the electric field components at point @xmath14 depend on the azimuthal angle @xmath31 as well as on @xmath15 .
this is because the @xmath29 limit is a singular one , which in general renders the limit dependent on the azimuth of the approach to point @xmath14 . written in detail , eqs .
( [ 2])-([4 ] ) appear as @xmath51 { |\epsilon l { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0 } ) + \mathbf{r}({s}_{\scriptscriptstyle 0})-\mathbf{r}(s)|}^{-3},\,\,\,\,\,\label{5}\\ & & { e}_{\textrm{ra}}(\alpha,{s}_{\scriptscriptstyle 0})=k\ , { \lim}_{\epsilon \rightarrow 0 } \int_{{s}_{1}}^{{s}_{2 } } ds \ , \lambda(s)\ { \epsilon l + { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0 } ) \nonumber \\ & & \cdot[\mathbf{r}({s}_{\scriptscriptstyle 0})-\mathbf{r}(s ) ] \ } { |\epsilon l { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0 } ) + \mathbf{r}({s}_{\scriptscriptstyle 0})-\mathbf{r}(s)|}^{-3},\,\,\,\,\,\label{6}\\ & & { e}_{\textrm{az}}(\alpha,{s}_{\scriptscriptstyle 0})=k\ , { \lim}_{\epsilon \rightarrow 0 } \int_{{s}_{1}}^{{s}_{2 } } ds \ , \lambda(s ) { \boldsymbol{\hat{\alpha}}}(\alpha,{s}_{\scriptscriptstyle 0})\nonumber \\ & & \cdot [ \mathbf{r}({s}_{\scriptscriptstyle 0})-\mathbf{r}(s ) ] { |\epsilon l { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0 } ) + \mathbf{r}({s}_{\scriptscriptstyle 0})-\mathbf{r}(s)|}^{-3}.\,\,\,\,\,\label{7}\end{aligned}\ ] ] at this point it is convenient to use the regularity conditions imposed above on the charged filament , namely the @xmath52 and @xmath0 properties of @xmath53 and @xmath0 respectively @xcite , to write @xmath54 where @xmath55 and @xmath56 and are `` remainder '' terms which vanish more rapidly than the terms preceding them as @xmath57 . in the following , we will use the representations of eqs .
( [ 8 ] ) and ( [ 9 ] ) to organize each electric field component into divergent ( leading ) and finite ( remainder ) parts as @xmath58 . furthermore , we are primarily interested in the divergent parts and will not explicitly calculate the finite remainder terms in the following .
we note in passing that the remainder terms would in general depend on the detailed structure of the filament and its charge distribution .
we will first consider the contributions to the radial component of the electric field near the filament , @xmath59 and @xmath60 , which arise from the first two terms in eq .
( [ 8 ] ) respectively .
let us first observe that @xmath28 ( the smaller of filament length and minimum radius of curvature ) is the only physically relevant length scale as point @xmath14 is approached . clearly then , the filament should appear as a long , straight line of charge as viewed from any point @xmath22 if @xmath61 .
consequently , we would intuitively expect @xmath62 to approach the electric field of an infinite , straight line of constant charge density @xmath63 through point @xmath14 in the direction of @xmath18 as @xmath64 .
in other words , @xmath65 as @xmath29 . similarly , we expect @xmath66 to approach the transverse field of a dipole of moment @xmath67 centered at point @xmath68 , where @xmath69 is a proportionality constant ( independent of @xmath26 ) .
however , due to symmetry , the transverse field of a dipole located at point @xmath14 vanishes at point @xmath21 .
thus we expect the leading contribution to the radial component of the electric field to equal @xmath70 . to verify the above expectations for @xmath71 , we find it convenient to start by making a change of variable from @xmath3 to @xmath72 .
the resulting expression can then be written as @xmath73 \int_{{\lambda}_{1 } } ^{{\lambda}_{2}}d\xi \nonumber \\ & & \times \big \ { 1 + \xi { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0 } ) \cdot [ \mathbf{r}({s}_{\scriptscriptstyle 0})- \mathbf{r}({s}_{\scriptscriptstyle 0}+\epsilon l \xi ) ] /\epsilon
l \xi \big \}\nonumber \\ & & \times { \big|{\mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0})+\xi [ \mathbf{r}({s}_{\scriptscriptstyle 0})- \mathbf{r}({s}_{\scriptscriptstyle 0}+\epsilon l \xi)]/\epsilon l \xi \big |}^{-3 } , \label{10}\end{aligned}\ ] ] where @xmath74/\epsilon l$ ] and @xmath75/\epsilon l$ ] . recalling from eq .
( [ 9 ] ) that @xmath76/\epsilon l \xi \rightarrow \mathbf{r}'({s}_{\scriptscriptstyle 0})+\frac{1}{2}\epsilon l \xi\mathbf{r}''({s}_{\scriptscriptstyle 0})= { \boldsymbol{\hat{\tau}}}({s}_{\scriptscriptstyle 0})+\frac{1}{2}\epsilon l \xi \kappa({s}_{\scriptscriptstyle 0 } ) \boldsymbol{\hat{\nu}}({s}_{\scriptscriptstyle 0})$ ] as @xmath29 , we can rewrite the above result as @xmath77 \int_{{\lambda}_{1}}^{{\lambda}_{2 } } d\xi [ 1-\frac{1}{2}\epsilon l \kappa({s}_{\scriptscriptstyle 0})\nonumber \\ & \times & \cos(\alpha){\xi}^2 ] { | { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0})-\xi { \boldsymbol{\hat{\tau}}}({s}_{\scriptscriptstyle 0})|}^{-3}. \label{11}\end{aligned}\ ] ] the first term in the above integral is readily found to equal @xmath78 by noting that @xmath79 and that @xmath80 in the limit .
we therefore find that the first term in eq .
( [ 11 ] ) is given by @xmath81 , confirming our intuitive expectation above .
the second term in eq .
( [ 11 ] ) is a correction to the leading contribution we anticipated , and involves the integral @xmath82 .
the asymptotic value of this integral is found to equal @xmath83 by noting that @xmath84 and @xmath85 are both of the order of @xmath86 as @xmath29 . using this result , we find that the second term in eq .
( [ 11 ] ) is given by @xmath87 .
combining the two contributions from eq .
( [ 11 ] ) , we find @xmath88 . \label{12}\ ] ] it is worth pointing out here that just as we anticipated the leading term in eq .
( [ 12 ] ) as corresponding to a long , straight line of charge , we could have arrived at the second contribution by noting that , in the second approximation , the filament can be considered a circular arc of radius @xmath89 , i.e. , the radius of curvature of the filament at the point in question .
the reader can verify this by reference to the treatment of the circular filament in iva . turning to @xmath60 next , and using a procedure similar to that followed above for @xmath59
, we find @xmath90 /\epsilon
\big \}\nonumber \\ & & \times { \big|{\mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0})+\xi [ \mathbf{r}({s}_{\scriptscriptstyle 0})- \mathbf{r}({s}_{\scriptscriptstyle 0}+\epsilon l \xi)]/\epsilon l \xi \big |}^{-3}.\,\,\,\,\,\ , \label{13}\end{aligned}\ ] ] using eqs .
( [ 8 ] ) and ( [ 9 ] ) , we implement the @xmath29 limit in the above expression to find @xmath91 which vanishes by symmetry as anticipated .
we have thus verified that @xmath92 as @xmath29 , where @xmath93 is the finite part of the radial electric field at point @xmath14 .
our next task is to deal with the axial component of the electric field , @xmath94 . for reasons of symmetry
, we would expect @xmath95 , the contribution arising from @xmath63 in ( [ 8 ] ) , to vanish since it corresponds to the axial electric field of a straight filament of uniform charge density as the filament is approached . indeed following a procedure parallel to that used above
, we find that @xmath95 is given by @xmath96 which vanishes by symmetry as surmised . the second contribution to the axial component of the electric field ,
the one arising from @xmath97 in ( [ 8 ] ) , corresponds to the longitudinal field of an electric dipole of moment @xmath98 centered at point @xmath14 , as stated earlier . indeed following the procedure established above
, we find @xmath99 which can be evaluated by using the fact that @xmath100 as noted above .
the result is @xmath101 a logarithmically divergent expression .
we have thus found that the axial component of the electric field @xmath102 as @xmath29 , where @xmath103 is the finite part of the axial electric field at point @xmath14 .
the calculation of the azimuthal component @xmath104 proceeds along similar lines , with @xmath105 reducing to @xmath106 and @xmath107 vanishing by symmetry .
note that unlike the other components , the azimuthal component of the electric field arises from the local curvature of the filament , and as such depends on the azimuthal angle @xmath31 .
we therefore have the result that the azimuthal component of the electric field @xmath108 as @xmath109 , where @xmath110 is the finite part of the azimuthal electric field at point @xmath14 . at this point
it is convenient to summarize the above results for the electric field components near a charged filament : _ the electric field at an interior point @xmath14 of a charged , regular , ideal filament is given by _
@xmath111 { \mathbf{\hat{n}}}(\alpha,{s}_{\scriptscriptstyle 0})\nonumber \\ & & + \lambda({s}_{\scriptscriptstyle 0 } ) \sin(\alpha)\kappa({s}_{\scriptscriptstyle 0 } ) \ln({\epsilon } ^{-1 } ) \boldsymbol{\hat{\alpha } } ( \alpha,{s}_{\scriptscriptstyle 0 } ) \bigr\ } , \ , \label{19}\end{aligned}\ ] ] _ where _ @xmath112 _ is the finite part of the electric field at point _
@xmath14 _ and the notation is as established above_. it is worth repeating here that the divergent parts of the electric field in eq .
( [ 19 ] ) arise from the local charge and dipole densities at point @xmath14 .
this is of course what is expected physically since , given the regularity requirements on the geometry and distribution of charge on the filament @xcite , there is no other source of divergence than the proximity of the charges to the field point .
more importantly , while divergent parts of the radial and azimuthal electric field components are generated by the local charge density , the divergent part of the axial component is generated by the local dipole density .
it is the latter feature of the fields and sources that will ensure the vanishing of the dipole density ( hence approach to uniformity ) for a conducting filament in the limit of vanishing diameter , as the argument leading to eq .
( [ 24 ] ) will show .
it should also be emphasized here that the results of this section apply to an ideal filament of charge regardless of its conductivity properties .
we conclude this section by deriving the electrostatic potential near the filament .
the divergent part of this potential is expected to correspond to a long , straight filament of constant charge density . indeed applying the methods established above
, we find the potential at point @xmath14 of the filament as @xmath113 where @xmath114 is the finite part of the potential at point @xmath68 .
as expected , the divergent part of this result is just the electrostatic potential of a straight filament of charge density @xmath63 , and is independent of the curvature of the filament at point @xmath14 . as such
, it does not contain any information on the azimuthal component of the electric field , or the subleading contribution to the radial component , both of which are present in eq .
( [ 19 ] ) , while the axial and the leading radial components are readily derivable from it , as will be shown below .
the potential terms corresponding to the missing components , being of the order of @xmath115 , are finite and vanish in the @xmath58 limit , which is the reason they do nt appear in eq .
( [ 20 ] ) . to extract the radial and axial electric field components from the potential
, we note that the variables @xmath15 , @xmath116 , and @xmath31 on which the terms on the right - hand side of eq .
( [ 20 ] ) depend are in effect the axial , ( rescaled ) radial , and azimuthal coordinates of a local cylindrical coordinate system .
consequently , we can write @xmath117 \lambda({{s}_{\scriptscriptstyle 0 } } ) \ln({\epsilon } ^{-1 } ) .
\label{21}\end{aligned}\ ] ] note that while @xmath114 is finite as @xmath58 , its gradient is not and in fact includes a logarithmically divergent azimuthal part as mentioned above . implementing the derivatives in eq .
( [ 21 ] ) , we find @xmath118 . \label{22}\end{aligned}\ ] ] note that @xmath119 , the difference being the two @xmath31-dependent terms in eq .
( [ 19 ] ) which are absent in eq .
( [ 20 ] ) , as anticipated earlier .
taking this into account , we find that eqs .
( [ 22 ] ) and ( [ 19 ] ) are in agreement with respect to the leading radial and axial components . since the approach to a uniform charge density for a thin conductor found in eq .
( [ 24 ] ) below is forged from these leading components , we are assured that the electrostatic potential contains sufficient information to imply the same result .
this is an important point , since the derivation of the uniformity result in iiic is based on energy considerations using the electrostatic potential .
anticipating the uniformity of charge distribution on a conducting filament in the limit of vanishing diameter , as expressed in eq .
( [ 24 ] ) below , we can use the asymptotic behavior given in eq .
( [ 20 ] ) to derive a universal asymptotic formula for the capacitance per unit length of a conducting filament .
for such a filament , charge per unit length tends to a constant , say @xmath120 , and the electrostatic potential to @xmath121 , as seen from eq .
. therefore the capacitance per unit length , @xmath122 , tends to @xmath123^{-1 } , \label{23}\ ] ] where , it may be recalled from ii , @xmath41 is the diameter of the filament . for a straight filament , eq .
( [ 23 ] ) reduces to the result first derived by maxwell @xcite ; cf . @xcite .
we are now in a position to consider the physical problem of a charged , conducting filament in the limit of vanishing diameter . we will define the physical filament by reference to the equipotential surfaces of the idealized charge distribution treated in the previous section @xcite .
specifically , given a value @xmath124 for the potential of the conductor ( with respect to infinity ) , we consider the equipotential surface @xmath125 $ ] consisting of all points @xmath42 such that @xmath126 to be the surface of the conducting , physical filament . here
@xmath127 is the ( exact ) electrostatic potential produced by the idealized filament .
the uniqueness property of the dirichlet problem would then assure us that the electrostatic field produced by @xmath53 in the exterior of @xmath128 is precisely the same as would be produced by the physical conductor if held at potential @xmath124 .
furthermore , the surface charge per unit length of the physical conductor would be given by @xmath53 @xcite .
having defined the physical conductor by reference to an assumed charge distribution , we have shifted the burden of our analysis to reconciling the resulting profile with the actual shape of the physical conductor .
in particular , we might wonder if , for a given @xmath1 , it is possible to choose @xmath53 so that @xmath129 $ ] has a desired form , e.g. , that of a curved , circular cylinder of fixed diameter defined in iia .
here we have the benefit of the insight gained in the study of straight filaments , especially in refs .
@xcite , where we find an affirmative answer .
in particular , we know that a conductor with an ellipsoidal surface corresponds to a uniform charge distribution along the symmetry axis @xcite . we also know , starting with maxwell @xcite , that a thin , cylindrical conductor acquires a near - uniform charge distribution except for endpoints where higher charge densities occur .
note that the field and charge configurations of a straight filament are azimuthally symmetric . in the case of curved filaments of finite diameter , however , there s no such symmetry and one can not in general expect a one - dimensional charge distribution to properly reproduce the azimuthal distribution of the charges on the filament s surface .
this shortcoming is of course a consequence of the mismatch between the dimensionality of an ideal filament and that of the surface of a conductor . by the same token
, however , any difference between the two is expected to be small for a thin filament and tend to zero as its diameter vanishes .
this expectation is in fact born out as the analysis below shows , and we will be able to determine the linear charge density of a sufficiently thin conductor regardless of its azimuthal variations . the standard electrostatic boundary condition on a conductor is the uniformity of the potential throughout its interior and surface , a condition which follows from the vanishing of the electric field in the interior of a conductor in electrostatic equilibrium .
our problem then is to constrain @xmath53 such that @xmath130 $ ] turns out to be the curved , circular cylinder of fixed diameter defined above .
equivalently , we will require the vanishing of the tangential components of the electric field on the surface of the conductor , starting with the axial component .
the axial component was calculated in
iib where we found that @xmath131 as @xmath132 .
the requirement that @xmath94 vanish in the limit immediately constrains the charge distribution according to @xmath133 recalling that @xmath134 is an arbitrary point of the filament ( with the exception of endpoints in case of open filaments ) , we conclude that , as its diameter tends to zero , the linear charge density on the filament tends to uniformity as the inverse of the logarithm of @xmath26 .
this result generalizes and closely parallels those found for straight filaments , starting with maxwell s @xcite .
next , we consider the azimuthal component of the electric field on the surface of the filament . at first blush , eq .
( [ 19 ] ) seems to imply that the vanishing of the azimuthal component requires the eventual disappearance of all charges on the filament ! in reality , however , what is conveyed by eq .
( [ 19 ] ) is simply that the equipotential surface @xmath135 $ ] generated by an ideal filament can not have a _ circular _ cross section except at points where the curvature of the filament vanishes .
this is of course exactly what is expected , since the deviation from circularity is needed to balance the electric field component generated by the local curvature of the filament .
this sort of behavior is already familiar from the study of straight conductors where variations in the diameter along the conductor are required to balance the effect of unequal distances from the ends , and has been discussed in detail in ref .
needless to say , this circumstance does not imply that circular cross - section filaments are unphysical , only that our one - dimensional filament within is not capable of masquerading them .
fortunately , it turns out that the deviation from circularity is quite small and disappears rapidly as @xmath58 .
this is the justification for our use of an ideal charge filament to describe the distribution of charge on a thin , physical filament . to establish the above assertion
, we will examine the shape of the non - circular cross section generated by the ideal filament .
the unit normal @xmath136 to the equipotential surface @xmath137 $ ] at point @xmath138 is defined by the direction of the electric field in eq .
( [ 19 ] ) , i.e. , @xmath139,\,\,\,\,\,\ , \label{25}\end{aligned}\ ] ] where we have only included the leading contribution in each component . simplifying further , we find @xmath140\epsilon \ln(\epsilon ) { \boldsymbol{\hat{\tau}}}({s}_{\scriptscriptstyle 0})\nonumber \\ & + & \frac{1}{2}\kappa({s}_{\scriptscriptstyle 0})l \epsilon \ln({\epsilon } ^{-1 } ) \sin(\alpha)\boldsymbol{\hat{\alpha } } ( \alpha,{s}_{\scriptscriptstyle 0 } ) \bigr\}. \label{26}\end{aligned}\ ] ]
let @xmath141 describe the equipotential surface @xmath137 $ ] in terms of our curved cylindrical coordinates @xmath142 .
since the gradient vector @xmath143 $ ] is normal to the surface in question , its direction must be the same as the unit normal @xmath136 .
therefore , using eq .
( [ 26 ] ) , we can write @xmath144/ \partial \alpha = \frac{1}{2}\kappa({s}_{\scriptscriptstyle 0})l \nonumber \\ & & \times { \eta(\alpha,{s}_{\scriptscriptstyle 0},{\phi}_{\scriptscriptstyle 0 } ) } \ln[{\eta(\alpha,{s}_{\scriptscriptstyle 0},{\phi}_{\scriptscriptstyle 0 } ) } ] \sin(\alpha),\,\,\,\,\,\,\,\,\,\,\ , \label{27}\end{aligned}\ ] ] which is valid to leading order as @xmath145 , or equivalently , as @xmath146 .
it is clear from eq .
( [ 27 ] ) , and can be readily verified numerically , that the maximum fractional variation in the diameter of the equipotential surface with respect to @xmath31 is of the order of @xmath147 $ ] , hence quite small for a sufficiently small diameter @xcite .
it is important to emphasize here that the equipotential surface @xmath148 is normal to the electric field by construction , so that the tangential component of the electric field vanishes identically at all points of this surface .
consequently , the relevant boundary condition in matching this equipotential surface to the surface of a curved , cylindrical conductor is the circularity of the transverse boundary of the former , i.e. , the vanishing of the partial derivative @xmath149 /\partial \alpha$ ] .
this is of course what is guaranteed by eq .
( [ 27 ] ) in the limit of vanishing diameter .
one can also examine the azimuthal variations of the surface charge density on the equipotential surface .
the latter is given by the standard formula @xmath150 .
using eqs .
( [ 19 ] ) and ( [ 27 ] ) , we find for the surface charge density , @xmath151 \sin(\alpha ) \bigr ] } ^{2 } \bigr\ } , \,\,\,\,\ , \label{28}\end{aligned}\ ] ] where we have kept the first two terms in @xmath152 and suppressed the dependence of the linear charge density @xmath63 on @xmath124 . the quantity of interest here is @xmath153
/\partial \alpha$ ] , the fractional variation of surface charge density with respect to the azimuthal angle .
using eqs .
( [ 27 ] ) and ( [ 28 ] ) , we find @xmath154 /\partial \alpha \cong \frac{1}{2}\kappa({s}_{\scriptscriptstyle 0 } ) l \nonumber \\ & & \times \eta(\alpha,{s}_{\scriptscriptstyle 0},{\phi}_{\scriptscriptstyle 0 } ) \ln[1/{\eta(\alpha,{s}_{\scriptscriptstyle 0},{\phi}_{\scriptscriptstyle 0 } ) } ] \sin(\alpha ) , \label{29}\end{aligned}\ ] ] to leading order .
in other words , the fractional azimuthal variations of charge density on the equipotential surface vanish as @xmath155 $ ] .
not surprisingly , this is the same rate at which the fractional variations of the diameter vanish [ cf .
( [ 27 ] ) ] , and significantly more rapid than the rate at which the linear charge density of the filament tends to uniformity .
the above results allow us to conclude that as the diameter of the filament decreases , the azimuthal nonuniformities in diameter and surface charge density disappear relatively rapidly , confirming our earlier statement to that effect .
equivalently , we can conclude that in the limit of vanishing diameter , the azimuthal component of the electric field on the surface of a thin , circular cylinder vanishes as required .
let us summarize the above results : _ the density of charge along a regular conducting filament of fixed diameter tends to uniformity according to eq .
( [ 24 ] ) , except near endpoints , as the ratio of its diameter to the smaller of its length and minimum radius of curvature tends to zero . _
this is the main result of this paper , and as stated earlier , generalizes the result of maxwell and others for straight conductors to curved filaments . as such , it confirms the intuition that the charge distribution along a sufficiently smooth , charged filament basically follows that of a straight one , and similarly converges to uniformity as the diameter of the filament approaches zero .
this is so despite the fact that there s an azimuthal variation of charge density on the surface of the conductor at points of non - zero curvature which serves to balance the azimuthal component of the electric field generated by deviation from axial symmetry .
the tendency of charges in equilibrium on thin , conducting conductors to become uniformly distributed is thus seen to be robust and survive the generalization to curved filaments .
we arrived at this result in the preceding section by a detailed consideration of the electric field components and equilibrium of charges , together with the boundary condition that all points of a conducting body must be at the same potential . at this juncture
it is instructive to reconsider the distribution of charge along the filament _ ab initio _ using energy methods .
specifically , we will look for that distribution of charge on the filament which minimizes the electrostatic energy , given that charges are free to move within a conducting medium . as a preliminary step
, we will provide a proof that the state of minimum energy for an assembly of charged conductors is one of uniform electrostatic potential throughout each conducting body @xcite .
consider a number of charged , insulated conductors in electrostatic equilibrium @xcite .
the charge on the @xmath156th conductor will be denoted by @xmath157 and the region it occupies by @xmath158 . since the charges are confined to the conductors , we may express the electrostatic energy of the system as @xmath159 where @xmath160 is the ( volume ) charge density and the integral is over the indicated region .
furthermore , since the total charge on each conductor is given and fixed , we have @xmath161 rewriting eq .
( [ 30 ] ) by expressing @xmath162 in terms of @xmath163 , we find @xmath164 the problem before us then is to minimize @xmath165 in eq .
( [ 32 ] ) subject to the constraints expressed in eq .
( [ 31 ] ) . following standard procedure , we associate a lagrange multiplier @xmath166 with the @xmath156th constraint and seek to minimize the quantity @xmath167
\label{33}\end{aligned}\ ] ] with respect to variations in the multipliers @xmath168 as well as the charge density @xmath160 . setting the first - order variation with respect to @xmath160 equal to zero
gives @xmath169 the analogous equations for the multipliers simply reproduce the constraint equations of eq .
( [ 31 ] ) .
note that if @xmath170 for any @xmath156 , we get @xmath171 instead of eq .
( [ 34 ] ) .
we conclude from eq .
( [ 34 ] ) that electrostatic energy is least when each @xmath158 is an equipotential region , i.e. , there s a uniform potential throughout the body of each conductor .
this immediately implies the vanishing of the electric field as well as the charge density at all but surface points of each conductor . at surface points , on the other hand
, the equipotential condition implies that the electric field is normal to the surface of the conductor .
we have thus arrived at the familiar results of the electrostatics of conductors by energy minimization @xcite . at this point
the argument continues as in iiib and culminates in eq .
( [ 24 ] ) , which conveys the inverse - logarithmic approach of charge density to uniformity .
we have thus recovered the main result of this paper by means of energy considerations .
to illustrate the main results of the foregoing analysis , we shall present the cases of straight and circular ideal filaments as examples of eqs .
( [ 19 ] ) , ( [ 20 ] ) , and ( [ 23 ] ) in the following subsection , and a detailed numerical and graphical study of an elliptical conducting filament demonstrating the approach of charge density to uniformity with vanishing diameter , eq .
( [ 24 ] ) , in ivb .
our objective here is an analytical calculation of the fields of ideal straight and circular filaments for specified charge densities , and a comparison of their asymptotic behavior in the limit of @xmath29 with those of eqs .
( [ 19 ] ) and ( [ 20 ] ) . as in the derivation of these equations
, we will present the limiting behavior as an asymptotic , singular part that depends on @xmath26 , plus a remainder part that is finite and independent of @xmath26 .
we will also verify the asymptotic formula for capacitance , eq .
. we shall start with the simplest example of a charge filament , namely a straight line of uniform charge density @xmath120 extending from @xmath172 to @xmath173 .
then a straightforward calculation gives @xcite @xmath174^{1/2 } } { -(\frac{1}{2}+ \zeta)+{[{(\frac{1}{2}+\zeta)}^{2 } + { \epsilon}^{2}]}^{1/2 } } \biggr\ } , \label{35}\ ] ] where @xmath175 .
this result can be used to find the asymptotic behavior of the potential as @xmath58 : @xmath176 , \label{36}\ ] ] for the electrostatic potential at point @xmath177 of the filament , and @xmath178 for the electric field , valid for points not too close to the ends of the filament @xcite .
one can verify by inspection that these results agree with their general counterparts , eqs .
( [ 20 ] ) and ( [ 19 ] ) , as well as ( [ 23 ] ) , for the case of a uniformly charged straight filament .
the second example is also a straight filament but with a uniform dipole density , i.e. , @xmath179 , where @xmath180 is a constant .
then we find by straightforward integration @xmath181^{1/2 } } { -(\frac{1}{2}+ \zeta)+{[{(\frac{1}{2}+\zeta)}^{2 } + { \epsilon}^{2}]}^{1/2 } } \biggr\ } \nonumber \\ & & + k g l
\bigl\ { \bigl[{(\frac{1}{2}-\zeta)}^{2 } + { \epsilon}^2\bigr]^{1/2 } - \bigl[{(\frac{1}{2}+\zeta)}^{2 } + { \epsilon}^2\bigr]^{1/2 } \bigr\}. \,\,\,\,\,\ , \label{38}\end{aligned}\ ] ] taking the limit of this equation as @xmath58 , we find @xmath182z , \label{39}\ ] ] for the electrostatic potential at point @xmath177 of the filament , and @xmath183{\boldsymbol{\hat{\tau } } } \nonumber \\ & & + 2k g\,{\lim}_{\epsilon \rightarrow 0 } [ \ln({\epsilon } ) { \boldsymbol{\hat{\tau } } } + z{({\epsilon l})}^{-1 } { \mathbf{\hat{n}}}(\alpha ) ] , \label{40}\end{aligned}\ ] ] for the electric field , valid for points not too close to the ends of the filament .
as in the above , inspection shows that these results agree with the general ones given in eqs .
( [ 20 ] ) and ( [ 19 ] ) .
our third example is the case of a circular filament of radius @xmath184 and uniform charge density @xmath185 , centered at the origin of a polar coordinate system @xmath186 . by symmetry ,
the electrostatic potential of such a filament is independent of @xmath187 and its axial electric field ( i.e. , the component parallel to @xmath188 ; see fig .
1 ) is identically zero . on the other hand , unlike straight filaments , the potential of a circular one has an azimuthal variation and a corresponding electric field component , as will be seen below . using the geometry depicted in fig . 1
, we find for the electrostatic potential of a circular filament @xmath189 + { \epsilon}^2 \ } } ^{-1/2 } \nonumber \\ & & \times k \big [ { \big ( [ 4(1-\epsilon \cos(\alpha ) ] / \ { 4[1-\epsilon \cos(\alpha ) ] + { \epsilon}^2 \ } \big ) } \big ] , \,\,\,\,\,\,\ , \label{41}\end{aligned}\ ] ] where @xmath190 is the complete elliptic integral of the first kind defined as @xmath191}^{-1/2}. \label{42}\ ] ] note that the azimuthal angle @xmath31 equals zero at points @xmath192 , and it equals @xmath193 at points @xmath194 .
the limiting behavior of the potential in eq .
( [ 41 ] ) can be found using the asymptotic formula @xmath195 $ ] , @xmath196 . keeping terms up to order @xmath115 ,
we find @xmath197 \ln(8/\epsilon ) -\frac{1}{2}\cos(\alpha)\epsilon \bigr\}. \label{43}\ ] ] up to terms of order @xmath198 , the right - hand side of eq .
( [ 41 ] ) reduces to @xmath199 , in agreement with eq .
( [ 20 ] ) .
note also that the capacitance of the circular filament is given by @xmath200 , which leads to @xmath201^{-1}$ ] for its capacitance per unit length , in agreement with eq .
( [ 23 ] ) as @xmath29 .
the electric field of the circular filament is found by calculating the negative gradient of eq .
( [ 43 ] ) .
the result is @xmath202 \nonumber \\ & & + { \lim}_{\epsilon \rightarrow 0}[\cos(\alpha)\ln(\epsilon ) + 2{\epsilon}^{-1 } ] \bigr ] { \mathbf{\hat{n}}}(\alpha,\theta ) \nonumber \\ & & -\sin(\alpha)[1-\ln(8)+ { \lim}_{\epsilon \rightarrow 0 } \ln(\epsilon ) ] \boldsymbol{\hat{\alpha}}(\alpha,\theta ) \bigr\ } , \label{44}\end{aligned}\ ] ] which , given the substitutions @xmath203 and @xmath204 , agrees with eq .
( [ 19 ] ) .
it is instructive to find the azimuthal variations of surface charge density for the circular filament using eq .
( [ 44 ] ) . following the procedure given in
iiib , we find @xmath205 /\partial \alpha = { \lim}_{\epsilon \rightarrow 0}\frac{1}{2 } \sin(\alpha ) \epsilon \ln({\epsilon}^{-1 } ) , \label{45}\ ] ] in agreement with the general result in eq .
( [ 29 ] ) . in particular
, this result confirms our conclusions in iiib that azimuthal nonuniformities vanish relatively rapidly as @xmath29 . as stated earlier
, the three cases treated above serve to exemplify the general asymptotic results derived in ii and iii . to illustrate the approach of charge density to uniformity for a curved filament whose axial field ( the electric field component parallel to @xmath188 ; see fig .
1 ) does not vanish identically due to symmetry , we will consider an elliptical filament and numerically investigate the behavior of the charge density and electrostatic potential as @xmath29 .
we will also calculate the capacitance of the filament as a function of its diameter ( thickness ) in that limit .
consider an the elliptical filament of axes @xmath206 centered at the origin of a polar coordinate system @xmath186 with its major axis at @xmath207 .
using the coordinate system depicted in fig . 1 , we will numerically calculate the axial component of the electric field for the first quadrant , @xmath208 , at the azimuthal angle @xmath209 , with @xmath26 decreasing from @xmath210 to @xmath211 . recall that the filament thickness is given by @xmath212 , where in this case @xmath213 is the minimum radius of curvature for the ellipse which obtains at points @xmath207 , the points of extremal distance from the foci .
note that @xmath214 corresponds to subatomic dimensions for a filament of , say , @xmath215 m. starting with the general representation of the axial field in eq .
( [ 5 ] ) , we make the replacements @xmath216 , @xmath217 , @xmath218 , where we have associated cartesian unit vectors with the polar plane in the standard manner .
note that @xmath219 , where @xmath220^{1/2}$ ] .
the three unit vectors constituting the triad of fig . 1 correspond here to @xmath221 as the unit tangent to the ellipse in the counter - clockwise direction , @xmath222 as the inward unit normal to the ellipse in the polar plane , and @xmath223 as the unit binormal , which is a fixed vector in this case .
note that we have chosen to look at the potential and the @xmath224-component of the electric field at the azimuthal angle @xmath225 , corresponding to the filament s `` ridge line . ''
however , the choice of this angle makes little difference for a sufficiently thin filament . for the charge density @xmath53
, we assume the form @xmath226 suggested by eq .
( [ 24 ] ) , and approximate @xmath227 by a three - term fourier series .
ideally , the charge density would be determined by the requirement that the potential @xmath228 is uniform , in which case we would find an @xmath26-dependent result , to wit , @xmath229 .
we have determined the three fourier coefficients by minimizing the variations of the potential at @xmath230 , and expect the potential to be reasonably uniform at @xmath230 , and less so at other values of @xmath26 .
the primary objective of the numerical investigation is to show that as the diameter of the filament decreases , the charge density tends to uniformity while the potential remains essentially uniform .
the results shown here confirm this expectation as the relative variations of the charge density and potential are found to be of the order of @xmath231 and @xmath232 for @xmath230 .
it is worth recalling here that the approach of charge density to uniformity is inverse - logarithmic , hence extremely slow .
( short dash , red ) , @xmath233 ( long dash , green ) , and @xmath230 ( solid , blue ) .
note the logarithmically slow rate of convergence to uniformity .
] figure 2 is a plot of charge density at three values of @xmath26 , namely @xmath232 ( short dash ) , @xmath234 ( long dash ) , and @xmath235 ( solid ) , for which the relative deviations of charge density from uniformity are expected to be in the ratio of @xmath236 respectively .
the slow rate of convergence to uniformity well known from studies of straight filaments is clearly in evidence here . , for the first quadrant of the elliptical filament , at @xmath237 ( short dash , red ) , @xmath233 ( long dash , green ) , and @xmath230 ( solid , blue ) .
] the behavior of the electrostatic potential for the same values of @xmath26 as above are shown in fig .
3 , where we have normalized the potential to unity at @xmath238 .
this is of course the quantity that must be uniform for a conducting filament .
indeed the potential varies about @xmath239 at @xmath230 and @xmath240 at @xmath237 .
given the approximate nature of our fit to @xmath229 , we find the results shown in fig . 3 in agreement with our expectations .
it is worth recalling that a strictly uniform charge density will in general produce a nonuniform potential along the filament regardless of how small its diameter is .
this nonuniformity is reflected in the axial part of @xmath112 in eq .
( [ 19 ] ) for the general case , and exemplified by the first term of eq .
( [ 37 ] ) for the straight filament , both of which represent the axial electric field in the limit of @xmath241 for the case of uniform charge density . that such a residual axial electric field persists in the limit of vanishing diameter
is of course the puzzle raised in ref .
@xcite and explained in detail in ref .
whether straight or curved , the residual axial field is caused by asymmetric contributions from charges not in the immediate neighborhood of a given point and requires an appropriate amount of local dipole density , @xmath242 , to counteract it .
points that receive symmetric contributions from all charges , such as the midpoint of a straight filament and points @xmath243 and @xmath244 of the elliptical filament ( as well as at the equivalent points @xmath245 and @xmath246 not shown in the figures ) , are therefore exceptions to the general statement . as seen in fig .
3 , the potential is stationary at these exceptional points for the three values of @xmath26 represented .
orders of magnitude ; see text . ] finally , remembering that maxwell s original contribution focused on the capacitance of the straight filament , we show the capacitance per unit length of the elliptical filament ( normalized to that of a circular one whose radius matches the minimum radius of curvature of the elliptical filament ) in fig . 4 for @xmath26 ranging from @xmath231 to @xmath235 . recall from
iic , eq .
( [ 23 ] ) , that as @xmath29 , the capacitance per unit length of a regular filament approaches a universal form that only depends on its diameter .
this result is born out by the behavior seen in fig .
4 , which also echoes the slow rate of convergence to uniformity characteristic of the problem .
we have posted an interactive version of the computer program used for the above calculations for use by interested readers @xcite .
it is a code written in _
mathematica _ and annotated by explanatory comments .
the user can manipulate the results or change the parameters , subject to the limitations resulting from the nearly singular nature of the integrations .
the exterior field of an infinitely long , conducting , charged , circular cylinder does not depend on its diameter , and is therefore the same as if its charges were to be concentrated on its axis .
when finite length effects for a thin conductor are taken into account , the correction terms to the uniform charge density of the infinite case turn out small , except near the endpoints , and vanish as the inverse of the logarithm of the ratio of the diameter to length of the conductor @xcite .
what we have found here is that the same conclusion holds for the charge density along a sufficiently smooth conducting filament even if curved , albeit with appropriate modifications arising from the nonzero curvature of the filament .
the intuition behind this result is the observation that at a point where the diameter of the filament is sufficiently small compared to either its length or radius of curvature , the filament is locally indistinguishable from a straight one . in both curved and straight cases ,
the leading behavior arises from local charge and dipole densities while the correction terms originate from charges farther away . beyond the main uniformity result ,
the two cases differ in important details .
for example , while the charge distribution on any straight filament can be described by an appropriate scaling of a universal function , such is not the case with curved filaments which can assume an enormous variety of forms . in particular ,
a curved filament can be closed , in which case complications introduced by endpoints would not be present . the regularity conditions imposed on the geometry of the filament in
iia are not mere mathematical niceties since the uniformity of charge distribution on the filament will break down in the presence of a corner or a kink on the filament @xcite . however , although we have excluded such points altogether , it should be noted that the uniformity results of this paper are still valid at any point of the filament whose neighborhood conforms to the smoothness conditions , even if they are violated elsewhere .
d. j. griffiths and y. li , `` charge density on a conducting needle , '' am . j. phys .
* 64 * , 706 ( 1996 ) .
r. h. good , `` comment on ` charge density on a conducting needle , ' '' am . j. phys . * 65 * , 155 ( 1997 ) .
m. andrews , `` equilibrium charge density on a conducting needle , '' am . j. phys .
* 65 * , 846 ( 1997 ) .
j. d. jackson , `` charge density on thin straight wire , revisited , '' am .
* 68 * , 789 ( 2000 ) .
j. d. jackson , `` charge density on a thin straight wire : the first visit , '' am .
* 70 * , 409 ( 2002 ) .
j. c. maxwell , `` on the electrical capacity of a long narrow cylinder , and of a disk of sensible thickness , '' proc .
london math .
ix * , 94 ( 1878 ) .
note that this regularity condition ( existence of a non - vanishing , continuously differentiable tangent ) is stronger than smoothness , which requires the existence of a non - vanishing , continuous tangent .
stated in a mathematically proper manner , this is our _ definition _ of a curved , circular cylinder .
we will refer to a charged filament statisfying these requirements as simply `` regular . ''
needless to say , smoothness characteristics of @xmath53 under equilibrium conditions are determined by those of @xmath1 and are not an independent issue .
the device of using a line charge to describe the fields of the physical filament was also used by andrews @xcite and jackson @xcite . here the length parameter for the physical conductor is defined to be the same as that of the curve @xmath1 within .
r. j. rowley , `` finite line of charge , '' am .
* 74 * , 1120 ( 2006 ) .
w. r. smythe , _ static and dynamic electricity _ ( mcgraw - hill , new york , 1968 ) , 3rd ed .
123 - 124 .
recall that @xmath247 is the diameter of the filament in units of @xmath28 , the smaller of the length and minimum radius of curvature of the filament , and that @xmath248 is a pure number that never exceeds unity .
this result is traditionally established by the argument that electrostatic equilibrium precludes the motion of charges within a conductor . if present , grounded conductors can be regarded as one with the earth and treated as another insulated conductor for the purpose of this derivation .
the vanishing of the first - order variations guarantees a stationary configuration , not necessarily a minimum .
however , the minimum nature of the result here is clear on physical grounds , and can also be ascertained mathematically by examining the second - order variations .
calculations of the fields of straight filaments appear in various forms in the literature .
see , e.g. , refs .
@xcite . specifically , we require that @xmath249 .
mathematically , the controlling scale parameter @xmath28 , the smaller of the length and minimum radius of curvature of the filament , would cease to exist , or perhaps vanish , at such points , in violation of our regularity conditions in iia .
physically , a `` corner '' essentially amounts to a point where the radius of curvature of the filament is smaller than , or comparable to , its diameter , so that @xmath250 , far from the limit required for uniformity .
the _ mathematica 7 _ program is posted to the url http://www.csus.edu/indiv/p/partovimh/pg.nb . | equilibrium charge distribution on a thin , straight conductor has been actively discussed in recent years , starting with griffiths and li @xcite , who suspected a uniform distribution , and culminating with maxwell s insightful words @xcite as reported by jackson @xcite .
the present work addresses the electrostatic problem of a thin , curved , cylindrical conductor , or a conducting _ filament _ , and shows that the corresponding linear charge density slowly tends to uniformity as the inverse of the logarithm of a characteristic parameter which is the ratio of the diameter to the smaller of the length and minimum radius of curvature of the filament .
an alternative derivation of this result directly based on energy minimization is developed .
these results are based on a general asymptotic analysis of the electric field components and potential near a charge filament in the limit of vanishing diameter whereby the divergent parts of these quantities are explicitly calculated .
it is found that the divergent parts of the radial and azimuthal electric field components , as well as the electric potential , are determined by the local charge density while the axial component is determined by the local dipole density . for a straight filament ,
these results reduce to those for conducting needles discussed in the literature . for curved filaments ,
the configuration of charges and fields is no longer azimuthally symmetric , and there is an additional length scale in the problem arising from the finite radius of curvature of the filament .
remarkably , the basic uniformity result survives the added complications , which include an azimuthal variation in the surface charge density of the filament . as with the variations of linear charge density along the filament ,
the azimuthal variations vanish with the characteristic parameter , only more rapidly .
these uniformity results yield an asymptotic formula for the capacitance of a curved filament that generalizes maxwell s original result
. the examples of a straight filament with uniform and linearly varying charge densities , as well as a circular filament with a uniform charge distribution , are treated analytically and found to be in agreement with the results of the general analysis .
numerical calculations illustrating the slow convergence of linear charge distribution to uniformity for an elliptical filament are presented , and an interactive computer program implementing and animating the numerical calculations is provided . |
the study of quantum phase transitions ( qpt ) remains a very active topic in many fields of physics , spurred by experimental progress to create novel tunable interacting systems .
qpt occur in quite different materials , including heavy fermion compounds , * ? ? ?
* unconventional superconductors , * ? ? ?
* mott insulators , * ? ? ?
* coupled spin systems , * ? ? ?
* and ultracold atoms . * ? ? ?
* in particular , the common phenomenon of bose einstein condensation ( bec ) of strongly interacting bosons by tuning the interaction or the chemical potential can now be found in a range of different physical systems .
ultracold atomic gases allow the tuning of interactions via feshbach resonances , * ? ? ?
* but also cross - dimensional phase transitions @xcite and berezinsky - kosterlitz - thouless ( bkt ) behavior @xcite have been observed recently .
phase transitions in coupled spin dimer systems are prime examples of bec of strongly interacting triplons,@xcite which allow easy tuning of the chemical potential via the magnetic field .
although qpt s occur at zero temperature as a function of a non - thermal control parameter such as the interaction , effective mass , or the chemical potential , a characteristic critical scaling with temperature can be observed in a large range above the critical point.@xcite in general a detailed analysis is necessary in order to understand how the critical behavior is reflected in the experiments and if the finite - temperature phase transition is affected in the vicinity the qpt , where thermal fluctuations are comparable to quantum fluctuations . compared to bosonic gases of atoms and magnons the temperature control is relatively easy in triplon gases , which allows a systematic analysis of the critical scaling behavior near the qpt . in this paper
we focus on the theoretical analysis of quantum critical points of antiferromagnetic spin dimer systems which are weakly coupled in two - dimensions .
two qpt s can be observed : as the field is increased through the lower critical value @xmath0 the spin dimers start to be occupied by triplons and the magnetization increases with characteristic two - dimensional logarithmic behavior .
the second qpt corresponds to the saturation field @xmath1 .
the intermediate phase is characterized by long - range phase coherence of triplons at @xmath2 and bkt behavior@xcite at finite @xmath3 .
similar phase transitions occur in two - dimensional hard - core boson systems@xcite and in distorted frustrated lattices.@xcite the schematic behavior is illustrated in fig .
[ scheme ] . in this paper
we show that the crossover from bkt behavior to critical scaling is rather well defined by the cooling rate and by characteristic maxima in the susceptibility .
however , this crossover occurs at distinctly higher temperatures than the bkt transition which can be determined by a careful analysis of the spin - stiffness .
there is no directly measurable signal for the bkt transition in experiments,@xcite but we find that magneto - caloric measurements are ideally suited to show the critical scaling and pinpoint the exact location of the qpt .
close to the qpt the bkt transition retains the characteristic logarithmic behavior , albeit with strongly renormalized parameters .
we find , however , that the low temperature behavior above the qpt s does not fully follow theoretical expectations .
we use a `` columnar '' arrangement of strongly coupled antiferromagnetic dimers ( @xmath4 ) on a two dimensional square lattice as shown in fig .
[ fig : squarelattice ] , described by the hamiltonian of localized spin-1/2 operators @xmath5 @xmath6 - b \sum_{i=1}^{n } \hat{s}_{i}^{z } , \end{split } \label{eqn : hxxz}\end{aligned}\ ] ] where the inter - dimer couplings @xmath7 and @xmath8 can be ferromagnetic or antiferromagnetic , but are assumed to be small @xmath9 .
assuming that the intra - dimer exchange interaction @xmath10 dominates over inter - dimer couplings @xmath7 and @xmath8 , it is natural to represent the system in the singlet and triplet basis at each dimer site @xmath11 at strong fields @xmath12 the last two states become nearly degenerate , while the other two higher energy states will be neglected for now .
it is therefore justified to work in a restricted hilbert space with only two states at each dimer site , which are represented by hard - core bosons on the vacuum @xmath13 and @xmath14 . in this hilbert space
the effective hamiltonian describes strongly interacting bosons on a rectangular lattice @xmath15 \label{heff1}\\ & & - \mu \sum_i n_i + u \sum_i n_i ( n_i-1 ) , \label{heff}\end{aligned}\ ] ] where the limit @xmath16 is implied to satisfy the hardcore constraint .
the effective chemical potential and the hopping in @xmath17 and @xmath18directions are given by @xmath19 note , that the hopping @xmath20 in eq .
( [ heff1 ] ) has been chosen to be positive , which can always be achieved by a local gauge transformation @xmath21 .
the nearest neighbor interaction in eq .
( [ heff1 ] ) is repulsive ( attractive ) for @xmath22 ( @xmath23 ) . by fourier
transforming the first term in the hamiltonian the kinetic energy becomes @xmath24 the position of the upper and lower band edges allows a straight - forward estimate of the critical fields @xmath0 and @xmath1 .
the lower critical field is determined by the chemical potential at which a single boson acquires positive energy @xmath25 , which gives @xmath26 this estimate is only correct to first order in @xmath27 , however , since the bosonic ground state ( vacuum ) is not an exact eigenstate of the full hamiltonian in eq . ( [ eqn : hxxz ] ) .
higher order corrections from the neglected triplet states @xmath28 and @xmath29 in eq .
( [ states ] ) will be determined from numerical simulations as described below .
the upper critical field is determined from the energy gain of removing a particle from the fully occupied band including the nearest neighbor interaction energy @xmath30 which is exact and corresponds to the saturation field of the original model ( [ eqn : hxxz ] ) . for intermediate fields @xmath31 the physics
is governed by the behavior of two - dimensional interacting bosons ( bkt phase ) as explained below .
we now focus on the lower qpt at @xmath0 which corresponds to the well studied case of a dilute interacting bose gas.@xcite at low filling the critical behavior of the lattice model in eqs .
( [ heff1])-([heff ] ) is believed to be in the xy - universality class independent of the microscopic details . * ? ? ? * in the continuum limit the nearest neighbor interaction can be neglected and the hard - core constraint can be replaced by a strong @xmath32 interaction of a complex bosonic field @xmath33 in a @xmath34-dimensional euclidean action @xmath35\text { , } \label{eqn : action}\end{aligned}\ ] ] where @xmath36 in our case .
the parameters can be obtained by approximating the sums in eqs .
( [ heff1])-([heff ] ) by integrals with a lattice spacing @xmath37 and then rescaling @xmath38 @xmath39 in what follows we set the lattice spacing to unity @xmath40 . the action in eq .
describes an interacting dilute bose gas with mass @xmath41 and chemical potential @xmath42 . for @xmath43 or @xmath44 a finite density of bosons appears even at zero temperature @xmath2 , which signals the qpt to the bkt phase .
analogously , the same model also applies at the upper critical field @xmath1 , where it describes bosonic singlet excitations on the saturated state with @xmath45 .
the upper critical dimensions is @xmath36 for this model so that logarithmic corrections appear in this case , which are described in terms of an ultraviolet cutoff @xmath46 ( of the order of the reciprocal rescaled lattice spacing ) .
this situation ( @xmath36 ) has been analyzed extensively in the literature@xcite for various quantities which we summarize below .
other dimensions are discussed in the textbook of sachdev.@xcite the density of bosons corresponds to the magnetization per site @xmath47 in the spin dimer system as a function of field @xmath48 , which is given at @xmath2 by@xcite @xmath49 \text{. } \label{mb}\ ] ] the susceptibility is therefore @xmath50 - 1 \right)\label{eqn : suscyboson2d}\,,\end{aligned}\ ] ] which is logarithmically divergent as the critical point is approached from above inside the bkt phase @xmath51 . for @xmath52 and @xmath53 it has been predicted that the density increases with temperature including a characteristic logarithmic correction@xcite @xmath54.\label{mt}\ ] ] the scaling as a function of @xmath3 can be used in order to identify the exact value of the critical field @xmath0 as outlined below .
finally , the bkt transition temperature has been predicted as a function of field@xcite @xmath55 } { \ln\left [ \ln\left [ \frac{\lambda_0 ^ 2}{2 m\tilde\mu } \right ] \right]}\label{eqn : tbkt}\,.\end{aligned}\ ] ] however , for this formula to be valid the double logarithm has to become very large , which does not correspond to physically relevant regions.@xcite in fact , it remains to be seen if the single logarithms in eqs .
( [ mb])-([mt ] ) are large enough so that the leading behavior can be observed in our numerical simulations below and in future experiments .
in order to analyze the quantum phase transitions , the exact locations of the critical fields have to be determined first . as mentioned above , the upper critical field @xmath1 is exactly the saturation field in eq .
( [ bs ] ) , but the lower field in eq .
( [ bc ] ) will in general have higher order corrections of the form @xmath56 the higher order corrections are due to virtual excitations to the neglected triplet states @xmath28 and @xmath29 in eq .
( [ states ] ) .
the exact values for @xmath57 and @xmath58 are known from higher order strong coupling expansions for the dimerized chain@xcite @xmath59 and for the ladder system@xcite @xmath60 , respectively . in order to determine the exact location of the qpt for general inter - dimer couplings , numerical simulations at @xmath2 in the thermodynamic limit
would be required .
this is obviously impossible , but large systems sizes at small finite temperatures are feasible with quantum monte carlo ( qmc ) simulations . in order to examine the model in eq .
( [ eqn : hxxz ] ) numerically , we therefore have implemented the stochastic series expansion algorithm@xcite with directed loop updates and using the so - called mersenne twister random number generator.@xcite at finite temperatures the discontinuity in eq .
( [ mb ] ) can not be observed directly , but the magnetization as a function of temperature becomes exponentially small for @xmath61 while it approaches a finite value for @xmath44 .
the critical field @xmath0 is then exactly defined as the point where critical scaling is obeyed , which can be determined rather accurately .
this behavior is illustrated in fig .
[ fig : magt26 ] .
0.9c|c|c|c|c case & @xmath62 @xmath63 $ ] & @xmath64 @xmath63 $ ] & @xmath65 $ ] & from eqs .
+ & & & @xmath66 & ( [ bc2 ] ) , ( [ coeff ] ) + @xmath67 & @xmath68 & @xmath69 & @xmath70 & 0.84625 + @xmath71 & @xmath72 & @xmath72 & @xmath73 & 0.83875 + @xmath74 & @xmath75 & @xmath76 & @xmath77 & 0.85225 + note , however , that the observed scaling in fig .
[ fig : magt26 ] at the exactly known upper critical field @xmath1 appears to be perfectly linear ( relative to the saturated state ) .
this means that the logarithmic correction in eq .
( [ mt ] ) must be very small , which puts a lower limit on the cutoff @xmath78 . to determine the lower critical field @xmath0
, we therefore use linear scaling as well . extrapolating the data to the thermodynamic limit and then determining the critical fields @xmath0 by the best linear fit gives the results for three different choices of inter - dimer couplings shown in table [ table : hoppings ] . ignoring higher orders , the values for the coefficients in eq .
( [ bc2 ] ) are then consistent with the following estimates @xmath79 before continuing our analysis we would also like to consider how the neglected higher energy triplet excitations @xmath28 and @xmath29 in eq .
( [ states ] ) affect physical observables like the magnetization .
we note that the effective hamiltonian ( [ heff1])-([heff ] ) is invariant under changing the inter - dimer coupling strengths @xmath7 and @xmath8 as long as all energies and the field @xmath80 are rescaled accordingly .
we therefore consider three different realization of the coupling strength @xmath81 , @xmath82 , @xmath83 and plot the susceptibility @xmath84 as a function of rescaled field @xmath85 at a given rescaled temperature @xmath86 in fig .
[ fig : univers1 ] .
we observe a finite susceptibility in the bkt phase with two characteristic maxima near the qpt .
while the three curves agree reasonably well , systematic deviations can be seen for larger @xmath27 , which can only come due to corrections from the higher energy triplet excitations . in what follows
we choose a coupling strengths of @xmath87 , which is a compromise between minimizing higher order corrections and efficient numerical simulations .
it is believed that the higher order triplet excitations do not change the form of the critical scaling in eqs .
( [ mb])-([eqn : tbkt ] ) . as a function of @xmath88 at inverse temperature @xmath86 for three inter - dimer coupling strengths
@xmath81 , @xmath82 and @xmath83 ]
we now turn to analyzing the scaling behavior of the susceptibility @xmath89 as a function of field @xmath90 in eq .
( [ eqn : suscyboson2d ] ) .
finite temperatures @xmath3 and system sizes @xmath91 play the role of an infrared cutoff @xmath92 which will give deviations from the predicted @xmath2 scaling in eqs .
( [ mb ] ) and ( [ eqn : suscyboson2d ] ) as the qpt is approached .
however , for fields @xmath93 the scaling can still be tested . at
each given temperature we first increase the system size until systematic convergence of the magnetization is obtained as shown in fig [ fig : scaling1 ] .
the resulting susceptibility in the thermodynamic limit near the qpt s is shown in fig .
[ fig : scaling2 ] as a function of the logarithm of @xmath94 for different temperatures .
the data confirms that the scaling approaches a logarithmic behavior for @xmath95 consistent with the form in eq .
( [ eqn : suscyboson2d ] ) .
we notice that the finite temperature susceptibility is actually rather small at the qpt @xmath53 , but then increases and _ overshoots _ the logarithmic divergence , before the logarithmic behavior is reached inside the bkt phase . in this way the field - integral of the susceptibility ( i.e. the magnetization ) remains largely temperature independent outside the critical region , since the smaller values at the qpt for finite @xmath3 are compensated by a corresponding overshooting of the maximum . in turn
this means that the characteristic maxima in fig .
[ fig : univers1 ] of the susceptibility are only indirectly related to the qpt .
the overshooting implies that the large fluctuations in the magnetization arise from a different mechanism at finite temperatures .
one may expect that the maxima are therefore related to the finite - temperature bkt transition , but this is _ not _ the case as we will see below .
instead we find that the susceptibility maxima are found for temperatures well above the bkt transition @xmath96 at the corresponding fields .
as we will see later the maxima coincide with maxima in the entropy , so that these points correspond to the crossover between quantum critical scaling to vortex physics . comparing with the expected form in eq .
( [ eqn : suscyboson2d ] ) quantitatively , we find rather small values of the effective mass @xmath97 at the lower qpt @xmath0 and @xmath98 at @xmath1 , which are strongly renormalized compared to the naive estimate @xmath99 according to eq .
( [ rescaling ] ) .
the value of @xmath100 remains finite in eq .
( [ eqn : suscyboson2d ] ) .
the value of @xmath41 from the fits at the lower qpt is rather sensitive to the exact location of the critical field @xmath0 . in general
all microscopic details such as the neglected next - nearest neighbor interaction in eq .
( [ heff1 ] ) will influence the exact value of the effective parameters in eq .
( [ rescaling ] ) .
the intermediate region between the two qpt s is dominated by the presence of interacting triplon excitations which form a condensate at @xmath2 with long range phase coherence .
we now consider the finite temperature behavior in this intermediate phase .
while the qpt s are driven by quantum fluctuations , the transition due to thermal fluctuation corresponds to classical behavior and is therefore not directly related to the scaling discussed above . the effective hard - core boson model in eqs .
( [ heff1])-([heff ] ) is exactly equivalent to the xxz - spin model with @xmath101 , which is known to be in the xy - universality class . at finite temperatures this system undergoes a bkt transition , which can be described in terms of classical two - dimensional spins as first explained in the works of berezinsky@xcite and kosterlitz and thouless.@xcite at low temperatures @xmath102 a quasi long - range ordered phase with power - law decay of correlations exists . above the phase transition temperature
@xmath103 the unbinding of vortices is energetically allowed leading to a disordered phase with exponential decaying correlations .
kosterlitz used the spin stiffness@xcite @xmath104 to identify a phase transition , where @xmath105 is the free energy of the system and @xmath106 is the angle between spins at opposite edges of the system . in order to determine the spin stiffness in qmc simulations
it is convenient to calculate the winding number fluctuations in each direction,@xcite which can be used to define a so - called helicity modulus@xcite @xmath107 the phase transition temperature @xmath103 is then determined by the value where the helicity modulus obeys@xcite @xmath108 generally the helicity modulus has a more complex relation to the spin - stiffness for anisotropic systems in low dimensions @xmath109 as discussed by prokofev and svistunov.@xcite for anisotropic systems it is therefore useful to define a separate helicity modulus for each direction@xcite @xmath110 where @xmath111 and @xmath112 are the edge lengths of the effective hard - core boson system in terms of the size of the original spin system @xmath113 . instead of taking the average in eq .
( [ eqn : definehelicity ] ) , only the largest one contributes @xmath114 , while the smaller one shows a linear behavior with edge length @xmath115.@xcite the energy of the vortices also depend logarithmically on the system size @xmath116 , so that the condition in eq .
( [ eqn : helidrop ] ) acquires a corresponding correction@xcite @xmath117 where @xmath118 is a fitting parameter and @xmath119 should take on the universal value of unity at the transition , but can also be used as a fitting parameter.@xcite following the procedure in ref .
[ ] the logarithmic corrections in eq .
( [ eqn : bktfit2 ] ) become only accurate at the phase transition , which can in fact be used to determine @xmath103 and @xmath120 . in fig . [
fig : helicity ] the helicity modulus is plotted at a given field @xmath121 for different system sizes . the bkt transition for each field
is determined by the best fit of eq .
( [ eqn : bktfit2 ] ) , i.e. when @xmath122 extrapolates to a limiting value linearly as a function of @xmath123 . for a classical isotropic spin model
a value of @xmath124 can be confirmed,@xcite but for the spin dimer model we find a field - dependent value for @xmath120 which is slightly larger than unity as given in table [ table : gammainflnl0 ] and shown in fig .
[ fig : helicity ] .
the fitting parameter @xmath118 also becomes field dependent .
the resulting transition temperature is shown in fig .
[ fig : tbkt ] , which shows a sharp drop near the qpt . as shown in the inset
the behavior is consistent with a logarithmic behavior @xmath125 but the double logarithmic correction in the asymptotic scaling at extremely small densities in eq .
( [ eqn : tbkt ] ) can not be confirmed numerically.@xcite .results of @xmath120 and @xmath126 for different magnetic fields .
[ cols="^,^,^,^",options="header " , ] the deviations from @xmath124 can be traced to two different sources : in the middle of the bkt phase we find that a nearly isotropic effective system with @xmath127 and @xmath128 gives values of @xmath129 , so that the detailed geometry appears to have some effect on the exact value of @xmath130 .
a second source may be higher order corrections in @xmath131 , which can be expected to become significant when the effective density of bosons per lattice site is small , which in turn leads to large distances between vortices .
therefore , the corrections must be largest close to the qpt , consistent with our findings .
using a constant value of @xmath124 in the fits changes the estimate for @xmath103 by up to 10 - 15% . as a function of field for different temperatures .
bottom : temperatures as a function of field for different values of constant entropy ( red isentropes @xmath132 ) near @xmath0 .
the shaded region is dominated by large entropy , corresponding to the minima in the isentropes , which are relatively close to the maxima in the susceptibility ( marked by green squares ) .
the bkt transition @xmath103 ( connected dots ) occurs at significantly lower temperatures .
, title="fig : " ] as a function of field for different temperatures .
bottom : temperatures as a function of field for different values of constant entropy ( red isentropes @xmath132 ) near @xmath0 .
the shaded region is dominated by large entropy , corresponding to the minima in the isentropes , which are relatively close to the maxima in the susceptibility ( marked by green squares ) .
the bkt transition @xmath103 ( connected dots ) occurs at significantly lower temperatures .
, title="fig : " ]
as we already discussed in sec .
[ sec : criticalfields ] , the behavior of the magnetization @xmath133 as a function of temperature plays an important role in determining the locations of the qpt .
the interplay of magnetization with temperature is often termed magneto - calorics , which has been a fruitful field ever since the discovery of adiabatic demagnetization by warburg in 1881.@xcite the central quantity of interest in this context is the cooling rate @xmath134 which describes the temperature change with the applied field under adiabatic conditions . using the cyclical rule and a maxwell relation the cooling rate is also directly related to @xmath133 and @xmath135 @xmath136 where @xmath137 is the heat capacity . therefore , the entropy is largest when @xmath138 . the cooling rate for different temperatures is plotted in fig .
[ fig : gamma ] ( top ) , which shows sharp features near the qpt . in ref .
[ ] it was predicted that the cooling rate diverges with a universal prefactor near the qpt , but we were not able to reach low enough temperatures to confirm this behavior . integrating the cooling rate in eq .
( [ gamma ] ) gives the temperature as a function of field for a given entropy @xmath139 .
the corresponding isentropes are shown in fig .
[ fig : gamma ] ( bottom ) .
the temperature reaches a minimum when the cooling rate is zero , which means that the entropy as a function of field ( horizontal path ) is maximal .
it is interesting to notice that the points of maximum entropy @xmath140 are relatively close to the maxima of the susceptibility .
however , the maximum entropy region is _ not _ exactly at the value of the critical field as is the case for other systems without an ordered phase , as in the ising chain.@xcite nor are those points associated with the finite temperature bkt phase transition as would be the case for ordered systems in @xmath141.@xcite the situation in @xmath36 is therefore special , since in this case the sign change in the cooling rate @xmath140 signals a maximum in the entropy in the crossover region where quantum critical behavior competes with vortex excitations in the shaded parameter range in fig .
[ fig : gamma ] ( bottom ) . in the @xmath3-@xmath90 parameter space for @xmath142 .
the bkt transitions @xmath103 is marked by connected dots ( black ) , points of maximum entropy @xmath140 by diamonds ( violet ) and maxima in the susceptibility by squares ( green ) . ]
in summary , the magneto - caloric quantity @xmath143 turns out to be a universal indicator of the quantum critical behavior .
we plot this quantity in fig .
[ fig : tbphase ] in the relevant @xmath3-@xmath90 parameter space . on the one hand we have seen in sec .
[ sec : criticalfields ] that the critical scaling is defined by a linear behavior of @xmath144 , which leads to a constant and large derivative @xmath143 .
the regions with quantum critical behavior therefore show up clearly in fig .
[ fig : tbphase ] as the lightest and darkest region in the phase diagram above @xmath0 and @xmath1 , respectively .
the points of @xmath145 mark the boundaries towards regions , which are dominated by bkt vortex excitations .
these points coincide with the maxima in the susceptibility , but are _ not _ directly associated with the qpt or the finite temperature bkt phase transition .
the bkt phase transition occurs at significantly lower temperatures and is not reflected by any directly measurable thermodynamic quantity.@xcite nonetheless , the predicted and well established behavior of the spin stiffness at the bkt transition holds also for the dimer system , but strong corrections start to appear at small magnetization ( i.e. boson density ) as discussed in sec .
[ sec : bkt ] .
we would like to emphasize that magneto - caloric measurements of @xmath143 not only allow a detailed analysis of the qpt , but also are potentially a very useful experimental tool in order to identify the effective dimensionality of the underlying spin systems due to the different density of states .
in particular , for quasi one - dimensional systems @xmath146 shows a characteristic divergence above the qpt * * ? ? ? * * while for @xmath147 we find an increase @xmath148 analogous to the famous @xmath149 bloch law .
we find in our numerical simulations that @xmath36 is characterized by perfectly linear behavior above the qpt , i.e. @xmath150 without any detectable logarithmic corrections in contrast to the field theory prediction in eq .
( [ mt]).@xcite as discussed in sec .
[ sec : criticalfields ] this can be used to determine the exact positions of the critical field , which in turn allows the quantitative estimate of higher order terms in the analytical expressions as a function of the antiferromagnetic coupling constants .
[ sec : ackn ] we are thankful for useful discussions with achim rosch , markus garst , denis morath , and axel pelster .
this work was supported by the sfb transregio 49 of the dfg and the allianz fr hochleistungsrechnen rheinland - pfalz ( ahrp ) . | spin dimer systems are a promising playground for the detailed study of quantum phase transitions . using the magnetic field as the tuning parameter it is in principle possible to observe a crossover from the characteristic scaling near critical points to the behavior of a finite temperature phase transition . in this work
we study two - dimensional coupled spin dimer systems by comparing numerical quantum monte carlo simulations with analytical calculations of the susceptibility , the magneto - caloric effect , and the helicity modulus .
the magneto - caloric behavior of the magnetization with temperature can be used to determine the critical fields with high accuracy , but the critical scaling does not show the expected logarithmic corrections .
the zeros of the cooling rate are an excellent indicator of the competition between quantum criticality and vortex physics , but they are not directly associated with the quantum phase transition or the finite temperature berezinsky - kosterlitz - thouless transition .
the results give a unified picture of the full quantum and finite temperature phase diagram . |
two important forms of carbon , fullerene ( c@xmath0 ) and carbon nanotubes , are closely related to each other by their structural commonality of the sp@xmath3 framework .
carbon nanotubes have been widely investigated for the last decade or so as one - dimensional ( 1d ) nanomaterial @xcite .
in contrast , fullerene 1d nanostructures , such as c@xmath0 nanowires , presently represent laboratory curiosities @xcite .
c@xmath0 nanowires are highly interesting nanomaterials because of the physical properties associated with their high surface area , low - dimensionality and quantum confinement effects @xcite . in comparison with carbon nanotubes
, however , c@xmath0 nanowires are mechanically fragile and electrically non - conducting on account of the large intermolecular spaces in the crystalline lattice although an individual c@xmath0 molecule is extremely hard , almost incompressible and electronically fully conjugated across all the carbon atoms @xcite .
it is expected that , in contrast to their parent nanowires or individual fullerene building blocks , polymerized c@xmath0 nanowires will exhibit remarkably different and/or improved physical properties owing to the formation of polymeric chains and interchain cross - linking networks . from an applications point of view
, a major advantage of using such polyfullerenes would be their bio - compatibility as they are totally free of metal .
notably , this clearly contrasts to the case of carbon nanotubes , the growth of which is catalyzed by metal nanoparticles , and from which by no means all the metal can be removed by post - purification processes @xcite . in this respect
, c@xmath0 nanowire polymers could be biologically even more attractive than carbon nanotubes for uses in , for example , drug delivery and biosensor preparation @xcite .
additionally , 1d c@xmath0 polymers are conjugated or nearly - conjugated materials that promise outstanding photo - electrical properties because of the large magnitude of the nonlinear optical response of c@xmath0 and its excellent photoinduced charge transfer behaviour @xcite . as such
, they promise applications as optical limiters and photoconductors for solar energy devices , fuel cells , and field - emission transistors @xcite .
all this makes them important materials for further study .
fundamental to the development of applications is the question of how to make such metal - free , c@xmath0-based nanowire polymers ? in seeking to answer this question , we note first that a limited number of polymerization reactions of c@xmath0 have been reported in literature @xcite . for example
, it has been shown that the polymerization of thin solid films of c@xmath0 may occur upon irradiation with visible - uv light . in this case , the thin - films were prepared by vacuum deposition of pure c@xmath0 on a substrate @xcite . on the basis of raman scattering studies , such polymerization has been hypothesized to result from a [ 2 + 2 ] cycloadduction linkage between adjacent c@xmath0 molecules in the structure @xcite .
single crystals of kc@xmath0 , prepared by the co - evaporation of k and c@xmath0 , have undergone polymerization via formation of the same [ 2 + 2 ] structure motif , but the whole polymer is a linear polyanion network , as suggested by electron spin resonance ( esr ) measurements , pointing to a possible donation of electrons from k atoms to the polymer chains @xcite .
an in - chain metal - fullerene ( -pdc@xmath0-)@xmath1 polymer , synthesized by the reaction of c@xmath0 with a palladium complex , pd@xmath4[(c@xmath5h@xmath6ch = ch)@xmath4c = o ] chcl@xmath7 , has been noted @xcite .
moreover , side - chain c@xmath0 copolymers formed through the attachment of c@xmath0 to amino polymer substrate have also been reported , with the coupling of c@xmath0 to the amino polymer being analogous to that observed for amines @xcite .
it has also been demonstrated that crystallization of c@xmath0 molecules from organic solvents yields c@xmath0 nanocrystals with varied shapes and structures , and that these shapes and structures are dependent upon the crystallization conditions and the solvent(s ) employed @xcite .
one example is the preparation of c@xmath0 nanowhiskers with a claimed face - centered cubic structure using a liquid - liquid interface method @xcite .
this method has also been extended to prepare the nanowhiskers of c@xmath8 , iodine - doped c@xmath0 , and the c@xmath0-based complex of ( @xmath9-c@xmath0)pt(pph@xmath7)@xmath4 @xcite . in the second example , crystallization of c@xmath0 molecules from a 1,2,4-trimethylbenzene ( 1,2,4-tmb ) solution of c@xmath0 has been found to be able to produce exceptionally long c@xmath0 nanowires under appropriate conditions , with a length - to - width aspect ratio as large as 3000 , and an orthorhombic nanostructure @xcite . in spite of these literature reports , it is clear that there has been little study in the related field of nanowire - oriented c@xmath0 polymer .
this new class of 1d nanomaterial exhibits similarities in structure to other important conjugated polymers such as polyacetylenes , polyanilines , and polyphenylenes .
here we report the synthesis and characterization of such a c@xmath0-based nanowire polymer , ( -c@xmath0tmb-)@xmath10 , which is formed by the polymerization occurring in their parent nanowire via a topochemical solid - state reaction . because the reactive monomers are pre - organized at a distance commensurate with the repeat distance in the final polymer , the application of thermal or photochemical energies to the solid induces polymerization . in order to understand the polymerization pathway , we have employed gas chromatography ( gc ) , mass spectrometry ( ms ) and @xmath2c nuclear magnetic resonance ( nmr ) spectroscopy to investigate the nature of the bonds formed during the polymerization process . additionally , to explain the polymerization mechanism in more detail we have theoretically analyzed the associated molecular orientations , structural arrangements and the favorable reaction pathways .
gc - ms analysis was carried out using a perkin elmer turbomass gc - ms .
this system utilizes an autosystem xl gc connected to a small quadrupole mass analyzer .
samples were dissolved in ccl@xmath11 prior to analysis , and were introduced to the system using a split injection technique . after a delay of 1 minute the mass spectrometer scan was initiated with the parameters set out below .
gc conditions : injection volume , 0.5 @xmath12l ; injection temperature , 200 @xmath13c ; helium carrier gas pressure , 40 psi .
column : perkin elmer elite pe-5ms ( 5 % phenyl , 95 % methyl polysiloxane ) ; length , 30 m ; i d , 0.25 mm ; film thickness , 0.25 @xmath12 m .
column gradient : start temperature , 60 @xmath13c , for 1 minute then ramped at 10 @xmath13c / minute to 200 @xmath13c , which temperature was maintained for 2 minutes .
total analysis time was 17 minutes .
ms spectrometer conditions : ionization mode , ei+ ; ionization energy , 70 ev ; source temperature , 200 @xmath13c ; transfer line temperature , 200 @xmath13c ; mass range scanned , 50 - 650 da ; scan time , 0.5 s ; interscan time , 0.1 s ; solvent delay time , 1 minute .
@xmath2c nmr spectra were acquired at 100.62 mhz on bruker avance 400 standard bore ( liquid state ) and wide bore ( solid state ) spectrometers with standard bruker broadband tuneable probes .
samples for solid state nmr were packed into 4 mm zirconia rotors padded with polytetrafluoroethylene tape to take up excess space due to limited sample volumes .
relevant spectral acquisition parameters were as below .
solution - state nmr spectra were obtained using pulse - acquire with 10 s @xmath14/2 pulses , a recycle delay of 5 s , and broadband @xmath15h decoupling , and referenced relative to tetramethylsilane at 0 ppm .
solid - state nmr spectra were acquired with standard cross polarization magic angle spinning ( cp - mas ) ( 12.5 khz , 2.5 @xmath12s @xmath15h pulse , ramped cp time 5 ms , recycle delay 1 s ) and high power broadband 1h decoupling during acquisition ( fullerene nanowires ) .
pure fullerene was observed using direct polarization mas ( 5 ms @xmath2c @xmath14/2 pulse ) , recycle delay 60 s ) .
chemical shifts are referenced to the methylene carbon signal of glycine at 43.1 ppm which is back referenced to tetramethylsilane .
transmission electron microscopy ( tem ) and selected area electron diffraction ( saed ) were performed using a jeol jem-3010x microscope operated at 300 kv and a jeol 200cxi microscope operated at 200 kv , respectively .
specimens were directly deposited on copper grids coated with holey carbon films without any prior suspension treatment in solution .
scanning electron microscopic ( sem ) examinations were performed using a leo-32 electron microscope operated at 5 kv .
samples were directly deposited on a specimen holder ( carbon mat ) without surface coating of a conducting material .
raman spectroscopic studies were performed using a micro - raman spectrometer with incident ar ion laser at @xmath16=514.5 nm and the renishaw software .
extra care was taken to avoid laser damage to the samples . to ensure this ,
the laser power employed was kept as low as possible , and optical microscopic examinations were carried out for the nanowires before and after each measurement .
the structures and properties of the molecules involved in the nanowire polymerization process were studied theoretically .
the calculations were performed with the gaussian 03 software package @xcite , within the framework of the _ ab initio _ density functional theory ( dft ) . within the framework of the dft one has to solve the kohn - sham equations @xcite , which read as ( in the atomic system of units ) : @xmath17 here the first term represents the kinetic energy of the i - th electron , and @xmath18 describes its attraction to the ions in the system , while @xmath19 is the hartree part of the interelectronic interaction : @xmath20 and @xmath21 is the electron density : @xmath22 @xmath23 in eq .
( [ eq1 ] ) is the local exchange - correlation potential , @xmath24 are the electronic orbitals and @xmath25 is the number of electrons in the system .
the exchange - correlation potential is defined as the functional derivative of the exchange - correlation energy functional : @xmath26}}{{\delta \rho ( \vec r ) } } \label{eq4}\ ] ] the functionals employed by dft methods partition the exchange - correlation energy into exchange and correlation parts .
the local exchange functional is virtually always defined as follows : @xmath27 this form was developed to reproduce the exchange energy of a uniform electron gas . by itself , however , it is not sufficient for an adequate description of a many - body system .
a more accurate parametrization of the exchange functional introduced by becke @xcite , which includes gradient - corrections and reads as : @xmath28 where @xmath29 and @xmath30 is a parameter chosen to fit the known exchange energies of the noble gas atoms . in spite of the success of pure dft theory in many cases
, one has to admit that the hartree - fock theory accounts for the electron exchange more naturally and precisely .
thus , becke has suggested hybrid functionals @xcite which include a mixture of hartree - fock and dft exchange along with dft correlations : @xmath31 where @xmath32 and @xmath33 are constants .
following this idea , a becke - type three parameter functional ( b3lyp ) can be defined as follows : @xmath34 here , @xmath35 , @xmath36 and @xmath37 are constants , which were derived by fitting to the atomization energies , ionization potentials , proton affinities and first - row atomic energies @xcite .
@xmath38 and @xmath39 are defined in eqs .
( [ eq5 ] ) and ( [ eq6 ] ) respectively .
@xmath40 is the functional corresponding to hartree - fock equations
. the explicit form for the correlation functional @xmath41 as well as for gradient - corrected correlation functional of lee , yang and parr , @xmath42 , can be found in ref .
@xcite and ref .
@xcite respectively .
we have used the b3lyp functional for all computations in the present paper . in gaussian 03
, the molecular orbitals , @xmath43 , are approximated by a linear combination of a pre - defined set of single - electron functions , @xmath44 , known as basis functions .
this expansion reads as follows : @xmath45 where coefficients @xmath46 are the molecular orbital expansion coefficients , and @xmath25 is the number of basis functions , which are chosen to be normalized .
the basis functions @xmath47 are defined as linear combinations of primitive gaussians : @xmath48 where @xmath49 are fixed constants within a given basis set , the primitive gaussians , @xmath50 , are the gaussian - type atomic functions ( for details see ref .
@xcite ) . in the calculations we accounted for all electrons in the system , and employed the standard sto-3 g , 6 - 21 g and 6 - 31g(d ) basis sets @xcite .
the sto-3 g and 6 - 21 g basis sets were used for geometry optimization and energy calculation , while the 6 - 31g(d ) was used for simulating the nuclear magnetic resonance properties . we have calculated chemical shielding tensors for different molecules of interest and compared the results with experimental measurements . according to nmr theory , the chemical shielding tensor can be computed as the second derivative of the electronic energy @xmath51 with respect to the external magnetic field @xmath52 and the nuclear magnetic moment of interest @xmath53 : @xmath54 indices @xmath55 and @xmath56 correspond to the @xmath55-th and @xmath56-th components of the vectors @xmath52 and @xmath53 respectively .
the gauge - origin problem inherent to finite basis set calculations of nmr shieldings can be solved by using gauge - including atomic orbitals ( giaos ) @xcite . in the giao method
the corresponding calculations are carried out with the explicitly field - dependent basis function defined as follows : @xmath57 here @xmath58 is the field independent basis function defined in eq .
( [ eq10 ] ) , @xmath59 denotes the center of the corresponding @xmath47 , @xmath60 describes the electron coordinate , and @xmath61 is the speed of light .
the chemical shielding tensor @xmath62 in eq .
( [ eq11 ] ) can be diagonalized to yield a tensor with three principal components , @xmath63 , which define the isotropic shielding as follows : @xmath64 isotropic shielding is an important quantity , because in a solution nmr experiment , rapid tumbling of the molecules commonly averages the chemical shielding tensor , and only the isotropic chemical shielding is detected . in solid samples ,
the presence of chemical shielding anisotropy often generates broad powder patterns , but in this case the magic - angle spinning ( mas ) technique can be used to partially or completely average the chemical shielding anisotropy powder patterns .
another important quantity is the chemical shift @xmath65 , which is defined as the difference in shielding between the nucleus in the species under investigation , @xmath66 , and the shielding of the same nucleus in a reference compound , @xmath67 : @xmath68 the latter approximation is often used because usually @xmath69 .
equation ( [ eq14 ] ) gives the chemical shift in parts per million ( ppm ) . in this work ,
we are interested in the chemical shifts for the @xmath2c nuclei , which we have calculated relative to the chemical shift of the carbon atoms in tetramethylsilane ( tms ) .
tmb nanowires . ( a ) gc analysis of the as - prepared sample .
( b ) gc profile of the aged sample .
the gc peaks at 1.34 minutes in ( a ) and 1.36 minutes in ( b ) are attributable to the remnant of ccl4 solvent , and the peaks at 2.93 or 2.95 minutes in the spectra are due to an impurity present in the system.,width=604 ] synthesis of the fullerene nanowire polymer required that we first prepared c@xmath0 nanowires using a 1,2,4-tmb solution of c@xmath0 by a previously described method @xcite .
we noted that the resulting nanowires were highly stable , as indicated by the fact that there was no detectable alteration in either their crystalline morphology , crystal color , or sample weight as a function of time .
however , the solubility of the material was found to change with time . unlike raw c@xmath0 powder ,
well known to be highly soluble in aromatic solvents , the as - made nanowires were only partially soluble in these solvents , and this solubility decreased further with time .
after ageing for a period of @xmath7010 months , the nanowires became totally insoluble in common organic solvents including benzene , toluene , 1,2,4-tmb , and carbon tetrachloride .
this initial observation led us to speculate that a polymerization reaction might have occurred within the system , and that this had led to interesting new physical and chemical properties , including excellent thermal stability and solvent - resistant behavior .
tmb nanowires .
the profiles ( a ) and ( b ) correspond to the gc peaks at 2.80 and 3.10 minutes ( elution times ) as shown in figure [ fig1]a.,width=604 ] observation of the loss of solubility was consistent with the results obtained from the gc - ms analysis .
tests were carried out using two suspensions of the solid in ccl@xmath71 , corresponding respectively to as - made and aged nanowires .
prior to the tests , the solids were resident in the solvent for 5 days .
the results are shown in figure [ fig1 ] .
as can be seen , gc successfully isolated 1,2,4-tmb molecules from the suspension containing the as - made nanowires , as indicated by the signals at 2.80 and 3.10 minutes ( see figure [ fig1]a ) .
the mass spectra of these eluent fractions , shown in figure [ fig2 ] , confirmed the presence of trimethylbenzene in both isomeric forms : the main gc peak ( at 2.80 minutes ) corresponded to 1,2,4-tmb , while the smaller peak ( at 3.10 minutes ) corresponded to either 1,3,5-tmb or 1,2,3-tmb , which existed as impurities in the solvent .
the major ions in the mass spectra were seen at 120 , 105 , 91 and 77 da and were consistent with those expected from trimethylbenzene .
in contrast , no tmb molecules could be extracted from the aged solid , as confirmed by the absence of the corresponding gc peaks from the spectrum ( see figure [ fig1]b ) .
these data support the observation of partial solubility behavior of the as - made solid , suggesting a possible partial polymerization in the sample , which appears to arise from the slow - growth of the nanowires .
the data also suggest a likely full polymerization in the aged solid , wherein all the tmb molecules appear to have chemically bonded to c@xmath0 , resulting in an extensive , non - soluble polymeric network .
c solid state spectrum of pure c@xmath0 .
( b ) solution - state ( in deuterochloroform ) @xmath2c nmr spectrum of pure 1,2,4-trimethylbenzene .
( c ) solid - state cp - mas @xmath2c nmr spectrum of the as - prepared c@xmath0 - 1,2,4-trimethylbenzene adduct .
( d ) solid - state cp - mas @xmath2c nmr spectrum of the aged c@xmath0 - 1,2,4-trimethylbenzene adduct .
( e ) the calculated spectra using the _ ab initio _ b3lyp/6 - 31g(d ) method .
solid line ( blue fill ) , dashed line ( green fill ) and dotted line ( red fill ) show the corresponding spectra for the c@xmath0-tmb - c@xmath0 complex as shown in ( f ) , pure 1,2,4-trimethylbenzene and pristine c@xmath0 respectively .
the numbers labelled in ( e ) show the chemical shifts of the corresponding carbon atoms marked in ( f).,width=619 ] in order to confirm the polymerization and elucidate the nature of the bonding mode involved , we performed @xmath2c nmr spectroscopic studies on both as - made and aged specimens , and on raw c@xmath0 powder ( 99.9 % , sers ltd . ) as well as 1,2,4-tmb solvent ( 98 % , aldrich ) ( see figure [ fig3 ] ) .
some assignments are immediately suggested by a comparison with the spectra of the pure precursor materials , namely , the pristine c@xmath0 ( see figure [ fig3]a ) and pure 1,2,4-tmb ( see figure [ fig3]b ) .
the broad peak at 22.5 ppm in the as - prepared c@xmath0tmb adduct ( see figure [ fig3]c ) results from the three methyl groups of 1,2,4-tmb ( 19.8 , 20.2 and 21.5 ppm ) , the signal at 129.9 ppm in the same spectrum to protonated benzyl carbons of the adducted 1,2,4-tmb ( c.f
. signals at 127.0 , 130.2 and 131.1 ppm in liquid 1,2,4-tmb ) , that at 135.2 ppm to the quaternary benzyl carbons ( c.f .
signals at 133.9 , 135.7 and 136.8 ppm in liquid 1,2,4-tmb ) , and that at 143.3 ppm to the c@xmath0 component of the adduct ( c.f . signal at the same chemical shift in pure c@xmath0 ) .
apart from these , it is crucial to note the new peak exhibiting a chemical shift of 30.4 ppm in the spectrum .
this is consistent with the presence of methylene carbon atoms directly bonded to aromatic rings @xcite , providing direct evidence for polymerization .
a comparison of the @xmath2c nmr spectra of the as - prepared and the aged ( see figure [ fig3]d ) materials clearly indicates that although signals in each sample are consistent with one another , in the latter case the peaks are much broader , consistent with the expected structural diversity and heterogeneity expected in the aged sample .
in addition , a characteristic fingerprint of the spectrum of the aged nanowires is the new emergence of three broad peaks at @xmath7056 ppm , @xmath7064 ppm , and @xmath7073 ppm , respectively .
these three peaks are consistent with what would be expected from the three carbons marked as 2 , 3 and 4 in figures [ fig3]e - f as the result of polymerization ( also see the theoretical analysis part below ) , providing further evidence of polymerization . .@xmath2c nmr spectroscopic data ( ppm ) of the molecules of c@xmath0 , 1,2,4-trimethylbenzene , 2,4,6,2@xmath72,4@xmath72,6@xmath72-diphenylmethane and the ( -c@xmath0tmb-)@xmath10 nanopolymer .
the values in brackets indicate chemical shifts calculated in this work using the b3lyp/6 - 31g(d ) method .
the values in bold numbers are the chemical shifts measured in this work . _
notes : a : reference @xcite ; b : reference @xcite ; c : reference @xcite ; * : quaternary carbons on the benzene ring .
_ [ cols=">,^,^,^,^ " , ] the fact that reaction ( [ eq15 ] ) is endothermic can be understood in terms of the changes in electron configuration for carbon atoms in c@xmath0 following the reaction .
every carbon atom in a fullerene has three covalent bonds with its neighbors , two of which are single , and one of which is double .
the carbon atoms in a fullerene therefore exhibit sp@xmath73 hybridization , but the framework is slightly distorted because of the surface curvature . to attach a tmb molecule to a fullerene , it is necessary to break a double bond in the carbon shell , leading to the formation of two unsaturated carbon atoms . in reaction ( [ eq15 ] )
the tmb@xmath74 complex caps one of these carbon atoms in each fullerene , rendering the carbon sp@xmath75 hybridized but leaving the neighboring carbon unsaturated an energetically unfavorable scenario .
in contrast , reaction ( [ eq16 ] ) is exothermic because the hydrogen atoms released from the tmb molecule cap the neighboring unsaturated carbon atoms in each of the c@xmath0 fragments .
thus , both carbon atoms that have undergone reaction in either fullerene have become sp@xmath76 hybridized ( as illustrated by the inset of figure [ fig7]b ) .
this result is consistent with our elemental analysis for a polymerized sample where parallel analyses yield an average hydrogen content of 1.35 wt% , consistent with the molar ratio of @xmath701:1 for 1,2,4-tmb over c@xmath0 in the solid , as measured previously by thermal gravimetric analysis ( tga ) @xcite .
reaction ( [ eq16 ] ) has some similarities to the prato and bingel reactions widely known in fullerene chemistry @xcite . in these reactions
an organic molecule reacts with a double bond in c@xmath0 and forms a ring peripheral to the fullerene superstructure .
the calculated reaction enthalpy suggests that reaction ( [ eq16 ] ) is energetically favorable , and therefore is likely to occur .
tmb complex as the product of reaction ( [ eq17 ] ) in the text .
the distances between fullerene centers correspond to the geometries optimized at the b3lyp / sto-3 g level of calculation.,width=604 ] a second and more complex polymerization pathway that can be envisaged involves three fullerenes and one tmb .
by analogy to equation ( [ eq16 ] ) , the reaction can be written as : @xmath77 the possible geometrical arrangements of two low - energy isomers of 3c@xmath0tmb ( b3lyp / sto-3 g method ) are shown in figure [ fig8 ] .
these structures are constructed in a similar way to the 2c@xmath0tmb complex , but in this case three c - h bonds have been broken in the tmb unit , and the resulting 3c@xmath0tmb complex adopts a triangular configuration .
the total energies calculated for two optimized isomeric structures of the 3c@xmath0tmb are shown in table [ tab2 ] , corresponding to the two complexes shown in figure [ fig8 ] .
the calculated enthalpies are @xmath78=-53.64 kcal mol@xmath79 and @xmath78=-55.11 kcal mol@xmath79 , for the product isomers ( a ) and
( b ) , respectively . in either case , it is clear that reaction ( [ eq17 ] ) is exothermic , and thus energetically favorable . in order to understand how the 1,2,4-tmb molecules bond to fullerenes in the nanowires ,
we have studied geometrical constraints of the unit cell of the crystal lattice .
figure [ fig9]a shows the geometry of such a unit cell and its dimensions , based on the experimental data @xcite . for simplicity ,
here we do not show the embedded tmb molecules . as discussed before ,
a tmb molecule may link two fullerenes together , and if this mode is repeated periodically , the crystalline nanowire would turn into a nanopolymer .
however , this is only possible if a tmb fits into the space between two fullerenes without significantly disturbing the unit cell structure .
tmb 1d nanocrystals . in ( a )
we show the geometry of the unit cell ( the tmb molecules embedded in the unit cell are not shown ) .
the distances between the centers of the fullerenes are indicated .
the geometry of a c@xmath0-tmb - c@xmath0 molecule is shown in ( b ) .
the distance between the two fullerenes in this case is 14.4 , very close to the dimensions of the unit cell .
plots ( c ) and ( d ) demonstrate the cross section of a nanowire , viewed along the c - axis , and illustrate two possible linking schemes between the fullerenes .
the bars between the fullerenes indicate the tmb molecules .
the polymerization can occur along either the a-
( plot c ) or the b - axis ( plot d ) of a nanowire @xcite.,width=604 ] in figure [ fig9]b we show a c@xmath0-tmb - c@xmath0 complex , where two fullerenes are covalently linked by a 1,2,4-tmb molecule . in this case
the distance between the two fullerenes is 14.4 , very close to one of the inter - fullerene spacings measured for the unit cell ( 14.5 , figure [ fig9]a ) .
the linkage is thus highly likely to occur in the unit cell and the c@xmath0-tmb - c@xmath0 complex is likely to be a possible building block for the polymerization .
figures [ fig9]c and [ fig9]d illustrate two possible linking schemes between the fullerenes , where the bars between fullerenes represent tmb molecules . based on the preferential crystal growth directions studied previously @xcite , this polymerization may occur along either the a- ( see figure [ fig9]c ) or the b - axis ( see figure [ fig9]d ) , with the linkages corresponding to the two schemes as shown in the figures . it is important to stress that a c@xmath0-tmb - c@xmath0 complex has several stable isomeric configurations with different c@xmath0-c@xmath0 distances .
to illustrate this fact , in figure [ fig10 ] we show how the the distance between two fullerenes in a c@xmath0-tmb - c@xmath0 complex depends on an angle @xmath80 , which is defined as the separation degree between the planes formed by atoms ( 1)-(2)-(3 ) and atoms ( 2)-(3)-(4 ) , as shown in the figure .
as the bonds connecting two c@xmath0 units and a tmb@xmath74 are not collinear , the center - to - center distance changes upon rotation of the fullerenes . as can be seen from figure [ fig10 ] , such a distance varies from 11.5 to 14.5 , and match the distances between fullerenes in the unit cell ( as indicated in figure [ fig9]a ) .
-tmb - c@xmath0 complex as a function of angle @xmath80 , which is defined as the angle between the planes formed by the atoms ( 1)-(2)-(3 ) and ( 2)-(3)-(4 ) , as shown in the inset.,width=604 ] based on this geometrical analysis , we have noted that there appear to be three possibilities for fusing a tmb@xmath74 molecule in between two fullerenes in the unit cell , as illustrated by the three sets of c@xmath0 dimers marked in figure [ fig11]a . except these possibilities , the distances between any other two fullerenes are either too small or too large and are thus incompatible with the requirement for a topochemical polymerization to occur . note that although the distance 16.4 in the unit cell is somewhat larger than the distances shown in figure [ fig10 ] , the related fusion mode is still possible because the distance between fullerenes in a c@xmath0-tmb - c@xmath0 complex may further change at certain conditions .
some deviations are expected because the c@xmath0-tmb - c@xmath0 complex was optimized in vacuo without accounting for other related constituents in the polymer .
also , the complex has many isomers because in the crystal unit cell a fullerene has many different atomic sites to which a tmb@xmath74 molecule may attach .
tmb complex ( a ) , and the 3c@xmath0tmb complex ( b ) , in the unit cell of the parent nanowire .
distances are given in ngstroms and correspond to experimental data @xcite.,width=604 ] similarly , figure [ fig11]b shows two possible fitting sites for a 3c@xmath0tmb complex in the unit cell .
this structural analysis clearly demonstrates that a 1,2,4-tmb molecule can easily stitch two fullerenes together within a nanowire without disturbing the lattice structure .
this conclusion is in excellent agreement with the experimental studies with sem , hrtem and saed , as described earlier .
our experimentally measured chemical shifts are supported by the calculated values as shown by figure [ fig3]e . to do the calculation , we have applied the _ ab initio _ b3lyp/6 - 31g(d ) method to pristine c@xmath0 ( dotted line , red fill ) , pure 1,2,4-tmb ( dashed line , green fill ) and the c@xmath0-tmb - c@xmath0 complex ( solid line , blue fill ) . since many of the calculated chemical shifts are degenerate , lorentzian broadening for each line
is assumed : @xmath81 here @xmath82 is the calculated chemical shift , and @xmath83 is the width of peaks at half maximum intensity .
the calculated @xmath2c - nmr spectra are plotted on a logarithmic scale for better visualization .
the calculated chemical shifts for pure c@xmath0 and 1,2,4-tmb are in good agreement with experiment . the observed signal for c@xmath0 at 143.3
ppm is only 4 ppm higher than the calculated , and the calculated shift for the ring carbons of tmb is only 7 - 8 ppm lower than the observed .
also , the calculated values for the methyl groups perfectly reproduce the measured data ( within an error of less than 1.0 ppm ) .
these comparisons enable us to conclude that the accuracy of the calculation is on the order of @xmath705 - 8 ppm , which is sufficient for an unambiguous characterization of the calculated nmr peaks in this work . in the previous section we suggested that the c@xmath0-tmb - c@xmath0 complex was a building block in the polymerization .
the calculated @xmath2c - nmr spectrum for the complex strongly supports this view by virtue of its excellent agreement with the experimental result .
crucially , the emergence of four peaks in the range of 45 - 80 ppm , consistent with the observed set of broad peaks at 55 - 85 ppm , corresponds to the chemical shifts of the carbon atoms ( 2 ) , ( 3 ) and ( 4 ) , respectively , as marked in figure [ fig3]f .
these three atoms participate in the bonds formed between c@xmath0 and 1,2,4-tmb molecules in the polymerized network .
c - nmr spectra calculated for three isomeric states of the tmb - tmb dimer using the _ ab initio _ b3lyp/6 - 31g(d ) method .
the geometries of the molecules used in the calculation are shown in the inset .
labels indicate the chemical shifts of different types of carbon atoms in the system .
note that the chemical shifts in ( c ) are much more degenerate than ( a ) and ( b ) due to the enhanced symmetry of the studied isomer.,width=604 ] the observed peak at 30.4 ppm can be interpreted as the response from methylene carbons cross - linking adjacent 1,2,4-tmb molecules due to polymerization .
this view is strongly supported by literature report of the structurally closely - related framework of multi - substituted diphenylmethane ( see table [ tab1 ] ) @xcite . within the accuracy of the calculation ,
it is also consistent with our calculations for tmb - tmb dimers , which predict the corresponding chemical shift to be in the range of 28 - 43 ppm , depending on the bonding mode adopted by the dimers and the relative orientations of the adjacent tmb molecules involved ( see figure [ fig12 ] ) .
figure [ fig12]a shows an example of a dimer in which the two benzene rings are perpendicular to each other , and the two tmb molecules are bonded together through two methylene carbons corresponding to the ring positions of 1- and 4-methyl , respectively . in this case , the two methylenes yield two closely placed peaks at 35.7@xmath847 ppm and 37.8@xmath847 ppm .
figure [ fig12]b shows another example of a dimer in which the two tmb molecules are linked by only one methylene carbon which exhibits a chemical shift of 38.4 ppm .
figure [ fig12]c shows the third example of such a dimer isomer , where the two tmb molecules are linked by two identical methylene carbons but the two benzene rings are parallel to each other . in this case the chemical shift calculated is 43.4 ppm .
these analyses suggest that the chemical shift of the methylene carbons is sensitive to the surrounding atoms , and that the peak observed at 30.4 ppm hint at some specific bonding mode and/or relative orientations between tmb molecules in the polymeric chains
. moreover , the enthalpy calculated for such a dimerization reaction using the b3lyp/6 - 21 g method is -19.2 kcal / mol , indicating that the dimerization reaction for polymerization is energetically favorable .
we have demonstrated for the first time an approach to the synthesis of a c@xmath0-based nanowire polymer and established the chemical bonding mode involved in the polymeric chains based on both experimental measurements and theoretical calculations .
importantly , the material adopts a crystalline 1d nanostructure which resembles carbon nanotubes in shape and other important conjugated polymers in structure .
since the material does not contain any metal but is simply composed almost entirely of carbon ( while it contains hydrogen , the content is only 1.4 wt % ) , it suggests biological compatibility and it is , perhaps , even more attractive than carbon nanotubes for bio - applications . in addition , the material has further important potential for applications in photo - electrical devices because of the intrinsically large magnitude of the nonlinear optical response of c@xmath0 and the excellence of its photoinduced charge transfer properties .
considering all these , we believe that this work represents a step toward true applications of c@xmath0 in nanotechnology by the ability of processing commercially available raw c@xmath0 powder into a one - dimensional , crystalline , and covalently - bonded fullerene nanopolymer .
we consider that applications of the reported nanopolymer may be facilitated by a wet - chemical approach through surface modification of the material using the rich chemistry of fullerene developed over the last 20 years .
since the nanopolymer is insoluble in common solvents , such surface modification or functionalization should be possible to achieve in either an aqueous or an organic solution without destructing its solid - state structure .
such a wet approach would benefit from low - cost processing , the need for only simple apparatus and the possibility of scaling - up to the industrial level .
moreover , the nanopolymer itself not only provides an example of phase transition of the parent nanowire driven by forming and breaking covalent bonds , but also illustrates the enduring significance of the original fullerene concept and its versatility as applied to new fullerene - related nanostructures .
finally , the host ( c@xmath0 ) and guest ( 1,2,4-tmb ) nature of the polymerization suggests a general host - guest route to the synthesis of new types of fullerene - based nanopolymers constructed by different organic monomers and fullerenes .
we thank support from the european commission project ( noe excell nmp3-ct-2005 - 515703 ) .
the assistance of dr .
wuzong zhou , school of chemistry , st .
andrews university , for carrying out tem analysis is acknowledged .
computer simulations were carried out at the frankfurt center for scientific computing .
baughman , r. h. ; zakhidov , a. a. ; de heer , w. a. _ science _ , * 2002 * , _ 297 _ , 787 .
miyazawa , k. ; minato , j. ; yoshii , t. ; fujino , m ; suga , t. _ j. mater .
_ , * 2005 * , _ 20 _ , 688 .
wang , l. ; liu , b. ; liu , d. ; yao , m. ; hou , y. ; yu , s. ; cui , t. ; li , d. ; zou , g. ; iwasiewicz , a. ; sundqvist , b. _ adv .
_ , * 2006 * , _ 18 _ , 1883 .
miyazawa , k. ; kuwasaki , y. ; hamamoto , k. ; nagata , s. ; obayashi , a. ; kuwabara , m. _ surf .
interface anal .
_ , * 2003 * , _ 35 _ , 117 .
liu , h. ; li , y. ; jiang , l. ; luo , h. ; xiao , s. ; fang , h. ; li , h. ; zhu , d. ; yu , d. ; xu , j. ; xiang , b. _ j. am .
_ , * 2002 * , _ 124 _ , 13370 .
dresselhaus , m. s. ; dresselhaus , g. ; eklund , p. c. _ science of fullerenes and carbon nanotubes _ * 1996 * , academic press : san diego .
lieber , c. m. ; wang , z. l. _ mrs bull .
_ , * 2007 * , _ 32 _ , 99 .
rodney , s. r. ; arthur , l. r. _ nature _ , * 1991 * , _ 350 _ , 663 .
geng , j. ; kinloch , i. a. ; sing , c. ; golovko , v. b. ; johnson , b. f. g. ; shaffer , m. s. p. ; li , y. ; windle , a. h. _ j. phys . chem .
b _ , * 2005 * , _ 109 _ , 16665 ; and references therein .
stoilova , o. ; jerome , c. ; detrembleur , c. ; mouithys - mickalad , a. ; manolova , n. ; rashkov , i. ; jerome , r. _ polymer _ , * 2007 * , _
48(7 ) _ , 1835 . ( a )
ghosh , s. ; sood , a. k. ; kumar , n. _ science _ , * 2003 * , _ 299 _ , 1042 .
( b ) zheng , m. ; jagota , a. ; strano , m. s. ; santos , a. p. ; barone , p. ; chou , s. g. ; diner , b. a. ; dresselhaus , m. s. ; mclean , r. s. ; onoa , g. b. ; samsonidze , g. g. ; semke , e. d. ; usrey , m. ; walls , d. j. _ science _ , * 2003 * , _ 302 _ , 1545 .
( c ) heller , d. a. ; jeng , e. s. ; yeung , t - k . ; martinez , b. m. ; moll , a. e. ; gastala , j. b. ; strano , m. s. _ science _
, * 2006 * , _ 311 _ , 508 .
guss , w. ; feldmann , j. ; gobel , e. o. ; taliani , c. ; mohn , h. ; muller , w. ; haubler , p ; ter meer , h. u. _ phys .
lett . _ * 1994 * , _ 72 _ , 2644 .
saito , r. ; dresselhaus , g. ; dresselhaus , m. s. _ phys .
b _ , * 1992 * , _ 46 _ , 9906 .
xu , m. ; pathak , y. ; fujita , d. ; ringor , c. ; miyazawa , k. _ nanotechnology _ , * 2008 * , _ 19 _ , 075712 .
( a ) wang , y. _ nature _ , * 1992 * , _ 356 _ , 585 .
( b ) sariciftci , n. s. ; smilowitz , l. ; heeger , a. j. ; wudl , f. _ science _ , * 1992 * , _ 258 _ , 1474 . ( a )
morana , m. ; wegscheider , m. ; bonanni , a. ; kopidakis , n. ; shaheen , s. ; scharber , m. ; zhu , z. ; waller , d. ; gaudiana , r. ; brabec , c. ; _ adv .
_ , * 2008 * , _ 18 _ , 1757 .
( b ) service , r. f. _ science _ , * 2001 * , _ 293 _ , 1570 .
rao , a. m. ; zhou , p. ; wang , k. a. ; hager , g. t. ; holden , j. m. ; wang , y. ; lee , w. t. ; bi , x. x. ; eclund , p. c. ; cornett , d. s. ; duncan , m. a. ; amster , l. j. _ science _ , * 1993 * , _ 259 _ , 955 .
eklund , p. c. ; rao , a. m. ; zhou , p. ; wang , y. ; holden , j. m. _ thin solid films _ , * 1995 * , _ 257 _ , 185 .
pekker , s. ; janossy , a. ; mihaly , l. ; chauvet , o. ; carrard , m. ; forro , l. _ science _ , * 1994 * , _ 265 _ , 1077 .
nagashima , h. ; nakaoka , m.a . ; saito , y. ; kato , m. ; kawanishi , t. ; itoh , k. _ j. chem .
soc . chem .
_ , * 1992 * , 377 .
geckeler , k. e. ; hirsch , a. _ j. am .
_ , * 1993 * , _ 115 _ , 3850 .
wudl , f. _ acc .
res . _ * 1992 * , _ 25 _ , 157 .
ceolin , r. ; tamarit , j. l. ; lopez , d. o. ; barrio , m. ; agafonov , v. ; allouchi , h. ; moussa , f. ; szwarc , h. _ chem .
_ , * 1999 * , _ 314 _ , 21 .
balch , a. l. ; lee , j. w. ; noll , b. c. ; olmstead , m. m. ; _ j. chem .
_ , * 1993 * , 56 .
burgi , h. b. ; restori , r. ; schwartzenbach , d. ; balch , a. l. ; lee , j. w. ; noll , b. c. ; olmstead , m. m. _ chem .
_ , * 1994 * , _ 6 _ , 1325 .
miyazawa , k. ; nishimura , c. ; fujino , m. ; suga , t. ; yoshii , t. _ transact .
japan _ , * 2004 * , _ 29 _ , 1965 .
miyazawa , k. ; suga , t. _ j. mater .
_ , * 2004 * , _ 19 _ , 2410 .
miyazawa , k. ; hamamoto , k. _ j. mater .
_ , * 2002 * , _ 17 _ , 2205 .
geng , j. ; zhou , w. ; skelton , p. ; yue , w. ; kinloch , i. ; windle , a. ; johnson , b. f. g. _ j. am .
_ , * 2008 * , _ 130 _ , 2527 .
geng , j. ; solovyov , i. a. ; zhou , w. ; solovyov , a. v. ; johnson , b. f. g. _ j. phys .
c _ , * 2009 * , _ 113 _ , 6390 .
frisch , m. j. ; trucks g. w. ; schlegel , h. b. ; scuseria , g. e. ; robb , m. a. ; cheeseman , j. r. ; vreven , j. a. m. jr . ; kudin , k. n. ; burant , j. c. ; millam , j. m. ; iyengar , s. s. ; tomasi , j. ; barone , v. ; mennucci , b. ; cossi , m. ; scalmani , g. ; rega , n. ; petersson , g. a. ; nakatsuji , h. ; hada , m. ; ehara , m. ; toyota , k. ; fukuda , r. ; hasegawa , j. ; ishida , m. ; nakajima , t. ; honda , y. ; kitao , o. ; nakai , n. ; klene , m. ; li , x. ; knox , j. e. ; hratchian , h. p. ; cross , j.b . ; adamo , c. ; jaramillo , j. ; gomperts , r. ; stratmann , r. e. ; yazyev , o. ; austin , a. j. ; cammi , r. ; pomelli , c. ; ochterski , j. w. ; ayala , p. y. ; morokuma , k. ; voth , g. a. ; salvador , p. ; dannenberg , j. j. ; zakrzewski , v. g. ; dapprich , s. ; daniels , a. d. ; strain , m. c. ; farkas , o. ; malick , d. k. ; rabuck , a. d. ; raghavachari , k. ; foresman , j. b. ; ortiz , j. v. ; cui , q. ; baboul , a. g. ; clifford , s. ; cioslowski , j. ; stefanov , b. b. ; liu , g. ; liashenko , a. ; piskorz , p. ; komaromi , i. ; martin , r. l. ; fox , d. j. ; keith , t. ; al - laham , m. a. ; peng , c. y. ; nanayakkara , a. ; challacombe , m. ; gill , p. m. w. ; johnson , b. ; chen , w. ; wong , m. w. ; gonzalez , c. ; pople , j. a. _ gaussian 03 , revision c.02 _ , gaussian , inc . , wallingford ct , 2004 .
haberland , h. ( ed . ) , _ clusters of atoms and molecules , theory , experiment and clusters of atoms _ , springer series in chemical physics , vol . 52 :
springer , berlin , * 1994*. de heer , w. a. _ rev .
_ , * 1993 * , _ 65 _ , 611 .
brack , m. _ rev .
_ , * 1993 * , _ 65 _ , 677 . ( a )
becke , a. _ phys .
a _ , * 1988 * , _ 38 _ , 3098 ; ( b ) becke , a. d. _ j. chem .
_ , * 1993 * , _ 98 _ , 1372 . foresman , j. b. ; frisch , a. _ exploring chemistry with electronic structure methods . _ pittsburgh , pa : gaussian inc .
* 1996*. vosko , s. h. ; wilk , l. ; nusair , m. _ can . j. phys .
_ , * 1980 * , _ 58 _ , 1200 .
lee , c. ; yang , w. ; parr , r. g. _ phys .
b _ , * 1988 * , _ 37 _ , 785 .
gauss , j. ; werner , h.j .
_ , * 2000 * , _ 2 _ , 2083 .
london , f. _ j. phys .
radium _ , * 1937 * , _ 8 _ , 397 .
hameka , h. f. _ mol .
_ , * 1958 * , _ 1 _ , 203 .
ditchfield , r. _ mol .
_ , * 1974 * , _ 27 _ , 789 .
buchanan , g. w. ; montaudo , g. ; finocchiaro , p. _ can .
_ , * 1974 * , _ 52 _ , 3196 .
woolfenden , w. r. ; grant , d. m. _ j. am .
_ , * 1966 * , _ 88 _ , 1496 .
matsuda , k. ; maniwa , y. ; kataura , h. _ physical review b _ , * 2008 * , _ 77 _ , 075421 1 - 6 .
hirsch a. ; brettreich m. , _ fullerenes : chemistry and reactions _ , wiley - vch , * 2005*. solovyov , i. a. ; geng , j. ; solovyov , a. v. ; johnson , b. f. g. _ chem .
_ , * 2009 * , _ 472 _ , 166 . | large - scale practical applications of fullerene ( c@xmath0 ) in nanodevices could be significantly facilitated if the commercially - available micrometer - scale raw c@xmath0 powder were further processed into a one - dimensional ( 1d ) nanowire - related polymer displaying covalent bonding as molecular interlinks and resembling traditional important conjugated polymers .
however , there has been little study thus far in this area despite the abundant literature on fullerene .
here we report the synthesis and characterization of such a c@xmath0-based nanowire polymer , ( -c@xmath0tmb-)@xmath1 , where tmb=1,2,4-trimethylbenzene , which displays a well - defined crystalline structure , exceptionally large length - to - width ratio and excellent thermal stability .
the material is prepared by first growing the corresponding nanowire through a solution phase of c@xmath0 followed by a topochemical polymerization reaction in the solid state .
gas chromatography , mass spectrometry and @xmath2c nuclear magnetic resonance evidence is provided for the nature of the covalent bonding mode adopted by the polymeric chains . theoretical analysis based on detailed calculations of the reaction energetics and structural analysis provides an in - depth understanding of the polymerization pathway .
the nanopolymer promises important applications in biological fields and in the development of optical , electrical , and magnetic nanodevices . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.