article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the theory of water waves now is among the most actively developing branches of the hydrodynamic science ( see , for instance , @xcite , and many references therein ) . in particular ,
two - dimensional ( 2d ) flows of an ideal incompressible fluid with a free surface are very attractive for theoretical study , since they possess many essential features in common with real water waves , and simultaneously are relatively simple for analytical and numerical treatment .
a powerful mathematical support for the 2d theory is provided by analytical functions and the corresponding conformal mappings .
the original idea of using conformal mappings was intended for waves on the deep water or over a static horizontal bottom : a geometrically simple region occupied by resting fluid ( it is the lower half - plane @xcite , or a horizontal stripe @xcite ) is mapped onto the flow region with a disturbed time - dependent free surface , so that the real axis is transformed into the moving boundary . with such conformal `` straightening '' of the free boundary , it is possible to obtain an exact analytical description for surface waves .
the corresponding ( 1 + 1)-dimensional nonlinear equations of motion are very convenient for numerical simulations by spectral methods .
important achievements in this direction were reported in works @xcite , including analytical and numerical results .
a generalization has been recently made to the case when the bottom shape is still static but strongly inhomogeneous in space @xcite .
the purpose of present paper is to derive exact equations of motion for the free boundary in 2d ideal flows , valid with arbitrary non - stationary bottom profile ( sec.ii ) , and to suggest a numerical method for solving the obtained highly nonlinear equations ( sec.iii ) .
this problem , besides being interesting by itself , is also important for studying such phenomena as tsunami generation by earthquakes and other similar processes in natural and laboratory conditions .
let us start our consideration with basic definitions . we will describe a nonstationary flow region @xmath0 in @xmath1-plane ( with @xmath2-axis up - directed ) by an analytical function @xmath3 , where @xmath4 is the time , and the complex variable @xmath5 occupies in @xmath6-plane the fixed infinite horizontal stripe @xmath7
. position of the free surface @xmath8 will be given in parametric form by the formula @xmath9 while the bottom profile @xmath10will be determined by @xmath11 inasmuch as the velocity potential @xmath12 satisfies the laplace equation @xmath13 , in @xmath0 the analytical complex potential @xmath14 is defined , with @xmath15 being the harmonically conjugate for @xmath16 .
we deal below with the complex analytical function @xmath17 defined in @xmath18 and taking the boundary values as written below : @xmath19 the velocity components are determined by the usual relations @xmath20 where @xmath21 denotes derivative over the complex variable @xmath22 .
now we proceed to equations of motion .
first of all , it is the dynamic bernoulli equation on the free surface , @xmath23 where @xmath24 is the gravitational acceleration .
this equation can be easily re - written in terms of the conformal variables as follows : @xmath25 two other equations of motion are the kinematic conditions on the free surface and on the bottom , expressing the fact that the boundary motion is determined by the normal component of the velocity field . in the conformal variables
these conditions take form @xmath26 @xmath27 where @xmath28 denotes complex conjugate value .
we should take into account that @xmath29 and @xmath30 are not mutually independent , but they are related to each other by a linear transform .
indeed , the following formula for analytical continuation is valid : @xmath31 here the fourier transform @xmath32 is defined in the usual way : @xmath33 therefore at the free surface , where @xmath34 , we have @xmath35 it is a proper point here to introduce some linear operators , necessary for further exposition .
these linear operators @xmath36 , @xmath37 , and @xmath38 are diagonal in fourier representation : @xmath39 it will be convenient for us to write @xmath30 in the form @xmath40 with @xmath41 and @xmath42 being some unknown real functions .
then for @xmath29 we will have the formula @xmath43 later we use in equations the complex function @xmath44 now it is necessary to look at the conformal mapping @xmath45 with more attention .
since the bottom motion is assumed to be prescribed , this function has the following structure ( compare with @xcite ) : @xmath46 where a known analytical function @xmath47 determines a conformal mapping for the upper half - plane of an intermediate complex variable @xmath48 onto the infinite region above the bottom in @xmath49-plane ( or , at least , the mapping @xmath47 should have no singularities within a sufficiently wide horizontal stripe above the real axis in @xmath48-plane , so there should be enough `` free space '' for large - amplitude waves in @xmath49-plane ) .
the intermediate analytical function @xmath50 takes real values at the real axis .
therefore we may write @xmath51 where @xmath52 is the fourier transform of a real function @xmath53 . on the bottom @xmath54 ,
thus the parametrically given curve @xmath55 coincides with the prescribed curve @xmath56 ( these two curves differ from each other just by parameterization ) . at the free surface @xmath57
@xmath58 since our purpose is to derive equations of motion for the unknown functions @xmath59 , @xmath42 , and @xmath60 , we have to substitute expressions ( [ phi_b ] ) , ( [ phi_s ] ) , ( [ zb ] ) , ( [ xi_def ] ) , and ( [ zs ] ) into eqs.([bernoulli_dw ] ) , ( [ kinematic_dw_s ] ) , ( [ kinematic_dw_b ] ) .
eq.([kinematic_dw_b ] ) is simple and does not need any trick .
eq.([kinematic_dw_s ] ) ( divided by @xmath61 ) takes the form @xmath62 from here we express @xmath63 ( see eq.([xi_t ] ) below and compare with @xcite ) and substitute proper expressions into eq.([bernoulli_dw ] ) , in order to find @xmath64 .
simple transformations , similar to those in work @xcite , lead to the following equations : @xmath65\!\ ! , \label{xi_t}\\ \psi_t\!\!&=&\!\!-\psi_u\,(\hat t\!+\!i)\!\!\left [ \frac{\mbox{im\,}\psi_u -\hat s f } { |z_\xi(\xi , t)\xi_u|^2}+\mbox{im}\left(\ !
\frac{z_t(\xi , t)}{z_\xi(\xi , t)\xi_u}\!\right)\right]\nonumber\\ & & + ( 1\!+\!i\hat r)\bigg [ \mbox{re}\left(\frac{\psi_u z_t(\xi , t)}{z_\xi(\xi , t)\xi_u}\right ) -\frac{|\psi_u|^2 - ( \hat s f)^2}{2|z_\xi(\xi , t)\xi_u|^2}\nonumber\\ & & \qquad\qquad\qquad\qquad\qquad -g\ , \mbox{im\,}z(\xi , t)\bigg ] . \label{psi_t}\end{aligned}\ ] ] this system of equations is the main result of present paper .
it provides exact description for potential water waves over a time - dependent bottom determined by an analytical function @xmath47 .
it should be noted that capillarity effects can be easily included here by adding the term @xmath66\ ] ] ( proportional to the surface curvature ) to the gravity term @xmath67 .
the above equations are valid not only for infinitely extended configurations ( when the fluid covers the whole bottom ) , but also for the flows with one or two non - stationary shore edges .
however , asymptotic properties of the intermediate mapping @xmath50 are different in these cases .
so , with two edges @xmath68 as @xmath69 , while for edge - less configurations the corresponding limits are @xmath70 .
equations ( [ f ] ) , ( [ xi_t ] ) , ( [ psi_t ] ) , though look quite complicated , in many cases are convenient for numerical simulations , as described below . instead of the function @xmath59 , we deal with another function ( similar to that proposed in @xcite ) , @xmath71 because this choice results in better numerical stability ( concerning the problem of optimal variables see @xcite ) .
let us consider a class of space - periodic flows ( without shore edges ! ) over @xmath72-periodic profiles of the bottom . with a proper choice for the length and time scales , we may write @xmath73 , and @xmath74 . as it was pointed in @xcite , in space - periodic
flows the variable @xmath75 comes into solutions in the combination @xmath76 , where a real function @xmath77 depends on time because the operator @xmath78 is singular on small wave - numbers . besides @xmath77 , solutions are determined by two other real functions , @xmath79 and @xmath80 , both having period @xmath81 on @xmath82 variable : @xmath83 @xmath84 @xmath85 the linear operator @xmath86 is diagonal in the discrete fourier representation : @xmath87 .
let us introduce the following quantities : @xmath88
\big|_{s=\vartheta+\hat { \mathsf s}_\alpha \rho } , \label{f}\\ u\!\!\!&=&\!\!(\hat{\mathsf t}_\alpha\!+\!i)\!\left [ \frac{\mbox{im\,}(\beta\xi ' ) \!-\!\hat { \mathsf s}_\alpha f } { |z_\xi(\xi , t)\xi'|^2}+\mbox{im}\!\left(\ !
\frac{z_t(\xi , t)}{z_\xi(\xi , t)\xi'}\!\right)\right]\!\ ! ,
\\ w\!\!\!&=&\!\!\frac{|\beta\xi'|^2\!-\!(\hat { \mathsf s}_\alpha f)^2 } { 2|z_\xi(\xi , t)\xi'|^2}+\mbox{im}z(\xi , t ) -\!\mbox{re}\!\left(\!\!\frac{\beta z_t(\xi , t ) } { z_\xi(\xi , t)}\!\!\right)\!\!,\end{aligned}\ ] ] where @xmath89 , @xmath90 . the linear operator @xmath91 is regular .
in the discrete fourier representation it is defined as follows : @xmath92 equations of motion for the real functions @xmath77 , @xmath79 , and @xmath80 follow from ( [ f ] ) , ( [ xi_t ] ) , ( [ psi_t ] ) and have the form @xmath93 where @xmath94 .
the above system ( [ dot_alpha])-([dot_chi ] ) can be efficiently simulated by the spectral method , since the multiplications are simple in @xmath82-representation , while the linear operators @xmath95 , @xmath86 , and @xmath91 ( also @xmath82-differentiation ) are simple in fourier representation .
the integration scheme can be based , for instance , on the runge - kutta 4-th order accuracy method , similarly to works @xcite .
efficient subroutine libraries for the fast fourier transform are now available . in the numerical experiments reported here , the fftw library was used @xcite .
the employed fourier harmonics had numbers @xmath96 in the limits @xmath97 ( the size of arrays for the fourier transform was even larger , @xmath98 ) , thus the obtained solutions are very accurate ( compare with @xcite ) . in the first numerical experiment [ referred to as ( i ) ] ,
the shape of the bottom was determined by analytical function @xmath99 , where @xmath100 , and function @xmath101 is expressed as follows : @xmath102.\end{aligned}\ ] ] here the parameters are @xmath103 , @xmath104 , @xmath105 , @xmath106 , @xmath107 , and @xmath108 .
the time - independent terms in the above formula describe the bottom profile as shown in fig.[i - t0 ] , with deep and shallow regions .
the time - dependent term in @xmath109 corresponds to formation of a cavity on the bottom .
the initial values were taken @xmath110 , @xmath111 , and @xmath112 , thus giving the horizontal free surface and zero velocity field at @xmath113 .
the corresponding numerical results are presented in figs.[i - t0]-[i - t3 ] in terms of the dimensionless quantities @xmath114 and @xmath115/[z_\xi(\xi , t)\xi']=v_x - iv_y$ ] .
this numerical experiment can be viewed as a rough model for a tsunami wave generation , subsequent propagation , and final breaking .
indeed , the inhomogeneous bottom displacement is an analog of an earthquake resulting in two opposite - propagating groups of surface waves .
when one of the groups first approaches a shallow region , the crest of its leading wave becomes higher and sharper , and finally singular . in real world instead true singularity one can observe vortices , splash , and foam on the crest .
further treatment of this situation is not possible within present ideal 2d theory , since 3d effects and dissipative processes become important . in the second numerical experiment [ referred to as ( ii ) ] , the bottom profile was determined by function @xmath116 , where @xmath117 with @xmath118 , @xmath104 , @xmath119 , @xmath120 .
the initial values were @xmath110 , @xmath111 , and @xmath112 , thus corresponding to the horizontal free surface @xmath121 , the horizontal bottom @xmath122 , and zero velocity field at @xmath113 . at later times , some pieces of the bottom rose and some pieces went down .
results of this simulation are presented in figs.[ii - t1]-[ii - t3 ] . in this case , a singularity forms in the places where the fluid flows from the shallow region to the deep region ( see fig.[ii - t3 ] ) . in the third numerical experiment [ referred to as ( iii ) ] , the following function @xmath47 was used : @xmath123,\ ] ] with @xmath124 , @xmath120 , @xmath125 .
as previously , @xmath110 , @xmath111 , and @xmath112 .
the results of this simulation are presented in figs.[iii - t1]-[iii - t3 ] .
in this paper exact nonlinear equations of motion for potential 2d ideal incompressible flows with a free surface in a uniform gravitational field have been obtained , taking into account a time - dependent curved bottom with a profile prescribed by an analytical function .
the mathematical theory of functions of a complex variable and conformal mappings was extensively used in the derivation . despite a complicated structure of the equations ,
an efficient numerical code has been developed for space - periodic solutions , which gives very accurate results .
kirby , in _ gravity waves in water of finite depth _ , j. n. hunt ( ed ) .
advances in fluid mechanics , * 10 * , 55 - 125 , computational mechanics publications , 1997 ; + j.t .
kirby , in _ advances in coastal modeling _ , v. c. lakhan ( ed ) , elsevier , 1 - 41 , 2003 ; + http://chinacat.coastal.udel.edu/~kirby/kirby_pubs.html | two - dimensional potential flows of an ideal fluid with a free surface are considered in situations when shape of the bottom depends on time due to external reasons .
exact nonlinear equations describing surface waves in terms of the so called conformal variables are derived for an arbitrary time - evolving bottom parameterized by an analytical function .
an efficient numerical method for the obtained equations is suggested . |
the transient , free decay of coupled , damped oscillators is not discussed in elementary physics courses and rarely , if ever , in advanced ones .
the discussion in advanced physics textbooks is cursory , typically suggesting that one would proceed just as one might imagine " but that the details are cumbersome .
the new features possessed by the generic or typical version of such systems relative to the well - studied ones are just not part of the basic physics education offered to all students of physical sciences and engineering . proportional " damping ( where each normal mode has its own individual damping constant ) is often presented and is easy to analyze , but it is a very special subset of possible damping in coupled systems , and it misses some of the most dramatic features that are actually quite common . with no basic understanding provided by the elementary physics canon , when analogous aspects arise in particular situations , the people involved sense an aspect of discovery . sometimes the newly " discovered perspective has had major impact . although the mechanics of such linear systems has been understood , in principle , for hundreds of years , rediscovery of their special features has occurred even into the @xmath0 century . here
are some examples of such rediscoveries .
understanding the @xmath1@xmath2 meson system@xcite in the 1950 s laid the groundwork for the experiments that identified cp and t ( time reversal symmetry ) violations in the fundamental interactions.@xcite@xmath3@xcite mechanical engineers in the 1960 s were interested in earthquake shaking of buildings,@xcite@xmath3@xcite which can exhibit something now known as transient growth after the initial quake ground shaking has ceased .
the origin of the characteristic sound of the piano@xcite ( in contrast to earlier stringed , keyboard instruments ) was elucidated in the 1970 s .
stability analyses in fluid mechanics and analogous problems in applied linear algebra witnessed a major revolution starting in the late 1990s.@xcite@xmath3@xcite
transient growth " is the generic term applied wherever all eigenvalues and eigenvectors decay exponentially in time and yet some combinations initially grow before ultimately dying off . and additional examples of systems and situations of this type continue to be identified . of course , numerical integration of differential equations has gotten easier and better over the years . and
it has often been observed that what are now identified as non - normal " linear systems sometimes exhibit surprising transient behavior , quite sensitive to parameters and initial conditions .
but that is not a substitute for thorough understanding of at least one simple , mechanical system .
what is needed is an explicit example that exhibits the phenomena particular to the _ simultaneous _ presence of _ both _ coupling and damping and yet is visualizable and analytically tractable . with respect to coupling , as simple as possible means only two oscillators . with respect to damping , as simple as possible means weak damping .
but solution of the general problem of this form involves solving a quartic polynomial .
quartic is the highest order polynomial for which a closed - form solution exists .
that solution has been known since the @xmath4 century , but it is far longer than most people can remember or comprehend at a glance .
reducing the algebra to quadratic equations requires weak coupling as well .
but even that is not enough .
we will need a system whose frequencies , before coupling and damping , are degenerate or nearly degenerate , i.e. , the same .
a bonus of having the coupling and damping be weak is that the combined behavior shows vestiges of the familiar separate problems of the single damped oscillator and of undamped coupled oscillators .
a mechanical system satisfying those requirements is illustrated in fig .
[ springs ] .
two identical oscillators are coupled with a weak spring , @xmath5 , but only one of the oscillators is weakly damped . with sufficient care , such an apparatus could be constructed .
in fact , bruce winstein@xcite , when he gave colloquia and talks about cp violation , brought along just such a mechanical model that he had made , which included finely adjustable damping and coupling .
makes explicit the connection of the neutral kaons to a two coupled degenerate mode damped system . on the other hand , plucking a typical stringed instrument readily provides the appropriate degrees of freedom and parameter ranges . as background
, it is useful to understand the generic case of two weakly coupled , degenerate oscillators . in typical demonstrations , damping is minimized . these could be springs,@xcite pendulums,@xcite or , most charmingly , the wilburforce pendulum.@xcite essential features are the small splitting of the normal frequencies , identification of the normal modes , and dependence of the qualitative motion on initial conditions in particular , the possibility of beats .
the other essential bit of background is the decomposition of string motion into normal modes .
the ideal string can be analyzed as a set of normal modes , which are uncoupled oscillators whose frequencies are integer multiples of the lowest , fundamental " frequency .
what is often not mentioned is that for a physical string , each of these modes is actually a degenerate pair , corresponding to the possible vertical and horizontal motion at the given frequency . in the simplest approximation ,
these are totally independent motions , which are then superposed into the motion of the single physical string . and for the string actually to be musical , " the damping and the coupling must be _ both _ small _ yet _ non - negligible .
there are several related meanings of the term non - normal . "
here , it refers simply to matrices with complete sets of eigenvectors which are not all orthogonal to one another .
they are common elements of descriptions of a variety of systems including ones that exhibit transient growth where all eigenvectors , individually , decay monotonically.@xcite@xmath3@xcite@xmath3@xcite as described in the following , the total energy of two coupled , damped oscillators decays in time , no matter what the initial conditions or parameter values . however , there are ranges of parameters and initial conditions for which the amplitude and energy of _ one _ of the oscillators can grow before eventually decaying . in such cases ,
the decay of the total system s energy can come in spurts , rather than a steady one- or two - rate exponential .
there are , of course , quantum mechanical analogs , where more sophisticated methods and different approximations yield a complementary set of solutions.@xcite
the ideal string stretched taut between two fixed points has normal modes with frequencies that are integer multiples of the fundamental . in a musical instrument , there are actually two degenerate modes for each frequency , reflecting the possibility of string displacements in the plane transverse to the string direction .
there are always some further interactions that couple those degenerate modes .
the result is that their frequencies are split an amount proportional to the strength of that coupling .
the coupling has the further consequence of establishing the form of the actual normal modes , i.e. , the motions that have precisely just one of those two frequencies . and
the modes thus determined are not coupled .
( it is said that the perturbation diagonalizes the interaction . ) in the specific context of a stringed instrument , it is important that the splitting be small so that any combination of the two is perceived as essentially a single pitch . by itself , the vibration of a string produces almost no sound because the tiny cross section of the string moves almost no air .
there must be some further transduction of the string motion to air motion . in acoustic instruments ,
that is accomplished by linking one end of the string to a sound board .
string oscillations force sound board motion , which in turn produces sound .
hence , at least one fixed " end of the string is not actually fixed .
furthermore , in a good musical instrument , that end motion is the string s primary loss of energy . and again , that damping must be weak so that the consequent width or spread in frequency due to the damping leaves a single discernible pitch rather than noise .
so , for the present , we focus on a single original frequency . the system is approximated as two initially degenerate oscillators which are weakly coupled and weakly damped .
the catch is that the coupling and the damping , represented as matrices in the configuration space of the two initial oscillators , are generally not simultaneously diagonalizable .
if the system is a particular mode of a string , we may choose as a basis the up - and - down motion ( relative to the sound board ) and the side - to - side .
the coupling of vibration to the sound board is typically far more effective for vertical motion than for horizontal . to make matters as simple as possible
, we will consider explicitly small vertical damping ( denoted by @xmath6 ) and no horizontal damping .
if the system is started out in one of the modes of the undamped , coupled problem , then the non - zero damping will decrease the amount of vertical motion relative to the horizontal .
unless the undamped modes were precisely pure vertical motion for one mode and pure horizontal motion for the other , this damping would disturb the balance that defined the normal mode .
thus , the damping mixes in some amount of the other mode . what is it that actually happens ? we choose the original restoring forces , the coupling , and the damping so that the system is linear ( e.g. , with the damping proportional to the velocities ) . with positive
damping , the total energy must decrease monotonically with time .
it is handy ( and virtually essential ) to use a complex number representation of the frequencies and displacements ; their superposition into real motions at the end is totally parallel to standard treatments of the free decay of the single damped harmonic oscillator .
we will find that there are , indeed , two eigenfrequencies .
they describe exponential decay multiplied by sinusoidal oscillation .
they have corresponding eigenvectors , which are possible motions that follow their single eigenfrequency .
however , these vectors have complex components , which , translated into the real motion of strings , means that their motion in the transverse plane is elliptical . and , finally , these eigenvectors are not orthogonal .
one consequence is that the total energy and the rate of energy dissipation ( which is the volume of produced sound in this simple model of a string ) are not necessarily the sum of two independent , decreasing exponentials .
newton s second law yields two coupled differential equations in time that are linear in the two displacements .
write the displacements , @xmath7 and @xmath8 in vector form with @xmath9 for the plucked string , @xmath7 is a given mode s vertical motion , and @xmath8 is the horizontal motion .
the equations of motion take the form @xmath10 where * k * and @xmath11 are @xmath12 matrices representing the zeroth order restoring forces , the coupling , and the damping .
( the general mathematical problem would include a mass matrix multiplying @xmath13 . ) the unit of time can be chosen to make the zeroth order * k@xmath14 = i * , the unit matrix .
the most general possible form of * k * is hermitian , which means it has three real parameters for this mode pair .
( the fourth parameter is absorbed into the average frequency , which is normalized to 1 . ) for the present , however , choose just the simplest possible mathematical form the has the feature that the eigenvectors of * k * are mixed by the damping .
that simplest , one parameter form is @xmath15 this particular * k * describes the mass / spring system of fig . [ springs ] . from the standpoint of actual stringed instruments ,
it is theoretically possible but may be an oversimplification .
however , with only a single parameter in * k * , the ensuing math is far easier to follow . at the end
, it will also be easy to see how the same calculations with the most general possible * k * would go through and see what aspects of the results are generic .
the damping , as described above , takes the form @xmath16 for weak coupling and damping , units are chosen such that all of the angular frequencies are close to 1 .
there is no basis in which * k * and @xmath11 are both diagonal .
a commonly recognized reflection of this is that the commutator @xmath17 = \mathbf{k } \cdot \mathbf{\gamma } - \mathbf{\gamma } \cdot \mathbf{k } = \begin{pmatrix } 0 & ~ ~ - \epsilon \gamma \\
\epsilon \gamma & ~ ~ 0 \end{pmatrix } \ne 0 ~ .\ ] ] ( of course , there are other forms of damping for which the damping and coupling matrices do commute and are simultaneously diagonalizable .
this is known as proportional damping , and the separation into normal modes is straightforward . )
we seek eigenvalues @xmath18 and time - independent eigenvectors @xmath19 such that e@xmath20 is a solution to eq .
( [ eq - mot ] ) . plugging that in yields @xmath21 where * i * is the identity matrix .
@xmath22 is a solution to eq .
( [ plug - in ] ) but not to the problem at hand . for all other solutions , the matrix factor in eq .
( [ plug - in ] ) can not have an inverse , and that requires that its determinant vanish , which is the following characteristic equation " : @xmath23 this would be easy to solve were either @xmath6 or @xmath24 zero , but as it stands it is a quartic equation .
its analytic solution is hundreds of characters long and contains many nested square and cube roots .
most people find it impossible to discern by inspection the qualitative properties for particular values of the parameters .
just as the single damped oscillator has cases with radically different qualitative behavior , i.e. , over damped , under damped , and critically damped , there are cases here , too only many more . for weak coupling and weak damping , we can anticipate the structure of the solutions from physical considerations . with @xmath25 , the four solutions for @xmath18 will be two complex conjugate pairs .
re[@xmath18 ] will be negative , reflecting the monotonic loss of energy .
im[@xmath18 ] comes in conjugate pairs .
these conjugate pair solutions can ultimately be superposed to get real solutions with sines and cosines of @xmath27 with the same frequency . and
this multiplicity of solutions allows fitting of any initial conditions of the two oscillators .
if _ both _ @xmath28 and @xmath29 , then _ all _ the frequencies will be near to 1 , i.e. _ i _ im[@xmath30 \approx \pm i$ ] .
this offers a way to approximate eq .
( [ det ] ) and reduce the algebra problem to a quadratic.@xcite in the term @xmath31 , approximating the first @xmath18 by @xmath32 leaves the whole term still as small as @xmath33 and @xmath34 , at least in the vicinity of the desired solutions .
so , using this approximation , eq .
( [ det ] ) becomes @xmath35 whose solutions are @xmath36 if the three @xmath37 s were chosen independently , there would appear to be eight solutions .
however , the coefficient @xmath32 of @xmath38 is an approximation to @xmath18 .
hence , that @xmath39 is @xmath40 when @xmath41 . that selects the four actual solutions out the apparent eight , which can also be identified by their having re@xmath42 < 0 $ ] for @xmath43 .
so there are really only two @xmath37 s coming from the sequential solution of two quadratic equations , and there is no ambiguity in the term @xmath44 . the separate limits @xmath45 and @xmath46 recover the previously understood behaviors of single damped oscillators and coupled , undamped oscillators .
there are evidently three qualitatively different regions , even with both @xmath28 and @xmath29 . with @xmath47 , the square root term contributes to the oscillation frequency ; there are two oscillation frequencies but only one decay rate , which is independent of @xmath24 . with @xmath48 ,
the square root term affects the two decay rates , and there is no splitting of the oscillation frequency degeneracy . and
for @xmath49 , the frequencies and decay rates of the two eigenmodes are nearly equal . in the appendix ,
evaluations of eq .
( [ approx ] ) for three numerical pairs of @xmath24 and @xmath6 , representative of the three regions , are compared to the exact values that come from solving eq .
( [ det ] ) .
let the components of the four eigenvectors be @xmath50 and @xmath51 : @xmath52 ( the eigenvalues @xmath18 and eigenvectors @xmath53 , with their components @xmath50 and @xmath51 , have a four - valued index @xmath54 to tell which goes with which .
that index @xmath54 is suppressed when that improves clarity . ) the lower component of eq .
( [ eq - mot ] ) tells us that @xmath55 this specifies the four eigenvectors , one for each @xmath18 .
( recall that @xmath56 only for @xmath57 and @xmath58 for @xmath59 . ) the first expression for @xmath60 with the = sign is exact relative to the initial statement of the problem , i.e. , eq .
( [ eq - mot ] ) ; the approximate solutions to eq .
( [ quad - det ] ) are used for the @xmath61 expression .
the ratio @xmath60 is of order 1 ( because @xmath62 ) . also , @xmath60 is complex .
the phase of each @xmath60 means that in the oscillatory part of the motion corresponding to a single eigenvalue , there is a fixed , non - zero phase difference between the @xmath7 and the @xmath8 . in the language of the plucked string : in the transverse plane , the eigen - motions are elliptical rather than linear ( which they would be in the absence of damping ) .
it is worth returning to the original physical problem and constructing the basis of real eigenfunctions .
the four eigenvalues @xmath18 are two pairs of complex conjugates . label them as @xmath63 and @xmath64 , where there are two , in general , different , negative real parts , each with a pair of conjugate imaginary parts .
the @xmath63 s can be assembled into two real eigenfunctions : @xmath65 @xmath66 these forms use the facts that @xmath67 and @xmath68 ( true in the original , exact formulation ) .
likewise , there are two real functions @xmath69 similarly constructed out of the conjugate pair @xmath64 .
appropriate superpostion of the four @xmath70 s can match the two initial positions and velocities .
the normal modes of linearly coupled , undamped oscillators behave themselves , essentially , like a set of uncoupled oscillators .
whatever superposition is determined by the initial conditions remains in force for all time .
many quantities of central importance , such as the kinetic energy , the potential energy , and the total energy of the system are , at any time , just the sum of the contributions from the normal modes .
since these particular quantities are quadratic in the dynamical variables , the reduction to a sum over modes requires that cross terms between the contributions of different modes vanish . and
that is the sense in which the normal modes are normal to each other .
however , for the generic case of coupled , damped oscillators , such cross terms are non - zero .
hence , there are no normal modes in spite of there being a complete set of solutions corresponding to the time - dependence eigenvalues .
as long as @xmath71 , for convenience we can choose all four @xmath72 .
then , for example , at @xmath73 : @xmath74\text{re}[b_{+2 } ] ) \nonumber \\
\nonumber & \ne & 0 \quad \textrm{for}\ \gamma / \epsilon \ne 0 ~.\end{aligned}\ ] ] once again , this result is not particularly surprising , except possibly to those imbued with an overwhelming respect for normal modes .
if the coupled , damped system were exactly describable by normal modes , then the total energy would decay steadily as the sum of one or two exponentials . however , if the corresponding undamped system could exhibit beats , then with the addition of very weak dissipation to only one of the ( pre - coupled ) degrees of freedom , dissipation should likewise come and go at the beat frequency . and
that is , indeed , one of the possible generic behaviors .
if the simple model analyzed thus far is interpreted as describing a mode doublet of a plucked string , then the produced sound volume would be proportional to the rate of energy lost to damping of the vertical motion of the particular string harmonic .
the instantaneous value of this lost power is @xmath75 . in fig .
[ pluck ] , the log of the several - cycle - averaged @xmath76 is plotted _ verus _ time for @xmath77 and @xmath78 with the @xmath73 condition that the pluck is purely in the horizontal ( undamped ) direction . the horizontal motion
@xmath8 is the lower component of @xmath79 .
the numerical parameter values in the appendix were chosen for convenience of computer entry , with the consideration that they be small but realistic for stringed instruments .
of these three pairs , the values used for fig . [ pluck ]
are the ones that exhibit beats , i.e. , there are two distinct oscillation frequencies with the beat period being distinctly shorter than the damping time . and the horizontal pluck was chosen for display because time must elapse after the pluck before the coupling to the dissipative vertical motion is substantial . hence , the dissipated power _ grows _ immediately after the pluck before it subsequently decays . with a vertical pluck
, the dissipation would be evident from the start but would also exhibit the beats between the two eigenfrequencies . the dissipated power for the other domains of @xmath24-@xmath6 space would look like single or double exponential decay without beats .
strictly speaking , the single exponential behavior arises at a point in the parameter space .
the two exponential rates become closer and closer as one approaches that point . on the other side of the point , with only a single exponential ,
the frequency splitting begins at zero and then grows .
the most general possible form of * k * that is close to the identity is @xmath80 where @xmath81 is a small detuning of the vertical and horizontal modes and @xmath82 is a small , ( in general ) complex coupling
. a non - zero phase or imaginary part of @xmath82 would give the undamped normal modes a phase shift between the vertical and horizontal components .
if the system under discussion were , indeed , a string , then that phase shift would give an elliptical motion of the undamped normal modes the plane transverse to the string .
when this general * k * is used in eq .
( [ plug - in ] ) , eq .
( [ quad - det ] ) is replaced by @xmath83 whose solutions are @xmath84 where the coefficients of @xmath6 are chosen to have the same sign of @xmath54 as the @xmath32 coefficient of the leading 1 term ( because their coefficient @xmath54 is an approximation to the same @xmath18 ) .
this coincides with the physical condition re[@xmath18 ] @xmath85 0 for @xmath43 .
if @xmath81 is negligible , the phase of @xmath82 would only contribute to the phase difference between the components of the eigenvectors , i.e. , the analog of eq .
( [ eigenvec ] ) .
this is in addition to the phase difference implicit in eq .
( [ eigenvec ] ) that arises from the simultaneous real coupling and damping .
the eigenvalues depend only on @xmath86 and would be qualitatively as before ( with real @xmath24 ) .
if @xmath86 were negligible , it would be the simple case of proportional damping .
the oscillation frequencies would be @xmath87 , and only the @xmath88 mode would decay , with decay constant @xmath89 . if both @xmath86 and @xmath90 are relevant , then the square root in the above expression for @xmath18 is generically complex with a @xmath37 coefficient .
this means that the most general behavior is that the two non - orthogonal modes have different frequencies and different decay rates . besides the trivial comment that the magnitudes of the frequency and decay splitting depend on the parameters , it is worth remembering that in practice the amount of time one has to observe the beats and the second decay rate is limited by the decay
the subject of what happens in real stringed instruments is a rich one .
the simplest features were recognized in antiquity .
details have been the subject of serious research over the past 150 years and continues to this day .
professional journals in which this research is reported include ( but are not restricted to ) the journal of the acoustical society of america , acta acoustica united with acoustica , and the journal of sound and vibration .
the present discussion has focused on effects that are small compared to the basic oscillatory motion ( here normalized to a frequency equal to 1 ) .
furthermore , attention was restricted to effects that are linear in terms of the differential equations .
this makes the mathematical exercise of relevance to a great number of other physical problems .
however , in terms of the behavior of an actual plucked string , there may be non - linear effects that are small enough to ignore relative to that 1 but are significant in the present context .
many effects have been discussed in the scientific literature .
but there is no consensus as to which is the correct one . " in fact , the relative importance of different mechanisms may depend on the kind of instrument and even the individual instrument .
the most basic lesson to take away is that , when dealing with the interactions of frequency degenerate systems , very tiny forces can lead to clearly visible effects if they can act over many cycles of the primary oscillation .
the following are just a few of possible contributions to the parameters in * k*. one contribution to the detuning " @xmath81 might be an approximate description of consequences of the up - down motion of the bridge end of the string being much larger than the horizontal motion . for a particular vertical mode , the bridge end is no longer an actual node , and the effective length of the string is longer than were it a node .
a small skewness of the elastic constants relative to vertical / horizontal ( i.e. , @xmath24 or re[@xmath82 ] ) can arise from imperfections and asymmetries in the materials of the string and its supports , which can rotate the principle axes of the restoring forces relative to exactly parallel and perpendicular to the head .
the longitudinal stretching of the string is usually successfully ignored at the level of the 1 , but its small effects can be relevant when one looks closer .
sometimes an effort is made to approximate those effects and represent them as contributions to the parameters of the generic linear system .
players of stringed instruments easily notice that the motion in the transverse plane is typically elliptical , and , furthermore , those ellipses typically precess .
this would happen even if the only deviation from degeneracy were a detuning of the two components and the initial pluck was not exactly in one of the normal mode directions .
a separate issue is whether the undamped normal modes themselves describe elliptical motion , i.e. , whether @xmath82 has a phase .
this would not arise from a combination of linear springs , but it can not be ruled out as an effective , approximate description of the combined effects of linear and non - linear dynamics . even given the uncertainties described above , it is interesting to look at a case of measured , real string motion where the investigators projected out the time evolution of particular , individual modes
. measurements of the sounds of plucked banjo strings were published by moore and stephey.@xcite for one aspect of their experimental survey , they damped all strings but the first , plucked it , first in the vertical direction ( perpendicular to the banjo head ) and then in the horizontal .
they did the same for the second string .
they recorded the sounds and analyzed them into fourier components . in fig .
[ plucked - strings ] ( copied from fig . 5 of ref . )
the ( logarithmic ) sound intensities for the first three harmonics of each plucked string are displayed as functions of time .
each harmonic acts as a separate nearly degenerate , coupled , damped pair . as those authors noted in their paper ,
evident are single exponential decays , double exponential decays , and decays with prominent beats modulating a single overall decay rate .
the qualitative agreement with the forms of the time dependence predicted by the initial model of * k * with a single , real parameter @xmath24 does not rule out the more general form .
rather , it suggests that the region of four dimensional parameter space of the generic linear model populated by the experimentally observed modes is roughly two dimensional and might be well - approximated by an effective @xmath24 .
more measurements would be needed to do better . with just the naked eye ,
some of this behavior is typically visible on a stringed instrument . in particular
, there is usually at least one string that after a pluck exhibits beats .
instead of decreasing steadily , its amplitude gets smaller and larger again a couple or even several times before it dies completely .
( generally , each maximum is smaller than the previous one . )
identifying double exponential decay is harder , just judging by eye .
however , there are typically some pluck responses with a dramatic initial fall off and a surprisingly long tail .
this phenomenon is actually a very important aspect of banjo sound .
the banjo is an instrument where the degeneracy is often four- and even six - fold , not just the two of a single string .
that is because , as normally tuned and played , the undamped strings are often in unison or share harmonics ( e.g. , the second harmonic of one string is degenerate with the third harmonic of another ) . and the design of the bridge facilitates coupling between between all of the strings .
computer math packages include the exact solution for the roots of the general quartic polynomial . hence , their numerical evaluations of eq .
( [ det ] ) are unassailable .
the table below gives comparison of the approximate eigenvalues from eq .
( [ quad - det ] ) to the exact eigenvalues from eq .
( [ det ] ) for @xmath77 and @xmath78 , 0.02 , and 0.03 .
rahul bale and rama govindarajan , transient growth and why we should care about it , " resonance * 15 * ( 5 ) 441457 ( 2010 ) ; this article offers a fine , elementary introduction to the subjects addressed in ref.s and .
m. bhattacharya , m. j. a. stoutimore , k. d. osborn and ari mizel , understanding the damping of a quantum harmonic oscillator coupled to a two - level system using analogies to classical friction , " am .
* 80 * ( 9 ) 810 - 815 ( 2012 ) .
the analysis of coupled , damped strings in ref .
likewise uses near degeneracy as a key approximation .
however , the methods and tools of analysis used there assume considerably greater sophistication on the part of the reader . also , with somewhat different goals , several of the very general features are not remarked upon as such . | motion of a single fourier mode of the plucked string is an example of transient , free decay of linear , coupled , damped oscillators .
it shares the rarely discussed features of the generic case , e.g. , possessing a complete set of non - orthogonal eigenvectors and no normal modes , but it can be analyzed and solved analytically by hand in an approximation that is appropriate to musical instruments plucked strings . |
one of the most intriguing discoveries made in solar physics early last century was the phenomenon known as ellerman bombs .
the common use of the h@xmath0 line to monitor the sun allowed the discovery of features with enhanced emission peaks in the wings of this line , not affecting the central absorption core ( @xcite @xcite ) .
off - band ( @xmath3 1 ) h@xmath0 filtergrams showed these features to be spatially point - like structures with a size of @xmath31 , lasting for little more than 10 minutes ( see @xcite @xcite and references therein ) .
as they are virtually absent from line core filtergrams , they have always been associated to events taking place in the low chromosphere .
this is also confirmed by their clear correlation with brightenings seen in the 1600 continuum images from the trace satellite ( @xcite @xcite ) .
sometimes , an x - ray transient brightening can also be observed at the same location ( @xcite @xcite ) .
recent reviews on ellerman bombs are given by @xcite ( @xcite ) and in the introduction of @xcite ( @xcite ) .
the physical process behind ellerman bombs has remained elusive . the blue asymmetry observed by the h@xmath0 profiles ( excess emission in the blue wing )
has sometimes been associated to outward motions of the solar material ( @xcite @xcite ) . but using two - component models separated along the line of sight , @xcite ( @xcite ) showed that the asymmetry could be originated from , either , hot emitting material moving upwards or cold absorbing material falling from the top of the chromosphere .
the relation to the magnetic field configuration has been controversial as well .
ellerman bombs have been detected in two distinct solar magnetic scenarios , but always inside active regions .
they occur in emerging regions between the opposite polarity footpoints and also surrounding isolated sunspots , beyond the penumbral boundary . for this later case , @xcite ( @xcite )
have shown that they can be identified with the so - called moving magnetic features ( mmfs ) that are seen in the moat region around sunspots ( for a description of mmfs see @xcite @xcite ) .
whereas this points towards a magnetic origin of ellerman bombs , a number of events observed by @xcite ( @xcite ) ( see also @xcite @xcite ) were not clearly linked to magnetic structures .
these authors point out that a number of cases showed a relationship with the boundaries of magnetic regions but others did not .
more recently the situation has been clarified by the analysis of the observations made by the flare genesis balloon experiment ( @xcite @xcite , @xcite @xcite ) .
observing an emerging flux region , these authors describe dipolar features that appear in the middle of the region , where the flux emerges at the surface ( below the so - called arch filament system ) . with better sensitivity magnetograms than the previous works
, @xcite ( @xcite ) found that almost all ellerman bombs studied could be identified with the neutral line above the dipolar features ( or in neutral lines inside supergranular boundaries ) .
those cases not linked with the presence of neutral lines could be associated with interacting field lines above the photosphere , a result confirmed by the extrapolations made by @xcite ( @xcite ) .
these authors proposed , then , as the most probably mechanism for the generation of ellermam bombs some form of magnetic reconnection in the low chromosphere .
that reconnection happens efficiently at these heights is supported by the fact that the temperature minimum is the region with the lowest electrical conductivity , and resistive processes occur there more easily than elsewhere ( @xcite @xcite ) . while the flare genesis observations correspond to the central portion of an emerging active region ,
these conclusions offer an interesting parallelism with the ellerman bombs observed surrounding sunspots and associated to mmfs .
@xcite ( @xcite ) pointed out that many of the bombs studied were spatially located at the dipped portions of undulatory field lines in the flux emerging area .
this undulatory or serpentine configuration is also used for the explanation of mmfs ( as first proposed by @xcite @xcite ) . in this way , a unifying scenario for the generation of these point - like energy release processes could be constructed . but a study similar to that made by @xcite ( @xcite ) for the moat of sunspots is still lacking . the comparison between the ellerman bombs seen in these two scenarios and the determination of their physical origin awaits for more solid observational tools than h@xmath0 line profiles .
in particular , techniques providing information of the magnetic field at the chromospheric heights where the energy release takes place , will be of great value .
linear polarization signals observed in ellerman bombs have been reported in the past ( @xcite @xcite ) , but their exact amount an origin remains controversial . in this paper , we propose the use of full stokes spectropolarimetry in more reliable diagnostic lines such as the caii infrared triplet lines . the most important advantage these lines offer is the availability of powerful inversion techniques that can be used to cover almost the full chromospheric region .
this paper presents a first characterization of the atmosphere of two regions close to a sunspot that display the properties ascribed to the ellerman bombs .
the observations used in this work are first - light data from the new instrument spectro - polarimeter for infrared and optical regions ( spinor , see @xcite @xcite ) , at the dunn solar telescope ( dst ) located at the sacramento peak observatory .
the dataset contains the two chromospheric caii lines at 849.8 and 854.2 nm and two photospheric fei lines at 849.7 and 853.8 nm . we observed active region noaa 0634 on june 16 at 15:16 ut , carrying out a 350-step scan with a spatial stepping of 0.22 . by using the adaptive optics system of the dst @xcite
, we attained a spatial resolution as good as 0.6 , although this figure varies slightly along the scanning direction ( the horizontal direction in the figures ) due to temporal seeing variations . in order to improve the signal - to - noise ratio in the stokes profiles , we averaged the observations in 3@xmath13 pixel boxes .
the noise is thus reduced to @xmath33@xmath110@xmath2 times the continuum intensity . at this level
there is barely no signal in stokes @xmath4 and @xmath5 , so we decided to leave them out of our analysis .
the chromospheric stokes @xmath6 profiles , however , reveal fascinatingly complex shapes with multiple peaks , while the photospheric lines exhibit the usual antisymmetric profiles .
fig [ fig : maps ] shows maps of the continuum intensity and a `` magnetogram '' taken in one of the photospheric lines .
the magnetic configuration surrounding the sunspots appears very different on both sides of it .
the eastern side exhibits a much more complex topology with an intricate pattern of plage fields extending from the sunspot boundary to the edge of the field of view .
images of the region ( fig [ fig : ha ] ) reveal considerable emission near the eastern penumbral boundary .
photospheric velocity maps of noaa 0634 were obtained by measuring the minimum position of the fe i line at 849.7 nm .
a similar strategy is not viable for the ca ii lines , however , because of the complicated patterns of emission reversals and self - absortions found in the line cores .
instead , we measured the intensity difference between two points symmetrically located at various distances from the line core .
this method is only an approximation but it works better than taking the minimum position .
fig [ fig : vels ] shows the velocity maps obtained at two heights , with the convention that positive velocities are away from the observer .
the cosine of the heliocentric angle for this region is @xmath7 .
notice the two prominent bright features marked with arrows in the lower panel of the figure .
these features are strong plasma flows directed away from the line of sight and located on the center side of the sunspot ( but outside the photospheric penumbral boundary ) .
detailed profiles of these features are shown in fig [ fig : velprofs ] .
both features exhibit very strong redshifts , visible in stokes @xmath8 and @xmath6 .
point a shows a very prominent redshift starting near 40 pm away from line center ( corresponding roughly to the high photosphere or low chromosphere ) all the way to the line core ( middle chromosphere ) .
this feature is well localized ( see lower panel ) and the doppler shift is @xmath314 km s@xmath9 .
point b exhibits somewhat weaker flows ( @xmath310 km s@xmath9 ) that are more localized in height ( upper photosphere / lower chromosphere ) , but has a more extended tail along the slit ( see lower panel ) .
notice how both points a and b coincide with h@xmath0 emissions in fig [ fig : ha ]
we performed detailed non - lte inversions of the area surrounding points a and b in fig [ fig : vels ] .
the inversions were carried out with the code of @xcite ( @xcite ) ( see also @xcite @xcite ) by applying it systematically on each pixel independently . in order to minimize the risk of the algorithm settling on a secondary minimum , we repeated each inversion 10 times with randomized initializations . the model atmosphere that yields the best fit to the observed stokes profiles
is taken as a good representation of the solar atmosphere under study .
the inversion code treats the caii lines in non - lte and the fei lines in lte .
blends are treated consistently by adding the opacities of all the relevant transitions at each wavelength and spatial point , and then solving the radiative transfer equation with the total opacity .
the non - lte calculation is based on the preconditioning strategy of @xcite ( @xcite ) ( see also @xcite @xcite ) assuming complete angle and frequency redistribution .
this last approximation is particularly good for the caii infrared triplet lines , as shown by @xcite ( @xcite ) .
let us consider in detail the models obtained for the points that we labeled as a and b. figures [ fig : modelsa ] and [ fig : modelsb ] show the atmospheric stratification inside the downflows , as well as an average over the surrounding , non - downflowing area for reference .
the most striking features in these models are the strong downwards velocities found in the upper photosphere , around @xmath10 .
the motions along the line - of - sight approach the speed of sound in point b and exceed it in point a. the other parameters show different behaviors in points a and b. in particular , the downflowing layer in point b is considerably hotter than the reference atmosphere .
however , in point a the downflowing layer is cooler ( although it then becomes hotter at higher layers ) . in both cases
there is another hot layer at @xmath11 .
our findings for the two regions are consistent with the results of @xcite ( @xcite ) who found ellermam bombs to be predominantly dominted by downflows with no indication of upflows .
the fits provided by the inversions can be seen in fig [ fig : profs ] .
note that the observed chromospheric profiles exhibit a very complex structure with multiple peaks in stokes @xmath6 ( the photospheric lines , on the other hand , produce rather normal profiles ) .
generally speaking , the inversion code yields a reasonably good fit to the overall features , but there is more structure in the observed profiles ( particularly noticeable in stokes @xmath6 with multiple peaks ) , than in the synthetic ones .
this is an indication that the actual atmospheres are even more complex than our models , probably due to the presence of unresolved atmospheric components .
thus far , we have not made a direct connection between the downflowing features analyzed here and ebs . to this aim
, we used the model atmospheres obtained from the inversion to synthesize the h@xmath0 profiles that these models produce . the result is plotted in fig [ fig : halpha ] , along with a reference quiet sun profile calculated from model c of @xcite ( @xcite ) ( val - c ) .
this synthetic profile has the same appearance traditionally observed in ebs .
@xcite ( @xcite ) describes the h@xmath0 spectra of his bombs as having emission features ( brighter than the continuum ) on both sides of the line center
. the red asymmetry ( higher intensity peak in the red wing ) shown here is not the one most commonly observed in ellerman bombs ( see the introduction ) , but this asymmetry is also present quite often ( 40% of the cases according to @xcite @xcite ) .
it is the result of a hot downflowing component as seen in fig [ fig : modelsa ] .
the resemblance of ebs with the synthetic profile in fig [ fig : halpha ] is a solid evidence that we are indeed observing the same phenomenon .
moreover , it also provides strong support for the reliability of the inversions .
the energetic phenomenon generally referred to as ebs has remained very ellusive to solar physicists for almost a century .
this is probably due in no small measure to the fact that it has only been observed in h@xmath0 thus far , which is a very difficult line to interpret . in this paper
we present the first observations of ebs in other lines , namely the caii infrared triplet .
these lines are much easier to model and have been successfully inverted in the past few years ( e.g. , @xcite @xcite ; @xcite @xcite ; @xcite @xcite ) and thus our results open a new window to the observational study of ebs .
we find that the ebs analyzed in this work are associated with supersonic ( or nearly so ) downflows in the upper photosphere .
no upflows are observed in our data .
@xcite ( @xcite ) determined the vector current densities in the large sunspot of fig [ fig : maps ] and found that the east penumbra is permeated by strong currents in the upper photosphere , with a magnitude of @xmath33@xmath110@xmath12 a km@xmath13 .
the fact that these ebs occur in a region with strong currents suggests that ebs may be associated with magnetic reconnection . in order to understand the nature of ebs and the physical processes that produce them
, it would be extremely valuable to have time - series observations of photospheric and chromospheric lines , ideally with full stokes polarimetry .
the authors wish to acknowledge the enthusiastic support from the nso staff at the sacramento peak observatory , especially d. gilliam , m. bradford and j. elrod .
thanks are also due to s. hegwer , s. gregory , r. dunbar , t. spence , s. fletcher , c. berst and w. jones . ,
d. , s. l. hegwer , t. rimmele , l. v. didkovsky , and p. r. goode : 2003 , ` optical design of high - order adaptive optics for the nso dunn solar telescope and the big bear solar observatory ' . in : _
innovative telescopes and instrumentation for solar astrophysics . edited by stephen l. keil , sergey v. avakyan .
proceedings of the spie , volume 4853 , pp .
593 - 599 ( 2003 ) .
_ pp . 593599 . | ellerman bombs are bright emission features observed in the wings of h@xmath0 , usually in the vicinity of magnetic concentrations . here
we show that they can also be detected in the ca ii infrared triplet lines , which are easier to interpret and therefore allow for more detailed diagnostics .
we present full stokes observations of the 849.8 and 854.2 nm lines acquired with the new spectro - polarimeter spinor .
the data shows no significant linear polarization at the level of 3@xmath110@xmath2 .
the circular polarization profiles exhibit measureable signals with a very intricate pattern of peaks . a non - lte analysis of the spectral profiles emerging from these features reveal the presence of strong downflows ( @xmath310 km / s ) in a hot layer between the upper photosphere and the lower chromosphere . |
fliess operators @xcite , also known as chen fliess functional expansions , play a crucial role in the theory of nonlinear control systems @xcite .
most systems found in science and engineering can be viewed as a set of simpler interconnected subsystems .
ferfera was the first to describe some of the mathematical structures that result when these interconnected subsystems are modeled as fliess operators @xcite .
four basic system interconnections are found in most control systems .
they are known as the parallel sum , parallel product , cascade and feedback connections .
( see @xcite for a review . )
the first two are rather simple to describe mathematically .
the third operation is equivalent to composing fliess operators . combining fliess operators through a feedback connection
as shown in figure [ fig : feedback - with - v ] is also based on composition . in short ,
the output of one system is fed into the input of another system , the output of which is then fed back into the input of the first .
this feedback loop is , mathematically speaking , the most interesting interconnection .
when one or both of the subsystems is nonlinear , the nature of the composite system can be remarkably complex . despite this , it turns out that given two nonlinear input - output systems written in terms of such chen fliess functional expansions , it is known that the feedback interconnected system is always well defined and in the same class @xcite . an explicit formula for the generating series of a single - input , single - output ( siso ) closed - loop system was provided by two of the authors , l. duffaut espinosa and w.s .
gray , in earlier work @xcite in the context of a fa di bruno type hopf algebra .
l. foissy @xcite made a crucial contribution by showing that this hopf algebra is connected .
as a result , a simple standard recursive formula could be used to calculate the antipode @xcite , which plays an important role in the analysis and design of feedback systems @xcite . in a more recent work ,
the full multivariable extension of the theory as it applies to control theory is presented @xcite .
an obvious but largely non - physical multivariable extension is provided in @xcite .
the focus here will be on the extension in @xcite which , as will be described later , has a richer structure and more realistic properties from the point of view of proper applications in control theory . and
@xmath0 foissy s insight regarding the grading in @xcite brought into focus the combinatorial nature of the fa di bruno type hopf algebra found by l. duffaut espinosa and w.s .
however , the central coproduct was presented in terms of two recursions , which made it difficult to understand its representation as a combinatorial algebra .
m. aguiar characterized the latter in terms of a connected graded vector space with homogeneous components spanned by finite sets of combinatorial objects , and the algebraic structures are given by particular constructions on those objects . in the present work , a set of combinatorial objects
is given that allows one to realize the algebraic structures of the aforementioned multivariable extension of the hopf algebra of l. duffaut espinosa and w.s .
gray in terms of combinatorial operations on those objects .
the elements in this set are based on the notion of rooted circle trees on which a connected graded bialgebra is defined in terms of extractions of rooted circle subtrees .
the latter can appear nested , disjoint or overlapping , which is remotely similar to the situation found in hopf algebras of feynman graphs @xcite .
motivated by this analogy , a zimmermann type forest formula on rooted circle trees is given , which provides a solution of the standard recursions for the antipode .
as a result , the antipode can be calculated in a way which is free of cancellations .
this is of interest with respect to applications in control theory such as system inversion and output tracking , where the antipode needs to be computed explicitly to high order .
thus , combinatorial efficient methods are needed for its computation .
the paper is organized as follows . in section [
sect : prelim ] a few basics on hopf algebras are collected mainly to fix the notation .
section [ sect : rcthopfalg ] describes the hopf algebra of decorated rooted circle trees .
a cancellation free zimmermann type forest formula is given for the antipode of this hopf algebra .
the pre - lie algebra structure on rooted circle trees is presented in subsection [ ssect : prct ] . in the last section , the link to the multivariable extension of the fa di bruno type hopf algebra of l. duffaut espinosa and w.s .
gray is described .
the second author was supported by ramn y cajal research grant ryc-2010 - 06995 from the spanish government .
the third author was supported by grant sev-2011 - 0087 from the severo ochoa excellence program at the instituto de ciencias matemticas in madrid , spain .
this research was also supported by a grant from the bbva foundation .
in this section , a few basics on hopf algebras are collected to fix the notation . for details ,
in particular on combinatorial hopf algebras , the reader is referred to @xcite , as well as the standard reference @xcite . in general
, @xmath1 denotes the ground field of characteristic zero over which all algebraic structures are defined .
a _ coalgebra _ over @xmath1 consist of a triple @xmath2 .
the coproduct @xmath3 is coassociative , that is , @xmath4 , and @xmath5 denotes the counit map . for a cocommutative coalgebra @xmath6 , where the flip map @xmath7 is defined on @xmath8 by @xmath9 .
a _ bialgebra _
@xmath10 is both a unital algebra and a counital coalgebra together with compatibility relations , such as both the algebra product , @xmath11 , and unit map , @xmath12 , are coalgebra morphisms @xcite .
the unit of @xmath10 is denoted by @xmath13 .
a bialgebra is called _ graded _ if there are @xmath1-vector subspaces @xmath14 , @xmath15 such that @xmath16 , @xmath17 and @xmath18 elements @xmath19 are given a degree @xmath20 .
moreover , @xmath10 is called _ connected _ if @xmath21 .
define @xmath22 .
for any @xmath19 the coproduct is of the form @xmath23 where @xmath24 is the _ reduced _ coproduct . by definition
an element @xmath25 is _ primitive _ if @xmath26 suppose @xmath27 is a @xmath1-algebra with product @xmath28 and unit @xmath29 , e.g. , @xmath30 or @xmath31 .
the vector space @xmath32 of linear maps from the bialgebra @xmath10 to @xmath27 together with the convolution product @xmath33 , where @xmath34 , is an associative algebra with unit @xmath35 . a _ hopf algebra _
@xmath36 is a bialgebra together with a particular @xmath1-linear map called an _
@xmath37 which satisfies the hopf algebra axioms @xcite .
when @xmath38 , the antipode @xmath39 is characterized as the inverse of the identity map with respect to the convolution product , that is , @xmath40 it is well - known that any connected graded bialgebra is a _ connected graded hopf algebra_. let @xmath41 be a connected graded hopf algebra .
suppose @xmath27 is a commutative unital algebra .
the subset @xmath42 of linear maps @xmath43 that send the unit to zero , @xmath44 , forms a lie algebra in @xmath45 .
the exponential @xmath46 is well defined and gives a bijection from @xmath47 onto the group @xmath48 of linear maps , @xmath49 , that send the unit of @xmath36 to the algebra unit in @xmath27 , @xmath50 .
an _ infinitesimal character _ with values in @xmath27 is a linear map @xmath51 such that for @xmath52 , @xmath53 .
the linear space of infinitesimal characters is denoted @xmath54 .
an @xmath27-valued map @xmath55 in @xmath45 is called a _ character _ if @xmath56 and for @xmath57 , @xmath58 .
the set of characters is denoted by @xmath59 .
it forms a pro - unipotent group for the convolution product with ( pro - nilpotent ) lie algebra @xmath60 @xcite . the exponential map @xmath61 restricts to a bijection between @xmath60 and @xmath62 . the neutral element @xmath63 in @xmath62
is given by @xmath64 and @xmath65 for @xmath66 .
the inverse of @xmath67 is given by composition with the hopf algebra antipode @xmath68 , i.e. , @xmath69 .
the presentation begins with a description of a class of rooted circle trees . the motivation for this setup will become clear once the theory of fliess operators in nonlinear feedback control theory and the related fa di bruno type hopf algebra
are presented in the next section .
[ def : rct ] a _ rooted circle tree _ ( rct ) @xmath70 consists of a set of vertices , @xmath71 , one of which is distinguished as the root indexed by @xmath72 , where @xmath73 , @xmath74 .
the vertices are strictly ordered and define a clockwise oriented closed edge @xmath75 .
the _ weight _ of @xmath70 is defined by its number of vertices , i.e. , @xmath76 .
the empty rct is denoted by @xmath77 , and @xmath78 .
the set of internal vertices of @xmath70 is denoted @xmath79 and is defined by excluding the root vertex from @xmath71 .
a rooted circle tree with @xmath80 vertices may be seen as an @xmath80-polygone , with edges oriented clockwise starting from the root , such that each vertex has exactly one incoming and one outgoing edge .
the last outgoing edge is the incoming edge of the root vertex . here
are the first few rooted circle trees of weights one up to five : @xmath81 the root is denoted by a black square to distinguish it clearly from the internal vertices .
a _ decorated rooted circle tree _ is a rct @xmath70 together with a mapping @xmath82 from the set of its internal vertices @xmath79 to the alphabet @xmath83 represented by @xmath84 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 45:0.4 cm ) circle(1.7pt ) ; \draw ( 30:0.73 cm ) node[above=-9pt]{${\hspace{-0.2cm}\phantom{l}^{l_1}}$ } ; \filldraw[black ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( -8:0.73 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{l_2}$ } ; \filldraw[black ] ( -45:0.4 cm ) circle(1.7pt ) ; \draw ( -46:0.74 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{l_3}$ } ; \filldraw[black ] ( -105:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -130:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -154:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -210:0.4 cm ) circle(1.7pt ) ; \draw ( -215:0.76 cm ) node[above=-9pt]{$\hspace{0.15cm}\phantom{l}^{l_k}$ } ; \end{tikzpicture}}\ ] ] the set ( resp .
the linear span ) of rooted circle trees will be denoted by @xmath85 ( resp .
@xmath86 ) . the set ( resp .
the linear span ) of rooted circle trees decorated by an alphabet @xmath87 will be denoted by @xmath88 ( resp .
@xmath89 ) . consider a particular grading on the set @xmath87 : @xmath90 and @xmath91 for all @xmath92 .
the _ degree _ of a decorated rooted circle tree @xmath93 of weight @xmath94 is defined by @xmath95 clearly this grading on @xmath87 distinguishes internal vertices of a rct decorated by @xmath96 from the rest .
they will play an important role in the hopf algebraic structure to be defined further below . to this end , it is natural to suppress the decoration by letters @xmath97 in favor of using colored vertices @xmath98 ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] and to consider bicolored rcts , i.e. , rcts with inner vertices being colored either black or white .
the white vertices correspond to a decoration by the letter @xmath96 , whereas the black vertices represent decorations by any of the letters @xmath97 , @xmath99 .
note that the decoration of the root is fixed to be @xmath72 , @xmath74 . here
are the first few bicolored rcts of degree one up to five : @xmath100 @xmath101 it will also be useful to have the white vertices of a rct @xmath70 enumerated consecutively in increasing order . for instance , the @xmath102 white vertices of the rct below are clockwise labeled from @xmath103 to @xmath102 @xmath84 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 45:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 25:0.33 cm ) circle(0.2pt ) ; \filldraw[fill = white , draw = black , thick ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( -8:0.6 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{1}$ } ; \filldraw[black ] ( -25:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -45:0.33 cm ) circle(0.2pt ) ; \filldraw[fill = white , draw = black , thick ] ( -75:0.4 cm ) circle(1.7pt ) ; \draw ( -70:0.6 cm ) node[above=-9pt]{$\hspace{-0.cm}\phantom{l}^{2}$ } ; \filldraw[black ] ( -105:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -115:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -125:0.33 cm ) circle(0.3pt ) ; \filldraw[fill = white , draw = black , thick ] ( -150:0.4 cm ) circle(1.7pt ) ; \draw ( -150:0.6 cm ) node[above=-9pt]{${\hspace{-0.1cm}\phantom{l}^s}$ } ; \filldraw[black ] ( -180:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -190:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -200:0.33 cm ) circle(0.3pt ) ; \end{tikzpicture}}\ ] ] denote by @xmath104 ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture}}$ ] and @xmath105 ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture}}$ ] the number of black and white vertices of @xmath70 , respectively .
the weight of a bicolored rct @xmath70 is defined as before by the number of its internal vertices plus one .
the degree of a bicolored rct @xmath70 is defined by : @xmath106 ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture } } + |c^i|_{\begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture } } + 1.\ ] ] the set ( resp . the linear span ) of bicolored rooted circle trees will be denoted by @xmath107 ( resp .
@xmath108 ) .
the use of bicolored rcts with white vertices enumerated consecutively significantly simplifies the description of the combinatorial structure to be defined next on decorated rooted circle trees .
the set of internal vertices @xmath109 of a rct @xmath70 is naturally ordered with the minimal element being the first vertex to the right of the root , and the maximal one being the vertex to the left of the root .
therefore , the elements in any subset @xmath110 can be written in strictly increasing order .
for instance , all subsets of the rct @xmath111 are obviously admissible .
the only admissible subset of the rct @xmath112 is the singleton @xmath113 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] . for the rct @xmath114
there are two admissible subsets , @xmath115 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] and @xmath116 ( 0,0 ) circle(2pt ) ; \end{tikzpicture},\begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] .
the rct @xmath117 has six admissible subsets : @xmath118 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{21}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{12}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{13}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{22}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{14}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}.\ ] ] a remark is in order regarding notation of the admissible subsets @xmath119 .
the first index corresponds to the label of the white vertex being the minimal element in @xmath119 .
the second index serves to distinguish among the possible admissible subsets that have the @xmath1-th white vertex as minimal element .
the rct @xmath120 has four admissible subsets all of which share the same white vertex as minimal element : @xmath118 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{12}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{13}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{14}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}.\ ] ] the rct @xmath121 has five admissible subsets : @xmath118 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{21}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{12}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{13}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{14}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}.\ ] ] [ bckhopf ] the notion of admissible subsets of rcts is remotely analogous to the notion of admissible cuts on non - planar rooted trees in the context of the butcher connes kreimer hopf algebra . see @xcite for reviews .
[ def : admextract ] for any rct @xmath122 with @xmath123 white vertices , define the set of all admissible subsets @xmath124 .
a subset @xmath125 is an _ admissible extraction _ if it consists of pairwise disjoint admissible subsets , i.e. , @xmath126 implies @xmath127 if @xmath128 .
the set of all admissible extractions of @xmath70 , including the empty set as well as all vertices , @xmath71 , is denoted @xmath129 .
the set of _ proper admissible extractions _ of @xmath70 , @xmath130 , excludes both the empty set and @xmath71 . for each admissible subset , @xmath131 , in an admissible extraction @xmath132 , one can associate an oriented polygon inside @xmath70 by connecting the vertices of @xmath131 by oriented edges according to the total order implied by @xmath70 inside each admissible subset @xmath131 .
for instance , returning to the previous example @xmath121 the following are all possible admissible extractions : @xmath133 @xmath134 pictorially these admissible extractions can be represented , respectively , by oriented polygons inscribed into the rct @xmath70 : @xmath135 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -50:0.4 cm ) arc ( -45:360:0.1 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture } \qquad\ \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -200:0.4 cm ) arc ( -200:180:0.1 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture } \qquad\ \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-40 ] ( -130:0.4 cm ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-5 ] ( -130:0.4 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture } \qquad\ \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -200:0.4 cm ) to [ bend left=-15 ] ( -50:0.4 cm ) ; \draw[thick = black ] ( -200:0.4 cm ) to [ bend left=15 ] ( -50:0.4 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture } \qquad\ \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-15 ] ( -130:0.4 cm ) ; \draw[thick = black ] ( -200:0.4 cm ) to [ bend left=15 ] ( -50:0.4 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -130:0.4 cm ) to [ bend left=-15 ] ( -200:0.4 cm ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture } \qquad\ \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -50:0.4 cm ) arc ( -45:360:0.1 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -200:0.4 cm ) arc ( -200:180:0.1 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture } \qquad\ \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-40 ] ( -130:0.4 cm ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-5 ] ( -130:0.4 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \draw[thick = black ] ( -200:0.4 cm ) arc ( -200:180:0.1 cm ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture}\ ] ] note that each polygon starts and ends at a particular white vertex , i.e. , the minimal element in the corresponding admissible subset .
let @xmath136 be an admissible extraction .
observe that the order of the white vertices implies an order in @xmath137 according to the order of the minimal elements @xmath138 .
the indices indicate the corresponding minimal element of the admissible subsets , that is , @xmath139 ( 0,0 ) circle(1pt ) ; \end{tikzpicture}}$ ] .
let @xmath140 be an admissible subset of the rct @xmath70 with the @xmath141-th white vertex of @xmath70 as minimal element .
define @xmath142 to be the rct with @xmath143 , that is , the rct formed by eliminating in @xmath71 the vertices corresponding to the admissible subset @xmath140 .
the rct @xmath144 is defined by replacing in @xmath70 the @xmath141-th white vertex corresponding to the minimal element of @xmath140 by a black vertex ( i.e. a vertex decorated by @xmath145 ) and eliminating in @xmath79 the vertices corresponding to the remaining elements in @xmath140 .
the rct @xmath146 is defined in terms of the vertices of @xmath140 by making the minimal element in @xmath140 the new root labeled @xmath147 , and the remaining elements of the admissible subset @xmath140 are the internal vertices ordered according to the order in @xmath140 .
note that @xmath148 and @xmath149 .
the rct @xmath146 may be considered a rooted circle subtree ( sub - rct ) of @xmath70 .
hence , to any admissible set @xmath140 there corresponds a unique pair of rcts @xmath150 .
this somewhat unwieldy notation becomes much clearer when working directly with the diagrammatical notation of rcts .
hence , continuing the example above , the admissible subset @xmath116 ( 0,0 ) circle(2pt ) ; \end{tikzpicture},\begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] @xmath151 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-40 ] ( -130:0.4 cm ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-5 ] ( -130:0.4 cm ) ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture}\ ] ] gives the rcts @xmath152 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 0:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -90:0.4 cm ) circle(1.7pt ) ; \draw(-60:0.45 cm ) node[above=-14pt]{$\hspace{-0.2cm}\phantom{l}^{n_1}$ } ; \filldraw[fill = white , draw = black , thick ] ( -180:0.4 cm ) circle(1.7pt ) ; } \end{tikzpicture } } \qquad\ c_{|j_{12}}^{n_1}= \raisebox{-10pt } { \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 65:0.63 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{n_1}$ } ; \filldraw[black ] ( -90:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture}}\ ] ] in the following , with the aim to simplify notation , an admissible subset of a rct is indexed only by its minimal element .
if several admissible subsets with the same minimal element are considered then again pairs of indices are used , where , as before , the first index indicates the white vertex and the second index labels the particular admissible subset . 1
. any admissible subset @xmath153 of a rct @xmath70 will be associated to a sub - rct of @xmath70 by identifying its minimal element as the root .
see for an example .
the basic assumption is that two admissible subsets shall not have the same minimal element , i.e. , roots .
apart from this , the following cases can occur when comparing two admissible subsets @xmath154 and @xmath153 , @xmath155 . if @xmath156 then the two sub - rcts are either strictly nested , @xmath157 , disjoint @xmath158 , or they overlap , i.e. , @xmath159
the first case is illustrated in the rct @xmath151 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \draw[thick = black ] ( 0:0.4 cm ) to [ bend left=-40 ] ( -40:0.4 cm ) ; \draw[thick = black ] ( -40:0.4 cm ) to [ bend left=-20 ] ( -90:0.4 cm ) ; \draw[thick = black ] ( -90:0.4 cm ) to [ bend left=-3 ] ( -180:0.4 cm ) ; \draw[thick = black ] ( -180:0.4 cm ) to [ bend left=20 ] ( -0:0.4 cm ) ; \draw[thick = black ] ( -40:0.4 cm ) to [ bend left=-20 ] ( -180:0.4 cm ) ; \draw[thick = black ] ( -40:0.4 cm ) to [ bend left=15 ] ( -180:0.4 cm ) ; \filldraw[black ] ( 45:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( 0:0.6 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{1}$ } ; \filldraw[fill = white , draw = black , thick ] ( -40:0.4 cm ) circle(1.7pt ) ; \draw ( -40:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{2}$ } ; \filldraw[fill = white , draw = black , thick ] ( -90:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -140:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -180:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -225:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture}\ ] ] where the admissible subset @xmath160 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2 , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] with the second white vertex as minimal element is nested inside the admissible subset @xmath161 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2 , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] . the following rct @xmath151 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \draw[thick = black ] ( 0:0.4 cm ) to [ bend left=-40 ] ( -90:0.4 cm ) ; \draw[thick = black ] ( -90:0.4 cm ) to [ bend left=10 ] ( -225:0.4 cm ) ; \draw[thick = black ] ( -225:0.4 cm ) to [ bend left=-3 ] ( 0:0.4 cm ) ; \draw[thick = black ] ( -0:0.4 cm ) to [ bend left=20 ] ( -0:0.4 cm ) ; \draw[thick = black ] ( -40:0.4 cm ) to [ bend left=-20 ] ( -180:0.4 cm ) ; \draw[thick = black ] ( -40:0.4 cm ) to [ bend left=15 ] ( -180:0.4 cm ) ; \filldraw[black ] ( 45:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( 0:0.6 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{1}$ } ; \filldraw[fill = white , draw = black , thick ] ( -40:0.4 cm ) circle(1.7pt ) ; \draw ( -40:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{2}$ } ; \filldraw[fill = white , draw = black , thick ] ( -90:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -140:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -180:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -225:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture}\ ] ] illustrates the case of disjoint admissible subsets @xmath161 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] and @xmath160 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2 , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] . the _ overlapping _ case , i.e. , @xmath159 , is represented by the example @xmath151 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ;
\draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \draw[thick = black ] ( 0:0.4 cm ) to [ bend left=-40 ] ( -90:0.4 cm ) ; \draw[thick = black ] ( -90:0.4 cm ) to [ bend left=1 ] ( -180:0.4 cm ) ; \draw[thick = black ] ( -180:0.4 cm ) to [ bend left=-100 ] ( -225:0.4 cm ) ; \draw[thick = black ] ( -225:0.4 cm ) to [ bend left=-3 ] ( 0:0.4 cm ) ; \draw[thick = black ] ( -0:0.4 cm ) to [ bend left=20 ] ( -0:0.4 cm ) ; \draw[thick = black ] ( -40:0.4 cm ) to [ bend left=-20 ] ( -180:0.4 cm ) ; \draw[thick = black ] ( -40:0.4 cm ) to [ bend left=15 ] ( -180:0.4 cm ) ; \filldraw[black ] ( 45:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( 0:0.6 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{1}$ } ; \filldraw[fill = white , draw = black , thick ] ( -40:0.4 cm ) circle(1.7pt ) ; \draw ( -40:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{2}$ } ; \filldraw[fill = white , draw = black , thick ] ( -90:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -140:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -180:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -225:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture}\ ] ] where @xmath161 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] and @xmath160 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2 , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\}$ ] share one black vertex .
[ rmk:2 ] the combinatorics described here are similar in flavor to those of feynman graphs in physics .
those graphs can have subgraphs , and they can be disjoint , nested or overlapping .
feynman graphs appear in perturbative quantum field theory , and their combinatorics are crucial in the context of the solution of the so - called renormalization problem @xcite . the analog of an admissible extraction is called a _ spinney _ in the context of feynman graphs .
the analog of the set of all admissible extractions is called a _
wood _ , which contains all spinneys @xcite .
observe that for any admissible subset @xmath162 the weight @xmath163 .
regarding the degrees of @xmath164 and @xmath165 , it follows that @xmath166 ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture } } - |j_l|_{\begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture } } + 1 $ ] , where @xmath167 ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture}}$ ] and @xmath168 ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture}}$ ] denote the number of white and black vertices in @xmath162 , respectively . for the extracted rct @xmath169 ( 0,0 ) circle(1.3pt ) ; \end{tikzpicture } } + |j_{l}|_{\begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(1.3pt);\end{tikzpicture } } -1\ ] ] such that @xmath170 consider next the rct @xmath165 corresponding to the admissible subset @xmath162 as a sub - rct in @xmath70 .
since an admissible extraction @xmath132 consists of strictly disjoint admissible subsets , one may extract several sub - rcts from a rct @xmath70 .
hence , to any admissible extraction @xmath171 , @xmath172 , there corresponds a unique rct @xmath173 with @xmath1 white vertices replaced by black vertices decorated by @xmath174 and a sequence of sub - rcts @xmath175 with roots indexed by @xmath176 .
returning to the earlier example , the admissible extraction @xmath177 corresponding to the rct with two sub - rcts @xmath151 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \draw[thick = black ] ( -200:0.4 cm ) arc ( -200:180:0.1 cm ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-40 ] ( -130:0.4 cm ) ; \draw[thick = black ] ( -50:0.4 cm ) to [ bend left=-5 ] ( -130:0.4 cm ) ; \filldraw[black ] ( 25:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -50:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -130:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -200:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture}\ ] ] gives @xmath178 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 0:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -90:0.4 cm ) circle(1.7pt ) ; \draw(-60:0.5 cm ) node[above=-14pt]{$\hspace{-0.2cm}\phantom{l}^{n_1}$ } ; \filldraw[black ] ( -180:0.4 cm ) circle(1.7pt ) ; \draw(-198:0.8 cm ) node[above=-14pt]{$\hspace{0.1cm}\phantom{l}^{n_2}$ } ; } \end{tikzpicture } } } \ ] ] and @xmath179 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.0cm}\phantom{l}^{n_1}$ } ; \filldraw[black ] ( -90:0.4 cm ) circle(1.7pt ) ; \end{tikzpicture } } \qquad c_{|j_{21}}^{n_2}=\raisebox{-10pt } { \begin{tikzpicture } { \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.0cm}\phantom{l}^{n_2}$ } ; } \end{tikzpicture}}\ ] ] the linear span of bicolored rooted circle trees @xmath108 is extended to a commutative @xmath1-algebra @xmath180 by using the disjoint union as a product , i.e. , @xmath181 .
this makes @xmath180 a polynomial algebra with @xmath107 as a generating set .
a coproduct is defined on @xmath182 in terms of admissible extractions .
let @xmath70 be a bicolored rct , the map @xmath183 is defined by summing over all admissible extractions on @xmath70 , namely , @xmath184 the coproduct is extended multiplicatively to @xmath182 .
the sum on the right - hand side is over all admissible extractions of the rct @xmath70 , including the empty set and the full rct .
note that it includes @xmath1 _ internal _ summations , i.e. , @xmath185 .
in other words , _ einstein s summation convention _ applies so that pairs of indices are implying a summation from @xmath103 to @xmath11 . on the left - hand side of the tensor product
there is always a single bicolored rct with root indexed by @xmath72 , where several internal vertices have been eliminated , and @xmath1 internal black vertices are indexed by @xmath186 .
the latter correspond to the minimal elements of the admissible subsets @xmath187 , i.e. , the white vertices enumerated @xmath188 that have been turned into black vertices . on the right - hand side of the tensor product
there is a monomial of @xmath1 sub - rcts with roots indexed by @xmath186 .
examples are in order to make this definition more transparent .
first , note that the rct with no internal vertices is primitive , since there are no proper admissible subsets .
that is , @xmath189 similarly , the rcts with only black internal vertices also has no proper admissible subsets , therefore , @xmath190 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 45:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 25:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 0:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -25:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -45:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -75:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -105:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -115:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -125:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -150:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -180:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -190:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -200:0.33 cm ) circle(0.3pt ) ; \end{tikzpicture}}\big ) \ ; = \ ; \raisebox{-10pt } { \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 45:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 25:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 0:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -25:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -45:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -75:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -105:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -115:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -125:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -150:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -180:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -190:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -200:0.33 cm ) circle(0.3pt ) ; \end{tikzpicture } } \ ; \otimes \ ; { \mathbf{1 } } \ ; + \ ; { \mathbf{1 } } \ ; \otimes \raisebox{-10pt } { \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[black ] ( 45:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 25:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( 0:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -25:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -35:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -45:0.33 cm ) circle(0.2pt ) ; \filldraw[black ] ( -75:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -105:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -115:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -125:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -150:0.4 cm ) circle(1.7pt ) ; \filldraw[black ] ( -180:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -190:0.33 cm ) circle(0.3pt ) ; \filldraw[black ] ( -200:0.33 cm ) circle(0.3pt ) ; \end{tikzpicture}}\ ] ] the grading of @xmath197 is given in terms of the degree of rcts , namely , @xmath198 . from
it follows that the coproduct is compatible with the grading , i.e. , @xmath199 .
next coassociativity is shown , i.e. , @xmath200 .
some notation is introduced to make the argument more transparent . suppressing the internal summations ,
the coproduct is written as @xmath201 here @xmath202 denotes the set of completely disjoint sub - rcts corresponding to an admissible extraction @xmath203 , and @xmath204 is the sub - rct corresponding to the admissible subset @xmath205 of @xmath70 .
the sum is taken over the set of sub - rct corresponding to @xmath206 . extracting the set of sub - rcts @xmath207
is denoted by @xmath208 .
this notation follows the one used in the context of hopf algebras of rooted trees or feynman graphs @xcite .
any sub - rct of @xmath70 may have itself sub - rcts .
indeed , for @xmath209 and admissible subsets @xmath210 , the two sub - rcts @xmath211 and @xmath212 of a rct @xmath70 for which @xmath211 is a sub - rct of @xmath213 yield @xmath214 in this notation , the identity to be verified is the equality between @xmath215 and @xmath216 hence , what needs to be shown is that for each @xmath217 @xmath218 the argument given here follows @xcite .
let @xmath217 such that @xmath219
. then @xmath220 . on the other hand , for @xmath221
there corresponds a @xmath222 such that @xmath223 and @xmath224 .
equality follows from ( [ nesting ] ) .
the reduced coproduct @xmath226 is defined as @xmath227 recall that @xmath130 is the set of proper admissible extractions , i.e. , the empty admissible set @xmath228 as well as the total admissible subset @xmath229 are excluded .
@xmath182 is a _ connected graded bialgebra _ , and therefore a hopf algebra .
the antipode @xmath230 is given in terms of the following recursive formulas starting with @xmath231 and a few examples are in order : @xmath232 @xmath233 @xmath234 in light of the last calculation and the cancellation appearing in it when using , one may wonder whether a cancellation free formula is available .
the answer is yes , and it is known as zimmermann s forest formula as discussed next @xcite .
zimmermann s celebrated forest formula first appeared in the context of the renormalization problem in perturbative quantum field theory @xcite .
it provides a closed form solution to bogoliubov s so - called counterterm recursion for feynman graphs @xcite .
the work of connes
kreimer unveiled the hopf algebraic structure underlying the theory of renormalization @xcite . as a result ,
a forest type formula , i.e. , a closed form solution for the antipode , can be given in the context of the butcher connes kreimer hopf algebra of rooted trees .
it is defined by extending the notion of admissible cuts on rooted trees to all cuts .
see @xcite for details . from a combinatorial point of view forest type formulas
are interesting since they provide a way to calculate antipodes which is free of cancellations @xcite . in control
theory applications they have the potential to significantly reduce computation times @xcite . in this section ,
a forest type formula for the antipode of the hopf algebra of rooted circle trees is presented .
it also provides a closed form solution for the recursions ( [ antipod2 ] ) , which is moreover free of cancellations .
it is defined by replacing admissible extractions by all extractions .
however , the situation is somewhat involved due to the internal summations over the decorations @xmath235 associated with extracting admissible subsets from a rct in the coproduct @xmath236 recall that in the definition of an admissible extraction @xmath237 of a rct @xmath70 all admissible subsets of @xmath132 have to be strictly disjoint .
this constraint is now relaxed by allowing for extractions @xmath238 , @xmath239 , which may contain admissible subsets that are either nested , @xmath240 for @xmath241 , or disjoint , @xmath242 for @xmath243 .
the set of _ all _ extractions of a rct @xmath70 is denoted @xmath244 .
note that @xmath245 . as an example , consider the rct @xmath151 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[fill = white , draw = black , thick ] ( -45:0.4 cm ) circle(1.7pt ) ; \draw ( -45:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{1}$ } ; \filldraw[fill = white , draw = black , thick ] ( -135:0.4 cm ) circle(1.7pt ) ; \draw ( -135:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{2}$ } ; \end{tikzpicture}\ ] ] and its admissible subsets : @xmath118 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1\ } , j_{12}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2\ } , j_{21}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2\}.\ ] ] then @xmath244 contains , among others , the extractions @xmath246 with @xmath247 .
another example is the rct @xmath151 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[fill = white , draw = black , thick ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( 0:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{1}$ } ; \filldraw[black ] ( -90:0.4 cm ) circle(1.7pt ) ; \filldraw[fill = white , draw = black , thick ] ( -180:0.4 cm ) circle(1.7pt ) ; \draw ( -180:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{2}$ } ; \end{tikzpicture}\ ] ] and its admissible subsets : @xmath118 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1\ } , j_{12}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}\ } , j_{13}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2\ } , j_{14}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[black ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture } , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2\ } , j_{21}=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2\}.\ ] ] then @xmath244 contains , among others , the extractions : @xmath248 which are nested , i.e. , @xmath249 and @xmath250 . in the rct @xmath251 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \filldraw[fill = white , draw = black , thick ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( 0:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{1}$ } ; \filldraw[fill = white , draw = black , thick ] ( -90:0.4 cm ) circle(1.7pt ) ; \draw ( -90:0.65 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{2}$ } ; \filldraw[fill = white , draw = black , thick ] ( -180:0.4 cm ) circle(1.7pt ) ; \draw ( -180:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{3}$ } ; \end{tikzpicture}\ ] ] one finds , among others , the extractions : for notational simplicity the second indices of the admissible subsets have been omitted .
elements in an extraction @xmath252 are either disjoint or strictly nested .
this implies a natural poset structure on each extraction , which is represented in terms of a non - planar rooted tree .
the tree reflects the resulting hierarchy of admissible subsets of a rct @xmath70 . recall that by definition a non - planar _ rooted tree _
@xmath253 is made out of vertices and nonintersecting oriented edges such that all but one vertex have exactly one outgoing line and an arbitrary number of incoming lines .
the particular vertex with no outgoing line is called the root of the tree .
it is drawn on top of the tree , whereas the leaves are the only vertices without any incoming lines .
the weight of a rooted tree @xmath253 is defined to be its number of vertices , @xmath254 .
the following list contains rooted trees up to weight four : @xmath255 ( 0,0 ) circle(2.5pt ) ; \end{tikzpicture } \qquad \arbrea \qquad \arbreb
\qquad \arbrec \qquad \arbree \qquad \arbref \qquad \arbred \qquad \arbreg\ ] ] the root has been colored green . the empty rooted tree is denoted by @xmath77 . the set ( resp .
space ) of rooted trees is denoted by @xmath256 ( resp .
@xmath257 ) .
( see @xcite for details . )
recall the map @xmath258 defined on rooted trees .
it maps rooted trees @xmath259 to the rooted tree @xmath260 by connecting each of the roots of @xmath259 via an edge to a new root .
the tree with only one vertex , i.e. , the root , and no edges is written @xmath261 ( 0,0 ) circle(2.5pt ) ; \end{tikzpicture } = b_+(\mathbf{1})$ ] . the rest of the above trees
can then be written @xmath262 , @xmath263 , @xmath264 , etc .
next , rooted trees are decorated in such a way that they represent the hierarchy of admissible subsets in an extraction of any rct .
this hierarchy reflects the internal summations in the coproduct on rcts .
let @xmath70 be a rct and @xmath265 an extraction .
define a map @xmath266 by decorating the root of @xmath267 by @xmath72 .
an edge in @xmath253 goes out from a vertex decorated by @xmath268 to a vertex decorated by @xmath269 if and only if @xmath240 .
regarding example ( [ exp : rct ] ) , the following rooted trees are associated with the extractions @xmath270 , @xmath271 , @xmath272 , @xmath273 respectively .
again , with @xmath274 ( 0,0 ) circle(2.5pt ) ; \end{tikzpicture}^i$ ] denoting the root with no edges and decorated by @xmath72 , any decorated rooted tree can be written using the map @xmath275 .
it adds a new root decorated by @xmath80 .
the previous list of decorated rooted trees then becomes @xmath276 , @xmath277 , @xmath278 , and @xmath279 .
note that rooted trees corresponding to admissible extractions @xmath280 are strictly of the form @xmath281 . using extractions ,
together with decorated rooted trees to encode the hierarchical structure of extractions of a rct , the antipode @xmath68 of @xmath282 is given in the following theorem .
the decorated rooted tree corresponding to an extraction @xmath283 is denoted @xmath284 .
[ thm : forestformula ] for any @xmath285 the antipode @xmath286 where @xmath287 is the decorated rooted tree corresponding to the hierarchy of admissible subsets underlying the extraction @xmath288 .
the set @xmath289 must not include @xmath290 .
the map @xmath291 is defined on the decorated rooted tree @xmath292 as @xmath293 and @xmath294 . here
@xmath295 and @xmath296 are pairwise disjoint .
the admissible subsets @xmath290 and @xmath297 correspond to @xmath298 and @xmath299 , respectively . note that the empty extraction @xmath300 corresponds to the rooted tree @xmath301 ( 0,0 ) circle(2.5pt ) ; \end{tikzpicture}^i$ ] .
hence , for a primitive rct @xmath70 @xmath302 as another example , consider the rooted tree @xmath303 corresponding to the extraction @xmath304 ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_1 , \begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_3\ } , j_2=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_2\ } , j_3=\{\begin{tikzpicture } \filldraw[fill = white , draw = black , thick ] ( 0,0 ) circle(2pt ) ; \end{tikzpicture}_3\}\}$ ] in example ( [ exp : rct ] ) .
then again , this gets much easier by working diagrammatically .
indeed , observe that @xmath305 corresponds to @xmath306 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.2cm}\phantom{l}^{i}$ } ; \draw[thick = black ] ( 0:0.4 cm ) to [ bend left=-44 ] ( -180:0.4 cm ) ; \draw[thick = black ] ( 0:0.4 cm ) to [ bend left=44 ] ( -180:0.4 cm ) ; \draw[thick = black ] ( -180:0.4 cm ) arc ( -180:360:0.08 cm ) ; \draw[thick = black ] ( -90:0.4 cm ) arc ( -90:290:0.08 cm ) ; \filldraw[fill = white , draw = black , thick ] ( 0:0.4 cm ) circle(1.7pt ) ; \draw ( 0:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{1}$ } ; \filldraw[fill = white , draw = black , thick ] ( -90:0.4 cm ) circle(1.7pt ) ; \draw ( -90:0.65 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{2}$ } ; \filldraw[fill = white , draw = black , thick ] ( -180:0.4 cm ) circle(1.7pt ) ; \draw ( -180:0.6 cm ) node[above=-9pt]{$\hspace{-0.1cm}\phantom{l}^{3}$ } ; \end{tikzpicture } } \;\ ; \qquad\;\ ; \raisebox{-20pt}{\arbreei } = b_+^i(b_+^1(b_+^3(\mathbf{1})),b_+^2(\mathbf{1}))\ ] ] such that @xmath307 ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.0cm}\phantom{l}^{n_1}$ } ; \filldraw[black ] ( -90:0.4 cm ) circle(1.7pt ) ; \draw ( 72:0.64 cm ) node[above=-42pt]{$\hspace{-0.1cm}\phantom{l}^{n_3}$ } ; \end{tikzpicture } } \raisebox{-9pt } { \begin{tikzpicture } \draw[thick ] ( 0cm,0 cm ) circle(0.4 cm ) ; \filldraw[black ] ( 90:0.4 cm ) ( 0.08,.48)rectangle(-.08,.34 ) ; \draw ( 75:0.60 cm ) node[above=-9pt]{$\hspace{-0.0cm}\phantom{l}^{n_3}$ } ; \end{tikzpicture } } \raisebox{-9pt}{\etk}\ ] ] note that the use of rooted trees solely reflects the need to keep track of the hierarchy of admissible subsets in an extraction , which is reflected in the extra summations in the coproduct ( [ cop ] ) .
the antipode of the rct consists of 26 terms corresponding to all possible extractions , including those mentioned above . for any @xmath285 define @xmath308 where @xmath309
the goal is to show that the map @xmath310 satisfies the recursion @xmath311 the uniqueness of the antipode then implies that @xmath312 .
recall that @xmath130 is the set of proper admissible extractions of the rct @xmath70 , i.e. , it contains neither @xmath313 nor @xmath314 .
in addition , recall that the admissible extractions are a subset of all extractions , i.e. , @xmath315 .
this allows one to break the sum ( [ sz ] ) into pieces such that @xmath316 the first term on the right - hand side corresponds to the admissible subset @xmath297 .
the first sum on the right - hand side is over strictly pairwise disjoint subsets @xmath280 , that is , proper admissible extractions .
recall that those extractions correspond to rooted trees of the form @xmath317 .
therefore , @xmath318 the second sum on the right - hand side of ( [ splitting ] ) is over extractions @xmath319 , which contain at least one pair of admissible subsets , @xmath320 , @xmath321 , such that either @xmath322 or @xmath323 .
assume that @xmath324 is an admissible subset of the rct @xmath70 with only one white vertex , i.e. , its minimal element .
this implies that the corresponding rct @xmath325 is primitive .
therefore , the rooted tree encoding the nesting is simply given by @xmath326 .
considering @xmath324 as an admissible extraction @xmath327 of @xmath70 yields a single term in @xmath310 @xmath328 which can be written @xmath329 since @xmath325 is primitive .
assume now that @xmath324 contains more than one white vertex , which implies that the rct @xmath325 corresponding to the admissible subset @xmath324 has itself non - trivial sub - rcts .
this means that there exist extractions @xmath330 such that @xmath331 for all @xmath332 , i.e. , @xmath333 , where @xmath334 are all proper extractions of @xmath325 .
this yields which can be written @xmath335 recall that einstein s convention is in place , which implies a summation over pairs of indices . for an arbitrary proper admissible extraction @xmath336 this generalizes to @xmath337 which then implies that @xmath311 calculations of the antipode using formula ( [ antipod1 ] ) show non - trivial cancellations of terms ( see table [ tbl : cancellations - left - antipode ] ) . on the other hand ,
the forest formula in theorem [ thm : forestformula ] is free of such cancellations .
indeed , this follows from the observation that each extraction @xmath338 corresponds to a monomial of @xmath1 rcts in formula ( [ antipode3 ] ) .
its sign depends solely on the number of elements in @xmath339 .
cancellations can only occur between monomials of the same number of rcts .
hence , the antipode formula is free of such cancellations .
( 0.5,4 )
( 3.5,4 ) ; ( 0,3.5 )
( 0,0.5 ) ; ( 4,3.5 )
( 4,0.5 ) ; ( 0.5,0 )
( 3.5,0 ) ; at ( 0,0 ) @xmath341,@xmath342 ; at ( 0,4 ) @xmath343,@xmath344 ; at ( 4,0 ) @xmath345 , @xmath346 ; at ( 4,4 ) @xmath347,@xmath348 ; at ( -0.9,2 ) @xmath349 ; at ( 4.9,2 ) @xmath349 ; at ( 2,-0.4 ) @xmath350 ; at ( 2,4.4 ) @xmath350 ; in this subsection , the various composition products appearing in the previous sections are examined in more depth to reveal their relationship to each other as well as to the convolution product on the dual of @xmath36 .
the situation is summarized in table [ tbl : composition - products ] .
the product @xmath351 is included here for completeness , but its role in the theory is relatively minor .
each composition induces a corresponding coproduct via identities like .
these coproduct are easily shown to be related by adding specific terms of the primitive @xmath352 as shown in figure [ fig : composition - products - commutative - diagram ] . with each @xmath353
, one can identify an element @xmath354 where by _ : a^i_a^i_()=(_i , ) , @xmath355 , and _
( a^i_a^j_)=a^i _ ( ) a^j_()=_(a^i_)_(a^j _ ) . given that @xmath356 , there is clearly a canonical embedding of @xmath357 in @xmath358 so that characters of @xmath36 can be identified with series in @xmath358 .
in addition , @xmath359 so in some sense an identification can also be made between convolution of characters and composition of series .
in fact , this can be done for any of the four series compositions described above .
this is not completely surprising from a system theory point of view given the fact that the composition of two linear time - invariant systems with respective impulse responses @xmath360 and @xmath361 has the impulse response @xmath362 , where in this case ` @xmath363 ' denotes function convolution . in the analytic case ,
function convolution is equivalent to the composition of the corresponding generating series , i.e. , @xmath364 , which in turn reduces to a convolution sum over a one letter alphabet @xcite .
put another way , the hopf convolution product induced by @xmath365 can be viewed as a nonlinear generalization of this classic convolution theorem in linear system theory .
continuing with the above development , since the composition inverse of @xmath366 in theorem [ th : allseriesdeltam - is - group ] is @xmath367 with @xmath368 , it follows that @xmath369 finally , the infinitesimal characters associated with @xmath36 can be identified with elements @xmath370 , where @xmath371 for any @xmath372 and @xmath373 .
a. ferfera , combinatoire du monode libre applique la composition et aux variations de certaines fonctionnelles issues de la thorie des systmes , doctoral dissertation , university of bordeaux i , 1979 . ,
a fa di bruno hopf algebra for analytic nonlinear feedback control systems , in _ fa di bruno hopf algebras , dyson - schwinger equations , and lie - butcher series _ , k. ebrahimi - fard and f. fauvet , eds . , irma lect .
21 , eur .
strasbourg , france , 2015 , pp .
149217 .
w. s. gray , l. a. duffaut espinosa , and k. ebrahimi - fard , recursive algorithm for the antipode in the siso feedback product , _ proc .
21st inter .
. mathematical theory of networks and systems _ ,
groningen , the netherlands , 2014 , pp . 10881093 . , a short survey on pre - lie algebras , _ noncommutative geometry and physics : renormalisation , motives , index theory _ , e. schrdinger institut lectures in math . phys .
, a. carey , ed .
soc , wien , austria , 2011 . | in this work a combinatorial description is provided of a fa di bruno type hopf algebra which naturally appears in the context of fliess operators in nonlinear feedback control theory .
it is a connected graded commutative and non - cocommutative hopf algebra defined on rooted circle trees .
a cancellation free forest formula for its antipode is given . |
magnetic reconnection is a fundamental process of energy release that lies at the core of many dynamic phenomena in the solar system such as solar flares , coronal heating events , geomagnetic substorms and flux transfer events .
reconnection in three dimensions has been shown to be completely different in many fundamental respects from the classically studied process in two dimensions @xcite .
the main thrust of reconnection theory at present is to understand the different ways in which it may take place in three dimensions ( e.g. , the books * ? ? ?
* ; * ? ? ?
a key point is that in three dimensions reconnection occurs where a component of the electric field parallel to the magnetic field is present and this can be in many different field configurations .
for example , reconnection may occur either at null points ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) or in the absence of null points at quasi - separatrix layers or hyperbolic flux tubes @xcite or it may occur along separators that join one null point to another @xcite .
null points are common in the solar atmosphere @xcite and are sometimes implicated in solar flares and coronal mass ejections @xcite .
three - dimensional collapse of a null has been described @xcite and stationary resistive flows near them have been modelled @xcite . in particular , for a linear null and uniform magnetic diffusivity , @xcite discovered field - aligned flows when the spine current is small and spiral field - crossing flows which do not cross the spine or fan when the spine current exceeds a critical value . a three - dimensional null point possesses two different classes of field lines that connect to the null : for a so - called _ positive null point _ , a surface of field lines ( called a _ fan _ by @xcite ) recede from the null , while an isolated field line ( called the _ spine _ of the null ) approaches it from two directions ; for a _ negative null point _ ,
on the other hand the fan approaches the null , while the spine recedes from it .
( for an alternative nomenclature see ref .
the different types of linear null were categorised by @xcite .
the generic null in a potential magnetic field is an improper radial null , with the fan perpendicular to the spine and the field lines in the fan approaching or receding from essentially two directions ( fig .
[ fig:1]b ) .
a particular case is the proper radial null in which the field lines in the fan are radial ( fig .
[ fig:1]a ) .
the effect of a current along the fan is to make the fan and spine no longer perpendicular ( fig .
[ fig:3]b ) , whereas a strong enough current along the spine makes the fan field lines spiral ( fig .
[ fig:3]a ) .
there have been three steps towards categorising reconnection at a null point due to ( i ) analytical ideal modelling , ( ii ) kinematic resistive modelling and ( iii ) computational experiments . the initial analytical ideal treatment by @xcite aimed to understand the types of ideal motions that are possible in the environment of a null point .
they supposed that the nature of reconnection is determined to a large extent by the nature of the large - scale flows : they suggested that an ideal flow across the fan would drive _ spine reconnection _ , in which a current forms along the spine , whereas an ideal flow across the spine would drive _ fan reconnection _ with a strong current in the fan .
they also proposed _
separator reconnection _ with a strong current along a separator joining two nulls . since then , as we shall see in this paper , although behaviour reminiscent of the early spine and fan models may be observed in certain limiting situations , recent numerical experiments have suggested different forms of spine and fan reconnection and also a hybrid spine - fan regime as being the generic modes that occur in practice .
however , the existence of separator reconnection has been well confirmed by a series of numerical experiments @xcite and its importance in the solar corona has been stressed @xcite . in addition , quasi - separatrix layer reconnection ( called slip - running reconnection by @xcite ) has been confirmed in numerical experiments @xcite and in bright point and flare simulations @xcite .
our aim here is simply to look more closely at the nature of reconnection at a 3d null point and to propose a new categorisation to replace spine reconnection and fan reconnection . in the next section it
is necessary to summarise the main results from theory and computational experiments on null - point reconnection and to reinterpret them in the light of the new regimes of reconnection that we are proposing . in the following sections we consider in turn the properties of the three new types of reconnection , namely , _ torsional spine reconnection _ , _ torsional fan reconnection _ and the most common regime _ spine - fan reconnection_.
the simplest linear null point ( for which the magnetic field increases linearly from the null ) has field components @xmath0 in cartesian coordinates or @xmath1 in cylindrical polars , so that @xmath2 identically , where @xmath3 and @xmath4 are constant .
the field lines are given by @xmath5 where @xmath6 and @xmath7 are constants .
the @xmath8-axis is the spine and the @xmath9-plane is the fan .
for this so - called _
proper radial null _ the fan field lines are straight ( figure [ fig:1]a ) .
it is a particular member ( with @xmath10 ) of a wider class of current - free improper radial null points ( @xmath11 ) with curved fan field lines , having field components @xmath12.\ ] ] this is the generic form for a current - free null , since the proper radial null is structurally unstable in the sense that it occurs only for a particular value of @xmath13 , but for simplicity much of the theory so far has used a proper radial null . more generally ,
each of the three field components of a linear null may be written in terms of three constants , making nine in all .
however , @xcite built on earlier work @xcite and showed , by using @xmath2 , by normalising and by rotating the axes , that the nine constants may be reduced to four constants @xmath14 such that @xmath15 where @xmath16 is the current parallel to the spine and @xmath17 is the current perpendicular to the spine
. furthermore , both nulls and separators are susceptible to collapse to form current sheets when the boundary conditions allow it @xcite .
the effects in the ideal region around a 3d null of steady reconnection were studied in the kinematic regime by @xcite extending earlier ideas @xcite .
they solved the equations @xmath18 and @xmath19 for @xmath20 and @xmath21 when @xmath22 is given by equation ( [ eq:1 ] ) and a variety of different boundary conditions are imposed .
in particular , eq .
( [ eq:3 ] ) implies that @xmath23 and then the component of equation ( [ eq:2 ] ) perpendicular to @xmath22 yields @xmath24 which , for certain imposed boundary conditions , may be integrated along field lines ( characteristics ) to determine the value of @xmath25 ( and therefore @xmath21 ) throughout the volume .
then the component of eq .
( [ eq:2 ] ) perpendicular to @xmath22 determines the plasma velocity normal to @xmath22 everywhere as @xmath26 if a continuous flow is imposed across the fan ( figure [ fig:2]a ) , singularities in @xmath21 and @xmath20 are produced at the spine .
@xcite speculated that this would produce a strong current at the spine in what they dubbed _
spine reconnection_. they considered the effect of diffusion in a preliminary manner , but they were unable at the time to resolve the singularities at the spine .
as an example , they considered flows with no @xmath27-component and an electric field of the form @xmath28 giving rise to a velocity @xmath29 for which @xmath30 is continuous at the fan @xmath31 , while @xmath32 is singular at the spine @xmath33 . if , on the other hand , a continuous flow is imposed across the spine ( figure [ fig:2]b ) , singularities are produced at the fan together with a strong _ flipping flow _ ( that @xcite had previously discovered ) .
@xcite suggested that this would produce a strong current at the fan in what they dubbed _
fan reconnection_. a particular example is given in terms of @xmath34 , @xmath35 , @xmath36 by a potential of the form @xmath37 $ ] , which produces a flow field @xmath38 for which @xmath39 is continuous on the planes @xmath40 , while @xmath41 and @xmath39 are singular at the fan ( @xmath42 ) .
however , this analysis left open the questions as to whether it is possible to resolve the singularity and also whether these pure states are likely to be set up in practice .
the next step in the theory was to consider the effect in 3d of an isolated diffusion region where frozen - in flux breaks down and the induction equation is typically of the form @xmath43 reconnection in 3d is very different in many respects from that in 2d . in 2d , a differentiable flux - transporting velocity @xmath44 @xcite satisfying @xmath45 always exists apart from at the x - point itself .
this velocity has a hyperbolic singularity at an x - type null point , where the reconnection takes place .
the magnetic flux moves at the velocity @xmath44 and slips through the plasma , which itself moves at @xmath20 .
furthermore , the mapping of the field lines in 2d is discontinuous at the separatrix field lines that thread the x - point .
this mapping discontinuity is associated with the fact that field lines break and reconnect at one point , namely , the x - point .
while they are in the diffusion region , field lines preserve their connections everywhere , except at the x - point .
two flux tubes that move into the diffusion region break and rejoin perfectly to form two new flux tubes that move out . in 3d ,
surprisingly , none of the above properties carry over and so the nature of reconnection is profoundly different @xcite .
first of all , a single flux tube velocity ( @xmath44 ) does not generally exist @xcite since @xmath46 , but it may be replaced by a pair of flux velocities describing separately what happens to field lines that enter or leave the diffusion region @xcite .
secondly , the mapping of field lines is continuous if there is no 3d null point or separatrix surface .
thirdly , as they move through a 3d diffusion region , magnetic field lines continually change their connections .
fourthly , two tubes do nt generally break and reform perfectly to give two other flux tubes : rather , when the two flux tubes are partly in the diffusion region and so are in the process of reconnecting , they split into four parts , each of which flips in a different manner , a manifestation of the continual change of connections .
( note that in general in 2d and 3d the flux velocity @xmath47 is non - unique .
we choose here to consider the case in which we select @xmath44 by insisting that @xmath48 in the ideal region .
the crucial distinction is that in 2d a single @xmath44 exists ( and is singular ) , while in 3d reconnection _ no single _
velocity @xmath44 exists that satisfies ( [ eq:9 ] ) together with the constraint that @xmath48 in the ideal region .
see refs .
@xcite for further discussion . )
the first attempt to model kinematically the effect of an isolated diffusion region was by @xcite who set up a formalism and applied it to a case without null points .
they solved @xmath49 where @xmath50 , @xmath51 and @xmath2 .
the idea was to impose a sufficiently simple magnetic field that both the mapping and the inverse mapping of the field can be found analytically .
then , after writing @xmath23 , the integral of the component of ( [ eq:6 ] ) parallel to @xmath22 determines @xmath25 everywhere as an integral @xmath52 along field lines , in terms of the values ( @xmath53 ) at one end of the field lines and the distance @xmath54 along field lines .
more simply in terms of a dimensionless stretched distance @xmath55 such that @xmath56 , @xmath57 one way of isolating the reconnection region in these kinematic solutions is by choosing a form of @xmath58 that is localised .
so - called _ pure _ solutions have @xmath59 and produce counter - rotating ( or flipping ) flows of field lines that link the diffusion region .
the rate of flux reconnection is calculated by evaluating the integral @xmath60 along a field line through the diffusion region @xcite .
then the flow normal to the field lines is determined by the component of equation ( [ eq:6 ] ) perpendicular to @xmath22 as @xmath61 these solutions may be regarded as either kinematic ( i.e. , satisfying just the induction equation ) or as fully dynamic in the limit of uniform density and slow flow ( since they also satisfy the equations @xmath62 and @xmath63 ) .
@xcite applied this formalism to determine the behaviour of the magnetic flux when an isolated diffusion region contains a spiral null point , i.e. a null with current directed parallel to the spine line .
the imposed magnetic field was @xmath64 or @xmath65 in cylindrical polars , with the spine and current both directed along the @xmath8-axis , where @xmath66 is a dimensionless current density .
the diffusion region was assumed to be a cylinder of radius @xmath13 and height @xmath67 ( figure [ fig:3]a ) .
first of all , a pure elementary solution which describes the core of the reconnection process was obtained by setting the flow to zero outside the volume defined by the ` envelope ' ( @xmath68 ) of flux that threads the diffusion region . inside @xmath68
the flow and flux velocities are purely rotational ( i.e. , in the @xmath27-direction ) , so that there is no flow across either the spine or the fan .
the reconnection rate is @xmath69 along the spine , and measures the rate of rotational mis - matching of the flux velocities of field lines entering and leaving the diffusion region .
to this solution any ideal solution ( @xmath70 ) may be added and in particular they considered a stagnation - point flow of the form @xmath71 , which brings flux into @xmath68 and carries it out again .
the result is a transition from o - type to x - type flow near the null when @xmath72 exceeds a critical value .
what this solution suggests , therefore , is that a type of spine reconnection with strong current along the spine direction is possible when there are twisting flows about the spine .
this is quite different from the spine reconnection that was envisaged in @xcite and so here we propose to call it _
torsional spine reconnection _ and discuss its properties further in section [ sec:3 ] . next , @xcite applied the same approach to a diffusion region ( @xmath73 ) containing a null point having a uniform fan - aligned current ( @xmath74 ) in the @xmath75-direction and field components @xmath76 the diffusion region was assumed to have the shape of a disc of radius @xmath13 and height @xmath67 ( see fig .
[ fig:3](b ) ) , inside which the magnetic diffusivity decreases smoothly and monotonically from the null to zero at its boundary .
outside @xmath73 it vanishes .
the resulting plasma flow was surprisingly found to be quite different from the fan reconnection of @xcite , since it is found to cross both the spine and fan of the null .
field lines traced from footpoints anchored in the fan - crossing flow are found to flip up and down the spine , whereas those that are traced from the top and bottom of the domain flip around the spine in the fan plane , as envisaged by @xcite .
the reconnection rate is again given by an integral of the form ( [ eq:8 ] ) , this time along the fan field line parallel to the direction of current flow ( here the @xmath75-axis ) .
for such a mode of reconnection this expression can be shown to coincide with the rate of flux transport across the fan ( separatrix ) surface @xcite .
it is possible to find a solution that has similar field line behaviour to the pure fan reconnection envisaged by @xcite , with flow across the spine but not the fan , by adopting instead a field of the form @xmath77 with a fan @xmath75-current @xmath78 ( see ref .
it is also possible to model pure spine reconnection with flow across the fan but not the spine by considering @xmath79 . with a fan @xmath75-current @xmath80 .
both of these fields have a vanishing current at the null .
however , a key property of a null point is the hyperbolic field structure , which tends to focus disturbances and thus generate non - zero currents at the null for the _ primary reconnection modes_. the above pure spine and fan solutions should therefore not be considered as fundamental or primary reconnection modes but as _ secondary reconnection modes _ in the sense that the current vanishes at the null .
it has been suggested that solutions for spine reconnection in incompressible plasmas @xcite may not be dynamically accessible , and while incompressible fan solutions @xcite are dynamically accessible @xcite , this breaks down when the incompressibility assumption is relaxed @xcite .
it turns out that the generic null point reconnection mode that is observed in numerical experiments in response to shearing motions is one in which there is a strong fan current with flow across both spine and fan , and which is in some sense a combination of the spine and fan reconnection of @xcite .
we propose here to call it _ spine - fan reconnection _ and discuss its properties further in section [ sec:5 ] .
several numerical experiments have been conducted in order to go beyond the constraints of analytical theory and to shed more light on the nature of reconnection at a 3d null .
the aim was also to see whether the types of reconnection envisaged qualitatively could indeed take place in practice and to discover whether any other regimes are possible .
first of all , @xcite investigated propagation of a helical alfvn wave towards the fan plane , launched by a rotational driving of the field lines around the spine .
this led to the concentration of current in the fan plane and suggests the possibility of _ torsional fan reconnection _ which we shall propose in section [ sec:4 ] .
( for highly impulsive driving coupling to a fast mode wave that wraps around the null was also observed . ) on the other hand @xcite used a resistive mhd code to show how rotational disturbances of field lines in the vicinity of the fan plane can also produce the strong currents along the spine that are symptomatic of _ torsional spine reconnection _
( section [ sec:3 ] ) .
then @xcite used an ideal lagrangian relaxation code to follow the formation of current sheets by the collapse of a line - tied 3d null in a compressible plasma .
this was a result of the focussing of externally generated large - scale stresses in the field in response to an initial shearing of either the spine axis or fan plane .
building on a previous linear theory by @xcite , they found that locally the fan and spine collapse towards each other to form a current sheet singularity .
( b ) the resulting collapse of spine and fan to form _ spine - fan reconnection _ , showing the current - density contours ( colour ) and flow velocity ( white ) in the @xmath81 plane ( after * ? ? ? * ) .,title="fig : " ] -axis .
( b ) the resulting collapse of spine and fan to form _ spine - fan reconnection _ , showing the current - density contours ( colour ) and flow velocity ( white ) in the @xmath81 plane ( after * ? ? ?
* ) .,title="fig : " ] this was followed up by @xcite , who used a resistive mhd code to investigate the formation and dissipation of the current sheet in response to shearing of the spine , as shown in figure [ fig:4 ] .
the results support the idea of _ spine - fan reconnection _ in which current concentrates around the null ( in a sheet spanning the spine and fan ) .
including compressibility does not affect the results qualitatively , except that in the incompressible limit the spine - fan current is found to reduce purely to a fan current @xcite with behaviour closely resembling earlier fan reconnection models @xcite .
so pure _ fan reconnection _ can be either an incompressible limit of _ spine - fan reconnection _ or , as we have seen in section [ sec:2c ] , the result of a secondary fan current which vanishes at the null .
the type of reconnection set up at a 3d null depends crucially on the nature of the flows and boundary conditions that are responsible for the reconnection .
let us suppose first that a rotation of the fan plane drives a current along the spine and gives rise to torsional spine reconnection , as sketched in figure [ fig:5]a .
the nature of the reconnection is that in the core of the spine current tube there is rotational slippage , with the field lines becoming disconnected and rotating around the spine ( see @xcite ) : figure [ fig:5]b shows on the left side a particular magnetic field line and its plasma elements at @xmath82 ; in the upper part of the figure ( above the shaded diffusion region ) this field line and its attached plasma elements rotate about the spine through positions at times @xmath83 , @xmath84 and @xmath85 ; in the lower part of the figure ( below the diffusion region ) the plasma elements that were on the field line at @xmath86 rotate to positions at @xmath83 , @xmath84 and @xmath85 that are on different field lines . , @xmath87 , @xmath88 , that initially ( @xmath82 ) lay on one field line .
( c ) the reconnection rate measures a rotational mismatching of flux threading the diffusion region , namely the difference between the rates of flux transport through surfaces a and b.,title="fig : " ] , @xmath87 , @xmath88 , that initially ( @xmath82 ) lay on one field line .
( c ) the reconnection rate measures a rotational mismatching of flux threading the diffusion region , namely the difference between the rates of flux transport through surfaces a and b.,title="fig : " ] , @xmath87 , @xmath88 , that initially ( @xmath82 ) lay on one field line .
( c ) the reconnection rate measures a rotational mismatching of flux threading the diffusion region , namely the difference between the rates of flux transport through surfaces a and b.,title="fig : " ] a steady kinematic solution may be found following the approach of section [ sec:2c ] . the electric field may be written as the sum ( @xmath89 ) of a nonideal pure ( elementary ) solution satisfying @xmath90 and an ideal solution satisfying @xmath91 consider a spiral null point ( equation [ eq:10 ] ) and suppose the diffusion region is a cylinder of radius @xmath13 and height @xmath67 and that the magnetic diffusivity has the form @xmath92 , where @xmath93 and @xmath94 vanishes on the boundary of the diffusion region and outside it .
the field lines for this spiral null may be obtained by solving @xmath95 suppose we start a field line at the point @xmath96 at @xmath97 .
then the field line equations are @xmath98 these give a mapping from an initial point @xmath99 to any other point @xmath100 along a field line .
the inverse mapping is @xmath101 where @xmath102 .
the pure elementary solution describes the core of the reconnection process .
it is obtained following refs .
@xcite by solving @xmath103 with @xmath50 , @xmath51 and @xmath2 .
thus we write @xmath104 with @xmath105 given by equation ( [ eq:7 ] ) and set @xmath59 so that the flow vanishes outside the diffusion region . inside the diffusion region
the flow and flux velocities have no component across either the spine or the fan . for the spiral magnetic field @xmath106 and the mapping ( [ eq:11 ] ) ,
@xmath105 becomes @xmath107 where @xmath108 .
then , once a form for @xmath58 is assumed , this may be integrated to give @xmath109 . after using the inverse mapping ( [ eq:12 ] ) , we can then deduce @xmath110 and therefore @xmath21 and @xmath111 everywhere .
if a diffusion region is isolated , a change of connectivity of field lines may be studied , by following field lines anchored in the ideal region on either side of the diffusion region .
a diffusion region is in general isolated if @xmath112 is localised in space . in practical cases in astrophysics ,
this is likely to be mainly because @xmath113 is localised but in addition sometimes because as a consequence @xmath58 is also localised .
some numerical simulations have a localised @xmath58 , whereas others have a uniform @xmath58 or a purely numerical dissipation . but
the important feature in all these cases is that the product @xmath114 is localised .
now , in each of our solutions below , we follow refs .
@xcite in choosing a spatially localised @xmath112 by imposing a spatially localised resistivity profile together with a @xmath113 that is not localised .
the reason for doing this is to render the mathematical equations tractable , since we have not yet discovered a way to do so with a localised @xmath113 .
the quantitative spatial profiles of physical quantities will depend on the @xmath58 profile , but the qualitative topological properties of the field line behaviour in such models are expected to be generic and independent of the particular profile chosen for @xmath58 . indeed , the topological properties of the reconnection models of refs .
@xcite have been verified by the numerical simulations @xcite .
there are four regions with different forms for @xmath105 , as illustrated in figure 6 , which shows a vertical cut in the first quadrant of the @xmath115-plane . in region
( 1 ) threaded by field lines that enter the diffusion region ( shaded ) from above , we assume @xmath116 , so that there is no electric field or flow . the same is true in region ( 2 ) which lies above the flux surface @xmath117 that touches the upper corner @xmath118 of the diffusion region .
we calculate below the forms of @xmath119 in the diffusion region ( 3 ) and in the region ( 4 ) threaded by field lines that leave the diffusion region through its sides .
-@xmath8 plane , showing 4 different regions ( 1)-(4 ) in which @xmath119 is calculated .
a magnetic field line whose projection intersects the top of the diffusion region in @xmath120 and the side in @xmath121 contains typical points @xmath122 inside and beyond the diffusion region . the bounding field line @xmath117 is shown dashed . ] for example ,
let us assume that @xmath58 vanishes outside the diffusion region ( @xmath73 ) and that inside @xmath73 it has the form @xmath123 which peaks at the origin and vanishes on the boundary of @xmath73 .
first , we use the mapping ( [ eq:11 ] ) to substitute for @xmath124 and @xmath8 , and integrate with respect to @xmath55 from the point @xmath120 on the top of @xmath73 to the point @xmath122 inside @xmath73 ( figure [ fig:6 ] ) .
then we use the inverse mapping ( [ eq:12 ] ) to replace @xmath125 and @xmath55 , and finally we obtain the potential throughout @xmath73 ( region ( 3 ) in figure [ fig:6 ] ) as @xmath126.\ \ \ \ \ \label{eq:13}\end{aligned}\ ] ] this then determines the components of the electric field ( @xmath127 ) everywhere in @xmath73 as @xmath128 @xmath129 in order to find @xmath130 in region ( 4 ) of figure 6 , we start with the values of @xmath105 at the point @xmath121 on the side of the diffusion region ( figure 6 ) and then calculate @xmath105 at any point @xmath122 that lies on the same field line in region ( 4 ) to the right of @xmath131 .
thus , after putting @xmath132=@xmath133 in the expression ( [ eq:13 ] ) for @xmath25 that holds in the diffusion region , we obtain @xmath134 . \label{eq:14}\
] ] since ideal mhd holds in region ( 4 ) , @xmath130 is constant along the field line ( @xmath135 ) joining @xmath131 to @xmath136 , and so the value of @xmath105 at @xmath136 is simply @xmath137 .
\label{eq:15}\end{aligned}\ ] ] the solution for @xmath138 can be obtained in a similar manner by integrating from @xmath139 .
we may now make various deductions from the solution .
the reconnection rate depends on the form of @xmath58 and is given in order of magnitude by @xmath140 where @xmath141 is the electric field at the centre of the diffusion region and @xmath142 is the dimension of the diffusion region along the magnetic field direction .
in our example , @xmath143 , where @xmath144 is the value of the current at the origin , and along the spine ( [ eq:13 ] ) implies that @xmath145 and so the reconnection rate becomes more accurately @xmath146 the other feature that we can deduce from the electric field components is the perpendicular plasma velocity given by eq .
( [ eq:9 ] ) . in particular , on the fan plane ( @xmath31 ) inside @xmath73 , @xmath147 , @xmath148 , @xmath149 and @xmath150 so that there is a rotational component given by @xmath151 where @xmath152.$ ] the nature of the flow becomes clear if we subtract a component parallel to @xmath22 in order that @xmath153 ( we are free to do this since the component of @xmath20 parallel to @xmath22 is arbitrary in the model ) . after doing this
we find that @xmath154 vanishes , leaving @xmath155 , i.e. , the flow corresponds to a pure rotation ( as in the solutions of refs .
@xcite ) . to the above pure diffusive solution
any ideal solution may be added satisfying @xmath156 and @xmath157 , for which the potential ( @xmath70 ) satisfies @xmath158 thus , once the functional form @xmath159 is chosen at the points @xmath99 on @xmath160 , say , that form of @xmath70 is constant along field lines given by the mapping ( [ eq:11 ] ) .
the resulting variation of @xmath161 throughout space is given by substituting for @xmath125 and @xmath72 from the inverse mapping ( [ eq:12 ] ) .
as an example , suppose @xmath162 on the plane @xmath160 .
then throughout the volume we find @xmath163 which implies electric field components @xmath164 then the plasma velocity components follow from @xmath165 as @xmath166 where @xmath167 . in particular , we notice that the flow vanishes on the spine @xmath33 , and that in the fan @xmath31 there is a rotational flow that linearly increases with distance @xmath168 .
the reconnection of field lines takes the form of a rotational slippage .
field lines entering the diffusion region have a flux velocity @xmath169 , while those that leave it have a flux velocity @xmath170 .
@xmath171 is obtained by integrating along field lines that enter from the ideal region on one side , while @xmath172 is obtained by integrating backwards along field lines that leave from the other side .
the rate of slippage between inward and outward flux bundles is given by @xmath173 and represents the rate of reconnection , which we have evaluated directly above in equation ( [ eq:16 ] ) .
this reconnection rate , obtained by integrating @xmath174 along the spine , measures the difference between the rates of flux transport across surface a and surface b in fig .
[ fig:5](c ) .
note that the extra ideal solution does not change the rate of relative slippage .
however , it does allow for different external conditions , such as rotation above and below the diffusion region in the same or opposite senses . to see the effect of a non - rotational ideal flow
see ref . @xcite . in the solution given above the physical quantities
@xmath175 and @xmath176 are continuous but not differentiable at the boundary between regions ( 3 ) and ( 4 ) .
this is a sacrifice made for tractability and pedagogic purposes . for a solution with differentiable physical quantities , see @xcite . in the above
, the diffusion region was imposed to be a cylinder whose width ( @xmath13 ) and height ( @xmath67 ) are parameters of the solution .
the formation , in a self - consistent fashion , of such a cylindrical diffusion region was observed in the simulations described by @xcite . in one of their simulations they imposed a twisting perturbation of the magnetic field in the vicinity of the fan plane .
as the disturbance propagated inwards towards the null , it was dominated by a helical alfv ' enic wave travelling along the field lines and thus stretching out along the spine .
the result was a tube of current focussed around the spine , giving a large aspect ratio to the diffusion region ( @xmath177 ) . during the process of torsional spine reconnection
the narrowing and elongation of the current tube is likely to continue until the rotational advection that twists the field and intensifies the current is balanced by the rotational slippage .
now suppose that we rotate the field lines near the spine in opposite directions above and below the fan .
then a current builds up in the fan . within the fan current sheet , field lines experience rotational slippage @xcite , in the opposite sense above and below the fan , in what we propose to term _ torsional fan reconnection _
( figure [ fig:7 ] ) . again there is no flow across either spine or fan .
the counter - rotation ( above and below the fan ) of the region around the spine builds up a double - spiral structure near the null point , with a current that possesses two components : an axial component that reverses sign at the fan plane and a radial component . a counter - rotating part to the diffusion velocity ( @xmath178 )
is set up in the @xmath27-direction that reverses sign at the fan . in order to model such reconnection
, we consider what we term a _
double - spiral null point _ with field components @xmath179 where @xmath180 and @xmath181 are positive integers and the corresponding current components are @xmath182 an alternative solution to the one presented below is outlined in the appendix .
the field line equations for a mapping from an initial point @xmath99 to any other point @xmath100 are @xmath183 and the inverse mapping is @xmath184 where @xmath102 .
let us follow the approach of section iiia and calculate the pure non - ideal solution
. we shall assume the diffusion region to be a disc of radius @xmath13 and height @xmath67 , with the same diagram as before ( figure 6 ) , except that the diffusion region is now expected to be in the shape of a thin disc ( with @xmath185 ) rather than a thin tube ( with @xmath186 ) .
assuming , as before , that @xmath187 vanishes in regions ( 1 ) and ( 2 ) , we evaluate it in region ( 3 ) by integrating from a point @xmath120 on the top of the disc to a point @xmath122 inside the diffusion region .
after using equation ( [ eq:7 ] ) and the mapping and setting @xmath188 , the expression for the potential at @xmath122 then becomes @xmath189 we adopt the following general form for the magnetic diffusivity inside the diffusion region ( d ) @xmath190 which peaks at the null point and vanishes on the boundary of d when @xmath191 and @xmath192 are positive and @xmath192 is even . after substituting into ( [ eq:21 ] ) and using the mapping and inverse mapping , we find the potential throughout the diffusion region .
in particular , it transpires that an important constraint on the constants @xmath180 , @xmath181 , @xmath191 and @xmath192 is that @xmath193 be finite and continuous at the fan plane . as an example , one set of such constants that works is @xmath194 , @xmath195 , @xmath196 and @xmath197 , for which @xmath198 the corresponding components of electric field are @xmath199 @xmath200 in order to find @xmath187 in region ( 4 ) of figure 6 , as before , we calculate its value at any point @xmath121 where a field line leaves the diffusion region , and then project that value along that field line .
thus , after putting @xmath201 into equation ( [ eq:22 ] ) , we obtain @xmath202 since ideal mhd holds in region ( 4 ) , @xmath130 is constant along the field line ( @xmath135 ) joining @xmath131 to @xmath136 , and so the value of @xmath105 at @xmath136 is simply @xmath203 the electric field components vanish in both the spine and the fan but are strong just above and below the fan , which is where the reconnection of field lines occurs by rotational slippage in a similar fashion to torsional spine reconnection .
near the spine and fan we have to lowest order in @xmath124 and @xmath8 @xmath204 the reconnection rate is the maximum value of ( @xmath205 ) along any field line ( @xmath206 ) , each of which enters the diffusion region from above at @xmath207 and leaves at @xmath121 .
along such field lines the integral is a function of @xmath208 and @xmath209 , namely : @xmath210 .
\end{aligned}\ ] ] when @xmath211 for a slender disc - shaped diffusion region , this reduces to @xmath212 .
\end{aligned}\ ] ] if @xmath13 and @xmath213 are held fixed and @xmath125 is varied , the maximum value of this occurs at @xmath214 , giving a reconnection rate of @xmath215 as for torsional spine reconnection , the reconnection rate is proportional to the potential @xmath108 , but in this case , as well as being proportional to the current density @xmath66 and diffusion region height ( @xmath213 ) , it also depends on its aspect ratio ( @xmath216 ) .
again , as before , a wide range of ideal solutions ( @xmath70 ) may be added to the diffusive solution . thus , if for example , @xmath217 on the top ( @xmath160 ) of the diffusion region , the fact that it remains constant along field lines ( @xmath206 ) determines @xmath218from which the electric field components can be deduced .
for instance , the case @xmath219 gives an electric field of@xmath220which implies a plasma velocity with components normal to the magnetic field of @xmath221where @xmath222b_{0}a^{4}b^{11}/l_{0}$ ] . in particular , it gives a rotational component ( @xmath223 ) that is odd in @xmath8 and so represents the kind of counter - rotation that is typical of torsional fan reconnection .
-direction ( large dark arrows ) : its width is @xmath224 , its total length @xmath225 in the @xmath226-plane , and its length @xmath227 common to spine and fan .
, title="fig : " ] -direction ( large dark arrows ) : its width is @xmath224 , its total length @xmath225 in the @xmath226-plane , and its length @xmath227 common to spine and fan .
, title="fig : " ] in general , if the driving motions tend to shear a null point rather than rotate it , then the result will be _ spine - fan reconnection_. a shear disturbance of either the spine or fan of the null will tend to make the null ` collapse ' .
that is , the resulting lorentz force acts to increase the displacement , just as at a 2d null ( see ref .
* ) and as at a separator @xcite .
this collapse is opposed by the line - tying at the boundaries , and what occurs is that the shear distortion tends to focus in the weak field region in the vicinity of the null point , forming a localised current sheet @xcite .
what distinguishes spine - fan reconnection from the other null point reconnection modes is that flux is transferred across both the spine and fan .
furthermore , the current concentration is in the form of a localised sheet that is inclined at an intermediate angle between the spine and fan indeed the current sheet contains part of both the spine and the fan ( see figures [ fig:4]b and [ fig:8]a ) . as mentioned above , the reconnection rate for this mode of reconnection is obtained by integrating @xmath228 along the fan field line with the maximum value of @xmath229 . by the symmetry of the simple models described herein , this is the field line parallel to the current orientation at the null ( perpendicular to the applied shear ) .
the reconnection rate thus obtained measures exactly the rate of flux transport in the ideal region across the fan separatrix surface . to illustrate the properties of this mode of null point reconnection
, we describe here briefly the results of resistive mhd simulation runs ( see @xcite for an initial description ) . in the simulations ,
a shear velocity is prescribed on the ( line - tied ) @xmath8-boundaries , which advects the spine footpoints , see fig .
[ fig:4](a ) ( the results are qualitatively the same if the fan is distorted instead ) . the current sheet that forms in response to the shearing
is localised in all three directions about the null
. however , in the plane of the applied shear ( perpendicular to the current orientation at the null ) the magnetic field and current patterns have a similar appearance to a 2d x - point configuration . as one moves away from the null in the fan along the direction of current flow ,
the magnetic field strength parallel to the current ( sometimes known as the ` guide field ' ) strengthens , while the current intensity weakens
see fig .
[ fig:8 ] .
the boundary shearing velocity is ramped up to a constant value ( @xmath230 ) at which it is held until being ramped down to zero , at @xmath231 ( space and time units in the code are such that an alfv ' en wave would travel one space unit in one time unit for uniform density and magnetic field @xmath232 ) .
the resistivity is uniform .
current focusses at the null during the period when the driving occurs , and when the driving ceases both the current and reconnection rate peak , after which the null gradually relaxes towards its original potential configuration . under continuous driving , it is unclear whether a true steady state would be set up , or whether the current sheet would continually grow in dimensions and intensity ( see ref .
this is an open question for future study . for the case of transient driving , the peak current and reconnection rate increase linearly with the driving velocity .
here we examine more closely the sheet geometry , and its scaling with the driving velocity , and also investigate the scaling of this geometry , the peak current and peak reconnection rate with resistivity .
as previously noted , the current sheet that forms in spine - fan reconnection is focussed at the null , locally spanning both the spine and fan .
the sheet has a tendency to spread along the fan depending on the parameters chosen ( spreading is enhanced by lowering @xmath230 or increasing the plasma-@xmath233 ) .
we examine four spatial measurements ( see figure [ fig:8]b ) associated with the current sheet , focussing on the time when the current magnitude is a maximum and defining the boundary of the sheet to be an isosurface at 50% of @xmath234 .
the sheet _ thickness _ is @xmath224 , the _ length _
@xmath225 is the total extension in the @xmath226-plane ( normal to @xmath113 ) , @xmath235 is the length of the ` collapsed ' section ( within which the sheet contains both spine and fan ) , and the _ width _ @xmath236 is the extension of the sheet along the @xmath75-direction ( parallel to @xmath113 ) . of ( top left
bottom right ) @xmath225 , @xmath235 , @xmath224 and @xmath237 ( see figure [ fig:8]b for notation ) . ] the scaling of these dimensions with ( peak ) driving velocity @xmath230 is shown in figure [ fig:9 ] ( we fix @xmath238 ) .
the angle @xmath237 between the current sheet and the @xmath31 plane can be seen to increase as the driving velocity increases .
this can be put down to the fact that the stronger driving creates a stronger lorentz force the force that is responsible for the collapse .
as expected , @xmath235 increases as @xmath230 increases .
this is a result of the fact that the spine footpoints are sheared further for larger @xmath230 , and there exists in fact a close correspondence ; @xmath239 .
in contrast to @xmath235 , @xmath225 shows a linear _ decrease _ with @xmath230 ( as does @xmath236 , see ref .
@xcite ) , showing that as the collapse becomes stronger the distortion of the magnetic field focusses closer and closer around the null itself .
the decline in @xmath225 with increasing @xmath230 must of course cease once @xmath240 , as is the case for the strongest driving considered .
examining finally the sheet thickness @xmath224 , any variation is within the error bars of our measurements , and moreover the resolution is not sufficient for firm conclusions to be drawn .
we turn now to consider the scaling of the current sheet with @xmath58 , setting @xmath241 , see figure [ fig:10 ] . as @xmath58 decreases ,
@xmath242 increases , while the reconnection rate decreases . in both cases , with the limited data of this preliminary study , the proportionality appears to be somewhere between power law and logarithmic . that the run with the largest resistivity does not seem to fit the trend for the reconnection rate is likely to be because the current significantly dissipates before reaching the null itself due to the high resistivity ( @xmath243 ) .
accompanying the increase in @xmath242 with @xmath58 is , as expected , a decrease in the thickness @xmath224 .
on the other hand , the overall dimensions of the sheet , @xmath225 and @xmath236 , seem to be unaffected by @xmath58 , to within our measurement accuracy .
finally , as @xmath58 decreases and the current becomes more intense , the collapse becomes more pronounced as evidenced by increases in both @xmath235 and @xmath237 . of ( top left
bottom right ) the peak current density , the peak reconnection rate , @xmath235 , @xmath224 and @xmath237 ( see figure [ fig:8]b for notation ) . ] the relationships discussed briefly above certainly warrant further investigation with carefully designed , higher resolution simulations , as do the corresponding scalings for the continuously driven case .
we have here outlined a new categorisation of reconnection regimes at a 3d null point . in place of the two previous types , namely , spine and fan reconnection
, we suggest that three distinct generic modes of null point reconnection are likely to occur .
the first two are caused by rotational motions , either of the fan or of the spine , leading to either _ torsional spine reconnection _ or _ torsional fan reconnection_. these involve slippage of field lines in either the spine or the fan , which is quite different from classical 2d reconnection , but does involve a change of magnetic connection of plasma elements .
even though pure spine or fan reconnection may occur in special situations ( such as when @xmath244 or there are high - order currents ) , it is much more likely in practice that a hybrid type of reconnection takes place that we refer to as _ spine - fan reconnection_. this is the most common form of reconnection that we expect to see in three dimensions at a null point , since it is a natural response to a shearing of the null point .
it is most similar of all the 3d reconnection regimes to classical 2d reconnection and involves transfer of magnetic flux across both the spine and the fan .
it possesses a diffusion region in the form of a current sheet that is inclined to the fan and spine and has current localised in both the spine and fan , focussed at the null . in future ,
much remains to be determined about these new regimes of reconnection that have been observed in numerical experiments .
one is the shape and dimensions of the diffusion regions and their relation to the driving velocity and the magnetic diffusivity .
another key question is : what is the rate of reconnection at realistic plasma parameters , and is there a maximum value ?
since the analytical theory is so hard in three dimensions , progress is likely to be inspired by future carefully designed numerical experiments .
we are grateful to guillaume aulanier , klaus galsgaard , and our colleagues in the st andrews and dundee mhd groups for stimulating discussions , especially gunnar hornig and clare parnell , and to the eu solaire network and uk particle physics and astronomy research council for financial support .
erp is also grateful to jiong qiu , dana longcope and dave mckenzie for inspiring suggestions in bozeman where this work was completed .
here we present another pure nonideal solution for torsional fan reconnection .
it possesses a much simpler electric current , but the assumed form for the magnetic diffusivity has to be more complex and vanish on the fan .
we follow section [ sec:4 ] but consider a different form for the magnetic field of a double - spiral point , namely , @xmath245 and corresponding current components @xmath246 assuming , as before , that @xmath187 vanishes in regions ( 1 ) and ( 2 ) , we evaluate it in region ( 3 ) of figure 6 by integrating from a point @xmath120 on the top of the diffusion region to a point @xmath122 inside the diffusion region . after using equation ( [ eq:7 ] ) and the mapping
( [ eq:23 ] ) and setting @xmath188 , the expression for the potential at @xmath122 then becomes @xmath249 as an example , we adopt the following form for the magnetic diffusivity inside the diffusion region ( d ) @xmath250 which peaks above and below the null point and vanishes on the boundary of d. we have also chosen it to vanish on the fan plane since @xmath251 vanishes there and in order to facilitate a closed form solution with continuous physical quantities . after substituting into ( [ eq:25 ] ) and using the mapping ( [ eq:23 ] ) and inverse mapping ( [ eq:24 ] ) , we find the potential throughout the diffusion region as @xmath252 and the corresponding components of electric field as @xmath253 @xmath254 we note that the solutions in the lower half plane @xmath138 may be obtained simply by replacing @xmath102 by @xmath255 .
there is a term in @xmath193 that behaves like @xmath256 and so is usually discontinuous at the fan plane .
this discontinuity may , however , be avoided by setting @xmath257 and balancing it with the term in @xmath258 when @xmath259 $ ] . as a simple example , let us consider @xmath260 and @xmath261 .
then , in order to find @xmath187 in region ( 4 ) of figure 6 , as before , we calculate its value at any point @xmath121 where the field line leaves the diffusion region , and then project that value along the field line .
thus , after putting @xmath201 into equation ( [ eq:26 ] ) , we obtain @xmath262 . \nonumber\end{aligned}\ ] ] as before ideal mhd holds in region ( 4 ) , and so @xmath130 can be calculated from the fact that it is constant along the field line ( @xmath135 ) joining @xmath131 to @xmath136 .
the electric field components follow and can be used to calculate the reconnection rate and the flow velocity in the usual way . comparing the two forms of torsional fan solution ,
the advantage of the one presented in section 4 is that the magnetic diffusivity peaks at the null point , but the slightly unwelcome feature is that the current density vanishes along the spine and fan . by comparison , the solution in this appendix has a current that vanishes at the null point but nowhere else in the spine or fan , but its disadvantage is that we had to choose the diffusivity to vanish in the fan . | recent advances in theory and computational experiments have shown the need to refine the previous categorisation of magnetic reconnection at three - dimensional null points points at which the magnetic field vanishes .
we propose here a division into three different types , depending on the nature of the flow near the spine and fan of the null .
the spine is an isolated field line which approaches the null ( or recedes from it ) , while the fan is a surface of field lines which recede from it ( or approach it ) .
so - called _
torsional spine reconnection _ occurs when field lines in the vicinity of the fan rotate , with current becoming concentrated along the spine , so that nearby field lines undergo rotational slippage . in _
torsional fan reconnection _ field lines near the spine rotate and create a current that is concentrated in the fan with a rotational flux mismatch and rotational slippage . in both of these regimes , the spine and fan are perpendicular and there is no flux transfer across spine or fan . the third regime , called _ spine - fan reconnection _
, is the most common in practice and combines elements of the previous spine and fan models . in this case , in response to a generic shearing motion , the null point collapses to form a current sheet that is focused at the null itself , in a sheet that locally spans both the spine and fan . in this regime the spine and fan are no longer perpendicular and there is flux transfer across both of them . |
mildly relativistic and non - relativistic absorption troughs are observed in the x - ray spectra of active galactic nuclei ( agns ) .
a qualitative separation is usually done between classical @xmath0 warm absorbers , we refer to here ( throughout paper ) as _ low - velocity _ outflows @xcite , and ultra - fast outflows ( ufos ) @xmath4 , we refer to here as _ high - velocity _ outflows @xcite . on one hand , the extracted velocity from the classical @xmath0 warm absorber is mainly obtained using the fe m - shell @xmath5 unresolved transition array ( uta ) @xcite , and o vii or o viii resonance lines ( e.g. , * ? ? ?
* the hallmark of the classical warm absorber ) .
the spatial location of this absorbing material is uncertain .
models of x - ray absorbers in agn place them at a wide range of distances from the central source .
specifically they are suggested to be winds originating from the accretion disk @xcite , located at the dusty ( @xmath6 pc ) torus scales @xcite or even beyond the narrow - line region ( e.g. * ? ? ?
on the other hand , blueshifted fe k absorption lines have been detected in recent years at @xmath7 kev in the x - ray spectra of several radio - quiet agns ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
they are usually interpreted as due to resonant absorption from fe xxv - xxvi associated with a zone of circumnuclear gas photoionized by the central x - ray source , with ionization parameter @xmath8 erg / s cm and column density @xmath9 .
the energies of these absorption lines are systematically blueshifted and the corresponding velocities can reach up to mildly relativistic values of @xmath10 .
these findings are important because they suggest the presence of previously unknown massive and highly ionized absorbing material outflowing from their nuclei , possibly connected with accretion disk winds / ejecta ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
several acceleration mechanisms have been proposed to explain these outflows : ( 1 ) thermally accelerated winds @xcite ; ( 2 ) radiation pressure through thomson scattering and magnetic forces ( mhd , * ? ? ?
* ) ; ( 3 ) and radiation pressure due to the absorption of spectral lines ( e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although the first one can explain the velocities we observe in the low - velocity outflows , it can be excluded because it can not explain the @xmath11 we observe in ufos @xcite .
@xcite seem to reproduce the velocities observed in _
low- _ and _ high - velocity _ outflows . on the other hand , @xcite ,
invoke radiation pressure due to lines to explain the @xmath12 outflow detected in a good s / n x - ray spectrum of a _ high - z _ quasar ( the broad absorption line [ bal ] , ) , and they reproduce , as part of the procedure , the fe xxv - xxvi lines detected at @xmath7 kev . here
we focus on this kind of approach , since it allows us to explore , very efficiently , a wide range of physical parameters of the system .
the goal of this letter is to place a proof of idea for a systematic study about the operating acceleration mechanisms in both ; x - ray _ low- _ and _ high - velocity _ outflows , using an anisotropic radiative pressure framework ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , beginning with line radiation pressure .
we present the observables from which we build the model in
[ obser1 ] .
the details of the proposed model are presented in [ model1 ] .
the results and the discussion are in [ result1 ] .
we summarize in [ summ1 ] .
in this section we present the observables of the two types of outflows , since they are the initial motivation for the proposed model . when describing the physical conditions of warm absorbers , it is common to use the definition of ionization parameter @xmath13 @xcite , where @xmath14 is the total ionizing flux ( @xmath15 ) , and @xmath16 is the gas density .
the source spectrum is described by the spectral ( specific , energy dependent @xmath17 ) luminosity @xmath18 , where @xmath19 is the integrated luminosity from 1 to 1000 ryd , and @xmath20 .
so we describe the @xmath21 warm absorber outflows as absorbing material around a supermassive black hole ( smbh ) with mass @xmath22 @xcite , column density of the absorbing material @xmath23 , flowing outwards at velocities @xmath24 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , at medium ionization states @xmath25 erg s@xmath26 cm @xcite . when computing the energetics they find mass loss rates of reported by @xcite in their table 4 , excluding ark 564 ( outlier @xmath27 m@xmath28 yr@xmath26 ) . ]
@xmath29 m@xmath28 yr@xmath26 ratios of @xmath30 to accretion rates , reported by @xcite in their table 4 , excluding ark 564 ( outlier @xmath31 ) . ] @xmath32 and kinetic luminosity @xmath33 , of the orders of @xmath34 erg / s , representing less than 1 % of the bolometric luminosity @xcite .
the main conclusion from these estimations is that these outflows contributes little to the energy injected in to the host galaxy .
but the amount of matter processed over the agn lifetime can be significant ( also in accordance with * ? ? ?
* for instance ) . the characteristics of the ultra - fast outflows with @xmath35@xmath3610,000 ( @xmath360.03c )
can be derived from the blue - shifted fe xxv - xxvi absorption lines detected by @xcite in a complete sample of local seyfert galaxies .
such features are detected in @xmath374060% of the sources , which suggests a covering fraction of @[email protected] .
tombesi et al .
( 2011 ) also performed a photo - ionization modeling of these lines .
they derived the distribution of the outflow velocities , which ranges from @xmath370.03c up to mildly relativistic values of @xmath370.3c , with a peak and mean value at @xmath370.14c . as expected , these absorbers are highly ionized , with log@xmath39@xmath3736 erg s@xmath26 cm , and have large column densities , in the range @xmath40@xmath37@xmath41@xmath42 @xmath43 .
the smbh masses of the seyferts in the @xcite sample have a mean value of @xmath44 @xmath45 @xcite .
when computing the energetics of these outflows , we see that these are more massive than the low - velocity ones , @xmath30@xmath37 @xmath46 @xmath45 yr@xmath26 @xmath37@xmath47 , and also much more powerful , with a mechanical power of @xmath37@xmath48@xmath49 erg / s ( e.g. , pounds et al .
2003 ; markowitz et al .
2006 ; braito et al .
2007 ; cappi et al .
2009 ; tombesi et al .
the latter value is @xmath37 @xmath50% of the bolometric luminosity .
therefore , the high - velocity outflows may potentially play an important role on the expected cosmological feedback from agns ( e.g. , * ? ? ?
* ; * ? ? ?
we build our model based on observational evidence , basically from two sources : @xcite for the classical @xmath0 warm absorbers , referred here as _
low - velocity _ outflows , and @xcite for the ultra - fast outflows ( ufos ) @xmath51 , referred here as _ high - velocity _ outflows . as we see , together the two classes of outflows cover a wide range in velocity and black hole masses .
we seek a physical model which provides the context to explain ( to first approximation ) , part of the observational evidence we have up to now . in order to do that , and to gain some insight into the relationship between outflow velocity ( @xmath52 ) and luminosity
, we invoke the velocity profile as a function of radius , given by hydrodynamical calculations @xcite .
the mathematical shape of the relation is given by assuming an outflow accelerated by radiation pressure from a central source with a bolometric luminosity @xmath53 , and a mass of @xmath54 : @xmath55 = \left [ 2gm_{bh}[i]\left ( { \gamma_{f}}{{l_{bol}[i]}\over{l_{edd}[i ] } } - 1\right ) \left ( { { 1}\over{r_{in } } } - { { 1}\over{r } } \right ) \right]^{1/2 } } , \label{luv1}\ ] ] where @xmath56 is the eddington luminosity , @xmath57 is the force multiplier , where the acceleration due to the absorption of discrete lines is encapsulated @xcite , @xmath58 is the radius at which the wind is launched from the disk , and @xmath59 is the distance of the accelerated portion of the outflow from the central source .
the index _ i _
, runs simultaneously over the two distributions : velocity / bh mass . assuming that we observe the gas when it has reached the terminal velocity of the wind , i.e. @xmath52 at @xmath60 , we can write : @xmath61 = { { l_{edd}[i]}\over{\gamma_{f } } } \left({{v_{out}[i]^2r_{in}}\over{2gm_{bh}[i]}}+1 \right ) .
\label{vout1e}\ ] ] now we have the basic ingredient of the model , and we are ready to build our two classes of outflows . _ set 1 : high - velocity_. this synthetic sample is composed from 1000 black holes with masses normally distributed with mean , @xmath62 m@xmath28 and standard deviation , @xmath63 m@xmath28 .
m@xmath28 ) between the mean extracted from @xcite ( @xmath64 m@xmath28 ) , excluding the mass of mrk 205 ( outlier @xmath65 m@xmath28 ) , and the mean extracted from table 5 of @xcite ( @xmath66 m@xmath28 ) , excluding the mass of iras 13349 + 2438 ( outlier @xmath67 m@xmath28 ) . also we take the larger value of the observed dispersion @xmath68 m@xmath28 . ]
also we use a normal distribution for the outflow velocity of the absorbing gas around the bh with @xmath69 ( this is the best - fit value for set 1 ) , and @xmath70 .
_ set 2 : low - velocity_. in this case we use 1000 black holes with masses normally distributed with the same mean , @xmath62 m@xmath28 and standard deviation , @xmath63 m@xmath28 as before but we use a normal distribution for the outflow velocity of the absorbing gas around the bh with @xmath71 ( this is the best - fit value for set 2 ) , and @xmath72 .
the best - fit values of @xmath73 and @xmath74 ( as well @xmath75 and @xmath76 , see next section ) , are estimated by comparing a grid of models computed using @xmath77 ( ) and @xmath78 ( ) , with the corresponding set of observation : set 3 _ vs _ @xcite and set 4 _ vs _ @xcite .
we performed the comparison using a kolmogorov - smirnov test .
we take the @xmath73 and @xmath74 ( as well @xmath75 and @xmath76 , see next section ) which give maximum p - values of the test . the final p - value for the comparison set 3 _ vs _ @xcite gives @xmath79 ( @xmath80 ) and for the set 4 _ vs _ @xcite @xmath81 ( @xmath82 ) .
the launching radius @xmath58 , for each set of simulations is based on the results of @xcite for set 2 ; i.e. , @xmath83 ( orders of magnitude value ) . for set 1 ,
we set @xmath84 , based on @xcite .
in this section we discuss the results of our simulated samples and summarize their physical context and implications . figure [ vel1 ] shows the theoretical relationship between outflow velocity and bolometric luminosity .
the curvature we see in @xmath52 of set 2 ( and also set 1 , filled circles , our line of sight is along the flow , see below ) , is due to the quadratic dependence of the luminosity with @xmath52 given by equation [ vout1e ] .
it is worth noticing that if we assume a force multiplier of @xmath85 ( this is the best - fit value for set 2 and set 4 , see section [ model1 ] ) ; we are able to reproduce the range of luminosities seen in figure 4 of @xcite .
on the other hand this model is not reproducing well the dispersion we observe in figure ( 4 ) of @xcite , which could be explained by the fact that equation [ vout1e ] will give radially accelerated flows , or in other words , that we are observing all the objects radially along the flow . to include the effect of observing random line - of - sights transversely through different sections of the flow we connect the observed velocity @xmath86 with the intrinsic radial velocity , i.e. @xmath87 , where @xmath88 is the angle between the outflow direction and the line of sight . then using the same luminosities produced by set 2 , but plotting ( open squares ) against 1000 @xmath86(s )
randomly computed using a random generator of numbers ( with @xmath89 ) , degrees , so we are able to observe the outflow only from @xmath90 . ] and equation [ vout1e ] with @xmath87 , we produce the sub - sample set 4 .
it is easy to see that set 4 , resembles better the plot shown in figure 4 of @xcite , taking into account the possible incompleteness of the sample .
doing the same for the ufos , we use the luminosities produced by set 1 ( @xmath91 , this is the best - fit value for set 1 and set 3 , see section [ model1 ] ) but plotting ( open squares ) against 1000 @xmath86(s ) randomly computed using a random generator of numbers ( with @xmath89 ) , and equation [ vout1e ] with @xmath87 , to produce the sub - sample set 3 . in this case
set 3 , better resembles the relationship between @xmath52 and luminosity ( see figure [ vel2 ] ) . in figure [ vel2 ]
we place both classes of outflows together along with observational points taken from two samples of objects : 14 points out of the 23 objects reported by @xcite ( the others either did not report outflow velocity or were unknown ) , in the right panel ; and 19 points from the sample of 19 objects were ufos have been detected by @xcite , in the left panel .
there are several interesting facts about the plot : ( 1 ) _ the observational points _
cover @xmath92 orders of magnitude in velocity ; ( 2 ) they cover @xmath93 orders of magnitude in bolometric luminosity ( from @xmath94 erg / s ) ; and ( 3 ) that our proposed model is able to reproduce most of the low - velocity points ( 11/14 ) ( two of the points might fall in the high - velocity set instead ) and , less stringently , half of the points for the high - velocity set ( 11/19 ) , using one physical acceleration mechanism and three free parameters : mass of the bh ( @xmath54 ) , outflow velocity ( @xmath52 ) and force multiplier ( @xmath95 ) . on the other hand , there are two sectors where the deviations between model and theory are large : ( 1 ) the high - velocity / low luminosity ; and ( 2 ) the low - velocity / high luminosity , both requiring a closer inspection to the completeness of the samples , and/or the addition of other acceleration mechanisms , like magnetic thrust for instance .
but , this may be the topic of future work .
we propose here that the low- and high - velocity outflows can be accelerated by the same physical mechanism ( as is suggested by figure [ vel2 ] ) ; i.e. , radiation pressure , and that the differences depend on the values of the 3 fundamental parameters : mass of the bh ( @xmath54 ) , the object luminosity ( @xmath53 ) and force multiplier ( @xmath95 ) . _
the values of @xmath95 we have used are of the orders of those found in observational ( e.g. , * ? ? ?
* ) and theoretical ( e.g. , * ? ? ?
* ) works_. however , detailed photoionization computations are required to verify if the opacity of the gas under the physical conditions presented here can overcome the over - ionization problem @xcite , and are left for future work . the anisotropic property of the radiation is basically demanded by the existence of obscured and un - obscured ( type 1 and type 2 ) agns @xcite , and it explains the decreasing of type 2 agns as a function of the x - ray luminosity @xcite .
it is also intrinsically linked to the existence of the dusty torus1000 k ) optically and geometrically thick gas in approximate rotational and virial equilibrium at @xmath371 pc @xcite . ] and it gives the natural frame to locate our two classes of outflows .
if comparable with the escape velocity of the system ( a black hole of @xmath96 m@xmath28 ) , fast - outflow , @xmath97c , locates escape radius at @xmath98 ( schwarzschild radius , @xmath99 ) .
this is very close to the smbh , and the origin is likely the accretion disk . again , using @xmath100 , as escape velocity from a @xmath101 @xmath45 bh , locates the gas at @xmath102 pc , well beyond the event horizon of the bh , and the broad line region ( blr ; * ? ? ?
* ) as well , where the dusty torus is thought to be located .
for the first time , an exploration on the relationship between bolometric luminosity and outflow velocity , for two classes of x - ray outflows in a large sample of active galactic nuclei has been performed .
we find that : ( 1 ) line radiation pressure is an efficient mechanism to accelerate the low - velocity ( @xmath103 ) gas we observe in the classical @xmath21 warm absorber .
( 2 ) it might also become efficient to accelerate the high - velocity ( @xmath104 ) gas we observe in the ufos .
( 3 ) they both might be placed in the same context of anisotropic radiation pressure ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . however , there are many open questions , which will require close investigation and detailed modeling . in the first place
the fact that we are assigning one type of outflow to different portions of the parameter space ( @xmath54 , @xmath53 and @xmath95 ) , does not preclude the existence of both in the same object .
also , careful studies of the connection between these outflows observed in x - rays and outflows seen in other bands of the electromagnetic spectrum , uv , infrared or optical ( multi - wavelength studies ) are important points to be addressed .
finally the close inspection of the main parameters of the system ( @xmath105 and @xmath95 ) with cosmological parameters ( like redshift ) is of high relevance , and might be the topic of future works .
jmr would like to thank t. kallman for a reading of the manuscript .
he also wants to thank the useful and constructive comments from the referee which helped to improve several aspects of the work .
ft provided the data based on observations obtained with the xmm - newton satellite , an esa funded mission with contributions by esa member states and usa .
ft acknowledge support from nasa through the adap / ltsa program . this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
s. , brandt w. n. , george i. m. , netzer h. , crenshaw d. m. , gabel j. r. , hamann f. w. , kaiser m. e. , koratkar a. , kraemer s. b. , kriss g. a. , mathur s. , mushotzky r. f. , nandra k. , peterson b. m. , shields j. c. , turner t. j. , zheng w. , 2002 , , 574 , 643 | an exploration of the relationship between bolometric luminosity and outflow velocity , for two classes of x - ray outflows in a large sample of active galactic nuclei has been performed .
we find that line radiation pressure could be one physical mechanism that might accelerate the gas we observe in warm absorber , @xmath0 , and on comparable but less stringent grounds the ultra - fast outflows ( ufos ) , @xmath1 . if comparable with the escape velocity of the system ; the first is naturally located at distances of the dusty torus , @xmath2 pc , and the second at sub - parsec scales , @xmath3 pc , in accordance with large set of observational evidence existing in the literature
. the presentation of this relationship might give us key clues for our understanding of the different physical mechanisms acting in the center of galaxies , the feedback process and its impact on the evolution of the host galaxy .
[ firstpage ] black hole physics - x - ray : galaxies - galaxies : active |
the observation of upward - going muons in underground detectors has long since been considered @xcite as an effective tool to study the neutrino component expected in atmospheric showers @xcite .
early anomalies in the muon event samples detected by the first two dedicated experiments @xcite soon sparked investigations of possible neutrino oscillation effects @xcite .
the upward - going muon total rates measured in recent underground experiments appear to be approximately in agreement with the expectations , thus disfavoring large neutrino oscillation effects .
this contrasts with the observed anomaly in the electron and muon contents of ( semi)contained events induced by atmospheric neutrinos , that might signal flavor oscillation effects ( see @xcite for a recent review ) .
the conflict between these two data sets is usually resolved by appealing to the large , noncancelable uncertainties affecting the calculation of the _ absolute _ muon rates @xcite . in terms of neutrino oscillations
, this implies that there are values of the oscillation parameters that explain the `` anomalous '' muon - to - electron flavor ratio without really conflicting with the `` regular '' upward muon data @xcite . in this work
we confirm and elucidate such point of view by performing an updated analysis of the upward - going muon data from five experiments : kamiokande @xcite , baksan @xcite , macro @xcite , imb @xcite , and superkamiokande @xcite .
the paper is structured as follows . in sec .
ii the available data on upward - going muons are reviewed , compared with each other and with the theoretical expectations ( in the absence of oscillations ) , and then combined . in sec .
iii the combined data on the muon flux angular distribution are used to derive bounds on the two - flavor and three - flavor neutrino oscillation parameters .
these bounds are compatible with the neutrino oscillation solution to the atmospheric neutrino anomaly .
it is shown that the inclusion of upward - going muon data in fits tends to favor the @xmath0 oscillation channel and to disfavor the threefold maximal mixing scenario . in sec .
iv we draw the conclusions of our work .
this study is part of a vast research program aiming to analyze the world neutrino oscillation data , including the results of solar @xcite , atmospheric , @xcite and laboratory ( accelerator and reactor ) @xcite neutrino experiments .
the experimental data on upward - going muons are usually given as distributions of either muon events or muon fluxes as a function of the zenith angle @xmath2 , with @xmath3 @xmath4 corresponding to horizontal ( vertical ) muons .
the distribution of _ muon events _ depends upon technical specifications of the detector such as the efficiency and the geometric acceptance , that are often unpublished .
the distribution of _ muon fluxes _
( when available ) is instead deconvoluted for these effects , and thus allows a more direct comparison with the theoretical expectations .
we describe in sec .
ii a the more recent available data ( either muon events or muon fluxes ) and in sec .
ii b the specific combination of data used in our oscillation analysis . in the last few years
, data on upward going muons have been collected by five experiments : kamiokande @xcite , baksan @xcite , macro @xcite , imb @xcite , and superkamiokande @xcite .
the imb experiment ( completed ) was performed in three distinct stages ( phases 1 , 2 , and 3 ) ; we will comment both on the global results ( imb-1,2,3 ) @xcite and on the results from the first two phases ( imb-1,2 ) @xcite .
data from the pioneering nusex experiment @xcite ( eight events only ) will not be considered .
since some experimental results have been reported only in conference proceedings , we compile the most recent data on upward - going muons for the benefit of the reader .
the compilation is reported in table [ tab:1 ] and in fig .
[ fig:1 ] , that we comment in parallel .
table [ tab:1 ] reports , for each experiment , the relevant detector characteristics and the experimental muon data , compared with the corresponding expectations in ten angular bins .
the theoretical muon fluxes refer to our calculation with the following inputs : bartol neutrino fluxes @xcite ( neglecting small , site - dependent geomagnetic effects ) , version `` g '' of the martin - stirling - roberts [ mrs(g ) ] structure functions @xcite ( i.e. , the default choice of the cern parton density function library @xcite ) , and lohmann - kopp - voss ( lkv ) muon energy losses in standard rock @xcite .
the use of the same input for the calculation of the expected muon fluxes will enable us to perform a uniform comparison with all the experiments .
in the normalization of the expected muon fluxes . this uncertainty will be included in the fits . ] notice that , for fixed inputs , the theoretical muon spectrum depends only on the energy threshold ; therefore , the baksan and macro experiments ( both with @xmath5 gev ) share the same expectations .
all other experiments have higher thresholds gev for phases 1 and 2 , to @xmath6 gev for phase 3 , and to an effective value @xmath7 gev for the total data sample ( imb-1,2,3 ) . ] and correspondingly lower muon fluxes . concerning the experimental data , we quote only the statistical errors , since the systematic errors in each bin are usually unpublished .
the only notable exception is the macro analysis @xcite , where the systematic uncertainties are estimated to range from a few percent near the vertical to about @xmath8 near the horizontal direction .
we also report ( last row in table [ tab:1 ] ) the total muon flux , defined as the sum of the individual fluxes in each bin times the bin width , with ( statistical ) errors added in quadrature .
figure [ fig:1 ] displays , in graphical form , the same information as table [ tab:1 ] .
the data are shown as dots with ( statistical ) error bars .
our calculations ( in the absence of oscillations ) are represented by the solid histograms .
in addition , we report as dashed histograms the original montecarlo simulations of the various collaborations .
the ingredients of these simulations were : ( 1 ) kamiokande @xcite : volkova neutrino fluxes @xcite , eichten - hinchliffe - lane - quigg ( ehlq ) structure functions @xcite , and lohmann - kopp - voss ( lkv ) muon energy losses @xcite ; ( 2 ) baksan @xcite : bartol neutrino fluxes @xcite , morfin - tung ( mt ) structure functions @xcite , and lkv muon energy losses ; ( 3 ) macro @xcite : bartol neutrino fluxes , mt structure functions , and lkv muon energy losses ; ( 4 ) imb @xcite : volkova neutrino fluxes , ehlq structure functions , and bezrukov - bugaev muon energy losses @xcite ; ( 5 ) superkamiokande @xcite : bartol neutrino fluxes , ehlq structure functions , and lkv muon energy losses .
the differences between our calculations and the montecarlo simulations are relatively small .
the last panel in fig . [ fig:1 ] shows the total number of muons that stopped in , or passed through , the imb detector ( phases 1 , 2 , and 3 combined ) .
unfortunately , it is difficult to calculate reliably the corresponding theoretical rates without detailed information about the imb detector .
the very definition of `` passing '' or `` stopping '' muon depends on unreproducible geometric cuts applied on an event - by - event basis .
their approximation with ( more manageable ) energy cuts ( see , e.g. , @xcite ) may bias the classification of events .
moreover , the absolute estimates of muon yields requires also the accurate knowledge of the detector geometrical acceptance and of its energy - dependent efficiency , which are not published .
therefore , we are not surprised that , just for imb-1,2,3 ( last panel of fig .
[ fig:1 ] ) , our tentative calculation of the `` passing '' muon yield is not in good agreement with the published imb montecarlo simulation @xcite . for these reasons ,
we have not used the imb-1,2,3 stopping and passing muon data in our analysis .
we have instead used the muon flux distribution ( published for imb-1,2 @xcite ) , that can be calculated more reliably and can also be compared with similar distributions from other experiments . as fig .
[ fig:1 ] shows , the angular distributions measured by kamiokande , baksan , macro , imb-1,2 , and superkamiokande are roughly in agreement with the theoretical expectations .
however , the data also show some `` bumps '' or `` dips '' that seem too large to be statistical fluctuations but also too random to suggest new physics
they are likely to be the effect of unknown experimental systematics .
let us consider , e.g. , the macro and baksan experimental spectra , that , as observed in the previous subsection , are characterized by the same threshold ( @xmath5 gev ) and thus can be directly compared by superposition .
both the kolmogorov - smirnov and the @xmath9 tests give a @xmath10 probability that these two distributions are the same within the _ statistical _ errors .
this suggests that the experimental systematics might be nonnegligible ( in some bins , at least ) , as indeed confirmed by the macro error analysis @xcite .
unfortunately , no analogous error analysis has been reported by the baksan collaboration .
the presence of relatively large `` dips and bumps '' in the individual experimental spectra of fig .
[ fig:1 ] poses some practical problems in fitting the data with theoretical curves ( that , conversely , are rather smooth even in the presence of neutrino oscillations ) .
in fact , the oscillation fits to the individual spectra become dominated by the bins that exhibit the largest fluctuations , thus producing rather unstable and unreliable results .
therefore , it seems more reasonable to combine first the data from different experiments , in order to `` average out '' their fluctuations , and then to fit the neutrino oscillation parameters .
we have combined the five muon flux spectra of fig .
[ fig:1 ] into a single spectrum as follows .
first , the experimental distributions have been `` rescaled '' to a common 1 gev threshold , by adding in each bin the theoretical contribution due to the change in threshold ( no correction is necessary for macro and baksan ) .
the results are shown in fig .
[ fig:2](a ) , together with the theoretical expectation ( solid histogram ) , now common to all experiments.(a ) is significant ( a factor of about two in the muon fluxes ) , suggesting again and independently that large and not well - understood systematics might affect ( at least some of ) the experimental results . ]
then the data have been combined in each bin , using only the known , statistical errors .
the resulting , combined distribution is shown in fig .
[ fig:2](b ) ( dots with small error bars ) : it appears smoother than each of the individual parent distributions , and thus more appropriate for fits .
a delicate point is the estimate of the ( nonnegligible but unknown ) systematic errors .
the particle data group recipe for scaling data errors as @xmath11 @xcite would give scaling factors between 1 and 2 in the various bins .
similarly , the addition of plausible , @xmath12 systematic errors would give total errors @xmath13 times larger than the statistical errors .
guided by these estimates , we have chosen to rescale the statistical errors by a uniform factor 1.5 [ large error bars in fig .
[ fig:2](b ) ] in order to obtain the `` total '' uncertainties .
these `` inflated '' experimental errors have been used in all the following fits .
we stress that any conceivable combination of the _ present _ muon data necessarily involves some approximations or subjective choices , both because spectra with different thresholds can not be compared directly and because no experiment ( excepted macro ) reports the systematic uncertainties in each bin .
for instance , the ( somewhat arbitrary ) choice of a common threshold for rescaling the experimental spectra has some influence on the sensitivity of the fits to low values of the neutrino mass differences .
therefore , great care should be taken in interpreting the results of _ any _ fit to the available muon data , including those presented in the following section .
hopefully , in the next few years the superkamiokande experiment will collect much higher muon statistics and will carefully estimate the systematic uncertainties of the flux distribution , so as to render such drawbacks gradually less significant in future analyses .
in this section we report the results of our two - flavor and three - flavor oscillation analysis of the upward muon spectrum shown in fig .
[ fig:2](b ) .
since this combined spectrum is in reasonable agreement with the expectations in the absence of oscillations , exclusion zones can be drawn in the neutrino oscillation parameter space .
this contrasts with the flavor anomaly observed in contained and semicontained atmospheric neutrino events , which suggests nonzero neutrino masses and mixings @xcite .
however , as we will see , positive and negative indications are compatible , within the experimental and theoretical uncertainties , in some zones of the parameter space where the neutrino mixing is `` intermediate '' ( i.e. , neither zero nor maximal ) .
we use the three - flavor oscillation formalism defined in our previous works @xcite .
the first two of the three neutrino mass eigenstates @xmath14 are assumed to be essentially degenerate @xmath15 , and separated from the third state @xmath16 by a square mass gap @xmath17 .
the flavor content of @xmath16 is parametrized as @xmath18 where the mixing angles @xmath19 and @xmath20 range between 0 and @xmath21 .
the cases @xmath22 , @xmath23 , and @xmath24 correspond to pure two - flavor oscillations in the channels @xmath1 , @xmath25 , and @xmath26 , respectively ( the latter case @xmath26 is uninteresting for atmospheric neutrinos and will not be specifically considered ) . as shown in @xcite , it is useful to chart the parameter space through the logarithmic coordinates @xmath27 .
our results will be presented in representative sections of this space . for each value of the neutrino mass and
mixing parameters , we calculate the theoretical muon flux spectrum with the inputs described in the caption to table [ tab:1 ] , including the matter oscillation effect for neutrino paths in the earth @xcite
. then we evaluate the goodness - of - fit to the experimental distribution in fig .
[ fig:2](b ) through a @xmath9 test .
the @xmath9 quadratic form includes both the uncorrelated total experimental errors and the correlated theoretical uncertainties of the absolute neutrino fluxes , cross sections , and muon energy losses in the rock . the @xmath28 total theoretical error is conservatively assumed to be @xmath29 , with a bin - by - bin correlation coefficient equal to @xmath30 . for both two - flavor and three - flavor oscillations the minimum value of @xmath9 is searched , and c.l .
contours are drawn at the values of @xmath31 appropriate to @xmath32 and @xmath33 , respectively ( see also @xcite ) .
we will combine in different ways the following data sets , in order to show their relative `` weight '' in driving the fit : upward - going muons ( u ) from kamiokande , imb-1,2 , baksan , macro , and superkamiokande ; sub - gev ( s ) @xmath34-like and @xmath35-like events from kamiokande , imb , nusex , and frejus @xcite ; multi - gev ( m ) @xmath34-like and @xmath35-like events from kamiokande @xcite , both binned ( mb ) and unbinned ( mu ) @xcite .
the data sets s , mu , and mb are analyzed as in @xcite .
figure [ fig:3 ] shows the results of our analysis of @xmath1 oscillations . in the first panel ( u )
only the upward - going muon data are fitted .
the symmetry of the c.l .
contours with respect to the vertical axis @xmath36 reflects the absence of matter effects in the @xmath1 channel @xcite . as expected , large neutrino mixing is excluded ( for @xmath37 ) .
this negative indication , however , does not rule out the evidence for oscillations coming from sub - gev and multi - gev binned or unbinned data ( s+mb and s+mu panels ) .
in fact , the combination of all atmospheric data still favors the oscillation hypothesis , as evidenced in the plots labeled `` s+mb+u '' ( sub - gev + multi - gev binned + up - going @xmath34 ) and `` s+mu+u '' ( sub - gev + multi - gev unbinned + up - going @xmath34 ) .
notice that the addition of the upward - going muon data reinforces the exclusion of large @xmath1 mixing ( required by the negative results of the nusex and frejus experiments @xcite ) , but does not alter dramatically the oscillation fit to sub - gev and multi - gev atmospheric neutrino data . in particular , for large @xmath38 the `` s+mu+u '' bounds on @xmath39 are only slightly narrower than those without upward @xmath34 data ( `` s+mu '' panel ) .
this observation is relevant for those phenomenological models in which the controversial zenithal anomaly of multi - gev data is discarded , such as the `` minimum sacrifice '' scenario studied in @xcite .
we have verified that the solution to the neutrino oscillation evidence proposed in such scenario survives after the inclusion of the upward - going muon data , although with a slightly worse @xmath9 .
figure [ fig:4 ] shows the results of our analysis of @xmath25 oscillations .
the panels are labeled as in fig .
[ fig:3 ] .
notice that the bounds from upward - going muon data in the @xmath25 channel are considerably weaker than in the @xmath1 channel , and basically exclude a relatively small zone at 90% c.l . for large mixing and @xmath40@xmath41 ev@xmath42 ( no constraints appear at 99% c.l . ) .
in fact , in the @xmath25 channel the @xmath43 disappearance is partially compensated by the conversion of @xmath44 ( initially present in the beam ) into @xmath43 .
the `` suppression '' of the muon disappearance is also influenced by matter effects , that are responsible of the left - right asymmetry in the plots ( see @xcite ) .
also in this case , the bounds from sub - gev and multi - gev data ( `` s+mu '' and `` s+mb '' panels ) are only slightly altered by the inclusion of upward - going muon data ( `` s+mu+u '' and `` s+mb+u '' panels , respectively ) . in conclusion ,
the upward - going muon data tend to disfavor large mixing in the @xmath1 channel and , more weakly , in the @xmath0 channel .
this indication reinforces the exclusion of large mixing required by the negative sub - gev results from nusex and frejus .
however , the indications for `` intermediate '' mixing derived by a global fit to the atmospheric data are not dramatically altered by the inclusion of the upward - going muon data . in particular ,
the @xmath1 oscillation mode is still compatible with the `` minimum sacrifice scenario '' of refs .
@xcite , which is thus able to accommodate all present neutrino oscillation data , with the only exception of the non - flat zenith angle distribution measured in the kamiokande multi - gev experiment @xcite . figure [ fig:5 ] reports the results of our analysis of upward - going muon data in the three - flavor space @xmath27 , shown as planar sections at representative values of @xmath38 .
we recall that , in each @xmath45 panel , the right side corresponds ( asymptotically ) to pure @xmath0 oscillations , while the lower side correspond ( asymptotically ) to pure @xmath1 oscillations . the excluded regions in fig . [ fig:5 ]
connect continuously these two subcases and show that also genuine three - flavor cases are disfavored by the available upward - going muon data .
these cases correspond to large neutrino mixings . in particular , the threefold maximal mixing scenario @xcite , obtained for @xmath46 ,
is rather strongly disfavored by the upward - going muon data .
the exact shapes of the excluded regions should be taken with some caution , since we have checked that they are rather sensitive to variations in the data and in their statistical treatment ; therefore , new and more precise upward - going muon data from the superkamiokande experiment might alter significantly the bounds shown in fig .
[ fig:5 ] . however , the following qualitative features seem to be relatively stable : 1 ) weaker bounds in the @xmath0 channel ; and 2 ) exclusion of large neutrino mixing and , in particular , of three - fold maximal mixing .
figure [ fig:6 ] shows the results of a combined analysis of all the available sub - gev , multi - gev ( binned ) , and upward - going muon data . as for fig .
[ fig:5 ] we recall that , because of the quality of the available upward - going muon data , these fits might change significantly with new , more accurate data .
nevertheless , some interesting , qualitative features are seen to emerge : 1 ) in general , @xmath25 oscillations tend to give a better fit than @xmath1 oscillations , as a result of the combined preference of multi - gev data @xcite and of weaker bounds from upward - going muon data for @xmath25 transitions ; and 2 ) the overall oscillation fit is reasonably good even for @xmath38 as small as @xmath47 ev@xmath42 .
these indications suggest that oscillation signals might be seen better in long - baseline _ reactor _ experiments , that can probe the @xmath44 disappearance mode down to @xmath48 ev@xmath42 .
if no oscillation signal is found in such experiments , large horizontal stripes will be excluded in the panels of figs .
[ fig:5 ] and [ fig:6 ] @xcite . in conclusion , both the positive and the negative indications for flavor oscillations coming from all the available atmospheric neutrino data ( including upward - going muons ) can be made compatible , within the uncertainties , at intermediate values of the neutrino mixing angles .
in fact , although the positive indications for @xmath49 oscillations favor large mixing , the negative results from upward - going muon experiments ( as well as from some sub - gev experiments like frejus and nusex ) , tend to drag the fit towards small neutrino mixings and to stabilize it at intermediate values . of course
, the overall compatibility of the different data sets might be no longer guaranteed if the total uncertainties were smaller .
therefore , it is very important that the magnitude of the experimental systematic errors in the muon flux distribution ( see sec .
ii b ) is clarified in the running underground experiments ( macro , baksan , and superkamiokande ) .
we have performed an updated analysis of the available upward - going muon data ( which disfavor neutrino oscillations ) , and a comparison with the atmospheric sub - gev and multi - gev data ( which globally favor neutrino oscillations ) .
we have worked out the oscillation constraints placed by the observed upward - going muon zenithal distribution , and identified the zones of the two- and three - flavor mass - mixing parameter space where all the atmospheric data can be reconciled . in general , the @xmath0 oscillation channel provides a better fit than the @xmath1 channel , thus privileging searches with long - baseline reactors .
large neutrino mixing , including threefold maximal mixing , is generally disfavored .
the `` minimum sacrifice '' fit of refs .
@xcite , that discards , in part , the atmospheric multi - gev data , is not significantly altered by the inclusion of the upward - going muon information .
we acknowledge fruitful discussions with t. montaruli .
we thank s. mikheyev and y. totsuka for useful correspondence . the work of a.m.
was supported in part by ministero delluniversit e della ricerca scientifica and in part by infn . .
detector characteristics , experimental data , and theoretical expectations for upward - going muons .
theoretical expectations refer to our calculations using bartol neutrino fluxes @xcite , mrs(g ) structure functions @xcite , and lohmann - kopp - voss muon energy losses in standard rock @xcite .
see the text for details and comments . [
cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] m. a. markov and i. m. zheleznykh , nucl .
* 27 * , 385 ( 1961 ) ; g. t. zatsepin and v. a. kuzmin , zh . eksp .
fiz . * 41 * , 1818 ( 1961 ) [ sov .
jetp * 14 * , 1294 ( 1962 ) ] ; see also : k. greisen , ann .
science * 10 * , 63 ( 1960 ) .
t. stanev , in _
taup 95 _ , proceedings of the 4th international workshop on theoretical and phenomenological aspects of underground physics , toledo , spain , edited by a. morales , j. morales , and j. a. villar [ nucl .
b ( proc . suppl . )
* 48 * , 165 ( 1996 ) ] .
kamiokande collaboration , a. suzuki _ et al .
_ in the _ proceedings of 7th international workshop on neutrino telescopes _ , venice , 1996 , edited by m. baldo ceolin ( university of padova , italy , 1996 ) p. 263 .
m. m. boliev , a. v. butkevich , a. e. chudakov , s. p. mikheyev , o. v. suvorova , and v. n. zakidyshev , in _
icrc 95 _ , proceedings of the 24th international cosmic ray conference , rome ( arti grafiche editoriali , urbino , italy , 1995 ) vol 1 , p. 686 ; _ ibidem _ , p. 722 .
m. m. boliev , e. v. bugaev , a. v. butkevich , a. e. chudakov , s. p. mikheyev , o. v. suvorova , and v. n. zakidyshev , in _
dark 96 _ , proceedings of the international workshop on aspects of dark matter in astrophysics and particle physics , heidelberg , germany , 1996 , to appear .
superkamiokande collaboration , k. young _ et al .
_ , in the _ joint aps / aapt 1997 meeting _ , washington dc , april 1997 , to appear in the proceedings ; also available at the url hhp://www.phys.washington.edu/@xmath50young / superk / drafts / aps97.html .
fig . 1 .
compilation of upward - going muon data and expectations for different experiments : kamiokande @xcite , baksan @xcite , macro @xcite , superkamiokande @xcite imb-1,2 @xcite , and imb-1,2,3 @xcite .
dots with error bars : experimental data with @xmath28 statistical uncertainties .
solid histograms : our calculations . dashed histograms : experimental simulations .
@xmath2 is the zenith angle .
see also table i. fig . 2 .
comparison of all upward - going muon spectra , rescaled to a common threshold of 1 gev .
( a ) separated data .
( b ) combined data ( as used in our analysis ) .
see the text for details .
fig . 3 .
results of the oscillation fit in the two - flavor oscillation channel @xmath1 , shown as c.l .
contours in the @xmath51 plane .
the allowed regions are marked by stars .
results of the oscillation fit in the two - flavor oscillation channel @xmath0 , shown as c.l .
contours in the @xmath52 plane .
the allowed regions are marked by stars .
fig . 5 . bounds placed by all upward - going muon data in the three - flavor oscillation space @xmath27 .
the allowed regions are marked by stars .
6 . as in fig .
[ fig:5 ] , but with the addition of the atmospheric sub - gev ( s ) and multi - gev binned ( mb ) data ( analyzed as in @xcite ) . | the available upward - going muon data from the kamiokande , baksan , macro , imb , and superkamiokande experiments are reviewed and combined .
bounds on the neutrino mass and mixing parameters are derived for oscillations in two and three flavors .
these bounds are not in significant conflict with the oscillation solution to the atmospheric neutrino flavor anomaly observed in the sub - gev and multi - gev energy range .
the combination of all the available atmospheric data tends to favor the @xmath0 channel with respect to the @xmath1 channel , and to disfavor the threefold maximal mixing scenario . |
cellular automata ( ca ) micro - simulation has emerged as a tool for simulating traffic flow and modelling transport networks @xcite . in ca , time and space are discrete .
the space is represented as a uniform lattice of cells with finite number of states , subject to a uniform set of rules , which drives the behavior of the system .
these rules compute the state of a particular cell as a function of its previous state and the state of the neighboring cells .
agents moving from origins to destinations across networks may represent several real entities , as for example : ants , biological organisms , small robots , transport in micro - mechanical systems , crowd flow , packets transport in the internet , etc .... it was found that the motion of the biological organisms is usually controlled by interactions with other organisms in their neighborhood and randomness also plays an important role @xcite .
real ants have been shown to be able to find shortest paths towards destinations using as only information the pheromone trail deposited by other ants @xcite .
the finding of shortest paths , inspired from ants behavior , are successfully applied to several physical problems such as , pedestrians @xcite , traffic flow @xcite , combinatorial optimization @xcite and circuit switched communications network problems @xcite .
the problems of movement of agents with origins and destinations were studied using a two dimensional cellular automata @xcite . in these models
an agent tries to reach its destination using simple rules .
transitions from the freely moving to the jamming states were studied .
a variant 2-dimensional ca model for simulation of agents moving from origins to destinations will be presented here .
agents moving across the network have sensors to perceive their local neighborhood and their destinations and then affect their environment .
this is done especially by estimating the distance metric to the destination site .
the concern herein will be with the movement , propagation , and interaction of agents in low and high density situations .
this will be done by exploring the patterns and behaviors of the spatio - temporal organization of agents .
the objective of this research is to provide insight into modelling complex dynamics using ca microsimulation and capturing general features of agents travelling from origins to destinations .
the paper is organized as follows . in sec .
@xmath3 , we describe our model for movements of agents with origins and destinations . in sec .
@xmath4 , we present our numerical results where we give the phase diagrams of the system . a detailed description of the cluster size and the travel time distributions are also presented .
finally , we conclude with some conclusions in sec . @xmath5 .
the ca model is a two - dimensional cellular automaton based on a square lattice with periodic boundary conditions .
there is a fixed number of agents on the lattice at all times .
only one agent at most can occupy a given site . at any time step ,
an agent can move at most to one of its 4 neighboring sites .
updating of the ca occurs in parallel where the rules are applied to all agents at the same time .
agents are associated with given origin - destination sites .
the origin and destination sites must be different .
an agent travels from the origin site towards its destination site , whereupon it disappears .
each disappeared agent is immediately replaced by a new agent , and so the agent number is always constant in the lattice .
a new origin - destination pair is then chosen randomly for this new agent .
if , however , there is an agent already present at the chosen origin , then another origin site is selected .
agents will move towards their destinations at all times by selecting an unoccupied neighboring site which has the minimal distance from that site to the destination site ( see figure 1 ) .
an agent examines the unoccupied neighboring sites . for each of these sites
a distance to the destination is evaluated .
then , a site with the minimal distance is selected as the next site to which the agent will move .
if all neighboring sites are occupied it will not move .
the stochasticity or noise is introduced in the model dynamics , through the uncertainty in estimating the distance from the destination .
so , with probability @xmath0 an agent moves towards an arbitrary empty neighboring site rather than the site of minimal distance .
the friction parameter @xmath2 is also introduced to control the probability that the movement of all agents involved to the same site ( conflict ) is denied at one time step .
this friction parameter which is essential for resolving the conflict arising in parallel update simulations , is applied for pedestrian traffic problems @xcite . in each time step ,
positions , speeds and directions of all agents are updated according to the following local rules : - with probability @xmath0 , an agent selects one arbitrary empty neighboring site
. - with probability @xmath6 agent selects an empty neighboring site corresponding to the minimal distance to the destination .
if two empty neighboring sites of one agent have the same minimal distance from the destination then one of these two allowed neighbors is chosen randomly . if two or more agents select the same site ( conflicts ) then : - with probability @xmath2 none of the agents
is allowed to move to their selected site . - with probability @xmath7 one of these agents is chosen randomly to move to its selected site ; the others agents do not move . if there exist no conflict , the agent moves to its selected site .
if all neighboring sites are occupied , the agent does not move .
we carry out our computer simulations of the model by considering a square lattice of size @xmath8 with periodic boundary conditions .
initially , we put randomly a number @xmath9 of agents into the lattice .
the density of agents is denoted as @xmath10 .
the velocity of each agent can be either 1 or 0 .
the duration of each simulation run is @xmath11 time steps with the first @xmath12 time steps to initiate the simulation and the latter @xmath13 used to generate performance statistics .
agents are only allowed to move to unoccupied nearest neighbor sites in one time step , i.e. @xmath14 cell / time step . in figure 2
, we carried out the plots of the mean velocity of agents as a function of the density , for several system size .
hence , the plots show that @xmath15 undergoes a sudden phase transition from a freely moving state @xmath16 to jammed state @xmath17 at a critical density @xmath18 . in the freely moving state ,
interaction between agents is weak and the propagation is important inside the network . in contrast , for large density , the interaction becomes strong and jamming takes place where agent movements become rare . as regards the variation with system size @xmath8 , we find that the critical density decreases with increasing @xmath8 and the phase transition becomes sharper . in figure 3
, we plot @xmath19 as a function of @xmath2 for different values of @xmath0 and @xmath20 . hence , for low density , the average speed remains almost constant if the probability @xmath0 is high .
however , if @xmath0 is low , @xmath15 undergoes a sudden decrease when @xmath2 exceeds a critical value @xmath21 .
this corresponds to a phase transition from the freely moving state to the jamming state .
consequently , the enhancement of the friction parameter can topple over from the freely moving to the jamming of agents even at low densities . at high densities , @xmath15 decreases gradually with @xmath2 for all values of @xmath0 . yet
, the speed @xmath15 remains always greater for larger @xmath0 .
the phase diagrams of the system is depicted in figure 4 , where we plotted the critical values @xmath21 as a function of @xmath0 for several fixed values of @xmath20 .
thus , for low densities and for a given value of @xmath0 , freely moving phase should exist if @xmath22 while jamming phase takes place if @xmath23 .
it was shown also from figure 4 that the jamming region is broaden as soon as the density is increased . yet ,
when the density exceeds some value @xmath24 , the freely moving phase should never exist .
it is clear that the density dependence of speed alone , can not give the whole information on the phase behavior of the system . to get more information on the microscopic structure of the phases
, one can determine the spatio - temporal organization of agents in the lattice .
this microscopic investigations can be obtained by plotting the organization patterns of agents and the distributions of cluster sizes and travel times . the cluster and the cluster size
mean here a connected bonds of unoccupied cells and a maximally number connected cells of agents respectively .
the travel time is the time it would take to travel from the origin to destination .
figures 5(a - b ) show a typical configurations of the organization of agents at low density .
so , for vanished value of @xmath0 and for low values of @xmath2 , the steady state corresponds to the freely moving .
however if @xmath2 is high , agents self - organize in a large cluster with few freely moving agents at the boundary . for higher values of @xmath0
, the freely moving phase should exist even for larger @xmath2 .
hence , it seems that the role of @xmath2 is to pile up the agents into one large cluster while @xmath0 tends to dispatch them in all directions . for high densities and low @xmath0
, agents pile up into one large cluster even if @xmath2 is vanished ( fig .
5c ) . in the other side , this agglomeration splits up into several clusters when @xmath0 becomes large ( fig .
5d ) .
the cluster size distributions of the model are given in figures 6(a - b ) . at low density and for vanished value of @xmath0 ,
only small clusters are present in the lattice whenever @xmath2 is low .
this is one characteristic of the freely moving phase . from results depicted in figure 6a , we observe the bimodal nature of the cluster size distribution as @xmath2 increases .
large clusters appear in the lattice but there are by far many more small - sized clusters than larger ones .
furthermore , with increasing @xmath2 , the probability of small clusters diminished while that of large cluster increases .
another important result is the discontinuity observed from the probability distribution when @xmath2 becomes very large .
as it was shown from figure 5b , almost all agents are congested in one large cluster with the exception of a few agents which are located at the boundary and moving towards their destinations . from figure 6a ( @xmath25 and @xmath26 )
, one can see a phase transition from the freely moving phase to the jamming phase , occurring at @xmath27 .
indeed , when @xmath28 , the cluster size distribution is a continuous function ; but it becomes discontinuous when @xmath29 .
this value of @xmath21 agree with that found in the phase diagram ( fig .
in the other hand , when @xmath0 is large enough , we see that large clusters can not exist even for large @xmath2 .
it is important to note also that the increase of @xmath0 increases the probability of small - sized clusters ( see figure 6b , higher part ) . at high densities
, it is not surprising that large clusters may be usually present in the lattice .
however , their sizes depend greatly to the parameter @xmath0 . for example , the most probable cluster size ( of large clusters ) is shifted towards the low size region ( fig .
6b , lower part ) .
in contrast to @xmath0 , the effect of @xmath2 on the cluster size distribution is negligible ( result not shown ) .
if the density is high , only a small value of @xmath2 may provoke a strong congestion of agents .
the second quantity we look at is the travel time , i.e. the time an agent needs to travel from its origin to its destination .
the probability distribution of the travel time presents a maximum which is considered as the most probable time , at which an agent finished its travel . at low density and for low @xmath0
, the distribution is sharply peaked around its maximum , whenever @xmath2 is zero ( fig .
if @xmath0 increases , one finds evidently a broad distribution of travel time , because agents are moving for quite long times before reaching their destinations .
the higher is @xmath0 , the higher is the travel time . as regards the variation of @xmath2 , figure 7b illustrates some graphs of the travel time distributions for low densities and when we set @xmath26 .
the interesting observed phenomena is the existence of a double asymptotic behavior .
so , when @xmath2 is not large enough , we see that infinite travel time can not exist .
in contrast , for large @xmath2 , some agent may take an infinite time to reach its destination . in this case , the agents situated in the interior of the large cluster do not move and rest inside for an indefinite time . when @xmath0 is large enough and the density is low , the travel time distribution changes slightly if @xmath2 is changed ( fig .
furthermore , we see that infinite travel time does not occur even for very large @xmath2 .
this shows again that the state is the freely moving one at low density and for large @xmath0 . for higher densities ,
the short travel times still remain , showing the presence of some agents situated at the boundary of a large cluster but the asymptotic behavior is rather increasingly wide ; reflecting the dynamics of agents inside the large clusters ( see figure 7a , lower part ) . indeed , in the presence of large clusters one has to distinguish between inner and outer regions of the cluster .
inside , one finds evidently a broad distribution of travel times , because agents are blocked for quite long times . in the outer region of a cluster , however , one finds shorter travel times .
furthermore , we find that infinite travel time exist even for large @xmath0 , because the capacity of the freely moving is reduced when there was a big crowd of agents ( congestion ) .
in summary , we have tried to identify the behavioral aspects of agents travelling from origins to destinations .
the microscopic ca model presented is capable of capturing the self - organization and complex dynamics of agents .
the model contains two parameters ( @xmath2,@xmath0 ) and displays two states ; namely the freely moving and the jamming state .
the agents speed @xmath15 depends strongly to the parameters ( @xmath2,@xmath0 )
. for low densities , phase transitions occur as the friction parameter @xmath2 exceeds a critical value @xmath21 , which depends on both the density @xmath20 and the noise @xmath0 . in the other hand
, it was found that the effect of @xmath2 is to gather different agents into a large cluster .
this leads to jamming even at low density .
however , the effect of @xmath0 is to disperse agents through the lattice .
thus , more mobility and fluidity will affect the whole system . when distinguished for different density ranges , cluster size and travel time distributions have interesting properties .
indeed , if @xmath20 and @xmath0 are low , a transition from the freely moving to the jamming states can occur at a critical value @xmath21 .
thus , when @xmath28 agents self - organized in small clusters and only short travel times can be taken by agents .
however , in the other side ( @xmath29 ) , agents self - organize in one large cluster with very small number of moving agents at the boundary .
this is the jamming state where the travel time can be either short and long .
so , short travel times concern agents at the boundary of the large cluster while the long times concern those in the inner of the cluster . at high densities , jamming occurs for all values of ( @xmath2,@xmath0 ) .
this is due to the reduction of the capacity of a freely moving when there was a big crowd of agents in the lattice . as a result ,
the speed show a drastic decrease with the density .
the cluster size distribution becomes a bimodal distribution which represents a coexistence of large clusters and small ones .
yet , the travel time distribution is much broad where infinite travel time exists for all values of the system parameters ( @xmath2,@xmath0 ) . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * fig.1*. illustration of agent movements in a square lattice with periodic boundary conditions .
circles with arrows represent agents while those without arrows represent their destination sites .
each arrow indicates the selected site that agent will choose ( here , we set @xmath26 ) .
the conflict situation is occurred for the `` white '' and the `` gray '' agents since they select the same site .
* fig.2*. diagrams of agents speed versus density for different values of lattice size @xmath8 .
* fig.3*. diagrams of agents speed versus @xmath2 for different values of @xmath20 and @xmath0 , ( @xmath30 ) . *
fig.4*. phase diagrams of the system for several fixed values of the density : squares , open circles , triangles and solid circles represent @xmath31 , @xmath32 , @xmath33 and @xmath34 respectively , ( @xmath30 ) .
* fig.5*. self - organization patterns of agents . black squares represent agents .
a ) @xmath35 , @xmath36 and @xmath25 , b ) @xmath37 , @xmath36 and @xmath25 , c ) @xmath38 , @xmath39 and @xmath40 , d ) @xmath38 , @xmath41 and @xmath40 . * fig.6*. cluster size distributions for several values of the system parameters , ( @xmath30 ) . * fig.7*. travel time distributions for several values of the system parameters , ( @xmath30 ) .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ | we develop a two - dimensional cellular automaton ( ca ) as a simple model for agents moving from origins to destinations .
each agent moves towards an empty neighbor site corresponding to the minimal distance to its destination .
the stochasticity or noise ( @xmath0 ) is introduced in the model dynamics , through the uncertainty in estimating the distance from the destination .
the friction parameter @xmath1 is also introduced to control the probability that the movement of all agents involved to the same site ( conflict ) is denied at one time step .
this model displays two states ; namely the freely moving and the jamming state .
if @xmath2 is large and @xmath0 is low , the system is in the jamming state even if the density is low .
however , if @xmath2 is large and @xmath0 is high , a freely moving state takes place whenever the density is low . the cluster size and the travel time distributions in the two states
are studied in detail .
we find that only very small clusters are present in the freely moving state while the jamming state displays a bimodal distribution . at low densities
, agents can take a very long time to reach their destinations if @xmath2 is large and @xmath0 is low ( jamming state ) ; but long travel times are suppressed if @xmath0 becomes large ( freely moving state ) .
pacs numbers : 45.70.vn , 02.50.ey , 05.40.-a _ keywords : _ traffic ; freely moving ; jamming ; clustering ; travel times . |
consider an advection diffusion equation of the form @xmath11 where @xmath0 is the non - dimensional strength of a prescribed vector field @xmath12 . under reasonable assumptions when @xmath4 , the solution @xmath13 becomes constant on the trajectories of @xmath14 . indeed , dividing by @xmath0 and passing to the limit @xmath15 formally shows @xmath16 which , of course , forces @xmath13 to be constant along trajectories of @xmath14 .
well known `` averaging '' results @xcite study the slow evolution of @xmath17 across various trajectories . on the other hand , if we fix @xmath18 , classical homogenization results @xcite determine the long time behavior of solutions of . for such results
it is usually convenient to choose @xmath19 small , and rescale to time scales of order @xmath20 , and distance scales of order @xmath21 .
this gives @xmath22 assuming @xmath14 is periodic , and that the initial condition varies slowly ( i.e. @xmath23 is independent of @xmath24 ) , standard homogenization results show that @xmath25 , as @xmath26 .
further , @xmath27 is the solution of the effective problem @xmath28 and @xmath29 is the effective diffusion matrix , which can be computed as follows . define the correctors @xmath30 , , @xmath31 to be the mean - zero periodic solutions of @xmath32 then @xmath33 @xmath34 is the period cell of the flow @xmath12 , and @xmath35 is the kronecker delta function . the main focus of this paper is to study a transition between the two well known regimes described above . to this end ,
rescale by choosing time scales of the order @xmath20 and length scales of order @xmath21 .
this gives @xmath36 where @xmath37 and @xmath19 are two _ independent _ parameters .
of course , if we keep @xmath24 fixed , and send @xmath4 , the well known averaging results apply . alternately , if we keep @xmath0 fixed and send @xmath26 , we are in the regime of standard homogenization results .
the present paper considers with _ both _ @xmath26 and @xmath4 .
our main result shows that if @xmath14 is a @xmath38 cellular flow , then we see a sharp transition between the homogenization and averaging regimes at @xmath39 . before stating our precise results ( theorems [ thmevalstrongflow ] and [ thmevalhomog ] below )
, we provide a brief explanation as to why one expects the transition to occur at @xmath39 . for simplicity and concreteness
, we choose the stream function @xmath40 , and define @xmath41 . even in this simple setting , to the best of our knowledge , the transition from averaging to homogenization has not been studied before .
first , for any _ fixed _
@xmath0 , we let @xmath42 denote the effective diffusion matrix obtained in the limit @xmath26 ( see @xcite for a comprehensive review ) . if @xmath43 is the mean zero , @xmath44-periodic solution to @xmath45 then the effective diffusivity ( as a function of @xmath0 ) is given by . as @xmath4 , the behaviour of the correctors @xmath43 is well understood @xcite . except on a boundary layer of order @xmath46 ,
each of the functions @xmath47 become constant in cell interiors .
using this one can show ( see for instance @xcite ) that asymptotically , as @xmath4 , the effective diffusion matrix behaves like @xmath48 here @xmath49 is the identity matrix , and @xmath50 is an explicitly computable constant .
consequently , if we consider , with the dirichlet boundary conditions on the unit square , we expect @xmath51 on the other hand , if we keep @xmath24 fixed and send @xmath4 , we know @xcite that @xmath13 becomes constant on stream lines of @xmath52 . in particular ,
because of the dirichlet boundary condition on the outside boundary , we must also have @xmath53 on the boundary of all interior cells .
since these cells have side length @xmath24 , we expect @xmath54 matching and leads us to believe @xmath55 marks the transition between the two regimes .
with this explanation , we state our main results .
our first two results study the averaging to homogenization transition for the principal dirichlet eigenvalue .
let @xmath3 be an even integer , @xmath56\times[-l/2,l/2]$ ] be a square of side length @xmath3 , and @xmath57 be given .
we study the principal eigenvalue problem on @xmath58 @xmath59 as both @xmath60 .
we observe two distinct behaviors of @xmath61 with a sharp transition . if @xmath62 , then the principal eigenvalue stays bounded , and can be read off using the variational principle in @xcite in the limit @xmath4 .
this is the averaging regime , and exactly explains . on the other hand ,
if @xmath63 , then the principal eigenvalue is of the order @xmath64 .
this is the homogenization regime , and when rescaled to a domain of size @xmath65 , exactly explains .
our precise results are stated below .
[ thmevalstrongflow]let @xmath66 be the solution of and @xmath67 be the principal eigenvalue .
if @xmath4 , and @xmath68 varies such that @xmath69 then there exist two constants @xmath70 , independent of @xmath3 and @xmath0 , such that @xmath71 for all @xmath0 sufficiently large .
[ thmevalhomog]as with theorem [ thmevalstrongflow ] , let @xmath66 be the solution of and @xmath67 be the principal eigenvalue .
if @xmath2 , and @xmath72 varies such that @xmath73 then there exists a constant @xmath74 , independent of @xmath3 and @xmath0 , such that @xmath75 for all @xmath3 sufficiently large . in the special case where @xmath76 , for @xmath77 , assumption is satisfied , and consequently the principal eigenvalue remains bounded and non - zero . for @xmath78 , assumption
is satisfied and the principal eigenvalue behaves like that of the homogenized equation . in the averaging regime ( theorem [ thmevalstrongflow ] ) , the proof of the upper bound in follows directly using ideas of @xcite .
the lower bound , however , is much more intricate .
the main idea is to control the oscillation of @xmath13 between neighbouring cells , and use this to show that the effect of the cold boundary propagates inward along separatrices , all the way to the center cell .
the techniques used are similar to @xcite .
the main new ( and non - trivial ) difficulty in our situation is that the number of cells also increases with the amplitude .
this requires us to estimate the oscillation of @xmath13 between cells in terms of energies localised to each cell ( proposition [ ppntotalosc ] , below ) . here
the assumption that @xmath3 is not too large comes into play . finally , the key idea in the proof is to use a min - max argument ( lemma [ lmagi ] , below ) to show that @xmath13 is small on the boundaries of all cells .
moreover , once smallness on separatrices is established , our proof may be modified to show that under a stronger assumption @xmath79 we have a precise asymptotics @xmath80 where @xmath34 is a single cell .
this is the same as the variational principle in @xcite .
we remark however that @xcite only gives for fixed @xmath3 as @xmath4 . turning to the homogenization regime ( theorem [ thmevalhomog ] ) , we remark first that homogenization of eigenvalues has not been as extensively studied as other homogenization problems .
this is possibly because eigenvalues involve the infinite time horizon .
we refer to @xcite that all study self - adjoint problems for some results on the homogenization of the eigenvalues in oscillatory periodic media .
the extra difficulties in the present paper come both from two sources .
first , since the problem is not self adjoint , a variational principle for the eigenvalue is not available .
second , as @xmath0 and @xmath3 tend to @xmath81 , we do nt have suitable aprori bounds because either the domain is not compact , or the effective diffusivity is unbounded .
our proof uses a multi - scale expansion to construct appropriate sub and super solutions .
when @xmath0 is fixed , it usually suffices to consider a multi - scale expansion to the first corrector .
however , in our situation , this is not enough , and we are forced to consider a multi - scale expansion up to the second corrector
. of course an asymptotic profile , and explicit bounds are readily available @xcite for the first corrector .
however , to the best of our knowledge , bounds on the second corrector as @xmath4 have not been studied .
there are two main problems to obtaining these bounds .
the first problem is appearance of that terms involving the slow gradient of the second corrector multiplied by @xmath0 .
in general , we have no way of bounding these terms .
luckily , if we choose our slow profile to be quadratic , then these terms idnetically vanish and present no problem at all !
the second problem with obtaining bounds on the second corrector is that it satisfies an equation where the first order terms depend on @xmath0 .
so one would expect the bounds to also depend on @xmath0 , which would be catastrophic in our situation .
however , for elliptic equations with a _ divergence free _ drift , we have apriori @xmath10 estimates which are _ independent _ of the drift @xcite .
this , combined with an explicit knowledge of the first corrector , allows us to obtain bounds on the second corrector that decay when @xmath63 .
the sub and super solutions we construct for eigenvalue problem are done through the expected exit time .
since these are interesting in their own right , we describe them below .
let @xmath82 be the solution of @xmath83 where @xmath14 and @xmath58 are as in .
though we do not use any probabilistic arguments in this paper , it is useful to point out the connection between @xmath84 and diffusions .
let @xmath85 be the diffusion @xmath86 where @xmath87 is a standard @xmath44-dimensional brownian motion .
it is well known that @xmath84 is the expected exit time of the diffusion @xmath85 from the domain @xmath58 .
numerical simulations of three realizations of @xmath85 are shown in figure [ figdiff ] .
note that for `` small '' amplitude ( @xmath88 ) , trajectories of @xmath85 behave similarly to those of the brownian motion . for a `` large '' amplitude ( @xmath89 ) ,
trajectories of @xmath85 tend to move ballistically along the skeleton of the separatrices .
similar to the eigenvalue problem , the behaviour of @xmath84 is described by two distinct regimes with a sharp transition . if @xmath62 , then the stirring is strong enough to force the diffusion @xmath85 to exit @xmath58 almost immediately along separatrices . in this case , we show that @xmath90 on separatrices , and is bounded everywhere else above by a constant independent of @xmath0 and @xmath3 . on the other hand ,
if @xmath63 , then the stirring is not strong enough for the effect of the cold boundary to be felt in the interior . in this case , it takes the diffusion @xmath85 a very long time to exit from @xmath58 , and @xmath91 as @xmath92 .
a numerical simulation showing @xmath84 in each of these regimes is shown in figure [ fgrtau ] .
the precise results are as follows .
[ ppnexittimestrongflow ] let @xmath82 be the solution to .
let @xmath4 , and suppose @xmath68 varies such that is satisfied .
there exists a constant @xmath93 , independent of @xmath0 , @xmath3 , such that for all @xmath0 sufficiently large @xmath94 consequently , if @xmath95 , then @xmath96 as @xmath4 , and @xmath97 is bounded uniformly in @xmath0 .
[ ppnexittimehomog]as with theorem [ ppnexittimestrongflow ] , let @xmath98 be the solution of on the square @xmath56\times[-l/2,l/2]$ ] .
suppose now @xmath2 , and @xmath72 varies such that is satisfied , for some fixed @xmath99 . then , for any @xmath100 , there exists a constant @xmath101 , independent of @xmath0 , @xmath3 , such that @xmath102 for all @xmath3 sufficiently large .
consequently , @xmath91 as @xmath2 , uniformly on compact sets .
the proof of theorem [ ppnexittimestrongflow ] is similar steps to that of theorem [ thmevalstrongflow ] . for the proof of theorem [ ppnexittimehomog ] ,
as mentioned earlier , we need to perform a multi - scale expansion up to two correctors , and choose the slow profile to be quadratic .
when the domain is a disk , a quadratic function is exactly the solution to the homogenized problem !
this gives us a sharper estimate for @xmath84 .
[ prop - disk]let @xmath103 be a disk of radius @xmath3 , and @xmath104 be the solution of in @xmath103 . if @xmath0 and @xmath3 satisfy the assumptions in theorem [ ppnexittimehomog ] then @xmath105 where @xmath106 is independent of @xmath0 and @xmath3 . here
@xmath107 is the effective diffusion matrix and @xmath108 denotes the trace of this matrix .
note that right hand side of tends to infinity as @xmath92 .
however , by the terms on the left are of order @xmath109 , which dominates the right hand side
. thus immediately implies . by fitting a disk inside , and outside a square , proposition [ prop - disk ] quickly implies theorem [ ppnexittimehomog ] .
further , since it is well known that the principal eigenvalue is bounded below by the maximum expected exit time , the lower bound in theorem [ thmevalhomog ] also quickly follows from proposition [ prop - disk ] .
the upper bound is a little more technical , however , also uses proposition [ prop - disk ] as the main idea .
we mention that we have chosen to use the particular form of the stream - function @xmath110 simply for the sake of convenience .
all our results may be generalized to other periodic flows with a cellular structure without any difficulty .
we also believe that for other flows the transition from the averaging to the homogenization regime happens when the effective diffusivity @xmath107 balances with the domain size @xmath3 .
that is , when the `` homogenized eigenvalue '' @xmath111 is of the same order as the `` strong flow '' eigenvalue : @xmath112 we leave this question for a future study .
this paper is organized as follows .
the averaging regime is considered in sections [ sec : eigfast ] and [ sec : exitfast ] .
the former contains the proof of theorem [ thmevalstrongflow ] and the latter of theorem [ ppnexittimestrongflow ] .
the rest of the paper addresses the homogenization regime .
the key step here is proposition [ prop - disk ] proved in section [ sxnhomogexittime ] . from this , theorem [ ppnexittimehomog ] quickly follows , and the proof is presented the same section .
theorem [ thmevalhomog ] is proved in section [ sec : homeig ] .
gi was supported by nsf grant dms-1007914 , tk by polish ministry of science and higher education grant nn 201419139 , an by nsf grant dms-0908011 , and lr by nsf grant dms-0908507 , and nsseff fellowship .
we thank po - shen loh for suggesting the proof of lemma [ lmagi ] .
in this section we present the proof of theorem [ thmevalstrongflow ] .
first , we discuss the proof of the upper bound in , followed by the proof of the corresponding lower bound , and , finally , of the limiting behavior in .
the upper bound for @xmath61 in follows directly from the techniques of @xcite .
we carry out the details below . following @xcite , given any test function @xmath113 , and a number @xmath114 ,
we multiply by @xmath115 and integrate over @xmath58 to obtain @xmath116 for the first term on the right , we have @xmath117 for the second term on the right of we have , since @xmath118 is incompressible , @xmath119 hence , equation reduces to @xmath120 now , choose @xmath121 to be any @xmath122 first integral of @xmath14 ( that is , @xmath123 ) .
then , equation reduces to @xmath124 upon sending @xmath125 , the monotone convergence theorem shows @xmath126 for any @xmath122 first integral of @xmath14 . choosing @xmath127 , which , of course , does not depend on @xmath3 , we immediately see that @xmath128 where @xmath129 is any cell in @xmath58 .
this gives a finite upper bound for @xmath61 that is independent of @xmath3 and @xmath0 .
the outline of the proof is as follows .
the basic idea is that if the domain size @xmath3 is not too large , and the flow is sufficiently strong , the eigenfunction @xmath13 should be small not only near the boundary @xmath130 but also on the whole skeleton of separatrices inside @xmath58 .
therefore , the dirichlet eigenvalue problem for the whole domain @xmath58 is essentially equivalent to a one - cell dirichlet problem , which gives the correct asymptotics for the eigenvalue for @xmath0 large . to this end
, we first estimate the oscillation of @xmath13 along a streamline of @xmath14 inside one cell that is sufficiently close to the separatrix , and show that this oscillation is small : see lemma [ lmaosconsep ] .
next , we show that the difference of the values of @xmath13 on two streamlines of @xmath14 ( sufficiently close to the separatrix ) in two neighbouring cells must be small , as in lemma [ lmaoscbetweencells ] below .
these two steps are very similar to those in @xcite , and their proofs are only sketched .
now , considering the ` worst case scenario ' of the above oscillation estimates , we obtain a pointwise upper bound on @xmath13 on streamlines of @xmath14 near separatrices in terms of the principal eigenvalue @xmath61 , and @xmath131 : see lemma [ lmagi ] .
next , we show that the streamlines above enclose a large enough region to encompass most of the mass of @xmath132 .
finally , we use the drift independent apriori estimates in @xcite to obtain the desired lower bound on @xmath61 . the basic reason behind the fact that the eigenfunction is constant on streamlines is the following estimate , originally due to s. heinze @xcite .
[ lmaheinz]there exists a constant @xmath133 so that we have @xmath134 let us use the normalization @xmath135 .
we multiply by @xmath136 , and integrate over @xmath137 .
this gives @xmath138 notice that @xmath139 since @xmath140 on @xmath141 and @xmath142 .
similarly , we have , as @xmath143 on @xmath130 : @xmath144 and consequently we obtain . the next lemma bounds
locally the oscillation on streamlines in terms of the @xmath1-norm of @xmath145 .
[ lmaosconsep]let @xmath137 be any cell . for any @xmath146
, there exists @xmath147 such that @xmath148 is a level set of @xmath52 , @xmath149 , and @xmath150 for some constant @xmath93 independent of @xmath151 .
we will see that @xmath152 is the ` width ' of the boundary layer , and will eventually be chosen to be @xmath153 .
the proof is straightforward and similar bounds have already appeared in @xcite .
we sketch the details here for convenience .
first we introduce curvilinear coordinates in the cell @xmath137 . for this ,
let @xmath154 be the center of @xmath137 , and @xmath155 be the solution of @xmath156 as usual , we extend @xmath157 to @xmath137 by defining it to be @xmath158 ( or @xmath159 ) on @xmath160 . in the coordinates @xmath161 given by the functions @xmath52 and @xmath155 , it is easy to check that @xmath162 assume , for simplicity , that @xmath163 on @xmath137 . then for any @xmath164
, we have @xmath165 the last inequality follows from the fact that @xmath166 as the length element along the contour @xmath167 .
now integrating over @xmath168 , we get @xmath169 and follows from the mean value theorem .
now , we consider two streamlines on which the solution is nearly constant and estimate the possible jump in the value of @xmath13 between them .
[ lmaoscbetweencells]let @xmath137 and @xmath170 be two neighbouring cells , @xmath147 , @xmath171 the respective level sets from lemma [ lmaosconsep ] , and let @xmath172 , @xmath173
. then there exists @xmath174 and @xmath175 such that @xmath176 assume again for simplicity that @xmath163 on @xmath137 , and @xmath137 is to the left of @xmath170 .
then , using the local curvilinear coordinates @xmath177 around the common boundary between the cells @xmath137 and @xmath170 , as in , we have @xmath178 now , let @xmath179 be fixed . in the region @xmath180 , and @xmath181 , we know that @xmath182 and @xmath183 .
hence , we have @xmath184 however , lemma [ lmaosconsep ] shows that @xmath185 this concludes the proof of .
for each cell @xmath137 , we set @xmath186 choosing @xmath187 , lemmas [ lmaosconsep ] and [ lmaoscbetweencells ] immediately give the following oscillation estimate .
[ ppntotalosc]if @xmath137 and @xmath170 are any two cells , and @xmath147 , @xmath171 the respective level sets from lemma [ lmaosconsep ] , then @xmath188 where the sum is taken over any path of cells that connects @xmath137 and @xmath170 , consisting of only horizontal and vertical line segments .
lemma [ lmaheinz ] implies that @xmath189 with the summation taken over all cells in @xmath58 .
now , the key to the proof of the lower bound in theorem [ thmevalstrongflow ] is to obtain an estimate on @xmath190 in terms of @xmath191 .
a direct application of cauchy - schwartz to proposition [ ppntotalosc ] is wasteful and does not yield a good enough estimate .
what is required is a more careful estimate of the ` worst case scenario ' for the values of @xmath192 .
this is the content of our next lemma .
[ lmagi]on any cell @xmath137 , we have @xmath193 where @xmath147 is the level set from lemma [ lmaosconsep ] .
for notational convenience , in this proof only , we will assume that @xmath194 is the square of side length @xmath195 centered @xmath196 , and @xmath197 is the cell with center @xmath198 .
note that in the present proof we label the cells , ( and contours @xmath199 inside the cell @xmath200 we use below ) by two indices that correspond to the coordinates of the center of the cell .
let @xmath201 denote the set of all paths of cells that join the boundary @xmath202 to the cell @xmath203 using only horizontal and vertical line segments .
let @xmath204 , @xmath205 , be a collection of non - negative numbers assigned to each cell , and denote @xmath206 we first claim there exists an explicitly computable constant @xmath93 , independent of @xmath207 such that @xmath208 which is an obvious improvement over the cauchy - schwartz estimate applied blindly to .
this improvement comes because we are taking the minimum over all such paths in . to prove
, we define @xmath209 where @xmath210 is any collection of ( possibly repeated ) paths in @xmath211 .
since the minimum of a collection of numbers is not bigger than the average of any subset , we certainly have @xmath212 for any collection @xmath213 .
the idea is to choose such a sub - collection in a convenient way .
we prove the claim for @xmath214 .
we choose @xmath215 to consist of @xmath216 paths , with the following property .
all paths stay in the upper - right quadrant .
the last cell visited by all paths is @xmath196 .
the second to last cell visited by @xmath217 paths ( half of the collection ) is @xmath218 , and the second to last cell visited by the remaining paths is @xmath219 . amongst the paths who s second to last cell
is @xmath218 , we choose @xmath215 so that two thirds of these paths have @xmath220 as the third to last cell , and one third have @xmath221 as the third to last cell .
symmetrically , we choose @xmath215 so that amongst all the paths who s second to last cell is @xmath219 , two thirds of these paths have @xmath222 as the third last cell , and one third have @xmath221 as the third to last cell .
consequently exactly @xmath223 paths have third to last cell @xmath220 , exactly @xmath223 paths in @xmath215 have third to last cell @xmath221 , and exactly @xmath223 paths in @xmath215 have third to last cell @xmath222 .
continuing similarly , we see that @xmath215 can be chosen so that for any cell @xmath198 with @xmath224 , _ exactly _
@xmath225 paths visit the cell @xmath198 as the @xmath226 to last cell .
finally , we assume that all paths in @xmath215 start on the top boundary and proceed directly vertically downward until they hit a cell of the form @xmath227 .
let us count how many times each term @xmath228 appears in the averaged sum @xmath229 . clearly , if @xmath224 , then the cell @xmath198 appears in exactly @xmath230 paths in @xmath215 . on the other hand , if @xmath231 , then the cell @xmath198 appears exactly @xmath232 paths
consequently , we have @xmath233 we now maximize the sum in with the constraint @xmath234 let @xmath235 denote the right side of , then at the maximizer of @xmath235 we have @xmath236 and @xmath237 the euler - lagrange equations now imply that @xmath238 and @xmath239 here @xmath240 is the lagrange multiplier that can be computed from the constraint : @xmath241 it follows that @xmath242 with @xmath243 as @xmath244 . hence , for the maximizer
we get @xmath245 and thus holds for @xmath246 .
however , it is immediate to see that the previous argument can be applied to any cell considering appropriate collection of paths that say up and to the right of @xmath247 , whence holds for all @xmath247 . with in hand
, we observe that proposition [ ppntotalosc ] implies @xmath248 where the last inequality follows from lemma [ lmaheinz ] .
this concludes the proof .
our next step shows that the mass of @xmath132 in the regions enclosed by the level sets @xmath148 is comparable to @xmath249 .
[ lmagi2]let @xmath137 be a cell , and @xmath147 the level set from lemma [ lmaosconsep ] .
let @xmath172 , and @xmath250 be a neighbourhood of @xmath141 .
let @xmath251 .
then , for @xmath0 sufficiently large , we have @xmath252 for any cell @xmath137 , the sobolev restriction theorem shows @xmath253 thus , using curvilinear coordinates with respect to the cell @xmath137 , and assuming , for simplicity , that @xmath254 , gives @xmath255 } \frac{1}{\abs{{\nabla}h}}\right ) \
, dh\\ & { \leqslant}c\left ( \alpha_i + \norm{\varphi}_{l^2(q_i)}^2 \right ) \int_{h=0}^{2 \delta_0 } \frac{1}{\sqrt{h } } \ , dh= c \sqrt{\delta_0 } \left ( \alpha_i + \norm{\varphi}_{l^2(q_i)}^2 \right).\end{aligned}\ ] ] summing over all cells gives @xmath256 where the last inequality follows using the upper bound in which was proved in section [ sxnstrongflowupperbound ] . since @xmath257 as @xmath4 , and @xmath258
inequality follows . our final ingredient is a drift _
independent _ @xmath10 estimate in @xcite .
we recall it here for convenience . [
lmadriftindlinfbound]let @xmath259 be a domain , @xmath121 be divergence free , and @xmath260 be the solution to @xmath261 with @xmath262 for some @xmath263 .
there exists a constant @xmath264 , _ independent of @xmath121 _ , such that @xmath265 .
we are now ready to prove the lower bound in theorem [ thmevalstrongflow ] .
using the notation from lemma [ lmagi2 ] , define @xmath266 , and let @xmath170 be a cell such that @xmath267 .
then @xmath268 and it follows from lemmas [ lmagi ] and [ lmadriftindlinfbound ] that @xmath269 where the last inequality follows from lemma [ lmagi2 ] .
consequently , @xmath270 where the last inequality follows from equation .
this proves the lower bound on @xmath61 in theorem [ thmevalstrongflow ] .
in this section we sketch the proof of theorem [ ppnexittimestrongflow ] .
the techniques in section [ sxnstrongflowlowerbound ] readily show that oscillation of @xmath84 on stream lines of @xmath14 becomes small .
now , the key observation in the proof of theorem [ ppnexittimestrongflow ] is an _ explicit , drift independent _ upper bound on the exit time .
we state this below .
[ lmadriftindtaubound]let @xmath271 be a bounded , piecewise @xmath272 domain , and @xmath273 a @xmath272 divergence free vector field tangential to @xmath274 .
let @xmath275 be the solution to @xmath276 then for any @xmath277 $ ] , @xmath278 where @xmath279 is a ball with the same lebesgue measure as @xmath280 , and @xmath281 is the ( radial , explicitly computable ) solution to on @xmath282 with @xmath283 . with this , we present the proof of theorem [ ppnexittimestrongflow ] .
following the same method as that in section [ sxnstrongflowlowerbound ] , we obtain ( analogous to lemma [ lmagi ] ) @xmath284 with @xmath192 now equal @xmath285 , where @xmath137 is the @xmath286 cell .
the sets @xmath147 appearing in are level sets of @xmath52 on which the oscillation of @xmath84 is small ( analogous to lemma [ lmaosconsep ] ) .
now observe that @xmath287 and so reduces to @xmath288 letting @xmath289 be the region enclosed by @xmath148 , we obtain ( similar to lemma [ lmagi2 ] ) @xmath290 for large enough @xmath0 . by lemma [ lmadriftindtaubound ]
we see @xmath291 and hence @xmath292 solving the above inequality quickly yields @xmath293 where the second inequality above follows from the assumption . substituting this in immediately shows that @xmath294 now , to conclude the proof , we appeal to lemma [ lmadriftindtaubound ] again .
let @xmath295 be the ( fattened ) skeleton of the separatrices .
observe that @xmath296 which , by assumption , remains bounded uniformly in @xmath0 .
consequently , by lemma [ lmadriftindtaubound ] , @xmath297 which immediately yields the desired result .
the key step in our analysis in the homogenization regime is proposition [ prop - disk ] , and we begin with it s proof .
the idea of the proof is to construct good sub and super solutions for the exit time problem in a disk of radius one .
let @xmath84 be the solution of in a ball of radius @xmath3 .
let @xmath298 be a ball of radius @xmath65 , and let @xmath299 .
then @xmath300 is a solution of the pde @xmath301 we begin by constructing an approximate solution @xmath302 , by defining @xmath303 where @xmath304 is the ` fast variable ' .
we define @xmath305 explicitly by @xmath306 and obtain equations for @xmath307 and @xmath308 using the standard periodic homogenization multi - scale expansion . using the identities @xmath309 we compute @xmath310 -{\delta}_x \tau_{11 } - 2 l { \nabla}_x \cdot { \nabla}_y \tau_{11 } - l^2 { \delta}_y \tau_{11}\\ + l a v \cdot { \nabla}_x \tau_{11 } + l^2 a v \cdot { \nabla}_y \tau_{11 } \bigr ) \end{multlined}\\ & + \frac{1}{l^2 } \left ( - l^2 { \delta}_y \tau_{12 } + l^2 a v \cdot { \nabla}_y \tau_{12 } \right).\end{aligned}\ ] ] we choose @xmath307 to formally balance the @xmath311 terms .
that is , we define @xmath307 to be the mean - zero , periodic function such that @xmath312 we clarify that when dealing with functions of the fast variable , we say that a function @xmath260 is periodic if @xmath313 for all @xmath314 .
this is because our drift @xmath14 is periodic , with period @xmath44 in the fast variable , and each cell is a square of side length 2 , in the fast variable .
now we choose @xmath308 to formally balance the @xmath315 terms .
define @xmath308 to be the mean - zero , periodic function such that @xmath316 where @xmath317 denotes the mean with respect to the fast variable @xmath318 .
observe that we had to introduce the term @xmath319 above to ensure that the right hand side is mean zero , to satisfy the compatibility condition .
we write @xmath320 where @xmath321 , @xmath322 are the mean zero , periodic solutions to @xmath323 using this expression for @xmath307 and we simplify to @xmath324 the key observation is that with our choice of @xmath305 , the right side of is _ independent _ of the slow variable . our aim is to show that @xmath325 satisfies the estimates and below . [ lmasupsolbis]there exists a positive constant @xmath326 independent of @xmath0 , @xmath3 , such that for @xmath325 defined by we have @xmath327 and @xmath328 here @xmath107 is the effective diffusion matrix , given by .
we first use the lemma to finish the proof of proposition [ prop - disk ] .
the key observation we obtain from lemma [ lmasupsolbis ] is that , except for the boundary condition , the function @xmath329 satisfies exactly .
this is a miracle that happens only when the domain is a disk
. then we get sub- and super - solutions for @xmath330 by setting @xmath331,\ ] ] and @xmath332.\ ] ] lemma [ lmasupsolbis ] implies that @xmath333 further , since @xmath334 on @xmath335 , equation implies that @xmath336 and @xmath337 on @xmath335 . consequently , @xmath338 is a super solution , and @xmath339 is a sub solution of , and hence @xmath340{\leqslant}\tau_1(x){\leqslant}\frac{1}{\operatorname{tr}(\bar \sigma(a))}\left[\tilde{\tau}(x)+\frac{2c_0}{l^{\alpha/4}}\right].\ ] ] rescaling to the ball of radius @xmath3 , we see @xmath341 now using and we obtain . it remains to prove lemma [ lmasupsolbis ] . by our definition of @xmath305 , @xmath307 , @xmath308
, we have @xmath342 = 2 + \average*{\abs{{\nabla}\chi_1}^2 + \abs{{\nabla}\chi_2}^2 } = \operatorname{tr}(\bar\sigma(a)),\ ] ] where the second inequality follows from .
this is exactly . to prove
, we will show @xmath343 for some constant @xmath344 , independent of @xmath0 and @xmath3 .
we will subsequently adopt the convention that @xmath345 is a constant , depending only on @xmath346 , which can change from line to line .
we first bound @xmath307 .
let @xmath347 be the fundamental domain of the fast variable .
let @xmath348 and @xmath349 denote the vertical and horizaondal boundaries of @xmath34 respectively . since @xmath350 is odd in @xmath351 and even in @xmath352 , by symmetry we have @xmath353 on @xmath348 , and @xmath354 on @xmath349 .
now if we consider the function @xmath355 , we have @xmath356 thus the hopf lemma implies @xmath355 does not attain it s maximum on @xmath349 , except possibly at corner points .
so by the maximum principle @xmath355 attains its maximum on @xmath348 , and so @xmath357 since @xmath358 is bounded similarly , we immediately have @xmath359 the last step is to prove a bound on @xmath360 . the crucial idea to bound
@xmath308 is to split the right hand side of into terms which are small in @xmath361 , and terms which can be absorbed by the convection term . to this end , write @xmath362 where @xmath363 , @xmath364 are mean - zero , periodic solutions to @xmath365 \\ -{\delta}_y \psi_2 + a v \cdot { \nabla}_y \psi_2 = -a \sum_{i=1}^2 v_i \left ( y_i - \frac{1}{2 } \operatorname{sign}(y_i ) \right ) . \ ] ] before estimating each term individually , we pause momentarily to explain this decomposition of @xmath308 .
the equation for @xmath363 is of course natural .
the equation for @xmath366 stems from the well known behaviour of the corrector @xmath30 . we know from @xcite that @xmath30 grows rapidly in a boundary layer of width @xmath367 and decreases _ linearly
_ in the cell interior . that is , for @xmath368 , @xmath369 away from the boundary layer .
further , by symmetry , @xmath370 is odd in @xmath371 , and even in the other variable .
thus , we expect the term @xmath372 to be away from zero only in the boundary layer ( see figure [ fgrxi ] ) , and hence should have a small @xmath361 norm !
now the equation for @xmath373 is chosen to balance the remaining terms , and thankfully the right hand side can be absorbed in the convection term . with this explanation
, we proceed to estimate each function individually , starting with @xmath363 .
since @xmath374 , lemma [ lmadriftindlinfbound ] guarantees @xmath375 we remark that while lemma [ lmadriftindlinfbound ] is stated for homogeneous dirichlet boundary conditions , the proof in @xcite goes through verbatim for periodic boundary conditions , provided , of course , we assume our solution is mean - zero . this justifies the application of lemma [ lmadriftindlinfbound ] in this context . since we know from @xcite that @xmath376
, we immediately obtain @xmath377 our bound for @xmath366 is similar in flavor .
let @xmath378 . then for any @xmath379
, we know from @xcite ( see also @xcite*theorem 1.2 ) that @xmath380 since @xmath381 , from lemma [ lmadriftindlinfbound ] we have @xmath382 for any @xmath383 . finally for @xmath373 , our aim is to absorb the right hand side into the drift . for @xmath368 ,
let @xmath384 be defined by @xmath385 and extended to be a @xmath272 , periodic function on @xmath386 in the natural way
. set @xmath387 , then @xmath260 is a periodic , mean - zero solution to @xmath388 where @xmath389 since for any @xmath379 , we can explicitly compute @xmath390 by lemma [ lmadriftindlinfbound ] we obtain @xmath391 for any @xmath383 . thus combining , and
, we see @xmath392 . thus using and choosing @xmath393 when @xmath394 , and @xmath395 for @xmath396 , we see @xmath397 proving .
this completes the proof .
theorem [ ppnexittimehomog ] follows immediately from proposition [ ppnexittimehomog ] .
the proof of theorem [ ppnexittimehomog ] is now trivial .
we simply inscribe a disk @xmath398 into the square @xmath399 ^ 2 $ ] , and circumscribe a bigger disk @xmath400 around @xmath58 .
the corresponding exit times satisfy the inequality @xmath401 using the bounds obtained from proposition [ prop - disk ] applied to @xmath402 and @xmath403 , the inequality follows .
we claim that in general , we have the principal eigenvalue and expected exit time satisfy @xmath404 to see this , pick any @xmath405 , and suppose for contradiction that @xmath406 .
then , @xmath407 also @xmath408 rescaling @xmath13 if necessary to ensure @xmath409 , we see have @xmath410 in @xmath58 .
thus perron s method implies the existence of a function @xmath411 such that this immediately implies @xmath412 equals the principal eigenvalue @xmath61 , which contradicts our assumption .
thus , for any @xmath405 , we must have @xmath413 . sending @xmath26
, we obtain .
applying theorem [ ppnexittimehomog ] concludes the proof .
the strategy is very similar to that for the exit time on a disk , we present the details for the convenience . for this section ,
we choose our domain as @xmath414\times[0,1]$ ] . rescaling the eigenvalue problem
gives @xmath415 where @xmath416 as before , we will use a multi - scale expansion to construct a super - solution to that yields the desired bound : set @xmath417 where @xmath304 is the ` fast variable ' .
we now define @xmath418 by @xmath419 where @xmath346 is as in the assumption , and the constant @xmath420 will be chosen later to make @xmath421 non - negative . following the previous strategy
, we define @xmath422 to be the mean - zero , periodic function such that @xmath423 and let @xmath424 be the mean - zero , periodic function such that @xmath425 where @xmath317 denotes the mean with respect to the fast variable @xmath318 . again , with our choice of @xmath418 , the right side of is _ independent _ of the slow variable .
our aim is to show that for a suitable constant @xmath431 , @xmath13 is a super - solution to .
the following lemma is proved identically to lemma [ lmasupsolbis ] .
choose @xmath436 so that @xmath437 for all @xmath438 .
we claim that @xmath439 . to see this ,
suppose for the sake of a contradiction , @xmath440 .
let @xmath441 be the principal eigenfunction of .
then @xmath442 also , from , we see @xmath443 if we normalize @xmath441 so that @xmath444 , inequality guarantees that @xmath445 .
thus , perron s method will provide the existence of a function @xmath446 such that @xmath447 consequently , @xmath446 is the principal eigenfunction of the operator @xmath448 and hence @xmath449 , which contradicts our assumption that @xmath440 .
thus , we have @xmath450 and using , we immediately get the lower bound in , except for the proof of lemma [ lmasupsol ] . in this section
we prove the upper bound in theorem [ thmevalhomog ] .
we will do this by using a multi - scale expansion of a sub - solution . as we have seen in the preceding sections ,
our multi - scale expansions are all forced to use a quadratic ` slow ' profile , in order to avoid extra terms in the expansion .
this makes the construction of the sub - solution slightly more difficult . as customary with homogenization problems , we rescale the problem so that the cell size goes to @xmath158 , and the domain is fixed . [ lmapsi]let @xmath451 , and @xmath452 be the solution of @xmath453 where @xmath454 , and @xmath455 is the characteristic function of the set @xmath235 .
assume that @xmath0 and @xmath3 vary such that holds .
then there exists @xmath451 , and @xmath456 such that @xmath457 provided @xmath0 and @xmath3 are sufficiently large
. let @xmath422 , @xmath458 be the principal eigenfunction and the principal eigenvalue respectively for the rescaled problem @xmath459 assume , for contradiction , @xmath460 , where @xmath345 is the constant in lemma [ lmapsi ] , then @xmath461 also , if @xmath452 is the function from lemma [ lmapsi ] , then by the maximum principle , @xmath462 in @xmath463 .
hence , @xmath464 by the hopf lemma , we know @xmath465 on @xmath335 , and so @xmath422 can be rescaled to ensure @xmath466 .
perron s method now implies that there exists a function @xmath411 that satisfies @xmath467 and , in addition , @xmath468 .
therefore , @xmath469 is the principal eigenvalue , which contradicts our assumption @xmath470 .
hence @xmath471 .
now rescaling back so the cell size is @xmath65 , let @xmath472 and @xmath473 be the principal eigenvalue and principal eigenfunction respectively of the problem on the ball of radius @xmath474 . since @xmath475 ,
we have @xmath476 . finally , let @xmath58 be the square with side length @xmath3 , and @xmath61 , @xmath13 are the principal eigenvalue and eigenfunction respectively of the problem on @xmath58 . then , since @xmath477 , the principal eigenvalues must satisfy @xmath478 , from which the theorem follows .
let @xmath479 , where @xmath300 is the solution of , the expected exit time problem from the unit disk .
now rescaling to the ball of radius @xmath65 , ( or directly using , which was what lead to ) , we obtain @xmath480 provided @xmath0 and @xmath3 are large enough and satisfy . here
@xmath481 is a fixed constant independent of @xmath0 and @xmath3 .
thus , will follow if we show that @xmath482 for some small @xmath483 .
observe that implies that the right hand side of is @xmath484 .
therefore , to establish , it suffices to show that there exists constants @xmath485 and @xmath106 such that for all @xmath486 , there exists @xmath487 and @xmath488 such that @xmath489 provided @xmath490 , @xmath491 and holds . above any power of @xmath492 strictly larger than @xmath65 will do ; our construction below obtains @xmath493 , however , in reality one would expect the power to be @xmath494 . we will obtain by considering a poisson problem on the annulus @xmath495 if we impose a large enough constant boundary condition on the inner boundary , the ( inward ) normal derivative will be negative on @xmath496 . now ,
if we extend this function inward by a constant , we will have a super - solution giving the desired estimate for @xmath497 .
we first state a lemma guaranteeing the sign of the normal derivative of an appropriate poisson problem .
[ lmatheta1bound]there exists @xmath498 and @xmath499 , such that for all @xmath500 , there exists @xmath501 , @xmath502 such that the solution @xmath503 of the pde @xmath504 satisfies @xmath505 provided @xmath491 , @xmath490 and holds . here
@xmath506 denotes the derivative with respect to the radial direction .
moreover , the function @xmath503 attains its maximum on @xmath507 , and @xmath508 for all @xmath509 .
now , postponing the proof of lemma [ lmatheta1bound ] , we prove .
choose @xmath492 small , and @xmath510 large , as guaranteed by lemma [ lmatheta1bound ] , and define @xmath511 by @xmath512 where @xmath503 and @xmath436 are as in lemma [ lmatheta1bound ] .
then @xmath513 , and @xmath514 further , when @xmath515 , @xmath516 where the second inequality follows from lemma [ lmatheta1bound ] .
thus @xmath511 is a viscosity super solution to the pde @xmath517 by the comparison principle , we must have @xmath518 , which immediately proves . from this follows , and using we obtain , concluding the proof .
it remains to prove lemma [ lmatheta1bound ] .
roughly speaking , if we choose the constant @xmath436 sufficiently large , the function @xmath503 is nearly harmonic .
the inhomogeneity of the boundary conditions dominates the right side of the equation .
a `` nearly harmonic '' function should attain its maximum on the boundary , implying the conclusion of lemma [ lmatheta1bound ] .
the reason we believe the constant @xmath519 is large enough , is because the homogenized exit time from the annulus is quadratic in the width of the annulus .
unfortunately , the slow profile is not quadratic in cartesian coordinates , and so the best we can do is obtain upper and lower bounds , which need not be sharp .
we begin by showing that the expected exit time from an annulus of width @xmath492 grows like @xmath493 . while we certainly do nt expect the exponent @xmath520 to be sharp , any exponent strictly larger than @xmath65 will suffice for our needs .
[ lmaannexittime]let @xmath521 be the annulus @xmath522 , and @xmath523 be the solution of the poisson problem @xmath524 suppose @xmath3 and @xmath0 vary so that holds .
then there exists constants @xmath485 and @xmath106 such that for all @xmath500 , there exists @xmath487 and @xmath488 such that @xmath525 provided @xmath491 , @xmath490 and holds . the main idea behind
the proof is that as @xmath92 , we know that @xmath300 tends to an explicit ( homogenized ) parabolic profile and is constant on @xmath526 . now
if we subtract off a harmonic function with these boundary values , then we should get a super solution for @xmath523 .
finally , we will show that a harmonic function with constant boundary values grows linearly near @xmath335 , at the same rate as @xmath300 .
thus the above super solution will give an upper bound for @xmath523 which is super - linear in the annulus width .
we proceed to carry out the details .
let @xmath527 be the solution of @xmath528 and define @xmath529 where @xmath530 is as in .
then @xmath531 satisfies @xmath532 the first boundary condition follows because both @xmath527 and @xmath300 are @xmath158 on @xmath335 .
the second follows from and the boundary condition for @xmath527 .
thus , the maximum principle immediately implies that @xmath533 .
[ lmaeta]let @xmath363 be the solution of @xmath535 then there exists a constant @xmath345 , independent of @xmath492 , @xmath0 and @xmath3 , such that @xmath536 when @xmath3 and @xmath0 are sufficiently large .
we will construct a sub - solution of equation in small rectangles overlapping @xmath539 . for convenience
, we now shift the origin to @xmath540 , and consider new coordinates @xmath541 . in these coordinates , let @xmath542 be the rectangle of height @xmath543 , width @xmath492 and top left corner @xmath544 , where @xmath545 ( see figure [ fgreta ] ) . before proving that the function @xmath546 exists , we remark that by the maximum principle , we @xmath549 on @xmath550 . moreover , as @xmath551 , the estimate can be extended to @xmath552 $ ] as well , possibly by increasing the constant @xmath345 .
this proves lemma [ lmaeta ] when @xmath553 is on the negative @xmath554-axis .
now , if is still valid when the coordinate frame is rotated , our proof of lemma [ lmaeta ] will be complete ! we will first prove that a function @xmath546 satisfying and exists .
we will do this by a multi - scale expansion .
let @xmath555 where @xmath556 is the fast variable , @xmath557 is given by @xmath558 and @xmath431 , and @xmath559 are constants , each independent of @xmath3 , @xmath0 and @xmath492 , to be chosen later .
as before , @xmath560 is @xmath561 and @xmath562 is the mean @xmath158 , periodic solution to @xmath563 { \partial}_{x_i'}{\partial}_{x_j ' } \eta_0,\ ] ] where @xmath370 are the solutions to .
again , the crucial fact here is that since @xmath557 is quadratic , the second derivatives are constant and @xmath562 becomes independent of the slow variable @xmath564 . using this
, a direct computation shows that @xmath565 note that by symmetry , @xmath566 , and @xmath567 .
further , by our choice of @xmath557 , we have @xmath568 .
hence , the previous equation reduces to @xmath569 as required by the first equation in .
next , we show that if @xmath559 is appropriately chosen , we can arrange the boundary conditions claimed in for @xmath557 .
notice that @xmath570 , and on the top and bottom boundary we have @xmath571 and @xmath572 .
thus @xmath573 so choosing @xmath559 large enough , we can ensure @xmath574 on the top and bottom of @xmath542 . on the left of @xmath542
, we have @xmath575 and @xmath576 .
so @xmath577 , and choosing @xmath559 large we can again ensure @xmath574 on the left of @xmath542 .
finally , on the right of @xmath542 , we have @xmath578 and @xmath579 , and we immediately see that for @xmath559 large enough , we have @xmath580 on the right of @xmath542 . thus @xmath557 satisfies the boundary conditions in . to see that @xmath546 also satisfies the boundary conditions in , we need to bound the correctors appropriately .
exactly as in the proof of lemma [ lmasupsolbis ] , we obtain @xmath581 where @xmath106 is independent of @xmath3 , @xmath0 and @xmath492 .
we remark that the extra @xmath582 factor arises because derivatives of @xmath557 are of the order @xmath582 and they appear as multiplicative factors in the expressions for @xmath560 and @xmath562 . consequently ,
if @xmath431 is chosen to be larger than @xmath345 , we have @xmath583 since @xmath557 already satisfies the boundary conditions in , this immediately implies that @xmath546 must also satisfy these boundary conditions .
finally , since @xmath557 certainly satisfies , it follows from that @xmath546 also satisfies .
this proves the existence of @xmath546 .
now , as remarked earlier , the only thing remaining to complete the proof of the lemma is to verify that if the rectangle @xmath542 , and the coordinate frame are both rotated arbitrarily about the center of the annulus , then there still exists a function @xmath546 satisfying and in new coordinates .
this , however , is immediate .
the new coordinates can be expressed in terms of the old coordinates as a linear function .
consequently , our initial profile for @xmath557 will still be a _
quadratic _ function of the new coordinates .
of course , by the rotational invariance of the laplacian , it will also be harmonic , and the remainder of the proof goes through nearly verbatim .
the only modification is that after the rotation the mixed derivative @xmath584 no longer vanishes , and the terms involving @xmath585 and @xmath586 do appear in and .
however , they can be treated in an identical fashion , as in the proof of lemma [ lmasupsolbis ] using the precise asymptotics for @xmath587 and @xmath358 from @xcite .
this concludes the proof . for a given @xmath436 , let @xmath527 be the solution of @xmath588 then
, we have @xmath589 and @xmath590 consequently @xmath591 , where @xmath523 is the solution of on the annulus @xmath592 . thus applying lemma [ lmaannexittime ] , we see @xmath593 the function @xmath527 decreases at most linearly with @xmath594 .
this can be seen immediately from an asymptotic expansion for a super solution . indeed , starting with @xmath595 and choosing @xmath560 and @xmath562 as in the proof of lemma [ lmaeta ]
, we immediately see that @xmath596 we remark again that the extra @xmath597 factor arises from the gradient of @xmath557 .
now by rotating the initial profile @xmath557 appropriately , we obtain the linear decrease @xmath598 as claimed .
we claim that and quickly conclude the proof . to see this , note first that equation and immediately give @xmath599 however , since @xmath600 on @xmath601 , the maximum principle implies that @xmath503 can not attain it s maximum in the interior of the annulus @xmath601 .
consequently must hold on the interior of the _ entire _ annulus @xmath602 .
now if we choose @xmath436 large enough so that @xmath603 , and then choose @xmath604 large enough so that @xmath605 , we see that @xmath503 is forced to attain it s maximum on the inner boundary @xmath496 , and the lemma follows immediately . | we consider an elliptic eigenvalue problem with a fast cellular flow of amplitude @xmath0 , in a two - dimensional domain with @xmath1 cells . for fixed @xmath0 , and @xmath2 , the problem homogenizes , and has been well studied .
also well studied is the limit when @xmath3 is fixed , and @xmath4 . in this case
the solution equilibrates along stream lines . in this paper
, we show that if _ both _ @xmath4 and @xmath2 , then a transition between the homogenization and averaging regimes occurs at @xmath5 .
when @xmath6 , the principal dirichlet eigenvalue is approximately constant . on the other hand ,
when @xmath7 , the principal eigenvalue behaves like @xmath8 , where @xmath9 is the effective diffusion matrix .
a similar transition is observed for the solution of the exit time problem .
the proof in the homogenization regime involves bounds on the second correctors .
miraculously , if the slow profile is quadratic , these estimates can be obtained using drift independent @xmath10 estimates for elliptic equations with an incompressible drift .
this provides effective sub and super - solutions for our problem . |
many measures including total number of passengers , total number of flights , or total amount of cargo quantifying the importance of the world airports are compiled and publicized @xcite .
we study here the _ oag max _ database , which comprises flight schedule data of more than 800 of the world s airlines for the period november 1 , 2000 to october 31 , 2001 .
this database is compiled by oag , a division of reed business information group , and includes all scheduled flights and scheduled charter flights , both for big aircrafts air carriers and small aircrafts air taxis .
we focus our analysis on a network of cities , not of airports for example , the newark , jfk and la guardia airports are all assigned to new york city .
we further restrict our analysis to _ passenger flights _ operating in the time period november 1 , 2000 to november 7 , 2000 . even though this data is more than four years old
, the resulting world - wide airport network is virtually indistinguishable from the network one would obtain if using data collected today .
the reason is that air traffic patterns are strongly correlated with : ( i ) socio - economic factors , such as population density and economic development ; and ( ii ) geo - political factors such as the distribution of the continents over the surface of the earth and the locations of borders between states @xcite .
clearly , the time scales associated to changes in these factors are much longer than the lag in the data we analyze here . during the period considered ,
there are 531,574 unique non - stop passenger flights , or flight segments , operating between 3883 distinct cities .
we identify 27,051 distinct city pairs having non - stop connections . the fact that the database is highly redundant that is , that most connections between pairs of cities are represented by more than one flight
adds reliability to our analysis .
specifically , the fact that unscheduled flights are not considered does not mean , in general , that the corresponding link between a certain pair of cities is missing in the network , since analogous scheduled flights may still operate between them .
similarly , even if some airlines have canceled their flights between a pair of cities since november 2000 , it is highly unlikely that all of them have .
we create the corresponding adjacency matrix for this network , which turns out to be almost symmetrical .
the very minor asymmetry stems from the fact that a small number of flights follow a `` circular '' pattern , i.e , a flight might go from a to b to c , and back to a. to simplify the analysis , we symmetrize the adjacency matrix . further , we build regional networks for different geographic regions ( table [ t - regions ] ) . specifically ,
we generate twenty - one regional networks at different aggregation levels . at the highest - aggregation level , we generate six networks ; one each for africa , asia and middle east , europe , latin america , north america , and oceania . for each of these regions , except for north america and oceania
, we generate between two and five sub - networks .
for instance , the asia and middle east network is further subdivided into south asia , central asia , southeast asia , northeast asia , and middle east .
a ubiquitous characteristic of complex networks is the so - called `` small - world '' property @xcite . in a small - world network ,
pairs of nodes are connected by short paths as one expects for a random graph @xcite .
crucially , nodes in small - world networks also have a high - degree of cliquishness , as one finds in low - dimensional lattices but not in random graphs . in the air transportation network ,
the average shortest path length @xmath0 is the average minimum number of flights that one needs to take to get from any city to any other city in the world .
we find that for the @xmath1 cities in the asia and middle east network @xmath2 , and that the average shortest path length between the @xmath3 cities in the giant component of the world - wide network is only about one step greater , @xmath4 . actually , most pairs of cities ( 56% ) are connected by four steps or less .
more generally , we find that @xmath0 grows logarithmically with the number @xmath5 of cities in the network , @xmath6 .
this behavior is consistent with both random graphs and small - world networks , but not with low - dimensional networks , for which @xmath0 grows more rapidly with @xmath5 . still , some pairs of cities are considerably further away from each other than the average .
the farthest cities in the network are mount pleasant , in the falkland islands , and wasu , in papua new guinea : to get from one city to the other , one needs to take fifteen different flights . from mount
pleasant , one can fly to punta arenas , in chile , and from there to some hubs in latin america . at the other end of the path , from wasu
one needs to fly to port moresby , which requires a unique sequence of eight flights . in the center of the path , between punta arenas and port moresby ,
six different flights are needed .
in contrast with what happens the ends of the path , in the central region of the path there are hundreds of different flight combinations , all of them connecting punta arenas and port moresby in six steps .
the clustering coefficient @xmath7 , which quantifies the local cliquishness of a network , is defined as the probability that two cities that are directly connected to a third city are also directly connected to each other .
we find that @xmath7 is typically larger for the air transportation network than for a random graph and that it increases with size .
these results are consistent with the expectations for a small - world network but not those for a random graph . for the world - wide network ,
we find @xmath8 while its randomization yields @xmath9 .
therefore , we conclude that the air transportation network is , as expected , a small - world network @xcite . another fundamental aspect in which real - world networks often deviate from the random graphs typically considered in mathematical analysis @xcite is the degree distribution , that is , the distribution of the number of links of the nodes @xcite . in binomial random graphs ,
all nodes have similar degrees , while many real world networks have some nodes that are significantly more connected than others .
specifically , many complex networks , termed scale - free , have degree distributions that decay as a power law . a plausible mechanism for such a phenomenon is preferential attachment @xcite , that is , the tendency to connect preferentially to nodes with already high degrees .
to gain greater insight into the structure and evolution of the air transportation network , we calculate the degree distribution of the cities .
the degree of a city is the number of other cities to which it is connected by a non - stop flight . in fig .
[ fig1]a , we show the cumulative degree distribution gives the probability that a city has @xmath10 or more connections to other cities , and is defined as @xmath11 , where @xmath12 is the probability density function . ] for the world - wide air transportation network .
the data suggest that @xmath13 has a truncated power - law decaying tail @xcite @xmath14 where @xmath15 is the power law exponent , @xmath16 is a truncation function , and @xmath17 is a crossover value that depends on the size of the network . the measured value of the exponent @xmath18 would imply that , as one increases the size of the network , the average degree of the cities are also expected to increase @xcite .
the degree of a node is a source of information on its importance .
however , the degree does not provide complete information on the _ role _ of the node in the network . to start to address this issue
, we consider the `` betweenness centrality '' of the cities comprising the world - wide air transportation network .
the betweenness @xmath19 of city @xmath20 is defined as the number of shortest paths connecting any two cities that involve a transfer at city @xmath20 @xcite .
we define the normalized betweenness as @xmath21 , where @xmath22 represents the average betweenness for the network .
we plot , in fig .
[ fig1]b , the cumulative distribution @xmath23 of the normalized betweenness for the world - wide air transportation network .
our results suggest that the distribution of betweennesses for the air transportation network obeys the functional form @xmath24 where @xmath25 is the power law exponent , @xmath26 is a truncation function , and @xmath27 is a crossover value that depends on the size of the network .
a question prompted by the previous results regarding the degree and the centrality of cities is : `` are the _ most connected _ cities also the _ most central _ ? '' to answer this question , we analyze first the network obtained by randomizing the world - wide air transportation network ( fig .
[ fig1]b ) .
we find that the distribution of betweennesses still decays as a power law but , in this case , with a much larger exponent value @xmath28 .
this finding indicates the existence of anomalously large betweenness centralities in the air transportation network .
for the randomized network , the degree of a node and its betweenness centrality are strongly correlated
, highly connected nodes are also the most central ( fig .
[ fig3]a ) .
in contrast , for the world - wide air transportation network it turns out that there are cities that are not hubs
, have small degrees but that nonetheless have very large betweennesses ( fig . [ fig3]a ) .
to better illustrate this finding , we plot the 25 most connected cities and contrast such a plot with another of the 25 most `` central '' cities according to their betweenness ( figs .
[ fig3]b and c ) .
while the most connected cities are located mostly in western europe and north america , the most central cities are distributed uniformly across all the continents .
significantly , each continent has at least one central city , which is typically highly - connected when compared to other cities in the continent johannesburg in africa or buenos aires and so paulo in south america .
interestingly , besides these cities with relatively large degree , there are others such as anchorage ( alaska , u.s . ) and port moresby ( papua new guinea)that , despite having small degrees are among the most central in the network ( table 1 ) .
nodes with small degree and large centrality can be regarded as anomalies .
other complex networks that have been described in the literature , like the internet @xcite , do not display such a behavior , and nodes with the highest degree are also those with the highest betweenness @xcite .
it is , in principle , easy to construct a network in which a node has small degree and large centrality
think , for example , of a network formed by two communities that are connected to one another through a single node with only two links .
the relevant question is , however , `` what general and plausible mechanism would give rise to scale - free networks with the obtained anomalous distribution of betweenness centralities ? '' to answer this question it is useful to consider a region such as alaska .
alaska is a sparsely populated , isolated region with a disproportionately large for its population size
number of airports .
most alaskan airports have connections only to other alaskan airports .
this fact makes sense geographically . however , distance - wise it would also make sense for some alaskan airports to be connected to airports in canada s northern territories .
these connections are , however , absent .
instead , a few alaskan airports , singularly anchorage , are connected to the continental us .
the reason is clear , the alaskan population needs to be connected to the political centers , which are located in the continental us , while there are political constraints making it difficult to have connections to cities in canada , even to ones that are close geographically @xcite .
it is now obvious why anchorage s centrality is so large .
indeed , the existence of nodes with anomalous centrality is related to the existence of regions with a high density of airports but few connections to the outside .
the degree - betweenness anomaly is therefore ultimately related to the existence of `` communities '' in the network .
the unexpected finding of central nodes with low degree is a very important one because central nodes play a key role in phenomena such as diffusion and congestion @xcite , and in the cohesiveness of complex networks @xcite .
therefore , our finding of anomalous centralities points to the need to ( i ) identify the communities in the air transportation network and ( ii ) establish new ways to characterize the role of each city based on its pattern of intra- and inter - community connections and not merely on its degree .
_ community structure _ to identify communities in the air transportation network , we use the definition of modularity introduced in refs .
the modularity of a given partition of the nodes into groups is maximum when nodes that are densely connected among them are grouped together and separated from the other nodes in the network . to find the partition that maximizes the modularity , we use simulated annealing @xcite .
we display in fig .
[ modules ] the communities identified by our algorithm in the world - wide air transportation network . as we surmised , both alaska and papua new guinea form separate communities .
this fact explains the large betweenness centrality of anchorage and port moresby , as they provide the main links to the outside world for the other cities in their communities .
another significant result is that even though geographical distance plays a clear role in the definition of the communities , the composition of some of the communities can not be explained by purely geographical considerations .
for example , the community that contains most cities in europe also contains most airports in asian russia .
similarly , chinese and japanese cities are mostly grouped with cities in the other countries in southeast asia , but india is mostly grouped with the arabic peninsula countries and with countries in northeastern africa .
these facts are consistent with the important role of political factors in determining community structure @xcite . _
global role of cities _ we characterize the role of each city in the air transportation network based on its pattern of intra- and inter - community connections .
we first distinguish nodes that play the role of hubs in their communities from those that are non - hubs .
note that cities like anchorage are hubs in their communities but they are not hubs if one considers all the nodes in the network .
thus , we define the within - community degree of a node .
if @xmath29 is the number of links of node @xmath20 to other nodes in its community @xmath30 , @xmath31 is the average of @xmath32 over all the nodes in @xmath30 , and @xmath33 is the standard deviation of @xmath32 in @xmath30 , then @xmath34 is the so - called @xmath35-score .
the within - community degree @xmath35-score measures how `` well - connected '' node @xmath20 is to other nodes in the community .
we then distinguish nodes based on their connections to nodes in communities other than their own .
for example , two nodes with the same @xmath35-score will play different roles if one of them is connected to several nodes in other communities while the other is not .
we define the participation coefficient @xmath36 of node @xmath20 as @xmath37 where @xmath38 is the number of links of node @xmath20 to nodes in community @xmath39 , and @xmath40 is the total degree of node @xmath20 .
the participation coefficient of a node is therefore close to one if its links are uniformly distributed among all the communities , and zero if all its links are within its own community .
we hypothesize that the role of a node can be determined , to a great extent , by its within - module degree and its participation coefficient @xcite .
we define heuristically seven different `` universal roles , '' each one corresponding to a different region in the @xmath41 phase - space . according to the within - module degree ,
we classify nodes with @xmath42 as module hubs and nodes @xmath43 as non - hubs .
both hub and non - hub nodes are then more finely characterized by using the values of the participation coefficient @xcite .
we divide non - hub nodes into four different roles : ( r1 ) _ ultra - peripheral nodes _ ,
i.e. , nodes with all their links within their module ( @xmath44 ) ; ( r2 ) _ peripheral nodes _ , i.e. , nodes with most links within their module ( @xmath45 ) ; ( r3 ) _ non - hub connector nodes _ ,
i.e. , nodes with many links to other modules ( @xmath46 ) ; and ( r4 ) _ non - hub kinless nodes _ , i.e. , nodes with links homogeneously distributed among all modules ( @xmath47 ) .
we divide hub nodes into three different roles : ( r5 ) _ provincial hubs_. i.e. , hub nodes with the vast majority of links within their module ( @xmath48 ) ; ( r6 ) _ connector hubs _ , i.e. , hubs with many links to most of the other modules ( @xmath49 ) ; and ( r7 ) _ kinless hubs _ , i.e. , hubs with links homogeneously distributed among all modules ( @xmath50 ) . for each city in the world - wide air transportation network
, we calculate its within - community degree @xmath51 and its participation coefficient @xmath36 .
then , we assign each city a role according to the definitions above ( figs .
[ f - roles]a and c ) .
significantly , 95.4% of the cities in the world - wide air transportation network are classified as either peripheral or ultra - peripheral .
additionally , there is a small fraction of non - hub connectors ( 0.5% ) .
this result suggests that cities which are not hubs in their respective communities rarely have links to many other communities in the air transportation network .
this situation is in stark contrast to what happens in some biological networks , in which non - hub connectors seem to be relatively frequent and to play an important role @xcite .
the remaining 4.1% of the nodes are hubs .
we find approximately equal fractions of provincial and connector hubs .
the former include cities that , for historical , political , or geographical reasons , are comparatively not well - connected to other communities .
examples are denver , philadelphia , and detroit , in north america ; stuttgart , copenhagen , istanbul , and barcelona , in the community formed by europe , north africa and the former soviet union ; adelaide and christchurch in oceania ; brasilia in south america ; fairbanks and juneau in alaska ; and the already discussed case of port moresby .
connector hubs include the most recognizable airport hubs in the word : chicago , new york , los angeles , and mexico city in north america ; frankfurt , london , paris , and rome in europe ; beijing , tokyo , and seoul in the south - eastern asian community ; delhi , abu dhabi , and kuwait in the community comprising india , the arabic peninsula and north - eastern africa ; buenos aires , santiago , and so paulo in south america ; melbourne , auckland , and sydney in oceania ; and anchorage in alaska .
the fractions of cities with each role in the world - wide air transportation network contrast with the corresponding fractions in a randomization of the network ( fig .
[ f - roles]b ) . in this case , the community identification algorithm still yields certain communities , but the network lacks `` real '' community structure .
the identification of roles enables one to realize that these communities are somehow artificial .
indeed , many cities are either kinless hubs or kinless non - hubs due to the absence of a real community structure , and the network contains essentially no provincial or connector hubs .
we carried out a `` systems '' analysis of the structure of the world - wide air transportation network .
the study enables us to unveil a number of significant results .
the world - wide air transportation network is a small - word network in which : ( i ) the number of non - stop connections from a given city , and ( ii ) the number of shortest paths going through a given city have distributions that are scale - free . surprisingly , the nodes with more connections are not always the most central in the network .
we hypothesize that the origin of such a behavior is the multi - community structure of the network .
we find the communities in the network and demonstrate that their structure can only be understood in terms of both geographical and political considerations .
our analysis of the community structure of the air transportation network is important for two additional reasons .
first , it allow us to identify the most efficient ways in which to engineer the structure of the network .
specifically , having identified the communities , one can identify which ones are poorly connected and the ways to minimize that problem .
second , cities that connect different communities play a disproportionate role in important dynamic processes such as the propagation of infections such as sars . as , we show , finding the communities is the first step toward identifying these cities . the existence of communities and the understanding that different cities may have very different impacts on the global behavior of the air transportation system , call for the definition of the role of each city .
we address this issue by classifying cities into seven roles , according to their patterns of inter- and intra - community connections .
we find that most of the nodes ( 95% ) are peripheral , that is , the vast majority of their connections are within their own communities .
we also find that nodes that connect different communities are typically hubs within their own community , although not necessarily global hubs .
this finding is in stark contrast with the behavior observed in certain biological networks , in which non - hub connectors are more frequent @xcite .
the fact that different networks seem to be formed by nodes with network - specific roles points to the more general question of what evolutionary constraints and pressures determine the topology of complex networks , and how the presence or absence of specific roles affects the performance of these networks .
we thank a. arenas , a. barrat , m. barthlmy , a. daz - guilera , a. a. moreira , r. pastor - satorras , m. sales - pardo , d. stouffer , and a. vespignani for stimulating discussions and helpful suggestions .
we also thank oag for making their electronic database of airline flights available to us , and landings.com for providing us with the geographical coordinates of the world airports .
is scaled by the average degree @xmath35 of the network .
the distribution displays a truncated power law behavior with exponent @xmath15 .
( b ) cumulative distribution of normalized betweennesses plotted in double logarithmic scale .
the distribution displays a truncated power law behavior with exponent @xmath25 . for a randomized network with exactly the same degree distribution as the original air transportation network
, the betweenness distribution decays with an exponent @xmath28 .
a comparison of the two cases clearly shows the existence of an excessive number of large betweenness values in the air transportation network .
, scaledwidth=40.0% ] phase - space corresponds to a city , and different colors indicate different roles .
most cities are classified as ultra - peripheral ( black ) or peripheral ( red ) nodes .
a small number of non - hub nodes play the role of connectors ( green ) .
we find approximately equal fractions of provincial ( yellow ) and connector ( brown ) hubs .
( b ) same as ( a ) but for a randomization of the air transportation network .
the absence of communities manifests itself in that most hubs become kinless hubs ( gray ) and in the appearance of kinless non - hubs ( blue ) .
( c ) non - hub connectors ( green ) , provincial hubs ( yellow ) , and connector hubs ( brown ) in the world - wide air transportation network . ] | we analyze the global structure of the world - wide air transportation network , a critical infrastructure with an enormous impact on local , national , and international economies .
we find that the world - wide air transportation network is a scale - free small - world network .
in contrast to the prediction of scale - free network models , however , we find that the most connected cities are not necessarily the most central , resulting in anomalous values of the centrality .
we demonstrate that these anomalies arise because of the multi - community structure of the network .
we identify the communities in the air transportation network and show that the community structure can not be explained solely based on geographical constraints , and that geo - political considerations have to be taken into account .
we identify each city s global role based on its pattern of inter- and intra - community connections , which enables us to obtain scale - specific representations of the network . like other critical infrastructures
, the air transportation network has enormous impact on local , national , and international economies .
it is thus natural that airports and national airline companies are often times associated with the image a country or region wants to project @xcite .
the air transportation system is also responsible , indirectly , for the propagation of diseases such as influenza and , recently , sars .
the air transportation network thus plays for certain diseases a role that is analogous to that of the web of human sexual contacts @xcite for the propagation of aids and other sexually - transmitted infections @xcite .
the world - wide air transportation network is responsible for the mobility of millions of people every day .
almost 700 million passengers fly each year , maintaining the air transportation system ever so close to the brink of failure .
for example , us and foreign airlines schedule about 2,700 daily flights in and out of ohare alone , more than 10% of the total commercial flights in the continental us , and more than the airport could handle even during a perfect `` blue - sky '' day . low clouds , for example , can lower landing rates at ohare from 100 an hour to just 72 an hour , resulting in delays and flight cancellations across the country .
the failures and inefficiencies of the air transportation system have large economic costs ; flight delays cost european countries 150 to 200 billion euro in 1999 alone @xcite .
these facts prompt several questions : what has led the system to this point ?
why ca nt we design a better system ? in order to answer these questions , it is crucial to characterize the structure of the world - wide air transportation network and the mechanisms responsible for its evolution .
the solution to this problem is , however , far from simple .
the structure of the air transportation network is mostly determined by the concurrent actions of airline companies both private and national that try , in principle , to maximize their immediate profit . however , the structure of the network is also the outcome of numerous historical `` accidents '' arising from geographical , political , and economic factors .
much research has been done on the definition of models and algorithms that enable one to solve problems of optimal network design @xcite . however , a world - wide , `` system '' level , analysis of the structure of the air transportation network is still lacking .
this is an important caveat .
just as one can not fully understand the complex dynamics of ecosystems by looking at simple food chains @xcite or the complex behavior in cells by studying isolated biochemical pathways @xcite , one can not fully understand the dynamics of the air transportation system without a `` holistic '' perspective .
modern `` network analysis '' @xcite provides an ideal framework within which to pursue such a study .
we analyze here the world - wide air transportation network .
we build a network of 3883 locales , villages , towns , and cities with at least one airport and establish links between pairs of locales that are connected by non - stop passenger flights .
we find that the world - wide air transportation network is a small - world network @xcite for which ( i ) the number of non - stop connections from a given city , and ( ii ) the number of shortest paths going through a given city have distributions that are scale - free .
in contrast to the prediction of scale - free network models , we find that the most connected cities are not necessarily the most `` central , '' that is , the cities through which most shortest paths go .
we show that this surprising result can be explained by the existence of several distinct `` communities '' within the air transportation network .
we identify these communities using algorithms recently developed for the study of complex networks , and show that the structure of the communities can not be explained solely based on geographical constraints , and that geo - political considerations must also be taken into account .
the existence of communities leads us to the definition of each city s global role , based on its pattern of inter- and intra - community connections . |
the reference interaction site model ( rism ) theory , formulated in the early 70 s by chandler and andersen @xcite , is still one of the most adopted structural theories of molecular fluids . in the rism approach , developed as a generalization of the ornstein - zernike theory of atomic fluids @xcite ,
the molecules are represented as a collection of spherical interaction sites , rigidly connected so to reproduce a given molecular geometry @xcite .
originally the theory was developed to deal only with hard - sphere fluids @xcite or with simple lennard - jones systems @xcite .
later on , the rism framework has been extended to take into account more realistic representations of complex fluids , including water , methanol and other compounds ( see refs @xcite and a general review in ref .
more recently , the rism approach ( along with its extensions and generalizations ) has been adopted for predicting structure , thermodynamic and phase equilibria of a large variety of systems , including dumbbell fluids @xcite , colloids @xcite , proteins @xcite , surfactants @xcite and water solutions @xcite . in the last years
further refinements of the original rism theory have made possible the development of more complex approaches , as for instance the three dimensional reference interaction site model theory @xcite and the multi - center molecular ornstein - zernike equation @xcite .
the structure and the phase behavior of complex fluids is of great interest in the field of liquid and soft matter physics .
this is expecially due to the peculiar thermodynamic and structural properties exhibited by some of these fluids , named anomalous liquids .
water is the most known example of such fluids ; its peculiar behaviors , including re - entrant melting and density anomalies , have been largely investigated both from experimental @xcite and theoretical @xcite points of view .
the existence of these anomalies has been traditionally related to the possibility to develop a network between water molecules through hydrogen bonds . within this framework
a modeling approach , which is arousing an increasing interest in the last years , is based on spherically symmetric core - softened ( cs ) potentials : according to this approach , formulated by hemmer and stell in 1970 @xcite , in these potentials the hard - core interactions turn to be softened and an attractive tail is set .
a large variety of studies has been carried out to investigate the peculiar physical properties of these potentials , with a particular focus on liquid - liquid phase transition @xcite and water - like anomalies @xcite . in this context , the rism theory has been successfully adopted , along with the mode - coupling theory , to study the pressure dependence of the diffusion coefficient for acetonitrile and methanol in water @xcite and to analyze the phase behavior of a number of cs molecular models @xcite . in this work
we investigate structure and thermodynamics of a simple model for a dimeric fluid interacting via a modified inverse - power potential ( mip ) @xcite by means of rism theory and monte carlo ( mc ) simulations . in a previous study @xcite
we have shown that a system of dimeric particles interacting through mip diplays typical water - like anomalies , like a temperature of maximum density ( tmd ) and a non - monotonic behavior of the pair translational entropy .
also , we have investigated how the existence of such anomalies depends on the elongation of the dimers and on the interaction potential parameters .
in particular , upon increasing the repulsion softening , it has been possible to study how the fluid structure is changed when going from one - scale behavior typical of lennard - jones fluids to a two - scale behavior characterizing the cs systems . here
we extensively employ the rism framework for calculating structural and thermodynamic properties of the dimer fluid interacting via mip , in the whole range of elongations and for two different values of the repulsion softening .
for this purpose we adopt and compare hypernetted chain ( hnc ) @xcite and kovalenko - hirata ( kh ) @xcite approximation closures and various routes from the structure to thermodynamics .
where possible , theoretical results are assessed against previous @xcite or newly generated mc simulations performed in the canonical ensemble . this work has been carried out with the aim to investigate the anomalous behaviors of the dimer fluid in a wide range of parameters ; also , to the best of our knowledge , this study constitutes the first application of the rism approach to investigate density and structural anomalies in a molecular fluid interacting via a cs potential .
the paper is organized as follows : in the next section we provide details of the model , the rism theory and the simulation approach .
results are presented and discussed in the third section and conclusions follow in the last section . [ cols="^,^,^ " , ] kh predictions for the tmd in the whole range of @xmath0 values
are collectively reported in fig [ fig : tmd ] : for @xmath3 ( a ) we observe a progressive shift of the density anomaly regions towards very low temperatures , with a `` saturation '' for @xmath4 .
at the same time , pressure attains progressively lower values .
a similar scenario holds for @xmath5 ( b ) , but in this case the saturation effect is not observed and the density anomaly region moves towards lower and lower values of @xmath6 . in comparison with simulations ,
the theory predicts the existence of the tmd for all values of @xmath7 for both @xmath3 and @xmath5 . on the other hand
, simulations predict the existence of the tmd only if @xmath8 for @xmath3 and @xmath9 for @xmath5 . at higher values of @xmath0
the development of the density anomaly is likely prevented by the onset of solid phases .
however , according to kh predictions , the system persists in a fluid phase down to very low temperatures and hence the development of the tmd is not hampered by a crystallization .
this observation can be checked by looking at the behavior of the molecular structure factor @xmath10 that , for a biatomic molecule , is equivalent to the structure factor between the centers of mass of two dimers . according to the qualitative criterion developed by hansen and verlet for the lennard - jones potential @xcite ,
a fluid system approaches a solid phase when the height of the first peak of the @xmath10 exceed 2.85 .
a similar criterion is expected to roughly hold even if a cs potential is adopted . in fig .
[ fig : sk ] we compare mc and kh @xmath10 obtained for @xmath11 , @xmath5 and @xmath12 : simulation data show a progressive increase of the first peak of the @xmath10 until it overcomes 3 for densities between 0.40 and 0.60 and the system becomes solid - like . for
density higher than 0.60 , the first peak decreases and the system becomes fluid again . according to the kh predictions ,
the increase of the peak is equally observed but it never becomes greater than 3 .
rather , it approaches a maximum at 2.5 and then decreases . from mc simulations ( circles ) and
kh ( squares ) for @xmath11 , @xmath5 and @xmath13 .
dashed lines are guides for the eye.,width=302 ] finally , we have calculated the residual multiparticle entropy @xmath14 defined as : @xmath15 with @xmath16 where @xmath17 is the excess free energy per particle and @xmath18 is the excess entropy per particle . according to the criterion developed by giaquinta and giunta
@xcite @xmath14 vanishes whenever a fluid system approaches the solidification .
it has been shown that the validity of this criterion holds for softly repulsive fluids also @xcite . in fig .
[ fig : deltas ] we report the residual multiparticle entropy obtained from mc and kh for @xmath11 , @xmath5 and @xmath13 .
simulation data indicate that @xmath14 crosses the zero line for @xmath19 , whereas kh theory never predicts that @xmath14 vanishes .
such a result agrees with the behavior of @xmath10 reported in fig .
[ fig : sk ] .
the emerging picture confirms that kh , while providing for the reentrant effect typical of anomalous fluids , underestimates the onset of the solid phase . summarizing , according to results presented in figs .
[ fig : sk ] and [ fig : deltas ] for @xmath11 , @xmath5 and @xmath13 , the system freezes for @xmath19 : on the other hand , the theoretical underestimation of the onset of the solid phase causes the density anomaly is still observed in the kh context .
we have investigated density and structural anomalies in a dimeric fluid interacting via a purely repulsive core - softened potential by means of integral equation theories . starting with a simple monomeric case
we have progressively increased the elongation @xmath0 till to obtain a tangent configuration .
we have also considered two different conditions , corresponding to one ( @xmath3 ) and two ( @xmath5 ) length scales of the intermolecular potential .
theoretical results have been systematically compared with already existing or newly generated monte carlo ( mc ) simulations .
we have found that in the monomeric case the ornstein - zernike equation , coupled with hypernetted chain and kovalenko - hirata ( kh ) closures is able to accurately reproduce simulation data for internal energy and pressure if the virial equation of state is adopted .
nevertheless , the loci of density anomalies are underestimated by both the closures . upon increasing @xmath0
we have implemented the reference interaction site model ( rism ) theory coupled with kh closure ; also , we have adopted a closed form for the calculation of pressure , available in the context of this closure .
the theory is still able to reproduce simulation results for pressure , being less predictive for the internal energy .
the fluid structure has been investigated by computing the site - site radial distribution function @xmath20 : mc and kh agree in providing a double - peak structure for @xmath5 , such a feature being strongly reminiscent of the double length scales of the interaction potential .
this is not the case of @xmath3 , where only a single peak in @xmath20 is observed .
structural and density anomalies are found for both @xmath3 and @xmath5 : the theory qualitatively agrees with simulations in predicting their existence , but systematically underestimates the temperatures where they occur . in comparison with simulations , where the density anomaly is observed for @xmath8 if @xmath3 and for @xmath9 if @xmath5 , kh predicts the existence of this anomaly in the whole range of elongations .
this is due to the underestimation of the solid phase , that in simulations prevents the development of this anomaly for higher values of @xmath0 .
overall , the theory proves to be reliable in providing the existence of structural and thermodynamic anomalies even if a quantitative prediction is not achieved .
we emphasize that this is the first application of the rism approach to investigate these anomalies in a fluid interacting through a core - softened potential .
the possibility to have a reliable theoretical tool able to provide a quick estimate of these anomalies , whereas simulations can become more time demanding , set integral equations as promising candidates in describing fluid structure and phase behaviors of many complex fluids , such as elongated molecules , polymers and colloidal dimers . | we study fluid structure and water - like anomalies of a system constituted by dimeric particles interacting via a purely repulsive core - softened potential by means of integral equation theories . in our model
, dimers interact through a repulsive pair potential of inverse - power form with a softened repulsion strength . by employing the ornstein - zernike approach and the reference interaction site model ( rism ) theory ,
we study the behavior of water - like anomalies upon progressively increasing the elongation @xmath0 of the dimers from the monomeric case ( @xmath1 ) to the tangent configuration ( @xmath2 ) . for each value of the elongation we consider two different values of the interaction potential , corresponding to one and two length scales , with the aim to provide a comprehensive description of the possible fluid scenarios of this model .
our theoretical results are systematically compared with already existing or newly generated monte carlo data : we find that theories and simulations agree in providing the picture of a fluid exhibiting density and structural anomalies for low values of @xmath0 and for both the two values of the interaction potential .
integral equation theories give accurate predictions for pressure and radial distribution functions , whereas the temperatures where anomalies occur are underestimated . upon increasing the elongation , the rism theory still predicts the existence of anomalies ; the latter are no longer observed in simulations , since their development is likely precluded by the onset of crystallization .
we discuss our results in terms of the reliability of integral equation theories in predicting the existence of water - like anomalies in core - softened fluids .
[email protected] ] |
surface resistance is very important in the operation of superconducting cavity @xcite .
as one can see from equation ( 1 ) , surface resistance is the function of a temperature dependent term and a temperature independent one .
temperature dependent resistance simply decreases as temperature goes down in a metal since the number of phonons decreases according to boltzmann factor @xcite .
@xmath0 practically , the operating temperature of a superconducting cavity is usually determined in such a way to achieve possibly low surface resistance as long as an economical issue allows . in other words
, the operating temperature can not be infinitely low since lowering temperature always demands substantial cost .
for the raon accelerator , a quarter wave resonator ( qwr ) cavity will be operated at 4k while the rest cavities will be operated at 2k @xcite .
unlike the temperature dependent resistance , the temperature independent resistance , which is called a residual resistance , is due to various factors .
it has been reported that defects acting scattering centers contribute to the residual resistance such as inclusions , voids , dislocations with a crystal lattice itself @xcite .
thus , lowering residual resistance is essential to achieve low surface resistance .
basically , the more pure material is , the lower residual resistance one can obtain . since the residual resistance ratio ( rrr ) is defined as the ratio of a resistance at 300k to a residual resistance ( resistance at or just above critical temperature , @xmath1 of nb @xmath2 9.2k ) @xcite , one can have high rrr with the more pure material .
fabrication process of superconducting cavities is very complicated since it requires mechanical ( pressing ) , electrical ( electron beam welding ) , and chemical ( polishing ) process @xcite . among these steps ,
the electrical process ( e - beam welding ) is of our interest in this paper .
generally , the rrr degrades during e - beam welding due to the introduction of impurities from surroundings such as oxygen , nitrogen , and hydrogen @xcite . although previous studies reported that a vacuum level of a chamber is the most critical , but still e - beam welding conditions , such as e - beam power and welding speed , are considered important as much as the vacuum level because they make a great effect on the heat affected zone ( haz ) in a welding part @xcite . in this paper , we report rrr degradation occurred during e - beam welding of niobium samples .
we measured rrr with different conditions ; vacuum level , e - beam power , and welding speed .
not only we compared rrr degradation , but also we analyzed a heat affected zone ( haz ) .
thus , we finally report how rrr degradation and heat affected zone were affected by different e - beam welding conditions .
a rrr 300 grade niobium sheet was from ati , wah chang inc .
( albany , usa . ) with the size of @xmath3 mm@xmath4 ( @xmath5 ) .
the chemical composition and the grain size of nb sheet satisfies each astm @xmath6 and astm @xmath7 .
samples for experiments were prepared from two different companies ( a & b ) and the sequences are as follows .
first , two companies cut the nb sheet of 3 mm in thickness into two different sets of width and length for butt welding @xcite : two nb pieces of @xmath8 mm@xmath9 by company a and two nb pieces of @xmath10 mm@xmath9 by company b. thus , the company a prepared the welded nb sheet of @xmath11 mm@xmath4 , and the company b prepared the welded nb sheet of @xmath12 mm@xmath9 . then , rrr samples were cut out of these welded nb sheets perpendicular to a welding line .
each nb sample was cut by electrical discharge machining ( edm ) @xcite with the size of @xmath13 mm@xmath4 .
[ p1 ] shows how the pieces of nb sheet were welded , and samples were cut out from the welded nb sheet by the company b. also , fig .
[ p2 ] shows the layout of cut samples , all samples were perpendicular to the welding line , and the distance from the weld seam increased from left to right ( 0 , 2 , 4 , 6 , 8 , 10 , and 15 , 20 , 25 mm ) . as shown in fig .
[ p2 ] , in order to compare differences in rrr between the welding part and the bead part , two types of samples were cut from the welding side and the bead side . since nine samples were cut out from each side , eighteen samples in total were prepared from both sides of one welded nb sheet .
welding conditions of two companies are summarized in table [ t1 ] . in this study
, we defined the rrr value as the ratio of the nb resistance at 300k to the nb resistance at the temperature just above critical temperature ( @xmath14 @xmath2 9.2k ) where nb still is in the normal conducting state . that temperature ranges from 9.5k to 10k .
the input current was 5 ma , and rrr measurement was performed with no magnetic field application . since the generated power in the sample ( @xmath2 10@xmath15 watt ) due to this level of input current is negligible compared to the cooling power of sample stage in ppms ( @xmath2 10@xmath16 watt ) , the 5ma input current level dose not vary the resistance of the sample while measurement . the time for stabilizing each temperature step was between 100 @xmath2 200 sec
the temperature increments were 0.1k up to 15k , 1k from 16k to 30k , and 5k from 35k to 300k .
rrr experiments were carried out with physical property measurement system ( ppms ) operated by the korea basic science institute ( kbsi ) .
mm@xmath9 was butt - welded by joining two nb sheet of @xmath10 mm@xmath9 .
weld side ( left ) , bead side ( right ) .
, width=604 ] mm@xmath4 for rrr measurement . ,
width=604 ] .welding conditions for sample preparation [ cols="^,^,^,^,^,^ " , ] [ t3 ] from the fig .
[ p6 ] , rrr degradation of the samples from company a occurred less than company b. this is in good agreement with previous studies @xcite since rrr degradation is greatly affected by a vacuum level .
as shown in table [ t1 ] , the vacuum levels of two companies each are @xmath17 torr for company a , and @xmath18 torr for company b. therefore , we could confirm that samples from company a showed higher rrr values ( less degradation ) than company b as the average rrr ( 292 vs. 264 ) , and the lowest rrr ( 253 vs. 234 ) .
the worst degradation rates based on the lowest points were each 20% and 44 % corresponding to company a and b. in fact , risp set the minimum rrr value as 275 , so only less than 8% of degradation is allowable in raon project .
although the degradation rate of samples from company a satisfied this criteria as 3% and 8% in the weld side ( these was based on the average rrr ) , the rest degradation rates did not satisfy the criteria .
we might think this was due to the vacuum level , that is to say , the vacuum level gauge did not read precisely the real vacuum level in the e - beam chamber , thus the real vacuum level during the welding did not reach @xmath17 torr .
so we should perform more experiments to confirm this result .
interesting results were that the rrr degradation in the weld side where the e - beam welding directly occurred was less than in the bead side for all samples regardless of companies . according to previous studies @xcite , the part where the e - beam welding directly occurred showed high rrr value since this part experienced purification due to the `` melting '' just like the zone melting method used in such as a nb purification and a semiconductor purification @xcite .
this result can be explained from the fact that a liquid phase can hold more solute atoms ( impurities ) than a solid phase , and the melting the solid into the liquid phase drives impurities from solid to liquid .
then , impurities in the liquid are driven into the solid quickly on cooling since the diffusivity of impurities in the liquid can be assumed evenly faster than in the solid .
thus , it makes adjacent solid part has more impurities @xcite .
therefore , the driven effect of impurities caused the weld side to be more purified than the bead side although these two parts experienced the melting event during the e - beam welding .
another result to be discussed here is a heat affected zone ( haz ) .
the rrr degradation occurs seriously in this region because grains in this region are greatly altered by a heat introduction depending on the e - beam power and the welding speed while rrr recovers as much as bulk material outside this region .
therefore , controlling ( narrowing ) the haz is an important issue for optimizing e - beam welding to keep rrr degradation above the target value . from the w. singer s work @xcite
, the haz was around 10 mm from the weld seam . by looking at fig .
[ p3 ] and fig .
[ p4 ] , the haz was around 15 mm where rrr started to recover as high as bulk s value for both companies .
in particular , the rrr of the bead side from company b was still far lower than the reference sample up to 25 mm , which means the haz expanded into the bulk level . one possible explanation is because too much heat was introduced in the welding part , and this heat did not spread out quickly through the whole sample during e - beam welding . in fact , nb has a low thermal conductivity , for example , the thermal conductivity of copper is larger than that of niobium by one order
. therefore , the welded part of nb did not have enough time to dissipate overheat quickly into the whole sample due to the low thermal conductivity .
consequently , the development of the optimized welding condition including the welding power and the welding speed should be established in order to achieve good haz ( narrow haz ) . as a next step ,
we need more experiments to analyze quantitatively how much heat was generated in the welding zone and how much grains were affected with the function of welding power and speed by using high quality optical microscope or sem .
the rrr measurements with the 300 rrr grade nb samples supplied from two companies of different welding conditions were carried out .
we confirmed that the vacuum level was critical factor to avoid rrr degradation since the degradation rate of company a having @xmath17 torr was lower than that of company b having @xmath18 torr .
also , we found that the degradation of the weld side where the melting directly occurred showed lower degradation than the bead side .
in addition , we found that the haz could be varied by controlling the welding condition , which was the heat introduction .
the haz expanded deeper into the bulk when the heat was introduced too much in the welding zone .
we thank kbsi for performing rrr tests for this study .
this work was supported by the rare isotope science project which is funded by the ministry of science , ict and future planning ( msip ) and the national research foundation ( nrf ) of the republic of korea under contract 2011 - 0032011 .
h. padamsee , j. knobloch , t. hays , _ rf superconductivity for accelerators _
( wiley - vch verlag gmbh @xmath19 co. kgaa , germany , 2008 ) .
h. padamsee , _ rf superconductivity , science , technolgoy , and applications _
( wiley - vch verlag gmbh @xmath19 co. kgaa , germany , 2009 ) . c. kittel , _ introduction to solid state physics _ ( john wiley @xmath19 son , inc . , united states , 1996 ) .
d. jeon _ et al .
_ , j. kor .
65 * , 1010 ( 2014 ) . c. r. barrett , _ the principles of engineering materials _
( prentice hall , united states , 1973 ) .
j. d. splett , d. f. vecchia , l. f. goodrich , j. research national inst .
standards and tech , * 116 * , 489 ( 2011 ) . c. xu , h. tian , c. e. reece , m. j. kelly , phys .
special topics - accelerators and beams , * 15 * , 043502 ( 2012 ) .
y. jung , h. j. kim , h. h. lee , h. c. yang , _ proc .
particle accel .
( dresden , germany , 2014 ) , p. 2549 . w. singer , x. singer , j. tiessen , h. wen , _ proc .
_ ( ny , usa , 2003 ) , p. 671 .
m. s. champion , _ et al .
_ , ieee trans .
superconductivity * 19 * , 1384 ( 2009 ) .
h. jiang , t. r. bieler , c. compton , t. l. grimm , _ proc .
ieee particle accel .
( oregon , usa , 2003 ) , p. 1359 . n. m. abbas , d. g. solomon , md .
f. bahari , int .
j. machine tools @xmath19 manufacturer , * 47 * , 1214 ( 2007 ) .
p. bauer , t. berenc , c. boffo , m. foley , m. kuchnir , y. tereshikin , t. wokas , _ proc .
11th workshop on rf superconductivity _ , ( lubeck , germany , 2003 ) , p. 588 . g. choi , j. lim , n. r. munirathnam , i. kim , metals and materials international , * 15 * , 385 ( 2009 ) . a. koethe , j. i. moench , materials trans , * 41 * , 7 ( 2000 ) . w. g. pfann , _ zone melting _ ( wiley , new york , 1966 ) . | the first heavy ion accelerator is being constructed by the rare isotope science project ( risp ) launched by the institute of basic science ( ibs ) in south korea .
four different types of superconducting cavities were designed , and prototypes were fabricated such as a quarter wave resonator ( qwr ) , a half wave resonator ( hwr ) and a single spoke resonator ( ssr ) .
one of the critical factors determining performances of the superconducting cavities is a residual resistance ratio ( rrr ) .
the rrr values essentially represent how much niobium is pure and how fast niobium can transmit heat as well . in general , the rrr degrades during electron beam welding due to the impurity incorporation .
thus it is important to maintain rrr above a certain value at which a niobium cavity shows target performance . in this study , rrr degradation related with electron beam welding conditions , for example , welding power , welding speed , and vacuum level will be discussed . |
in the first part we give a rigorous and concise construction of a space of transseries adequate for the study of a relatively large class of ordinary differential and difference equations .
the main stages of the construction and the notations are similar to those in @xcite , but the technical details rely on a generalized contractivity principle which we prove for abstract multiseries ( theorem [ pfp ] ) . we then discuss how various classes of problems are solved within transseries . the last part of the paper is devoted to a brief overview of generalized borel summation , an isomorphism used to associate actual functions to transseries , in the context of odes , difference equations and the extension of some of these techniques to pdes .
various spaces of transseries were introduced roughly at the same time in logic ( see @xcite ) , analysis @xcite , @xcite , and the theory of surreal numbers @xcite ; similar structures were introduced independently by berry and used as a powerful tool in applied mathematics , see @xcite .
informally , transseries are _
asymptotic , where @xmath0 if @xmath1 , finitely generated _ combinations powers , logarithms and exponentials .
it is a remarkable fact that a wide class of functions can be asymptotically described in terms of them .
as hardy noted @xcite `` no function has yet presented itself in analysis the laws of whose increase , in so far as they can be stated at all , can not be stated , so to say , in logarithmic - exponential terms '' .
this universality of representation can be seen as a byproduct of the fact that transseries are closed under many common operations .
it is convenient to take @xmath2 , @xmath3 , as the small parameter in the transseries .
when corresponding functions are considered we change variables so that the limit is again @xmath4 .
a simple example of a transseries generated by @xmath2 and @xmath5 is @xmath6 where @xmath7 .
this transseries is without logarithms or log - free , and has _ level _
one since it contains no iterated exponential ; the simplest nontrivial transseries of level two is @xmath8 .
a more complicated example of a transseries of exponential level two , with level zero _ generators _ @xmath2 and @xmath9 , level one generator @xmath10 , and level two generators @xmath11 and @xmath12 is @xmath13 some examples of transseries - like expressions for @xmath14 , which are in fact _ not _ transseries , are @xmath15 ( it fails the asymptoticity condition ) and @xmath16 ( it does not have finitely many generators , this property is described precisely in the sequel ) .
the underlying structure behind the condition of asymptoticity is that of _ well ordering_. in order to formalize transseries and study their properties , it is useful to first introduce and study more general abstract expansions , over a well ordered set
. let @xmath17 be an ordered set , with respect to @xmath18 .
if @xmath19 we write @xmath20 or @xmath21 .
@xmath17 is * totally ordered * if any two elements are comparable , i.e. , if for any @xmath22 we have @xmath23 or @xmath24 .
if @xmath17 is not totally ordered , it is called * partially ordered*. the set @xmath17 is * well ordered * with respect to @xmath25 if every nonempty totally ordered subset ( _ chain _ ) of @xmath17 has a minimal element , i.e. @xmath26 if any nonempty totally ordered subset of @xmath17 has a * maximal * element , we say that @xmath17 is well ordered with respect to @xmath27 .
@xmath17 has the * finite descending chain * property if there is no _ infinite strictly decreasing _ sequence in @xmath17 , in other words if @xmath28 is decreasing , then @xmath29 is constant for large @xmath30 .
[ wofc ] @xmath17 is well ordered with respect to `` @xmath25 '' _ iff _ it has the finite descending chain property .
a strictly decreasing infinite sequence is obviously totally ordered and has no minimal element . for the converse ,
if there exists @xmath31 such that @xmath32 then for @xmath33 , the sequence @xmath34 is an infinite descending chain in @xmath17 .
_ example : multi - indices .
_ @xmath35 is well ordered with respect to @xmath25 , and so is @xmath36 with respect to the order relation @xmath37 iff @xmath38 .
indeed an infinite descending sequence @xmath39 would be infinitely descending on at least one component . [ va ] let @xmath40 be fixed .
any infinite set @xmath17 in @xmath41 contains a _ strictly increasing _ ( infinite ) sequence .
the set @xmath17 is unbounded , thus there must exist at least one component @xmath42 so that the set @xmath43 is also unbounded ; say @xmath44 .
we can then choose a sequence @xmath45 so that @xmath46 is strictly increasing .
if the set @xmath47 is bounded , then there is a subsequence @xmath48 of @xmath49 so that @xmath50 is a constant vector .
then @xmath48 is a strictly increasing sequence . otherwise , one component , say @xmath51 is unbounded , and we can choose a subsequence @xmath52 so that @xmath53 is increasing .
the argument continues in this fashion until in at most @xmath54 steps an increasing sequence is constructed . [ nononcomp ]
any infinite set of multi - indices in @xmath55 contains at least two comparable elements .
[ lowerset ] let @xmath17 be a nonempty set of multi - indices in @xmath56 .
there exists a unique and * finite * minimizer set @xmath57 such that none of its elements are comparable and for any @xmath58 there is an @xmath59 such that @xmath60 .
consider the set @xmath61 of all _ maximal _ totally ordered subsets of @xmath17 ( every chain is contained in a maximal chain ; also , in view of countability , zorn s lemma is not needed ) .
let @xmath57 be the set of the least elements of these chains , i.e. @xmath62 . then @xmath57 is finite . indeed ,
otherwise , by corollary [ lowerset ] at least two elements in @xmath57 such that @xmath63 . but
this contradicts the maximality of the chain whose least element was @xmath64 .
it is clear that if @xmath65 is a minimizer then @xmath66 .
conversely if @xmath67 then @xmath68 contradicting the definition of @xmath57 .
if @xmath69 is a commutative group with an order relation , we call it an * abelian ordered group * if the order relation is compatible with the group operation , i.e. , @xmath70 ( e.g. @xmath71 or @xmath72 with addition ) .
let @xmath69 be an abelian ordered group and let @xmath73 be a _ decreasing group morphism _
, i.e. , \(1 ) @xmath74 .
\(2 ) @xmath75 .
\(3 ) @xmath76 .
then @xmath77 is the subgroup finitely generated by @xmath78 where @xmath79 is the unit vector in the direction @xmath80 in @xmath72 , and since @xmath81 it follows that @xmath82 in view of our final goal , a simple example to keep in mind is the multiplicative group of _ monomials _ @xmath83 , generated by the functions @xmath84 , @xmath85 and @xmath5 , for large positive @xmath86 .
the order relation on @xmath83 is @xmath87 if @xmath88 for all large @xmath86 . when @xmath89 we choose @xmath90 .
[ r0 ] the relation @xmath91 induces an equivalence relation on @xmath92 ; we denote it by @xmath93 .
for instance in @xmath83 , since @xmath94 and @xmath95 are not rationally independent , there exist distinct @xmath96 so that @xmath97 .
[ finigen ] clearly any choice of @xmath98 with @xmath99 for @xmath100 defines an order preserving morphism via @xmath101 .
ordered morphisms preserve well - ordering : [ ( 4 ) ] let @xmath102 be _ well ordered _
( an important example for us is @xmath103 ) and @xmath104 an order preserving morphism .
then @xmath105 is well ordered . assuming the contrary ,
let @xmath106 be such that @xmath107 is an infinite strictly _ ascending _ chain in @xmath105 .
then the index set @xmath108 is clearly infinite , and then , by proposition [ va ] it has a strictly increasing subsequence @xmath109 .
then @xmath110 is a _ descending _ subsequence of @xmath111 , which is a contradiction .
[ bilateral ] for any @xmath112 , the set @xmath113 is finite .
in particular , given @xmath114 , the set @xmath115 is finite . by proposition
[ va ] , the contrary would imply the existence of an strictly increasing subsequence of @xmath116 , for which then @xmath117 would be strictly decreasing , a contradiction .
the last part follows if we take @xmath118 .
[ 2.9 ] the space of formal series with usual addition and convolution ( [ innerprod1 ] ) .
] @xmath119 is an algebra with respect to componentwise multiplication by scalars , componentwise addition , and the inner multiplication @xmath120 where @xmath121 the same is true for @xmath122 factored by the equivalence relation @xmath123 replacing @xmath124 with @xmath125 in ( [ innerprod ] ) ( note that by corollary [ bilateral ] the equivalence classes have finitely many elements ) .
straightforward .
we define @xmath126 the linear subspace of @xmath127 for which @xmath128 is fixed .
the sum @xmath129 is in _ collected _ form if , by definition , @xmath130 , where the maximum is with respect to the lexicographic order .
( in other words the coefficients are collected and assigned to the earliest @xmath104 in its equivalence class . )
it is then natural to represent the equivalence class ( v. ( [ eq : defequi ] ) ) @xmath131 of @xmath132 , in @xmath133 , by @xmath134 .
[ collected ] by proposition [ 2.9 ] every nonzero sum can be written in collected form .
a topology is introduced in the following way : [ topology ] the sequence @xmath135 in @xmath126 converges in the asymptotic topology if for any @xmath136 , @xmath137 becomes constant ( with respect to @xmath30 ) eventually .
this induces a natural topology on @xmath133 .
this topology is metrizable .
indeed , any bounded @xmath138 such that @xmath139 iff all @xmath140 ( e.g. @xmath141 ) provides a translation - invariant distance @xmath142 where @xmath143 if @xmath144 and is one otherwise .
a cauchy sequence in @xmath126 is clearly convergent , and in this sense @xmath126 is a complete topological algebra . from this point on ,
we assume @xmath69 is a * totally ordered abelian group*. let @xmath145 .
a subgroup of @xmath69 generated by @xmath30 elements @xmath146 is totally ordered and well ordered , and thus can be indexed by a set of ordinals @xmath147 , in such a way that @xmath148 implies @xmath149 .
a sum @xmath150 where we agree to omit from @xmath151 all ordinals for which @xmath152 is called the _
asymptotic _ form of @xmath132 .
* dominant term . *
assume @xmath153 is presented in collected form .
the set @xmath154 is then totally ordered and must have a maximal element @xmath155 , by proposition [ ( 4 ) ] .
we say that @xmath156 is the dominant term of @xmath132 and @xmath157 is the _ ( dominant ) magnitude _ of @xmath132 ( equivalently , @xmath158 ) .
we allow for @xmath159 to be zero , * iff * @xmath160 .
the following property is an immediate consequence of corollar[collected ] : [ r2.19 ] for any nonzero @xmath132 we can write @xmath161 where all the terms in @xmath162 are less than one . [ rfinitude ] the magnitude is continuous : if @xmath163 and @xmath164 in the asymptotic topology , then @xmath165 ( i.e. @xmath166 so that @xmath167 ) .
this follows immediately from the definition of the topology and of @xmath168 .
the proposition below discusses the closure of @xmath126 under restricted _ infinite sums_. [ genclos ] let @xmath169 with @xmath170 and consider the sequence in @xmath126 @xmath171 and a fixed @xmath172 , @xmath173 then the sum ( `` blending '' ) @xmath174 obtained by replacing each @xmath175 in @xmath176 by the product @xmath177 is well defined in @xmath178 as the limit as @xmath179 of the convergent sequence of _ truncates _
@xmath180}(\tilde{s})=\sum_{\mathbf{l}_0\le\bk\le\bk'}c'_{\bk}\mu_{\bk}\cdot \tilde{s}^{(\mathbf{k})}= \sum_{\mathbf{j}\ge\bj_0}\mu_{\mathbf{j}}\sum_{b^{(\bk')}_\mathbf{j}}c'_{\mathbf{m}_1}c_{\mathbf{m}_2}^{(\mathbf{m}_1)}= \sum_{\mathbf{j}\ge\bj_0}d_{\mathbf{j}}^{(\bk')}\mu_{\mathbf{j}}\end{gathered}\ ] ] where @xmath181 [ p ] given @xmath182 , the coefficient @xmath183 is constant for large @xmath116 .
indeed , in the expression of @xmath184 we have @xmath185 and similarly @xmath186 and therefore there is a bound independent of @xmath116 on the number of elements in the set @xmath187 . on the other hand
, we obviously have @xmath188 if @xmath189 .
thus the set @xmath187 is constant if @xmath116 is large enough , and thus @xmath183 is constant for all large @xmath116 , which means the sums @xmath190}(\tilde{s})$ ] are convergent in the asymptotic topology . *
note*. the condition that @xmath191 decreases strictly in @xmath192 does * not * suffice for @xmath193 to be well defined .
indeed , the terms @xmath194 ( cf .
[ sfst ] ) have strictly decreasing magnitudes and yet the formal expression @xmath195 is meaningless .
[ dfp ] let @xmath108 be a linear operator from @xmath126 or from one of its subspaces , to @xmath126 , @xmath196 then @xmath108 is called asymptotically contractive on @xmath126 if @xmath197 thus @xmath198 we note that by ( [ econtractformula ] ) and proposition [ genclos ] , @xmath108 is well defined .
[ nonlinear.contrac ] the linear or nonlinear operator @xmath108 is ( asymptotically ) contractive in the set @xmath199 if @xmath200 and the following condition holds .
let @xmath201 and @xmath202 in @xmath17 be arbitrary and let @xmath203 then @xmath204 where @xmath205 for some @xmath206 .
[ rlinearity ] the sum of asymptotically contractive operators is contractive ; the composition of contractive operators , whenever defined , is contractive .
( i ) if @xmath108 is linear and contractive on @xmath126 then for any @xmath207 the fixed point equation @xmath208 has a unique solution @xmath209 .
\(ii ) in general , if @xmath210 is closed and @xmath200 is a ( linear or nonlinear ) contractive operator on @xmath17 , then @xmath211 has a unique solution is @xmath17 .
\(i ) uniqueness : if we had two solutions @xmath162 and @xmath212 we would get @xmath213 by ( [ econtractformula ] ) . we show that @xmath214 in the asymptotic topology implying that @xmath215 is convergent .
we have @xmath216 and since @xmath217 and since the set @xmath218 is finite by corollary [ bilateral ] , for each @xmath136 the condition @xmath219 becomes impossible if @xmath30 exceeds some @xmath220 , and then the coefficient of @xmath221 in @xmath222 is zero for @xmath223 .
\(ii ) uniqueness follows in the same way as in the linear case . for existence , prove the convergence of the recurrence @xmath224 . with @xmath225
we have , for some coefficients @xmath226 @xmath227 and therefore @xmath228 has an expression similar to ( [ ejn(delta ) ] ) , @xmath229 consequently @xmath230 and , as before , it follows that @xmath231 converges . [ compo_n ]
let @xmath232 be arbitrary and @xmath233 for @xmath234 .
then the operator defined by @xmath235 is contractive in the set @xmath236 .
we have @xmath237 let @xmath69 be a totally ordered abelian group .
we now define the algebra : @xmath238 modulo the obvious inclusions , and with the induced topology ( convergence in @xmath239 means convergence in one of the @xmath240 ) .
_ product form_. [ pstd ] any @xmath241 can be written in the form @xmath242 i.e. , @xmath243 where @xmath244 for some @xmath245 ( cf . also remark [ r2.19 ] ) and @xmath246 . we have , by remark [ r2.19 ] , @xmath247 where all the elements in the last sum are less than one .
let @xmath17 be the set of multi - indices in the sum in ( [ casei ] ) for which _ some _ @xmath248 and let @xmath249 be its minimizer in the sense of corollary [ lowerset ] , a finite set .
we now consider the extended set of generators @xmath250 we clearly have @xmath251 . by the definition of @xmath249 , for any term in the sum in ( [ casei ] ) either @xmath252 or else @xmath253 with @xmath254 and @xmath255 . in both cases we have @xmath256 with @xmath257 and @xmath258 .
thus @xmath132 can be rewritten in the form @xmath259 where the assumptions of the proposition are satisfied .
[ field ] @xmath239 is a field .
the only property that needs verification is the existence of a reciprocal for any nonzero @xmath132 . using proposition [ pstd ] we only need to consider the case when @xmath260 since multiplication by @xmath261 is manifestly contractive ( see [ subcontr ] ) , @xmath262 is the solution ( unique by proposition [ pfp ] ) of @xmath263 _ closure under infinite sums .
_ [ infisum ] ( i ) let @xmath264 and @xmath209 and @xmath265 be any sequence .
then @xmath266 \(ii ) more generally , if @xmath267 are of the form @xmath268 and @xmath269 is a multi - sequence of constants , then @xmath270 is well defined . for some @xmath271 we have @xmath272 .
since @xmath273 we have @xmath274 and by proposition [ genclos ] @xmath275 _ formal series with real coefficients .
order relation_. let @xmath276 the subfield consisting in @xmath277 which have real coefficients .
we say that @xmath278 if @xmath279 .
then every nonzero @xmath280 is either positive or else @xmath281 is positive .
this induces a _ total _ order relation on @xmath276 , by writing @xmath282 if @xmath283 .
transseries are constructed as a special instance of abstract series in which the ablelian ordered group is constructed inductively . in constructing spaces of transseries ,
one aims at constructing differential fields containing @xmath2 , closed under all operations of importance for a certain class of problems operations .
smaller closed spaces can be endowed with better overall properties .
the construction presented below differs in a number of technical respects from the one of calle , and the transseries space constructed here is smaller than his .
still some of the construction steps and the structure of the final object are similar enough to calle s , to justify using his terminology and notations .
* @xmath284 small transmonomial .
* @xmath285 large transmonomial .
* @xmath286 any transmonomial , large or small .
* @xmath287 small transseries .
* @xmath288 large transseries . *
@xmath289 any transseries , small or large .
let @xmath86 be large and positive and let @xmath69 be the totally ordered multiplicative group @xmath290 , with @xmath291 if @xmath292 as @xmath293 , i.e. , if @xmath294 .
the space of level zero log - free transseries is by definition @xmath295}=\stackrel{\approx}{s}(\mathcal{g})$ ] . by proposition
[ field ] , @xmath295}$ ] is a field . if @xmath296}$ ] , then @xmath297 iff @xmath298 for some @xmath299 , @xmath300 if @xmath301 and @xmath302 if @xmath303
. the general element of @xmath295}$ ] is a level zero transseries , @xmath304}$ ] or @xmath289 in short .
we have @xmath305 there are two order relations : @xmath27 and @xmath306 on @xmath296}$ ] .
we have @xmath307 iff @xmath308 ( the sign of the leading coefficient is immaterial ) and @xmath309 if ( @xmath310 and ) the real number @xmath311 is positive . [ smalllarge ]
a transseries is * small * , i.e. @xmath312 iff in ( [ defanytr ] ) we have @xmath313 whenever @xmath314 .
correspondingly , transseries is * large * , i.e. @xmath315 iff in ( [ defanytr ] ) we have @xmath313 whenever @xmath316 .
we note that @xmath312 iff @xmath317 ( there is an asymmetry : the condition @xmath318 does _ not _ imply @xmath315 , since it does not prevent the presence of small terms in @xmath289 ) .
any transseries can then be written uniquely as @xmath319 the set @xmath320}$ ] of transmonomials of exponentiality one consists by definition in the formal expressions @xmath321}=\anymo^{[0]}\exp({\largetr^{[0]}}),\ \
\anymo^{[0]},\largetr^{[0]}\in \tilde{\mathcal{t}}^{[0]}\ ] ] where we allow for @xmath322}=0 $ ] and set @xmath323 . with respect to the operation @xmath324}\exp(\largetr_1^{[0 ] } ) \anymo_2^{[0]}\exp(\largetr_2^{[0]})= ( \anymo_1^{[0]}\anymo_2^{[0]})\exp(\largetr_1^{[0]}+\largetr_2^{[0]})\ ] ] we see that @xmath320}$ ] is a commutative group .
the order relations are introduced in the following way . @xmath325 in particular , if @xmath326}$ ] is positive , then @xmath327}\exp(\largetr_1^{[0]})\gg 1 $ ] .
the second order relation , @xmath25 , is defined by @xmath324}\exp(\largetr_1^{[0]})>0\iff\anymo_1^{[0]}>0\ ] ] it is straightforward to check that @xmath328},\cdot,\gg)$ ] is an abelian ordered group .
the abelian ordered group of zero level monomials , @xmath329},\cdot,\gg)$ ] , is naturally identified with the set of transmonomials for which @xmath322}=0 $ ] .
the space @xmath330}$ ] of level one transseries is by definition @xmath331})$ ] . by proposition
[ field ] , @xmath330}$ ] is a field . by construction , the space @xmath295}$ ]
is embedded in @xmath330}$ ] .
formula ( [ defanytr ] ) is the general expression of a level one transseries , where now @xmath284 is a transmonomial of level one .
the two order relations on transseries are the ones induced by transmonomials , namely @xmath332 assuming the transseries of level @xmath333 are constructed , transseries of level @xmath30 together with the order relation , are constructed exactly as in [ level1 ] , replacing @xmath334 $ ] by @xmath335 $ ] and @xmath336 $ ] by @xmath337 $ ] .
the group @xmath320}$ ] of transmonomials of order at most @xmath30 consists in expressions of the form @xmath338}=x^{\sigma}\exp(\largetr^{[n-1]})\ ] ] where @xmath339}$ ] is either zero or a large transseries of level @xmath340 with the multiplication : @xmath341})x^{\sigma_2}\exp(\largetr_2^{[n-1 ] } ) = x^{\sigma_1+\sigma_2}\exp(\largetr_1^{[n-1]}+\largetr_2^{[n-1]})\ ] ] the order relation is given by @xmath342})\gg x^{\sigma_2}\exp(\largetr_2^{[n-1 ] } ) \iff \\
\big(\largetr_1^{[n-1]}>\largetr_2^{[n-1]}\big ) \ \mbox{or } \ \big(\largetr_1^{[n-1]}=\largetr_2^{[n-1 ] } \ \mbox{and}\ \sigma_1>\sigma_2\big ) \end{aligned}\ ] ] @xmath343}=\sum_{\bk\ge \bk_0}c_{\bk}(\smallmo^{[n]})^{\bk}\ ] ] as in [ level1 ] , @xmath344}$ ] is naturally embedded in @xmath345}$ ] .
this is the space of arbitrary level transseries , the inductive limit of the finite level spaces of transseries : @xmath347}\ ] ] clearly @xmath346 is a field .
the order relation is the one inherited from @xmath345}$ ] .
the topology is also that of an inductive limit , namely a sequence converges iff it converges in @xmath345}$ ] for _ some _ @xmath30 . .
the level @xmath348 of @xmath289 is @xmath30 if @xmath349}$ ] and @xmath350}$ ] .
[ rform0 ] if @xmath351 then @xmath352 .
we may clearly take @xmath353 . since ( by definition )
@xmath354 we must have , in particular , @xmath355 with @xmath356 .
by induction , and the assumption @xmath357 we must have @xmath358 and @xmath359 .
the proposition follows since , by again by the induction step , @xmath360 .
[ rform ] if @xmath286 is of level no less than @xmath361 , then either @xmath286 is large , and then @xmath362 or else @xmath286 is small , and then @xmath363 .
[ rfinige1 ] we can define _ generating monomials _ of @xmath364}$ ] a minimal subgroup @xmath365 of @xmath366}$ ] with the following properties : * @xmath367 ; * @xmath368 implies @xmath369 and , if @xmath370 , then @xmath371 . by induction
we see that @xmath372 is finitely generated for any @xmath373}$ ] .
[ pdifferentiation ] @xmath346 and @xmath374};\ n\in\nn$ ] are differential fields .
differentiation @xmath375 is introduced inductively on @xmath346 , as term by term differentiation , in the following way .
differentiation in @xmath295}$ ] is defined as : @xmath376 where , as mentioned in [ level0 ]
we have @xmath377 for some @xmath378 and , in the natural way , we set @xmath379 .
this makes @xmath380}$ ] , and the generating transmonomials of @xmath381 are those of @xmath289 together with @xmath2 .
we assume by induction that differentiation @xmath382}\mapsto \tilde{\mathcal{t}}^{[n-1]}$ ] has been defined for all transseries of level at most @xmath340 .
( in particular , @xmath383 is finitely generated . )
we define @xmath384}\right)={\mathcal{d}}\left(x^{\sigma}\exp(\largetr^{[n-1]})\right)= \sigma x^{\sigma-1}\exp(\largetr^{[n-1 ] } ) \\+x^{\sigma}{\mathcal{d}}\largetr^{[n-1]}\exp(\largetr^{[n-1]})\end{gathered}\ ] ] a level @xmath30 transseries is @xmath385 and we write in a natural way @xmath386 } = \sum_{\bk\ge \bk_0}c_{\bk } \sum_{m=1}^m k_m\smallmo_m^{k_m-1 } { \mathcal{d } } \smallmo_m \prod_{m\ne j=1}^{m}\smallmo_j^{k_j}\\= \sum_{m=1}^m\smallmo_m^{-1}\sum_{\bk\ge \bk_0}c_{\bk } k_m\smallmo_m^{k_m-1 } { \mathcal{d } } \smallmo_m \prod_{m\ne j=1}^{m}\smallmo_j^{k_j}\end{gathered}\ ] ] and the result follows from the induction hypothesis , since @xmath387}\ ] ] [ cfinigen ] if @xmath388 is the group ( finitely ) generated by all generators in any of the levels of @xmath289 , then @xmath383 is generated by the transmonomials of @xmath388 together possibly with @xmath2 .
if @xmath389 then @xmath390 .
immediate induction ; cf .
also the beginning of the proof of proposition [ pdifferentiation ] .
the properties of differentiation are the usual ones : [ pdiffprop ] @xmath391 , @xmath392 and @xmath393 ( for composition , see [ scompo ] ) .
the proof is straightforward induction . in the space of transseries ,
differentiation is also compatible with the order relation , a property which is not true in general , in function spaces .
[ porrder ] for any @xmath394 and @xmath395 we have @xmath396 @xmath397 @xmath398 the proof is by induction .
it is true for power series , which are the level zero transseries .
assume the property holds for transseries of level @xmath333 , and first prove the result for _ transmonomials _ of order @xmath30 , i.e. for the case when for @xmath399 @xmath400}=x^{\sigma_{i}}\exp(\largetr_{i})$ ] , where at least one of @xmath401 has level @xmath340 .
we have to evaluate @xmath402 and it is plain that we can assume without loss of generality that @xmath403 .
\(1 ) if @xmath404 then @xmath405 by proposition [ rform0 ] .
the remaining case is that for @xmath399 we have @xmath406 but @xmath401 are equal or else @xmath407 ( which obviously requires @xmath408 ) .
let @xmath409 .
we have @xmath410 and also @xmath411 for some @xmath412 , thus by proposition [ rform0 ] and the induction hypothesis we have @xmath413 @xmath414 thus @xmath415 \(2 ) we now let @xmath289 be arbitrary with the property @xmath416 and use proposition [ pstd ] to write @xmath417 where @xmath418 and thus , by step ( 1 ) we have @xmath419 the rest of the proof is immediate .
we have @xmath420 we have to show that if @xmath421 then @xmath422 . if @xmath423 then ( for instance ) @xmath424 and then @xmath425 . if instead @xmath426 then @xmath427 and we see that @xmath428 which , by the above , implies @xmath429 which gives @xmath430 , a contradiction .
[ p dom ] assume @xmath315 or @xmath312 .
then : \(i ) if @xmath431 then @xmath432 .
\(ii ) @xmath433 .
straightforward induction .
complex transseries @xmath434 are constructed in a similar way as real transseries , replacing everywhere @xmath435 by @xmath436 .
thus there is only one order relation in @xmath434 , @xmath437 .
difficulties arise when exponentiating transseries whose dominant term is imaginary .
operations with complex transseries are then limited .
we will only use complex transseries in contexts that will prevent these difficulties .
the theory of differential equations in @xmath438 is similar to the corresponding theory for functions . _ example_. the general solution of the differential equation @xmath439 in @xmath438 ( for @xmath440 ) is @xmath441 .
indeed , the fact that @xmath442 is a solution follows immediately from the definition of the operations in @xmath438 . to show uniqueness , assume @xmath443 satisfies ( [ eqei1 ] ) .
then @xmath444 is a solution of @xmath445
. then @xmath446 satisfies @xmath447 i.e. , @xmath448 the particular solution @xmath449 is the unique solution of the equation @xmath450 which is manifestly contractive in the space of level zero transseries ( cf .
[ subcontr ] ) .
however this same equation is not contractive for transseries of positive level , ( because e.g. @xmath451 ) ; this could also have been anticipated noting that the solution is not unique . the right composition @xmath452 is defined on @xmath453 , if @xmath454 and @xmath455 .
the definition is inductive .
we first define the power and the exponential of a transseries .
assume powers and exponentials have been defined for all transseries of level @xmath333 .
let @xmath456}$ ] be any transseries such that @xmath457 , cf .
proposition [ pstd ] . by the definition of @xmath168 and ( [ defanytrn ] ) ,
@xmath458 is a transmonomial , @xmath459}\exp(\largetr^{[n-1]})$ ] .
we let @xmath460}\right)^{\sigma } \exp(\sigma\largetr^{[n-1 ] } ) ( 1+\smalltr)^{\sigma}\\= c^{\sigma}\left(\largemo^{[n-1]}\right)^{\sigma } \exp(\sigma\largetr^{[n-1 ] } ) ( 1+\smalltr)^{\sigma}\\= c^{\sigma}\left(\largemo^{[n-1]}\right)^{\sigma } \exp(\sigma\largetr^{[n-1 ] } ) \sum_{k=0}^{\infty}\binom{n}{\sigma}\smalltr^{n}\end{gathered}\ ] ] where @xmath461 are the generalized binomial coefficients , the infinite sum is well defined , by proposition [ infisum ] , and thus @xmath462 is well defined as well . then , if @xmath463 and if @xmath304}=\sum_{\bk\ge\bk_0}c_\bk x^{-\boldsymbol\sigma\cdot\bk}$ ] is a level zero transseries , we write @xmath464}\circ \anytr= \sum_{\bk\ge\bk_0}c_\bk ( \anytr^{-1})^{\boldsymbol\sigma\cdot\bk}\ ] ] which is well defined by proposition [ infisum ] ( ii ) and proposition [ pstd ] .
we note that , under our assumptions for @xmath289 , @xmath304}\circ\anytr>0 $ ] is positive iff @xmath304}>0 $ ] similarly , we write cf . definition [ smalllarge ] @xmath465}\anytr^{[n]}\end{gathered}\ ] ] well defined by the definition of a transmonomial , proposition [ infisum ] ( ii ) and proposition [ pstd ] .
now the definition of general composition is straightforward induction .
we assume that composition is defined at all @xmath333 levels , and that in addition @xmath466}\circ\anytr>0 $ ] if @xmath466}>0 $ ]
. then @xmath467}\circ\anytr=\sum_{\bk\ge\bk_0}c_{\bk}(\smallmo^{[n ] } \circ\anytr)^{\bk } \\=\sum_{\bk\ge\bk_0}c_{\bk}(\anymo^{[n-1]}\circ\anytr)^{\bk } \exp(-\largetr^{[n-1]}\circ\anytr)\\ = \sum_{\bk\ge\bk_0}c_{\bk}(\anymo^{[n-1]}\circ\anytr)^{\bk } \left[\exp(-l(\largetr^{[n-1]}\circ\anytr)c\exp(-s(\largetr^{[n-1]}\circ\anytr)\right]\\ = \sum_{\bk\ge\bk_0}c'_{\bk}(\anymo^{[n-1]}\circ\anytr)^{\bk } \smallmo^{[n]}\exp(-s(\largetr^{[n-1]}\circ\anytr)\end{gathered}\ ] ] and the last sum exists by the induction hypothesis and proposition [ genclos ] .
we define @xmath468 with the convention @xmath469 .
we write @xmath470 and then any log - free transseries can be written as @xmath471 .
this defines right composition with @xmath472 in this trivial case , as @xmath473 . more generally , we define @xmath438 , the space of general transseries , as a set of formal compositions @xmath474 with the algebraic operations ( symbolized below by @xmath475 ) inherited from @xmath453 by @xmath476\circ l_{n+k}\end{aligned}\ ] ] and using ( [ defop ] ) , differentiation is defined by @xmath477({\mathcal{d}}\anytr)\circ l_n\ ] ] [ pbigfield ] @xmath438 is an ordered differential field , closed under restricted composition .
the proof is straightforward , by substitution from the results in [ slogfree ] .
we will denote generically the elements of @xmath438 with the same symbols that we used for @xmath453 .
[ pintegr ] @xmath438 is closed under integration .
the idea behind the construction of @xmath478 is the following : we first find an invertible operator @xmath108 which is _ to leading order _ @xmath479 ; then the equation for the correction will be contractive .
let @xmath480 . to unify the treatment , it is convenient to use the identity @xmath481 where the last integrand
, @xmath482 is a log - free transseries and moreover @xmath483 the case @xmath484 is trivial and it thus suffices to find @xmath485 , where @xmath486 .
then @xmath487 for any @xmath192 and thus also @xmath488 for all @xmath192 .
therefore , since @xmath489 we expect that @xmath490 and we look for a @xmath491 such that @xmath492 then @xmath491 satisfies the equation @xmath493 by propositions [ rform0 ] , corollary [ cfinigen ] and proposition [ p dom ] , ( [ eq : intcontrac ] ) is contractive in @xmath374}$ ] .
the proof now follows from proposition [ genclos ] . in the following we also use the notation @xmath494 and we write @xmath495 for the antiderivative
@xmath479 constructed above .
[ noconst ] @xmath495 is an antiderivative without constant terms , i.e ,
@xmath496 this follows from proposition [ rform0 ] , together with the fact that @xmath497 and @xmath498 in ( [ defdelta ] ) belong to @xmath374}$ ] .
[ ppropint ] we have @xmath499 where @xmath500 all the properties are straightforward ; preservation of inequalities uses proposition [ porrder ] .
[ rexpcontrac ] let @xmath501 .
the operators defined by @xmath502 are contractive on @xmath438 . for ( [ ej1 ] )
it is enough to show contractivity of @xmath503 .
this is a straightforward calculation similar to the proof of proposition [ pintegr ] .
we have for some @xmath30 @xmath504 where @xmath505
. @xmath506\circ l_{n+2}\\ \ll \prod_{1\le j\le n+2}e_j\exp(\largetr\circ e_2)\end{gathered}\ ] ] the proof of ( ii ) is similar .
the general contractivity principle stated in theorem [ pfp ] , which we have used in proving closure of transseries with respect to a number of operations can be used to show closure under more general equations .
our main focus is on differential systems .
we start with an example , a first order equation : @xmath507 where @xmath508 where the sums are assumed to converge and @xmath509 .
we see that @xmath510 is well defined if @xmath511 ( cf .
proposition [ genclos ] ) , and it is under this assumption that we study @xmath510 . ) , @xmath510 is not in @xmath438 if @xmath512 ( since , in this case , @xmath513 is unbounded ) . ] ( 1 )
. solutions of ( [ edefterms ] ) in @xmath295}$ ] .
the equation @xmath514 is contractive in @xmath295}$ ] ( this follows immediately from [ subcontr ] ) .
thus there exists in @xmath295}$ ] a unique solution @xmath515 . since ( [ elev0 ] ) is also contractive in the subspace of @xmath295}$ ] of series of the form @xmath516 we have @xmath517 * note . *
the iteration @xmath518 , @xmath519 is convergent in @xmath438 and , if @xmath520}x^{-k}$ ] then @xmath521}=c_k$ ] for @xmath522 , and this is a very convenient way to calculate the coefficients @xmath523 .
\(2 ) let now @xmath524
. then @xmath525 with @xmath526 or @xmath527 since by assumption @xmath528 we have @xmath529 and thus @xmath530 for any @xmath531 so that @xmath532 whence , by composition with @xmath533 we get @xmath534 equation ( [ eeqdelta1 ] ) implies @xmath535 and ( [ eequint ] ) is contractive by remark [ rlinearity ] and remark [ rexpcontrac ] .
in particular , for every @xmath61 there is a unique @xmath536 satisfying ( [ eequint ] ) .
[ rspecialform ] the general transseries solution of ( [ eqord1 ] ) is @xmath537 where @xmath538 with @xmath539}$ ] and @xmath540 this is a straightforward consequence of discussion of this section and of ( [ eequint ] ) .
let @xmath541 .
we have @xmath542 a direct calculation shows that @xmath543 . the transformation @xmath544 formally linearizes ( [ eqord1 ] ) .
consider the differential system @xmath545 we look at solutions @xmath546 such that @xmath547 as @xmath293 along some direction @xmath548 .
the following conditions are assumed ( a1 ) the function @xmath549 is analytic at @xmath550 .
( a2 ) nonresonance : the eigenvalues @xmath551 of the linearization @xmath552 are linearly independent over @xmath553 ( in particular nonzero ) and such that @xmath554 are different from each other ( i.e. , the stokes lines are distinct ; we will require somewhat less restrictive conditions , see
3.1 ) . by relatively straightforward algebra , ( @xcite and also 3.2 where all this is exemplified in a two - dimensional case ) , the system ( [ eqor1 ] ) can then be brought to the form @xmath555 where @xmath556 are constant matrices , @xmath557 [ invert ] ( i ) if @xmath558 . in this case
the system ( [ eqor ] ) is linear and has the general transseries solution @xmath559 \(ii ) more generally , if @xmath560 is a transseries , then the general solution of ( [ eqor ] ) is @xmath561 in both cases the system is diagonal and the result follows immediately from the case when @xmath562 , i.e. from proposition [ rspecialform ] .
the general solution of ( [ eqor ] ) in @xmath434 ( cf . [ complextr ] ) is an @xmath563 parameter transseries , as shown in the sequel .
[ gts ] let @xmath564 be a ray in @xmath565 .
the general solution of ( [ eqor ] ) in @xmath566 with the restriction @xmath567 is of the form @xmath568 where @xmath569 for all @xmath570 so that @xmath571 as @xmath293 in @xmath564 .
if @xmath546 is a solution of ( [ eqor ] ) then we have , by remark [ invert ] @xmath572 for some @xmath573 .
since @xmath567 we have @xmath574 and thus @xmath575 again since @xmath567 , we then have @xmath576 for all @xmath570 for which @xmath577 . .
with the condition @xmath567 , eq . ( [ eq : invert ] )
has a unique solution .
indeed , the difference of two solutions @xmath578 satisfies the equation @xmath579\big)\ ] ] since @xmath580 we have @xmath581 which by proposition [ ppropint ] implies @xmath582 , i.e. , @xmath583 . using remark [ rexpcontrac ]
it is easy to check that ( [ eq : invert2 ] ) is an asymptotically contractive equation , in the space of @xmath546 which are @xmath584 thus it has a solution @xmath585}$ ] with this property .
since the previous note shows the solution of ( [ eqor ] ) with @xmath567 is unique , we have @xmath586}$ ] .
formula ( [ transsf ] ) is obtained by straightforward iteration of ( [ eq : invert2 ] ) .
in this section we discuss a borel summation induced isomorphism between transseries and functions in the setting [ sysnonlin ] .
the goal is to show , on simple examples , how borel summation is proved and used . for more general results we refer to @xcite . 1 .
the truncated borel transform @xmath589 of @xmath590 has a nonzero radius of convergence , 2 .
@xmath591 can be analytically continued along @xmath592 and 3 .
the analytic continuation @xmath591 grows at most exponentially along @xmath592 and is therefore laplace transformable along @xmath592 .
_ example _ the formal solution for large @xmath86 of the equation @xmath597 is borel summable .
indeed @xmath598 and it can be checked that the borel sum is a solution of the given equation ; see also ( i ) in the note below .
\(i ) it can be shown that borel summation is an extended isomorphism : in particular it commutes with algebraic operations , including multiplication and complex conjugation , and with differentiation @xcite .
this can be expected from the fact that , formally , borel summation is the composition of laplace transform with its inverse , and the identity obviously commutes with the operations mentioned above .
\(ii ) borel summation has to be generalized in a number of ways in order to be used for solving more general equations .
indeed , an equation as simple as @xmath599 has a formal solution which is not borel summable : @xmath600 is not laplace transformable . if the laplace transform integral is taken along a ray above or below @xmath601 , then @xmath602 is a solution of the given equation .
none of these path choices however yields a real valued function , whereas the formal series has real coefficients .
the resulting summation operator would not be an isomorphism since commutation with complex conjugation fails . in this simple example
the half sum of the upper and lower integrals is real valued and solves the equation , but this ad hoc procedure would not work for nonlinear equations .
more complicated averages need to be introduced , see [ ave ] . where @xmath610 are integer power series .
it is important to note that in ( [ eq : gtrns ] ) only the constant @xmath61 depends on the solution .
we will see that the @xmath610 are simultaneously borel summable , to @xmath611 , and that the sum which is contractive in the sup norm on @xmath620 , for @xmath621 large enough , in a ball of radius @xmath622 small enough .
the solution of ( [ eq : intequ ] ) is thus unique and it is easy to see that @xmath623 for large @xmath86 ( @xmath624 is not in general equal to @xmath61 ) .
whence @xmath626}\sim const .
\gamma(k-\beta)x^{-k}$ ] and thus borel summation appears natural .
since @xmath515 is defined through a differential equation it is natural to borel transform the equation itself .
the result is and clearly , if @xmath641 and @xmath642 are large enough , the image @xmath643 of the ball @xmath644 is contained in @xmath645 .
similarly , it can be seen that @xmath646 is contractive in @xmath645 .
thus the following conclusion .
this however does not yet imply borel summability of the series @xmath515 ; to show this we must prove appropriate analyticity properties for @xmath631 and this can be done in essentially the same manner .
vanishing at @xmath649 , continuous in @xmath650 , and such that @xmath635 with the norm @xmath651 is a banach algebra with respect to convolution .
in addition , if @xmath652 for any @xmath653 and @xmath654 is analytic in @xmath49 then we have for large enough @xmath641 and for any @xmath54 there is a unique analytic solution in @xmath659 , @xmath656 , and @xmath656 is thus independent of @xmath54 .
since we have @xmath660 for @xmath661 it follows that @xmath662 .
using proposition [ l1loc ] the proof is complete . where in the sum @xmath665 the indices satisfy @xmath666 while in @xmath667 the condition is @xmath668 . we have already seen that @xmath515 is summable .
the equation for @xmath669 is special , and we treat it separately . to ensure borel transformability , since @xmath670 , it is convenient to take @xmath671 , which gives it is easy to see that ( [ eq : eqy1 ] ) is contractive in a space of analytic functions for @xmath681 if @xmath682 is small . to find exponential bounds for large @xmath683 we first restrict to a ray @xmath684 with @xmath685 .
note now that the operator on the r.h.s . of ( [ eqeqyk ] ) is contractive in the norms introduced , for large @xmath641 .
it is then clear that @xmath706 are analytic and laplace transformable for large enough @xmath641 .
it remains to show they are simultaneously laplace transformable .
note that for large @xmath641 , denoting @xmath707 we have , with @xmath708 arbitrarily small if @xmath641 is large , @xmath709and the coefficients @xmath710 are majorized by the taylor coefficients of the analytic solution of the algebraic equation @xmath711 thus , for large enough @xmath641 we have that @xmath712 ( @xmath713 for large @xmath641 ) and thus , for @xmath86 large enough @xmath714 ( an inductive calculation shows that @xmath715 for small @xmath683 . ) the sum @xmath716 is then uniformly convergent .
it is easy to see that it gives therefore a solution of ( [ eq : eq * ] ) .
: there is an asymmetry at this point : the opposite inequality can not be achieved , in general , as simply and without violating analyticity at infinity . in some instances
this transformation is not convenient since it makes more difficult the study of certain properties , see @xcite .
we construct over @xmath735 a surface @xmath736 , consisting of homotopy classes of smooth curves in @xmath735 starting at the origin , moving away from it , and crossing at most one stokes line , at most once ( fig .
1 ) : the borel transforms of @xmath743 are analytic at zero and the only possible singularities are at multiples of the eigenvalues @xmath744 of the linearized system .
these singularities are `` regular '' in the sense that there exist _ convergent _ local expansions at these singularities in terms of powers and possibly logarithms .
the following theorem makes this statement precise .
@xmath770^{(lm_j)}+l\mathbf{b}_{\bfk lj}(z)=\cr \big[z^{\bfk\cdot{\bfbet}'+l\beta_j'-1}(\ln z)^{0,1}\,\mathbf{a}_{\bfk l j}(z)\big]^{(lm_j)}+l\mathbf{b}_{\bfk l j}(z)\ ( l=0,1,2,\ldots ) \end{gathered}\ ] ] it is possible to take laplace transforms of @xmath778 along a ray avoiding @xmath601 from above or below .
however it can be checked that , assuming @xmath778 is singular at @xmath779 and that the series @xmath780 has real coefficients , neither of them would yield a real valued function .
the sum above coincides with the one in which @xmath784 is exchanged with @xmath785 , accounting for the reality - preserving property .
clearly , if @xmath755 is analytic along @xmath782 , then the terms in the infinite sum vanish and @xmath786 ; we also let @xmath786 if @xmath787 , where again @xmath788 is analytic .
it follows from ( [ defmed ] ) and theorem [ ceq ] below that the laplace integral of @xmath789 along @xmath601 can deformed into contours as those depicted in fig .
1 , with weight @xmath790 for a contour turning around @xmath791 . more generally , we consider the averages with @xmath794 , this represents the most general family of averages of borel summation formulas which commute with complex conjugation , with the algebraic and analytic operations and have good continuity properties @xcite .
the value @xmath795 is special in that it is the only one compatible with optimal truncation .
for rank one odes the balanced average mentioned above can be shown to coincide with the `` median '' summation of ecalle , who has introduced and studied a wide variety of averages @xcite , suitable in more complicated settings .
[ ceq ] ( i ) the branches of @xmath796 in @xmath738 have limits in a @xmath797-algebra of distributions , @xmath798 .
their laplace transforms in @xmath799 @xmath800 exist simultaneously and with @xmath801 and for any @xmath802 there is a constant @xmath803 and an @xmath804 large enough , so that for @xmath805 we have @xmath806 . if @xmath812 and @xmath813 then @xmath814 is continuous in @xmath815\backslash\{\phi_i : i\le n\}$ ] in the @xmath816 topology and has ( at most ) jump discontinuities for @xmath817
. for each @xmath818 , @xmath819 and any @xmath803
there is an @xmath820 and an @xmath192 such that @xmath821 are continuous in @xmath815\backslash\{\phi_i ;
-\bfk'\cdot\bflam+\lambda_i : i\le n , \bfk'\le\bfk\}$ ] in the @xmath822 topology and have ( at most ) jump discontinuities on the boundary .
\(ii ) the sum ( [ defmed ] ) converges in @xmath809 ( and coincides with the analytic continuation of @xmath755 when @xmath755 is analytic along @xmath601 ) . for any @xmath823
there is a large enough @xmath804 _ independent of @xmath818 _ so that @xmath824 with @xmath825 are laplace transformable for @xmath826 and furthermore @xmath827 .
in addition , if @xmath828 , then for large @xmath641 , @xmath829 . converges uniformly for @xmath834 , and @xmath546 is a solution of ( [ eqor ] ) . when the direction of @xmath683 is not the real axis then , by definition , @xmath835 , @xmath596 is the usual laplace transform and ( [ soleqn ] ) becomes for some @xmath841 ( depending on the solution and chosen form ) . with the convention binding the directions of @xmath86 and @xmath683 and the representation form being fixed
the representation of a solution is unique .
consider the expression of a fixed solution @xmath546 of ( [ eqor ] ) as a borel summed transseries ( [ soleqn ] ) .
as @xmath845 varies , ( [ soleqn ] ) changes only through @xmath841 , and that change occurs when the stokes lines are crossed .
we have , in the neighborhood of @xmath601 , with @xmath846 defined in ( [ defsj ] ) : [ r1 ] in view of ( [ mainresur ] ) the different analytic continuations of @xmath747 along paths crossing @xmath601 at most once can be expressed in terms of @xmath848 .
the most general formal solution of ( [ eqor ] ) that can be formed in terms of @xmath849 with @xmath850 is ( [ eqformgen , n ] ) with @xmath851 arbitrary and @xmath852 for @xmath853 . any true solution of ( [ eqor ] ) based on such a transseries
is given in ( [ soleqnp ] ) with @xmath841 as above .
any average @xmath854 along paths going forward in @xmath601 such that @xmath855 is thus of the form ( [ defmedc ] ) . [ stokestr ]
assume only @xmath721 lies in the right half plane .
let @xmath856 be two paths in the right half plane , near @xmath857 such that @xmath858 as @xmath293 along @xmath856 .
consider the solution @xmath546 of ( [ eqor ] ) given in ( [ soleqn ] ) with @xmath859 and where the path of integration is @xmath860 .
then [ asymptrick ] i ) let @xmath862 and @xmath863 be solutions of ( [ eqor ] ) so that @xmath864 for large @xmath86 in an open sector @xmath49 ( or in some direction @xmath564 ) ; then @xmath865 for some constants @xmath866 , where the indices run over the eigenvalues @xmath867 with the property @xmath868 in @xmath49 ( or @xmath564 ) .
if @xmath869 for all @xmath80 , then @xmath870 .
\ii ) let @xmath862 and @xmath863 be solutions of ( [ eqor1 ] ) and assume that @xmath871 has differentiable asymptotics of the form @xmath872 with @xmath873 and @xmath874 , for large @xmath86 .
then @xmath875 for some @xmath570 .
given @xmath546 , the value of @xmath884 can change only when @xmath885 , @xmath886 , i.e. when crossing one of the ( finitely many by ( c1 ) ) stokes lines .
the procedure is similar to the medianization proposed by calle , but ( due to the structure of ( [ eqor ] ) ) requires substantially fewer analytic continuation paths .
resurgence relations are found and in addition we provide a complete description , needed in applications , of the singularity structure of the borel transforms of @xmath887 .
many classical equations which are not presented in the form studied in
[ sysnonlin ] can be brought to that form by elementary transformations .
we look at a few examples , to illustrate this and the normalization process
. for large @xmath86 is not of the form ( [ eqor1 ] ) .
it can nevertheless be brought to that form by simple changes of variables .
the normalizing change of variables of a given is most conveniently obtained in the following way . at level zero , differentiation is contractive and thus , within power series there is a unique solution of ( [ erf ] ) , obtained as a fizxed point of @xmath889 denoting this series by @xmath201 we look for further transseries solutions in the form @xmath890 .
we get where we look for higher level transseries solutions ( we choose to ignore the fact that ( [ delta1 ] ) is solvable explicitly ) .
we therefore take @xmath892 and look for level zero solutions .
we have @xmath893 with the one parameter family of solutions @xmath894 .
therefore @xmath895 the exponent is supposed to be linear in @xmath86 which suggests the change of variable @xmath896 . in terms of @xmath750 we get @xmath897 which is not of the required form because of the noninteger power of @xmath750 .
this can be easily take care of by the substitution @xmath898 which leads to @xmath899 which is a contractive equation in the space @xmath295}$ ] , cf . [ deg1 ] .
it follows in particular that @xmath903 and thus @xmath904 .
since for ( [ eqor1 ] ) the exponentials have linear exponents , the natural variable of ( [ eq : schr ] ) , for our purpose , is @xmath905 . in
this variable , ( [ eq : schr ] ) becomes the matrix @xmath909 can ( always ) be diagonalized , by taking @xmath910 for a suitable matrix @xmath911 , and the diagonalized matrix is @xmath912 .
indeed , the equation for @xmath911 is @xmath913 , which can be solved whenever the eigenvalues of @xmath914 are distinct
. the normal form of ( [ eq : schrsyst ] ) is thus .
once we determined @xmath921 we look for possible small corrections to @xmath921 ; since these are usually exponentially small , formal wkb is again useful .
we substitute @xmath922 .
the equation for @xmath654 since the exponentials in a transseries solution of a normalized system have _ linear _ exponent , with negative real part , the independent variable should be @xmath926 and @xmath927 .
then @xmath928 @xmath929 , which compared to ( [ transsf ] ) suggests the change of dependent variable @xmath930 .
choosing for convenience @xmath931 yields the next step is to achieve leading behavior @xmath933 .
this is easily done by subtracting out the leading behavior of @xmath934 ( which can be found by maximal balance , as above ) .
with @xmath935 we get the normal form the first normalization applies for the manifold of solutions such that @xmath952 ( for @xmath953 @xmath593 is exponentially small and behaves like an airy function ) while the second one corresponds to @xmath954 .
transseries and borel summation techniques can be successfully used for difference equations .
consider difference systems of equations which can be brought to the form @xmath955 where @xmath956 and @xmath957 are constant coefficient matrices , @xmath958 is convergently given for small @xmath959 by @xmath960 with @xmath961 analytic in @xmath30 at infinity and @xmath962 under nonresonance conditions : let @xmath963 and @xmath964 where @xmath965 are the eigenvalues of @xmath914 and the @xmath966 are the eigenvalues of @xmath967 .
then the nonresonance condition is @xmath968 we consider the solutions of ( [ eq : bra1 ] ) which are small as @xmath30 becomes large .
braaksma and kuik @xcite @xcite showed that the recurrences ( [ eq : bra1 ] ) posess @xmath969-parameter transseries solutions of the form @xmath970 with @xmath971 where @xmath972 are formal power series in powers of @xmath973 and @xmath974 is chosen such that , after reordering the indices , we have @xmath975 for @xmath976 . it is shown in @xcite , @xcite that these transseries are generalized borel summable in any direction and borel summable in all except @xmath192 of them and that @xmath977 is a solution of ( [ eq : bra1 ] ) , if @xmath978 , @xmath979 large enough . for partial differential equations with analytic coefficients which can be transformed to equations in which the differentiation order in a distinguished variable , say time ,
is no less than the one with respect to the other variable(s ) , under some other natural assumptions , cauchy - kowalevski theory ( c - k ) applies and gives existence and uniqueness of the initial value problem .
a number of evolution equations do not satisfy these assumptions and even if formal power series solutions exist their radius of convergence is zero .
the paper @xcite provides a c - k type theory in such cases , providing existence , uniqueness and regularity of the solutions .
roughly , convergence is replaced by borel summability , although the theory is more general . unlike in c - k , solutions of nonlinear evolution equations develop singularities which can be more readily studied from the local behavior near @xmath980 , and this is useful in determining and proving spontaneous blow - up . in the following ,
@xmath981 , @xmath982 , @xmath983 is in a poly - sector @xmath984 in @xmath985 where @xmath986 , @xmath987 is a function analytic in @xmath988 near @xmath989 vanishing as @xmath990 .
the results in @xcite hold for @xmath30-th order nonlinear _ quasilinear _ partial differential equations of the form @xmath991 where @xmath992 , for large @xmath993 in @xmath49 .
generically , the constant coefficient operator @xmath994 in the linearization of @xmath995 is diagonalizable .
it is then taken to be diagonal , with eigenvalues @xmath996 .
@xmath495 is subject to the requirement that for all @xmath997 and @xmath998 in @xmath999 with @xmath1000 we have @xmath1001}_j ( -{\bf p } ) > 0\ ] ] where @xmath1002}(\partial_{\bf x})$ ] is the principal symbol of @xmath1003
. then the following holds .
( the precise conditions and results are given in @xcite . )
determining asymptotic properties of solutions of pdes is substantially more difficult than the corresponding question for odes .
borel - laplace techniques however provide a very efficient way to overcome this difficulty .
the paper shows that formal series solutions are actually borel summable , a fortiori asymptotic , to actual solutions .
[ trb ] if condition [ cond 2 ] and the assumptions of theorem [ t1 ] are satisfied , then the unique solution @xmath1014 found there can be written as @xmath1015 where @xmath1016 is ( a ) analytic at zero in @xmath1017 ; ( b ) analytic in @xmath1018 in the poly - sector @xmath1019 , @xmath1020 ; and ( c ) exponentially bounded in the latter poly - sector .
eariler , borel summability has been shown for heat equation by lutz , miyake and schfke @xcite and generalized to linear pdes with constant coefficients by balser @xcite and in special classes of higher order nonlinear pdes in @xcite .
the simple normalizing procedure described above does not always work ; although this situation is in some sense nongeneric it plays an important role in some applications . in some instances
mixed type transseries ; for instance , by taking @xmath1021 which is a solution of a rank one differential equation as the rhs of ( [ erf ] ) we get which can be brought to a second order meromorphic equation by using the equation of @xmath1021 , i.e. by adding ( [ erfmix ] ) to its derivative , @xmath1023 the power series solution can be obtained contractively as in the previous examples .
for the complete transseries solution , the substitution @xmath1024 now yields @xmath1025 which has two solutions of level zero , @xmath1026 and thus @xmath1027 with @xmath1028 power series .
clearly , no change of independent variable brings ( [ gentr ] ) to the form ( [ transsf ] ) . on the other hand ,
neither of variable @xmath86 nor @xmath1029 ( equivalently , @xmath1030 ) can be used for borel summation : \(i ) borel transform in the variable @xmath1030 ( or @xmath1029 ) yields a function that has still a _ divergent _ power series at the origin .
indeed the the coefficients @xmath1031 of the power series @xmath201 , which satisfy the recurrence relation @xmath1032\ ] ] ( with @xmath1033 ) grow like @xmath1034 , while a borel transform in @xmath1030 only decreases the growth of the coefficients by a factor of , roughly , @xmath1035 . at a closer look it can be shown seen that the factorial growth of the coefficients and the presence of the term @xmath5 are interrelated .
\(ii ) it can be checked that in the variable @xmath86 , the borel transform @xmath631 of the power series solution @xmath29 of ( [ erfmix2 ] ) is not laplace transformable because of faster than exponential growth .
indeed , @xmath631 satisfies the equation @xmath1036 the superexponential growth , in turn , can be related to the presence of @xmath1037 in the transseries .
this suggests decomposing the formal power series solution into two parts , one summable in the variable @xmath86 , the other one in the variable @xmath1030 .
we can adjust the parameters @xmath1038 and @xmath1039 so that the formal series solution of @xmath1043 is thus decomposed in a sum of separately borel summable series in the variables @xmath1030 and @xmath86 respectively . that this ( rather ad - hoc ) procedure works is a reflection of general theorems in multisummability .
a powerful and general technique , that of _ acceleration and multisummability _ introduced by calle , see @xcite and @xcite , adequately deals with mixed divergences in very wide generality .
the procedure is universal and works in the same way for very general systems . in the case of solutions of meromorphic differential equations ,
mixed types of diveregence relate to the presence of exponential terms with exponents of different powers .
then multisummation consists in borel transform with respect to the lowest power of @xmath86 in the exponents of the transseries , usual summation @xmath1044 , a sequence of transformations called accelerations ( which mirror in borel space the passage from one power in the exponent to the immediately larger one ) followed by a final laplace transform in the largest power of @xmath86 .
more precisely ( @xcite ) : @xmath1045 where @xmath1046 , @xmath1047 is the formal inverse of @xmath1048 , @xmath1049 and the acceleration operator @xmath1050 is formal the image , in borel space , of the change of variable from @xmath1051 to @xmath86 and is defined as @xmath1052 and where , for @xmath1053 , the kernel @xmath1054 is defined as @xmath1055 where we adapted the notations in @xcite to the fact that the formal variable is large . in our example , @xmath1056 .
\(i ) multisummability of type ( [ multisum ] ) can be equivalently characterized by decomposition of the series into terms which are ordinarily summable after changes of independent variable of the form @xmath1057 .
this is shown in @xcite where it is used to give an alternative proof of multisummability of series solutions of meromorphic odes , closer to the cohomological point of view of ramis and sibuya .
\(iii ) we expect that the finite generation property of transseries allows for the cohomological approach ( i ) combined with ( ii ) to prove multisummability of transseries solutions of more general contractive equations , in the sense of theorem [ pfp ]
. this would give a rigorous proof of closure of analyzable functions under all natural operations in the sense of ecalle .
berry , m. v. ; howls , c. j. 1990 hyperasymptotics .
a _ * 430 * , 653668 .
berry , m. v. ; keating , j. p. 1992 a new asymptotic representation for @xmath1058 and quantum spectral determinants .
london ser .
a * 437 * ( 1992 ) , no .
1899 , 151173 .
_ j h conway , _ on numbers and games _ , academic press ( 1976 ) .
o. costin , s. tanveer , existence and uniqueness for a class of nonlinear higher - order partial differential equations in the complex plane , comm .
pure appl .
math , vol .
liii , 10921117 ( 2000 ) .
o. costin , s. tanveer _ on the existence and uniqueness of solutions of nonlinear evolution systems of pdes in @xmath1059 , their asymptotic and borel summability properties _ preprint ( 2003 ) .
les fonctions resurgentes , vol .
i , ii and iii _
orsay , 1985 .
j calle , _ in bifurcations and periodic orbits of vector fields , nato asi series , vol .
408 , 1993 _ h flashka and a c newell _ monodromy and spectrum preserving transformations _ commun .
phys . * 76 * pp . 65116 ( 1980 ) .
j ecalle , f menous , _ well - behaved convolution averages and the non - accumulation theorem for limit - cycles . _
the stokes phenomenon and hilbert s 16th problem , 71101 , world sci .
publishing ( 1996 ) l fuchs _ sur quelques quations diffrentielles linaires du second ordre _ c. r. acad .
sci . , paris * 141 * pp .
555 - 558 ( 1905 ) .
r kuik , transseries in differential and difference equations , phd thesis , university of groningen , isbn 90 - 367 - 1771-x ( 2003 ) .
d. a. lutz , m. miyake and r. schfke on the borel summability of divergent solutions of the heat equation , nagoya math .
j. * 154 * , 1 , ( 1999 ) . | transseries in the sense of calle are constructed using a topological approach . a general contractive mapping principle is formulated and proved , showing the closure of transseries under a wide class of operations . in the second part
we give an overview of results and methods reconstruction of actual functions and solutions of equations from transseries by generalized borel summation with in ordinary and partial differential and difference equations . |
analysis of the sunyaev - zeldovich effect ( sze ) and x - ray data provides a method of directly determining distances to galaxy clusters at any redshift .
clusters of galaxies contain hot ( @xmath1 kev ) gas , known as the intracluster medium ( icm ) , trapped in their potential wells .
cosmic microwave background ( cmb ) photons passing through a massive cluster interact with the energetic icm electrons with a probability of @xmath2 .
this inverse - compton scattering preferentially boosts the energy of a scattered cmb photon , causing a small ( @xmath3 mk ) distortion in the cmb spectrum , known as the sunyaev - zeldovich effect @xcite .
the sze is proportional to the pressure integrated along the line of sight , @xmath4 .
x - ray emission from the icm has a different dependence on the density @xmath5 , where is the x - ray cooling function .
taking advantage of the different density dependences and with some assumptions about the geometry of the cluster , the distance to the cluster may be determined .
sze and x - ray determined distances are independent of the extragalactic distance ladder and provide distances to high - redshift galaxy clusters .
this method does not rely on clusters being standard candles or rulers and relies only on relatively simple properties of highly ionized plasma .
the promise of direct distances has been one of the primary motivations for sze observations .
efforts over the first two decades after the sze was first proposed in 1970 @xcite yielded few reliable detections . over the last decade ,
new detectors and observing techniques have allowed high - quality detections and images of the effect for more than 60 clusters with redshifts as high as one .
sze observations are routine enough to build up samples of clusters and place constraints on cosmological parameters .
the sze offers a unique and powerful observational tool for cosmology .
distances to galaxy clusters yield a measurement of the hubble constant , @xmath6 . with a sample of high - redshift clusters , sze and x - ray distances
can be used to determine the geometry of the universe .
in addition , the sze has been used to measure gas fractions in galaxy clusters ( e.g. , * ? ? ? * ; * ? ? ? * ) , which can be used to measure the matter density of the universe , @xmath7 , assuming the composition of clusters represents a fair sample of the universal composition .
upcoming deep , large - scale sze surveys will measure the evolution of the number density of galaxy clusters , which is critically dependent on the underlying cosmology . in principle , the equation of state of the `` dark energy '' may be determined from the evolution of the number density of clusters . in this review ,
we first outline the properties of the sze in [ sec : sze ] and provide a brief overview of the current state of the observations in [ sec : obs_status ] .
sze / x - ray determined distances are discussed in [ sec : cosmic_da ] , briefly discussing the current state of sze / x - ray distances , sources of systematics , and future potential .
sze surveys are briefly discussed in [ sec : surveys ] and a summary is given in
[ sec : summary ] .
the physics of the sze is covered in previous reviews @xcite , with @xcite and @xcite providing reviews of the observations . @xcite
provide a review with focus on cosmology from the sze , with special attention to sze surveys .
the sze is a small spectral distortion of the cmb spectrum caused by the scattering of the cmb photons off a distribution of high - energy electrons .
we only consider the sze caused by the hot thermal distribution of electrons provided by the icm of galaxy clusters .
cmb photons passing through the center of a massive cluster have a @xmath8 probability of interacting with an energetic icm electron .
the resulting inverse - compton scattering preferentially boosts the energy of the cmb photon , causing a small ( @xmath3 mk ) distortion in the cmb spectrum . to illustrate the small effect ,
figure [ fig : sze_cmb ] shows the sze spectral distortion for a fictional cluster that is over 1000 times more massive than a typical cluster .
the sze appears as a decrease in the intensity of the cmb at frequencies below @xmath9 ghz and as an increase at higher frequencies .
the sze spectral distortion of the cmb , expressed as a temperature change @xmath10 at dimensionless frequency @xmath11 , is given by @xmath12 where @xmath13 is the compton @xmath13-parameter , which for an isothermal cluster equals the optical depth times the fractional energy gain per scattering . here , @xmath14 is the thomson cross - section , @xmath15 is the electron number density , @xmath16 is the electron temperature , @xmath17 is the boltzmann s constant , @xmath18 is the electron rest - mass energy , and the integration is along the line of sight .
the frequency dependence of the sze is @xmath19 , \label{eq : fx}\ ] ] where @xmath20 is the relativistic correction to the frequency dependence .
note that @xmath21 in the nonrelativistic and rayleigh - jeans ( rj ) limits .
it is worth noting that @xmath22 is independent of redshift , as shown in equation [ eq : deltat1 ] .
this unique feature of the sze makes it a potentially powerful tool for investigating the high - redshift universe .
expressed in units of specific intensity , common in millimeter sze observations , the thermal sze is @xmath23 where @xmath24 and the frequency dependence is given by @xmath25 .
\label{eq : gx}\ ] ] @xmath10 and @xmath26 are simply related by the derivative of the blackbody with respect to temperature , @xmath27 .
the spectral distortion of the cmb spectrum by the thermal sze is shown in figure [ fig : sze_di - dt ] ( solid line ) for a realistic massive cluster ( @xmath28 ) , in units of intensity ( left panel ) and rj brightness temperature ( right panel ) . the rj brightness is shown because the sensitivity of a radio telescope is calibrated in these units .
it is defined simply by @xmath29 , where @xmath30 is the intensity at frequency @xmath31 , @xmath17 is boltzmann s constant , and @xmath32 is the speed of light .
the cmb blackbody spectrum , @xmath33 , multiplied by 0.0005 ( dotted line ) , is also shown for comparison .
note that the spectral signature of the thermal effect is distinguished readily from a simple temperature fluctuation of the cmb .
the kinetic sze distortion is shown by the dashed curve ( [ sec : kinetic_sze ] ) . in the nonrelativistic regime , it is indistinguishable from a cmb temperature fluctuation .
the gas temperatures measured in massive galaxy clusters are around @xmath34 kev @xcite and are measured to be as high as @xmath35 kev in the galaxy cluster 1e @xmath36 @xcite . at these temperatures , electron velocities are becoming relativistic , and small corrections are required for accurate interpretation of the sze .
there has been considerable theoretical work to include relativistic corrections to the sze @xcite .
all of these derivations agree for @xmath37 kev , appropriate for galaxy clusters . for a massive cluster with @xmath38 kev ( @xmath39 ) , the relativistic corrections to the sze
are of order a few percent in the rj portion of the spectrum , but can be substantial near the null of the thermal effect .
convenient analytical approximations to fifth order in @xmath40 are presented in @xcite .
the measured sze spectrum of abell 2163 , spanning the decrement and increment with data obtained from different telescopes and techniques , is shown in figure [ fig : a2163_spec ] @xcite .
also plotted is the best - fit model ( solid ) consisting of thermal ( dashed ) and kinetic ( dotted ) sze components .
the sze spectrum is a good fit to the data , demonstrating the consistency and robustness of modern sze measurements .
the most important features of the thermal sze are : ( 1 ) it is a small spectral distortion of the cmb , of order @xmath41 mk , which is proportional to the cluster pressure integrated along the line of sight ( eq . [ eq : y ] ) ; ( 2 ) it is independent of redshift ; and ( 3 ) it has a unique spectral signature with a decrease in the cmb intensity at frequencies @xmath9 ghz and an increase at higher frequencies .
if the cluster is moving with respect to the cmb rest frame , there will be an additional spectral distortion due to the doppler effect of the cluster bulk velocity on the scattered cmb photons .
if a component of the cluster velocity , @xmath42 , is projected along the line of sight to the cluster , then the doppler effect will lead to an observed distortion of the cmb spectrum , referred to as the kinetic sze . in the nonrelativistic limit ,
the spectral signature of the kinetic sze is a pure thermal distortion of magnitude @xmath43 where @xmath42 is along the line of sight ; that is , the emergent spectrum is still described completely by a planck spectrum , but at a slightly different temperature , lower ( higher ) for positive ( negative ) peculiar velocities @xcite .
figure [ fig : sze_di - dt ] illustrates the kinetic sze ( dashed ) for a typical galaxy cluster with a peculiar velocity of 500 @xmath44 .
figure [ fig : a2163_spec ] shows the sze spectrum of the galaxy cluster a2163 along with the best - fit model consisting of thermal ( dashed ) and kinetic ( dotted ) sze components .
relativistic perturbations to the kinetic sze are due to the lorentz boost to the electrons provided by the bulk velocity @xcite .
the leading term is of order @xmath45 , and for a 10 kev cluster moving at 1000 the effect is about an 8% correction to the nonrelativistic term .
the @xmath46 term is only about 1% of the nonrelativistic kinetic sze , and the @xmath47 term is only 0.2% . in the 20 years following the first papers by sunyaev & zeldovich ( 1970 , 1972 ) , there were few firm detections of the sze despite a considerable amount of effort ( see * ? ? ?
* for a review of early experiments ) .
over the last several years , however , observations of the effect have progressed from low signal - to - noise ratio detections and upper limits to high - confidence detections and detailed images . in this section
we briefly review the current state of sze observations .
the dramatic increase in the quality of the observations is due to improvements both in low - noise detection systems and in observing techniques , usually using specialized instrumentation to control carefully the systematics that often prevent one from obtaining the required sensitivity .
the sensitivity of a low - noise radio receiver available 20 years ago should have easily allowed the detection of the sze toward a massive cluster .
most attempts , however , failed due to uncontrolled systematics .
now that the sensitivities of detector systems have improved by factors of 3 to 10 , it is clear that the goal of all modern sze instruments is the control of systematics .
such systematics include , for example , the spatial and temporal variations in the emission from the atmosphere and the surrounding ground , as well as gain instabilities inherent to the detector system used .
the observations must be conducted on the appropriate angular scales .
galaxy clusters have a characteristic size scale of order a mpc . for a reasonable cosmology ,
a mpc subtends an arcminute or more at any redshift .
low - redshift clusters will subtend a much larger angle ; for example , the angular extent of the coma cluster ( @xmath48 ) is of order a degree ( core radius @xmath49 ; herbig et al .
the detection of extended low - surface brightness objects requires precise differential measurements made toward widely separated directions on the sky .
the large angular scale presents challenges to control offsets due to differential ground pick - up and atmospheric variations .
there are three main sources of astronomical contamination and confusion that must be considered when observing the sze : ( 1 ) cmb primary anisotropies , ( 2 ) radio point sources , and ( 3 ) dust from the galaxy and external galaxies .
for distant clusters with angular extents of a few arcminutes or less , the cmb anisotropy is expected @xcite and found to be damped considerably on these scales ( @xcite ; see also holzapfel et al .
1997b and laroque et al .
2003 for cmb limits to sze contamination ) . for nearby clusters , or for searches for distant clusters using beams larger than a few arcminutes , the intrinsic cmb anisotropy must be considered .
the unique spectral behavior of the thermal sze can be used to separate it from the intrinsic cmb in these cases .
note , however , that for such cases it will not be possible to separate the kinetic sze effects from the intrinsic cmb anisotropy without relying on the very small spectral distortions of the kinetic sze due to relativistic effects .
historically , the major source of contamination in the measurement of the sze has been radio point sources .
it is obvious that emission from point sources located along the line of the sight to the cluster could fill in the sze decrement , leading to an underestimate .
radio point sources can also lead to overestimates of the sze decrement for example , point sources in the reference fields of single - dish observations .
the radio point sources are variable and therefore must be monitored .
radio emission from the cluster member galaxies , from the central dominant ( cd ) galaxy in particular , is often the largest source of radio point source contamination , at least at high radio frequencies @xcite . at frequencies near the null of the thermal sze and higher , dust emission from extragalactic sources as well as dust emission from our own galaxy must be considered . at the angular scales and frequencies of interest for most sze observations , contamination from diffuse galactic dust emission
will not usually be significant and is easily compensated .
consider instead the dusty extragalactic sources such as those that have been found toward massive galaxy clusters with the scuba bolometer array @xcite .
spectral indices for these sources are estimated to be @xmath50 @xcite .
sources with 350 ghz ( 850 @xmath51 m ) flux densities greater than 8 mjy are common , and all clusters surveyed had multiple sources with flux densities greater than 5 mjy .
this translates into an uncertainty in the peculiar velocity for a galaxy cluster of roughly 1000 @xmath44 ( see @xcite for details ) . as with sze observations at radio frequencies
, the analyses of high - frequency observations must also consider the effects of point sources and require either high dynamic angular range , large spectral coverage , or both to separate the point source emission from the sze .
the first measurements of the sze were made with single - dish radio telescopes at centimeter wavelengths .
advances in detector technology made the measurements possible , although early observations appear to have been plagued by systematic errors that led to irreproducible and inconsistent results .
eventually , successful detections using beam - switching techniques were obtained . during this period , the pioneering work of birkinshaw and collaborators with the ovro 40 m telescope stands out for its production of results that served to build confidence in the technique @xcite .
figure [ fig : ovro40m_cl0016 ] shows the ovro 40 m scan through the galaxy cluster cl @xmath52 .
more recently , leading and trailing beam - switching techniques have been used successfully with the ovro 5 m telescope at 32 ghz to produce reliable detections of the sze in several intermediate - redshift clusters @xcite .
the sest 15 m and iram 30 m telescopes have been used with bolometric detectors at 140 ghz and chopping mirrors to make significant detections of the sze in several clusters @xcite .
the nobeyama 45 m telescope has also been been used at 21 , 43 , and 150 ghz to detect and map the sze @xcite .
the sunyaev - zeldovich infrared experiment ( suzie ) uses its six - element 140 ghz bolometer array to observe in a drift - scanning mode , where the telescope is fixed and the rotation of the earth moves the beams across the sky . using this drift - scanning technique , the suzie experiment has produced high signal - to - noise ratio strip maps of the sze emission in several clusters @xcite . because of the high sensitivity of bolometric detectors at millimeter wavelengths , single - dish experiments are ideally suited for the measurement of the sze spectrum . by observing at several millimeter frequencies , these instruments should be able to separate the thermal and kinetic szes from atmospheric fluctuations and sources of astrophysical confusion .
the measured sze spectrum of abell 2163 , spanning the decrement and increment with data obtained from different telescopes and techniques , is shown in figure [ fig : a2163_spec ] @xcite .
the sze spectrum is a good fit to the data , demonstrating the consistency and robustness of modern sze measurements .
single - dish observations of the sze are just beginning to reach their potential , and the future is very promising .
the development of large - format , millimeter - wavelength bolometer arrays will increase the mapping speed of current sze experiments by orders of magnitude . to the extent that atmospheric fluctuations are common across a bolometric array
, it will be possible to realize the intrinsic sensitivity of the detectors .
operating from high astronomical sites with stable atmospheres and exceptionally low precipitable water vapor , future large - format bolometer arrays have the potential to produce high signal - to - noise ratio sze images and search for distant sze clusters with unprecedented speed .
the stability and spatial filtering inherent to interferometry has been exploited to make high - quality images of the sze .
the stability of an interferometer is due to its ability to perform simultaneous differential sky measurements over well - defined spatial frequencies .
the spatial filtering of an interferometer also allows the emission from radio point sources to be separated from the sze emission .
there are several other features that allow an interferometer to achieve extremely low systematics .
for example , only signals that correlate between array elements will lead to detected signal . for most interferometers
, this means that the bulk of the sky noise for each element will not lead to signal .
amplifier gain instabilities for an interferometer will not lead to large offsets or false detections , although , if severe , they may lead to somewhat noisy signal amplitude . to remove the effects of offsets or drifts in the electronics , as well as the correlation of spurious ( noncelestial ) sources of noise , the phase of the signal received at each telescope is modulated , and then the proper demodulation is applied to the output of the correlator .
the spatial filtering of an interferometer also allows the emission from radio point sources to be separated from the sze emission .
this is possible because at high angular resolution ( @xmath53 ) the sze contributes very little flux .
this allows one to use long baselines , which give high angular resolution , to detect and monitor the flux of radio point sources , while using short baselines to measure the sze .
nearly simultaneous monitoring of the point sources is important , as they are often time variable .
the signal from the point sources is then easily removed , to the limit of the dynamic range of the instrument , from the short - baseline data , which are sensitive also to the sze . for the reasons given above ,
interferometers offer an ideal way to achieve high brightness sensitivity for extended low - surface brightness emission , at least at radio wavelengths .
most interferometers , however , were not designed for imaging low - surface brightness sources .
interferometers have been built traditionally to obtain high angular resolution and thus have employed large individual elements for maximum sensitivity to small - scale emission . as a result , special - purpose interferometric systems have been built for imaging the sze @xcite .
all of them have taken advantage of low - noise hemt amplifiers @xcite to achieve high sensitivity .
the first interferometric detection @xcite of the sze was obtained with the ryle telescope ( rt ) .
the rt was built from the 5 kilometer array , consisting of eight 13 m telescopes located in cambridge , england , operating at 15 ghz with east - west configurations .
five of the telescopes can be used in a compact e - w configuration for imaging of the sze @xcite .
the ovro and bima sze imaging project uses 30 ghz ( 1 cm ) low - noise receivers mounted on the ovro and bima mm - wave arrays in california .
they have produced sze images toward 60 clusters to date @xcite .
a sample of their sze images is shown in figure [ fig : szpanel12 ] .
all contours are multiples of 2 @xmath54 of each image , and the full - width at half maximum ( fwhm ) of the synthesized beam ( psf for this deconvolution ) is shown in the lower left - hand corner of each image .
figure [ fig : szpanel12 ] also clearly demonstrates the independence of the sze on redshift .
all of the clusters shown have similarly high x - ray luminosities , and , as can be seen , the strength of the sze signals are similar despite the factor of 5 in redshift .
the ovro and bima arrays support two - dimensional configurations of the telescopes , including extremely short baselines , allowing good synthesized beams for imaging the sze of clusters at declinations greater than @xmath55 degrees .
the rt , ovro , and bima sze observations are insensitive to the angular scales required to image low - redshift ( @xmath56 ) clusters .
recently , however , the cosmic background imager ( cbi ; padin et al .
2001 ) has been used to image the sze in a few nearby clusters @xcite .
the cbi is composed of 13 0.9 m telescopes mounted on a common platform , with baselines spanning 1 m to 6 m. operating in 10 1 ghz channels spanning 26 36 ghz , it is sensitive to angular scales spanning 3@xmath57 to 20@xmath57 .
the large field of view of the cbi , 0.75 degrees fwhm , makes it susceptible to correlated contamination from terrestrial sources ( i.e. , ground emission ) . to compensate
, they have adopted the same observing strategy as for single - dish observations ( [ sec : single_dish ] ) , by subtracting from the cluster data , data from leading and trailings fields offset by @xmath5812.5 minutes in right ascension from the cluster .
interferometric observations of the sze , as for single - dish observations , are just beginning to demonstrate their potential .
upcoming instruments will be over an order of magnitude more sensitive .
this next generation of interferometric sze instruments will conduct deep sze surveys covering tens , and possibly hundreds , of square degrees . while not as fast as planned large - format bolometric arrays , the interferometers will be able to survey deeper and provide more detailed imaging .
in particular , the high resolution and deep imaging provided by future heterogeneous arrays will provide a valuable tool for investigating cluster structure and its evolution .
such studies are necessary before the full potential of large sze surveys for cosmology can be realized .
several years after the sze was first proposed @xcite it was recognized that the distance to a cluster could be determined with a measure of its sze and x - ray emission @xcite .
the distance is determined by exploiting the different density dependences of the sze and x - ray emissions .
the sze is proportional to the first power of the density ; @xmath59 , where @xmath15 is the electron density , @xmath16 is the electron temperature , and @xmath60 is along the line - of - sight . the distance dependence is made explicit with the substitution @xmath61 , where @xmath62 is the angular - diameter distance of the cluster .
the x - ray emission is proportional to the second power of the density ; @xmath63 , where is the x - ray cooling function .
the angular - diameter distance is solved for by eliminating the electron density , yielding @xmath64 where these quantities have been evaluated along the line of sight through the center of the cluster ( subscript 0 ) and @xmath65 refers to a characteristic scale of the cluster along the line of sight , whose exact meaning depends on the density model adopted .
only the characteristic scale of the cluster in the plane of the sky is measured , so one must relate the characteristic scales along the line of sight and in the plane of the sky . for detailed treatments of this calculation ,
see @xcite and @xcite . combined with the redshift of the cluster and the geometry of the universe , one may determine the hubble parameter , with the inverse dependences on the observables as that of @xmath62 . with a sample of galaxy clusters ,
one fits the cluster distances versus redshift to the theoretical angular - diameter distance relation , with the hubble constant as the normalization ( see , e.g. , fig .
[ fig : da ] ) . to date
, there are 41 distance determinations to 26 different galaxy clusters from analysis of sze and x - ray observations .
all of these sze / x - ray distances use _ rosat _ x - ray data and model the cluster gas as a spherical isothermal @xmath66 model @xcite .
the _ rosat _
data do not warrant a more sophisticated treatment . in figure
[ fig : da ] we show all sze - determined distances from high signal - to - noise ratio sze experiments .
the uncertainties shown are statistical at 68% confidence .
the theoretical angular - diameter distance relation is shown for three cosmologies assuming @xmath67 @xmath68 .
there are currently three homogeneously analyzed samples of clusters with sze distances : ( 1 ) a sample of seven nearby ( @xmath69 ) galaxy clusters observed with the ovro 5 m telescope ( green solid triangles ) that finds @xmath70 @xcite , ( 2 ) a sample of five intermediate - redshift ( @xmath71 ) clusters from the rt interferometer that finds @xmath72 ( cyan solid squares ) @xcite , and ( 3 ) a sample of 18 clusters with @xmath73 from interferometric observations by the ovro and bima sze imaging project , which infers @xmath74 ( blue solid stars ) @xcite .
the above hubble constants assume a @xmath75 , @xmath76 cosmology , and the uncertainties are statistical followed by systematic at 68% confidence .
the treatment of uncertainties varies among the three cluster samples .
a fit to the ensemble of 41 sze - determined distances yields @xmath0 km s@xmath77 mpc@xmath77 for an @xmath75 , @xmath76 cosmology , where the uncertainties are statistical followed by systematic at 68% confidence . since many of the clusters are at high redshift , the best - fit hubble constant will depend on the cosmology adopted ; the best - fit hubble constant shifts to 57 km s@xmath77 mpc@xmath77 for an open @xmath75 universe , and to 54 km s@xmath77 mpc@xmath77 for a flat @xmath78 geometry .
the systematic uncertainty , discussed below , clearly dominates .
the systematic uncertainty is approximate because it is complicated by shared systematics between some distance determinations .
for example , including multiple distance determinations to a single cluster overstresses any effects of asphericity and orientation from that galaxy cluster .
statistical uncertainty includes contributions from the icm shape parameters , the electron temperature , point sources in the field , and the cooling functions , which depends on @xmath16 , metallicity , and the column density .
the largest sources of statistical uncertainty are the icm shape parameters ( from model fitting ) and the x - ray determined electron temperature ( @xmath79 ) .
uncertainty from model fitting is roughly 20% in the distance and is dominated by the uncertainty in the central decrement .
the contribution from @xmath16 on the distances varies greatly from @xmath80% to @xmath81% , with 10%20% being typical values .
as expected , nearby cluster temperatures are more precisely determined than those of distant galaxy clusters .
the absolute calibration of both the sze and x - ray observations directly affects the distance determinations .
in addition to the absolute calibration uncertainty from the observations , there are possible sources of systematic uncertainty that depend on the physical state of the icm and other sources that can contaminate the cluster sze emission .
table [ tab : h0_sys ] summarizes the systematic uncertainties in the hubble constant determined from 30 ghz interferometric sze observations of a sample of 18 clusters @xcite , but are typical of most sze experiments .
the entries marked with asterisks are expected to average out for a large sample of clusters and include a @xmath82 factor reflecting the 18 clusters used in this work . for detailed discussions of systematics
see @xcite and @xcite .
most clusters do not appear circular in radio , x - ray , or optical wavelengths . under the assumption of axisymmetric clusters , the combined effect of cluster asphericity and its orientation on the sky
conspires to introduce a roughly @xmath83% random uncertainty in @xmath6 determined from one galaxy cluster @xcite .
when one considers a large , unbiased sample of clusters , with random orientations , the errors due to imposing a spherical model are expected to cancel , resulting in a precise determination of @xmath6 .
numerical simulations using triaxial @xmath66 models support this assumption @xcite .
departures from isothermality in the cluster atmosphere may result in a large error in the distance determination from an isothermal analysis .
the _ rosat _
band is fairly insensitive to temperature variations , showing a @xmath84% change in the pspc count rate for a factor of 2 change in temperature for @xmath85 kev gas @xcite .
a mixture of simulations and studies of nearby clusters suggests a 10% effect on the hubble parameter due to departures from isothermality ( e.g. , * ? ? ?
* ; * ? ? ?
clumping of the intracluster gas is a potentially serious source of systematic error in the determination of the hubble constant .
unresolved clumps in an isothermal intracluster plasma will enhance the x - ray emission by a factor @xmath86 , where @xmath87 if significant substructure exists in galaxy clusters , the cluster generates more x - ray emission than expected from a uniform icm , leading to an underestimate of the angular - diameter distance ( @xmath88 ) and therefore an overestimate of the hubble parameter by a factor @xmath86 .
unlike orientation bias , which averages down for a large sample of clusters , clumping must be measured in each cluster or estimated for an average cluster .
there is currently no observational evidence of significant clumping in galaxy clusters .
if clumping were significant and had large variations from cluster to cluster , we might expect larger scatter than is seen in the hubble diagrams from sze and x - ray distances ( fig .
[ fig : da ] ) .
in addition , the agreement between sze ( e.g. , * ? ? ? * ) and x - ray ( e.g. , * ? ? ?
* ) determined gas fractions from galaxy clusters also suggests that clumping is not a large effect .
undetected point sources near the cluster mask the central decrement , causing an underestimate in the magnitude of the decrement and therefore an underestimate of the angular diameter distance .
point sources in reference fields and for interferometers , the complicated synthesized beam shapes , may cause overestimates of the angular diameter distance .
massize clusters typically have central dominant ( cd ) galaxies , which are often radio bright .
therefore , it is likely that there is a radio point source near the center of each cluster .
typical radio sources have falling spectra , roughly @xmath89 , where @xmath90 . at 30 ghz
, possible undetected point sources just below the detection threshold of the observations introduce a @xmath84% uncertainty .
cluster peculiar velocities with respect to the cmb introduce an additional cmb spectral distortion known as the kinetic sze ( see [ sec : kinetic_sze ] ) . for a single isothermal cluster ,
the ratio of the kinetic sze to the thermal sze is @xmath91 where @xmath42 is the peculiar velocity along the line of sight . at low frequencies ( rayleigh - jeans regime ) , @xmath92 and
the kinetic sze is @xmath84% of the thermal sze .
recent observational evidence suggests a typical one - dimensional rms peculiar velocity of @xmath93 @xmath44 @xcite , and recent simulations found similar results @xcite . in general , the kinetic effect is @xmath94% that of the thermal sze , except near the thermal null at @xmath95 ghz where the kinetic sze dominates ( see fig . [
fig : sze_di - dt ] ) .
cluster peculiar velocities are randomly distributed , so when averaged over an ensemble of clusters , the effect from peculiar velocities should cancel .
cmb primary anisotropies have the same spectral signature as the kinetic sze .
the effects of primary anisotropies on cluster distances depend strongly on the beam size of the sze observations and the typical angular scale of the clusters being observed ( nearby versus distant clusters ) ; the cmb effects on the inferred hubble constant should average out over an ensemble of clusters .
recent bima observations provide limits on primary anisotropies on scales of a few arcminutes @xcite . on these scales ,
cmb primary anisotropies are an unimportant ( @xmath3% ) source of uncertainty . for nearby clusters , or for searches for distant clusters using beams larger than a few arcminutes ,
the intrinsic cmb anisotropy must be considered .
the unique spectral signature of the thermal sze can be used to separate it from primary cmb anisotropy .
however , it will not be possible to separate primary cmb anisotropies from the kinetic sze without relying on the very small spectral distortions of the kinetic sze due to relativistic effects . the sze may be masked by large - scale diffuse nonthermal radio emission in clusters of galaxies , known as radio halos .
if present , radio halos are located at the cluster centers , have sizes typical of galaxy clusters , and have a steep radio spectrum ( @xmath96 ; hanish 1982 ; moffet & birkinshaw 1989 ; giovannini et al .
1999 ; kempner & sarazin 2001 ) . because halos are rare , little is known about their nature and origin , but they are thought to be produced by synchrotron emission from an accelerated or reaccelerated population of relativistic electrons ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
conservative and simplistic modeling of the possible effects of these halos implies a @xmath97% overestimate on the inferred hubble parameter from radio halos @xcite .
@xcite show very large array nvss contours overlaid on 30 ghz interferometric sze images , suggesting that radio halos have little impact on sze observations and therefore on sze / x - ray distances .
imprecisely measured beam shapes affect the inferred central decrements and therefore affect the hubble constant . for interferometric observations ,
the primary beam is determined from holography measurements .
conservative and simple modeling of the effects of the primary beam suggests that the effect on the hubble constant is a few percent ( @xmath98% ) at most . in theory , this is a controllable systematic with detailed measurements of beam shape and is currently swamped by larger potential sources of systematic uncertainty . the prospects for improving both the statistical and systematic uncertainties in the sze distances in the near future are promising .
note , from equation [ eq : dadepend ] , that the error budget in the distance determination is sensitive to the absolute calibration of the x - ray and sze observations .
currently , the best absolute calibration of sze observations is @xmath99% at 68% confidence , based on observations of the brightness of the planets mars and jupiter .
efforts are now underway to reduce this uncertainty to the 1% level ( 2% in ) .
uncertainty in the x - ray intensity scale also adds another shared systematic .
the accuracy of the _ rosat _ x - ray intensity scale is debated , but a reasonable estimate is believed to be @xmath100% . it is hoped that the calibration of the _ chandra _ and _ xmm - newton _ x - ray telescopes will greatly reduce this uncertainty .
possible sources of systematic uncertainty are summarized in table [ tab : h0_sys ] .
these values come from 30 ghz interferometric sze observations and _ rosat _ data for a sample of 18 galaxy clusters @xcite , but are typical of most sze distance determinations .
the largest systematic uncertainties are due to departures from isothermality , the possibility of clumping , and possible point source contamination of the sze observations ( for detailed discussion of systematics , see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
_ chandra _ and _ xmm - newton _ are already providing temperature profiles of galaxy clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the unprecedented angular resolution of _ chandra _ will provide insight into possible small - scale structures in clusters .
in addition , multiwavelength studies by existing radio observatories , for example the very large array , can shed light on the residual point source contamination of the radio wavelength sze measurements .
therefore , though currently at the 30% level , many of the systematics can and will be addressed through both existing x - ray and radio observatories and larger samples of galaxy clusters provided from sze surveys .
the beauty of the sze and x - ray technique for measuring distances is that it is completely independent of other techniques , and that it can be used to measure distances at high redshifts directly .
since the method relies on the well - understood physics of highly ionized plasmas , it should be largely independent of cluster evolution .
inspection of figure [ fig : da ] already provides confidence that a large survey of sze distances consisting of perhaps a few hundred clusters with redshifts extending to one and beyond would allow the technique to be used to trace the expansion history of the universe , providing a valuable independent check of the recent determinations of the geometry of the universe from type ia supernovae @xcite and cmb primary anisotropy experiments @xcite .
a promising way of finding high - redshift clusters is to perform deep , large - scale sze surveys , taking advantage of the redshift independence of the sze .
such surveys will provide large catalogs of galaxy clusters , many of which will be at high redshift .
the redshift evolution of the number density of galaxy clusters is critically dependent on the underlying cosmology , and in principle can be used to determine the equation of state of the `` dark energy '' ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
figure [ fig : dndz ] illustrates the dependence of the evolution of cluster number density on cosmology @xcite .
all three cosmologies are normalized to the local cluster abundance .
notice that for low-@xmath101 cosmologies , there are a significant number of galaxy clusters with @xmath102 .
there are a number of dedicated sze experiments under construction that will perform deep , large - scale sze surveys . in the next few years
, there are a number of interferometric approaches that should find hundreds of galaxy clusters .
bolometer array based sze experiments should find roughly thousands of clusters just one year after the interferometric experiments .
the following generation of bolometer array based sze experiments should measure tens of thousands of galaxy clusters on a roughly five - year time scale .
the possible systematics that could affect the yields of sze surveys are presently too large to realize the full potential of a deep sze survey covering thousands of square degrees .
these systematics include , for example , the uncertainties on the survey mass detection limit owing to unknown cluster structure and cluster gas evolution .
these systematics can begin to be addressed through detailed follow - up observations of a moderate - area sze survey .
high - resolution sze , x - ray , and weak lensing observations will provide insights into the evolution and structure of the cluster gas .
numerical simulations directly compared and normalized to the sze yields should provide the necessary improvement in our understanding of the mass function .
soon after it was proposed , it was realized that sze observations could yield distances to galaxy clusters when combined with x - ray data .
sze / x - ray distances are independent of the extragalactic distance ladder and do not rely on clusters being standard candles or rulers .
the promise of direct distances helped motivate searches for the sze in galaxy clusters .
it is only in recent years that observations of the small ( @xmath3 mk ) sze signal have yielded reliable detections and imaging , mostly due to advances in observational techniques .
sze / x - ray distances have progressed from distance determinations one cluster at a time to samples of tens of galaxy clusters . to date , there are 41 sze / x - ray determined distances to 26 different galaxy clusters .
the combined 41 cluster distances imply a hubble parameter of @xmath103 , where the approximate uncertainties are statistical followed by systematic at 68% confidence
. shared systematics between the determined distances make it difficult to determine them precisely . for example , multiple distance determinations to a single galaxy cluster overstress any asphericity and orientation effects .
there are three homogeneously analyzed samples of clusters consisting of five , seven , and 18 galaxy clusters .
even with these small samples , systematics clearly dominate .
systematic uncertainties are approachable with current x - ray and radio observatories .
sze / x - ray determined direct distances to galaxy clusters are beginning to trace out the theoretical angular - diameter distance relation .
it is clear that with a large sample of galaxy clusters at high redshift , sze / x - ray distances will be able to determine the geometry of the universe .
in fact , if for some ( highly unlikely ) reason there was a systematic offset with the sze / x - ray distances , the shape of the curve could still be determined and cosmological parameters inferred .
the redshift independence of the sze will be exploited with surveys of large patches of sky with next generation , dedicated sze experiments .
such surveys will provide large catalogs of clusters , many of which will be at high redshift ( @xmath102 ) , providing large enough samples to use sze / x - ray distances to determine both the hubble constant and the geometry of the universe .
in addition , the cluster yields from deep , large - scale sze surveys are critically dependent on the cosmology , potentially allowing a determination of the equation of state of the `` dark energy . ''
high - resolution sze , x - ray , and optical ( including weak lensing ) observations will provide insights into the evolution and structure of the cluster gas necessary to exploit fully the evolution in cluster yields from sze surveys . | combined with x - ray imaging and spectral data , observations of the sunyaev - zeldovich effect ( sze ) can be used to determine direct distances to galaxy clusters .
these distances
are independent of the extragalactic distance ladder and do not rely on clusters being standard candles or rulers . observations of the sze have progressed from upper limits to high signal - to - noise ratio detections and imaging of the sze .
sze / x - ray determined distances to galaxy clusters are beginning to trace out the theoretical angular - diameter distance relation .
the current ensemble of 41 sze / x - ray distances to galaxy clusters imply a hubble constant of @xmath0 , where the uncertainties are statistical followed by systematic at 68% confidence . with a sample of high - redshift galaxy clusters ,
sze / x - ray distances can be used to measure the geometry of the universe .
[ 1996/06/01 ] a&a |
during the last half - century a great attention is being paid to studies of different quantum models of nonlinear optics since they enable to reveal new physical effects and phenomena ( see , e.g. , @xcite and references therein ) . in view of the hamiltonian nonlinearity these models
are mainly analyzed with the help of the numerical calculations @xcite or some linearization procedures @xcite which are not adapted to reveal many peculiarities of model dynamics @xcite .
however , recently a new universal lie - algebraic approach , essentially improving both analytical and numerical solutions of physical problems , has been suggested in @xcite and developed in @xcite for the class of nonlinear quantum models whose hamiltonians @xmath2 have invariance groups @xmath3=0 $ ] .
this approach is based on reformulations of models under study in terms of dual algebraic pairs @xmath4 @xcite where dynamic symmetry algebras @xmath5 are polynomial lie algebras ( pla ) @xmath6 completely describing model dynamics .
specifically , for a wide class of quantum - optical models with @xmath7 both exact and approximate methods were developed in @xcite to obtain adequate ( for physical applications ) forms of the model evolution operators @xmath8 .
exact methods include appropriate lie - algebraic path - integral schemes @xcite and @xmath9-techniques @xcite based on solving difference and difference - differential equations which are fairly complicated for practical calculations because they yield only algorithms rather than closed analytical expressions for solving physical problems .
therefore , for this aim two analytical approximations of @xmath8 were found in @xcite . they enable to examine model dynamics at cluster quasiclassical levels described by `` large '' values of certain characteristic parameters @xcite .
however , up to now systematic examinations of their accuracy and efficiency were absent . in the present work we cancel in part this lacuna for the simplest and most widely examined ( see , e.g. @xcite and references therein ) example of such models describing process of second - harmonic generation which is given by the hamiltonian @xmath10 , \label{1.1a}\ ] ] and by the fock space @xmath11 as the model hilbert space @xmath12
; here @xmath13 are operators of field modes with frequencies @xmath14 , @xmath15 are coupling constants .
classical solutions of the model ( [ 1.1a ] ) were first found ( in terms of elliptical functions ) in @xcite , and later its different quantum features were examined by many investigators ( see , e.g. , @xcite and references therein ) . specifically , in @xcite some peculiarities , like the collapse - revival phenomena in the janes - cummings model @xcite , were revealed for its model dynamics at the quantum and quasiclassical levels .
the work is organized as follows . after some preliminaries ( section 2 ) we give exact and approximate calculation schemes and routines to determine `` diagonal '' ( spectral ) representations of the model evolution operator by means of the @xmath1-techniques @xcite ( section 3 ) . in section 4
we implement computer experiments to exhibit their efficiency and to investigate the accuracy of appropriate quasiclassical approximations obtained in @xcite . in conclusion
we briefly discuss some of ways to improve the results obtained .
as was shown in @xcite , the invariance of the model hamiltonian ( [ 1.1a ] ) with respect to the group @xmath16 enables to re - formulate the model under study in terms of two sets of collective operators : integrals of motion @xmath17\,(n_{i}= a_{i}^+a_{i } , [ a]$ ] is the entire part of @xmath18 ) and @xmath19-invariant dynamic variables @xmath20 : @xmath21 they obey the commutation relations @xmath22= \pm v_{\pm},\quad [ v_{\alpha } , r_1 ] = 0 = [ v_{\alpha } , k ] , \nonumber \\\ ; & & [ v_- , v_+ ] = \phi ( v_0 ; r_1 ) \equiv \psi ( v_0 + 1 ; r_1 ) -\psi ( v_0 ; r_1 ) , \nonumber \\ & & \psi ( v_{0};r_1 ) = ( r_1 + 2 v_{0})(r_1 + 2 v_{0}-1 ) ( r_1 + 1 - v_{0 } ) , \label{2.2a}\end{aligned}\ ] ] that identifies @xmath23 as generators of pla @xmath1 having the casimir operator @xmath24=0 , \label{2.2b}\end{aligned}\ ] ] acting on @xmath12 complementarily to @xmath19 ( in view of the relationship @xmath25 ) and , hence , forming dynamic symmetry algebra @xmath5 in the dual algebraic pair @xmath4 @xcite . in terms of these collective operators the hamiltonian ( [ 1.1a ] ) is expressed in the form @xcite @xmath26,\;\nonumber \\ & & c = ( \omega_1 + \omega_0 ) r_1,\quad \delta= 2 \omega_1- \omega_0,\quad [ v_{\alpha } , c ] = 0 , \label{2.3a}\end{aligned}\ ] ] and the hilbert space @xmath12 is decomposed into the infinite direct sum @xmath27 of @xmath28-irreducible @xmath29 - dimensional subspaces @xmath30 specified by eigenvalues @xmath31 of the @xmath28 invariant operators @xmath32 . herewith @xmath31 are expressed through the numbers @xmath33 determining , respectively , a maximal population of the fundamental ( pump ) mode and a minimal population of the harmonic within a fixed `` optical atom '' @xmath30 as it follows from the structure of the new ( collective ) basis in @xmath12 @xcite : @xmath34^{1/2 } , \nonumber \\ & & v_0 |k , s;f\rangle = ( l_0+f ) |k , s;f\rangle,\nonumber \\ & & |k , s\rangle= |n_1= k , n_0= s\rangle , \quad v_-\,|k , s\rangle = 0,\quad k=0,1,\,\,s\geq 0 . \label{2.4b}\end{aligned}\ ] ] evidently , eqs .
( [ 2.4b ] ) explicitly manifest the @xmath1 cluster structure of the fock states @xmath35 and specify @xmath36 as the `` lowest '' weight operator and @xmath37 as the `` lowest '' weight state .
this @xmath1 cluster reformulation of the model enables one to use the @xmath1 formalism for getting representations of the model evolution operator @xmath38 which facilitate analysis of the model dynamics and calculations of temporal dependences @xmath39 $ ] for arbitrary physical observables @xmath40 ; @xmath41 is the density operator of the initial quantum state @xcite .
one of such adequate representations has a diagonal form @xcite @xmath42 , \label{2.5}\end{aligned}\ ] ] where @xmath43 and @xmath44 are eigenvalues and complete sets of orthonormalized eigenvectors of the hamiltonian ( [ 2.3a ] ) respectively : @xmath45 in the basis @xmath44 quantum expectations @xmath46 are given as follows @xcite : @xmath47}\ , , \label{2.7}\end{aligned}\ ] ] and , hence , the main problem is in solving the eigenproblem ( [ 2.6 ] ) .
in the case , when the pla structure polynomial @xmath48 is quadratic in @xmath49 and pla @xmath1 is reduced to the familiar lie algebra @xmath50=\pm y_\pm,\,[y_-,y_+]=2 y_0 \}$ ] the eigenvalue problem ( [ 2.6 ] ) is solved exactly with the help of the @xmath0 displacement operators @xmath51 @xcite in terms of simple analytical expressions @xcite . however , it is not the case for the model under study in view of the absence of explicit expressions for matrix elements @xmath52 @xcite .
nevertheless , the @xmath1 formalism enables one get convenient ( for physical applications ) calculation schemes , algorithms and analytical expressions for exact and approximate solutions of this problem .
a lie - algebraic scheme for finding exact solutions of the eigenproblem ( [ 2.6 ] ) is based on looking for eigenfunctions @xmath53 on each subspace @xmath30 in the form @xcite @xmath54 where amplitudes @xmath55 satisfy the orthonormalization and completeness conditions : @xmath56 then , inserting eq .
( [ 3.1a ] ) for @xmath53 and eq .
( [ 2.3a ] ) for @xmath2 in eq .
( [ 2.6 ] ) and using eqs .
( [ 2.2a ] ) , ( [ 2.4b ] ) , one gets a set of recurrence relations at fixed @xmath57 @xmath58\tilde q_f^{v}(k , s ) ) - g\tilde q_{f-1}^{v}(k , s),\nonumber\\ & & f , v=0,\ldots , s,\ ; \nonumber \\ & & \psi ( l_0+f+1 ; l_1)= ( k + 2f+2)(k + 1 + 2f)(s - f ) .
\label{3.2}\end{aligned}\ ] ] these relations along with the boundary conditions @xmath59 determine amplitudes @xmath60 and eigenenergies @xmath61 from solutions of the sturm - liouville spectral problem @xcite @xmath62 p_{f}(\lambda ) - f=0,\ldots , s ; \qquad p_{0}(\lambda ) = 1 , \nonumber\\ & & p_{-1}(\lambda ) = 0 = [ \lambda - \delta ( s + l_0 ) ] p_{s}(\lambda ) - |g|^{2 } \psi ( l_0+s ; l_1 ) p_{s-1}(\lambda ) , \label{3.3b}\end{aligned}\ ] ] for finding non - classical orthogonal ( in view of ( [ 3.1b ] ) ) polynomials @xmath63 of the discrete variable @xmath64 on the non - uniform lattice @xmath65 @xcite . indeed , eqs .
( [ 3.3a ] ) , ( [ 3.3b]),([3.4 ] ) provide the following algorithm for solving the eigenproblem ( [ 2.6 ] ) .
+ i ) using the recursive formula ( [ 3.3a ] ) with the boundary values from eq .
( [ 3.3b ] ) one calculates the polynomial sequence @xmath66 .
+ ii ) inserting @xmath67 in the last equality in ( [ 3.3b ] ) one gets the algebraic equation with respect to @xmath64 ; its solution yields the sequence @xmath68 of admissible values of the spectral parameter @xmath64 and the appropriate energy spectrum @xmath69 .
+ iii ) for each value @xmath70 using @xmath66 and eq . ( [ 3.4 ] ) one finds the sequence @xmath71 of all amplitudes as functions of the only undetermined quantity @xmath72 which , in turn , is found from the normalization condition of eqs . ( [ 3.1b ] ) . + this algorithm has been realized in @xcite with the help of the reduce procedure solve for @xmath73 and the conventional fortran subroutines of eispark package ( with applying the multiprecision package @xcite ) for @xmath74 .
the routine package developed enables us to implement numerical calculations of model dynamics for @xmath75 but it is unsuitable for practical calculations with larger @xmath76 in view of multiprecision computer limitations .
therefore in @xcite an approximate analytical solution of the problem ( [ 2.6 ] ) has been suggested .
it is given by the @xmath0-quasiclassical eigenfunctions @xmath77^{-\frac{1}{2}}=(y_-)^+ , \quad\!\!\ !
2j = s,\nonumber \\ & & y_{\pm , 0}\ , \in\ , su ( 2 ) , \quad s^j_{f v}(\xi)\equiv q_f^{v;\ , ap}(k , s)\,=\,(\frac{g}{|g|})^{f - v}\ , d^{j}_{-j + f
,- j + v}(2r ) , \label{3.5b}\end{aligned}\ ] ] and eigenenergies @xmath78,\nonumber \\ & & \lambda_v^{qc } ( k , s;r)\ , = \ , \delta [ j + l_0 - ( j - v)\cos 2r]+ \nonumber \\ & & + 2|g| \sum_{f=0}^s \sqrt{(s - f)(f+1)2(2k+1 + 2f ) } \,d^{j } _ { -j+f , -j+v } ( 2r ) \ , d^{j}_{-j + f+1 , -j + v } ( 2r ) \nonumber \\ & & \approx \delta [ j + l_0 - ( j - v)\cos 2r]- \nonumber \\ & & - 2|g| ( j - v ) \sin 2r \sqrt{2 [ s+2k+1+(- s+2v)\cos 2r ] } = \lambda_v^{cmf}(k , s;r),\qquad \label{3.6}\end{aligned}\ ] ] where @xmath79 are the @xmath80- functions expressed in terms of the gauss hypergeometric function @xmath81 @xcite .
approximate values @xmath82 in ( [ 3.6 ] ) are calculated in the cluster mean - field approximation : @xmath83 , and values of the parameter @xmath84 in ( [ 3.5a])-([3.6 ] ) are found from energy - stationarity - conditions and/or from minimizing a proximity measure between exact hamiltonian @xmath2 and its @xmath0 - quasiclassical approximation @xmath85 a standard measure for such estimates on the subspaces @xmath30 is defined with the help of the unitarily invariant euclidean operator norm @xcite as follows @xcite @xmath86}{\sum_{v}(\lambda_v ( k , s))^2}. \label{3.8}\ ] ] this approximation has been used in @xcite for calculating approximate expressions @xmath87 of the temporal dependences @xmath88 determining , in accordance with eqs .
( [ 2.1 ] ) , ( [ 3.5b ] ) , the dynamics of the field - mode populations : @xmath89 for different types of initial states ; here @xmath90,\ , \bar k= tr [ \rho k ] $ ] and the quantity @xmath91 is calculated with the help of eq .
( [ 2.7 ] ) for @xmath92 .
specifically , in the case of the @xmath0-quasiclassical cluster initial state @xmath93 of the form ( [ 3.5a ] ) , ( [ 3.5b ] ) with @xmath94 , belonging to a fixed `` optical atom '' @xmath30 with @xmath95 , eq . ( [ 2.7 ] ) yields an approximate analytical expression @xcite @xmath96\right\ } \label{3.10a}\end{aligned}\ ] ] with @xmath97^{\frac{s-1}{2}}\!\!\!\!\!\ ! , \quad \!\tan \phi_{k , s } ( t)= \frac{\tan\omega_{l}(k , s ) t}{\sqrt s } , \label{3.10b}\end{aligned}\ ] ] which exhibits a high - frequency ( @xmath98 ) periodic dynamics with a slow ( @xmath99 ) periodic modulation in the phase and amplitude , i.e. an occurrence of a specific temporal coherent structure ( described in terms of elliptic functions too ) @xcite .
at the same time for general initial states @xmath100 and having non - zero projections on all subspaces @xmath30 , e.g. , for glauber coherent states @xcite , analogous calculations lead to series containing weighted sums of terms like those given by eqs .
( [ 3.10a]),([3.10b ] ) that corresponds to occurrences of coherence - decoherence phenomena like `` collapse - revivals '' revealed in @xcite by means of other methods .
according to the general quasiclassicality theory @xcite all approximations ( [ 3.5a])-([3.6 ] ) ( and , hence , ( [ 3.10a ] ) , ( [ 3.10b ] ) ) are valid only for large values of @xmath76 , and , besides , the measure ( [ 3.8 ] ) gives only a global rather than local characteristic of the approximate energy spectra @xmath101 that does not allow to feel their important symmetry properties and local peculiarities related to `` energy errors '' @xmath102 \equiv \delta
e_v(k , s ) \cdot e_v(k , s ) .
\label{3.11}\ ] ] therefore , we implemented numerical comparisons of both exact and approximate results in order to estimate the applicability range of the quasiclassical approximation ( [ 3.5a])-([3.6 ] ) .
in order to examine the efficiency of calculation schemes and the algorithm given above we tested them by means of computer experiments for the resonance case determined by @xmath2 from eq .
( [ 3.1a ] ) with @xmath103 .
first of all we calculated exact values @xmath104 according to the algorithm of section 3 and their approximations @xmath105 according to eqs .
( [ 3.5a])-([3.6 ] ) for @xmath106 .
values of the fitting parameter @xmath84 were determined from energy - stationarity - conditions : @xmath107 ( optimizing only the upper part of spectra ) @xcite , @xmath108 ( quasi - linear approximation ) @xcite and from minimizing the proximity measure ( [ 3.8 ] ) : @xmath109 ( `` smooth '' cluster mean - field approximation ) @xcite ; herewith @xmath110 means that we take @xmath111 in the first half of spectra and @xmath112 in the second one . to estimate the accuracy of approximations we also used non - invariant measures @xmath113 ^ 2 } { \sum_{v=0}^s ( \lambda_v ( k , s))^2},\nonumber\\ & & \delta^2_{e_{up}}(k , s ) = \frac{\sum_{v = s/2}^s [ ( \lambda_v ( k , s ) - \lambda_v^{cmf}(k , s;r)]^2 } { \sum_{v = s/2}^s ( \lambda_v ( k , s))^2 } , \label{4.1}\end{aligned}\ ] ] to characterize more precisely ( in comparison with eq . (
[ 3.8 ] ) ) energy spectra and standard ( related to the fubini - study metric @xcite in @xmath114 ) measures @xmath115;v}\equiv \sum_{f}\ , s_{fv}^j\ , q_f^v ( k , s ) , \quad \delta^2_{ef}(k , s;v)= 1-|\cos ( { \bf s},{\bf q})_{k , s;v}|^2 \label{4.2}\ ] ] ( or associated graphic representations via `` overlap areas '' ) to estimate an `` approximation quality '' for eigenfunctions . some of typical results of these numerical calculations are presented in table 1 and figs.1,2 .
+ table 1 .
multiplets @xmath116 with the level step @xmath117 for @xmath118 . @xmath119 = @xmath120=@xmath121 = @xmath122=@xmath123 = @xmath124 + 0 -1536.9 -1096.7 -1545.3 -1482.4 -1421.2 + 10 -1151.7 -919.6 -1205.2 -1175.2 -1137.0 + 20 -798.1 -720.0
-880.0 -873.3 -852.7 + 30 -480.3 -499.3 -570.2 -576.7 -568.5 + 40 -205.5 -259.0 -276.7 -285.6 -284.2 + 50 0.0 0.0 0.0 0.0 0.0 + 60 205.5 276.7 276.7 280.0 284.2 + 70 480.3 570.2 570.2 554.3 568.5 + 80 798.1 880.0 880.0 822.8 852.7 + 90 1151.7 1205.2 1205.2 1085.5 1137.0 + 100 1536.9 1545.3 1545.3 1342.3 1421.2 + @xmath125 10.222 -12.220 0.010 -1.000 + @xmath126 2.563 0.670 0.806 0.657 + @xmath127 0.670 0.670 0.944 0.657 + as is seen from data given in table 1 and fig .
1 we have an acceptable consent of exact eigenenergies and their approximations ( [ 3.6 ] ) at @xmath128 almost everywhere for @xmath129 and @xmath130 .
discrepancies between exact and approximate results in the middle parts of spectra , probably , are due to the availability of the square - root singularities in the model hamiltonian ( [ 2.3a ] ) re - written ( with the help of eqs .
( [ 3.5b ] ) ) in terms of @xmath131 that is , actually , ignored in the `` smooth '' the @xmath132 - quasiclassical approximation ( [ 3.5a])-([3.6 ] ) @xcite .
( note also that some negative values of @xmath133 are due to using @xmath134 instead of @xmath135 in eq .
( [ 3.8 ] ) and because of calculation errors ) .
however , the approximation with @xmath130 breaks the orthogonality of eigenfunctions belonging to opposite ends of spectra whereas the quasi - linear approximation with @xmath136 leads to equidistant spectra within fixed subspaces @xmath30 .
therefore , in spite of the spectrum symmetry breaking , the most satisfactory quasiclassical approximation is given by eqs .
( [ 3.5a])-([3.6 ] ) with @xmath137 that minimizes @xmath133 .
note that the spectrum asymmetry at @xmath137 and related shifts between amplitude values @xmath60 and @xmath138 ( see fig .
2 ) are due to using `` smooth '' @xmath132 - quasiclassical eigenfunctions ( [ 3.5a ] ) @xcite . besides the verifications above we also performed calculations of temporal dependences of the quantity @xmath139 and related dynamics of the normalized average photon numbers @xmath140 .
herewith exact dependencies were calculated with the help of routine package above , whereas approximate calculations were implemented using approximate expressions ( [ 3.5a ] ) , ( [ 3.5b ] ) , ( [ 3.6 ] ) for eigenvalues and eigenfunctions .
results of such calculations for @xmath141 against the dimensionless time @xmath142 are plotted in fig .
3 where we compare exact results with the quasiclassical approximations obtained with the help of eqs .
( [ 3.5a])-([3.6 ] ) with @xmath143 and ( [ 3.10a])-([3.10b ] ) .
+ evidently , the graphic representations of fig.3 enable us to reveal transparently a double - periodic component in the exact multi - frequency dynamics of @xmath144 and @xmath145 that is rather well described by eqs .
( [ 3.10a ] ) , ( [ 3.10b ] ) or ( [ 3.5a])-([3.6 ] ) at @xmath137 ( and at @xmath146 ) . note that an availability of this important dynamic feature is displayed clearer when the characteristic parameter @xmath76 increases ( in accordance with the general quasiclassical theory @xcite ) .
so , our numerical calculations given in section 4 show a rather good qualitative consent of exact and approximate results at @xmath147 and at relevant choices of the fitting parameter @xmath84 in ( [ 3.5a])-([3.6 ] ) .
however , partial quantitative discrepancies of them require further improvements of the quasiclassical approximations used . in particular , the approximate solutions of the eigenproblem ( [ 2.6 ] )
can be improved by means of : 1 ) using less smooth ( in comparison with ( [ 3.5a ] ) , ( [ 3.5b ] ) ) generalized coherent states of the @xmath1 algebra as quasiclassical eigenfunctions ( cf .
@xcite ) and 2 ) exploiting the standard @xcite or special ( e.g. , developed in @xcite ) algebraic perturbative and iterative algorithms or modifications of the algebraic `` dressing '' schemes @xcite
. then these improvements ( along with the exact calculation schemes developed above ) can be used for a more detail analysis ( like those implemented in @xcite ) of the model under consideration in all ranges of the parameter @xmath76 and for arbitrary initial states .
it is also of interest to compare results obtained ( and their improvements ) with those of @xcite based on an alternative @xmath0 form of @xmath148 as well as with calculations performed in @xcite using the formalism of the @xmath149 - deformed lie algebra @xmath150 .
the work along these lines is in progress .
armstrong , n. bloembergen , j. ducuing , and p.s .
pershan , phys .
, * 127 * 1918 ( 1962 ) ; n. bloembergen , _ nonlinear optics _ , w. a. benjamin , new york ( 1965 ) .
eberly , n.b .
narozhny , j.j .
sanchez - mondragon , phys .
lett . , * 44 * , 1329 ( 1980 ) .
shen , the principles of nonlinear optics ( wiley , new york , 1984 ) j. perina , quantum statistics of linear and nonlinear optical phenomena .
( reidel , dordrecht 1984 ) .
karassiov , l.a .
shelepin , trudy fian [ p.n .
lebedev inst .
proc ] , * 144 * , 124 ( 1984 ) ; v.p .
karassiov , j. sov .
laser res . ,
* 12 * , 147 ( 1991 ) .
v.p . karassiov and a.b .
klimov , phys .
lett . , * a 189 * , 43 ( 1994 ) .
ou , phys . rev . , * a 49 * 2106 ( 1994 ) .
li , p. kumar , phys .
, * a 49 * 2157 ( 1994 ) .
s.m . chumakov and kozierowski , quantum semiclass
* 8 * , 775 ( 1996 ) ; a. bandilla , g. drobny and i. jex , phys . rev .
, * a 53 * 507 ( 1996 ) .
n. debergh , j. phys . , * a 30 * , 5239 ( 1997 ) ; * a 31 * , 4013 ( 1998 ) .
karassiov , phys .
lett . , * a 238 * , 19 ( 1998 ) ; j. rus .
laser res . ,
* 20 * , 239 ( 1999 ) .
karassiov , j. rus .
laser res . , * 21 * , 370 ( 2000 ) ; phys . atom .
* 63 * , 648 ( 2000 ) ; optika i spektr . ,
* 91 * , 543 ( 2001 ) .
a.b . klimov , l.l .
sanchez - soto , phys .
, * a 61 * , 063802 ( 2000 ) .
karassiov , teor . mat .
fiz . , * 95 * , 3 ( 1993 ) ; j. phys . , * a 27 * , 153 ( 1994 ) .
karassiov , rep . math . phys . , * 40*,235 ( 1997 ) ; czech .
j. phys.,*48 * , 1381 ( 1998 ) .
m. kozierowski , r. tanas , opt . commun . * 21 * 229 ( 1977 ) l. mandel , opt
* 42 * 437 ( 1982 ) .
averbukh , n.f .
perelman , sov .
phys. jetp * 96 * 818 ( 1989 ) ; phys .
lett . , * a 139 * , 449 ( 1989 ) .
nikitin , a.v .
masalov , quantum opt . * 3 * 105 ( 1991 ) .
olsen , et al .
, * a 61 * 021803 ( 2000 ) .
a.m. perelomov , generalized coherent states and their applications .
( nauka , moscow , 1987 ) .
karassiov , a.a .
gusev , s.i .
vinitsky , e - archive : quant - ph/ 0105152 ( 2001 ) ; in : proc .
xxiii inter .
icgtm-23 ( dubna , july 28-august 5 , 2000 ) .
( jinr , dubna , in press ) .
bailey , acm trans .
* 19 * , 288 ( 1993 ) ; * 21 * , 379 ( 1995 ) .
t. kato , perturbation theory for linear operators .
( springer , berlin e.a . , 1965 ) ; p. lankaster , theory of matrices .
( academic , new york - london , 1969 ) .
jaffe , rev .
phys . , * 54 * 407 ( 1982 ) ; a. chatterjee , phys . rep . , * 186 * , 249 ( 1990 ) .
d.n . page , phys .
, * a 36 * 3479 ( 1987 ) ; s. kobayashi and k. nomizu , differential geometry , vol . 2 ( interscience , new york,1969 )
. a. gusev , v. samoilov , v. rostovtsev and s. vinitsky , in : computing algebra in scientific computing : proc .
workshop casc 2000 , v.g .
ganzha , e.w .
mayr , e.v .
vorozhtsov ( eds . ) ( springer , berlin e.a . , 2000 ) , p. 219 .
grebenikov , yu . a. mitropolsky , yu .
a. ryabov , introduction into the resonance analytic dynamics .
( janus - k , moscow,1999 ) .
a. ballesteros and s.m .
chumakov , j. phys . , * a32 * 6261 ( 1999 ) . | we compare exact and @xmath0-cluster approximate calculation schemes to determine dynamics of the second - harmonic generation model using its reformulation in terms of a polynomial lie algebra @xmath1 and related spectral representations of the model evolution operator realized in algorithmic forms .
it enabled us to implement computer experiments exhibiting a satisfactory accuracy of the cluster approximations in a large range of characteristic model parameters .
second - harmonic generation model , polynomial lie algebra methods 42.50 , 42.65 , 02.20 , 03.65 |
ls peg ( s193 ) was first identified as a cataclysmic variable ( cv ) by downes & keyes ( 1988 ) .
it is among the brightest cvs in the sky ( v @xmath0 13.0 szkody et al .
1997 ) which lacks a well - established orbital period .
it is apparently a novalike system ; such systems often show relatively high excitation emission lines , long - term brightness variations , and , in the case of the sw sex stars , orbital variations in the narrow absorption components of and radial velocity phase shifts with respect to the photometric eclipse phase ( warner 1995 ; thorstensen et al .
1991 ) .
the classification of sw sextantis ( sw sex ) stars was created by thorstensen et al .
this classification has been somewhat controversial because it is based on spectroscopic properties while other variable - star classifications are based on photometric properties ; therefore the sw sex classification overlaps with other classifications ( dhillon 1996 ) .
sw sex stars were originally defined as novalike variables with 1 ) eclipses , 2 ) orbital periods between 3 and 4 hours , 3 ) @xmath74686 emission around half the strength of h@xmath8 , 4 ) a balmer - line radial velocity phase that lags substantially behind the orbital phase , 5 ) absorption in balmer and lines opposite the eclipse .
this last characteristic is known as `` phase 0.5 absorption '' , referring to eclipse phase as phase @xmath9 over time , the definition has simplified somewhat , to : novalikes , with phase - dependent absorption around phase 0.5 , in which the spectroscopic phase lags significantly behind the orbital phase ( warner 1995 , dhillon 1996 ) . since inclination depends on line of sight from earth , the `` eclipsing '' characteristic
has increasingly fallen away .
in addition to the defining characteristics , sw sex stars tend to have continuum and/or high - excitation lines eclipsed more strongly than low excitation lines ( warner 1995 and references therein ; young , schneider , & shectman 1981 ; baptista , steiner , & horne 1996 ) , relatively narrow fwhm ( @xmath10 km s@xmath1 ) and large fwzi in the balmer lines , no observable circular polarization ( young , schneider , & shectman 1981 ; grauer et al .
1994 ; thorstensen et al .
1991 ; dhillon & rutten 1995 ; stockman et al .
1992 ) , and a radial temperature profile in the inner disk which is noticably flatter than @xmath11 while in the high state ( baptista , steiner , & horne 1996 ; rutten , van paradijs , & tinbergen 1992 ) .
there is no single comprehensive study of ls peg in the literature , but many observations have been published and we summarize these here .
no stable photometric periods are known , but plenty of quasi - periodic oscillations have been seen in the light curve at different periods : a @xmath12 min period from a one - night optical light curve , an intermittent 18.7 min period over several nights , a weaker quasi period at 32 min , and other possible periods ranging from 5 to 50 min ( garnavich , szkody , & goldader 1988 ; garnavich & szkody 1992 ; szkody et al .
no significant circular polarization was detected ( stockman et al .
analysis of the harvard plate collections from 1897 1988 showed long term high and low states of b @xmath0 12 and b @xmath0 14 ( garnavich & szkody 1992 ) .
_ iue _ and _ voyager _ spectra of ls peg in a high state were similar to sw sex candidates v795 her and pg0859 + 415 ( bp lyn ) , with all these showing deep uv absorption lines .
the line ratios and slope of the uv flux distribution are not typical of dwarf novae at outburst .
the low - state spectrum is most similar to 1h0551 - 819 , with a flat flux distribution , only and lines in emission , and and in deep absorption ( szkody 1991 ; szkody et al .
no x - ray period was determined from _ ginga _ and _ rosat _ data , but the data place an upper limit of 40% on the sinusoidal amplitude .
the differences in column densities and temperatures between the _ rosat _ and _ ginga _ data indicated either a two - component system or a change of accretion characteristics ( szkody et al . 1994a ; szkody et al .
szkody ( 1991 ) reported that the velocities of the balmer - line wings @xmath13 km s@xmath1 , which implies a low inclination .
in addition , very prominent , highly variable ( on timescales @xmath14 minutes ) , and random absorptions in @xmath74471 and balmer emission were seen , with the absorption always redshifted by 250 - 300 km s@xmath1 . by 1995
, balmer - line wings had appeared showing an extremely large velocity amplitude of @xmath151700 km s@xmath1 ; these were the only sinusoidal features seen ( szkody 1995 ) .
no orbital period has been formally published , but szkody et al .
( 1994b ) and szkody ( 1995 ) refer to an orbital period of @xmath154.2 hours from the balmer emission - line wings .
we used the telescopes at mdm observatory on kitt peak , during five observing runs 1996 july , 1996 september , 1996 october , 1996 december , and 1997 june .
table [ tbl-1 ] lists the different combinations of telescopes , spectrographs , and ccd chips , as well as the spectral resolution ( fwhm ) .
the exposure times ranged from 5 to 8 minutes .
the exact dates of observation are implicit in the radial velocity timeseries ( table [ tbl-2 ] ) . to avoid daily cycle count ambiguity , we observed over a wide range of hour angle .
we derived the wavelength scale from comparison lamp spectra bracketing the observations as the telescope tracked .
flux standards were observed when conditions warrented . for data reduction we used standard iraf routines .
the pixel - to - wavelength transformations had typical rms residuals of @xmath16 .
the stability of the 5577.350 night - sky line in the reduced data was @xmath17 km s@xmath1 , and the systematic uncertainty was on the order of 10 km s@xmath1 .
we estimate our flux calibrations to be accurate to about 30 per cent . because most of the lines have an asymmetric wing component
, we could not use the standard algorithms described by schneider & young ( 1980 ) to measure the radial velocities .
since the base of the lines have an approximately constant width ( see figure [ fig6 ] ) and the wings shift back and forth , the radial velocities were measured by hand using the midpoint of the full width zero intensity ( fwzi ) .
the spectra were first rectified using the iraf task _ sfit _ and the fwzi were measured in _
splot_. the times and velocities were corrected to the solar center .
figure [ fig1 ] shows results of a period search conducted using the ` residual - gram ' method described by thorstensen et al .
the best fit is at 5.722 cycles / day .
the monte carlo test of thorstensen & freed ( 1985 ) shows that the cycle count is unambiguous on all scales .
a least - squares sinusoid fit of @xmath18\ ] ] yielded @xmath19 where @xmath20 is the uncertainty of a single measurement inferred from the goodness - of - fit .
figure [ fig2 ] shows the velocities derived from the h@xmath2 wings folded on the best - fit period with the sinusoidal fit superposed .
the @xmath21 value is far too large for any plausible motion of the white dwarf .
interpreted literally , this @xmath21 value yields a secondary mass function of @xmath22 , which , as an illustration , gives @xmath23 assuming @xmath5 and @xmath24 .
we carried out time - series photometry of ls peg for 12 nights ( distibuted over an interval of 18 nights ) in 1996 october , using the telescopes of the center for backyard astrophysics ( harvey et al .
we accumulated 55 hr of coverage , during which the star varied irregularly in the range v=12.5 - 13 .
a sample light curve is shown in the top frame of figure [ fig3 ] .
we searched for periodic signals by calculating the power spectrum ( from the discrete fourier transform ) , and found no stable periods .
the full amplitude upper limit to any signal in the vicinity of @xmath25 was 0.04 mag , and the upper limit at @xmath25 exactly was 0.027 mag .
the individual nights showed features in the power spectrum near 70 cycles d@xmath1 , and a weighted average over the 6 longest nights of coverage yielded the mean power spectrum in the lower frame of figure [ fig3 ] .
the highest peak occurs at a period of @xmath26 min , and corresponds to a full amplitude of 0.041 mag in the 12-night time series .
however , the structure and precise location of the peak are not stable , but drift about on a timescale of hours .
thus this is probably a manifestation of the signal reported at 19 min by garnavich & szkody ( 1992 ) , with the period difference merely arising from erratic drifts in the star .
since the signal is not stable in phase , the measured amplitude of 0.041 mag underestimates the true amplitude of the ( incoherent ) signal , which occasionally appears to reach 0.2 mag in the raw light curve . in cv parlance ,
such things are called quasi - periodic oscillations ( qpos ) .
they are commonly seen in novalike variables ( tt ari : andronov et al .
1998 ; v795 her : zhang et al . 1991
; ah men : patterson 1995 ) , and are oddly clustered near 20 min .
but there is still no understanding of the origin of these signals .
figure [ fig4 ] is an average of 23 spectra from the 1996 october run .
the spectrum has typical cv lines ( see table [ tbl-3 ] ) , including the and blend near @xmath74640 .
there are two weak unidentified lines at approximately @xmath75000 and @xmath76344 .
the line near @xmath75000 is most likely a blend .
the feature near @xmath76344 could be due to or .
interstellar na d lines are also present .
we spend most of this section discussing phase - dependent features in the line profiles which reproduce from orbit to orbit , but the profiles also show secular variations .
figure [ fig5 ] shows rectified spectra from four different orbits at phase @xmath27 = 0.7 . the middle two spectra show absorption in the h@xmath2 and h@xmath8 line wings as well as in the @xmath76678 and @xmath74920 lines , but the upper and lower spectra show little or no absorption . in our data ,
the most pronounced orbit - to - orbit variation occurs roughly between phases @xmath28 , with slight variations mainly in the lines at other phases .
these line profile variations are evidently due to the variable phasing of the appearance and disappearance of the absorption features .
szkody & pich ( 1990 ) have seen these variations in the sw sex stars dw uma , sw sex , and v1315 aql .
figures [ fig6]-[fig8 ] show single - trailed spectra of ls peg .
the single - trailed image was constructed by taking 155 continuum - subtracted spectra from the 1996 september and october runs , computing phases from the ephemeris above , and binning the spectra into 150 phase bins , each of which corresponds to a horizontal line in the image . to compute the spectrum in each phase bin , spectra falling near the central phase of the bin
were combined using weights computed from a gaussian in phase , with @xmath20 = 0.02 cycle .
thus a feature appearing in one input spectrum will be smeared over several lines of the image .
the image lines were stacked to create the 2-d trailed spectrum with the first 50 lines repeated to provide continuity .
note that since there is no eclipse by the secondary , the phase is based on the blue - to - red crossing of the broad wings , as described above . in figure [ fig6 ]
the broad emission wings in the balmer lines are quite prominent .
the width of the emission wings remains approximately constant .
the lines all show braided emission and absorption features , with the absorption crossing over at phase 0.650.9 , suggestive of sw sex stars .
the @xmath29 line shows a similar braided feature .
there also seems to be a faint , broad , high - velocity ( @xmath15 1500 km s@xmath1 ) , blueshifted emission feature in the lines starting at phase 0.4 and ending around phase 1.05 ; it is most prominent in @xmath30 ( see figure [ fig7 ] ) , and is barely visible in @xmath31 , @xmath32 , and @xmath33 .
unfortunately , the interstellar na d lines obscure any possible corresponding redshifted emission of @xmath30 , though close examination of the other lines suggests that there is no redshifted emission .
also , the absorption crosses from red to blue at phase @xmath0 0.8 .
both the / blend and lines show no phase - dependent features .
the constant velocity and sharpness of the na d lines confirms that they are interstellar .
figure [ fig8 ] is a close up of the h@xmath2 and @xmath31 region of the single trailed spectrum .
the h@xmath2 wings extend out to @xmath342000 km s@xmath1 and have a width of approximately 3000 km s@xmath1 . at phases
@xmath35 , there is relatively strong blueshifted absorption in the h@xmath2 wing and the edge of the line core .
this occurs at the same phase as the absorption red - to - blue crossing , most strongly seen in figure [ fig7 ] .
there appears to be fainter redshifted absorption in the h@xmath2 wing and line core at phase @xmath36 .
the faint , broad , high - velocity , blueshifted emission in @xmath31 occurs at the same phase as the blueshifted h@xmath2 wing .
note that the maximum blueshift of the faint emission in @xmath31 occurs at the same phase as absorption crossing from red to blue ( see figure [ fig7 ] for a clearer example in @xmath30 ) .
we created doppler tomograms of h@xmath2 ( figure [ fig9 ] ) and @xmath31 ( figure [ fig10 ] ) using the fourier - filtered back - projection algorithm ( marsh & horne 1988 , horne 1991 ) . unlike the maximum entropy method ( mem ) algorithm , the fourier - filtered back - projection algorithm does not require only positive values in the single - trailed spectrum ( see the appendix of harlaftis & marsh 1996 for a comparison of the two algorithms ) .
we subtracted unity from our single - trailed spectrum ( which was created from continuum - divided spectra ) to obtain a mean of zero , as required by the doppler tomogram algorithm .
the doppler tomogram is not very sensitive to the assumed system parameters ; we took @xmath37 , @xmath38 , @xmath39 km s@xmath1 , @xmath40 km s@xmath1 , and @xmath41 .
@xmath42 and @xmath43 are plausible values for sw sex stars , @xmath44 is determined from the mass - period relationship ( @xmath45(h ) ; warner 1995 ) , and @xmath46 is determined from our model described in 5 .
the model was also used to estimate the orbital phasing , so the tomograms could be rotated slightly . the difference between the h@xmath2 core and wing tomograms is shown in figure [ fig9 ] .
the core tomogram ( lefthand panel ) is similar to v795 her ( casares et al .
1996 ; dickinson et al .
1997 ) , px and ( hellier & robinson 1994 ; still , dhillon , & jones 1995 ) , and v1315 aql ( dhillon , marsh , & jones 1991 ; hellier 1996 ) .
it suggests that the line core emission is from the disk edge .
the wing tomogram ( righthand panel ) is most similar to v795 her wing tomogram ( figure 13b in dickinson et al . 1997 ) , and to a lesser extent , sw sex ( dhillon , marsh , & jones 1997 ) and v1315 aql ( hellier 1996 ) .
the high - velocity wing emission appears to originate from both the stream and the inner disk suggesting the stream - disk interaction at the re - impact site in the disk - overflow models .
however , the exact stream trajectory is uncertain since we do nt know @xmath43 .
the circular pattern is an artifact of the phase - smeared noise in the single - trailed spectra .
the predicted single - trailed spectra reasonably match the data , which suggests the emission is near the plane of the disk . in our @xmath47 tomogram ( figure [ fig10 ] ) , and to a lesser extent in the h@xmath2 tomograms ,
we violate the basic premise of doppler tomography : that the emission / absorption is seen at all phases .
this violation is clearly seen in figure [ fig10 ] in the predicted trailed spectrum , in which the absorption extends in phase to produce more of an `` s '' wave not seen in the original data .
the emission is similar to v795 her @xmath47 tomograms ( casares et al .
1996 ; haswell et al .
the absolute phasing is not known in ls peg or v795 her , so again some rotation of the tomograms may occur .
in an effort to understand the origin of the very high - velocity balmer - line wings , we used a modified version of the disk - overflow model of hellier and robinson ( 1994 ) .
our purpose was not to exhaustively explore parameter space , nor to consider radiative transfer problems , but simply to arrive at a plausible kinematic model .
the stream motions were approximated by particle trajectories , which were computed in a roche potential using code described by thorstensen et al .
the stellar masses were held fixed at plausible values ( @xmath37 , @xmath38 ) . in hellier & robinson s scenario ,
a portion of the accretion stream flows over the top of the disk and strikes the disk close to the white dwarf . at the point where the stream re - impacts the disk
, we assumed the emission to be a broad gaussian with a @xmath20 equal to 50% of the average of the keplerian disk and the stream free - fall velocities .
as in hellier and robinson s model , our re - impact region is slightly upstream from the re - impact point interpolated from table 1 in lubow ( 1989 ) for our adopted parameters .
we also assume that prior to re - impact the stream creates absorption , which we model as a narrower gaussian with @xmath20 equal to 25% of the free - fall velocity .
we then added a double - peaked keplerian disk profile .
note that in the model , zero phase corresponds to inferior conjunction of the secondary , while the phase of the fit to the broad line wing motion in the data ( hereafter the `` data phase '' ) has no clear dynamical interpretation .
our simulations bore a reasonable resemblence to the data using @xmath48 , disk radius = @xmath49 cm , and stream re - impact region @xmath50 cm from the white dwarf .
figures [ fig11 ] and [ fig12 ] show comparisons of the single - trailed h@xmath2 spectrograms and model .
the model phase was matched by eye to the data phase .
the model matches the gross features we observe , in particular the very high velocity broad wings of the balmer lines .
it does not explain , however , the relatively strong absorption in the wing and edge of the line core at data phase @xmath0 0.8 ( model phase @xmath0 1.4 ) and to a lesser extent at data phase @xmath0 0.25 ( model phase @xmath0 0.85 ) . also , in the model the line core shows an orbital sinusoidal velocity variation , while the observed line core remains at nearly constant velocity ( see figure [ fig12 ] ) .
the absorption in phase with the emission wings suggests some absorption associated with the re - impact point .
however when we extended the absorption along the entire stream to include the re - impact region as in the model of v795 her by dickinson et al .
( 1997 ) , we get absorption in the broad wings at later phases than that is seen in the data .
but it is important to note that our model is relatively simple .
how does this system compare to others in which large - amplitude motion and absorption effects occur ?
the effects seen in the sw sex stars are referred in phase to the eclipse ( sw sex stars as defined by thorstensen et al .
( 1991 ) always show eclipses ) . because ls peg does not eclipse ,
we do not know the absolute orbital phase . however , our model is computed using mass points to generate the potential , so the orbital phase in the model is always known to arbitrary accuracy .
our model is matched by eye to the data , giving a rough relationship between orbital phase and spectroscopic phase . to the extent
our model is realistic , we can ask whether the line wing motion might be offset in phase from the ` expected ' motion of the white dwarf center of mass , and whether the absorption events occur opposite the expected ` eclipse ' phase .
both these effects are seen in sw sex stars ( thorstensen et al . 1991 ) .
using the by - eye match between the model and our spectroscopic ` data phase ' ( for which @xmath51 at blue - to - red crossing ) , the inferior conjunction of the red star ( which would yield an eclipse in a high - inclination system ) occurs at data phase 0.4 . if the line base traced the white dwarf motion ( unlikely in this case , but often assumed ) , the eclipse would be expected at data phase 0.5 . thus the spectroscopic phase ` lags ' the phase expected for white - dwarf motion by @xmath52 cycle .
the sense of the offset is the same as seen in sw sex stars , but it is somewhat smaller than the offsets seen in most sw sex stars , in which offsets of 0.15 to 0.2 cycles are common ( garnavich et al . 1990 ; honeycutt et al . 1986 ; thorstensen et al . 1991 ; casares et al . 1996 ) . again using the model - to - data phase offset , we find that absorption occurring opposite the eclipse would appear around data phase @xmath53 .
the absorption is actually seen near data phase @xmath54 . in view of the uncertain correspondence between binary phase and data phase , this is fair agreement ; alternatively , the match to standard sw sex behavior could be improved for both the line motion and the absorption if a slightly different phase offset were assumed .
the model suggests , then , that the relationship between dynamical ( eclipse ) phase and spectroscopic phase is similar to that found in sw sex stars . the absorption seen in the lines
also superficially resembles that seen in sw sex stars . to explore
how accurately the absorption in ls peg resembles the sw sex phenomenon , we constructed figure [ fig13 ] , which compares the ls peg spectrogram to a single - trailed spectrogram of px and , which was constructed from data obtained 1988 december and 1989 february ( already discussed by thorstensen et al .
the px and data are of lower spectral resolution than the present ls peg data , but the presentation is otherwise similar .
the resemblance between behaviors of the @xmath75015 absorption in these two stars is detailed and striking .
note especially in px and how the weak redshifted absorption appears soon after eclipse , grows in strength , and moves toward the blue , with maximum strength occurring opposite the eclipse .
the absorption in ls peg mimics this behavior perfectly .
note that @xmath55 is plotted on the left axes of both panels ; these phases are derived from fits to the h@xmath2 velocities , and the two panels are aligned with respect to this phase .
the eclipse phase plotted for px and is derived from the observed eclipse , whereas ( again ) the ls peg eclipse phase is _ inferred _ from the dynamical model of the emission lines .
again , there is fair agreement between these two left axes
they differ by 0.11 cycle .
table [ tbl-4 ] gives the estimated phases of the absorption events in both stars in both phase conventions . in an attempt to constrain the absorption phases , hellier ( 1998 ) adds a flared disk to the disk - overflow model and , unlike the previous versions of the model , includes spatial obscuration of the disk and stream by the secondary , obscuration of parts of the disk by the stream , obscuration of the stream by a flared disk , and obscuration of the stream by other other regions of the stream .
previous versions of the model ( hellier & robinson 1994 ; hellier 1996 ) produced the absorption features at all phases . in hellier
s recent model , the strength and phase dependence of the absorption depends on the departure of @xmath56 from @xmath57 , where _ i _ is the inclination and @xmath2 is the flare angle of the disk .
he found if @xmath58 , the phase - dependent absorption features weaken and become visible at all phases .
based on theoretical and observational estimates , hellier assumes a flare angle , @xmath2 @xmath59 .
we can estimate an upper limit for the inclination of ls peg .
the firmest limit comes from the absence of eclipses . because we do nt know the exact masses , we use the values assumed in our model , @xmath37 and @xmath60 , and estimate from table 1 in bailey ( 1990 ) that @xmath61 , which corresponds to a grazing eclipse . for hellier s model to work at this inclination
, the flare angle must be greater than @xmath62 , at least twice as large as the theoretical and observational estimates .
in fact , the inclination of ls peg is probably much lower than @xmath63 since there is no photometric signal at the orbital period .
our model provides a reasonable fit at an inclination of @xmath64 , which would imply an unrealistic flare angle of over @xmath65 .
we obtain a radial velocity period of @xmath3 d. this is almost certainly the orbital period .
the phase - dependent absorption is reminiscent of sw sex stars , and a comparison between our data , our model , and archival data of the well - studied sw sex star px and suggests that the absolute binary phase in ls peg is about as expected for the sw sex phenomenon .
there is a striking qualitative resemblance between the phase - dependent absorption features observed in ls peg and those observed in sw sex stars . as new objects are discovered , the sw sex behaviors listed by thorstensen et al .
( 1991 ) continue to appear together as a suite .
although eclipses are formally necessary in the original definition of sw sex stars , phase - dependent absorption appears to be fairly common in non - eclipsing bright novalikes ( such as wx ari ; beuermann et al .
the formal criterion that a system be eclipsing for admission to the sw sex class was always somewhat unsatisfactory , as it depends on our accidental viewing angle rather than on a physical characteristic of the system ; in this sense the phase - dependent absorption features are a more fundamental criterion .
we believe that ls peg should be classified as an sw sex star even though it does not eclipse .
the ( modest ) success of our simple kinematic model in accounting for the broad line wings and the indirect evidence that the absolute binary phase predicted by the model is as expected given the behavior of other sw sex stars suggests strongly that the disk - overflow model proposed by hellier & robinson ( 1994 ) is essentially correct .
the striking broad line wings seen in figures [ fig6 ] , [ fig8 ] , and [ fig11 ] are almost certainly coming from material associated with the stream overflowing the disk and the stream re - impact point .
however , as casares et al .
( 1996 ) pointed out , hellier & robinson s disk - overflowing stream scenario does not limit the phase of the absorption to around @xmath66 .
the inclination of ls peg is so low that a flared disk such as that proposed by hellier ( 1998 ) _ can not _ explain the phase dependence of the absorption , so a new explanation must be sought .
a clue may come from recent hydrodynamical simulations of the stream - disk interaction by armitage & livio ( 1996 , 1998 ) .
for inefficent cooling , an explosive stream - disk interaction produces a `` halo '' of material above and below the disk as well as a bulge along the disk rim downstream of the stream impact .
because of the surface brightness of the disk is highest toward its center , the absorption produced by such out - of - plane material could be highly dependent on viewing angle .
we see no evidence of a strictly periodic modulation in the photometry . in view of the lack of polarization ,
the lack of a strict photometric period , and the similarity to sw sex stars ( none of which are known to be magnetic ) , there is little to corroborate suggestions that this is a magnetic cv .
thus we can consider it unlikely that the magnetic accretion models of casares et al .
( 1996 ) and williams ( 1989 ) apply in this case .
we warmly thank bob fried at braeside observatory and the columbia university astronomy c3997 fall 1995 class for taking the photometric data .
we also thank keith horne for providing the fourier - filtered back - projection doppler tomography code .
we thank the nsf for support through grant ast-9314787 and the mdm staff for their usual excellent support .
this research made use of the simbad database , operated at cds , strasbourg , france .
andronov , i.l .
1998 , , in press armitage , p.j . &
livio , m. 1996 , , 470 , 1024 armitage , p.j . & livio , m. 1998 , , 493 , 898 bailey , j. 1990 , , 243 , 57 baptista , r. , steiner , j.e .
, & horne , k. 1996 , , 282 , 99 beuermann , k. , thorstensen , j.r . ,
schwope , a.d . , ringwald , f.a .
, & sahin , h. 1992 , , 256 442 casares , j. , martinez - pais , i.g . , marsh , t.r . ,
charles , p.a . , & lazaro , c 1996 , , 278 , 219 dhillon , v. 1996 , in cataclysmic variables and related objects , ed .
evans , a. & wood , j. h. ( dordrecht : kluwer academic ) , 3 dhillon , v.s . , marsh , t.r .
, & jones , d.h.p .
1991 , , 252,342 dhillon , v.s . , marsh , t.r . , & jones , d.h.p .
1997 , , 291,694 dhillon , v.s . &
rutten , r.g.m .
1995 , , 277 , 777 dickinson , r.j . ,
prinja , r.k . , rosen , s.r . , king , a.r . , hellier , c. , & horne , k. 1997 , , 286 , 447 downes , r.a .
& keyes , c.d .
1988 , , 96 , 777 garnavich , p. , szkody , p. , & goldader , j. 1988 , , 20 , 1020 garnavich , p.m. , szkody , p. , mateo , m. , feinswog , l. , booth , j. , goodrich , b. , miller , h.r . ,
carini , m.t . , & wilson , j.w .
1990 , , 365 , 696 garnavich , p. , & szkody , p. 1992 , aavso j , 21 , 81 grauer , a.d . , ringwald , f.a . ,
wegner , g. , liebert , j. , schmidt , g.d .
, & green , r.f .
1994 , , 108 , 214 harlaftis , e.t . & marsh , t.r .
1996 , , 308 , 97 harvey , d. , skillman , d. , patterson , j. , & ringwald , f.a .
1995 , , 107 , 551 haswell , c. , horne , k. , thomas , g. , patterson , j. , & thorstensen , j.r .
1994 , in interacting binary stars , asp conference series vol .
a shafter ( san francisco , astronomical society of the pacific ) , p. 268 hellier , c . 1996 , , 471 , 949 hellier , c. & robinson , e.l . 1994 , , 431 , l107 honeycutt , r.k . , schlegel , e.m . , & kaitchuck , r.h .
1986 , , 302 , 388 horne , k. 1991 in fundamental properties of cataclysmic variable stars : proc .
12th north american workshop on cataclysmic variables and low mass x - ray binaries , ed .
a. w. shafter ( san diego , san diego state univ . ) , p. 23
lubow , s.h .
1989 , , 340 , 1064 marsh , t.r . &
horne , k. 1988 , , 235 , 269 patterson , j. 1995 , , 107 , 657 rutten , r.g.m . ,
van paradijs , j. , & tinbergen , j. 1992 , , 260,213 still , m.d . ,
dhillon , v.k .
, jones , d.h.p .
1995 , , 273 , 863 stockman , h.s . ,
schmidt , g.d . ,
berriman , g. , liebert , j. , moore , r.l . , & wickramasinge , d.t .
1992 , , 401 , 628 szkody , p. 1991
, in fundamental properties of cataclysmic variable stars , ed .
a. shafter ( san diego , san diego state university ) , p. 56
szkody , p. 1995 , in cape workshop on magnetic cataclysmic variables , asp conference series vol .
d. a. h. buckley & b. warner ( san francisco , astronomical society of the pacific ) , p. 54
szkody , p. , hoard , d.w . , patterson , j. , moulden , m. , howell , s. , & garnavich , p. 1994a
, in interacting binary stars , asp conference series vol .
56 , ed . a shafter ( san francisco , astronomical society of the pacific ) , p. 350 szkody , p. , garnavich , p. , castelaz , m. , & makino , f. 1994b , , 106 , 616 szkody , p. , garnavich , p. , holberg , j. , silber , a. , & pastwick , l. 1997 , , 113 , 2276 szkody , p. & pich , f. 1990 , , 361 , 235 schneider , d. & young , p. 1980
, , 238 , 946 thorstensen , j.r . & freed , i.b .
1985 , , 90 , 2082 thorstensen , j.r . ,
ringwald , f.a . ,
wade , r.a . ,
schmidt , g.d . , & norsworthy , j.e .
1991 , , 102 , 272 thorstensen , j.r . , patterson , j.o . , shambrook , a. , & thomas , g. 1996 , , 108 , 73 warner , b. 1995 , cataclysmic variable stars ( cambridge , cambridge university press ) williams , r.e .
1989 , , 97 , 1752 young , p. , schneider , d.p .
, & shectman , s.a .
1981 , , 244,259 zhang , e. , robinson , e.l . ,
ramseyer , t.f .
, shetrone , m.d . ,
& stiening , r.f .
1991 , , 381 , 534 clccccc 1996 jul . & 19 & 2.4 & modular & tek @xmath67 thinned & 4610 - 6688 & 3 1996 sep .
& 71 & 1.3 & mkiii & loral @xmath68 unthinned & 4310 - 7184 & 5 1996 oct .
& 68 & 2.4 & modular & loral @xmath68 unthinned & 4585 - 7171 & 3 1996 dec .
& 17 & 1.3 & mkiii & tek @xmath67 thinned & 4472 - 6760 & 5 1997 jun .
& 4 & 2.4 & modular & tek @xmath68 thinned & 4000 - 7502 & 3.5 lrlrlrlr @xmath69 & @xmath70 & @xmath71 & @xmath72 & @xmath73 & @xmath74 & @xmath75 & @xmath76 @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 @xmath85 & @xmath86 & @xmath87 & @xmath88 & @xmath89 & @xmath90 & @xmath91 & @xmath92 @xmath93 & @xmath94 & @xmath95 & @xmath96 & @xmath97 & @xmath98 & @xmath99 & @xmath100 @xmath101 & @xmath102 & @xmath103 & @xmath104 & @xmath105 & @xmath106 & @xmath107 & @xmath108 @xmath109 & @xmath110 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath116 @xmath117 & @xmath118 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 & @xmath124 @xmath125 & @xmath126 & @xmath127 & @xmath128 & @xmath129 & @xmath130 & @xmath131 & @xmath132 @xmath133 & @xmath134 & @xmath135 & @xmath136 & @xmath137 & @xmath138 & @xmath139 & @xmath140 @xmath141 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & @xmath146 & @xmath147 & @xmath148 @xmath149 & @xmath150 & @xmath151 & @xmath152 & @xmath153 & @xmath108 & @xmath154 & @xmath155 @xmath156 & @xmath128 & @xmath157 & @xmath158 & @xmath159 & @xmath160 & @xmath161 & @xmath162 @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath90 @xmath170 & @xmath171 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath162 @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 & @xmath182 & @xmath183 & @xmath184 @xmath185 & @xmath186 & @xmath187 & @xmath188 & @xmath189 & @xmath190 & @xmath191 & @xmath192 @xmath193 & @xmath194 & @xmath195 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 & @xmath208 @xmath209 & @xmath210 & @xmath211 & @xmath96 & @xmath212 & @xmath213 & @xmath214 & @xmath215 @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath138 & @xmath221 & @xmath222 @xmath223 & @xmath224 & @xmath225 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 @xmath231 & @xmath232 & @xmath233 & @xmath188 & @xmath234 & @xmath235 & @xmath236 & @xmath74 @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath251 & @xmath252 @xmath253 & @xmath248 & @xmath254 & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath175 & @xmath265 & @xmath266 @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 & @xmath272 & @xmath273 & @xmath274 @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 @xmath283 & @xmath188 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & @xmath289 @xmath290 & @xmath285 & @xmath291 & @xmath292 & @xmath293 & @xmath138 & @xmath294 & @xmath285 @xmath295 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath98 & @xmath300 & @xmath301 @xmath302 & @xmath261 & @xmath303 & @xmath144 & @xmath304 & @xmath280 & @xmath305 & @xmath306 @xmath307 & @xmath308 & @xmath309 & @xmath238 & @xmath310 & @xmath311 & @xmath312 & @xmath313 @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 @xmath322 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 & @xmath337 @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & @xmath345 @xmath346 & @xmath323 & @xmath347 & @xmath348 & @xmath349 & @xmath350 & @xmath351 & @xmath352 @xmath353 & @xmath348 & @xmath354 & @xmath355 & @xmath356 & @xmath350 & @xmath357 & @xmath232 @xmath358 & @xmath120 & @xmath359 & @xmath348 & @xmath360 & @xmath361 & @xmath362 & @xmath363 @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath319 & @xmath369 & @xmath219 @xmath370 & @xmath371 & @xmath372 & @xmath373 & @xmath374 & @xmath98 & @xmath375 & @xmath376 @xmath377 & @xmath378 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 @xmath385 & @xmath325 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 & @xmath391 @xmath392 & @xmath150 & @xmath393 & @xmath387 & @xmath394 & @xmath382 & | we present time - resolved spectroscopy and photometry of the bright cataclysmic variable ls peg (= s193 ; v @xmath0 13.0 szkody et al . 1997 ) .
the balmer lines exhibit broad , asymmetric wings doppler - shifted by about 2000 km s@xmath1 at the edges , while the lines show phase - dependent absorption features strikingly similar to sw sextantis stars , as well as emission through most of the phase .
the / emission blend does not show any phase dependence . from velocities of h@xmath2 emission lines , we determine an orbital period of @xmath3 d (= 4.1946 h ) , which agrees with szkody s ( 1995 ) value of approximately 4.2 hours .
no stable photometric signal was found at the orbital period .
a non - coherent quasi - periodic photometric signal was seen at a period of @xmath4 min .
the high - velocity balmer wings most probably arise from a stream re - impact point close to the white dwarf .
we present simulated spectra based on a kinematic model similar to the modified disk - overflow scenario of hellier & robinson ( 1994 ) .
the models reproduce the broad line wings , though some other details are unexplained . using an estimate of dynamical phase based on the model
, we show that the phasing of the emission- and absorption - line variations is consistent with that in ( eclipsing ) sw sex stars .
we therefore identify ls peg as a low - inclination sw sex star .
our model suggests @xmath5 , and the observed absence of any photometric signal at the orbital frequency establishes @xmath6 .
this constraint puts a severe strain on interpretations of the sw sex phenomenon which rely on disk structures lying slightly out of the orbital plane . |
the coolest and lowest mass objects have historically been discovered as companions to low - luminosity stars .
these objects include the two lowest luminosity spectral classes of low mass stars and brown dwarfs - the l and t dwarfs ( * ? ? ?
* and references within ) .
the first l dwarf , gd165b , was discovered as the companion to a white dwarf @xcite while the first widely accepted brown dwarf , gliese229b , was the companion to an m dwarf @xcite .
hence it is quite possible that the first of the ultracool brown dwarfs with effective temperatures less than @xmath2700k will also be discovered as a companion .
these objects would likely populate a new spectral type beyond t with masses overlapping the planetary regime .
the hunt today for even cooler objects benefits from advances in high resolution imaging with the _ hubble space telescope _ ( hst ) and large ground - based telescopes fitted with adaptive optics ( ao ) . with the spectral energy distributions of these cool objects peaking in the near - infrared ( 1 - 6 ) , observing at these wavelengths are advantageous for their detection and characterization .
thus observational strategies have relied on targetting continuously lower luminosity objects to further improve the contrast differential obtained in the near - infrared .
an advantage in using ao over the hst is that they can be attached to larger ground - based telescopes attaining higher angular resolution and increased sensitivity to fainter sources .
the challenge , however , exists in locating natural guide stars sufficiently bright ( @xmath313.5mag , @xmath412mag ) and near one s science target ( isoplanatic angular distance @xmath530@xmath0 ) to provide sufficient wavefront correction . there is less than 10% chance of finding a natural guide star ( ngs ) meeting these requirements at 30galactic latitude @xcite .
the probability improves little even with the use of infrared wavefront sensors when targeting ultracool objects such as mid - l dwarfs ( limiting magnitude of the ngs infrared wavefront sensor on naos at the very large telescope is @xmath612mag ) .
slightly better ngs sensitivity performance has been achieved using curvature wavefront sensors with avalanche photodiodes where @xmath412.3mag @xcite .
the search for substellar and planetary - mass objects through direct detection from ground - based telescopes now has a new technique - laser guide stars ( lgss ) .
lgss serve as artificial beacons for ao systems by exciting sodium atoms in the earth s mesosphere at their resonant d - line frequency .
these beacons serve as artificial ( and steerable ) guide stars which provide sufficient flux density for wavefront sensing and correcting .
while lgs ao still requires a ngs to help correct both the lowest wavefront orders ( `` tip / tilt '' , @xmath22khz ) and the higher orders ( `` low - band wavefront sensor '' , @xmath20.01hz ) , its flux density requirement is comparatively small ( @xmath718mag ) .
this results in @xmath22/3 of the night sky accessible to high spatial resolution imaging @xcite and opens the door to probing the regions around ultracool l and t dwarfs never previously observed by ground - based telescopes .
while the keck ii is the first of the 8 - 10 m class telescopes to have an operational lgs ao system @xcite , several more are expected to be commissioned within just the next 2 years ( see * ? ? ?
several recent investigations using keck ii lgs ao have discovered companions to previously unresolved faint sources ushering in this new era of high - resolution imaging ( eg . * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in this investigation we observe six nearby ultracool2700k ) ; @xcite . ]
field dwarfs which we target for very faint companions .
the objects were selected from the literature satisfying the following criteria : spectral type later than m6 , never observed at high spatial resolution , too faint for current ground - based ngs ao systems , _ and _ spectrophotometric distances less than 30pc .
we present here the discovery of one of the targets , 2massj21321145 + 1341584 @xcite as a closely - separated ( 0.066@xmath0 ) l dwarf binary resolved by the keck ii telescope nirc2 infrared camera in combination with lgs ao .
the binary is hereafter referred to as 2m2132 + 1341ab .
the five other targets not found with near - equal mass companions are listed in table [ tbl-1 ] .
this discovery demonstrates the power of lgs ao - the ability to resolve a faint ( @xmath8@xmath920mag ; @xmath1016mag ) binary very near the diffraction limit ( 50mas ) of a 10 m telescope using an artificial beacon for wavefront correction .
the discovered binary system 2m2132 + 1341ab was observed on ut 2006 june 17 with the 10 m keck ii telescope on mauna kea , hawaii .
it was the lone binary discovered from our sample of six ultracool dwarf targets . to optimize the resolution capabilities of our observations
, we used the facility ir camera nirc2 in the narrow ( 0.01/pixel ) camera mode with a 10@xmath1110@xmath0 field of view , in combination with the sodium lgs ao system @xcite .
all targets were observed through the broadband _ k@xmath12
_ filter ( 2.15@xmath13 ) where strehl ratios are improved over _ j _ ( 1.25@xmath13 ) and _ h _ ( 1.63@xmath13 ) . in the case of 2m2132
+ 1341ab , observations were also made in both @xmath14 and @xmath15 .
all filters are of the mauna kea observatories ( mko ) filter consortium @xcite .
conditions were photometric for the majority of the night with better than 0.6@xmath0 seeing in the optical but with occasional windy periods .
higher - order ao corrections were produced using the laser s on - axis light in the direction of the science target .
this produced an emission similar to a _
v_@xmath1610 point source .
lower order tip / tilt corrections were obtained using natural guide stars within 60@xmath0 of the targets . in the case of 2m2132 + 1341ab ,
the natural guide star used was 1036 - 0598908 ( _ r_=14.2mag ) from the usno - b1.0 catalog @xcite , located 13.6@xmath0 away .
table [ tbl-1 ] lists the five ultracool dwarf targets observed with no near - unity mass ratio companions detected at separations @xmath90.05@xmath0 .
figure [ fig1 ] shows the resolved discovered binary 2m2132 + 1341ab .
both components are elongated along the telescope elevation axis projected to @xmath1645@xmath17 , attributed to windshake during the observations ( a common problem with lgs ao on windy nights ; see also * ? ? ?
the lgs ao - corrected images have full width at half - maximum of 0.06@xmath0 , 0.07@xmath0 , 0.07@xmath0 at _ j , h , _
@xmath18 , respectively .
each of the images shown in figure [ fig1 ] was made by dithering a few arcsecs over three different quadrant positions on the nirc2 narrow camera detector .
three images were taken per filter per dither position resulting in 2.5min for total on - source integration time per filter .
the object was easily resolved into two components in all our data .
the images were reduced in a consistent manner using an ao data reduction pipeline written in the iraf language as first described in @xcite . modified for the nirc2 narrow camera ,
the pipeline produces final unsaturated 15@xmath19 exposures in _ j , h , _ and @xmath18 with the highest signal - to - noise in the inner 5@xmath20 region . the photometric reduction pipeline uses the iraf task allstar in the daophot point spread function ( psf ) fitting photometry package .
the central regions of the pipeline s output is shown in figure [ fig1 ] .
two different unsaturated single objects observed during the same night , but from a different program , were selected as psf stars .
these objects were observed with the same instrumental setup and showed similar strehl ratios , fwhm , elongation due to windshake , and air mass .
both psf sources , incidentally , are resolved primary objects of newly discovered wide ( 1.7@xmath0 - 1.9@xmath0 ) binaries . in both cases ,
the a and b components are sufficiently separated such that there is no flux contamination between them .
the two psfs used are 2massj16233609 - 2402209a , also shown in figure [ fig1 ] , and 2massj16223020 - 2322240a @xcite .
2m2132 + 1341ab was fit independently with both psf objects leaving behind clean residuals .
the differential photometry in magnitudes is reported in table [ tbl-2 ] and the photometric uncertainties are produced from the differences in the photometry between the two psfs .
these dominate the overall uncertainty .
unresolved , low - resolution near - infrared spectroscopy of 2m2132 + 1341ab was obtained on 2005 october 17 ( ut ) using the spex spectrograph mounted on the 3 m nasa infrared telescope facility @xcite .
conditions during the night were clear with moderate seeing ( 0@xmath211@xmath220 at @xmath14-band ) .
data were obtained utilizing the spex prism mode , providing a single - order spectrum spanning 0.8 - 2.5@xmath13 spectrum with a dispersion of 20 - 30 / pixel .
use of the 0@xmath225 slit , aligned with the parallactic angle , provided resolution @xmath23 @xmath16 120 across the near - infrared band .
six exposures of 2m2132 + 1341ab were obtained in an abba dither pattern along the slit , with individual exposure times of 150s .
the system was observed at an airmass of 1.02 .
the a0 star hd 210501 was observed immediately after the target exposures at a similar air mass , followed by internal flat - field and ar arc lamps for pixel response and wavelength calibration .
data were reduced using the spextool package , version 3.2 @xcite .
the raw science data were processed by performing linearity corrections , pairwise subtraction , and division by a normalized flat field .
the spectra were then extracted using the spextool default settings for point sources , and wavelength solutions were calculated using the ar arc calibration frames .
extracted spectra from the same source were scaled to match the highest singal - to - noise spectrum of the set , and the scaled spectra were median - combined .
telluric and instrumental response features were removed following the procedure of @xcite .
the reduced spectrum of 2m2132 + 1341ab is shown in figure [ fig2 ] ( _ black line _ ) .
strong absorption features of h@xmath24o are present at 1.4 and 1.9@xmath25 m , co is prominent at 2.3@xmath25 m , and feh is present at 0.99 , 1.2 , and 1.6@xmath25 m .
the @xmath14-band spectral region exhibits a number of features that can be attributed to ki and nai lines in addition to feh .
there is no indication of ch@xmath26 in the spectrum of this source .
these features are all indicative of a late - type l dwarf , as also indicated by the optical spectrum of @xcite .
the key binary properties of 2m2132 + 1341a and b are derived here and summarized in table [ tbl-3 ] .
individual apparent magnitudes are calculated from the observed @xmath27 magnitudes ( table [ tbl-2 ] ) and the integrated apparent magnitudes ( unresolved ) measured by the two micron all sky survey ( 2mass ) @xcite .
since the differential photometry observed with nirc2 was measured with the mko filter system , we converted the integrated _ j _ and _ h _ photometry from the 2mass filter system to mko using the color transformations of @xcite . while they provide no transformation between _
k@xmath12 _ mko and _ k@xmath12 _ 2mass , the transmission curves are very similar ( 1 - 2% difference , s. leggett private communication ) and therefore we apply no correction .
uncertainties in the transformations and photometry are propagated in quadrature and reported in table [ tbl-3 ] . with measured differential photometry , derived apparent magnitudes , and a measured combined [email protected] optical spectrum from @xcite ,
what physical properties of the individual components can we infer ?
since there is no known trigonometric parallax for the object , we rely on first estimating the component spectral types to derive absolute magnitudes using an empirical relation .
this then enables estimates of the distance modulus , bolometric luminosities , and ultimately , with the aid of theoretical evolutionary tracks , masses and a period .
the better constrained the component spectral types are , the more constrained ( and meaningful ) will the derived physical properties be .
the component near - infrared colors listed in table [ tbl-2 ] by themselves provide only rough constraint on the individual spectral types ( eg . * ? ? ?
while the primary is certainly a mid - l dwarf , l3-l8 , the possible spectral types for the secondary extend into the t - range , l3-t1 . while the combined light spectrum is similar to that of an l6
, the components may have very discrepant spectral types .
the secondary could even be a t dwarf without the characteristic ch@xmath26 bands appearing in the combined light spectrum . to derive more precise estimates of the individual component spectral types , we used a spectral synthesis technique based on that used by @xcite to study l dwarf plus t dwarf binaries .
a large sample of composite spectra were generated by combining various pairings of l5-t6 spex prism spectra obtained by a. burgasser & k. cruz ( 72 individual spectra in all ) .
the spectral types of the template spectra are based on optical classifications for l dwarfs ( e.g. * ? ? ?
* ) and near - infrared classifications for the t dwarfs ( e.g. * ? ? ?
the components of these spectra were constrained to have the same relative @xmath18-band magnitudes as measured for 2m2132 + 1341ab , and to simultaneously be within 3@xmath28 of the measured @xmath29 ( 0.27mag ) and @xmath30 ( 0.12mag ) .
the best matches between the composite spectra and the observed ( unresolved ) spectrum of 2m2132 + 1341ab were quantitatively determined by comparing both relative @xmath14 and @xmath15 magnitudes and h@xmath24o and ch@xmath26 spectral ratios ( defined in * ? ? ? * ) .
no assumption was made on the absolute magnitudes of the individual components in this analysis so that the absolute magnitude / spectral type scale was left as a free parameter .
figure [ fig2 ] illustrates the three best - fit composite spectra based on both the relative magnitudes and spectral ratio comparisons . in all three cases ,
an l5 spectral classification is selected for the primary along with an l7 or l8 for the secondary .
in fact , this was the case for the best twenty fits . while the uncertainty is dominated by one subclass of uncertainties in the individual library spectral classifications , the consistency in the matches likely average out the overall uncertainty .
hence we conclude that the primary is a likely [email protected] and its companion a likely [email protected] .
the fourth fit shown in figure [ fig2 ] ( bottom right ) shows one of the combinations that was disqualified due to the disparity between the predicted and measured @xmath27 magnitudes .
the components of this system are the unusually red l5 2massj062445.95 - 452154.8 ( reid et al .
, in prep . ) and the unusually blue l7 2massj09083803 + 5032088 @xcite . despite a good morphological fit , this kind of analysis that includes @xmath27 magnitudes as constraints to the properties of individual components is sufficiently robust to remove atypical component spectra . of course ,
resolved near - infrared spectroscopy is required to verify the accuracy of these classifications .
are the components of 2m2132 + 1341ab physical companions ? calculating spectrophotometric distances of the two sources separately results in equal values , 28@xmath14pc .
we used the fitted spectral types to independently obtain intrinsic flux densities ( m@xmath31 ) from the polynomial fit of @xcite .
the distance s uncertainty includes those in the spectral types and in a spectral type / absolute magnitude relation ( see next section ) taken in quadrature . in addition , assuming a surface density of order 10@xmath32 deg@xmath33 @xcite for all nearby l dwarfs , the probability of two lying within 0.1@xmath0 is @xmath1610@xmath34 .
hence random alignment is very unlikely .
lastly , the 2m2132 + 1341 pair , or at least the primary , shows a large proper motion of 0.4@xmath0/yr ( nomad ; * ? ? ?
a 2mass _ k@xmath12 _ image of 2m2132 + 1341 observed in ut 1998 appears single . with sensitivity to point source brightness of k@[email protected] , 2m2132 +
1341b would have been detectable and resolved at separations @xmath91.5@xmath0 @xcite .
therefore we rule out 2m2132 + 1341b as an unrelated background object since its projected position nearly 8yrs ago would have been resolved in the 2mass image .
these factors provide strong evidence that the two sources are physical companions . with well - constrained spectral types in hand
, we can now derive many of the binary s physical properties summarized in table [ tbl-3 ] .
an absolute @xmath35 magnitude for the primary is obtained using an m@xmath31-spectral type relation from figure 3 of @xcite , where binaries have been excluded .
the companion absolute magnitude is then obtained by applying our measured @xmath36 ( @xmath37 ; s. leggett , priv .
comm . ) . using our constrained component spectral types
, we acquire the _
k_-band bolometric corrections from @xcite , apply them to our @xmath38 photometry , and calculate the bolometric luminosity ratio between the components to be [email protected] .
individual bolometric luminosities in units of solar luminosity , estimated from the component m@xmath31 and bc@xmath31 values , are 6.3@xmath3910@[email protected]@xmath3910@xmath40l@xmath41 for the primary and 3.0@xmath3910@[email protected]@xmath3910@xmath40l@xmath41 for the companion .
individual masses of 2m2132 + 1341a and b can be estimated from theoretical evolutionary models using our derived bolometric luminosities and estimated ages of the system .
the system s age , however , is less constrained . the binary does not appear affiliated with any moving - group or open cluster .
its optical spectrum shows no lithium or h@xmath42 spectral features @xcite suggesting that the source is more consistent with old field l dwarfs @xcite .
neither is there near - infrared color or optical spectrum evidence of sub - solar metallicity ( eg . * ? ? ?
* ; * ? ? ?
* ) indicating the system is probably not a member of the galaxy s thick disk or halo populations ( @xmath910gyr ; * ? ? ?
in addition , the system s tangential motion of 53@xmath12kms@xmath43 ( nomad ; * ? ? ? * ) is inconsistent with a young object ( @xmath51gyr ) . the lack of lithium absorption in the optical spectrum can help place a lower mass limit to 2m2132 + 1341a of approximately 0.065m@xmath41 depending on the system s age @xcite .
objects less than this limiting mass will always have central temperatures below the lithium - burning temperature .
slightly more massive objects will undergo lithium burning such that the element is observable at only younger ages .
for example , using the models of @xcite , a 0.075m@xmath41 object will undergo complete lithium burning in about 140myr . in figure [ fig3 ]
we show their theoretical evolutionary tracks where we place the lower mass limit of 2m2132 + 1341a along a constant lithium abundance line of 1% of the original abundance ( similar to @xcite , we assume that a decrease in the initial lithium abundance by a factor of 100 marks the lithium absorption detection limit ) .
this provides a lower age limit of 0.8 - 1.3gyr , consistent with a weak or absent lithium absorption feature . assuming the companion is coeval with the primary , this lower age along with the uncertainties in the secondary s luminosity predicts masses of 0.040 - 0.054m@xmath41 . a 10gyr upper limit results in a primary mass of 0.077 - 0.079m@xmath41 and a secondary of 0.076 - 0.077m@xmath41 . according to a theoretical analysis conducted by @xcite of the age distribution of nearby field l dwarfs ,
there is a @xmath44 probability that 2m2132 + 1341a and b are less than @xmath21gyr and a @xmath45 chance that they are younger than @xmath25gyr .
we list the median mass estimates for three ages in table [ tbl-3 ] including a 5 gyr best guess for stars in the solar neighborhood . both objects
likely straddle the hydrogen burning mass threshold ( @xmath160.072 - 0.075m@xmath41 ; * ? ? ?
* ; * ? ? ?
* ) with the secondary most likely substellar .
the projected separation between the two components is only [email protected] ( at a distance of 28@xmath14pc ) .
we estimate the semimajor axis of 2m2132 + 1341ab by assuming that on average the true semimajor axis is 1.26 times larger than the projected separation @xcite or @xmath46=2.3au . using kepler s third law and the range of possible masses
, we estimate an orbital period of 7 - 12yr .
hence , this system is a good candidate target for astrometric monitoring to derive orbital mass measurements @xcite .
vlm binaries are characterized by near - unity mass ratios ( [email protected] - 1.0 ) and tight separation distributions peaking between 3 - 10au ( * ? ? ? * and references within ) . according to the very low - mass binaries
this mass limit is arbitrary and corresponds to binary m6 field dwarfs ( slightly earlier spectral types for younger objects ) .
the website is maintained by nick siegler at http://paperclip.as.arizona.edu/@xmath2nsiegler/vlm@xmath47binaries ] , about a third of these systems are l / l binaries .
the binary properties of 2m2132 + 1341ab , [email protected] and projected separation of [email protected] , are consistent with these distributions .
currently there are 16 known vlm binaries with angular separations less than the mean 66mas separation of 2m2132 + 1341ab .
the tightest nine are spectroscopic binaries and are as yet unresolved .
the subsequent seven were all discovered with the hst . despite large aperture ground - based telescopes achieving
ao corrected resolutions at _ k _ typically twice that of the hst , the space telescope s more stable point spread function allows for the identification of undersampled binaries .
interestingly , _ 2m2132 + 1341ab is the tightest resolved very low - mass binary discovered by a ground - based telescope and the tightest using lgs ao_. the clear separation of this system into two well - resolved components indicates that with good ao correction , ground - based facilities can indeed achieve a superior resolution in the near - infrared compared to hst .
theoretical evolutionary models relating mass - luminosity - age relations are still largely uncalibrated for the lowest mass objects .
in fact , only three vlm systems with constrained ages ( all young ) have had reliable orbits _ and _ resolved fluxes leading to derived individual kinematic masses - ab dor c ( @xmath250 - 100myr ; * ? ? ? * ; * ? ? ? * ) , the eclipsing brown dwarf binary 2massj05352184 - 0546085ab found in orion ( @xmath21 - 2myr ; * ? ? ? * ) , and gj569bab ( @xmath2500myr ; * ? ? ? * ) .
unfortunately , neither 2m2132 + 1341ab s age nor distance is sufficiently well constrained to be used as a high accuracy luminosity - mass calibrator . however ,
if future high resolution optical spectroscopy ( eg . repaired space telescope imaging spectrograph on hst ) shows the presence of lithium in the companion , the system s age could be further constrained to @xmath20.8 - 2.5gyr making it a useful system for dynamical mass measurements .
this would likely require , however , a widening in the components projected separation .
the hst and/or ground - base lgs ao observations should be able to measure significant orbital motion over the next @xmath26yrs , similar to the study of 2masswj0746425 + 2000321 @xcite , the only dynamical mass measurement of an l dwarf binary .
keck ii lgs ao observations of 2massj21321145 + 1341584 show that this very low - mass dwarf is a binary system
. observed differential near - infrared photometry and integrated spectra ( optical and near - infrared ) indicate that both components are consistent with mid - l dwarfs .
based on modeling the integrated optical spectra with spectra from 72 known l and t dwarfs , we identify 2m2132 + 1341a as an an [email protected] and 2m2132 + 1341b as an [email protected] .
the lack of lithium in the optical spectra suggests the primary s age is older than 800myr .
the system s very close separation ( 0.066@xmath0 ) and common proper motion from 2mass infers a physical association . with conservative age estimate of 5gyr ,
model - dependent masses suggest a system whose components straddle the hydrogen burning limit threshold with the primary likely stellar and the secondary likely substellar .
the close projected separation ( [email protected] ) and near unity mass ratio of the system are consistent with previous results for field vlm binaries .
the relatively short estimated orbital period of this system ( @xmath27 - 12yr ) make it an ideal target for dynamical mass measurements . at the time of this writing , 2m2132 + 1341ab s angular separation is
the tightest for any vlm binary discovered from a ground - based telescope and is the tightest using lgs ao .
the authors would like to acknowledge nasa and the nasa tac for making this time available and the entire keck lgs ao team for having set the bar for lgs performance .
we also thank the referee kevin luhman for a thorough reading and suggested improvements .
n.s . would like to thank mike cushing for discussions regarding l and t spectral classifications and adam burrows for providing model calculations .
c.m . and b.m .
note that their research was performed under the auspices of the u.s .
department of energy by the university of california , lawrence livermore national laboratory under contract w-7405-eng-48 , and also supported in part by the nsf science and technology center for ao , managed by the university of california at santa cruz under cooperative agreement ast 98 - 76783 .
this research has made use of the simbad and vizier databases operated at cds in strasbourg , france ; the two micron all sky survey ( 2mass ) data services , a joint project of the university of massachusetts and the infrared processing center / california institute of technology , funded by nasa and the nsf ; the u.s .
naval observatory ( usno ) naval observatory merged astrometric dataset ( nomad ) .
iraf is distributed by the national optical astronomy observatories , which is operated by the association of universities for research in astronomy , inc .
, under contract to the nsf .
allen , p. r. , kroener , d. w. , reid , i. n. , & trilling , d. e. 2005 , , 625 , 385 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 1998 , , 337 , 403 basri , g. , marcy , g. w. , & graham , j. r. 1996 , , 458 , 600 becklin , e. e. & zuckerman , b. 1988 , 336 , 656 bouchez , a. h. , et al .
2004 , , vol .
5490 , 321 , advancements in adaptive optics , ed .
d. calia , b. l. ellerbroek , & r. ragazzoni bouy , h. , et al .
2004 , , 423 , 341 burgasser , a. j. , et al .
2003 , , 592 , 1186 burgasser , a. j. 2004 , , 614 , l73 burgasser , a. j. , kirkpatrick , j. d. , & lowrance , p. j. 2005a , 129 , 2849 burgasser , a. j. , reid , i. n. , leggett , s. k. , kirkpatrick , j. d. , liebert , j. , & burrows , a. 2005b , , 634 , 177 burgasser , a. j. & kirkpatrick , d. j. 2006 , , 645 , 1485 burgasser , a. j. , et al .
2006a , , 166 , 585 burgasser , a. j. , geballe , t. r. , leggett , s. k. , kirkpatrick , j. d. , & golimowski , d. a. 2006b , , 637 , 1067 burgasser , a. j. , reid , i. n. , siegler , n. , close , l. m. , allen , p. , lowrance , p. j. , & gizis , j. e. , in planets and protostars v , eds b. reipurth , d. jewit , and k. keil ( univ .
arizona press : tucson ) , 951 , 2007 , 427 burgasser , a. j. 2007 , , in press burrows , a. et al .
1997 , , 491 , 856 burrows , a. , hubbard , w. , b. , lunine , j. , i. , & liebert , j. 2001 , review of modern physics , 73 , 719 chabrier , g. , baraffe , i. , & plez , b. 1996 , 459 , l91 chappelle , r. j. , pinfield , d. j. , steele , i. a. , dobbie , p. d. , & magazzu , a. 2005 , 361 , 1323 chiu , k. , et al .
2006 , , 131 , 2722 close , l. m. , et al .
2002 , , 566 , 1095 close , l. m. , et al .
2005 , , 433 , 286 close , l. m. , et al .
2007 , , in press cruz , k. l. , reid , i. n. , liebert , j. , kirkpatrick , j. d. , & lowrance , p. j. 2003 , 126 , 2421 cruz , k. l. , et al .
2004 , , 604 , l61 cruz , k. l. , et al .
2007 , , 133 , 439 cushing , m. c. , vacca , w. d. , & rayner , j. t. 2004 , , 116 , 362 cushing , m. c. & vacca , w. d. 2006 , , 131 , 1797 cutri , r. m. et al . 2003 , 2mass all - sky catalog of point sources ( pasadena : ipac ) dahn , c. c. , et al .
2002 , , 124 , 1170 fischer , d.a . & marcy , g. w. 1992 , , 396 , 178 gelino , c. r. , kulkarni , s. r. , & stephens , d. c. 2006 , , 118 , 611 golimowski et al .
2004 , , 127 , 3516 hawley , s. l. et al .
2002 , , 123 , 3409 kirkpatrick , j. d. , henry , t. j. , & simons , d. a. 1995 , , 109 , 797 kirkpatrick , j. d. , et al .
1999 , , 519 , 802 kirkpatrick , j. d. , reid , i. n. , liebert , j. , gizis , j. e. , burgasser , a. j. , monet , d. g. , dahn , c. c. , nelson , b. , williams , r. j. 2000 , , 120 , 447 kirkpatrick , j. d. 2005 , , 43 , 195 lane , b. f. , zapatero osorio , m. r. , britton , m. c. , martn , e. l. , & kulkarni , s. r. 2001 , , 560 , 390 leggett , s. k. , et al .
2007 , , in press lpine , s. , rich , r. m. , & shara , m. m. 2003 , , 591 , l49 liu , m. c. & leggett , s. k. 2005 , , 634 , 616 liu , m. c. et al .
2006 , , 647 , 1393 liu , m. c. 2006 , , vol .
6272 , in press , advances in adaptive optics ii , b. l. ellerbroek & c. d. bonaccini luhman , k. l. , stauffer , j. r. & mamajek , e. e. 2005 , , 628 , l69 maxted , p. f. l. & jeffries , r. d. 2005 , , 326 , l45 mclean , i.s . , graham , j. r. , becklin , e. e. , figer , d. f. , larkin , j. e. , levenson , n. a. & teplitz , h.l .
2000 , , vol . 4008 , 1048 , optical and it telescope instrumentation and detectors , ed
. m. iye & a. moorwood monet , d. g. et al .
, 2003 , , 125 , 984 nakajima , t. , oppenheimer , b. r. , kulkarni , s. r. , golimowski , d. a. , matthews , k. , & durrance , s. t. 1995 , , 378 , 463 oppenheimer , b. r. , kulkarni , s. r. , matthews , k. , & nakajima , t. 1995 , science , 270 , 1478 pinfield , d. j. et al . , 2003 , , 342 , 1241 rebolo , r. , martn , e. l. , & magazzu , a. 1992 , , 389 , l83 rayner , j. t. , toomey , d. w. , onaka , p. m. , denault , a. j. , stahlberger , w. e. , vacca , w. d. , cushing , m. c. , & wang , s. 2003 , , 115 362 reid , i. n. & hawley , s. l. 2005 , new light on dark stars , ( 2nd ed ; chichester : praxis publishing ltd reid , i. n. , lewitus , e. , allen , p. r. , cruz , k. l. , & burgasser , a. b. 2006 , , 132 , 891 roddier , f. 2004 , adaptive optics in astronomy ( cambridge , uk : cambridge university press ) siegler , n. , close , l. m. , & freed m. 2002 , vol . 4839 , 114 , adaptive optics systems technologies ii , ed p. wizinowich & d. bonaccini simons , d. a. & tokunaga , a. 2002 , , 114 , 169 simons , d. a. & tokunaga , a. 2002 , , 114 , 169 stassun , k. g. , mathieu , r. d. , & valenti , j. a. 2006 , , 440 , 311 thompson , l. a. & gardner , c. s. 1987 , , 328 , 229 tokunaga , a. , simons , d. a. , & vacca , w. d. 2002 , , 114 , 180 vacca , w. d. , cushing , m. c. , & rayner , j. t. 2003 , , 115 , 389 west , a. a. , et al . 2004 , , 128 , 426 wizinowich , p. l. , et al . 2004 , , vol . 5490 , 1 , advancements in adaptive optics , ed .
d. calia , b. l. ellerbroek , & r. ragazzoni wizinowich , p. l. , et al .
2006 , , 118 , 297 zacharias , n. , monet , d. g. , levine , s. e. , urban , s. e. , gaume , r. , & wycoff , g. l. 2004 , , 205 , # 48.15 zapatario - osorio , m. r. , lane , b. f. , pavlenko , ya . , martn , e. l. , britton , m. , & kulkarni , s. r. 2004 , , 615 , 958 lsrj1610 - 0040 & 12.02 & sdm : sdl : &
1,2 + 2massj17210390 + 3344160 & 12.47 & l3 & 3 + sdssj202820.32 + 005226.5 & 12.79 & l3 & 4 + 2massj20343769 + 0827009 & 13.08 & m : l & 5 + 2massj22490917 + 3205489 & 13.59 & l5 & 6 + lc @xmath48 & [email protected] + @xmath49 & [email protected] + @xmath50 & [email protected] + @xmath51 & [email protected] + @xmath52 & [email protected] + @xmath53 & [email protected] + @xmath54 & [email protected] + @xmath55 & [email protected] + @xmath56 & [email protected] + separation & 66@xmath14mas + position angle & [email protected]@xmath17 + date observed & ut 2006 jun 17 lc spectral types & + a + b ( optical ) & l6@xmath11 + a & [email protected] + b & [email protected] + @xmath57 & [email protected] + @xmath58 & [email protected] + distance & 28@xmath14pc + luminosities : + l@xmath59 & 6.3@xmath3910@[email protected]@xmath3910@xmath40l@xmath41 + l@xmath60 & 3.0@xmath3910@[email protected]@xmath3910@xmath40l@xmath41 + masses ( a / b ) : & + 0.8gyr & 0.065/0.048m@xmath41 + 5gyr & 0.077/0.075m@xmath41 + 10gyr & 0.078/0.076m@xmath41 + proper motion ( nomad ) : & + @xmath61cos@xmath62 & [email protected]@xmath43 + @xmath63 & [email protected]@xmath43 + separation ( projected ) & [email protected] + orbital period & 7 - 12yr + | we present the discovery of 2massj21321145 + 1341584ab as a closely separated ( 0.066@xmath0 ) very low - mass field dwarf binary resolved in the near - infrared by the keck ii telescope using laser guide star adaptive optics .
physical association is deduced from the angular proximity of the components and constraints on their common proper motion .
we have obtained a near - infrared spectrum of the binary and find that it is best described by an [email protected] primary and an [email protected] secondary .
model - dependent masses predict that the two components straddle the hydrogen burning limit threshold with the primary likely stellar and the secondary likely substellar .
the properties of this sytem - close projected separation ( [email protected] ) and near unity mass ratio - are consistent with previous results for very low - mass field binaries .
the relatively short estimated orbital period of this system ( @xmath27 - 12yr ) makes it a good target for dynamical mass measurements .
interestingly , the system s angular separation is the tightest yet for any very low - mass binary published from a ground - based telescope and is the tightest binary discovered with laser guide star adaptive optics to date . |
core - holes are created when matter is irradiated with energetic photons .
the net binding energy is obtained from the kinetic energy of the photoelectron upon subtracting the photon energy .
as the target size increases from the isolated atom via molecules and clusters to , ultimately , bulk matter , the shift of the net binding energy contains information about the screening capability of the other electrons in the system and allows to investigate , e.g. , metal - to - nonmetal transitions .
in fact , x - ray photoelectron spectroscopy ( xps ) is well - established and work has been extensively devoted to measure core - level binding energy shifts in solids and at surfaces @xcite .
however , the systematic study of core - level shifts as a function of the target size became possible only recently because of the necessity of having size - selected clusters and tunable , powerful x - ray sources such as free - electron lasers available . from the theoretical point of view , the calculation of core - level binding energies in many - electron systems is non - trivial both conceptually and computationally @xcite .
the study of atoms , clusters or bulk , in which suddenly a core - electron is removed by single - photon ionization , has a long history , starting probably with slater s transition state theory @xcite and various kinds of self - consistent field methods ( @xmath0scfs ) @xcite .
@xmath0scf - approaches are based on energy balances between electronic configurations , and it is not always clear whether relaxed or unrelaxed configurations should be taken . here , `` relaxed '' means that the other electrons are allowed to assume the new , energetically most favorable configuration albeit without _ filling _ the core - hole .
`` unrelaxed '' means that the other electrons are kept frozen .
it is intuitively clear that the unrelaxed energy difference between final and initial configuration should be considered if the photoelectron escapes too rapidly to notice the relaxation dynamics of the other electrons @xcite .
methodologically , self - consistent field methods like hartree - fock ( hf ) or density functional theory ( dft ) ( see , e.g. , @xcite ) are usually employed for the calculation of energy differences between final and initial configurations in @xmath0scf approaches . as both methods introduce orbitals and orbital energies ,
the interesting question arises how these orbital energies are related to the ( relaxed or unrelaxed ) binding energies @xcite . clearly , the energy difference depends on initial _ and _ final state while the orbital energies do not , unless fractional occupation numbers , as in slater s transition state theory , are introduced .
we use an extended na - cluster jellium model with an _ all - electron _ na atom in the center in order to have core - levels in the system at all .
this model may be viewed as a hybrid of the pure cluster jellium model @xcite and bulk - models in which atoms are immersed in homogeneous electron gas @xcite .
we employ dft to calculate relaxed and unrelaxed energy differences between the initial ground state and the final configuration with either a valence electron or a 2p - core electron removed .
the results as a function of the cluster size ( up to 58 atoms ) are compared with the prediction of the metal sphere - model .
time - dependent dft ( tddft ) ( see @xcite for a state - of - the - art account ) is used to investigate the screening _ dynamics _ after the sudden removal of a 2p electron .
we apply our model in this first study to na clusters ( instead of pb@xmath1-clusters in the experiments @xcite ) in order to keep the number of inner electrons in the embedded atom and thus the numerical effort managable .
the paper is organized as follows . the basic methodology and theoretical background
is introduced in section [ basictheory ] .
the dft results on the size - dependence of binding energies and the tddft results on the screening dynamics are presented in section [ results ] .
we conclude in section [ concl ] .
atomic units @xmath2 are used , unless noted otherwise .
we use dft in exchange - only local spin - density approximation ( xlsd ) with perdew - zunger ( pz ) self - interaction correction ( sic ) @xcite . the electronic kohn - sham ( ks ) equation reads _ i_i = ( t + v(r ) + + - \ { + } ) _ i .
[ ks1]here , @xmath3 is the energy of ks orbital @xmath4 , @xmath5 with @xmath6 the number of electrons with spin - projection @xmath7 .
@xmath8 is the kinetic energy operator , @xmath9 is the spherically symmetrical external potential introduced below . @xmath10 and @xmath11 are the hartree potential and the exchange - correlation ( xc ) potential in xlsd approximation , respectively ( see , e.g. , @xcite ) .
both are functionals of the ( spin ) density .
the term in the curly bracket is the pz - sic , i.e. , in the ks equation for orbital @xmath4 the hartree - xc potential @xmath12 evaluated with the corresponding spin density @xmath13 is subtracted from the full hartree - xc potential @xmath14 evaluated with n=_n_= n_+ n _ , n_=_i=1^n _ n_i .
it is known that the pz - sic applied to xlsd significantly improves the results as compared to uncorrected xlsd @xcite .
the central - field approximation is applied to @xmath15 in so that the single - particle orbital angular momentum quantum numbers @xmath16 and magnetic quantum numbers @xmath17 remain good quantum numbers .
although the pz - sic in general improves the uncorrected ks results quantitatively , there have been several critical issues discussed in the literature @xcite , one of them being the orbital - dependent effective hamiltonian in , leading to non - orthogonal ks orbitals . in our calculations
we therefore force the ks orbitals to be orthogonal
. an extended version of the qprop code @xcite is used to solve and for the dynamics in section [ tddftresults]its time - dependent version ( in which @xmath3 is replaced by @xmath18 and the respective adiabatically time - dependent potentials are used @xcite ) .
the external potential @xmath9 in in this work is v(r)= - + \ { ccc - ( 3- ) & & r < r + - & & rr . . [ potential ] the first term
@xmath19 is the pure coulomb potential of the central atom for which all electrons are taken into account .
this atom provides the ionizable core - shell .
the other @xmath20 atoms in the cluster are treated in jellium approximation .
the jellium potential is that of a homogeneously charged sphere of radius r = a^1/3,[radius ] mimicking the net ionic background of the cluster , as `` seen '' by the cluster valence electrons .
a wigner - seitz radius for na of @xmath21 was used . for @xmath22
the potential clearly reduces to the single - atom case . cluster .
( a ) total density ( strong increase in center because of explicitly treated all - electron na atom there ) .
( b ) ks potentials for the various pz - ks orbitals ( gray to black ) , @xmath23 dotted ; horizontal lines indicate ks levels ( central atomic ones solid , cluster levels dashed ) .
[ na58resdensandpot],scaledwidth=50.0% ] figure [ na58resdensandpot ] shows the results for electron density , the pz - ks potentials , and the ks levels for the neutral closed - shell na@xmath24 cluster . because all electrons in the central atom are considered there is a peak in the electron density in the center
. what would be the 3s level in the isolated na atom is now the 1s jellium cluster level .
the higher the jellium level the closer the respective ks eigenvalue is to the one in the corresponding pure valence electron jellium model .
note that the pz - sic yields the correct @xmath23-behavior of the ks potentials , unlike pure xlsd .
the experimental observable is the kinetic energy @xmath25 of photoelectrons which are emitted upon the irradiation of clusters with x - rays of photon energy @xmath26 .
energy conservation requires e_0 + = e _ + _ , e_0 < e _
< 0 so that the binding energy can be determined as e = e _ - e_0 = - _ , e > 0 . [ experimentalbinding ] @xmath27 is the initial energy of the target . in what follows
we assume that the target is initially in its ground state .
however , it is far from obvious what @xmath28 is .
as long as auger decay occurs on time - scales longer than the photoelectron escape times the final target configuration of interest to us is a cluster in which one atom has a core - hole .
we assume that the photon energy is such that most likely one of the 2p electrons is removed .
if the escape time is longer than the screening time the photoelectron may pick up the relaxation energy = e_- e _ > 0 , [ relaxenerg ] and the net binding energy is e _
= e _ - e_0 = e _ - e_0 - . [ bindingenergy1]on the other hand , if the photoelectron escapes before the system is relaxed , the binding energy reads e_= e _ - e_0 = e _ - e_0 + >
e_. according to koopmans theorem the unrelaxed energy difference equals the hartree - fock orbital energy , @xmath29 .
the fact that in the ultra - high photon energy - limit the centroid of the photoelectron spectrum is located at @xmath30 is known as the manne - berg theorem @xcite or the lundqvist sum rule @xcite .
however , koopmans theorem does not hold in dft . instead , the `` non - koopmans energy '' tends to cancel a part of the relaxation energy @xcite so that ks orbital energies @xmath31 are shifted towards @xmath32 . whether @xmath33 is above or below @xmath32
depends on the xc energy functional used . introducing fractional occupations @xmath34 ,
janak s theorem establishes a connection between the _ relaxed _ binding energy and the _ relaxed _ ks orbital energies @xmath35 ( see , e.g. , @xcite ) . using the extended cluster jellium model introduced in section [ clustermodel ] we determined the binding energy e_i = e_i - e_0 of a core - shell electron @xmath36 by switching the fractional occupation number of the respective ks orbital @xmath37 from 1 to 0 . without relaxing the final configuration with the core - hole we obtain e_i , = e_i , - e_0 , and with relaxation e_i , = e_i , - e_0 < e_i , . the relaxation energy _
i = e_i , - e_i,= e_i , - e_i , , in general , depends on the orbital from which the electron is removed .
clearly , in the case of the removal of a valence electron the relaxation energy is expected to be very small because mainly the electrons _ above _ the level from which the electron is removed contribute to screening . on the other hand , for the screening by the cluster valence electrons only the net positive , almost point - like hole of charge
@xmath38 counts , and it should not matter much from _ which _ inner - atomic shell a core - electron is removed .
possible differences in the relaxation energy are expected to be due to other core - shell electrons .
we call this _ inner - atomic screening _ , in contrast to _ inter - atomic screening _ by cluster valence electrons .
a widely used approach in the dft community is to use the ks orbital energies as a zeroth - order approximation for the binding energy of electrons .
however , the initial state ks orbital energy may be above or below @xmath39 , depending on the xc potential used @xcite . energy differences @xmath40 are less sensitive and , even more so , binding energy _ shifts _ , i.e. , the difference of energy differences .
the energy required to remove one electron from a metal sphere of radius @xmath41 which , initially , has @xmath42 negative elementary _ excess _
charges @xmath43 reads e _ , = e_-1 - e_= w + .
[ metsph]here , @xmath44 is the work function . in the bulk
limit @xmath45 one has @xmath46 , independent of @xmath42 .
equation can be simply derived by calculating the initial and final energies . as the interior of the metal sphere is , by definition , field - free one has e_= _
r^r 4r^2 ^2(r)= _
r^r 4r^2 ( ) ^2 = ( with @xmath47 the electric field ) so that e_-1 - e_= = .
the work @xmath44 needed to remove an electron from the bulk is added `` by hand , '' leading to .
improved metal sphere models yield also @xmath44 and helped to resolve the `` image charge paradox '' @xcite .
figure [ workf ] summarizes schematically the prediction of the metal sphere model .
starting from the bulk value @xmath44 for the binding energy for @xmath48 eq . predicts a linear behavior in @xmath49 with a negative slope for initially negatively charged clusters , a positive slope for initially neutral clusters , and increasingly steep positive slopes for more and more positively charged clusters . while the cluster radius is
apart from a small spill - out well - defined for big clusters this is not the case for small clusters . in the following
we will use down to the single , isolated atom ( where @xmath50 ) and hence plot the respective binding energy vs @xmath51 .
even with the uncertainty in the definition of the single atom or ion radius kept in mind , the binding energy for the single atom or ion in general does not lie on the metal sphere line predicted by . hence , there is a transition region ( hatched area in fig .
[ workf ] ) in which the binding energy is expected to deviate from the metal sphere line towards the single atom or ion result ( filled colored circles in fig .
[ workf ] ) .
( bulk ) with the work function @xmath44 the slope is negative for electron removal from anions ( dark red , @xmath52 ) , positive for neutrals ( blue ) , and steeper positive for cations ( green , @xmath53 ) .
possible single , isolated atom or ion results are indicated by filled circles , which , in general do not lie on the metal sphere lines .
the transition from the perfect metal behavior to the single , isolated atom or ion result occurs in the hatched region .
[ workf],scaledwidth=50.0% ]
table [ tab1 ] summarizes the pz - sic xlsd results for the smallest and largest system we consider in this work , namely the single , isolated na atom ( @xmath22 ) , and the na@xmath24 cluster ( @xmath54 ) .
one sees that in the single atom - case the removal of the 3s@xmath55 valence electron and subsequent relaxation gives @xmath56 , in good agreement with the experimental ionization potential @xmath57 , showing that the pz - sic greatly improves the result . ] . the same procedure for a 2p@xmath58 yields a binding energy @xmath59 , i.e. , @xmath60ev , in very good agreement with the experimental result for the @xmath61p@xmath62 xps peak @xcite .
the removal of a 2p@xmath55-electron gives the slightly different and higher binding energy @xmath63 .
looking at the ks orbital energies one recognizes that @xmath64 is slightly below @xmath56 . hence , koopmans theorem is almost fulfilled for the valence electron , which is not surprising because relaxation effects are expected to be small if a loosely bound outer electron is removed .
the difference between ks orbital energy and relaxed energy difference is larger for the 2p - electron removal : @xmath65 .
obviously , the relaxation energy does _ not _ fully cancel the `` non - koopmans energy '' @xcite . for na@xmath24
we see in the right table [ tab1 ] that the binding energy of the valence electron @xmath66 is reduced compared to the atomic case , as expected from the metal sphere model for an initially neutral cluster .
the binding energy @xmath67 also shifts to lower values with increasing cluster size .
however , the absolute value of the ks orbital energy increased with respect to the atomic value . hence , naively estimating the photoelectron peak position in an xps spectrum to be @xmath68 is not only inaccurate but even may predict peak shifts into the wrong direction
. figure [ sizedep1]b shows @xmath69 , @xmath70 , and @xmath71 vs @xmath49 .
we observe the well - known zig - zag behavior of the binding energy because of shell - closures @xcite before the metal - sphere result is approached for increasing cluster size .
for the removal of a valence electron the ks orbital energies @xmath71 are very close to @xmath69 , and both develop towards the metal sphere behavior as @xmath49 decreases .
the relaxation energy @xmath72 is very small . ) , unrelaxed energy - differences ( blue @xmath73 ) , and absolute value of ks orbital energies ( green @xmath74 ) vs inverse cluster radius @xmath75 for the removal of an atomic 2p electron ( a ) and a cluster valence electron ( b ) .
the small annotations close to the symbols in ( b ) indicate the number of atoms @xmath76 in the cluster .
red crosses indicate experimental data points , red lines [ several dashed in ( a ) and one solid in ( b ) ] the metal sphere - model prediction for @xmath77 . for the valence case ( b )
there is almost no difference between relaxed binding energy , unrelaxed binding energy , and the ks orbital energy .
however , for the removal of a 2p core - shell electron the differences are large and increase with increasing cluster size .
the difference between unrelaxed and relaxed binding energy is the relaxation energy , as indicated by the curved arrow in ( a ) .
the purple + symbols in ( a ) show the minimum absolute value of the 2p ks orbital energy for the relaxed configuration as the 2p ground state fractional occupation number is varied between 1 and 0 ( see text for discussion ) .
[ sizedep1],scaledwidth=60.0% ] the typical valence electron result for binding energies in clusters shown in fig .
[ sizedep1]b is known for decades @xcite .
the qualitatively new aspect coming into play when core - electrons are removed is the reorientation of the @xmath78 electrons after the emission of the core - shell photoelectron .
there are ( at least ) three options : * the photoelectron is slow enough to gain the relaxation energy which is set free due to the screening of the core - hole by the other electrons but fast enough to escape before the core - hole itself decays ; * the photoelectron escapes so rapidly that it can not pick up the relaxation energy ; * the photoelectron is slow enough to gain energy not only from relaxation but also from auger - like core - hole decay .
signatures in xps spectra that would support option ( iii ) were not observed in experiments @xcite .
instead , slow `` secondary '' electrons are measured , which are likely generated via auger decay of the core - hole _
after _ the photoelectron is already gone .
indeed the estimated photoelectron escape times are sub - femtosecond for the cluster sizes and photon energies considered in this work and thus faster than auger decay @xcite . option ( i ) is supported by the experiment because the experimental binding energies for sufficiently big clusters follow again the metal sphere prediction with only the work function shifted to the respective bulk binding energy .
option ( ii ) would manifest itself as an _ increased _ binding energy of the photoelectron as compared to the situation where all the relaxation energy goes to the photoelectron .
instead , the experimentally measured binding energies in @xcite drop _ below _ the metal - sphere result , which ultimately must be so because the single ion value ( pb@xmath1 in @xcite ) lies below the metal - sphere line .
hence , one has to distinguish ( at least ) two effects here : the ability of the other electrons to screen and the ability of the photoelectron to pick up the relaxation energy due to screening .
the first one seems to be relevant in the experiments @xcite so that indeed the metal - to - nonmetal transition in pb@xmath79 around @xmath80 could be probed .
the situation for initially neutral na clusters is quantitatively different . in fact , we find
apart from the shell - oscillations little deviation from the metal - sphere behavior down to the single - atom limit in fig .
[ sizedep1]b , in agreement with experiment @xcite .
figure [ sizedep1]a shows the pz - sic xlsd result for the removal of a 2p core - shell electron from the central , all - electron na atom .
again , the relaxed energy difference follows approximately the trend predicted by the metal sphere model , although with substantial fluctuations on the ev - level ( @xmath81ev @xmath82a.u . ) .
these fluctuations are also correlated with the filling of electronic shells but in a less obvious way as compared to the valence case where maxima in the ( absolute value of the ) binding energy occur at half - filled and full shells .
it would be desirable to move towards bigger cluster in order to see whether our model really approaches the metal - sphere result for the screened 2p core - hole .
however , it is numerically very demanding to proceed significantly towards the bulk limit .
note that @xmath83 already amounts to @xmath84 .
figure [ sizedep1]a clearly shows the importance of the relaxation energy @xmath85 ( i.e. , the difference between blue triangles and black squares ) . as expected , the relaxation energy increases with the cluster size . for the biggest cluster treated in this work , na@xmath24
, we have @xmath86 , i.e. , @xmath87ev . the ks 2p - orbital energies @xmath88 of the ground state ( green diamonds ) are _ between _ the relaxed and the unrelaxed energies , as it is expected from pz - sic xlsd @xcite .
we checked how the ks orbital energy shifts as the fractional occupation number @xmath89 is decreased from 1 to 0 in the always _ relaxed _ system .
we observed that @xmath90 decreases below @xmath91 ( purple + symbols in fig .
[ sizedep1]a ) and then , for @xmath92 proceeds towards the value for the core - shell - ionized system .
if the unknown exact xc potential could be employed the orbital energy would _
jump _ discontinuously as @xmath89 changes from @xmath93 to @xmath94 .
this is the so - called `` derivative discontinuity '' @xcite .
the pz - sic xlsd mimics this behavior but smoothes - out the discontinuity .
we see that the metal sphere result in fig .
[ sizedep1]a lies between the relaxed binding energy and the minimum ks 2p - orbital energies @xmath95 .
fractional occupation numbers are related to slater s transition state theory where the idea is to consider a ( relaxed ) configuration `` half way '' between initial and final state .
indeed , @xmath96 would predict a photoelectron peak in xps spectra close to what is expected from the metal - sphere model whereas @xmath97 would be completely off and apart from fluctuations erroneously _ increasing _ with increasing cluster size . concerning the metal - sphere result for the 2p core - level binding energy
, it should be noted that the 2p bulk limit is not precisely known .
moreover , the experimental peaks are typically splitted over @xmath98ev because of fine - structure .
we estimated the 2p bulk value from extrapolating the binding energy obtained from the large - cluster xps spectrum in @xcite .
as a consequence , the core - level metal sphere result in fig .
[ sizedep1]a may have an uncertainty of @xmath99ev @xmath100atomic units ( red dashed area ) .
the experimental results for the core - level binding energies as a function of the cluster size @xcite follow the metal - sphere prediction down to a certain minimum cluster size and then deviate , ultimately approaching the single atom or ion result , which in the experiment with pb@xmath1 clusters was _ below _ the metal sphere binding energy for the 5d and 4f core - shells ( unlike for the 2p shell in na above ) .
a too rapid escape of the photoelectron would lead to an increased binding energy .
it thus seems that the relaxation energy that is set free due to screening goes indeed to the photoelectron and that the deviation from the metal sphere - model is an element - specific electronic structure - effect .
in fact , it is expected that from a certain minimal cluster size on the spherical jellium approximation breaks down .
however , our extended jellium model has the exact single atom - limit , so it may deviate at a too small cluster size from the metal sphere line but it ultimately will . the only exception arises when the single atom or ion binding energy itself lies perhaps by chance on the metal sphere line . indeed ,
for sodium this is almost the case , as is seen in fig .
[ sizedep1 ] . in order to understand why the relaxation energy that is set free due to screening goes to the photoelectron we determined the relevant time scales . to that end
we removed at time @xmath101 instantaneously a 2p@xmath58 electron from the central , all - electron na atom in the na@xmath24-cluster , and propagated the ks orbitals thereafter .
figure [ diffdens ] shows the difference in the radial spin densities 4r^2[n^(r , t ) - n^_,(r)]as a function of time @xmath102 and radial position @xmath103 .
the spin - down density @xmath104 shows a very similar behavior and is therefore not shown . at small @xmath105 mainly the charge density of the inner electrons ( up to the 2p - shell from which the photoelectron has been removed ) screens on a very fast time - scale ( @xmath106a.u .
@xmath107as ) and continues to oscillate .
this is simple to understand : after the sudden creation of the core - hole , the ks orbitals experience a new potential in which they are not eigenstates anymore .
if the potential was stationary one would expand the ks orbitals in the new eigenstates , and the oscillation periods would be given by the inverse eigenenergy - separations . in tddft
the situation is more involved because the potential is nonlinear , i.e. , it depends on the ks orbital densities .
moreover , we can not employ linear response theory here because the change in the fractional occupation @xmath108 number is not small but unity . nevertheless , the energy width spanned by the inner - electronic levels is such that it yields the screening time - scale visible in fig . [ diffdens ] . as can be inferred from the density oscillations for bigger @xmath103 , the valence electrons screen on a time - scale of @xmath109a.u .
@xmath110fs , i.e. still sub - fs , in agreement with ref .
the decay of the density oscillations seen in fig .
[ diffdens ] is probably underestimated because spontaneous emission , electron ion collisions , energy transfer to ionic degrees of freedom , and perhaps other possible dissipation channels are absent in our tddft approach .
the time - dependent calculations clearly reveal that the screening dynamics , and thus the relaxation , are fast enough to provide the escaping photoelectron with the relaxation energy . for instance , the kinetic energy of a 2p electron in a na@xmath24-cluster that is removed by a @xmath111ev photon is @xmath112ev , and it takes @xmath113fs to escape from the cluster .
hence , it certainly has enough time to pick up the relaxation energy from its own shell ( and the shells below ) , and even the time - scale for the valence electron screening is comparable .
$ ] in na@xmath24 vs time @xmath102 and radius @xmath103 ( both logarithmically ) after the sudden removal of a 2p@xmath58 electron at time @xmath101 .
( b ) line - outs for @xmath114 ( black ) , @xmath115 ( orange ) , @xmath116 ( blue ) , showing the different time - scales involved , i.e. , `` inner - atomic '' and `` inter - atomic '' , respectively , as indicated in ( a ) .
[ diffdens],scaledwidth=70.0% ]
a jellium model with a central all - electron atom has been introduced in order to model recent experiments on core - shell ionization of metal clusters as a function of the cluster size .
density functional theory in exchange - only local spin - density approximation with a perdew - zunger self - interaction correction has been employed to calculate the self - consistent electron configurations of the model applied to sodium clusters .
the _ relaxed _ binding energy follows the trend predicted by the metal sphere model and confirmed experimentally . however , shell - structure - related fluctuations more pronounced than those for valence electron - removal are observed .
both the _ unrelaxed _ binding energy and the initial , ground state kohn - sham orbital energy are completely off , demonstrating a substantial relaxation energy , the importance of screening , and the irrelevance of initial - state core - shell kohn - sham eigenvalues .
predicting a photoelectron peak by simply adding the photon energy to the initial kohn - sham core - level energy is , in general , doomed to fail . instead , for the exchange - correlation potential used in this work kohn - sham eigenvalues evaluated for fractional occupation numbers ( in the spirit of slater s transition state theory ) are closer to the relaxed binding energy and metal - sphere result .
time - dependent density functional theory has been applied in order to investigate the screening
_ dynamics _ after the sudden removal of a na 2p core - electron . _ inner - atomic
_ screening due to electrons in the same and other core - shells of na occurs within a few tens of attoseconds .
_ inter - atomic _ screening by the cluster valence electrons occurs rather on the hundreds - of - attoseconds time scale .
the inner - atomic time scale is , of course , core - shell dependent and may be even faster for deeper - lying shells and heavier elements while the screening time scale in metal clusters due to valence electrons is less sensitive and depends mostly on the number of valence electrons per atom . whether the photoelectron `` has enough time '' to pick up the relaxation energy depends on the ratio of escape time to relaxation time .
the higher the photon energy and the smaller the cluster the higher should be the net binding energy .
however , the deviation from the metal - sphere model observed so far experimentally is most likely dominated by electronic - structure effects . in order to verify this assertion future studies will go beyond the jellium model
fruitful discussions with k - h meiwes - broer are gratefully acknowledged .
this work has been supported by the collaborative research center sfb 652 at the university of rostock .
10 egelhoff w f 1987 core - level binding energy shifts at surfaces and in solids _ surface science reports _ * 6 * 253 puska m j and nieminen r m 1982 density - functional calculations of auger and x - ray photoemission shifts for metallic elements _ physica scripta _ * 25 * 708 jones r o and gunnarsson o 1989 the density functional formalism , its applications and prospects _ rev .
phys . _ * 61 * 689 perdew j p and zunger a 1981 self - interaction correction to density - functional approximations for many - electron systems _ phys .
b _ * 23 * 5048 mundt m , kmmel s , huber b and moseler m 2006 photoelectron spectra of sodium clusters : the problem of interpreting kohn - sham eigenvalues 2006 _ phys .
b _ * 73 * 205407 seidl m and perdew j p 1994 size - dependent ionization energy of a metallic cluster : resolution of the classical image - potential paradox _ phys .
b _ * 50 * 5744 peredkov s _ et al . _ 2007 free nanoscale sodium clusters studied by core - level photoelectron spectroscopy _ phys .
b _ * 75 * 235407 engel e and perdew j p 1991 theory of metallic clusters : asymptotic size dependence of electronic properties _ phys .
b _ * 43 * 1331 rander t , schulz j , huttula m , mkinen a , tchaplyguine m , svensson s , hrwall g , bjrneholm o , aksela s and aksela h 2007 core - level electron spectroscopy on the sodium dimer na 2p level _ phys .
rev . a _
* 75 * 032510 bowlan j , liang a and de heer w a 2011 how metallic are small sodium clusters ? _ phys .
lett . _ * 106 * 043401 perdew j p , parr r g levy m and balduz j l 1982 density - functional theory for fractional particle number : derivative discontinuity of the energy _ phys .
_ * 49 * 1691 | the screening of a 2p core - hole in na clusters is investigated using density functional theory applied to an extended jellium model with an all - electron atom in the center .
the study is related to recent experiments at the free electron laser at desy in which photoelectron spectra from mass - selected , core - shell ionized metal clusters have been recorded .
relaxed and unrelaxed binding energies as well as kohn - sham orbital energies are calculated in perdew - zunger self - interaction - corrected exchange - only local spin - density approximation for valence and 2p core electrons in na clusters up to 58 atoms .
the relaxed binding energies follow approximately the metal - sphere behavior . the same behavior is seen in the experiment for sufficiently big clusters , indicating perfect screening and that the relaxation energy due to screening goes to the photoelectron . instead , calculating the kinetic energy of the photoelectrons using unrelaxed binding energies or kohn - sham orbital energies yields wrong results for core - shell electrons . the screening
_ dynamics _ are investigated using time - dependent density functional theory .
it is shown that screening occurs on two time scales , a core - shell - dependent inner - atomic and an inter - atomic valence electron time scale . in the case of
na 2p ionization the remaining electrons in the 2p shell screen within tens of attoseconds while the screening due to cluster valence electrons occurs within several hundreds of attoseconds .
the screening time - scales may be compared to the photon energy and cluster size - dependent escape times of the photoelectron in order to estimate whether the photoelectron is capable of picking up the relaxation energy or whether the residual system is left in an excited state . |
the relativistic and ultrarelativistic jets arise as important components in the structure and dynamics involved in several astrophysical scenarios . among the different types of astrophysical jets , the most energetic ones are potential candidates to give rise to emission of gravitational waves .
for example , highly relativistic jets should be associated with some sources of gamma ray bursts ( grbs ) . in the last few years
, observations that indicate the presence of jets in grbs have been reported by many authors @xcite .
the expansion of the jet in the burst leads to very high lorentz factors which is at least of hundreds , but higher values such as @xmath0 are taken into account in modelling the grb sources @xcite . in the present paper , we consider a model in which the gamma lorentz factor takes the limit @xmath1 .
this assumption can be used to solve exactly the gravitational field equations .
then , we can take this analytic solution to represent the gravitational wave generated in a strong burst . to study the impact of an ultrarelativistic jet over the spacetime metric , we start from the extreme situation where the velocity of the particles in the beam is assumed to be equal to the velocity of light . the jet is then represented by a beam of null particles . for jets which start with a small opening angle @xmath2 @xcite , we assume that the width of the beam remains constant during the first stage of the jet expansion .
then , we calculate the effect of this jet over the spacetime metric in a flat background . we take advantage of the use of exact methods to provide solutions to higher - derivative gravity equations @xcite and
compute the solution to quadratic gravity field equations .
quadratic gravity is an example of the higher order theories of gravity which are generally covariant extensions of the general relativity .
these theories are constructed by the addition of terms nonlinear in the curvature to the einstein - hilbert action .
the coupling parameters of the quadratic curvature terms in the lagrangian must be determined by experiments .
observational constraints on these parameters can be traced out from sub - millimeter tests of the inverse square law @xcite or by the bending of light by a gravitational field @xcite . are there a way to obtain an observational constraint on quadratic gravity coupling parameters from gravitational radiation ?
the application of the traditional methods of ordinary gravitational wave theories , which deals with the linearized version of the gravitational theories , are until the present date , unable to clarify the role played by the quadratic curvature terms in the gravitational radiation @xcite . in a recent paper @xcite we deal with the linearized version of quadratic gravity treating the higher - curvature terms perturbatively .
we conclude that the effects of quadratic curvature terms are so small that can not be measured by the current interferometric or mass resonant detectors for @xmath3 @xcite , this difference becomes @xmath4 .
this difference can not be measured by the current gravitational wave detectors . ] . in the present paper , we obtain a solution which can be able to explain the effect of higher - curvature invariants on gravitational radiation by dealing with exact solutions to gravitational field equations .
we obtain exact shock wave solutions .
a gravitational shock wave is truly a discontinuity of the spacetime metric which propagates with the velocity of light in a given direction @xcite .
the notion of wave amplitude as a finite quantity which varies with the distance from the source can not be applied here .
the gravitational shock wave must be taken instead as a propagating singularity .
according to @xcite a modification of the gravitational wave detectors would be necessary to observe gravitational shock waves .
we do not intend to discuss the detection of gravitational shock waves in the present work .
the tidal field of a gravitational shock wave is an issue which we are dealing with and will be presented in a forthcoming paper .
the plane of the present paper is the following . in section [ sec-2 ]
we write the field equations for the shock wave metric in the frameworks of quadratic gravity and einstein s general relativity .
this is carried out by assuming that the solution can be described by a pp - wave metric @xcite . in section [ sec-3 ]
we consider a homogeneous and finite beam of null particles and obtain the gravitational shock wave generated by it .
we also compare the solutions obtained in both theories . in section [ conc ]
we summarize the principal results obtained and make some comments concerning to the application of these results in astrophysics .
we start with an impulsive pp - wave metric given by the following line element @xmath5 where @xmath6 and @xmath7 ( i=1,2 ) denote the cartesian coordinates in the plane perpendicular to the wave propagation , which we call transverse coordinates .
the metric ( [ shockmetric ] ) represents a gravitational shock wave propagating in @xmath8 direction .
this is a variation of the peres s wave metric @xcite and was used by many authors to study wave - like exact solutions to einstein gravitational equations @xcite .
the coefficient of @xmath9-distribution is the wave profile function .
this function must be determined by the field equations and will depend on the characteristics of the source and on the underlying theoretical model .
the nonvanishing christoffel symbols for the metric ( [ shockmetric ] ) are given by @xmath10 where @xmath11 denotes partial derivatives with respect to the coordinate @xmath7 , and the dot denotes the derivative with respect to the @xmath12 coordinate .
the only nonvanishing components of the riemann tensor , apart from the ones obtained by symmetry properties , are given by @xmath13 the only nonvanishing components of the ricci tensor for the metric ( [ shockmetric ] ) are @xmath14 where @xmath15 denotes the laplacian operator in the transverse space @xmath16 .
the curvature scalar @xmath17 and the elementary quadratic invariants vanish identically @xmath18 the quadratic gravity theory in four dimensions can be derived from the action @xmath19 since , in four spacetime dimensions , the gauss - bonnet invariant can be used to eliminate the @xmath20 term . in the above action
, @xmath21 stands for the presence of matter fields .
the units are such that @xmath22 .
we remark that the gravitational actions built from lagrangian densities which are arbitrary functions of @xmath17 are conformally equivalent to general relativity interacting with scalar fields .
however , this equivalence can not be traced for the theory derived from the action ( [ ac1 ] ) due to the presence of the @xmath23 invariant @xcite . requiring the action @xmath24 to be stationary leads to the following field equations @xmath25 where @xmath26 and @xmath27 are the operators associated respectively with the @xmath28 and @xmath23 invariants and are given by @xmath29 and @xmath30 where @xmath31 refers to the curved space dalambert operator .
the theory defined by the action ( [ ac1 ] ) leads to the following non - relativistic gravitational potential @xcite : @xmath32 where @xmath33 this potential reflects the three modes of the linearized theory , one with a vanishing mass which gives the newtonian force and two massive modes which create yukawa - type interactions . to obtain an acceptable newtonian limit we restrict @xmath34 and @xmath35 to real values , leading to the , so called , no - tachyon constraints @xmath36 experiments which tests the inverse square law at sub - millimeter distances indicate an upper limit to the absolute value of the @xmath37 or @xmath38 parameters of @xmath39 @xcite
. a direct observational constraint on the @xmath38 parameter can be obtained from the bending of light by the sun s gravitational field . according to a semiclassical computation of the bending of light in quadratic gravity @xcite
, the higher - curvature coupling parameter @xmath38 must satisfy the same constraint , namely @xmath40 .
this bound improves by several orders of magnitude the early accepted value for the upper limit on @xmath3 @xcite .
now , let us write the field equations for the metric given in ( [ shockmetric ] ) .
by virtue of the identities ( [ invvanish ] ) the operator @xmath26 vanishes , ensuring that the square curvature scalar does not contribute to the impulsive wave solution .
the field equations ( [ fieldeq ] ) , reduces to the following linear fourth - order two dimensional partial differential equation : @xmath41h(x^i , u)=8\pi gt_{uu},\label{feqh}\end{aligned}\ ] ] where @xmath42 . in the einstein gravity , @xmath43 ,
the equation ( [ fieldeq ] ) reduces to @xmath44 taking into account the above equation we can rewrite ( [ feqh ] ) as @xmath45h(x^i , u)=h_{{\rm e}}(x^i , u)+ah(x^i , u),\ ] ] where @xmath46 is a harmonic function of the transverse coordinates @xmath7 and , @xmath47 an arbitrary constant .
this result means that we can add to the final solution a harmonic function of the transverse coordinates .
the particular dependence of @xmath48 on @xmath7 is dictated by the symmetries of the problem and the dependence on the @xmath12 coordinate is determined by the source term . by making a decomposition of @xmath49 as @xmath50 where the index @xmath51 refers to the purely quadratic part of the solution , we obtain the following second order partial differential equation for @xmath52 : @xmath53h_{q}(x^i , u)=16\pi gt_{uu}.\label{qeqh}\ ] ] thus , the problem of solve the fourth order equation ( [ feqh ] ) was reduced to the problem of solve the second order equations ( [ einsppeq ] ) and ( [ qeqh ] ) .
in this section , we compute the solution of equations ( [ einsppeq ] ) and ( [ qeqh ] ) when the source is given by a distribution of null particles moving along the same direction .
the relevant component of the corresponding energy momentum tensor in polar coordinates is given by @xcite : @xmath54 where @xmath55 is the energy density of the distribution of null particles and @xmath56 is a dimensionless constant which varies within the range @xmath57 and comes from the equation of state @xmath58 which describes an ultrarelativistic perfect fluid @xcite .
the line element ( [ shockmetric ] ) can be written as @xmath59\delta(u)du^2+dr^2+r^2d\theta^2 , \label{linerteta}\end{aligned}\ ] ] where the functions @xmath60 and @xmath61 must satisfy the following equations : @xmath62 and @xmath63f_{\rm q}(r,\theta)=16\pi g\lambda\varrho(r,\theta).\label{fqeq}\end{aligned}\ ] ] we assume that the energy density is a constant @xmath64 within a certain radius @xmath65 and zero outside .
thus , the source represents a cylindrical beam of null particles with width @xmath66 .
this beam is a simple generalization of a single null particle .
we know that the wave profile of a gravitational shock wave generated by a single null particle in quadratic gravity is given by @xmath67 , \label{spsol}\ ] ] where @xmath68 is momentum of the particle , @xmath69 a modified bessel function and @xmath70 some integration constant @xcite .
this is a continuos function of @xmath71 which is regular at the origin an diverges logarithmically at @xmath72 . to obtain a solution which generalizes the gravitational shock wave generated by a single null particle we proceed by solving the equations ( [ feeq ] ) and ( [ fqeq ] ) choosing the integration constants in such a way that the wave profile @xmath73 fulfill the same boundary and regularity conditions that the single null particle shock wave profile ( [ spsol ] )
therefore , @xmath74 must be _ i ) _ a continuous function of @xmath71 ( even at @xmath75 , where the source have a discontinuity ) , _ ii ) _ regular at the origin , namely @xmath76 , and _
iii ) _ logarithmically divergent at @xmath72 . let us solve the equation ( [ feeq ] ) .
the cylindrical symmetry implies that @xmath60 will depend only on the @xmath71 coordinate .
it is easy to see that the solution to ( [ feeq ] ) can be written as @xmath77 where @xmath78 , @xmath79 , @xmath80 and @xmath81 are integration constants . by intergrating both sides of ( [ feeq ] ) , with @xmath82 given by ( [ fesolc ] ) , over a circle of radius @xmath83 and applying the divergence theorem to the left - hand side term we get @xmath84 .
now , integrating over a circle of radius @xmath85 and applying the divergence theorem to the left - hand side term we obtain @xmath86 .
the remaining integration constants @xmath79 and @xmath81 , are determined by the boundary and regularity conditions . as a regularity condition
we impose the continuity of @xmath82 at @xmath75 .
this condition imply that @xmath87 the constant @xmath79 , which gives the value of @xmath60 at @xmath88 , is the only arbitrary constant that will be determined by the boundary conditions .
the regularity at the origin is satisfied by given any finite value to @xmath79 , then , for simplicity and without loose of generality we choose @xmath89 .
thus , the solution of equation ( [ feeq ] ) for any @xmath90 is given by @xmath91\qquad & for$\qquad r > r_0$. } \label{fesol}\ ] ] let us turn out to the purely quadratic gravity part of the wave profile , @xmath92 .
integration of the equation ( [ fqeq ] ) leads to the following result @xcite : @xmath93 where @xmath94 and @xmath95 are modified bessel functions , @xmath96 and @xmath97 , @xmath98 , @xmath99 and @xmath100 are integration constants . to obtain a solution @xmath74 regular at the origin and logarithmically divergent at @xmath72 , we set @xmath101 and @xmath102 in ( [ fqsolc ] ) . as we want
that @xmath74 be a continuous function of @xmath71 , @xmath103 must be continuous at @xmath75 .
this condition is fulfilled by taking @xmath104 and @xmath105 thus , @xmath103 takes the following form @xmath106\qquad & for$\qquad r\leq r_0$\\ \\ -16\pi g\lambda\varrho_0r_0{b}i_1\!\left(\frac{r_0}{b}\right)\!k_0\!\left ( \frac{r}{b}\right)\qquad & for$\qquad r > r_0 $ , } \label{foutqsol}\ ] ] the solution to equation ( [ feqh ] ) with the source term ( [ tuu ] ) is given by @xmath107\delta(u ) , \label{solh}\ ] ] where @xmath73 is the wave profile and @xmath108 is a harmonic function .
the cylindrical symmetry implies that @xmath109 .
note that @xmath110 diverges at @xmath88 and @xmath111 .
if we choose @xmath112 the wave profile approaches to a constant value of @xmath113 as @xmath72 .
however , the solution will diverge at @xmath88 unless that we put @xmath114 .
but , this choice violate the previous choice , @xmath102 , and implies in the lack of continuity of @xmath103 at @xmath75 . therefore , to fulfill the boundary and regularity conditions chosen for @xmath74 we must set @xmath115 in ( [ solh ] ) .
the final result diverges logarithmically at @xmath72 and resembles the wave profile for a gravitational shock wave generated by a single null particle @xcite the figure [ figur1 ] , shows the @xmath71 dependence of the wave profiles @xmath82 ( dashed line ) and @xmath73 ( solid line ) . the non zero value of @xmath74 at @xmath88 is given by @xmath116.\ ] ] this quantity results from the ricci - squared term in the gravitational lagrangian .
it carries the information of the beam width @xmath66 and tends to zero as @xmath117 .
we also observe that the wave profile obtained for quadratic gravity agrees with the prediction of einstein s gravity at large distances from the beam axis .
a comparison between the first and second derivatives of the wave profiles obtained in the two theories shows that the higher curvature terms of quadratic gravity contributes to smooth out the @xmath71 dependence of these functions at @xmath75 .
this `` smoothing '' property of quadratic gravity is related to the fact that the field equations are of fourth order in quadratic gravity , whereas they are of second order in einstein s gravity . in figure [ figur2 ]
we plot the first derivative with respect to @xmath71 of the wave profiles in both theories .
the dashed line represents @xmath118 and the solid line represents @xmath119 ( the _ prime _ denotes the derivative with respect to @xmath71 ) . figure [ figur3 ] shows the @xmath71 dependence of the second derivatives of the wave profiles .
we note the continuity and softness of @xmath120 at the point in which @xmath121 has a discontinuity .
therefore , we can conclude that the quadratic curvature component associated to the ricci - squared term in the gravitational action smooths out the functional dependence of the derivatives of the wave profile .
furthermore , this higher - curvature term contribute to remove the discontinuity of the second derivative of the einstein s gravity wave profile at @xmath75 .
we must bear in mind that a gravitational shock wave can not be treated in the same way as a linearized gravitational wave .
the effect of a gravitational shock wave on test particles is a subject that we will treat in an incoming paper .
nevertheless , we can remark that the test particles would be be sensitive to the analytic proprieties of the wave profile function @xcite .
obviously , these properties remains the same whatever the particular values assumed for the @xmath38 parameter .
thus , even at the limit of @xmath122 the quadratic gravity could in principle be distinguished from einstein s gravity .
let us now summarize the main results obtained in our investigation and give a suggestion for application of these results in astrophysics .
the first remarkable result is that the @xmath28 quadratic invariant does not contribute to the wave solution .
the only contribution , which comes from quadratic part of the theory , comes from the ricci - squared term in the quadratic gravity action .
the effect of this term over the gravitational wave emission is to regularize the discontinuity associated to the einstein s gravity solution .
we recall that at large distances from the beam axis the predictions derived in both theories coincide .
the gravitational shock wave is very different as compared to an ordinary gravitational wave @xcite .
the geodesics and the geodesic deviations in spacetimes of gravitational shock waves have been studied in @xcite .
the authors have showed that the behavior of small geodesic deviations will be dependent on the combinations of the wave profile and theirs derivatives in the transverse space with products and powers of the dirac @xmath9 distribution and the `` kink '' function , defined by @xmath123 , where @xmath124 is the step function .
therefore , the effect of a gravitational shock wave on test particles will be dependent on these results .
we consider that the issue of observation of gravitational shock waves deserves a separated investigation and we intend to do so in another paper , now in preparation , to appear elsewhere .
an important question remains to be discussed .
can the model studied in present the work be addressed to the study of astrophysical sources ?
we know that some grbs are generated in the expansion of ultrarelativistic jets @xcite
. then a cylindrical beam of null particles can be used as a frist approximation to an ultrarelativistic jet . in this approximation
an exact shock wave solution was obtained .
we compute the wave forms in the framework of einstein s and quadratic gravity .
if a gravitational shock wave takes place in a grb scenario it may have important consequences to the physical process at the source . according to @xcite
a gravitational shock wave would be relevant to the high energy scattering of quantum particles .
therefore it can be espected that the gravitational shock wave contributes to the dynamics of a grb source and to the composition of the ejected material . furthermore ,
once a gravitational shock wave related to a grb can be detected , the results derived in the present work could contribute to the understanding of the ultrarelativistic jets features from their gravitational shock wave profiles .
999 sari r , piran t and halpern j p 1999 _ ap
* l17 * 519 ; sari r 2000 beaming and jets in gamma ray bursts , astro - ph/0002165 ; piran t 2000 _ phys .
rept . _ * 333 * , 529 - 553 ; mszros p 1999 origin of gamma ray bursters , astro - ph/9912546 .
usov v 1994 _ mont . not .
soc . _ * 267 * , 1035 ; usov v and smolsky m 1997 wide ultrarelativistic plasma beam - magnetic barrier collision and astrophysical applications , astro - ph/9704152 ; rees m j and mszros p 1992 _ mont . not .
soc . _ * 258 * , 41 .
buchdahl h a 1983 * 16 * , 1441 .
long j c , chan h w and price j c 1999 _ nucl .
_ b * 539 * , 23 - 34 , hep - ph/9805217 ; hoyle c d , schimidt u , heckel b r , adelberger e g , gundlach j h , kapner d j and swanson h e 2001 _ phys .
lett . _ * 86 * , 1418 , hep - ph/0011014 .
accioly a and blas h 2001 _ phys .
_ d * 64 * 067701 , gr - qc/0107003 .
will c m 1993 _ theory and experiment in gravitational physics _
( new york : cambridge university press ) .
de rey neto e c , araujo jcn and aguiar o d , _ class . quantum grav .
_ , submitted .
barrabs c and israel w 1991 _ phys .
_ d * 43 * n. 4 , 1129 .
barrabs c and hogan p a 2001 _ int .
_ * d 10 * , 711 - 722 .
kramer d , stephani h , herlt e and maccallum m 1980 _ exact solutions of einstein s field equations _
( london : cambridge university press ) p 233 .
peres a 1959 , _ phys .
lett . _ * 3 * n. 12 , 571 .
campanelli m , lousto c o 1996 _ phys .
_ d * 54 * , 3854 .
gottlber s , schimidt h j and starobinsky a a 1990 _ class .
quantum grav . _ * 7 * , 893 - 900 .
accioly a , azeredo a , mukai h and de rey neto e c 2000 _ prog .
phys . _ * 104 * , 103 .
stelle k s 1978 _ gen . rel . and grav .
_ * 9 * , 353 - 371 ; donoghe j f 1994 _ phys .
lett . _ * 72 * , 2996 .
neilsen d w , choptuik m w 2000 * 17 * , 733 .
kunzinger m and steinbauer r 1999 _ j. math .
_ * 40 * , 1479 ; steinbauer r 1998 _ j. math . phys . _ * 39 * , 2201 .
gradshteyn i s and ryzhik i m 1965 _ table of integrals , series and products _ , ( new york : academic press ) .
garriga j and verdaguer e 1991 _ phys .
_ d * 43 * , 391 ; t hooft g 1987 _ phys .
_ b * 198 * , 61 ; loust c o snchez 1994 scattering processes at the planck scale , gr - qc/9410041 . | in the present work we approximate an ultrarelativistic jet by a homogeneous beam of null matter with finite width . then , we study the influence of this beam over the spacetime metric in the framework of higher - derivative gravity .
we find an exact shock wave solution of the quadratic gravity field equations and compare it with the solution to einstein s gravity .
we show that the effect of higher - curvature gravity becomes negligible at large distances from the beam axis .
we also observe that only the ricci - squared term contribute to modify the einstein s gravity prediction .
furthermore , we note that this higher - curvature term contribute to regularize the discontinuities associated to the solution to einstein s general relativity . |
the experimental research devoted to the detection of gravitational waves is nowadays entering a critical phase . after having accomplished a first stage of successful operation , the largest ground - based interferometric detectors are presently undergoing a major technical upgrade towards a tenfold sensitivity improvement over the whole detection band @xcite .
hopefully , this will lead to the first signal detection within the next few years , thus triggering the beginning of gravitational - wave astronomy . in those detectors ,
laser interferometry is used to measure the relative displacement of large , massive suspended mirrors playing the role of gravitational - field probes : the test masses @xcite . in the central and most sensitive region of the detection band ,
around 200 hz , the advanced virgo and advanced ligo interferometers will be limited by the thermal noise of the mirror high - reflection coatings .
also the future einstein telescope @xcite , planned to be one order of magnitude more sensitive than advanced detectors , will be affected by this noise source as well .
the reduction of coating thermal noise is of fundamental importance . for gravitational - wave interferometers , lowering such noise would yield an increase of the detection horizon ( i.e. the radius of the accessible volume of the universe ) and hence of the expected detection rate .
more generally , lowering the coating thermal noise would be highly beneficial also for other precision experiments using high - finesse optical cavities , such as frequency standards for laser stabilization @xcite , quantum computing devices @xcite and optomechanical resonators @xcite .
thermal noise arises from fluctuations of the mirror surface under the random motion of particles in coatings and substrates @xcite .
its amplitude is determined by the amount of internal friction within the mirror materials , via the fluctuation - dissipation theorem @xcite : the higher the loss , the higher the thermal noise level . at the same time , since the coating loss is usually several orders of magnitude larger than the substrate one , the thermal noise of the thin high - reflection coating turns out to be largely dominant over that of the thick underlying substrate .
therefore , since the first works that addressed this relevant source of noise @xcite , a considerable research effort has been committed to the investigation and the reduction of the mechanical loss of optical coatings .
the high - reflection coatings of gravitational - wave interferometers are bragg reflectors composed of alternate layers of ion - beam - sputtered ( ibs ) silica ( sio@xmath0 ) and tantala ( ta@xmath0o@xmath1 ) , as low- and high - refractive - index materials , respectively .
the historical reason for this choice is the low optical absorption of tantala at the typical wavelength of operation of the detectors ( @xmath2 nm ) , as the use of this material allowed reducing the coating absorption down to about 1 ppm @xcite . from the point of view of the mechanics , however , tantala proved to be substantially more dissipative than silica @xcite , making it the dominant source of loss in the coating stack .
eventually , the coating loss was remarkably improved by applying a titania ( tio@xmath0 ) doping to tantala @xcite .
this procedure , developed by the laboratoire des matriaux avancs ( lma ) in lyon , significantly decreased the mechanical loss of the high - index material of the stack .
moreover , it proved to be beneficial to its optical performances as well @xcite . very recently , for example , we measured an average absorption of 0.14 @xmath3 0.05 ppm on the central area ( @xmath4 160 mm ) of a coating deposited on a test mass of advanced ligo . in the meantime , an optimization technique to dilute the loss contribution of the high - index material has also been developed @xcite , further decreasing the resulting coating thermal noise @xcite . in this paper
, we will present the results of mechanical characterizations of state - of - the - art coatings of titania - doped tantala and silica , carried out at the lma since 2009 .
this activity is part of a wide series of follow - up quality checks ( surface quality , absorption and scattering among others ) of the optics manufactured at the lma , especially for advanced gravitational - wave interferometers . within this frame , we measured the mechanical loss and the internal residual stress of a large series of samples with different features .
coatings of assorted thickness ( ranging from few hundreds of nanometers to several microns ) and morphology have been characterized on substrates of different geometries . because of the large number and variety of the samples , our results constitute an important database and set the basis for further research and development of coatings of high optical quality and low mechanical loss .
besides , our measurements reveal the presence of an excess loss whose origin is still not clearly understood .
a possible interpretation of this phenomenon will be addressed , and the consequences of our results will be discussed .
in order to characterize the intrinsic loss of several high - reflection optical coatings , we applied the resonance method @xcite on a large set of different mechanical resonators .
for each of these samples , we measured the ring - down time @xmath5 of several vibrational modes , before and after the deposition of a coating stack .
the calculation of the coating mechanical loss then follows straightforwardly from the measurements , provided that the the _ dilution factor _ @xmath6 the ratio of the elastic energy of the coating to the elastic energy of the coated substrate is known @xcite .
this coefficient is critical since it accounts for the fact that the energy loss of the coated resonator is diluted by the relatively lossless substrate . to compute this dilution factor
, we performed a mechanical finite - element simulation of the coated samples .
we also used these simulations to identify the vibrational modes measured in our setups . in parallel
, we used an interferometric microscope to measure the curvature induced by coating internal stress in the coated samples . by associating each value of loss to a radius of curvature
, we could study a potential correlation with mechanical loss . to determine the coating morphology
, we also carried out a series of analyses with a transmission electron microscope .
a detailed description of our samples , experimental setups , measurement techniques and simulations is reported in the following .
the results of the measurements are presented in section [ secexpres ] .
the large majority of our samples is constituted by fused - silica cantilever blades provided by heraeus and optically polished by gooch and housego and by htm , having the following dimensions : 5 mm x 42 mm , for a nominal thickness of 110 @xmath7 m .
high - purity fused silica ( suprasil 311 ) was chosen because of its very low loss at room temperature @xcite , which makes it very well suited for coating characterization .
these cantilevers are welded on a corner to a cylindrical clamping block of the same material , used to suspend the sample .
the welding has been performed by hand , with a hydrogen torch , by wisepower .
the other few substrates are fused - silica disks ( @xmath4 3 " ) of corning 7980 of a nominal thickness of 2.54 mm , micro - polished ( @xmath8 flatness , roughness @xmath9 nm rms ) on both surfaces by general optics .
the purpose of such high - quality polishing is to test the coatings on substrates with specifications as close as possible to those of substrates of gravitational - wave interferometers .
the high - reflection coatings are stacks of titania - doped tantala ( tio@xmath0ta@xmath0o@xmath1 ) and silica for operation at @xmath10 nm ( except for only two samples designed for @xmath11 = 1529 nm ) . in the doped - tantala layers ,
the atomic ratio of ti to ta is approximately 1/3 , as measured through rutherford backscattering spectrometry .
the thickness of each layer is either one - quarter of the laser wavelength ( i. e. the typical quarter - wave coating ) or optimized @xcite . in the latter case , the thickness of each doped - tantala layer
is reduced to decrease the loss of the stack and thus its expected thermal noise amplitude , but the overall optical performances are preserved . nonetheless , even in the optimized configuration , the coating stack consists in a periodic structure of doped tantala and silica pairs , or ` doublets ' , with the exception of the terminal layers ( the bottom one is of doped tantala and slightly thicker , while the top one is the thinnest of the stack and is of silica for protective purposes ) .
one example of such coating is the one developed by the lma for the end test masses of the over - coupled fabry - perot arm cavities of advanced ligo and advanced virgo @xcite : it features a measured transmissivity of @xmath12 ppm and a measured absorption of @xmath9 ppm , having at the same time nearly the minimum expected thermal noise achievable for this given high reflectivity . a small subset of substrates are also coated with anti - reflection stacks , for operation at the same @xmath13 .
in such coatings , whose transmissivity is highly sensitive to thickness errors , the number @xmath14 of alternating layers is reduced to the minimum ( usually @xmath15 ) , and the design of each layer thickness is only driven by the desired optical performances .
another subset of substrates is coated with a quarter - wave single layer ( _ mono - layer _ coating ) either of titania - doped tantala or silica of varying thickness , in order to measure the intrinsic loss of each coating material composing the stacks . finally , some substrates have been coated on both surfaces ( either with stacks or with mono - layers ) in order to change the resulting internal stress of the coated samples themselves .
the complete list of the samples used in this work can be found in the tables of section [ secexpres ] .
all the coatings presented here have been deposited at the lma through ibs with the same deposition parameters ( source settings , substrate temperature , etc . ) , into a large ( @xmath16 m@xmath17 ) custom - developed coating chamber .
this chamber has been built specifically to coat twin pairs of large mirrors ( each one having @xmath4 350 mm and being 200 mm thick ) for gravitational - wave interferometers , with very high thickness uniformity over a large surface @xcite . as part of the standard post - deposition process , all the coated samples underwent in - air annealing at @xmath18c over 10 hours .
this is necessary to decrease both the internal stress and the optical absorption of the coatings .
all the measurements of curvature and mechanical loss presented hereafter have been carried out on annealed samples .
we used a custom optical profiler developed by micromap , shown on fig .
[ figmicromap ] , to measure the roc of the annealed coated cantilever blades .
the instrument is a mirau interferometer designed specifically for the characterization of large optics , and is equipped with remotely driven micrometric translation stages on the three axes and with manually tunable micrometric screws for tilting ( roll and pitch ) the sample support plate . with standard settings ( monochromatic light at @xmath19 nm , phase - shifting interferometry ) ,
the surface of the sample probed by the laser beam is typically less than 1 mm@xmath20 .
( color online ) micromap optical profiler used for the measurement of the roc of coated cantilever blades.,scaledwidth=50.0% ] ( color online ) scheme of the technique used to measure the roc of coated cantilever blades .
see section [ secroc ] for more details.,scaledwidth=75.0% ] the measurement technique is shown on fig .
[ figroc ] .
the longitudinal axis of the cantilever blade ( roughly aligned along the abscissa of the micromap translation stage frame ) is swept with steps of a few mm , and at each position the distance to the sample ( the @xmath21 coordinate ) is adjusted with steps of 5 @xmath7 m to find the focal plane .
once the displacements @xmath22 and @xmath23 are known , the roc @xmath24 of the sample can be readily computed : @xmath25 as an estimation of the sample roc is obtained for each step along @xmath26 , an average roc with its standard deviation can be calculated .
we used two distinct setups , both based on the resonance method , to perform ring - down measurements of all the cantilever blades and disks .
the two setups only differ in the system adopted to suspend the samples . in the first apparatus , a clamping system was used to suspend the cantilever blades from the base block . in the second one , a system named gentle nodal suspension @xcite ( gens ) was used to suspend the disks from the center . both systems
are shown in fig .
[ figsuspensions ] . in the gens system ,
a sphere provides a mechanically stable support for the disc to be measured , thus minimizing its contact surface .
therefore , the system allows to avoid the clamping of samples , suppresses any potential additional source of damping and exhibits a high reproducibility of the results @xcite .
all the measurements have been performed in vacuum enclosures at @xmath27 mbar , to prevent residual - gas damping . in both setups , the vibrational modes @xmath28 of the resonator
are excited with an ac voltage ( ranging from 0.2 to 3 kvpp , depending on the sample ) applied through a comb - shaped capacitor @xcite placed a few millimeters away from the surface of the sample , in order to avoid any contact . in turn
, each mode @xmath29 is excited well above its ambient noise level , then the excitation is turned off , leaving the resonator free to ring down . the amplitude of the resonating mode is continuously read out through an optical - lever system , where a he - ne laser is reflected at the surface of the sample towards a quadrant photodiode used as a displacement sensor . a custom - developed software based on labview
is then used to : ( i ) : : acquire the signal of the photodiode ( ii ) : : perform its fast fourier transform and filter it with a narrow ( 1 hz ) band - pass filter centered around the mode frequency ( iii ) : : compute the exponential fit of the envelope @xmath30 of the free decay amplitude of the filtered signal ( iv ) : : calculate the ring - down time @xmath31 and the corresponding loss angle @xmath32 .
typically , we measured three to five modes per sample on cantilever blades , in the 10@xmath33 - 10@xmath17 hz band , and five modes on disks , sampling the 10@xmath20 - 10@xmath34 band .
as the coating loss is expected to be frequency independent @xcite , the loss angles @xmath35 of each sample are averaged to yield a measured loss @xmath36 with its standard deviation .
a jeol 2010f transmission electron microscope of the centre lyonnais de microscopie ( clym ) has been operated at 200 kv to analyze the structure of a coating stack , and in particular the interface between the layers . to this purpose
, a thin transverse section of a stack has been extracted from a test sample using a focused - ion - beam ( fib ) setup ( zeiss nvision 40 ) .
the preparation of the sample is depicted in fig .
[ figclymfib ] .
this sample has then been used for scanning and transmission electron microscopy ( sem and tem , respectively ) , energy dispersive x - ray spectrometry ( edx ) and electron energy loss spectroscopy ( eels ) .
sem images of the preparation of a thin transverse section of a coating stack through fib etching : a ) chemical - vapor deposition of a 1 @xmath7m - thick protective layer of carbon over the sample to be extracted ; b ) etching of the coating section by means of a high - energy beam of gallium ions ; c ) extraction of the section with a tungsten micro - manipulator ( tungsten coating for tip welding ) ; d ) welding of the section to a copper grating via a carbon coating ; e ) thinning of the section by means of the high - energy ion beam ; f ) front view of the thin coating section ( bright : tio@xmath0ta@xmath0o@xmath1 layers ) .
courtesy of the clym.,scaledwidth=100.0% ] the calculation of the coating loss requires the knowledge of the dilution factor , @xmath37 with @xmath38 and @xmath39 being the elastic energy stored in the coating layers and in the substrate , respectively .
preferably , we numerically computed these energies with the finite - element software ansys . alternatively , a simple analytic model is also available @xcite .
a comparison between the finite - element simulations and the analytic model is presented in section [ secexpres ] .
table [ tabsimpar ] lists the parameters used in our simulations . to model the coated samples , we used three - dimensional structural elements ( _ solid186 _ ) for the substrates and two - dimensional structural elements ( _ shell281 _ ) for the coatings .
exemplary meshed models are shown on fig .
[ figsimmodels ] .
.[tabsimpar ] parameters used in the finite - element simulations for bulk and coating materials : density @xmath40 , young s modulus @xmath41 , poisson ratio @xmath42 , shear modulus @xmath43 and refractive index @xmath44 at @xmath45 nm . [ cols="^,^,^,^,^,^",options="header " , ] in summary , by looking at the results of tables [ tab2stacksblades ] and [ tabstacksdisks ] , it is clear that the residual stress is correlated with the measured excess loss .
however , our attempts to build up a relatively simple model of the excess loss ( like @xmath46 or @xmath47 ) proved to be inconclusive because of the poor quality of the fitting , yielding very large uncertainties on the parameters @xmath48 and @xmath49 . certainly , the loss excess does not have a trivial explanation .
its unexpected complexity calls for further tests and deeper analyses which seems to exceed the boundaries of this present work .
we used the fib sample from a coating stack with @xmath50 and @xmath51 for the analyses .
the sem and tem imagery revealed a regular structure with no appreciable defect . according to the images ,
the thickness of the coating blade penetrated by the electron - beam probe is @xmath52 nm .
an edx cartography showed a sharp transition between the layers , once the background noise neglected .
some exemplary pictures are shown on fig .
[ figclymtem ] .
analyses of a coating stack : a ) sem image ( bright : tio@xmath0ta@xmath0o@xmath1 layers ) ; b ) bright - field tem image ( dark : tio@xmath0ta@xmath0o@xmath1 layers ) ; c)-e ) from left to right , edx cartography of the confinement of si , ta and ti within 5 layers .
courtesy of the clym.,scaledwidth=100.0% ] for a more detailed analysis of the interface , we used the eels data collected from eighteen consecutive line scans at the edge of two layers , as shown on fig .
[ figeelsscans ] .
the line scans have been performed with an electron - beam probe of nominal width of 1 nm , separated by 1 nm steps along the direction perpendicular to the layer thickness .
the acquired signal was integrated to yield a single spectrum per line scan .
two additional reference spectra have been acquired in the middle of a silica and of a titania - doped tantala layer , respectively . in these spectra
, we took the k edge of oxygen ( k - o ) in sio@xmath0 and in ta@xmath0o@xmath1 for quantification in the next analysis .
an additional spectrum for the ti - l@xmath53 line has been acquired separately on a rutile powder , to have a reference for the tio@xmath0 signature in the spectrum of the titania - doped tantala layer .
finally , a linear combination of the three reference spectra was fitted to each line scan to quantify the content of sio@xmath0 and ta@xmath0o@xmath1 at the interface .
all the spectra used in the analysis ( raw line scans , line scans with background noise subtracted , fitting spectra ) are shown on fig .
[ figeelsscans ] .
( color online ) normalized atomic ratios ( from fitting coefficients of k - o edges in eels spectra , in arbitrary units ) of sio@xmath0 ( blue ) and ta@xmath0o@xmath1 ( red ) , as a function of the position across an interface in a stack with @xmath50 and @xmath51.,scaledwidth=75.0% ] the final results of the eels analysis is presented on fig .
[ figeelsconcentr ] . here
, the normalized amplitude of the fitting coefficients of k - o edges in sio@xmath0 and ta@xmath0o@xmath1 are shown as a function of the position in the interface ( the redundant information from the tio@xmath0 coefficients is neglected ) .
the graph indicates clearly that there is a gradient of species concentration at the interface between layers .
looking at the spatial width of the gradient , the interface appears to be @xmath54 nm thick . once corrected from the systematic enlargement experienced by the beam in penetrating the sample , the actual estimated thickness of the interface
is reduced to @xmath55 nm at most .
if our analysis is correct , this value is rather large .
it would correspond to a few percent of each layer thickness , yielding ( for @xmath56 ) a consistent fraction of mixed material in the coating .
however , this value is compatible with a tilt of the coating section of about 8@xmath57 with respect to the impinging beam , so the accuracy of our measurements appears to be limited by the intrinsic configuration of the experimental setup .
further measurements will be needed to address the issue of the interface thickness definitively .
our findings indicate that the loss of mono - layers is independent of the coating thickness and of the stress of the sample . furthermore , the loss from the interface with the substrate is eventually negligible .
the average measured loss of silica mono - layers is @xmath58 , in agreement with the outcome of independent measurements .
the average measured loss of titania - doped tantala mono - layers is @xmath59 , which , for reasons yet to be understood , is in contrast with the result from independent experiments .
the synthesis concerning our results from stacks is far more complex .
we measured several thin cantilever blades , with a single or a double coating , and a few thicker disks with a single coating .
a clear outcome is that the measured loss is larger than expected ( provided that the expectation is correct , of course ) , though the cause of this loss excess is still not known .
double - coated cantilever blades and disks show similar results , but the loss excess is smallest when measured on disks . in the attempt to understand the origin of this phenomenon
, we looked first for the presence of an unknown dissipative mechanism at the interface between layers .
then , pushed by some new unexpected evidence from our results , we tested also the hypothesis of a stress - dependent loss .
unfortunately , we could not conclusively prove any of these speculations .
our tem analyses could not determine unequivocally the steepness of the interface between the layers , though we know now that it should be 5 nm wide at the most .
if any , the interface loss would be only a tiny fraction ( between 1 and 1% , depending on the sample ) of the observed excess loss .
on the other hand , we could not find any simple model to satisfactorily describe the dependence of the excess loss on stress .
regrettably , it seems that we can not proceed deeper in our analysis with the data presently in our possession . in this work we presented the results of several years of coating characterization carried out at lma
this database comprehends results from a large variety of titania - doped tantala and silica coatings of different thickness , number of layers and thickness ratios .
the coatings have been characterized on fused - silica substrates of different geometries on different suspension systems .
the results for mono - layer coatings appear well assessed , globally .
the measured loss of each material is constant with respect to the specificity of the samples ( thickness , residual stress ) , as one would expect for an intrinsic loss mechanism in the material . for coating stacks
, we could not unequivocally prove the existence of a lossy interface of mixed materials between the layers .
nevertheless , our data demonstrate the existence of an excess loss which was totally unexpected from a linear combination of measured mono - layer loss .
moreover , this excess loss appears to be related to the residual stress inside the samples .
some of the coatings presented here have the same design as those deposited on the input ( @xmath50 , @xmath51 ) and the output ( @xmath60 , @xmath61 ) test masses of advanced ligo and advanced virgo , so our measurements which are their most up - to - date characterization have a particular relevance for the noise budget of those detectors .
though the origin of the excess loss still remains unknown , it is possible to draw some conclusions from our analysis . from now on ,
the stress inside the coating should be considered as a relevant parameter for the characterization of mechanical loss . according to our results , the measured loss is closest to the expectation if the coating stress is not relaxed , on relatively stiff substrates .
this configuration surely does have the advantage of being closer to the actual operation condition of coatings in interferometric gravitational - wave detectors , which currently use very thick ( 20 cm ) substrates .
of course , for coating characterization , the substrate thickness should be determined as a trade - off between the need to keep the coating stressed and the requirement of minimizing the dilution factor to enhance the effect of the coating loss . finally ,
we hope that this work might help in developing a new protocol of mechanical - loss characterization and possibly trigger new studies on the unsolved issues we raised .
the authors acknowledge t. epicier and b. van de moortele of the clym for the electron microscope analyses , and are grateful to the labex lyon institute of origins ( grant no .
anr-10-labx-0066 ) of the universit de lyon for its financial support within the program `` investissements davenir '' ( grant no .
anr-11-idex-0007 ) of the french government operated by the national research agency ( anr ) .
r. weiss , electromagnetically coupled broadband gravitational antenna , massachusetts institute of technology technical report , https://dcc.ligo.org/public/0038/p720002/001/p720002-00.pdf[ligo-p720002-00-r ] ( 1972 ) .
t. kessler , c. hagemann , c. grebing , t. legero , u. sterr , f. riehle , m. j. martin , l. chen and j. ye , a sub-40-mhz - linewidth laser based on a silicon single - crystal optical cavity , nature photonics 6 , 687 ( 2012 ) .
r. flaminio , j. franc , c. michel , n. morgado , l. pinard and b. sassolas , a study of coating mechanical and optical losses in view of reducing mirror thermal noise in gravitational wave detectors , classical and quantum gravity 27 , 084030 ( 2010 ) .
j. agresti , g. castaldi , r. de salvo , v. galdi , v. pierro , and i. m. pinto , optimized multilayer dielectric mirror coatings for gravitational wave interferometers , proceedings of spie 6286 , 628608 ( 2006 ) . c. michel , n. morgado , l. pinard , b. sassolas , r. bonnand , j. degallaix , d. forest , r. flaminio and g. billingsley , realization of low - loss mirrors with sub - nanometer flatness for future gravitational wave detectors , proceedings of spie 8550 , 85501p ( 2012 ) .
e. cesarini , m. lorenzini , e. campagna , f. martelli , f. piergiovanni , f. vetrano , g. losurdo and g. cagnoli , a `` gentle '' nodal suspension for measurements of the acoustic attenuation in materials , review of scientific instruments 80 , 053904 ( 2009 ) . c. comtet et al .
, reduction of tantala mechanical losses in ta2o5/sio2 coatings for the next generation of virgo and ligo interferometric gravitational waves detectors , proceedings of the 42th rencontres de moriond gravitational waves and experimental gravity , la thuile , italy , 11 - 18 march , 2007 ( to be published ) , http://hal.in2p3.fr/in2p3-00177578/document[in2p3-00177578 ] .
m. r. abernathy , j. hough , i. w. martin , s. rowan , m. oyen , c. linn and j. e. faller , investigation of the young s modulus and thermal expansion of amorphous titania - doped tantala films , applied optics 53 , 3196 ( 2014 ) . | we present the results of mechanical characterizations of many different high - quality optical coatings made of ion - beam - sputtered titania - doped tantala and silica , developed originally for interferometric gravitational - wave detectors .
our data show that in multi - layer stacks ( like high - reflection bragg mirrors , for example ) the measured coating dissipation is systematically higher than the expectation and is correlated with the stress condition in the sample .
this has a particular relevance for the noise budget of current _ advanced _ gravitational - wave interferometers , and , more generally , for any experiment involving thermal - noise - limited optical cavities . |
the intracluster medium ( icm ) in clusters of galaxies is now understood to evolve significantly over the lifetime of a cluster due to the complex interactions induced by mergers with other clusters .
these mergers are an important part of the cluster formation process in present hierarchical models of large - scale structure .
mergers supply a substantial amount of energy to a cluster ( two @xmath5 clusters colliding at @xmath6 km s@xmath7 yield an energy @xmath8 erg ) .
thus they have been suspected of contributing to many different cluster phenomena , including the production of radio halos ( harris et al .
1980 ; burns et al .
2000 ) , extreme ultraviolet and nonthermal x - ray emission ( sarazin & lieu 1998 ; blasi & colafrancesco 1999 ; blasi 2000 ) , the bending of narrow - angle - tail and wide - angle - tail radio galaxies ( burns et al .
1994 ; gmez et al .
1997 ; bliton et al . 1998 ) , and the disruption of cooling flows ( mcglynn & fabian 1984 ; stewart et al . 1984 ) .
while some evidence for mergers can be obtained from optical data alone , it is the x - ray - emitting gas which provides the strongest evidence for mergers .
the energy gained by the icm of one cluster as it falls into the potential well of another is easily enough to raise its temperature to the observed values of @xmath9 k. simulations have shown that the resulting shocks can produce long - lived substructure in x - ray surface brightness maps ; however , they have also shown that x - ray temperature and velocity maps are likely to be much better diagnostics , partly because shock brightness contrasts are small when seen in projection ( roettiger , burns , & loken 1996 ; ricker 1998 ) .
chandra _ and _ xmm _ satellites are producing icm temperature maps of much higher spatial and spectral resolution and energy sensitivity than was previously possible ( e.g. , markevitch et al .
2000 ; vikhlinin , markevitch , & murray 2001 ; mazzotta et al .
while generally confirming that mergers produce large temperature variations in clusters , these new observations have raised new questions .
in addition to showing high - temperature ridges characteristic of shocks , many of these observations show cold cluster cores moving through low density , shock - heated intracluster gas . the intracluster gas is repeatedly stirred during a cluster s life by merger shocks ( norman & bryan 1998 ) and galaxy wakes ( stevens , acreman , & ponman 1999 ; sakelliou 2000 ) .
the reynolds number @xmath10 associated with these motions is typically of the order ( sarazin 1988 ) @xmath11 where @xmath12 is the mach number , @xmath13 is the size of the stirring agent , and @xmath14 is the mean free path for ions in the icm .
merger shocks ( @xmath15 ) typically span lengths comparable to the size of a cluster ( @xmath16 ) , so for them @xmath17 .
galaxy motions ( @xmath18 ) are transonic relative to the icm , yielding @xmath19 . these influences on
the icm may produce turbulence ; if so , it is in a range of reynolds number which is accessible to present simulations .
however , definitive observational detection of turbulence in the icm will require measurements of the local gas velocity field .
x - ray observations are just beginning to enter the regime in which it will be possible to constrain these velocities using x - ray emission lines ( roettiger & flores 2000 ) .
turbulence may contribute to a number of the proposed secondary consequences of mergers , including the formation of constant - density gas cores , the destruction of cooling flows , and _ in situ _ acceleration of cosmic - ray particles .
it can also affect the stripping of gas from galaxies as they orbit in the cluster potential . unless tangled small - scale magnetic fields limit diffusion ( chandran et al .
1999 ) , this picture is further complicated by the breakdown of the fluid description of the icm on kpc scales .
nevertheless , we can make some headway with this problem by asking whether merger - induced hydrodynamical effects on scales of tens to hundreds of kpc can feed energy to scales on which collisionless processes become important .
it may also be possible to include collisionless transport effects as a correction to fluid calculations on small scales .
offset cluster mergers are a likely source of vorticity and turbulence in the intracluster medium .
the bulk angular momentum that feeds this vorticity originates in tidal torques produced during the linear phase of structure formation ( hoyle 1949 ; peebles 1969 ; white 1984 ) and should result in mergers with nonzero impact parameters .
one example of such an off - center merger is abell 754 ( zabludoff & zaritsky 1995 ; henriksen & markevitch 1996 ; roettiger , stone , & mushotzky 1998 ; valinia et al .
1999 ) . to distinguish the effects of offset mergers from those of the many other simultaneous influences on the icm , it is necessary to study them in isolation
( see schindler [ 2000 ] for a brief review ) .
offset mergers have been studied by only a few groups .
roettiger et al .
( 1998 ) and roettiger & flores ( 2000 ) treat them primarily in the context of attempts to model specific observed clusters .
ricker ( 1998 ) performed a parameter study of offset mergers between equal - mass clusters but neglected the effects of the collisionless dark matter component thought to dominate the mass of most clusters .
takizawa ( 2000 ) considers electron - ion nonequipartition effects in a parameter study of offset mergers , but these smoothed - particle hydrodynamics calculations use too few particles to adequately characterize merger shocks and energy transfer between the gas and dark matter . also , published studies of controlled mergers have used dark matter density profiles with constant - density cores , despite the fact that simulations of hierarchical structure formation produce halos with cuspy central density profiles ( e.g. , navarro , frenk , & white 1997 [ nfw ] ; moore et al .
1998 ; jing & suto 2000 ) .
finally , none of these systematic studies of controlled mergers emphasizes the distribution of entropy or the development of turbulence . in this paper
we describe a parameter study of offset mergers that addresses these issues .
we ask the following questions .
can constant - density gas cores survive in collisions between clusters with steep central total density profiles ?
how long can the pressure peaks associated with these cores survive in off - center collisions ? how do typical offsets affect the luminosity and temperature jumps expected during mergers ?
how is entropy generated and redistributed in mergers ?
how much do merger - driven motions mix the cluster cores and contribute to nonthermal support of the gas ?
in [ sec : numerical ] we describe the simulation code and isolated merger model we use in our calculations .
our single - cluster convergence study and the merger simulations themselves are described in [ sec : simulations ] . in [ sec : discussion ] , we discuss the structure of the merger remnants , the evolution of luminosity and temperature , and the survival of distinct pressure peaks in the simulations . finally ,
[ sec : conclusions ] presents our conclusions .
we will discuss at greater length merger - induced turbulence and methods for detecting off - center collisions in a future paper .
when cosmological scaling is required , in this paper we have assumed a hubble constant @xmath20 km s@xmath7 mpc@xmath7 with @xmath21 . unless otherwise specified
, quantities are expressed using length units of @xmath22 , mass units of @xmath5 , time units of gyr , and temperature units of kev .
for our merger simulations we have used the cosmos @xmath0-body / hydrodynamics code ( ricker , dodelson , & lamb 2000 ) .
cosmos solves the euler equations for the intracluster medium using the piecewise - parabolic method ( ppm ; colella & woodward 1984 ) .
the code follows the evolution of collisionless dark matter by means of the particle - mesh method ( hockney & eastwood 1988 ) , solving the poisson equation for the total gravitational potential using a full multigrid algorithm ( brant 1977 ) .
it also incorporates radiative cooling of the gas , although the simulations described in this paper do not make use of this feature .
all three of the primary code components ( hydrodynamics , gravity , and dark matter ) make use of static , nonuniform grids , allowing us to use smaller zones in regions of interest than would be possible with a uniform - grid code .
our code is fully compressible and uses a three - dimensional cartesian grid , so shocks and convective effects are intrinsically included .
cosmos has been tested thoroughly against a suite of standard test problems , and we have already used its hydrodynamical and gravitational modules to study cluster mergers without dark matter ( ricker 1998 ) . the combination of ppm and particle - mesh is widely used in cluster dynamics and large - scale structure simulation .
ppm is to be contrasted with another widely - used hydrodynamical method , smoothed particle hydrodynamics ( sph ; gingold & monaghan 1977 ; lucy 1977 ) , a lagrangian method which uses particles as moving interpolation centers for the hydrodynamical variables .
we use ppm because it resolves shocks extremely well .
sph handles shocks poorly but compensates in part by achieving better overall spatial resolution in high - density regions .
both algorithms have been shown to give consistent results when run with the same cluster - formation initial conditions ( frenk et al .
1999 ) .
clusters formed in cosmological large - scale structure calculations do not evolve in isolation .
generally they are subject to tidal torques in the linear phase of structure growth , and during the nonlinear phase they accrete a significant number of smaller objects , from galaxies to other clusters . however , we are interested here in separating out the effects of merger - generated shocks .
because we are studying controlled mergers in isolation , it makes sense to use outflow boundary conditions for the gas and dark matter and isolated boundaries for the gravitational field . as noted in
[ sec : cluster scaling ] , this requires us to cut off the initial density profiles at a finite radius , allowing particles and gas to escape through this outer boundary as the simulation progresses . our single - cluster convergence study ( [ sec : convergence ] ) shows that this leads to a steepening of density profiles at large radius which is unrelated to the merger - induced changes in the clusters .
for this reason we confine our conclusions to the region within the initial cutoff radius , focusing particularly on the behavior of the clusters within the innermost few core radii of their centers . here
we briefly describe our implementation of isolated boundary conditions ; further details appear in ricker et al .
( 2000 ) . for hydrodynamics
we use standard zero - gradient boundary conditions wherever the mass flux through the external boundary is positive . for edge zones with inward - directed velocities , we use zero - gradient boundary conditions for all variables except the component of velocity normal to the boundary ; for this component we use a dirichlet boundary condition .
this procedure prevents the destabilizing artificial inward flows which can result when matter near the boundary begins to fall back toward the center . in the case of dark matter particles
we discard those particles which leave the grid .
to handle isolated boundaries with our poisson solver we have developed a method which first obtains a zero - boundary solution , uses this to obtain boundary values for the required correction to this solution , then solves for the correction using the given boundary values . in this paper
we approach the merger problem by simulating isolated collisions between idealized model clusters .
our model clusters are initially spherically symmetric , with specified density profiles for the gas and dark matter . for the questions we address ,
this approach has two main advantages over studying clusters formed in large - scale structure simulations .
we eliminate the complication of multiple mergers and focus on the effects of a single collision , allowing us to isolate important physical mechanisms .
we also achieve high spatial resolution , which is critical for obtaining suitably converged results that track the length scales of interest in cluster cores .
however , this approach does not consider the effects of multiple simultaneous mergers , external tidal fields , and infalling matter .
additionally , we neglect radiative cooling and choose our cluster parameters to ensure that the cooling timescale is larger than a hubble time .
we do this to avoid the computational difficulties that arise when the gas is allowed to cool very quickly ( the so - called ` cooling catastrophe ' ) .
the effects of mergers on cooling flows present some very interesting and challenging problems , which we will take up in a future paper . in the current picture ,
the formation of a cluster of galaxies is a process of steady gravitational accretion and relaxation toward equilibrium , periodically interrupted by mergers with other partially virialized systems .
these periodic interruptions destroy any pre - existing equilibrium .
after such a merger , the cluster resumes relaxing toward a new equilibrium state , continuing until it is again interrupted .
previous calculations have shown that this relaxation time @xmath23 is long compared to the likely merger interval @xmath24 ( roettiger et al .
1993 ; schindler & mller 1993 ; pearce , thomas , & couchman 1994 ; ricker 1998 ) , meaning that mergers usually involve clusters which are not fully virialized .
however , studying the case of large @xmath25 is a necessary preliminary to study of the case in which a merger perturbs a partially virialized initial state . in this paper
we are interested in understanding the physical mechanisms important in individual mergers rather than simulating a specific observed cluster .
accordingly , we begin with an initial model that represents our best current knowledge of the structure of fully virialized , noninteracting clusters .
we assume functional forms for the gas density and total density profiles , then set the dark matter profile , the gas temperature profile , and the dark matter velocity dispersion profile in such a way as to establish hydrostatic equilibrium .
the gas density profile is taken from the @xmath1-model ( cavaliere & fusco - femiano 1976 ) , @xmath26^{-3\beta/2}\ , \ ] ] where the central density @xmath27 , core radius @xmath28 , and asymptotic slope parameter @xmath1 are fitting parameters . when used with the assumption of isothermality ,
the @xmath1-model with @xmath29 produces an x - ray surface brightness profile which is consistent with those observed in many clusters without cooling flows ( jones & forman 1984 ) . in this paper
we take @xmath30 as a representative value .
the @xmath1-model is similar in shape to the nonsingular isothermal sphere profile ( binney & tremaine 1987 ) , which is a solution of the gas hydrostatic equation under the assumption of isothermality . we take guidance for the form of the dark matter density profile from models of hierarchical large - scale structure formation . using @xmath0-body simulations , nfw
have shown that , in such models , dark matter halos on scales from individual galaxies to rich clusters form with density profiles given by @xmath31^{-1}\ , \ ] ] where @xmath32 and @xmath33 are , respectively , a scaling density and radius , both of which depend upon the mass of the halo in question .
the mass dependency in these simulations arises because of the differing redshift of formation of halos with differing masses , with the sense of the dependency being that less massive halos , which form earlier , have higher central densities and are thus more concentrated .
the exact shape of this universal density profile has been called into question by moore et al .
( 1998 ) , who favor a steeper profile near the centers of halos , arguing that previous simulations have insufficiently resolved the halos . nevertheless , the clear prediction of these models is that self - similar , purely gravitational evolution leads to a steeper density profile than that predicted by the @xmath1-model .
the addition of collisional , self - gravitating gas most likely modifies this profile somewhat , though if dark matter dominates as we expect it to , the total density profile should still be steeper than that of the gas near the center . therefore , we choose a total density profile using the nfw functional form .
the dark matter density profile is given by the difference between the total profile and the gas profile : @xmath34 this can only be an approximation to the true situation , because the nfw profile has been established by @xmath0-body simulations which do not include gasdynamics .
if the nfw profile were to be maintained for the total density upon including gasdynamics , the shape of the dark matter profile would need to be different enough to exactly compensate for the @xmath1-model shape of the gas .
however , because the dark matter dominates the total density , particularly in the center of the cluster where the nfw profile continues to rise while the @xmath1-model tends to a constant value , this is most likely an acceptable approximation in the inner regions of non - cooling - flow clusters .
we choose the gas temperature and dark matter velocity dispersion profiles from the equations of hydrostatic equilibrium for the two matter components .
a more complete discussion will be given in a subsequent paper ( ricker & sarazin 2001 , in preparation ) . with the total density profile given by equation ( [ eqn : nfw model ] ) ,
the total mass @xmath35 enclosed within radius @xmath36 is given by @xmath37\ , \ ] ] where @xmath38 .
the hydrostatic equation for the gas temperature profile @xmath39 is @xmath40 = -{gm(r)\over r^2}\rho_g(r)\ , \ ] ] where @xmath41 is the average mass per particle in the gas .
it is useful to define a characteristic temperature scale @xmath42 as @xmath43 then the dimensionless temperature function @xmath44 is defined by @xmath45 if we define @xmath46 as the ratio of the total mass scale length to the gas core radius , @xmath47 then the gas density can be written as @xmath48 where we have used our assumption that @xmath30 . using equations ( [ eqn : enclosed mass ] ) through ( [ eqn
: dimensionless gas density ] ) , the dimensionless form of the hydrostatic equation becomes @xmath49 = - { ( 1+x)\ln(1+x ) - x\over x^2(1+x)(1+\eta^2x^2)}\ .\ ] ] in the case of @xmath30 which we are considering here , the solution to this equation ( with the boundary conditions discussed below ) can be written in closed form in terms of clausen s integral . however , for the purpose of initializing our simulations it is more straightforward to solve this equation numerically .
the outer boundary condition for equation ( [ eqn : dimensionless gas hydrostatic ] ) depends on the pressure at large radii in the cluster .
one class of solutions would have the pressure approach a constant value at large radii to maintain pressure equilibrium with surrounding intercluster gas .
another might have the pressure go to zero at a finite radius , which would represent the outermost extent of the hot gas in the cluster .
this condition might approximate a cluster with a strong accretion shock at its outer radius , with the gas pressure being very small outside of this radius .
however , the overdensity in clusters is large ( @xmath50 in the core ) , so the outer pressure is likely to be very much smaller than the central pressure .
thus , we adopt the boundary conditions that apply to an isolated cluster that extends to large radii , so that the pressure approaches zero at large radii . assuming the adopted form for the gas density profile ( eq . [
[ eqn : beta model ] ] ) , this implies that @xmath51 as @xmath52 . with this boundary condition
, the temperature profile requires no additional scaling parameters beyond those supplied by the assumed density profiles , and it is independent of the normalization of the gas profile ( @xmath27 ) .
we constrain the allowable values of @xmath46 by noting that convective stability requires @xmath53 while in order for the central cooling time to be larger than a hubble time ( to avoid producing a cooling flow ) , we must have @xmath54 ( ricker & sarazin 2001 ) . by experimenting with typical values of @xmath32 , @xmath27 , and @xmath33 ,
we have found @xmath55 to be a safe value given the resolutions we obtain in the simulations described in
[ sec : simulations ] .
this is comparable to the sizes of the numerical core radii seen in some large - scale structure calculations ( e.g. , nfw ) .
the resulting gas temperature profile ( scaled to the central temperature ) is shown as the dotted line in figure [ fig : single cluster plots]e ( below ) .
for @xmath56 the temperature drops slowly ( roughly as @xmath57 ) , then flattens out and eventually tends toward an asymptotic @xmath58 behavior . by the radius @xmath59 ,
the temperature has dropped to 25% of its central value .
we follow a similar procedure to determine the dark matter velocity dispersion profile .
we assume the dark matter velocity components to be isotropic and gaussian distributed with velocity dispersion @xmath60 .
the virial equation for the dark matter particles is then @xmath61 = -{gm(r)\over r^2}\rho_{\rm dm}(r)\ .\ ] ] for the dark matter the density profile we use contains @xmath32 , @xmath27 , @xmath33 , and @xmath28 ; hence no simple , natural scaling for @xmath62 is available ( as in eq .
[ [ eqn : temperature scaling ] ] below ) .
therefore we choose the velocity dispersion at the cutoff radius @xmath63 , @xmath64 , and constrain its value using the assumption of virial equilibrium ( eq .
[ [ eqn : virial sigma ] ] ) . as with the gas hydrostatic equation
, we numerically integrate the resulting dimensionless virial equation .
the result ( for @xmath65 , which we motivate in the next section ) appears as a dotted line in figure [ fig : single cluster plots]b .
the velocity dispersion rises from its central value to about twice this value at @xmath66 , then drops slowly back to its central value at @xmath63 .
the difference in behavior at small radii in comparison to the gas temperature is due to the differing behaviors of the dark matter and gas densities near the center .
the dark matter requires a ` colder ' center to maintain its cuspy density profile , while the gas requires a hotter center in order to maintain its constant - density core . by choosing the dark matter and gas density profiles , we introduce as model parameters
the density scales @xmath32 and @xmath27 , and the length scales @xmath33 and @xmath28 . because both density laws ( for typical values of @xmath1 ) yield divergent total masses for @xmath67 , we introduce a cutoff in the density profile at @xmath63 , making @xmath68 an additional parameter .
we examine the effects of this cutoff in the single - cluster test calculations discussed in [ sec : convergence ] .
the velocity dispersion profile introduces the additional ` thermal ' parameter @xmath69 , yielding a total of six parameters .
( the scaling parameter @xmath42 for the gas temperature profile is determined by @xmath32 and @xmath33 via equation [ [ eqn : temperature scaling ] ] . )
while we could examine the evolution of clusters with chosen values of the above parameters , it is more useful to select values for the total mass @xmath35 , x - ray luminosity @xmath70 , emission - weighted temperature @xmath71 , and gas fraction @xmath72 , expressing the density , thermal , and length scales in terms of these quantities .
this enables us more readily to compare our results with real clusters .
in addition , it permits us to use observed correlations between these global quantities to reduce the dimensionality of the parameter space we must study .
we begin by defining the dimensionless density and dark matter velocity dispersion profiles : @xmath73 the dimensionless gas temperature profile is defined in equation ( [ eqn : dimensionless temperature ] ) .
here we define @xmath74 if we take @xmath68 to be the virial radius @xmath75 , @xmath76 becomes the familiar halo concentration parameter .
while nfw find in their simulations that @xmath76 has a weak dependence on halo mass , we adopt the simplification of a constant value @xmath77 , consistent with their results for cluster - mass halos . together with our constraint on @xmath46 ( [ sec : temperature profiles ] ) , this reduces the number of free parameters to four .
we can now write expressions for our desired quantities @xmath35 , @xmath70 , @xmath71 , and @xmath72 : @xmath78 \\ t_x & = & { 4\pi\rho_s^2 r_s^3\over l_x}\int_0^c dx\,x^2 \tilde{\rho}_g(x , f_g,\eta)^2\lambda[t(x)]t(x)\\ f_g & = & { \int_0^c dx\,x^2\tilde{\rho}_g(x , f_g,\eta ) \over
\int_0^c dx\,x^2\tilde{\rho}(x)}\ .\end{aligned}\ ] ] here @xmath79 is the plasma emissivity function ; we use the emissivity function of the mekal model in xspec ( mewe , kaastra , & liedahl 1995 ) with half - solar abundances . with the chosen values of @xmath46 ( [ sec : temperature profiles ] ) and @xmath76 , only three of these equations are independent ; one of the four quantities on the left - hand side , say @xmath71 , depends on the others . therefore , given @xmath35 , @xmath70 , @xmath72 , and our constraints on @xmath46 and @xmath76 , we can numerically solve the three independent equations for @xmath32 , @xmath33 , and @xmath80 .
we also have @xmath69 as a remaining free parameter .
we can further reduce the number of free parameters by considering ` typical ' virialized clusters . because @xmath35 and @xmath72 are observationally the most poorly determined quantities , it makes sense to allow them to vary .
we constrain @xmath70 in terms of @xmath35 using the observed luminosity - temperature relation and the virial mass - temperature relation : @xmath81 the observed exponent of the @xmath82 relation lies in the range @xmath832.5 to @xmath833.0 ( markevitch 1998 ; arnaud & evrard 1999 ; allen & fabian 1999 ) when corrected for effects due to cooling flows . we have used the bolometric relation determined by markevitch ( 1998 ) from a sample of 35 nearby clusters . in our notation markevitch s relation corresponds to @xmath84 erg s@xmath7 kev@xmath85 and @xmath86 .
mushotzky and scharf ( 1997 ) show the cluster @xmath82 relation to be nearly independent of redshift up to @xmath87 ; hence this is likely to be a good approximation even for clusters at moderate redshifts .
the virial mass - temperature relation we use is taken from evrard , metzler , & navarro ( 1996 ) , who find in an ensemble of 58 simulated clusters that equation ( [ eqn : virial m - t ] ) with @xmath88 kev@xmath89 and @xmath90 predicts cluster masses to within a standard deviation of 15% , even for fairly irregular clusters .
finally , we constrain @xmath69 , the scaling factor for the dark matter velocity dispersion , using the assumption of virial equilibrium within @xmath63 .
we also assume that dark matter particle orbits at @xmath63 are nearly isotropic .
this yields the constraint @xmath91 in setting up the initial particle distribution for each cluster , we choose the @xmath92- , @xmath93- , and @xmath94-components of the velocity for each particle from a gaussian distribution with dispersion @xmath60 . with the assumptions discussed above
, the properties of each cluster are completely determined by the assumed mass @xmath35 and gas fraction @xmath72 .
we chose gas fractions which fell within the range of values permitted by x - ray cluster observations ( david , jones , & forman 1995 ; white & fabian 1995 ) and which yielded central cooling times @xmath95 larger than a hubble time , @xmath96 .
table [ table : cluster parameters ] presents the parameter values for the two cluster models studied in the collision runs .
any other cluster properties can be derived from these values using the scaling relations discussed above .
the temperatures of the two clusters were chosen to fall in the range of ` typical ' rich clusters seen at low redshift ( e.g. , edge et al .
1990 ; edge , stewart , & fabian 1992 ) .
we then chose masses using the virial mass - temperature relation as described in section [ sec : cluster scaling ] .
the values of the scaling radius @xmath33 and central density ( given in table [ table : cluster parameters ] in terms of the central electron number density , @xmath97 ) were constrained as described in [ sec : cluster scaling ] .
table [ table : cluster parameters ] also shows the calculated values of the sound - crossing time @xmath98 and the 210 kev x - ray luminosity @xmath99 .
the two clusters , which we label ` a ' and ` b ' , have a mass ratio of approximately 2.8 . among real nearby clusters , our simulated clusters a and b are most nearly similar to virgo and awm7 , respectively , although both of these real clusters have weak cooling flows .
ccccccccc a & 0.199 & 2.0 & 0.125 & 118 & 3.95 & 1.89 & 17.7 & 0.117 + b & 0.562 & 4.0 & 0.174 & 169 & 5.31 & 1.93 & 19.7 & 1.25 + [ table : cluster parameters ] our collision model , depicted in figure [ fig : initial conditions ] , is appropriate for two clusters that have fallen together from a large distance with nonzero angular momentum .
this model introduces three new parameters : the initial separation @xmath100 , the initial relative velocity @xmath101 , and the impact parameter @xmath102 .
because clusters of given masses do not all merge at the same time or with the same total angular momentum , we have some freedom in choosing the values of these parameters .
however , only certain ranges of values are physically well - motivated . here
we discuss the scaling of our choices with the properties of our model clusters . for computational reasons
, it makes sense to fix the separation @xmath100 at the start of the calculation , balancing the need to save computational time in following the relatively uninteresting pre - merger cluster evolution with the need to start the clusters from a virialized state . in each case
we set @xmath103 where @xmath104 and @xmath105 are the cutoff radii of the two clusters .
we assume the clusters to have fallen to this separation from a greater separation @xmath106 , at which their relative radial velocity is zero .
the value of @xmath107 corresponds to the turn - around distance for the clusters ( sarazin 2001 ) , assuming that the cluster merger ( the first core crossing ) occurs when the age of the universe is @xmath8311 gyr ( the present age for @xmath108 and @xmath21 ) .
the exact value of @xmath107 does not affect the total energetics much as long as the relative velocity of the clusters at closest approach is close to the value expected for free - fall from infinity . for the purpose of computing the initial relative velocity
, we approximate the two clusters as point masses . comparing the angular momentum and energy predicted with this approximation to the actual value determined from the simulations
shows that this estimate is good to within 1% because of the highly concentrated nature of the cluster density profiles .
we also neglect the clusters spin angular momenta . at the separation @xmath107 ,
the clusters are assumed to have zero relative radial velocity ; hence their angular momentum and energy are @xmath109 where their reduced mass is @xmath110 and @xmath111 is their relative azimuthal velocity . at the separation @xmath100 ,
the relative velocity @xmath101 is perpendicular to the direction of @xmath102 , so we can write @xmath112 conserving angular momentum and energy , we eliminate @xmath111 and find @xmath113^{-1}\ .\ ] ] we estimate the range of impact parameters @xmath102 to study by using the linear - theory result for the dimensionless spin of dark - matter halos .
the spin parameter @xmath114 is defined as ( peebles 1969 ) @xmath115 here @xmath116 is the angular momentum of the halo , @xmath117 is its total energy , and @xmath35 is its mass .
recently , sugerman , summers , and kamionkowski ( 2000 ) have performed a detailed comparison of linear - theory predictions to actual angular momenta of galaxies formed in cosmological @xmath0-body / hydro calculations .
these simulations did not include cooling or star formation , so at the upper end of the mass range they studied their results should carry over to the clusters we are simulating .
they find , in agreement with white ( 1984 ) , that linear theory overpredicts the final angular momentum of galaxies by roughly a factor of three , with a large ( @xmath118 ) dispersion in the ratio of the linear - theory prediction to the actual value .
the importance of nonlinear effects on the spin parameter is less clear , as uncertainties in their group - finding algorithm strongly affect the determination of @xmath116 , @xmath117 , and @xmath35 for each galaxy and produce a large scatter in @xmath114 .
however , given this uncertainty , the mean value of @xmath114 they obtain is about 0.05 , close to the linear - theory result .
we take the halo to be the final merger remnant .
its final total angular momentum is the sum of the spin angular momenta of the two subclusters plus the orbital angular momentum @xmath119 .
we estimate the initial spins of the subcluster by applying equation ( [ eqn : dimensionless spin ] ) to each of them , and we assume that the initial spins are correlated .
the orbital angular momentum is then the difference between the final angular momentum of the merger remnant , and the total spin angular momenta of the subclusters ( sarazin 2001 ) . substituting equations ( [ eqn : init ang mom and energy])-([eqn : initial velocity ] ) into equation ( [ eqn : dimensionless spin ] ) and solving for the impact parameter
, we find that @xmath120 here , the function @xmath121 corrects for the internal angular momenta and energy of the subclusters , and can be written as @xmath122^{3/2 } \
, , \ ] ] but it only depends on the ratio @xmath123 of the smaller to larger mass of the two subclusters .
it varies between @xmath124 , so that @xmath125 . in our merger models , @xmath126 ( @xmath127 is the larger of the two scale radii of the clusters ) , and @xmath128 .
if we assume @xmath129 and approximate @xmath125 , we find that the typical impact parameter is @xmath130 .
since this argument is statistical and a range of values are expected in mergers , we consider values of @xmath131 in our simulations .
cccccccccc c1 & aa & @xmath132 & 0 & 0.436 & 7.10 & 4.73 & 4.73 & 0.144 + c2 & aa & @xmath132 & 2 & 0.436 & 7.10 & 4.97 & 4.73 & 0.144 + c3 & aa & @xmath132 & 5 & 0.436 & 7.10 & 5.32 & 4.73 & 0.144 + c4 & ab & @xmath133 & 0 & 0.548 & 8.63 & 6.78 & 6.78 & 0.213 + c5 & ab & @xmath133 & 2 & 0.548 & 8.63 & 7.12 & 6.78 & 0.213 + c6 & ab & @xmath133 & 5 & 0.548 & 8.63 & 7.63 & 6.78 & 0.213 + + [ table : collision parameters ] table [ table : collision parameters ] presents the parameter values used in the collision runs . the simulations are labeled ` c1 ' to ` c6 ' ; runs c1c3 are equal - mass collisions using cluster a , while runs c4c6 use clusters a and b. each run is followed for approximately 15 gyr , more than the age of the universe for @xmath108 and @xmath21 . for the hydrodynamics and gravitational potential calculations
, we use a nonuniform mesh with @xmath134 zones covering a volume measuring approximately @xmath135 , where @xmath136 is the larger of @xmath104 and @xmath105 .
the minimum zone spacing is approximately equal to one - seventh of the larger of the two dark matter scale radii , or @xmath137 kpc for runs c1c3 and @xmath138 kpc for runs c4c6 . for the dark matter
, each run also uses @xmath139 particles of equal mass , distributed between the merging clusters in proportion to their total dark mass .
the results of these runs are described in
[ sec : equal - mass ] and [ sec : unequal - mass ] . in [ sec : convergence ] we first discuss the results of a convergence study involving a single cluster , which we used to set the grid and particle parameters used in the collision runs .
we carried out several calculations of the motion of a single cluster with different spatial resolutions and numbers of particle in order to evaluate numerical effects due to our boundary conditions and finite - radius cutoff .
these runs also enabled us to choose grid and particle parameters for the collision calculations .
these runs used the cluster parameters appropriate to cluster a ( see [ sec : choice of parameters ] ) together with the numerical parameters listed in table [ table : single - cluster grids ] .
six runs were performed with two different grid spacings and three different numbers of particles .
one additional run was performed at a grid spacing and particle number slightly better than the collision runs to be described later .
the computational volume measured @xmath140 , and the cluster started at the position ( 2.36,2.36,2.36)@xmath141 with an initial velocity of 0.436@xmath142 in the @xmath92-direction . in each
run the cluster was permitted to move for 5 gyr , yielding an expected final displacement of 2.18@xmath141 .
crrr s1 & @xmath143 & 65536 & 0.80 + s2 & @xmath143 & 524288 & 0.80 + s3 & @xmath143 & 1048576 & 0.80 + s4 & @xmath144 & 65536 & 0.25 + s5 & @xmath144 & 524288 & 0.25 + s6 & @xmath144 & 1048576 & 0.25 + s7 & @xmath145 & 1048576 & 0.12 + [ table : single - cluster grids ] figure [ fig : single cluster plots ] compares final average profiles of various quantities in the single - cluster tests to the initial profiles .
included are the dark matter density and velocity dispersion , gas density , enclosed gas fraction , gas temperature and specific entropy , and the gas radial mach number .
we also plot the ratio of the gas pressure gradient to the gravitational force per unit volume on the gas , which shows the extent to which the gas is in hydrostatic equilibrium .
the results show the effects of finite zone spacing and particle number on small scales and outflow boundary conditions on large scales . for gridded ( gas ) quantities ,
each profile was generated by averaging profiles interpolated along 1000 different directions from the gas centroid .
the radial sample spacing was taken to be equal to the size of the smallest zones in the calculation . for particle ( dark matter )
quantities , particles were binned in radial bins about the dark matter centroid ; then the particle count or the average value of the velocity dispersion in each bin was used .
the bin spacing used was the same as the sample spacing used for the gas profiles . with the exception of the gas fraction , the pressure - potential ratio , and the radial mach number ,
all quantities are scaled in terms of the initial values of the appropriate scaling parameters , such as the core radius or central gas density .
the dark matter density profile ( figure [ fig : single cluster plots]a ) shows little sensitivity to the number of particles or zone spacing for the ranges of these parameters considered here .
resolution effects appear in the innermost part of the cluster ; the density in the innermost radial bin tends to be some 15% below the initial value .
this demonstrates the tendency of particle - mesh force smoothing to produce an artificial constant - density core on scales smaller than twice the mesh spacing .
thus in order to observe the turnover to @xmath146 behavior at small scales characteristic of the nfw profile , the mesh spacing should be no larger than about 1/4 the scale radius @xmath33 .
for @xmath147 , the escape of particles through the outer boundary of the cluster makes the resulting final profile slightly steeper than its initial @xmath148 shape .
runs s1-s3 display a somewhat shallower profile between @xmath149 and @xmath150 than that in runs s4-s7 .
the dark matter velocity dispersion profile ( figure [ fig : single cluster plots]b ) also reflects this evaporative cooling : within @xmath151 in run s7 , @xmath152 is about 9% lower than its initial value , with roughly the same @xmath36 dependence .
outside @xmath151 the final profile drops off more steeply , whereas outside @xmath63 it appears to rise again due to the radial motion of escaping particles .
( we estimate @xmath60 by computing @xmath153 for each velocity component @xmath154 , averaging within radial shells , then sum the results for @xmath92 , @xmath93 , and @xmath94 and divide by 3 .
this procedure assumes isotropic particle velocity dispersions and does not subtract out radial expansion or contraction . ) the central dispersion in runs s1-s3 is substantially lower than in run s7 , but runs s4-s6 agree with s7 .
the lack of an external confining pressure produces expansion of the gas at the outer boundary of the cluster .
this has a major effect on the gas density profile ( figure [ fig : single cluster plots]c ) , causing it to become steeper at large radii .
the profile also becomes slightly shallower at small radii .
however , unlike the dark matter profile , once the gas core radius is resolved ( runs s4-s7 ) , the gas profile converges to a central density slightly below the initial value . between @xmath155 and
@xmath156 it agrees with the initial profile , but the effective core radius ( defined as that radius at which the density drops to one - half its central value ) at the end of the simulation is @xmath157 smaller than the initial value , so the overall shape at small radii differs slightly from a @xmath1-model . when the core is not resolved ( runs s1-s3 ) , the behavior is more like the dark matter density , with the innermost density point well below its initial value . again
, the number of particles makes little difference for the range of values considered here .
the different behaviors of the gas and dark matter are apparent from the enclosed gas fraction profile ( figure [ fig : single cluster plots]d ) .
the enclosed gas fraction @xmath158 is computed from the average gas and dark matter densities via the relation @xmath159 in all cases the gas fraction at a given radius is smaller than its initial value for radii larger than about two zones . within two zones of the center
the gas fraction flattens out to a constant value rather than dropping to zero because of the artificial core in the dark matter profile .
the final gas fraction rises more slowly than its initial profile , reaching the initial total value of 12.5% only at @xmath160 . the spreading of the gas due to the lack of external pressure confinement leads to a steepening of the temperature profile ( figure [ fig : single cluster plots]e ) and the advection of entropy out of the outer parts of the cluster ( figure [ fig : single cluster plots]f ) .
whereas the initial temperature profile drops to @xmath161 times its central value at @xmath63 , by the end of the single - cluster calculation it has dropped to about 0.2 times its central value . in runs s1-s3 the gas core
is not resolved , and the innermost temperature point is about two - thirds its initial value at that radius . for runs s4-s7
the final temperature agrees much better with its initial profile , although run s7 does not reproduce the upturn within @xmath162 present in the initial conditions .
the two sets of core - resolved runs also disagree slightly between @xmath155 and @xmath163 , with runs s4-s6 agreeing better with the initial profile .
all of the simulated profiles agree with each other beyond about five core radii .
similar results can be seen in the specific entropy profile .
note that the specific entropy is measured relative to the central value given the imposed density and temperature profiles .
the resulting imposed entropy profile passes through a minimum near @xmath164 , so at smaller radii the gas would be convectively unstable if our calculations were to use smaller zones . the radial mach number ( figure [ fig : single cluster plots]g ) and pressure support profiles ( figure [ fig : single cluster plots]h ) show that the innermost parts of the cluster approximate hydrostatic equilibrium much better as grid resolution improves . both quantities are formed from angle - averaged values ; thus the radial mach number is computed using the average radial velocity and the average temperature .
when the core is not resolved , the innermost radial velocity point is as much as 0.2 times the sound speed , and the pressure provides about 10% of the gravitational force there . in runs
s4-s7 the maximum mach number inside @xmath165 is less than 5% , and at the innermost point the pressure contributes 30% of the force due to gravity .
the pressure support ratio increases immediately to a value slightly larger than unity outside the center , indicating that the outer parts of the cluster are nearly in hydrostatic equilibrium but are still slowly expanding . in all cases
the gas outside @xmath165 is in nearly free expansion , with the radial velocity increasing to a sonic point at @xmath166 .
the material inside this sonic point is slowly decelerating and will eventually fall back onto the cluster . in summary ,
the single - cluster test calculations show that the values of particle number and zone spacing used in the collision runs are adequate to produce a control cluster which remains in nearly hydrostatic equilibrium .
the code is especially well - converged with respect to particle number for the value ( @xmath167 ) used in the collision runs , and the quantities of interest have converged down to radii of about two zones for the zone spacing used ( @xmath168 ) .
the accuracy of the converged solution is affected by the particle - mesh smoothing length on small scales and by the lack of an external confining pressure on large scales . these effects lead to an artificial core radius of 12 zones in the dark matter and a nearly free expansion of the gas and dark matter outside the initial density cutoff which steepens the density and temperature profiles near the cutoff . in this section
we describe the simulations of mergers between clusters of equal mass ( runs c1c3 ) .
figure [ fig:1:1 mass plot ] shows as functions of time the projection of the gas and dark matter densities along the @xmath94-axis , which is perpendicular to the plane of the collision in the offset runs .
we refer to the projected densities as @xmath169 and @xmath170 , respectively .
figure [ fig:1:1 xray plot ] shows the 210 kev x - ray surface brightness @xmath171 and x - ray emission - weighted temperature @xmath71 , also projected along the @xmath94-axis .
figure [ fig:1:1 lt plot ] shows the total luminosity @xmath70 and average emission - weighted temperature for the entire computational grid as functions of time .
the zero of time in each plot has been chosen to be the instant of maximum luminosity ( in each case approximately 3.5 gyr or @xmath172 after the start of the calculation ) .
the abscissa of the luminosity and temperature plots is also scaled to the sound - crossing time @xmath173 . as the outer regions of the two clusters begin to interact early in the merger , a pair of roughly planar shocks forms at an angle to the collision axis , moving in opposite directions . by 0.5 gyr
prior to maximum luminosity , these planar shocks have passed through the cluster cores and are on their way out of the interaction region .
they are readily apparent in the projected gas mass plot ( figure [ fig:1:1 mass plot ] ) and are barely visible in the surface brightness contours ( figure [ fig:1:1 xray plot ] ) . between the clusters
the temperature is higher than the pre - shock value by about a factor of three , and the material in this region is expelled from the collision axis at about 1/2 the initial relative velocity .
the material between the clusters is adiabatically compressed after the passage of the shocks . in previous gas - only merger calculations ( ricker 1998 ) , the planar shocks later developed into spiral shocks which acted to dissipate the angular momentum of the cluster cores into large - scale rotational motion . with the addition of collisionless dark matter , however , the planar shocks play less of a role in the later stages of the collision , and we see more small - scale rotation as early as 34 gyr after maximum brightness and continuing for at least 3 gyr beyond that . although we have not attempted to reproduce any specific clusters with these calculations , we note that the simulations do resemble some merging clusters with particularly simple geometries . the temperatures and x - ray surface brightnesses in the early phases of the zero - offset , equal mass mergers strongly resemble the observed _ rosat _ image and _ asca _ temperature map of the cluster surrounding cygnus - a ( markevitch et al .
the offset 1:1 cases , seen just after first core passage , resemble the structure of abell 3395 ( markevitch et al .
1998 ) , which appears to be observed from above the merger plane .
the merging clusters undergo a dramatic increase in luminosity and temperature as the cluster cores collide .
the luminosity jump ranges from a factor of 4 for the most off - center case to a factor of 10 for the head - on case .
the average temperature increases by a factor of two to three . in the projected mass and x - ray maps ,
the period of core collision corresponds to the formation of a strong , crudely ellipsoidal shock which dissipates some of the initial kinetic energy of the cores into thermal energy in the surrounding gas .
the greater increase in total luminosity and average temperature for the head - on case results from the fact that the dark matter cores , whose gas content produces the majority of the x - ray emission , pass directly through one another and the adiabatically heated gas between them . in the offset cases this hot gas is driven ahead of the cluster cores as they slip past one another , producing an s - shaped region of high temperature . in all cases , however , the dark matter cores continue to have associated peaks in the gas density even after they pass one another , a significant departure from the behavior in the gas - only collisions considered by ricker ( 1998 ) . here
the central potential is dominated by collisionless dark matter , which dissipates energy in a relatively inefficient manner compared to the gas during core passage .
thus the dark matter cores , which are able to pass through one another even in the head - on case , drag gas density peaks along with them .
the extra oscillation due to the dark matter allows the dark matter to ` pump ' the gas , keeping the system disrupted longer than would be expected if both matter components were collisional .
offsets on the order of a few hundred kpc are sometimes observed between the dark matter and gas density maxima , particularly as the cluster cores reach maximum separation after their first interaction .
the first period of enhanced luminosity and temperature lasts for approximately one sound - crossing time . after their close interaction , the cores recede from one another , causing the temperature to drop to about 80% of its initial value . by the time the cores reach their maximum separation , the luminosity decreases much more , reaching about 15% of its initial value in the most offset run and about 30% in the head - on run . as one might expect due to the relative efficiency of gas dissipation in the three cases , the clusters require longer to reach maximum separation in run c3 ; this occurs about @xmath173 after core passage in this run , compared to only 1/2 sound - crossing time for run c1 . in each case
, a second core passage occurs between 1.5 and 2.5 sound - crossing times after the first , producing smaller brightness and temperature jumps .
following this second passage , the total luminosity drops to between 30 and 70 percent of its initial value , while the average temperature drops to 1.4 times its initial value .
the cores continue to oscillate , producing minor variations in the luminosity and temperature , but the values of these quantities change very little for the remainder of the calculation . during this time
several weaker sets of shocks are generated by core passages , as is apparent at 3.5 gyr after the first core passage in the projected mass and x - ray plots ( figures [ fig:1:1 mass plot ] and [ fig:1:1 xray plot ] ) .
we discuss the evolution of luminosity and temperature in more detail in
[ sec : luminosity - temperature ] .
after the cores have coalesced , angular momentum continues to be redistributed in the gas by turbulent eddies which are fed by material falling back in the wake of the ellipsoidal shock .
the remnant requires several billion years to reach equilibrium , and indeed has not quite reached equilibrium by @xmath174 , the end of the calculation ( and more than a hubble time for @xmath21 ) .
we consider the properties of the merger remnant at @xmath174 in
[ sec : merger remnants ] . in this section
we describe the simulations of mergers between clusters with a mass ratio of approximately 1:3 ( runs c4c6 ) .
figure [ fig:1:3 mass plot ] shows mass maps projected along the @xmath94-axis ( @xmath169 and @xmath170 ) for these runs at several different times .
figure [ fig:1:3 xray plot ] shows the 210 kev x - ray surface brightness @xmath171 and x - ray emission - weighted temperature @xmath71 , also projected along the @xmath94-axis .
figure [ fig:1:3 lt plot ] shows the total luminosity @xmath70 and average emission - weighted temperature as functions of time . in these figures
the more massive cluster initially approaches from the right . as for the 1:1 merger runs , the zero of time in each plot
has been chosen to be the instant of maximum luminosity .
the abscissa of the luminosity and temperature plots is also scaled to the sound - crossing time @xmath173 of the more massive cluster .
numerically , the values of the time of maximum luminosity and the sound - crossing time are similar to those obtained in the 1:1 collisions ( @xmath175 and 1.9 gyr , respectively ) . in the 1:3
runs the initial pair of shocks is not planar ; instead , the shock sweeping through the less massive cluster curves around its core , and the adiabatically compressed region between the clusters takes on an arc shape . for nonzero impact parameters , the shock being driven into the more massive cluster is stronger on the side from which the less massive cluster approaches .
however , at 0.5 gyr prior to maximum luminosity the right - hand shock has not yet penetrated the more massive cluster core , as it has by this point in the equal - mass collisions .
peak luminosity and temperature occur as the core of the less massive cluster passes through ( in the head - on case ) or swings past ( in the offset cases ) that of the more massive cluster . as in the equal - mass runs , in the offset cases here
an s - shaped region of hot gas is apparent in the region between the cluster cores .
the heating of the gas is more severe in the region in front of the less massive cluster .
also , the less massive cluster undergoes extreme distortions at this time .
gas velocities of up to 2000 km s@xmath7 are present behind it , and as it plows into the hotter , denser icm of the larger cluster , it is flattened perpendicular to its direction of motion .
the core of the larger cluster is less affected by the interaction , as expected . shortly after first core passage
, the offset case c5 resembles abell 754 , consistent with the results of roettiger et al .
the offset cases also resemble abell 2256 just at core passage if one uses the _
asca _ or _
bepposax _ temperature maps ( markevitch 1996 ; molendi , de grandi , & fusco - femiano 2000 ) and assumes the merger axis to be perpendicular to the plane of the sky .
although the smaller cluster undergoes severe distortion during the core passage , the merger s effect on the total luminosity and average emission - weighted temperature of the system is noticeably less than in the 1:1 collisions
. the luminosity increases by a factor of 24 , and the temperature by a factor of 1.42.2 , for about 1/2 sound - crossing time during the initial core passage . while maximum separation following core passage
is reached at about the same time as in the 1:1 cases , the drop in luminosity and temperature is not as great .
this period is again followed by a weaker secondary peak in luminosity and temperature as the cores fall back together . here the most offset case ( run c6 ) spends more time at maximum separation , causing the spread in secondary peak times to be 50% greater for the 1:3 runs than in the 1:1 cases .
the second core passage for this run takes place almost 7 gyr after the first , and in all cases this second interaction takes place on a nearly radial trajectory due to the dissipation caused by the first interaction .
( the last two panels of figures [ fig:1:3 mass plot ] and [ fig:1:3 xray plot ] for this run do not show the less massive cluster because at these times it is at a separation which places it outside the plotted region . ) both luminosity and temperature settle down to final values of 50% and 110% , respectively , of their initial values following the second core passage .
a few oscillations are again seen as the cores settle into a common remnant .
unlike the 1:1 cases , here there is little variation in the final luminosity as impact parameter is varied .
each merger simulation was followed for approximately 15 gyr , or about six sound - crossing times after the first core passage . although this time is greater than the age of the universe for @xmath108 and @xmath21 , it is nevertheless insufficient time for the remnants of either the 1:1 or the 1:3 mergers to come to equilibrium .
this point is illustrated by figure [ fig : final entropy ] , which shows the gas specific entropy and velocity fields in the @xmath176-plane of each run at @xmath177 gyr .
although in each case the core of the remnant has lower entropy than its surroundings , the gas at the margins ( @xmath178 ) continues to be disturbed by small - scale convective motions as matter accretes onto the remnant behind the shocks that have been driven out of the core region .
the offset mergers all show evidence of bulk rotation , with velocities of @xmath179 km s@xmath7 , but in even the most offset cases this provides at most a few percent of the support needed to keep the remnant in hydrostatic equilibrium .
the 1:1 cases are closest to equilibrium , having settled into roughly spherical distributions within @xmath180 of the center of mass .
the 1:3 cases are relatively further from equilibrium , with distinct streams of material falling in from the right ( in the head - on case ) and along the top edge ( in the offset cases ) .
these streams continue to feed high - entropy gas from the smaller cluster into the remnant as the simulations come to an end .
departures from equilibrium can also be described in terms of total energetics . in describing the results of gas - only merger simulations , ricker ( 1998 ) defined a ` virial disequilibrium ' parameter as @xmath181 where @xmath182 , @xmath183 , and @xmath184 are the total kinetic , internal , and potential energies .
when the system is in virial equilibrium , @xmath185 , and when it is not , @xmath186 is equal to the second time derivative of the moment of inertia , divided by @xmath184 .
thus a useful equilibrium criterion is to require @xmath187 for an extended period ( say @xmath188 ) . in the gas - only mergers , the clusters required between 5.5 and 6 sound - crossing times to reach the level @xmath189 , after which @xmath186 was roughly constant . in figure
[ fig : virial plot ] we plot @xmath186 versus time for the six gas - plus - dark matter mergers considered in this paper .
note that , as in the earlier work , @xmath186 passes through a maximum during the first core passage , when the system is furthest from equilibrium .
@xmath186 then declines steeply but oscillates for several sound - crossing times as the dark matter continues to pump the gas . finally the system begins to tend steadily toward @xmath185 around @xmath190 after first core passage .
however , by the end of the simulations only runs c4 and c5 have satisfied the equilibrium criterion just described .
we plot angle - averaged profiles of various quantities from the merger runs at @xmath177 gyr in figures [ fig:1:1 profile plots ] and [ fig:1:3 profile plots ] .
profiles are generated from the raw simulation data as for the single - cluster test runs described in [ sec : convergence ] .
although the merger remnant is not in hydrostatic equilibrium , the average profiles still serve to characterize the effect of the collision on the structure of the clusters . as for the single - cluster runs ,
the profiles here are scaled to the initial values of @xmath32 , @xmath33 , and so forth , with the exception of the enclosed gas fraction , the gas radial mach number , and the pressure support ratio . for the 1:3 mergers the scaling parameters used are those pertaining to the more massive cluster ( b ) . the dark matter density profile ( figures [ fig:1:1 profile plots]a and [ fig:1:3 profile plots]a ) shows very little change in shape as a result of the collisions . of all of the properties of the merger remnant ,
its shape is also the least affected by variations in mass ratio and impact parameter .
the normalization ( @xmath191 since dark matter dominates the potential ) increases in each case because the merger remnant is more massive than the individual colliding clusters , but the scale radius changes very little .
the asymptotic slope of the profile does not increase as in the single - cluster test runs .
figures [ fig:1:1 profile plots]b and [ fig:1:3 profile plots]b give the profile of the dark matter velocity dispersion at the end of the simulations . at the end of the 1:1 collisions ,
the central velocity dispersions are all increased as a result of the deeper potential well of the merger remnant .
a similar but smaller effect is seen in the 1:3 mergers . at large radii ,
the velocity dispersions are all small as a result of the evaporation of faster moving particles and the outflow - only outer boundary conditions in the simulations .
a similar effect was seen in the single cluster simulations ( figure [ fig : single cluster plots]b ) .
although the dark matter density profiles are not strongly affected by the mergers , the gas density profiles ( figures [ fig:1:1 profile plots]c and [ fig:1:3 profile plots]c ) change substantially as a result of the mergers .
the central gas density in each case actually shows a substantial decrease from its initial value , dropping to 3070% of its initial value for the 1:1 runs and to 2030% of its initial value for the 1:3 runs . using the half - density radius as an estimate for the core radius
, we find that @xmath28 increases by 2025% .
the final value of @xmath27 shows a significant dependence on impact parameter . for the equal - mass collisions , the final gas density is about twice as high for head - on mergers as for the most offset mergers . for the 1:3 collisions ,
the variation is only 50% , but the sense is reversed , with head - on collisions resulting in a lower central gas density .
in all cases we see a steepening of the asymptotic slope from @xmath192 to @xmath148 , as in the single - cluster runs , because of the initial cutoff at @xmath63 combined with the outflow boundary conditions .
the flattening of the gas density profile occurs because the shocks generated by the merger add entropy to the system and increase the buoyancy of the icm in the relatively constant dark matter potential .
the resulting profile for the enclosed gas fraction ( figures [ fig:1:1 profile plots]d and [ fig:1:3 profile plots]d ) typically begins below the initial value of @xmath193 for radii comparable to the zone spacing , then increases roughly as @xmath194 and almost meets the initial profile at @xmath195 before flattening out to an asymptotic value slightly below the initial value .
the change in curvature at small radii is also seen in the single - cluster runs and is a numerical effect due to the influence of particle - mesh force smoothing on the cuspy dark matter profile . likewise , the drop from the initial profile between @xmath195 and @xmath180 is a numerical effect due to the expansion of material beyond the imposed initial density cutoff .
the drop in the asymptotic value of the gas fraction occurs because more gas than dark matter is lost through the outflow boundary @xmath196 versus @xmath197 as shocks tend to eject collisional material while leaving the collisionless dark matter alone .
the addition of an external confining pressure would most likely allow the asymptotic gas density slope to be flatter but leave the inner core unchanged .
this is apparent from the gas temperature profiles ( figures [ fig:1:1 profile plots]e and [ fig:1:3 profile plots]e ) , which show an inner region whose behavior depends on impact parameter and mass ratio and an outer region that is roughly independent of impact parameter and corresponds to the single - cluster result .
the boundary between these two regions occurs at @xmath198 in each of the runs and is @xmath199 hotter relative to @xmath200 in the 1:1 runs than in the 1:3 runs .
in all cases the final temperature profile declines more slowly at the very center than does the initial profile .
offset mergers display opposite trends with increasing @xmath102 for the two mass ratios : for 1:1 mergers , increasing impact parameter increases the final temperature over the head - on case , while for 1:3 mergers it decreases the temperature , reproducing the initial profile in the most offset 1:3 case .
this behavior occurs because of the increase in core entropy associated with mixing of shocked gas from the outer icm into the core material , as discussed in [ sec : pressure peaks ] .
the specific entropy profile ( figures [ fig:1:1 profile plots]f and [ fig:1:3 profile plots]f ) also demonstrates a parameter - independent outer region and parameter - dependent core , with the central entropy increase being @xmath201 over the initial central value in the hottest runs .
note that the magnitude of this increase is roughly the same for the two mass ratios .
although the gas has not reached virial equilibrium in the simulations by @xmath177 gyr , it is relatively quiescent .
figures [ fig:1:1 profile plots]g and [ fig:1:3 profile plots]g show the radial mach number profiles for each run . within @xmath165
the radial mach number is less than 0.06 in magnitude , indicating that radial motions near the center are fairly subsonic at this point .
outside @xmath165 , the magnitude of the mach number increases rapidly , reflecting the continued accretion of gas that was previously ejected from the center by the merger shocks .
the gas within the initial cutoff radius is also quite close to hydrostatic equilibrium in angle average .
the angle - averaged ratio of the pressure force to gravitational force per unit mass is shown in figures [ fig:1:1 profile plots]h and [ fig:1:3 profile plots]h .
except for the innermost few zones where the small number of zones in the radial bins and truncation error in the gradient calculation dominate this ratio , it is within 15% of unity for @xmath202 . in the simplest self - similar picture of their assembly ,
clusters undergo spherical collapse and virialization and then radiate via thermal bremsstrahlung ( kaiser 1986 ) .
this model fails to predict the observed x - ray luminosity - temperature relationship : the predicted relation is @xmath203 , whereas the observed relation is closer to @xmath204 ( mushotzky 1984 ) .
cooling flows have larger luminosities and lower average temperatures than average and are known to steepen this relation and increase its dispersion ( fabian et al .
however , removing them from cluster samples in various ways does not eliminate the discrepancy ( markevitch 1998 ; allen & fabian 1999 ; arnaud & evrard 1999 ) . various solutions to this problem have been proposed .
most recently these have focused on raising the entropy of the icm via nongravitational heating , though bryan ( 2000 ) has recently suggested that galaxy formation may do the same thing by removing low entropy gas . preheating models increase the entropy of the gas before it falls into cluster potentials ( kaiser 1991 ; evrard & henry 1991 ) .
these models require substantial amounts of heating : @xmath205 kev / baryon appears to be needed ( lloyd - davies , ponman , & cannon 2000 ) .
note that the luminosity - temperature relation used to scale the initial conditions for our simulations ( eq . [ [ eqn : observed l - t ] ] ) is not retained as the mergers progress .
comparing @xmath206 to @xmath70 as time passes in run c1 , we find that immediately following the initial brightness peak , the ratio of these two quantities jumps to twice its initial value , fluctuates here for a bit , increases briefly to almost six times its initial value during the second core passage , then settles down to about three times its initial value , where it remains roughly constant for the remainder of the calculation .
the behavior of the other runs is similar .
thus the trajectory of the system in the @xmath82 plane is roughly parallel to the observed statistical relation both before and after the merger , but during the merger the system undergoes a shift to a higher temperature normalization .
the normalization of the @xmath82 relation is complicated by the presence of two clusters within the computational volume . in this context
we note that luminosity is an ` extensive ' quantity , which must be divided by the number of clusters one is counting , while temperature is an ` intensive ' quantity , which does not scale with the number of clusters .
for equal - mass mergers , the cluster cores merge to form a single core ( to within the spatial resolution of our grid ) approximately 57 gyr after the initial core passage .
long before this point the clusters would be considered a single cluster with ` bimodal ' structure if seen along the @xmath94-axis , because the twin brightness peaks lie within an abell radius of each other and are gravitationally bound . moreover , from most lines of sight the cluster would appear to be unimodal by at most 3 gyr after initial core passage .
indeed , the head - on collision seen along the collision axis appears to be unimodal throughout the calculation .
comparing the initial and final states of the system is less problematic .
initially we have two clusters following the imposed @xmath82 relation ; at the end , we have one cluster with temperature 1.4 times the initial value and luminosity between 80% and 160% of the initial value of the individual clusters .
hence the final ratio of @xmath206 to @xmath70 is in the range 1.5 ( run c1 ) to 3 ( run c3 ) .
alternatively , one may say that the final temperature is @xmath207 the value expected given the final luminosity and the initial @xmath82 relation .
let us define @xmath208 in the highest - resolution single - cluster run ( s7 ) we see @xmath209 due to numerical effects .
nearly all of the x - ray emission arises within a few scale radii of the center of the cluster , so the increase in run s7 ( and presumably a comparable portion of the increase in the merger runs ) is probably due to spatial resolution .
the lack of an external confining pressure ( such as might be expected due to continuously infalling material from the low - density intergalactic medium ) allows material from the center of the cluster to expand more than it might otherwise , causing unwanted adiabatic cooling .
although this changes @xmath210 , adiabatic changes yield @xmath203 , so cooling due to the outflow boundaries should _ decrease _ this ratio , not increase it : @xmath211 shocks might have a much larger effect on @xmath210 because they produce less of a density jump than a temperature jump . for shocks of arbitrary strength , the inverse of the shock compression is given by ( markevitch , sarazin , & vikhlinin 1999 ) : @xmath212^{1/2 } -\frac{1}{2}\frac{\gamma+1}{\gamma-1}\left(\frac{t_2}{t_1}-1\right ) , \ ] ] where @xmath213 ( @xmath214 ) and @xmath215 ( @xmath216 ) are the preshock ( postshock ) density and temperature , respectively , and @xmath217 is the adiabatic index . for @xmath218 ,
this gives @xmath219^{1/2 } -2 \left ( \frac{t_x^{\rm final}}{t_x^{\rm initial } } -1 \right ) \right\}^{1
/ a_l } \nonumber \\ & \approx & \left ( \frac{t_x^{\rm final}}{t_x^{\rm initial } } \right)^{0.81 } \left\ { \left [ 4 \left ( \frac{t_x^{\rm final}}{t_x^{\rm initial } } -1\right)^2 + \frac{t_x^{\rm final}}{t_x^{\rm initial } } \right]^{1/2 } -2 \left ( \frac{t_x^{\rm final}}{t_x^{\rm initial } } -1 \right ) \right\}^{0.38 } \ , .
\label{eq : fltshock}\end{aligned}\ ] ] the average temperature increases by about 40% during the equal - mass merger runs , so the value of @xmath210 lies between 1.2 and 1.5 . in the 1:3 cases
, the less massive cluster contributes only about 10% as much luminosity as the more massive cluster at the start of the calculation , so in this case @xmath220 , consistent with the range of values in the 1:1 collisions . on the other hand , equations ( [ eq : fltadiab ] ) and ( [ eq : fltshock ] ) give @xmath221 and @xmath222 .
thus , the temperature increase is larger ( or the x - ray luminosity increase is smaller ) than expected from either shocks or adiabatic compression .
this probably indicates that the luminosity increase is limited by the finite numerical resolution in the cores of the clusters and would be larger if the calculations were better resolved .
we close this section with the following note .
the observation of massive clusters at high redshift ( e.g. , ms1054 : donahue et al .
1998 ) is often taken as strong evidence for low @xmath223 because of the exponential sensitivity of the cluster mass function to @xmath223 ( e.g. , oukbir & blanchard 1992 ) .
however , in deriving such constraints one directly measures an x - ray temperature , then uses the mass - temperature relation to find the mass .
our results show that this procedure is dangerous : merger shocks produce short - lived , but significant , increases in the x - ray temperature of a cluster . for a few percent of its life
, a merging cluster appears much hotter , and thus more massive , than it really is ( figure [ fig : temphist ] ) .
furthermore , because this period corresponds to a time of significantly enhanced luminosity , merging clusters are also easier to detect .
we note , for example , that ms1054 now shows evidence of substructure and a lower intrinsic temperature than first reported , suggesting that its apparent high temperature is due to an ongoing merger ( neumann & arnaud 2000 ) .
as another example , gmez , hughes , and birkinshaw ( 2000 ) have recently suggested that abell 665 , one of the richest , hottest , and most x - ray luminous clusters at redshifts @xmath224 , is currently undergoing a merger and is at the epoch when the cluster cores collide .
this could explain its very high luminosity and temperature .
we are using our results with a merger tree code to estimate the effect of merger shocks on the significance of distant , massive cluster constraints on @xmath223 and will address this question in a subsequent paper ( randall , sarazin , & ricker 2001 , in preparation ) .
a cluster is defined by the presence of a distinct potential well within which dark matter , gas , and galaxies are confined .
mergers produce dramatic fluctuations in these potentials , leading eventually to the combination of the merging clusters into a single remnant .
some merging clusters in which the mergers appear to be well - advanced show a double core structure at their centers ( e.g. , abell 2065 ; markevitch et al .
1999 ) , with two x - ray surface brightness peaks , indicating that the cores of the individual cluster potentials are still intact
. if the original subclusters have cooling flows at their centers , the x - ray surface brightness of the surviving peaks in the subcluster potentials can be particularly large .
cluster cores lose their distinct identity in several ways . as with galaxies that pass too close to a cluster s center , tidal
stripping can remove material from a shallower cluster potential and add it to a deeper one .
in addition , violent relaxation ( for collisionless matter ) and ram - pressure stripping ( for diffuse gas ) convert the bulk kinetic energy of a cluster into dispersive or thermal energy , enabling material to become unbound from its own potential , while remaining bound within the combined potential of the material from both clusters .
dynamical friction offers another means for converting organized bulk velocities to dispersive ones , but the timescale on which it operates is typically large compared to the timescales associated with these other mechanisms .
our results shed some light on the mechanism by which ram pressure affects the survival of cluster cores in mergers . rather than ablating gas from the edges of a core , ram pressure disrupts cores by pushing the gas as a unit away from the center of the potential well in which it originates , causing it to become convectively unstable .
the resulting convective motions destroy the spherical symmetry of the core and mix the core gas with higher - entropy shocked gas in its surroundings .
this also appears to be an important mechanism by which mergers initiate turbulent motions .
figure [ fig : convective plume ] shows an example of this process using the specific entropy field .
this example begins just after the first core interaction of run c5 ( the less massive cluster core has exited to the right at 4.00 gyr ) .
the diffusivity of the gas is low , so the core interaction pushes low - entropy gas out of the more massive core , where it finds itself out of convective equilibrium . beginning around 5.00 gyr this
gas flows in a plume back toward the more massive potential center .
a large eddy also forms in the wake of the less massive cluster ; its remnants persist until 7.00 gyr .
the plume overshoots the potential center and begins to decelerate and spread along equipotential lines .
it reaches maximum overshoot around 7.00 gyr , when it is met by high - entropy gas from the returning less massive core .
the remaining gas bound to this core has higher entropy and , stripped from its potential well , is forced to flow around the plume . as the dark matter associated with the less massive cluster passes through the more massive potential well , it drives a new shock through the low - entropy gas from the more massive cluster that did not flow into the plume ; this shock exits to the lower left after 8.50 gyr .
the remnants of the plume mix with the newly reshocked gas to form the core of the merger remnant , with higher entropy than was present in the original clusters .
the turbulent velocity field created by a merger lasts for several sound - crossing times . in figure [ fig : rms gas velocities ] we plot the rms gas velocity profile in each of the collision runs at @xmath225 gyr as a fraction of the average sound speed and the circular ( keplerian ) velocity . throughout the merger remnant ,
turbulent velocities are subsonic : for @xmath226 they are typically @xmath227% of the sound speed in the 1:1 cases , while the rms mach number is about twice as large in the 1:3 cases . outside this radius
the rms velocity rises to between 60% and 100% of the sound speed , with the increase much steeper in the 1:1 cases . throughout most of the cluster ,
turbulence supplies @xmath228% of the hydrostatic support . in certain cases
this contribution appears to be larger within the innermost @xmath28 or so of the core , though it is likely that here the rms velocity is overestimated due to the relatively small number of zones contributing to radial velocity averages .
our results support the finding of others ( e.g. , roettiger , burns , & loken 1996 ; norman & bryan 1998 ) that , following a merger , the intracluster medium becomes turbulent , and that turbulent pressure and bulk rotation provide less than 20% of the support needed for hydrostatic equilibrium .
the picture which emerges from our calculations is that this turbulence is initiated by convective instability and , to a lesser extent , by baroclinic vorticity generation at oblique shocks .
it is sustained for several gyr by pumping due to the oscillation of the dark matter , which relaxes more slowly than the gas . while roettiger et al .
( 1996 ) also argue for this mechanism using isolated merger simulations , norman and bryan ( 1998 ) find higher rms velocities in a cluster simulation involving multiple mergers and accretion , and they argue that minor mergers are the primary pumping mechanism . our results suggest that individual major mergers and cumulative minor mergers induce comparable amounts of turbulence , since the turbulent pressure in our simulations is @xmath229 that of norman & bryan s result even several gyr after the collision . to date the grid - based merger simulations used to argue for turbulence in clusters have reached spatial resolutions on the order of 10 kpc . in fully developed turbulence , the range of eddy length scales is of order @xmath230 where @xmath231 is the integral scale and @xmath232 is the dissipation scale ( richardson 1922 ; kolmogoroff 1941 ) .
thus , directly simulating three - dimensional turbulent flow with a hydrodynamic code requires at least of order @xmath233 zones . for many astrophysical applications ,
this is impossible for example , in stellar convective layers @xmath234 .
however , as we have already argued , the reynolds number of the intracluster medium during mergers is of order @xmath235 , and turbulence , if it is present at all , operates over a much smaller range of length scales .
we therefore have some hope of performing direct numerical simulations of merger - driven turbulence within the very near future . by comparison ,
the simulations described in this paper achieve a spatial dynamic range of @xmath236 .
several approximate methods for numerically treating turbulence exist ( see canuto 1994 for a partial review ) . among these methods , large - eddy simulation ( les )
has achieved some success in reproducing characteristics of turbulence in laboratory experiments ( canuto 1997 ) .
les attempts to simulate directly only the largest - scale eddies in a turbulent flow , since these depend on the boundary conditions and geometry of the flow at the integral scale , and it uses a subgrid model to handle the ( presumably universal and geometry - independent ) turbulent dissipation on unresolved scales .
most grid - based astrophysical simulation algorithms are , on some level , les , but they do not use subgrid turbulence models , instead relying on numerical dissipation at the zone scale .
this numerical dissipation arises due to the truncation error of each individual method and bears no relation to the real dissipation mechanism . for most methods this practice should yield an incorrect energy dissipation spectrum , since it links large- and small - scale eddies in an unphysical way ( canuto 1997 ) .
however , it has been argued by sytine et al .
( 2000 ) that the piecewise - parabolic method , which we employ in this paper , accurately reproduces the energy and enstrophy power spectra produced by solving the full navier - stokes equations in direct simulations of compressible turbulence , because it is much less dissipative than other techniques and because its dissipation is confined to narrow regions near flow discontinuities .
estimates of the dissipation in ppm for some simple flow fields ( porter & woodward 1994 ) suggest that convective features spanned by 32 or more zones are well - converged .
this is consistent with recent results of calder et al .
( 2001 ) for single - mode rayleigh - taylor instabilities .
thus , for example , the convective plume shown in figure [ fig : convective plume ] , which measures approximately 32@xmath23710 zones at maximum extent , is probably just sufficiently resolved to yield the correct growth rate .
further high - resolution calculations using adaptive mesh refinement will be useful in verifying this result ( ricker et al .
we have presented results from a controlled parameter study of off - axis mergers between clusters of galaxies .
we have extended previous work by using the nfw density profile and nonisothermal temperature profiles , systematically controlling merger parameters and justifying their scaling , carefully studying numerical effects and convergence requirements , and examining the survival of pressure peaks and the mechanism for the onset of turbulence in mergers .
we have also presented results on the variation of relaxation times , merger remnant properties , and luminosity and temperature peaks with merger parameters .
the calculations presented here do not include radiative cooling or magnetic fields .
we anticipate adding these physical processes in future calculations ( particularly cooling , which appears to play a very important dynamical role in many observed clusters ) .
the morphological changes , relative velocities , and temperature jumps we observe agree well with previous studies of collisions between clusters modeled using the king profile ( e.g. , roettiger et al .
1998 ) . in particular ,
velocities of @xmath238 km s@xmath7 are to be expected , even in quite offset mergers , for cosmologically reasonable initial conditions .
the peak temperature occurs for a period @xmath239 at first core passage and is a factor @xmath240 greater than the initial temperature .
we observe a larger jump in x - ray luminosity ( @xmath2 ) than previous studies including dark matter , and we argue that this increase is most likely a lower limit due to our spatial resolution .
second core passage , with corresponding secondary luminosity and temperature jumps , typically occurs @xmath241 after the first . in the most offset 1:3 case
, this secondary peak is delayed by an additional sound - crossing time .
we emphasize that luminosity and temperature jumps due to mergers may have an important bearing on constraints on @xmath3 derived from the observation of hot clusters at high redshift .
we will address this issue further in a subsequent paper .
shocks play an important dissipative role in mergers , but they are relatively weak in the highest - density regions . as a result they do not directly raise the entropy of the cluster cores .
instead , shocks create entropy in the outer parts of the clusters , and this high - entropy gas is mixed with the core gas during later stages of the merger . mixing is initiated by ram pressure :
the core gas is displaced from its potential center and becomes convectively unstable .
the resulting convective plumes initiate large - scale turbulent motions with eddy sizes up to several 100 kpc .
this turbulence is pumped by oscillations in the gravitational potential , which in turn are driven by the more slowly relaxing collisionless dark matter . even after nearly a hubble time these motions persist as subsonic turbulence in the cluster cores , providing @xmath4 of the support against gravity .
the dark matter oscillations are also reflected in the extremely long time following a merger required for the remnant to come to virial equilibrium . because of the increase in core entropy , if a constant - density gas core is initially present , it is not destroyed by a merger , even if the dark matter density profile has a central cusp . this research has been supported in part by nasa astrophysical theory program grant nag 5 - 3057 and _ chandra _ award numbers go0 - 1019x , go0 - 1141x , go0 - 1158x , and go0 - 1173x at the university of virginia , and by the asci flash center at the university of chicago under doe contract b341495 .
calculations were performed using the computational resources of the pittsburgh supercomputing center and the san diego supercomputer center .
m. r. would like to acknowledge useful conversations with d. lamb , m. c. miller , j. stone , and e. ostriker .
allen , s. w. , & fabian , a. c. 1999 , mnras , 297 , l57 arnaud , m. , & evrard , a. e. 1999 , mnras , 305 , 631 blasi , p. 2000 , aph , to appear ( astro - ph/0008113 ) blasi , p. , & colafrancesco , s. 1999 , aph , 12 , 169 bliton , m. , rizza , e. , burns , j. o. , owen , f. n. , & ledlow , m. j. 1998 , mnras , 301 , 609 brandt , a. 1977 , math .
, 31 , 333 bryan , g. 2000 , apjl , 544 , 1 burns , j. o. , rhee , g. , owen , f. n. , & pinkney , j. 1994 , apj , 423 , 94 burns , j. o. , roettiger , k. , loken , c. , norman , m. , & bryan , g. 2000 , in cluster mergers and their connection to radio sources , 24th meeting of the iau , joint discussion 10 calder , a. c. , et al .
2001 , to appear in proc .
20th texas symposium on relativistic astrophysics , december 10 - 15 , 2000 , austin , tx ( astro - ph/0102239 ) canuto , v. m. 1994 , apj , 428 , 729 _ _ _ _ _ _ _ _ .
1997 , apj , 478 , 322 cavaliere , a. , & fusco - femiano , r. 1976 , a&a , 49 , 137 chandran , b. d. g. , cowley , s. c. , ivanushkina , m. , & sydora , r. 1999 , apj , 525 , 638 colella , p. & woodward , p. 1984 , jcp , 54 , 174 david , l. p. , jones , c. , & forman , w. 1995 , apj , 445 , 578 donahue , m. , voit , g. m. , gioia , i. , lupino , g. , hughes , j. p. , & stocke , j. t. 1998 , apj , 502 , 550 edge , a. c. , et al . 1990 , mnras , 245 , 559 edge , a. c. , stewart , g.
c. , & fabian , a. c. 1992 , mnras , 258 , 177 evrard , a. e. , & henry , j. p. 1991 , apj , 383 , 95 evrard , a. e. , metzler , c. a. , & navarro , j. f. 1996 , apj , 469 , 494 frenk , c. s. et al .
1999 , apj , 525 , 554 gingold , r. a. & monaghan , j. j. 1977 , mnras , 181 , 375 gmez , p. l. , pinkney , j. , burns , j. o. , wang , q. , owen , f. n. , & voges , w. 1997 , apj , 474 , 580 gmez , p. l. , hughes , j. p. , & birkinshaw , m. 2000 , apj , 540 , 726 harris , d. e. , lari , c. , vallee , j. p. , & wilson , a. s. 1980 , a&as , 42 , 319 henriksen , m. j. , & markevitch , m. l. 1996 , apjl , 466 , 79 hockney , r. w. & eastwood , j. w. 1988 , computer simulation using particles ( bristol : iop ) hoyle , f. 1949 , in problems of cosmical aerodynamics ( dayton : central air documents office ) , 195 jing , y. p. , & suto , y. 2000 , apjl , 529 , 69 jones , c. , & forman , w. 1984 , apj , 276 , 38 kaiser , n. 1986 , mnras , 222 , 323 _ _ _ _ _ _ _ _ .
1991 , apj , 383 , 104 kolmogoroff , a. n. 1941 , dokl .
nauk sssr , 32 , 16 lloyd - davies , e. j. , ponman , t. j. , & cannon , d. b. 2000 , mnras , 315 , 689 lucy , l. b. 1977 , aj , 82 , 1013 markevitch , m. 1996 , apjl , 465 , 1 markevitch , m. 1998 , apj , 504 , 27 markevitch , m. , forman , w. r. , sarazin , c. l. , & vikhlinin , a. 1998 , apj , 503 , 77 markevitch , m. , sarazin , c. l. , & vikhlinin , a. 1999 , apj , 521 , 526 markevitch , m. , et al .
2000 , apj , 541 , 542 mazzotta , p. , markevitch , m. , vikhlinin , a. , forman , w. r. , david , l. p. , & van speybroeck , l. 2001 , apj , in press ( astro - ph/0102291 ) mcglynn , t. a. , & fabian , a. c. 1984 , mnras , 208 , 709 mcnamara , b. r. , et al .
2000 , apjl , 534 , 135 mewe , r. , kaastra , j. s. , & liedahl , d. a. 1995 , legacy , 6 , 16 molendi , s. , de grandi , s. , & fusco - femiano , r. 2000 , apjl , 533 , 43 moore , b. , governato , f. , quinn , t. , stadel , j. , & lake , g. 1998 , apj , 499 , l5 mushotzky , r. f. 1984 , phys .
scripta , t7 , 157 mushotzky , r. f. , & scharf , c. a. 1997 , apj , 482 , l13 navarro , j. f. , frenk , c. s. , & white , s. d. m. 1997 , apj , 490 , 493 ( nfw ) neumann , d. , & arnaud , m. 2000 , apj , 542 , 35 norman , m. l. , & bryan , g. l. 1998 , in ringberg workshop on m87 , eds .
k. meisenheimer and h .- j .
rser ( springer ) oukbir , j. , & blanchard , a. 1992 , a&a , 262 , l21 pearce , f. r. , thomas , p. a. , & couchman , h. m. p. 1994
, mnras , 268 , 953 peebles , p. j. e. 1969 , apj , 155 , 393 porter , d. h. , & woodward , p. r. 1994 , apjs , 93 , 309 richardson , l. f. 1922 , weather prediction by numerical process ( cambridge : cup ) ricker , p. m. 1998 , apj , 496 , 670 ricker , p. m. , dodelson , s. , & lamb , d. q. 2000 , apj , 536 , 122 ricker , p. m. , et al .
2001 , to appear in proceedings of the vii international workshop on advanced computing and analysis techniques in physics research ( acat 2000 ) , october 16 - 20 , 2000 , fermi national accelerator laboratory , batavia , il ( astro - ph/0011502 ) roettiger , k. , burns , j. , & loken , c. 1993 , apjl , 407 , 53 _ _ _ _ _ _ _ _ .
1996 , apj , 473 , 651 _ _ _ _ _ _ _ _ .
1997 , apjs , 109 , 307 roettiger , k. , & flores , r. 2000 , apj , 538 , 92 roettiger , k. , stone , j. , & mushotzky , r. 1998 , apj , 493 , 62 sakelliou , i. 2000 , mnras , 318 , 1164 sarazin , c. l. 1988 , x - ray emission from clusters of galaxies ( cambridge : cup ) sarazin , c. l. 2001 , in merging processes in clusters of galaxies , ed .
l. feretti , i. m. gioia , & g. giovannini ( dordrecht : kluwer ) , in press ( astro - ph/0105418 ) sarazin , c. l. , & lieu , r. 1998 , apjl , 494 , 177 schindler , s. 2000 , in cluster mergers and their connection to radio sources , 24th meeting of the iau , joint discussion 10 ( astro - ph/0010006 ) schindler , s. , & mller , e. 1993 , a&a , 272 , 137 stevens , i. r. , acreman , d. m. , & ponman , t. j. 1999 , mnras , 310 , 663 stewart , g. c. , fabian , a. c. , jones , c. , & forman , w. 1984 , apj , 285 , 1 sugerman , b. , summers , f. j. , & kamionkowski , m. 2000 , mnras , 311 , 762 sytine , i. v. , porter , d. h. , woodward , p. r. , hodson , s. w. , & winkler , k .- h .
2000 , jcp , 158 , 225 takizawa , m. 2000 , apj , 532 , 183 valinia , a. , henriksen , m. j. , loewenstein , m. , roettiger , k. , mushotzky , r. f. , & madejski , g. 1999 , apj , 515 , 42 vikhlinin , a. , markevitch , m. , & murray , s. s. 2001 , apj , submitted ( astro - ph/0008496 ) white , d. a. , & fabian , a. c. 1995 , mnras , 273 , 72 white , s. d. m. 1984 , apj , 286 , 38 zabludoff , a. i. , & zaritsky , d. 1995 , apj , 447 , l21 | we present a parameter study of offset mergers between clusters of galaxies . using the eulerian hydrodynamics/@xmath0-body code cosmos
, we simulate mergers between nonisothermal , hydrostatic clusters with a steep central dark matter density profile and a @xmath1-model gas profile .
we constrain global properties of the model clusters using observed cluster statistical relationships .
we consider impact parameters between zero and five times the dark matter scale radius and mass ratios of 1:1 and 1:3 .
the morphological changes , relative velocities , and temperature jumps we observe agree with previous studies using the king profile for the dark matter .
we observe a larger jump in x - ray luminosity ( @xmath2 ) than in previous work , and we argue that this increase is most likely a lower limit due to our spatial resolution . we emphasize that luminosity and temperature jumps due to mergers may have an important bearing on constraints on @xmath3 derived from the observation of hot clusters at high redshift .
shocks are relatively weak in the cluster cores ; hence they do not significantly increase the entropy there .
instead , shocks create entropy in the outer regions , and this high - entropy gas is mixed with the core gas during later stages of the merger .
ram pressure initiates mixing by displacing the core gas from its potential center , causing it to become convectively unstable .
the resulting convective plumes produce large - scale turbulent motions with eddy sizes up to several 100 kpc .
this turbulence is pumped by dark matter - driven oscillations in the gravitational potential . even after nearly a hubble time
these motions persist as subsonic turbulence in the cluster cores , providing @xmath4 of the support against gravity .
the dark matter oscillations are also reflected in the extremely long time following a merger required for the remnant to reach virial equilibrium . |
in recent years the field of nonlinear atom optics @xcite has become firmly established both theoretically and experimentally with the demonstration of four - wave mixing in bose - einstein condensates @xcite , coherent matter wave amplification @xcite , the creation of dark @xcite and bright @xcite atomic solitons , the generation of correlated atomic pairs and squeezing @xcite , and the creation of a molecular condensate the analog of second - harmonic generation in optics @xcite .
so far , this work has been largely limited to bosonic atoms and has been based to a great extent on analogies with nonlinear optics : ultracold bosons interact primarily via @xmath0-wave collisions and typically occupy only a few quantum states , so that their mean - field description in terms of a gross - pitaevskii equation includes a nonlinearity of the same form as a third - order optical kerr nonlinearity , and thus many of the results from nonlinear optics can be directly applied to condensates . in particular ,
bose - einstein condensates are often viewed as a matter - wave equivalent of the optical laser the so called `` atom laser '' .
the last few years have also seen substantial progress in the cooling and trapping of atomic fermi gases to the quantum degenerate regime , with temperatures reaching as low as @xmath1 where @xmath2 is the fermi temperature @xcite .
it is therefore only a matter of time before experiments start to probe the dynamics of these gases far from equilibrium and perform fermionic nonlinear atom optics experiments .
however , the extension of the ideas developed in nonlinear optics to fermi systems is by no means obvious or straightforward because of the different quantum statistics obeyed by fermions and bosons .
whereas for bosons the analysis can typically be limited to a few macroscopically occupied modes , the pauli exclusion principle prohibits the number of modes from ever being less than the number of particles for fermions .
this makes both analytical and numerical studies of the dynamics of fermi systems rather challenging , except for very limited situations ( e.g. @xmath3 and weak perturbations ) .
consequently , the idea of doing nonlinear atom optics with fermions is still new and largely unexplored .
it is only recently that it has been convincingly argued that four - wave mixing is possible with fermions @xcite .
these works showed how four - wave mixing can be interpreted in terms of incident particles undergoing bragg scattering off of a fermionic density grating , but were limited to an incident beam consisting of a single test particle .
they also neglected the dynamics of the density grating .
related work has shown that four - wave mixing between a degenerate fermion beam and a bosonic density grating is also possible @xcite . while these results are promising , a complete analysis of purely fermionic four - wave mixing is still lacking . in a previous paper @xcite , we studied fermionic four - wave mixing using the matter - wave analog of the undepleted pump approximation in optics , i.e. we neglected the back - action of the diffracted atoms on the density grating .
the density grating was formed by atoms in a single hyperfine state prepared in a coherent superposition of the momentum states @xmath4 and @xmath5 for all @xmath6 such that @xmath7 , where @xmath8 is the fermi wave number of the grating atoms and @xmath9 the wave number of the density modulation . the third incident wave consisted of a degenerate fermi gas in a different hyperfine state so that the atoms in the beam and grating could interact via @xmath0-wave collisions .
these atoms could then be diffracted by the density grating to produce a fourth , scattered wave .
because of the range of energy states occupied by its constituent atoms , the amplitude of the grating would decay away due to the dephasing of the atoms .
this led to a finite time during which four - wave mixing could occur and placed serious constraints on the properties of the fermions involved . in this paper , we extend the results of ref .
@xcite by taking into account the back - action of the incident beam on the fermionic density grating .
we find that when the number of atoms in the incident beam becomes a sizable fraction of the number of atoms forming the density grating , it no longer decays away due to the dephasing .
instead , it exhibits large nonlinear amplitude oscillations that are coupled to the bragg oscillations of the beam .
this leads to the efficient generation of the fourth scattered wave , even for times much longer than the grating dephasing time .
we interpret this behavior in terms of a bloch vector picture .
we also evaluate the efficiency for the generation of the scattered wave and show that it should be possible to produce a macroscopic fourth wave .
interestingly enough , we find that the efficiency of the four - wave mixing can be improved significantly by initially preparing a density grating with an amplitude that is less than maximal .
the organization of this paper is as follows : in section ii , we briefly review the model developed in @xcite .
section iii presents numerical results that show how the dephasing of the grating can be eliminated when the back - action becomes significant .
the efficiency of the four - wave mixing is also calculated and it is shown that a macroscopic fourth wave can be readily produced . finally , section iv is a summary and conclusion .
the appendix presents an analytically soluble model in terms of coupled bloch vectors that yields a qualitative interpretation of our results .
in this section we briefly review the one - dimensional model developed in @xcite to study four - wave mixing between fermions at zero temperature .
we consider two spin - polarized species of fermions a density grating composed of spin - up polarized atoms and an incident beam of spin - down polarized atoms , see fig .
[ model ] .
atoms in different spin states interact via elastic @xmath0-wave scattering .
the cross - section for @xmath10-wave scattering at @xmath3 is orders of magnitude smaller than that of @xmath0-wave scattering , thus @xmath10-wave scattering is neglected in our model @xcite .
-wave ) by a spin - up polarized grating.,width=302,height=188 ] for clarity and simplicity we restrict ourselves to one spatial dimension .
it is usually possible to reduce the problem of four - wave mixing to a two - dimensional planar geometry , with one dimension parallel to the density grating .
since the momentum transfer to the beam atoms is parallel to the density modulation of the grating , the essential dynamics of the wave mixing process occurs only in that @xmath11-direction .
we assume that the grating is confined between @xmath12 to @xmath13 along the perpendicular @xmath14 axis . if the average momentum , @xmath15 , of the atomic beam along that direction is sufficiently large that @xmath16 , then the requirements of energy and momentum conservation insure that @xmath17 remains practically unchanged by the scattering . in this case ,
wave mixing in two or three dimensions is in essence the same as in the one - dimensional model that we consider .
( the transverse momenta of the beam atoms can at most change sign due to reflection off the grating at @xmath12 @xcite , but the reflected wave can be minimized by using a smooth grating density profile in the @xmath14-direction . ) the second quantized hamiltonian describing the situation at hand is @xmath18 with @xmath19 and @xmath20 .
here , @xmath21 is the scattering length , @xmath22 is the transverse cross - sectional area , @xmath23 the mass of one atom and @xmath24 the quantization length . the operators @xmath25 and @xmath26 are annihilation and creation operators for atoms with momentum @xmath27 and spin @xmath28 , satisfying standard fermionic anticommutation relations .
the dynamics of the system is conveniently described in terms of the particle - hole operators @xmath29 that create a particle in state @xmath30 and a hole in @xmath6 for the beam and grating , respectively .
they obey the heisenberg equations of motion @xmath31 with @xmath32 .
we consider a beam of @xmath33 atoms propagating along the @xmath34-direction , but with a small momentum component around a central value @xmath35 parallel to the grating , which is taken to be in a superposition of the states @xmath4 and @xmath5 for @xmath36 . here
, @xmath37 is the fermi momentum ( in one dimension ) and @xmath38 is the number of atoms in the grating .
such a grating can be created from a stationary homogeneous gas by a two - step process where first a two - photon bragg @xmath39-pulse imparts a momentum @xmath40 to each atom , followed by a second bragg pulse with momentum @xmath41 that creates the desired superposition ( see e.g. ref .
( note that the bandwidth of the pulses should be greater than @xmath42 . )
thus the initial state of the beam - grating system is @xmath43 where @xmath44 @xmath45 being the fermi momentum of the beam .
the initial expectation values of the particle - hole operators @xmath46 and @xmath47 are readily found to be @xmath48 we note that the spatial density of the fermi gas is given by @xmath49 from which it is clear that the density for the initial state of the spin - up polarized grating atoms exhibits as expected a spatial modulation with the grating wave number @xmath9 and an amplitude @xmath50 . in contrast , the spin - down polarized incident beam has a uniform density .
it is convenient to introduce the slowly varying particle - hole operators @xmath51 whose heisenberg equation of motion simply become @xmath52 in the following we factorize the expectation values of these operators for different spin states , @xmath53 this ansatz results in the truncation of the bbgky - type hierarchy of equations of motion for the higher - order moments of the particle - hole operators after the first moments @xcite , yielding the closed set of @xmath54-number equations @xmath55 where @xmath56 and @xmath57 .
these equations can be readily integrated numerically .
implicit in eqs .
( [ integrated_c ] ) and ( [ initial_state ] ) are the three fundamental time scales that govern the dynamics of the interaction between the beam and the grating .
the first time scale is associated with the kinetic energy @xmath58 gained by an atom scattered by the grating .
the other two time scales are associated with the density grating itself . if , for the moment , we revert to the undepleted pump approximation used in ref .
@xcite , namely @xmath59 , then we obtain the linear equations of motion for the beam , @xmath60 here @xmath61 where the second equality was obtained in the continuum limit , @xmath62 .
the coupling constant @xmath63 is proportional to the amplitude of the density grating , which decays to zero in a time of the order of @xmath64 this decay is caused by the spread in energies of the fermions that comprise the grating , @xmath65 , so that the different energy states in the grating get out of phase with each other over time .
this leads to an effective grating lifetime of the order of @xmath66 , which corresponds to the relative phases of the different states being of order unity .
the third time scale is given by the characteristic time @xmath67 for atoms to scatter off of the grating in the limit of a static grating ( @xmath68 ) .
for an incident particle with momentum @xmath69 , @xmath70 is the frequency of oscillations between the degenerate states @xmath71 and @xmath72 . in the bragg regime that we consider in this paper , both the kinetic energy and the momentum of the atoms are conserved by the scattering process .
this corresponds to the condition , @xmath73 the population that develops in states centered around @xmath74 is then at most on the order of @xmath75 .
if we choose the average momentum of the beam as @xmath76 , with @xmath77 , only states centered around @xmath72 are phase - matched to the states in the incident beam , so that we can restrict ourselves to those @xmath6 states within a width @xmath78 of @xmath69 .
( this is in contrast to the raman - nath regime , where the change in the kinetic energy of an atom as result of scattering off of the grating is negligible compared to the interaction energy with the grating @xcite . ) in the undepleted pump approximation it is important that the scattering time be less than the lifetime of the grating , @xmath79 otherwise there is no significant scattering beam generated .
however , as we show in the next section , four - wave mixing can continue for times much longer than @xmath80 when the effects of the back - action on the grating become significant .
in this section we study the phenomena in fermionic four - wave mixing resulting from the back - action of the atomic beam on the grating .
we proceed by numerically solving eqs .
( [ integrated_c ] ) , subject to the initial conditions ( [ initial_state ] ) with @xmath81 , using a fourth - order runge - kutta algorithm with fixed time steps chosen to produce a relative error smaller than @xmath82 . as a function of time ( in units of @xmath83 ) for ( a ) @xmath84 and ( b ) @xmath85 . in both cases @xmath86 and @xmath87=2.,width=302,height=415 ] the back - action of the beam on the density grating is negligible for all times when the number of beam atoms is much less than the number of atoms in the grating , @xmath88 .
the reason is that for every atom in the beam that suffers a momentum transfer of @xmath89 , one atom of the grating sees its wave function `` reduced '' from the superposition state @xmath90 to the state @xmath91 . since only those atoms in a superposition of @xmath4 and @xmath5 ( for all possible @xmath6 )
contribute to the density modulation , the fractional reduction in the amplitude of the grating will be at most @xmath92 and the undepleted pump approximation is valid for all times .
this approximation also holds for arbitrary @xmath33 for times short enough that the number of atoms scattered by the grating is much less than @xmath38 , that is , for times @xmath93 . as a function of time scaled by @xmath94 for the ratios @xmath95 and @xmath96 in the subplot ( a ) , ( b ) and ( c ) , respectively .
, width=302,height=264 ] fermionic four - wave mixing is conveniently characterized in terms of the occupation numbers of the momentum states of the beam and grating atoms @xmath97 and the amplitude of the density modulation of the grating , @xmath98 fig .
[ fig2]a plots @xmath99 as a function of time for the case of a `` large '' grating , @xmath84 and @xmath86 , and fig .
[ fig3]a shows the evolution of the density modulation of the grating @xmath100 for these same parameters .
the atoms in the beam undergo bragg oscillations between the @xmath6 states centered around @xmath72 and @xmath101 up until a time of the order of @xmath80 , when the amplitude of the oscillations dies out .
it is clear from the behavior of @xmath100 that the disappearance of the bragg oscillations corresponds to the decay of the grating amplitude due to dephasing . in this example
, the back - action of the beam on the grating manifests itself in the small amplitude oscillations superimposed on the sinc function behavior of @xmath100 .
these oscillations reflect the coupling between individual modes of the beam and of the grating , which all occur at roughly the same frequency @xmath70 for @xmath102 , as further discussed later on . figs .
[ fig2]b and [ fig4]b are the same as fig .
[ fig2]a except that `` smaller '' gratings , with @xmath85 and @xmath103 , respectively , have been used . for these parameters
there is a dramatic change in the behavior of @xmath100 as the back - action becomes more pronounced . instead of decaying away , the density grating now undergoes large amplitude oscillations , as illustrated in fig .
[ fig3 ] . at the same time
, the beam atoms undergo nearly perfect bragg oscillations that persist for times much longer than @xmath80 .
note that the amplitude of the oscillations is equal to 1 for @xmath85 , but only to about @xmath104 for @xmath103 ( see the scales of the grey - scale plots ) although there is no evidence of any decay in the bragg oscillations .
if one ignores the weak population of the states centered around @xmath105 ( an approximation valid in the bragg regime ) these coupled oscillations involve the oscillation of a beam atom between the states @xmath106 and @xmath107 , while an atom in the grating oscillates between @xmath108 and @xmath109 , @xmath110 out of phase with the beam atom , as follows from the existence of the constants of motion @xmath111 ) for the case of equal particle numbers @xmath103 . the upper plots ( a ) show the density of the grating , @xmath112 , and the lower plots ( b ) show the beam evolution , @xmath113 . ,
width=302,height=415 ] we can gain a better understanding of fermionic four - wave mixing by recasting it in terms of the dynamics of an inhomogeneously broadened ensemble of coupled two - level systems . in this picture ,
an individual two - level system consists of the manifold of beam and grating states @xmath114 for all states with @xmath7 and @xmath115 for the grating and beam atoms , respectively .
we proceed by introducing the associated bloch vectors @xcite , @xmath116 with the components @xmath117 it is easily seen that the bloch vectors associated with the atoms in the incident beam obey the equations of motion @xmath118 that is , they precess around the effective field @xmath119 where @xmath120 here , @xmath121 and @xmath122 are the components of the @xmath123-th bloch vector for the beam , @xmath124 .
they are defined in the same way as eqs .
( [ bvector_g])-([wg ] ) for the grating .
clearly , the grating bloch vector @xmath125 also obeys eq .
( [ bloch ] ) , with the interchange @xmath126 .
the components of the bloch vector in the equatorial @xmath127 plane of the bloch sphere are a measure of the coherence between the momentum states @xmath108 and @xmath109 and therefore the sum of the projections of the grating bloch vectors onto the equatorial plane is equal to the amplitude of the density grating . of the atomic densities ; dotted line : analytic result of eq .
( [ period ] ) .
time in units of @xmath128 .
, width=302,height=188 ] initially all of the bloch vectors for the grating point along the @xmath14 axis , while those of the atomic beam , which does nt exhibit any coherence , point toward the north pole ( @xmath129 ) of the bloch sphere .
( that they point north rather than south follows from the fact that the initial momenta of the beam are in an interval of width @xmath130 about @xmath131 , see eq .
( [ initial_state ] ) .
we have seen that the bragg regime requires additionally that @xmath132 , so that @xmath133 is initially equal to zero and @xmath134 . ) in the absence of any coupling to the beam , @xmath135 , the version of eq .
( [ bloch ] ) for the grating shows that the evolution of the bloch vectors is simply a rotation in the equatorial plane at frequency @xmath136 .
this `` inhomogeneous broadening '' causes the grating to dephase because of the spreading out of the bloch vectors in the equatorial plane .
consider now the other extreme situation where the collisional coupling between the beam and grating atoms completely dominates the dynamics , i.e. @xmath137 or equivalently , @xmath138 .
the amplitude of the effective field that drives the grating , @xmath139 , is initially equal to zero since the initial bloch vectors of the beam point towards the north pole ( @xmath129 ) of the bloch sphere .
however , those vectors start rotating around the @xmath14-axis under the influence of the instantaneous grating effective field @xmath140 , leading in turn to a projection along the @xmath141-axis .
this then allows the bloch vector of the grating to start rotating about that axis towards the north pole of the bloch sphere at the instantaneous frequency @xmath142 . as @xmath125 approaches the @xmath141-axis
, @xmath143 reaches the north pole and its projection on the equatorial plane vanishes and then changes sign . this reverses the sense of rotation of the beam bloch vectors , which begin to move back toward the pole , at which point their projections on the @xmath141-axis likewise change sign .
this reverses the direction of rotation of @xmath143 , which then passes back through the north pole , reversing the sense of rotation of @xmath125 , etc .
we see that the beam atoms are transformed into a grating - like superpositions state , and at the same time the atoms forming the initial grating take on a state similar to the initial beam state .
thus , as time passes , the labels `` beam '' and `` grating '' become more and more interchangeable .
one result of this back - and - forth is that the oscillation frequency of the populations is twice that of the coherences .
furthermore , their amplitude is only @xmath104 , both for the beam and the grating , since the bloch vectors are confined to the northern hemisphere of the bloch sphere , @xmath144 .
this explains the reduced amplitude of the oscillations in fig .
[ fig4 ] . the analytic solution of this model given in the appendix yields , for @xmath145 , an oscillation period of @xmath146 where @xmath147 is the complete elliptic integral of the first kind @xcite . as a function of time ( in units of @xmath83 ) for the case of short dephasing time @xmath148 , with @xmath103 and @xmath149 .
the upper plots are for the grating and the lower plots for the beam.,width=302,height=415 ] we have numerically calculated the fourier transform of @xmath150 for @xmath151 and @xmath152 and confirmed that the populations of the grating and the beam oscillate at the same frequency @xmath153 independently of @xmath6 . fig .
( [ fig5 ] ) compares the period of oscillations @xmath154 as a function of @xmath33 for fixed @xmath38 to the analytical result of eq .
( [ period ] ) , showing a very good agreement .
we find that the dominant frequency in the evolution of @xmath155 is approximately @xmath156 , in agreement with the predictions of bloch vector model .
the small amplitude , high - frequency oscillations in @xmath155 are due to a residual coupling to momentum states around @xmath157 . the picture that we have developed indicates that the effects of dephasing can only be suppressed when @xmath158 . for @xmath159 , the bloch vectors of the grating no longer precess around the @xmath160-axis but rather around an axis at an angle @xmath161 relative to the @xmath11-axis at the `` generalized rabi frequency '' @xmath162 .
this again leads to the spreading out of the bloch vectors in the equatorial plane and hence a collapse of the density grating .
this is confirmed in fig .
[ fig6 ] , which shows the dynamics of the populations forming ( a ) the grating and ( b ) the beam for @xmath103 and @xmath163 .
it illustrates clearly the frequency dependence on @xmath6 familiar from the generalized rabi oscillations of two - level physics .
the four - wave mixing efficiency can be quantified by the maximum number of beam atoms that can be scattered by the grating .
initially the beam atoms are scattered into negative momentum states , hence the peak efficiency is given by the first maximum of @xmath164 fig .
[ fig7 ] shows @xmath165 for different dephasing times .
we observe a rapid decrease in efficiency when @xmath166 .
this result can be explained in terms of the bloch vector model : we show in the appendix that when @xmath145 , the maximum number of beam atoms that can be scattered into states around @xmath72 is @xmath167 . in practice , the efficiency can exceed this maximum value due to the residual coupling to states around @xmath157 , which enhances the density modulation of the grating .
[ fig7 ] also illustrates the dramatic reduction in efficiency caused by the dephasing of the grating . even in that case , though , the achievable efficiencies indicate that four - wave mixing should still be observable when the fermi gases used for the grating and the beam contain comparable numbers of atoms . as a function of @xmath168 for @xmath169 @xmath170 ( dotted - line ) , @xmath171 ( dashed line ) , and @xmath96 ( dash - dot line ) , with @xmath149 .
the solid line corresponds to maximum efficiency calculated from the bloch vector model .
, width=302,height=188 ] interestingly , one can improve the efficiency significantly by choosing unequal probability amplitudes for the components of the grating , that is , by replacing the initial state @xmath172 of eq .
( [ initialconditions ] ) by the more general superposition state @xmath173 with @xmath174 in our case . fig .
[ fig8 ] plots the bragg efficiency for several values of @xmath175 and @xmath176 , showing a maximum improvement of about @xmath177 as compared to the case of equal initial amplitudes , @xmath178 .
even though @xmath179 produces an initial density grating with the largest possible amplitude , subsequent scattering of beam atoms will reduce the size of the grating . in contrast , while the amplitude of the density grating produced by an unequal initial superposition state is less than optimal initially , the subsequent scattering of beam atoms from @xmath108 to @xmath109 increases @xmath175 while reducing @xmath176 , resulting in a net increase over time in the grating amplitude . as a function of initial probability amplitudes for the grating @xmath180 , with @xmath148 , @xmath181.,width=302,height=226 ]
in this paper we have investigated purely fermionic four - wave mixing in a manner that accounts for the back - action of the scattered beam on the density grating .
our numerical results indicate that the lifetime for fermionic four - wave mixing can be dramatically improved by making the number of atoms in the beam comparable to the number of atoms that form the density grating .
although the effects of the beam on the grating does reduce the overall efficiency of the process , the efficiencies that we have calculated indicate that it should be possible to convert in excess of @xmath182 of the atoms in the incident beam into a fourth scattered beam .
it is well known that four - wave mixing with bosons can lead to the generation of squeezed states @xcite .
therefore a central feature of future work will examine the statistical properties of the scattered beam generated by the four - wave mixing to see if it is possible to generate novel fermionic states .
this will involve going beyond the factorization ansatz in eq .
( [ factorize ] ) to calculate higher order moments for the particle hole operators .
phase conjugation should also be possible if one uses a superfluid fermi gas due to the presence of anomalous moments , @xmath183 , resulting from the formation of cooper pairs .
this work is supported in part by the us office of naval research , by the national science foundation , by the us army research office , by the national aeronautics and space administration , and by the joint services optics program .
h. c. acknowledges the support of the studienstiftung des deutschen volkes .
one of the author ( t. m. ) is supported by jsps of japan .
in this appendix we solve the coupled bloch equations for the beam and grating with @xmath184 . by introducing the normalized vectors @xmath185 we obtain the equations of motion @xmath186 subject to the initial conditions @xmath187 . here
, the @xmath188 represents differentiation with respect to time .
we note that the constants of motion given by ( [ constant1 ] ) and ( [ constant2 ] ) can be expressed as @xmath189 for our particular initial conditions we have @xmath190 for all times .
this allows us to parameterize the solution in terms of the angles @xmath191 and @xmath192 , of the beam and grating bloch vectors relative to the @xmath11-axis in the @xmath193- and @xmath194-planes , respectively , @xmath195 by introducing the sum and difference angles @xmath196 we obtain the uncoupled equations @xmath197 where @xmath198 .
( [ pendulum1 ] ) is the equation of motion for a plane pendulum @xcite .
since the initial conditions for these pendula are @xmath199 and @xmath200 , a small amplitude linearization of the equations of motion is not justified .
however it is still possible to find a solution .
we can express the total `` energy '' for the pendula as @xmath201 where @xmath202 and @xmath203 are the turning points for the pendula . if @xmath204 , then there are no turning points and the pendula undergo complete revolutions about `` their '' axes . in this case
the bloch vector for the beam undergoes complete revolutions in the @xmath193-plane as can be seen from eq .
( [ pop ] ) .
if , on the other hand , @xmath145 , then we have the oscillatory solution @xcite , @xmath205 where @xmath206 is a jacobian elliptic function @xcite and @xmath207 is determined by the initial condition , @xmath208 . in this case the beam bloch vector has a turning point at @xmath209 , so that the maximum number of beam atoms that can be scattered into states around @xmath72 is @xmath167 .
the period of the oscillations for the elliptic function is given by eq .
( [ period ] ) .
99 g. lenz , p. meystre , and e. m. wright , phys .
lett . * 71 * , 3271 ( 1993 ) .
l. deng , e. w. hageley , j. wen , m. trippenbach , y. band , p. s. julienne , j. e. simsarian , k. helmerson , s. l. rolston , and w. d. phillips , nature * 398 * , 218 ( 1999 ) .
s. innouye , t. pfau , a. p. chikkatur , a. grlitz , d. e. pritchard and w. ketterle , nature * 402 * , 641 ( 1999 ) ; m. kozuma , y. suzuki , y. torii , t. sugiura , t. kuga , e. w. hagley , and l. deng , science * 286 * , 2309 ( 1999 ) . s. burger , k. bongs , s. dettmer , w. ertmer , k. sengstock , a. sanpera , g. v. shlyapnikov , and m. lewenstein , phys . rev .
83 * , 5198 ( 1999 ) .
k. e. strecker , g. b. partridge , a. g. truscott , and r. g. hulet , nature * 417 * , 150 ( 2002 ) .
j. m. vogels , k. xu , and w. ketterle , phys .
* 89 * , 020401 ( 2002 ) .
j. m. vogels , j. k. chin , and w. ketterle , cond - mat/0209067 .
r. wynar , r. s. freeland , d. j. han , c. ryu , and d. j. heinzen , science * 287 * , 1016 ( 2000 ) .
c. mckenzie _ et al .
* 88 * , 120403 ( 2002 ) .
a. g. truscott , k. e. strecker , w. i. mcalexander , g. b. partridge , r. g. hule , science * 291 * , 2570 ( 2001 ) ; f. schreck , g. ferrari , k. l. cubizolles , l. khaykovich , m.o .
mewes , and c. salomon , phys . rev .
a * 64 * , 011402 ( 2001 ) .
b. demarco , and d. s. jin , science * 285 * , 1703 ( 1999 ) .
z. hadzibabic , c. a. stan , k. dieckmann , s. gupta , m. w. zwierlein , a. grlitz , and w. ketterle , phys .
lett . * 88 * , 160401 ( 2002 ) .
g. roati , f. riboli , g. mondugno , and m. ingusco , phys .
lett . * 89 * , 150403 ( 2002 ) .
k. m. ohara , s. l. hemmer , m. e. gehm , s. r. granade , and j. e. thomas , science * 298 * , 2179 ( 2002 ) . w. ketterle , and s. inouye , phys . rev
. lett . * 86 * , 4203 ( 2001 ) .
m. g. moore , and p. meystre , phys .
. lett . * 86 * , 4199 ( 2001 ) .
p. villain , p. hberg , l. santos , and m. lewenstein , phys .
a * 64 * , 023606 ( 2001 ) .
h. christ , c. p. search , and p. meystre , cond - mat/0212384 .
b. demarco , and d. s. jin , phys .
a * 58 * , 4267 ( 1998 ) .
et al . _ ,
a * 61 * , 041602 ( 2000 ) .
a. g. rojo , j. l. cohen , and p. r. berman , phys .
a * 60 * , 1482 ( 1999 ) .
kerson huang , `` statistical mechanics '' ( john wiley & sons , new york , 1987 ) .
p. meystre and m. sargent , `` elements of quantum optics '' third edition ( spinger - verlag , berlin , 1999 ) .
h. p. yuen and j. h. shapiro , opt
. lett . * 4 * , 334 ( 1979 ) .
r. e. slusher _ et al .
lett . * 55 * , 2409 ( 1985 ) .
j. b. marion and s. t. thorton , `` classical dynamics of particles and systems , third edition '' ( harcourt brace jovanovich , san deigo , 1988 ) .
e. t. whittaker , `` a treatise on the analytical dynamics of particles and rigid bodies '' ( cambridge university press , cambridge , uk , 1927 ) .
m. r. spiegel , `` mathematical handbook '' ( mcgraw hill , new york , 1997 ) . | we analyze the full nonlinear dynamics of the four - wave mixing between an incident beam of fermions and a fermionic density grating .
we find that when the number of atoms in the beam is comparable to the number of atoms forming the grating , the dephasing of that grating , which normally leads to a decay of its amplitude , is suppressed . instead , the density grating and the beam density exhibit large nonlinear coupled amplitude oscillations . in this case
four - wave mixing can persist for much longer times compared to the case of negligible back - action .
we also evaluate the efficiency of the four - wave mixing and show that it can be enhanced by producing an initial density grating with an amplitude that is less than the maximum value .
these results indicate that efficient four - wave mixing in fermionic alkali gases should be experimentally observable . |
it has long been known that galaxies are shaped by the environment in which they reside .
works by e.g. , @xcite , @xcite , and @xcite showed that galaxies in high - density environments are preferentially red and early - type compared to those in lower density regions .
the more recent advent of large scale photometric and spectroscopic surveys confirmed with large statistics those early findings @xcite .
meanwhile , space and ground based missions have probed the geometry of our universe .
those observations coupled to cosmological models have built the solid lambda cold dark matter ( @xmath6cdm ) framework @xcite , in which lower mass haloes are the building blocks of more massive structures .
one of the major tasks for modern studies of galaxy formation is therefore to understand how and when galaxy evolution is driven by internal processes or the evolving environment that each galaxy experiences during its lifetime .
while internal mechanisms , including ejective feedback from supernovae or active galactic nuclei , are deemed responsible for suppressing star formation in all galaxies @xcite , a galaxy can also directly interact with its environment when falling into a massive , gas- and galaxy - rich structure such as a galaxy cluster . at low redshift detailed studies of poster child objects @xcite coupled with state - of - the - art models and simulations
@xcite have started to explore the rich physics governing those processes ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for reviews ) .
broadly speaking , they can be grouped into two classes .
the first of them includes gravitational interactions between cluster or group members @xcite or with the potential well of the halo as a whole @xcite , or their combined effect known as `` galaxy harrassment '' @xcite .
the second class includes hydrodynamical interactions between galaxies and the hot and dense gas that permeates massive haloes .
this class includes the rapid stripping of the cold gas via ram pressure as the galaxy passes through the hot gas medium @xcite .
ram - pressure stripping is known to effectively and rapidly suppress star formation in cluster galaxies in the local universe @xcite .
less directly influencing the galaxy s current star formation , the multi - phase medium ( e.g. warm , hot gas ) associated to the galaxy ( known as the `` reservoir '' ) should be easier to strip than the cold gas .
even easier , the filamentary accretion onto the galaxy from the surrounding cosmic web will be truncated as the galaxy is enveloped within the hot gas of a more massive halo @xcite .
both of these processes will suppress ongoing accretion onto the cold gas disk of the galaxy and lead to a more gradual suppression of star formation , variously labelled `` strangulation '' or `` starvation '' ( e.g. * ? ? ?
* ; * ? ? ?
* ) these processes are complicated in nature and the exact details of their efficiency and dynamics are still poorly understood .
the situation is even more complicated when several of those processes are found to act together @xcite . a different approach to disentangle the role of environment from the secular evolution is to study large samples of galaxies and correlate their properties ( e.g. star formation activity ) to internal properties ( e.g. stellar mass ) and environment . in the local universe
, the advent of the sloan digital sky survey ( sdss ) has revolutionized the field of large statistical studies and allowed for the effects of the environment on the galaxy population as a whole to be studied @xcite .
one of the main results is that environmental quenching is a separable process that acts on top of the internal processes that regulate the star formation activity of galaxies .
a crucial parameter to understand the collective effect of the several environmental processes is the timescale over which the star formation activity is quenched .
several authors took advantage of excellent statistics to estimate the average timescale for environmental quenching , accounting for internal quenching processes , and found that in the low redshift universe this is generally long ( @xmath7 gyr ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , while possibly shorter in clusters of galaxies ( @xmath8 gyr ; * ? ? ?
* ; * ? ? ?
* ) . at higher redshift ,
the situation is made more complex due to the more limited availability of spectroscopic redshifts which are paramount to depict an accurate picture of the environment . in the last decade ,
several ground based redshift surveys started to address this issue @xcite . by exploiting the multiplexing of spectroscopic instruments at 8 - 10 meter class telescopes ( e.g. vimos and gmos , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , these works showed that the environment plays a role in quenching the star formation activity of galaxies accreted onto massive haloes ( satellite galaxies ) up to @xmath9 @xcite , although the samples are limited to massive galaxies or a small number of objects .
low - resolution space - based slitless spectroscopy is revolutionising this field providing deep and highly complete spectroscopic samples .
the largest of those efforts is the 3d - hst survey @xcite which , by combining a large area , deep grism observations and a wealth of ancillary photometric data , provides accurate redshifts to @xmath10 @xcite for a large sample of objects down to low stellar masses ( @xmath11 , and @xmath12 at @xmath13 and @xmath14 respectively ) .
the public release of their spectroscopic observations @xcite , in synergy with deep photometric observations @xcite has opened the way to an accurate quantification and calibration of the environment over the redshift range @xmath15 .
another source of uncertainty in the interpretation of correlations of galaxy properties with environment is the inhomogeneity of methods used for different surveys ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and the lack of calibration of important parameters such as halo mass . in @xcite , we studied how to link a purely observational parameter space to physical quantities ( e.g. , halo mass , central / satellite status ) by analysing a stellar mass limited sample extracted from semi - analytic models of galaxy formation . to do so , we computed a projected density field in the simulation box and we tested different definitions of density at different redshift accuracy .
our method is bayesian in nature ( galaxies have well - defined observational parameters , while the calibration into physical parameters is probabilistic ) .
this approach is best suited to statistical studies where the application of selection functions and observational uncertainties can be fully taken into account . in this paper
, we extend this method to the 3d - hst survey by building up an environment catalogue which we make available to the community with this work .
we then explore the efficiency and timescales for quenching of satellite galaxies over cosmic time ( @xmath16 ) by combining the 3d - hst data at high redshift with sdss data in the local universe in a homogeneous way .
we also address the long standing issue of impurity and contamination of the calibrated parameters ( the fact that the observations do not perfectly constrain the halo mass of the parent halo for each galaxy or its central / satellite status ) by recovering the `` pure '' trends using the mock sample as a benchmark .
the paper is structured as follows . in section [ obssample ]
, we introduce the 3d - hst dataset . in section [ sec_environment ] ,
we derive the local density for 3d - hst galaxies including accurate edge corrrections .
section [ sec_overdensities ] presents the range of environments in the 3d - hst area and how they compare to known galaxy structures from the literature . in section [ sec_models ] , we introduce the mock galaxy sample and how we calibrate it to match the 3d - hst sample . we then link models and observations in section [ sec_halomass ] , and assign physical quantities to observed galaxies . in section [ sec_quenching ]
we study the quenching of satellite galaxies at @xmath17 , and derive quenching efficiency and timescales .
lastly , we discuss the physical implications of our findings in section [ sec_discussion ] and summarize our work in section [ sec_conclusions ] .
all magnitudes are given in the ab system @xcite and we assume a flat @xmath6cdm universe with @xmath18 , @xmath19 , and @xmath20 unless otherwise specified . throughout the paper
, we use the notation @xmath21 for the base 10 logarithm of @xmath22 .
in this work we aim at a quantification and calibration of the local environment for galaxies in the five candels/3d - hst fields @xcite namely cosmos , goods - s , goods - n , aegis and uds .
the synergy of these two surveys represents the largest effort to obtain deep space - based near - infrared photometry and spectroscopy in those fields . for a description of the observations and reduction techniques , we refer the reader to @xcite and @xcite for the photometry and spectroscopy respectively .
the candels observations provide hst / wfc3 near infrared imaging in the f125w and f160w filters ( @xmath23 and @xmath24 hereafter ) for all the fields , while 3d - hst followed - up a large fraction of this area with the f140w filter ( @xmath25 hereafter ) and the wfc3/g141 grism for slitless spectroscopy .
the novelty of this approach is to obtain low resolution ( @xmath26 ) spectroscopy for all the objects in the field .
taking advantage of the low background of the _ hst _ telescope , it is possible to reach a depth similar to traditional slit spectroscopy from 10 m class telescopes on earth .
hereafter , we use the term `` 3d - hst '' sample to refer to the combination of candels and all the other space- and ground - based imaging datasets presented in @xcite , plus the grism spectroscopy of the 3d - hst program .
the 3d - hst photometric catalog @xcite used @xmath24 or @xmath25 as detection bands and its depth varies from field to field and across the same field due to the observing strategy of candels .
however , even in the shallowest portions of each field , the 90@xmath27 depth confidence level is @xmath28 mag . beyond this magnitude limit
, the star / galaxy classification ( which is a key parameter for the environment quantification ) becomes uncertain .
the 3d - hst spectroscopic release @xcite provides reduced and extracted spectra down to @xmath29 mag .
the spectra are passed through the eazy template fitting code @xcite along with the extensive ground- and space - based multiwavelength photometry .
this results in `` grism '' redshifts , which are more accurate than photometric redshifts thanks to the wealth of stellar continuum and emission line features present in the spectra .
however , only objects brighter than @xmath30 mag have been visually inspected , and have a ` use_grism ` flag that describes if the grism spectrum is used to compute the redshift .
incomplete masking of contaminating flux from nearby sources in the direction of the grism dispersion , residuals from spectra of bright stars , and corrupted photometric measurements can lead to this flag being set to 0 ( `` bad '' ) .
we include in the present analysis all galaxies brighter than @xmath31 mag , therefore limiting our footprint to the regions covered by grism and @xmath25 observations .
we limit the redshift range to @xmath32 .
the lower limit roughly corresponds to the redshift where the @xmath33 line enters the grism coverage and the upper limit is chosen such that the number density of objects above the magnitude cut allows a reliable estimate of the environment .
it also allows follow - up studies targeting the rest - frame optical features from ground based facilities in the @xmath34 , @xmath35 , and @xmath36 bands ( e.g. @xmath37 , @xcite and mosfire , @xcite ) we exclude stars by requiring ` star_flag ` to be 0 or 2 ( galaxies or uncertain classification )
. we do not use the ` use_phot ` flag because it is too conservative for our goals .
indeed this flag requires a minimum of 2 exposures in the f125w and f160w filters , and the object not being close to bright stars .
the quantification of environment requires a catalog which is as complete as possible even at the expenses of more uncertain photometry ( and photo - z ) for the objects that do not meet those cuts .
nonetheless a @xmath38 mag cut allows a reliable star / galaxy separation for 99% of the objects and is at least 1 mag brighter than the minimum depth of the mosaics , thus alleviating the negative effects of nearby stars on faint sources .
the final sample is made of 18745 galaxies . as a result of the analysis in @xcite , each galaxy
is assigned a `` best '' redshift .
this is : 1 . a spectroscopic redshift from a ladder of sources as described below . 2
a grism redshift if there is no spectroscopic redshift and ` use_grism = 1 ` 3 . a pure photometric redshift if there is no spectroscopic redshift and ` use_grism = 0 ` .
a ` zbest_type ` flag is assigned to each galaxy based on the conditions above .
the best redshift is the quantity used to compute the environment for each galaxy in the 3d - hst fields .
spectroscopic redshifts are taken from the compilation of @xcite which we complement with newer data .
for the cosmos field we include the final data release of the zcosmos bright survey @xcite .
we find 253 new sources with reliable redshifts in the 3d - hst / cosmos footprint mainly at @xmath39 . in cosmos and goods - s , we include 95 objects from the dr1 @xcite of the vimos ultra deep survey ( * ? ? ?
* vuds , [ ) .
this survey mainly targets galaxies at @xmath40 therefore complementing zcosmos .
we include 105 redshifts from the mosfire deep evolution field survey ( mosdef , * ? ? ?
* ) which provides deep rest frame optical spectra of galaxies selected from 3d - hst .
for the uds field , we also include 164 redshifts from vimos spectroscopy in a narrow slice of redshift ( @xmath41 , galametz et al . in prep . )
lastly , we include 376 and 33 secure spectroscopic redshifts from @xmath42 @xcite and virial @xcite respectively .
those large surveys use the multiplexing capability of the integral field spectrometer kmos on the eso very large telescope to follow - up 3d - hst selected objects .
the former is a mass selected survey of emission line galaxies at @xmath43 , while the latter observed passive massive galaxies at @xmath44 . in the selected sample ,
@xmath45 of the galaxies have a spectroscopic redshift , @xmath46 have a grism redshift , and only @xmath47 have a pure photometric redshift . in the next section ,
we explore the accuracy of the grism and photometric redshifts as a function of the galaxy brightness and the @xmath48 of emission lines in the spectra .
stellar masses and stellar population parameters are estimated using the fast code @xcite , coupled with @xcite stellar population synthesis models .
those models use a @xcite initial mass function ( imf ) and solar metallicity .
the best redshift is used for each galaxy together with the available space- and ground - based photometry .
the star formation history is parametrized by an exponentially declining function and the @xcite dust attenuation law is adopted .
a careful quantification of the grism and photometric redshift accuracy is paramount for a good calibration of the environmental statistics into physically motivated halo masses . in section [ sec_halomass ]
, we will show how these masses are obtained from mock catalogues selected to match the number density and redshift uncertainty of 3d - hst galaxies .
the low - resolution spectra cover different spectral features as a function of galaxy properties and redshift .
the most prominent features are emission lines , which are however limited to star forming objects . on the other hand ,
stellar continuum features ( balmer break , absorption lines ) are present in the spectra of all galaxies with a @xmath48 that depends on the galaxy magnitude . because all those features contribute to the redshift fitting procedure , we explore their impact on the redshift accuracy in bins of @xmath48 of the strongest emission line in the spectrum and @xmath25 total magnitude .
given the limited spectral coverage of the g141 grism , it is common to find only one prominent emission line feature in the spectrum @xcite ; this justifies our approach of using the @xmath48 of the strongest line .
we define the redshift accuracy ( @xmath49 ) as half the separation of the 16@xmath50 and 84@xmath50 confidence levels obtained from the probability density function ( pdf ) of grism redshifts as derived from the eazy template fitting procedure . in the case of fits
obtained without including the spectral information , it becomes a pure photometric redshift uncertainty .
a comparison of grism redshifts to spectroscopic redshifts shows that @xmath51 should be added to the formal uncertainty on the grism redshifts to obtain a scatter in @xmath52 with a @xmath53 width of unity .
this `` intrinsic grism '' uncertainty can arise from morphological effects , i.e. the light - weighted centroid of the gas emission can be offset from that of the stars ( see * ? ? ?
* ; * ? ? ?
* ) . in this analysis
, we added the intrinsic uncertainty of the grism data in quadrature to the formal uncertainty from the fitting process .
figure [ fig1 ] shows @xmath49 , for bins of emission line @xmath48 and @xmath25 magnitude .
the bottom right panel ( f ) shows the accuracy of photometric redshifts for the same sources highlighting the significant improvement on the redshift quality when the spectra are included . from the top panels of figure [ fig1 ] ,
it is clear how an emission line detection narrows the redshift pdf to the intrinsic uncertainty , irrespective of the stellar continuum features . at @xmath48 where the emission line becomes less dominant ,
we start to witness a magnitude dependence of the redshift accuracy .
brighter galaxies have better continuum detections and therefore a more accurate redshift . even when there is no line detection ( panel e ) , the typical redshift uncertainties are a factor 2 - 3 lower than pure photometric redshifts .
the inclusion of the spectra helps the determination of the redshifts even when the spectra are apparently featureless .
the grism redshift accuracy is comparable to the pure photometric redshift accuracy only for the faintest objects ( @xmath25@xmath54 23 mag ) with no emission line detection ( @xmath55 ) , a population which accounts for @xmath56 of our grism sample . as a final note of caution
, we highlight that whenever the information in the spectra is limited , the final grism redshift accuracy depends largely on the photometric data , whose availability depends on the field .
indeed , cosmos and goods - s have been extensively observed with narrow or medium band filters @xcite resulting in better photometric redshifts compared to the other fields . however , as shown in section [ sec_halomass ] these field - to - field variations have negligible effects on our calibration of halo mass .
there are many ways to describe the environment in which a galaxy lives ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in this work , we apply to observational data the method we explored and calibrated in @xcite and based on the work of @xcite .
we use the number density of neighbouring galaxies within fixed cylindrical apertures because it is more sensitive to high overdensities , less biased by the viewing angle , more robust across cosmic times , and easier to physically interpret and calibrate than the n@xmath50 nearest neighbour methods @xcite .
we consider all 3d - hst galaxies selected in section [ obssample ] to be part both of the primary ( galaxies for which the density is computed ) and neighbour samples .
we calculate the projected density @xmath57 in a combination of circular apertures centered on the primary galaxies with radii @xmath58 .
the apertures range from 0.25 to 1.00 mpc in order to cover from intra - halo to super - halo scales . for a given annulus defined by @xmath58 ,
the projected density is given by @xmath59 where @xmath60 is the sum of the weights of galaxies in the neighbour sample falling at a projected distance on the sky @xmath61 from the primary galaxy and within a relative rest - frame velocity @xmath62 .
for the 3d - hst galaxies with a grism or spectroscopic redshift , the weights are set to unity ( non weighted sum ) , while for galaxies with pure photometric redshifts , we apply a statistical correction for the less accurate redshifts as described in section [ edgecor ] . the primary galaxy is not included in the sum therefore isolated galaxies have @xmath63 .
we set the velocity cut at @xmath64 .
this value is deemed appropriate for surveys with complete spectroscopic redshift coverage @xcite and for 3d - hst given the quality of grism redshifts shown in figure [ fig1 ] .
a small value of @xmath65 avoids the peaks in the environmental density to be smoothed by interlopers in projection along the redshift axis . on the other hand ,
if only less accurate redshifts are available , a larger cut must be used to collect all the signal from overdense regions which is artificially dispersed along the redshift axis ( see figure 4 in * ? ? ?
* ; * ? ? ?
* ) .
because the mean number density changes continuously with redshift , it is not possible to compare the local density ( @xmath66 ) across time .
instead we define a relative overdensity @xmath67 , which is given by : @xmath68 where @xmath69 is the average surface density of galaxies at a given redshift .
this is obtained by computing the volume density of galaxies ( per mpc@xmath70 ) in the whole survey and parametrising the redshift dependence with a third degree polynomial .
this value is multiplied by the depth of the cylindrical aperture at redshift @xmath71 to obtain the surface density @xmath69 . throughout the paper , we will mainly use the overdensity in terms of the logarithmic density contrast defined as @xmath72 .
the calculation of the environment of primary galaxies at the edges of the 3d - hst footprint ( see figure [ figfootprints ] , red areas ) suffers from incomplete coverage of neighbours that results into an underestimated density in the considered aperture .
in large scale surveys ( e.g. sdss , * ? ? ?
* ) , it is common practice to remove galaxies too close to the edges of the observed field . in the case of deep fields , however , the observed area is relatively small and the removal of such galaxies would reduce total number of objects significantly .
one possible solution is to normalize the densities by the area of the circular aperture which is within the survey footprint in equation [ eqsigma ] .
although this is a simple choice , it assumes a constant density field and neglects possible overdense structures just beyond the observed field .
a more accurate solution consists of building up galaxy catalogues for a more extended area than 3d - hst and then use galaxies within these areas as `` pure neighbours '' for the environment of the primary 3d - hst galaxies .
given the amount of publicly available data , this is possible in goods - s , cosmos and uds ( see figure [ figfootprints ] , blue areas ) . in appendix [ edgecor_cats ]
we describe the data , depth and redshift quality of the catalogues we built in those fields . here
we present the edge correction method we developed and how it was tuned to perform the edge corrections in the other two fields goods - n and aegis .
the availability of spectroscopic redshifts in the extended area catalogs is limited ( from @xmath73 in cosmos and uds to @xmath74 in goods - s ) .
we thus need to deal with the limited accuracy of photometric redshifts for the galaxies in those fields .
the photo - z accuracy , which varies from field to field and depends on the redshift , brightness and color of the objects @xcite , is such that most of the sources which are part of the same halo in real space would not be counted as neighbours of a primary galaxy , simply due to the redshift uncertainty .
@xcite show that increasing the depth of the velocity window would recover most of the real neighbours but at the expense of a larger fraction of interlopers ( galaxies which are not physically associated to the primary ) . here , we thus exploit a different method .
we assume that galaxies which are at small angular separation and whose redshifts are consistent within the uncertainties are , with a high probability , physically associated ( e.g. , * ? ? ?
* ; * ? ? ?
if one of them has a secure spectroscopic redshift , we assign this to the others .
our method works as follows : * for each galaxy with a photometric redshift , we select all neighbours with a redshift within @xmath75 .
this value is chosen to recover most of the real neighbours given the average photo - z uncertainties .
* among those neighbours , we select the closest ( in spatial coordinates ) which has a secure spectroscopic ( or grism ) redshift . here
we assume grism redshifts to have a negligible uncertainty compared to photo - z . *
we replace the photo - z of the galaxy of interest with this spec - z ( or grism - z ) .
since the statistical validity of the assumption of physical association depends on the distance of the neighbour , for increasing distances we underestimate the true clustering .
we correct for the bias by assigning a weight @xmath76 to each galaxy .
the weight is evaluated on a training sample made of galaxies in 3d - hst with @xmath77 mag . for each galaxy in the three fields , we compute the `` real '' density ( @xmath78 in a 0.75mpc radius and @xmath79 ) using spec - z or grism - z from 3d - hst .
we then take for each galaxy its photometric redshift , and follow the procedure described above , but , instead of choosing the closest neighbour with a secure redshift , we select a random neighbour in different bins of projected sky distance ( from 0 to 3 mpc in bins of 0.5 mpc width ) .
then we compute densities with each of those distance replacements separately and the fractional bias ( @xmath80 ) as : @xmath81 where the @xmath82 subscript denotes the replacement with a spec - z of a galaxy found at distance @xmath82 . by using the 3d - hst data ,
we make sure that there are always a large number of neighbours with a secure redshift , and we repeat this procedure 1000 times in order to uniformly sample the neighbours . figure [ figedgecor ] left panel shows @xmath80 as a function of the real density in four bins of @xmath82 . clearly , the larger @xmath82 is , the more underestimated the real density will be , due to a decreasing fraction of correct redshift assignments .
we then derive the median weight @xmath83 where the median is computed among all galaxies that have @xmath84 ( see the vertical dashed line in figure [ figedgecor ] left panel ) .
the density dependence of @xmath80 is negligible at these densities , therefore by avoiding underdense regions ( where the uncertainty on @xmath80 is large ) we obtain a robust determination of @xmath85 .
figure [ figedgecor ] middle panel shows @xmath85 versus @xmath82 , which we fit with a quadratic relation obtaining : @xmath86 with the additional constraint that @xmath87 which corresponds to @xmath88 for @xmath89 mpc .
we tested that this relation , although obtained combining all fields , holds within the uncertainties when each field is considered separately .
lastly we show in figure [ figedgecor ] right panel how the systematic bias is removed when the weight is applied to all neighbours when computing the density .
this is consistent with no bias within the uncertainties for all the distance bins .
the goods - n and aegis fields do not have deep and extended near - infrared public catalogues that can be used to derive the edge corrections as presented above . as shown in figure [ figfootprints ] ( light blue shaded areas ) the 3d - hst / candels footprint slightly extends beyond the area covered by g141 grism observations ( the main requirement for our primary sample ) .
therefore the 3d - hst / candels catalogue itself can be used to perform edge corrections .
we derive @xmath25 magnitudes from the @xmath23 magnitudes using a linear function derived from the five 3d - hst fields ( @xmath90 ) .
we then use 3d - hst photometric redshifts ( or spec - z where available ) and apply the method described in section [ edgecormethod ] . however , the 3d - hst / candels photometric catalogues do not extend enough beyond the primary sample area to ensure the apertures used to compute the density are entirely covered by the photometric catalog footprint . for this reason ,
we compute the densities using the area of the circular aperture within the photometric catalogue .
we test this method by comparing the density ( @xmath78 ) in a 0.75mpc aperture measured using the extended catalogues for cosmos , goods - s , uds and the density ( @xmath66 ) measured correcting for the fraction of the aperture ( @xmath91 ) in the 3d - hst / candels footprint .
the result is shown in figure [ edgecor_frac ] .
we note that although the median ( red solid line ) is consistent with no bias , the area correction introduces a scatter ( dotted and dashed lines ) which increases by decreasing the fraction of the aperture in the footprint .
in conclusion , the environment catalogue released with this work includes all the primary galaxies in the five 3d - hst fields .
the structure of the catalogue is described in appendix [ envcatalogue ] . however , in the rest of this work we only include galaxies for which @xmath92 for the goods - n and aegis fields . the total number of objects in the primary 3d - hst sample with a robust determination of the environmental density is therefore reduced to 17397 ( @xmath93 of the original sample ) . ) in a 0.75mpc aperture measured using the extended catalogues for cosmos , goods - s , uds and the density ( @xmath66 ) measured using the fraction of the aperture in the 3d - hst footprint ( @xmath91 ) as a function of the latter quantity .
the solid line is the median while dashed and dotted lines mark the 1@xmath94 and 2@xmath94 confidence intervals respectively .
the offset between the two methods is zero , with a scatter which increases with decreasing fraction of the aperture in the 3d - hst footprint .
, width=302 ]
in order to explore correlations of galaxy evolution with environment , we need to make sure the 3d - hst fields span a wide range of galaxy ( over-)densities , and use known structures as a sanity check of our density estimates .
figures [ densitygoodss ] , [ densitycosmos ] , [ densityuds ] , [ densityaegis ] , and [ densitygn ] present the primary sample of 3d - hst galaxies in the five fields color coded by their overdensity in the 0.75 mpc aperture in different redshift slices .
this aperture corresponds to the typical virial radius of massive haloes ( @xmath95 ) in the redshift range under study .
the range of density probed is wide and spans from isolated galaxies to objects for which the local number of neighbours is up to ten times larger than the mean at that redshift , reaching the regime of clusters or massive groups . in each figure , we overplot the position and extent of x - ray extended emission from the hot intragroup ( and intracluster ) medium that fills massive haloes . the exquisite depth of x - ray data in the deep fields @xcite allows the detection of the hot gas from haloes down to @xmath96 .
we find a very satisfactory agreement between our overdensities and the x - ray emission position . indeed , most of the x - ray groups are coincident with large overdensities in our maps . on the other hand ,
not all the overdense structures identified in our work are detected in x - ray .
we speculate this is mainly due to the presence of more than one massive structure along the line of sight or that low mass groups may not yet be virialized .
lastly , we note that the redshift of the x - ray emission is assigned based on the photometric or spectroscopic information available at the epoch of the publication of the catalogue ; these data might not have been as accurate as the density field reconstruction performed in this work .
our analysis has therefore the potential to spectroscopically confirm more x - ray groups and improve the quality of previous redshift assignments .
several other works have also analysed , with different techniques , the presence of overdense structures in the deep fields .
we overplot on figure [ densitygoodss ] the position of overdensities found in the goods - s field by @xcite .
these have been derived from the goods - music catalogue using a smoothed 3d density searching algorithm .
these data have 15% spectroscopic redshifts and photometric redshifts for the remaining fraction . because the smoothing technique is less able to constrain the size of the structure , we plot circles with an arbitrary radius .
the structures within the 3d - hst footprint ( except those at @xmath40 ) are confirmed with our data to be at least a factor of @xmath97 denser than the mean .
the differences in samples and techniques hamper a more quantitative comparison .
our data confirm with a high degree of significance the detection of two well known super - structures , one at redshift @xmath98 @xcite and one at redshift @xmath99 first detected by @xcite .
the latter is made of 5 peaks in the photo - z map ( which correspond to putative positions for the x - ray emission , see table 1 in @xcite ) .
the main structure is robustly recovered by our analysis while the other sub - structures are only mild ( @xmath100 ) overdensities . in the cosmos field
( see figure [ densitycosmos ] ) , @xcite applied an adaptive smoothing technique ( similar to * ? ? ?
* ) to find large scale structures at @xmath39 . while their results do not constrain the size of the structure and are less sensitive to very compact overdensities , we do find that their detections in the 3d - hst footprint correspond to high overdensities in our work .
similarly in the uds field , we do detect a very massive cluster surrounded by filaments and less massive groups ( upper left panel of figure [ densityuds ] ) at @xmath101 ( galametz et al . in prep . ) .
another well known structure in this field is located at @xmath102 @xcite . despite being
only partially covered by the 3d - hst grism observations ( isolated pointing on the left of the contiguous field ) , we do find it corresponds to a large overdensity of galaxies thanks to our accurate edge corrections using ukidss - uds photometric data . in summary ,
our reconstruction of the density field in the 3d - hst deep fields recovers the previously known massive structures across the full redshift range analysed in this work .
the goal of this work is to understand the environment of galaxies in the context of a hierarchical universe . to reach this goal , we need to calibrate physically motivated quantities using observed metrics of environment by means of semi - analytic models ( sam ) of galaxy formation .
we make use of light cones from the latest release of the munich model presented by @xcite .
this model is based on the millennium n - body simulation @xcite which has a size of @xmath103 mpc .
the simulation outputs are scaled to cosmological initial conditions from the _ planck _ mission ( planck collaboration xvi , @xcite ) : @xmath104 .
although those values are slightly different from those used in our observational sample , the differences in cosmological parameters have a much smaller effect on mock galaxy properties than the uncertainties in galaxy formation physics @xcite .
this model includes prescriptions for gas cooling , size evolution , star formation , stellar and active galactic nuclei feedback and metal enrichment as described by e.g. @xcite .
the most significant updates concern the reincorporation timescales of galactic wind ejecta that , together with other tweaks in the free parameters , reproduce observational data on the abundance and color distributions of galaxies from @xmath2 to @xmath105 @xcite .
our choice of this model is therefore driven by those new features which are critical for an accurate quantification of the environment .
we make use of the model in the form of 24 light cones , which are constructed by replicating the simulation box evaluated at multiple redshift snapshots . before deriving the density for the light cones ,
as described in section [ subsec_density ] , we first match the magnitude selection and redshift accuracy of the 3d - hst survey .
sams are based on n - body dark matter only simulations .
therefore ( and opposite to observations ) , the galaxy stellar masses are accurate quantities , while observed magnitudes are uncertain and rely on radiative transfer and dust absorption recipes implemented in the models . on the other hand , magnitudes are direct observables in a survey ( like 3d - hst ) are therefore known with a high degree of accuracy . to overcome these limitations and the fact that @xmath25 magnitudes are not given in @xcite cones ,
we employ a method that generates observed magnitudes for sam galaxies by using observational constraints from 3d - hst .
each model galaxy is defined by its stellar mass ( @xmath106 ) , @xmath107 rest frame color ( @xmath108 ) and redshift ( @xmath109 ) .
similarly , 3d - hst galaxies are defined by stellar mass ( @xmath110 ) , @xmath107 rest frame color ( @xmath111 ) , redshift ( @xmath112 ) , and magnitude ( @xmath113 ) .
the method works as follows : * for each bin of stellar mass ( 0.25 dex wide ) and redshift ( 0.1 wide ) we select all the model and 3d - hst galaxies . * for each model galaxy in this bin we rank the @xmath108 and we find the @xmath111 that corresponds to the same ranking .
* we assign to the model galaxy a randomly selected stellar mass - to - light ratio @xmath114 at @xmath115 drawn from the distribution of 3d - hst galaxies in the stellar mass and redshift bin of the mock galaxy of interest . * from @xmath114 , @xmath106 , and @xmath109 , we compute @xmath116 for the model galaxy .
this method generates @xmath25 magnitudes for all the model galaxies down to @xmath117 .
this is much deeper than the 3d - hst magnitude limit even at the lower end of our redshift range .
we then select model galaxies down to a @xmath118 magnitude that matches the total number density of the primary targets ( @xmath25@xmath119 mag ) in the five 3d - hst fields to that in the 24 lightcones .
this protects us from stellar mass function mismatches between the models and the observations ( although those differences are very small in @xcite ) .
we employ a @xmath120 mag , which is very close to 24 mag further supporting the quality of the stellar mass functions in the models . after the model sample
is selected , the next goal is to assign to each galaxy a redshift accuracy that matches as closely as possible to the one in 3d - hst .
to do so , we should not only assign the correct fraction of spec - z , grism - z and photo - z as a function of observed magnitude but also assign an accuracy for the grism - z and photo - z as a function of physical properties such that the final distributions resemble those in figure [ fig1 ] .
we showed in section [ 3dhstzaccuracy ] that the grism redshift accuracy depends on the signal - to - noise of the strongest emission line and the galaxy magnitude .
for the latter , we use @xmath25 as derived above , while the former quantity needs to be parametrized in terms of other quantities available in the models . for 3d - hst galaxies as a function of the @xmath25 magnitude and line magnitude .
right panel : emission line @xmath48 obtained using the parametrization from equation [ eqlinesn].,width=340 ] figure [ sngrism ] left panel shows how the emission line @xmath48 depends both on the line flux and the @xmath25 magnitude for galaxies with a measured line flux . for each galaxy , we take the flux ( in units of erg @xmath121 s@xmath122 ) of the strongest line and we define the line magnitude as @xmath123 . at fixed line flux ,
brighter galaxies have more continuum , thus decreasing the line @xmath48 .
this relation is well reproduced by the following parametrization : @xmath124 figure [ sngrism ] right panel shows the line @xmath48 obtained with this equation .
the small differences between the two panels can be due to additional variables not taken into account ( e.g. dust extinction or grism throughput ) .
we tested ( by perturbing the @xmath48 assigned to each model galaxy ) that a more accurate parametrization of this relation is not required for the purpose of this paper . in order to obtain a synthetic line @xmath125 for the model galaxies ,
we first convert the star formation rate ( sfr ) of the model galaxies into an @xmath126 flux ( or @xmath127 flux where @xmath126 is redshifted outside the grism wavelength range ) by inverting the relation given in @xcite .
we then obtain @xmath128 from @xmath129 , and @xmath116 using equation [ eqlinesn ] .
the rank in @xmath125 ( and not the absolute value ) is then matched to that in @xmath48 for the 3d - hst galaxies .
lastly , we assign to each mock galaxy a random grism redshift accuracy such that the observed distributions shown in figure [ fig1 ] are reproduced for the mock sample .
a photometric redshift accuracy is also generated using the same distributions ( as a sole function of @xmath116 ) . each model galaxy
is then defined by three redshifts : a spectroscopic redshift which is derived from the geometric redshift ( @xmath130 ) of the cones plus the peculiar velocity of the halo , a grism like redshift which is derived from the spec - z plus a random value drawn from a gaussian distribution with sigma equal to the grism redshift accuracy derived above , and a photometric redshift derived as the previous but using the photometric redshift accuracy .
the last step in this procedure requires that for each galaxy only one of these three redshifts is selected to generate a `` best '' redshift .
to do so , we work in bins of @xmath25 magnitude . for each bin of magnitude
, the fraction of 3d - hst galaxies with spec - z , grism - z and photo - z is computed . then in order of descending @xmath125 ,
the spec - z is taken for a number of galaxies matching the fraction of galaxies with spec - z in the observational catalog , a grism - z is taken for an appropriate number of galaxies and lastly a photo - z is taken for the galaxies with the lowest @xmath125 which mimic line non - detections in the grism data .
we stress that since the grism redshift accuracy is a function of @xmath125 , the quality of grism redshifts for objects with marginal line detections is preserved by this method .
once a catalog of model galaxies is selected and their redshift accuracy matches the 3d - hst catalog , we compute the environment parameters as described in section [ subsec_density ] .
the only minor difference is that , as the number of model galaxies is very large , we can remove objects closer than 1.0 mpc from the edges of the cone , to avoid edge biases .
the local density of galaxies is not the only parameter that describes the environment in which a galaxy lives .
another important parameter is whether a galaxy is the dominant one within its dark matter halo ( central ) , or if it orbits within a deeper potential well ( satellite ) .
the definition of centrals and satellites in the mock sample is obtained from the hierarchy of subhaloes ( the main units hosting a single galaxy ) .
first , haloes are detected using a friends - of - friends ( fof ) algorithm with a linking length @xmath131 @xcite .
then each halo is decomposed into subhaloes running the algorithm subfind @xcite , which determines the self - bound structures within the halo .
as time goes by , the model follows subhaloes after they are accreted on to larger structures .
when two haloes merge , the galaxy hosted in the more massive halo is considered the central , and the other becomes a satellite . in this section , we describe how we use the mock catalog to assign a halo mass probability density function ( pdf ) and a probability of being central or satellite to 3d - hst galaxies .
the method builds on the idea of finding all the galaxies in the mock lightcones that match each 3d - hst galaxy in redshift , density , mass - rank ( described below ) , and stellar mass ( within the observational uncertainties ) .
the main advantage of using multiple parameters is to break degeneracies which are otherwise dominant if only one parameter is used ( e.g. to account for the role of stellar mass at low density , where halo mass depends more significantly on stellar mass than density ; * ? ? ?
@xcite explored how the rank in stellar mass of a galaxy in an appropriate aperture can be a good discriminator of the central / satellite status for a galaxy .
this method , which complements the one usually used in local large scale surveys of galaxies based on halo finder algorithms , is more effective with the sparser sampling of high redshift surveys .
we refer the reader to @xcite for the details of how this method is calibrated . here , we recall that we define a galaxy to be central if it is the most massive ( mass - rank = 1 ) within an adaptive aperture that only depends on the stellar mass . otherwise , if it is not the most massive ( mass - rank @xmath132 ) ,
it is classified as a satellite .
the adaptive aperture is motivated by the fact that ideally , the aperture in which the mass - rank is computed should be as similar as possible to the halo virial radius to maximize the completeness of the central / satellite separation and reduce the fraction of spurious classifications .
@xcite , defined this aperture as a cylinder with radius : @xmath133}\ ] ] where @xmath134 is the stellar mass , @xmath135 , and @xmath136 are the parameters which describe the dependence of the virial radius with stellar mass .
these values are calibrated using the models ( see * ? ? ?
we also limit the aperture between 0.35 and 1.00 mpc .
the lower limit is set to avoid small apertures which would result in low mass galaxies being assigned mass - rank = 1 even if they are satellites of a large halo .
the upper limit is approximately the radius of the largest haloes in the redshift range under study .
the adaptive aperture radius ( in mpc ) is therefore defined as : @xmath137 in this work , we have to consider the variable redshift accuracy of 3d - hst galaxies . therefore fixing the depth of the cylinder to @xmath138
does not optimize the central versus satellite discrimination .
we set the depth of the adaptive aperture cylinder ( in @xmath139 ) to : @xmath140 where @xmath49 is the redshift accuracy of the primary galaxy . by using the mock sample
, we tested that this combination of upper and lower limits gives a pure yet sufficiently complete sample of central galaxies .
the simple classification of centrals and satellites based on mass - rank only is subject to a variety of contaminating factors .
for instance in galaxy pairs or small groups ( where the mass of the real central and satellites are very close ) , it is difficult to use the stellar mass to robustly define which galaxy is the central . on the other hand , in the infalling regions beyond the virial radius of massive clusters , many central galaxies would be classified as satellites as analysed in detail in @xcite . in this work ,
we go beyond the simple dichotomic definition that each galaxy is either central or satellite using the mass - rank only .
we combine multiple observables to derive a probability that each 3d - hst galaxy is central or satellite by matching observed galaxies to mock galaxies .
this probabilistic approach naturally takes into account all sources of impurity and is of fundamental importance to separate the effects of mass and environment on the quenching of galaxies . in this section , we describe how we match individual 3d - hst galaxies to the mock sample to access physical quantities unaccessible from observations only .
our method heavily relies on the fact that the distributions of stellar mass and density ( and their bivariate distribution ) are well matched between the mocks and the observations across the full redshift range . in the upper and right panels of figure [ mstar_dens_compare ]
, we show the distributions of density and stellar mass respectively , while the main panel shows the 2d histogram of both quantities .
the overall agreement is very satisfactory and relates to the agreement of the observed stellar mass functions to that from @xcite , and to our careful selection of objects .
the match of the density distributions also confirms that the redshift assignment for mock galaxies is accurate enough to reproduce the observed density distributions . a good match between models and observations
is found for other apertures as well . in the future , it should be possible to improve our method by combining density information on several scales by means of machine learning algorithms .
to match observed galaxies to mock galaxies , we also require an estimate of the uncertainty on both the density and the stellar mass .
for the stellar mass , we use @xmath141 dex @xcite . for the density ,
the error budget is dominated by the redshift uncertainty of each galaxy and the fact that for a sample of galaxies with given @xmath25 and emission line @xmath48 , the redshift accuracy has a distribution with non zero width .
this means that the redshift uncertainty of mock galaxies can only match the observational sample in a statistical sense . to test how the densities of individual galaxies are affected by the redshift uncertainty , we repeat 50 times the process of assigning a redshift to mock galaxies described in section [ subsec_redshiftaccuracy ] .
we then compute the density for each of those samples independently and analyse the distribution of densities for each galaxy .
we find that the distribution roughly follows a poissonian distribution : @xmath142 based on this evidence , we match each 3d - hst galaxy to the mock galaxies within @xmath143 in redshift space and within @xmath144 and @xmath145 for the stellar mass and density on the 0.75 mpc scale respectively .
the local density is a quantity that depends on the redshift accuracy both of the primary galaxy and of the neighbours , which in turn depends on the emission line strength in the grism data and the galaxy brightness ( see section [ 3dhstzaccuracy ] ) . as a result
the density peaks are subject to different degrees of smoothing if the neighbouring galaxies have a systematically poorer redshift accuracy in a given environment .
our mock catalogue is a good representation of the observational sample only if the sfr ( from which the syntetic line @xmath48 is derived ) and the stellar mass distributions as a function of environment are well reproduced by the sam . @xcite
have shown that the h15 model is qualitatively able to recover the observed trends of passive fraction as a function of environment . by matching model galaxies with a redshift accuracy within @xmath146 to that of the observed galaxy
we introduce no bias in the halo mass distributions : for galaxies with less accurate redshifts , we simply obtain broader pdfs of halo mass .
lastly , we restrict the match for the most massive galaxies ( mass - rank @xmath147 ) to the most massive mock galaxies .
the rest of the population ( mass - rank @xmath148 ) was matched to the same population in the mocks .
the central and satellite fractions of those matched mock galaxies are used to define a probability that the 3d - hst galaxy under consideration is central ( @xmath149 ) or satellite ( @xmath150 ) : @xmath151 figure [ pcenpsat ] shows the average values of those quantities in bins of logarithmic density contrast ( see section [ subsec_density ] ) in the 0.75 mpc aperture and stellar mass for all the 3d - hst galaxies included in our sample .
the average value of @xmath149 decreases with increasing density and decreasing stellar mass , and the opposite trend occurs for @xmath150 .
high mass haloes ( high density regions ) are indeed dominated by the satellite population , but objects with high stellar masses are more likely to be centrals .
galaxies in low density environments ( @xmath152 ) are almost entirely centrals . however , in the analysis performed in the next sections we use the values of @xmath149 and @xmath150 computed for each galaxy instead of the average values ( @xcite performs instead an average correction as a function of galaxy density ) .
this takes into full account possible second order dependencies on mass - rank , redshift , or redshift accuracy .
we also examine how @xmath150 varies as a function of the distance from the center of over - dense structures , such as massive groups or clusters of galaxies .
to do so , we take the haloes more massive than @xmath153 in the mock lightcones .
we then select all galaxies in a redshift slice centered on the redshift of the central galaxy and within @xmath154 and compute their projected sky positions with respect to the central galaxy .
we normalize their positions to the virial radius of the halo and remove the central galaxy .
figure [ psat_2dstack ] , top panel , shows the average value of @xmath150 as a function of normalized r.a . and dec . offset from the center of the haloes .
the black solid circles mark @xmath155 and @xmath156 .
figure [ psat_2dstack ] , bottom panel , shows the average value of @xmath150 ( black solid line ) as a function of radial distance from the center of the haloes .
the red solid line shows the fraction of satellites in the same radial bins but using the mock definition of satellites .
lastly , the red dashed line shows the value of @xmath157 including only sam satellites living in the same halo of the central galaxy .
our bayesian definition tracks well the sam definition of satellites as a function of halo mass .
however the real trend is smoothed due to both the transformation from real to redshift space , and the intrinsic uncertainty of our method to extract @xmath150 based on observational parameters .
moreover , @xmath150 only drops to @xmath158 at @xmath159 .
this is caused by satellites from nearby haloes , while the contribution from satellites belonging to the same halo becomes negligible at @xmath160 .
indeed , massive structures are embedded in filaments and surrounded by groups which will eventually merge with the cluster .
therefore , even at large distances from the center , the density is higher than the mean density ( at @xmath159 the density is @xmath161 times higher than the average density ) . as a reference we show in figure [ psat_2dstack ] , bottom panel , the value of @xmath150 for a stellar mass and redshift matched sample of galaxies living in average density environments ( @xmath162 , horizontal dashed line ) .
similarly , we use the halo masses of matched central and satellite model galaxies to generate the halo mass pdfs given their type ( @xmath163 and @xmath164 respectively ) .
figure [ halopdfexamples ] shows three examples of such pdfs for one object with high @xmath149 , one with high @xmath150 , and one object with an almost equal probability of being a central or a satellite .
the vertical dashed lines mark the median halo mass for a given type .
although the total halo mass pdf can be double peaked ( middle panel ) , the degeneracy between the two peaks is broken once the galaxy types are separated , making the median values well determined for each type independently .
we test the halo mass calibration by comparing the halo mass distributions of the mock sample to the 3d - hst sample . in both panels of figure [ mhalodistr ]
, we plot the halo mass histograms for centrals and satellites of the entire mock sample .
the number counts are scaled by the ratio of the volume between the 24 lightcones and the five 3d - hst fields . in the left panel of figure [ mhalodistr ]
the dashed lines are the halo mass distributions of 3d - hst galaxies obtained by summing the full halo mass pdfs for centrals ( @xmath163 , red dashed ) and satellites ( @xmath164 , blue dashed ) weighted by @xmath149 and @xmath150 for each galaxy .
the agreement with the mock sample distributions is remarkable .
although this is in principle expected because the halo mass pdfs for observed galaxies are generated from the mock sample , it should be noted that we perform the match in bins of redshift , redshift accuracy , stellar mass , density and mass - rank . the good agreement for the whole sample between the derived pdfs and the mock distributions ( for centrals and satellites separately )
should therefore be taken as an evidence that our method has not introduced any bias in the final pdfs .
we take the median value of the halo mass pdfs given that each galaxy is a central ( @xmath165 ) or a satellite ( @xmath166 ) as an estimate of the `` best '' halo mass , weighted by @xmath149 and @xmath150 .
these values are shown in figure [ mhalodistr ] , right panel .
the agreement with the mock distributions is good . for central galaxies ,
the shape and extent of the distribution is well preserved . for satellite galaxies ,
the halo mass range is less extended than the one in the mocks ; values above @xmath167 and below @xmath168 indeed only contribute through the tails of the pdfs , and therefore do not appear when the median of the pdfs are used . in the next section , we make use of the full pdfs to derive constraints on the environmental quenching of satellite galaxies .
however , the satisfactory agreement of single value estimates of halo mass with the mock distributions makes them a valuable and reliable estimate in science applications when the use of the full pdfs is not possible or feasible .
in this section , we explore the role of environment in quenching the star formation activity of galaxies over @xmath170 using 3d - hst data .
it was first proposed by @xcite that the fraction of passive galaxies depends both on stellar mass and environment in a separable manner .
@xcite , using the sdss and zcosmos surveys , extended the independence of those processes to @xmath171 .
more recently , @xcite interpreted these trends in the local universe by suggesting that central galaxies are only subject to `` mass quenching '' while satellites suffer from the former plus an `` environmental quenching '' .
@xcite similarly found that satellite galaxies are the main drivers of environmental quenching up to @xmath172 using zcosmos data . here ,
we extend these analysis to higher redshift by exploring the dependence of the fraction of passive galaxies on stellar mass , halo mass and central / satellite status in order to derive the efficiency and timescale of environmental quenching . in appendix [ app_passfrac_dens ]
, we show that we obtain consistent results using the observed galaxy density as opposed to calibrated halo mass .
the populations of passive and star - forming galaxies are typically separated either by a specific star formation rate cut ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or by a single color or color - color selection ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in this work , we use the latter method and select passive and star forming galaxies based on their position in the rest - frame uvj color - color diagram @xcite . following @xcite , passive galaxies are selected to have : @xmath173 @xmath174\ ] ] @xmath175\ ] ] where the colors are rest - frame and are taken from @xcite .
figure [ uvjdiagram ] shows the distribution of 3d - hst galaxies in the rest frame uvj color - color plane .
the red solid line shows the adopted division between passive and star forming galaxies .
the fractions of passive centrals and satellites in bins of @xmath176 and @xmath177 are computed as the fraction of passive objects in a given stellar mass bin where each galaxy is weighted by its probability of being central or satellite and the probability of being in a given halo mass bin for its type .
algebraically : @xmath178 where ty refers to a given type ( centrals or satellites ) , @xmath179 is 1 if a galaxy is uvj passive and 0 otherwise , @xmath180 is 1 if a galaxy is in the stellar mass bin and 0 otherwise , @xmath181 is the probability that a galaxy is of a given type ( see section [ pcensatcalib ] ) and @xmath182 is the halo mass pdf given the type integrated over the halo mass bin limits ( see section [ halomasscalib ] ) .
the data points in figure [ passfrac_mhalo ] show the passive fractions in two bins of halo mass ( above and below @xmath183 ) and in three independent redshift bins . the median ( log ) halo masses for satellites are 12.36 , 13.53 at @xmath184 for the lower and higher halo mass bin
respectively ; 12.41 , 13.44 at @xmath185 ; and 12.43 , 13.34 at @xmath186 .
the uncertainties on the data points can not be easily evaluated assuming binomial statistics because the number of galaxies contributing to each point is not a priori known .
indeed , @xmath181 and @xmath182 act as weights and all galaxies with a stellar mass within the mass bin do contribute to the passive fraction . to assess the uncertainties we use the mock lightcones ( where each mock galaxy has been assigned a @xmath149 and @xmath150 and halo mass pdfs as if they were observed galaxies ) . in a given stellar mass bin
we assign each model galaxy to be either passive or active such that the fraction of passive galaxies matches the observed one
. then we randomly select a number of model galaxies equal to the number of observed galaxies in that bin and we compute the passive fraction of this subsample using equation [ eqpassive_mhalo ] .
we repeat this procedure 50000 times to derive the @xmath53 errorbars shown in figure [ passfrac_mhalo ] .
this method accounts for uncertainties in the estimate of @xmath181 and @xmath182 as well as cosmic variance .
the vertical dashed lines , in figure [ passfrac_mhalo ] , mark the stellar mass completeness limit derived following @xcite . in brief
, we use the 3d - hst photometric catalog ( down to @xmath25 = 25 mag ) and we scale the stellar masses of the galaxies as if they were at the spectroscopic sample limit of @xmath25= 24 mag ( which defines the sample used in this work ) .
the scatter of the points is indicative of the @xmath187 variations in the population at a given redshift .
we then take the upper @xmath188 percentile of the distributions as a function of redshift as the stellar mass limit , which is approximately @xmath189 and @xmath190 for old and red galaxies at @xmath191 and @xmath192 respectively . below this mass
we limit the upper edge of the redshift slice such that all galaxies in the stellar mass bin are included in a mass complete sample .
a stellar mass bin is included only if the covered volume is greater than @xmath193 of the total volume of the redshift slice .
this typically results in only one stellar mass bin below the completeness limit being included in the analysis . in the highest halo mass bin of figure
[ passfrac_mhalo ] at @xmath184 , the satellite passive fraction ( integrated over all galaxies ) is higher than the central passive fraction , with a marginal significance .
the same trend can be observed in the other halo mass and redshift bins , although the separation of the observed satellite and central passive fractions becomes more marginal . in each redshift bin
we also identify a sample of `` pure '' central galaxies ( @xmath194 , irrespective of overdensity or halo mass ) , which provides a reference for the passive fraction of galaxies subject only to mass - quenching .
the passive fraction of this sample @xmath195 of centrals ( which has an average @xmath196 ) is shown as the thick red line in both halo mass bins .
the separation of the observed satellite passive fraction from that of the pure sample of centrals is more significant ( especially at @xmath197 ) .
indeed , the passive fractions derived using equation [ eqpassive_mhalo ] can be strongly affected by impurities in the central / satellite classification and by cross - talk between the two halo mass bins , given that each galaxy can contribute to both bins and types ( see equation [ eqpassive_mhalo ] ) .
any contribution of central galaxies to the satellite passive fraction , and vice versa , will reduce the observed difference between the two populations with respect to the `` pure '' , intrinsic difference . in order to recover the `` pure '' passive fraction for satellite galaxies as a function of halo mass
, we perform a parametric model fitting to our dataset .
we start by parametrizing the probability of a satellite galaxy being passive independently in each stellar mass bin as a function of log halo mass , using a broken function characterized by a constant value ( @xmath198 ) below the lower break ( @xmath199 ) and another constant value ( @xmath200 ) above the upper break ( @xmath201 ) . in between the breaks ,
the passive fraction increases linearly .
algebraically , this 4-parameter function is defined as : @xmath202 where @xmath203 .
this function is chosen to allow for a great degree of flexibility .
we make the assumption that satellite galaxies are not subject to environmental quenching below @xmath199 , and therefore treat @xmath198 as a nuisance parameter of the model with a gaussian prior centered on the observed passive fraction of pure centrals @xmath195 and a sigma equal to its uncertainty . for @xmath200
, instead we assume a semi - gaussian prior with the same center and sigma as above , but only extending below the observed passive fraction of central galaxies ( this implies that satellites are affected by the same mass - quenching as centrals ) .
above this value we assume a uniform prior .
for the break masses we assume uniform priors .
table [ modelpars ] summarizes the model parameters , their allowed range , and the number of bins in which the range is divided to compute the posterior .
.table of the model parameters .
[ cols="^,^,^,^",options="header " , ] the probability that each 3d - hst galaxy , @xmath204 is passive is : @xmath205 where @xmath206 is from equation [ eqpassive_satmodel ] and @xmath207 .
the likelihood space that the star forming or passive activity of 3d - hst galaxies in a stellar mass bin is reproduced by the model is computed as follows : @xmath208 we compute the posterior on a regular grid covering the parameter space .
we then sample the posterior distribution and we apply the model described in equation [ eqpassive_satmodel ] to obtain the median value of @xmath206 and its @xmath53 uncertainty as a function of halo mass .
lastly , we assign the probability of being passive to mock satellites in each stellar mass bin according to their model halo mass , and we compute the average passive fraction in the two halo mass bins ( above and below @xmath209 ) .
this results in the thick blue ( @xmath210 ) lines with @xmath53 confidence intervals plotted as shaded regions in figure [ passfrac_mhalo ] .
we illustrate in appendix [ app_decontamination ] an example of this procedure applied to a single redshift bin .
we verify that the separation seen in the pure passive fractions in figure [ passfrac_mhalo ] is real .
to do so we randomly shuffle the position in the uvj diagram for galaxies in each stellar mass bin ( irrespective of environmental properties ) to break any correlation between passive fraction and environment .
then we compute the observed passive fractions of centrals and satellites , and for the pure sample of centrals and we perform again the model fitting procedure . at @xmath211
we find that the pure satellite passive fraction is inconsistent with the null hypothesis ( no satellite quenching ) at a @xmath212 level in each stellar bin at high halo mass and 7 out of 8 stellar mass bins at low halo mass .
the combined probability of the null hypothesis is @xmath213 in either halo mass bin .
the difference is smaller , but still very significant ( @xmath214 ) at @xmath215 .
at @xmath216 the hypothesis of no satellite quenching is acceptable ( @xmath217 ) in the low halo mass bin , while it can be ruled out ( @xmath214 ) at higher halo mass .
@xcite , @xcite , and @xcite have found that the environment plays an important role in determining the star formation activity of satellites , at least up to @xmath218 .
however these works have only probed relatively massive haloes ( @xmath219 ) .
the depth of the 3d - hst sample allows us , for the first time , to extend these results to higher redshift , to lower mass galaxies and to lower mass haloes . in order to further understand the increased passive fractions for satellite galaxies we compute the `` conversion fractions '' as first introduced by @xcite
this parameter , sometimes called the satellite quenching efficiency , quantifies the fraction of galaxies that had their star formation activity quenched by environment specific processes since they accreted as satellites into a more massive halo ( see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is defined as : @xmath220 where @xmath221 and @xmath195 are the corrected fractions of quenched centrals and satellites in a given bin of @xmath134 and @xmath222 obtained as described above . in equation [ eqconvfrac ]
we compare the sample of centrals at the same redshift as the satellites .
this builds on the assumption that the passive fraction of central galaxies only depends on stellar mass and that the effects of mass and environment are independent and separable .
the conversion fraction then represents the fraction of satellites which are quenched due to environmental processes above what would happen if those galaxies would have evolved as centrals of their haloes .
a different approach would be to compare the passive fraction of satellites to that of centrals at the time of infall in order to measure the total fraction of satellites quenched since they were satellites ( e.g. , * ? ? ?
* ; * ? ? ?
however this measurement includes the contribution of mass - quenched satellite galaxies , which we instead remove under the assumption that the physical processes driving mass quenching do not vary in efficiency when a galaxy becomes a satellite .
we also caution the reader that equation [ eqconvfrac ] has to be taken as a simplification of reality as it does not take into account differential mass growth of centrals and satellites which can be caused by tidal phenomena in dense environments or different star formation histories .
figure [ convfrac_mhalo ] shows the conversion fractions in the same bins of @xmath176 , @xmath177 and redshift as presented in figure [ passfrac_mhalo ] .
previous results from galaxy groups and clusters from @xcite , and @xcite are plotted in our higher halo mass bin ( colored points with errorbars ) .
we also add the conversion fractions from @xcite obtained from zcosmos data as a function of local galaxy overdensity .
we plot their overdensity bins above the mean overdensity in our higher halo mass bin and the others in our lower halo mass bin following the overdensity to halo mass conversion given in @xcite .
the agreement of our measurements with other works is remarkable considering that different techniques to define the environment ( density and central / satellite status ) and passiveness are used in different works .
the satellite quenching efficiency tends to increase with increasing stellar mass and to decrease with increasing redshift at fixed stellar mass . in the lower halo mass bin ,
we note the presence of similar trends as at higher halo masses although the uncertainties are larger due to the smaller number of satellites . in our probabilistic approach
this is due to the lower @xmath157 in low density environments as shown in figure [ pcenpsat ] .
moreover , @xmath223 is poorly constrained at @xmath224 due to small number statistics of high mass satellites in the 3d - hst fields .
a positive satellite conversion fraction can be interpreted in terms of a prematurely truncated star formation activity in satellite galaxies compared to field centrals of similar stellar mass .
we define the quenching timescale ( @xmath225 ) as the average time elapsed from the first accretion of a galaxy as satellite to the epoch at which the galaxy becomes passive , and we estimate it by assuming that galaxies which have been satellites for longer times are more likely to be quenched @xcite .
indeed , the quenching can be interpreted to happen a certain amount of time after satellite galaxies cease to accrete material ( including gas ) from the cosmic web ( see section [ sec_discussion ] ) . in practice
, we obtain quenching timescales from the distribution of @xmath226 for satellite galaxies , which we define as the time the galaxy has spent as a satellite of haloes of any mass since its first infall ( e.g. , * ? ? ?
* ) . for each bin of @xmath176 , @xmath177 , and
redshift we select all satellite galaxies in our mock lightcones which define the distribution of @xmath226 .
then we select as the quenching timescale the percentile of this distribution which corresponds to @xmath227 .
this method builds on the assumption that the infall history of observed satellites is well reproduced by the sam .
systematic uncertainties can arise in the analytic prescriptions used for the dynamical friction timescale of satellites whose parent halo has been tidally stripped in the n - body simulation below the minimum mass for its detection ( the so - called `` orphan galaxies '' ) .
when this time is too short , too many satellites merge with the central galaxy and are removed from the sample , and vice - versa when the time is too long .
@xcite explored the dynamical friction timescale in multiple sams , finding a wide range of timescales . however a dramatically wrong dynamical friction recipe impacts the fraction of satellites , the stellar mass functions , and the density - mass bivariate distribution , which we found to be well matched between the mocks and the observations . in principle
low stellar mass galaxies ( @xmath228 ) are more affected by the resolution limit of the simulation and their derived quenching timescales might be subject to a larger uncertainty compared to galaxies of higher stellar mass .
we verified that this is not the case by comparing the distribution of @xmath226 in two redshift snapshots ( @xmath229 and @xmath230 ) of @xcite built on the millennium - i and the millennium - ii simulations .
the latter is an n - body simulation started from the same initial conditions of the original millennium run but with a higher mass resolution at the expense of a smaller volume .
the higher resolution means that the sub - haloes hosting low mass satellites galaxies , which can be tidally stripped , are explicitly tracked to lower mass and later times : while these are detected the recipe for dynamical friction is not invoked .
we obtain consistent quenching timescales for millennium - i and millennium - ii based mock catalogues , and therefore we conclude that the analytical treatment of orphan galaxies does not bias our results .
figure [ tquench_mhalo ] shows our derived quenching timescales ( black points ) in the same bins of @xmath176 and @xmath177 and redshift as presented in figures [ passfrac_mhalo ] and [ convfrac_mhalo ] .
the observed trend of @xmath223 with stellar mass that is found in both redshift bins turns into a trend of @xmath225 .
quenching timescales increase to lower stellar mass in all redshift and halo mass bins , mainly a consequence of the decreasing conversion fraction .
this parameter ranges from @xmath231 gyr for low mass galaxies to @xmath232 gyr for the most massive ones , and is in agreement with that found by @xcite .
remarkably , the dependence of quenching timescale on halo mass is very weak .
we overplot in each panel , as gray symbols , the quenching timescales obtained from our sample with the same procedure described above but without separating the data in two halo mass bins . in most of the stellar mass bins we find a good agreement , within the uncertainties , between the black and the gray points .
the lack of a strong halo mass dependence is a consequence of the typically shorter time since infall for satellite galaxies in lower mass haloes which largely cancels the lower conversion fraction in low mass haloes , and suggests that the physical process responsible for the premature suppression of star formation in satellite galaxies ( when the universe was half of its present age ) is largely independent of halo mass .
a mild redshift evolution is also seen when comparing the redshift bins : passive satellites at higher redshift are quenched on a shorter timescale . in the next section
we will further explore the redshift evolution of the quenching timescales from @xmath233 by combining the 3d - hst sample with a local galaxy sample from sdss .
figures [ convfracz ] and [ tquenchz ] show the evolution of the conversion fraction and the quenching timescale from redshift 0 to 2 .
we now concentrate on three bins of stellar mass , each of 0.5 dex in width , and ranging from @xmath234 to @xmath235 . given that @xmath223 ( and consequently @xmath225 ) are poorly constrained at @xmath224 due to the low number statistics of massive satellites , we exclude more massive galaxies from these plots .
similarly , galaxies at @xmath236 are only included in the mass limited sample at the lowest end of the redshift range under study , therefore the redshift evolution of @xmath223 , and @xmath225 can not be derived for those low mass galaxies .
a stellar mass bin appears in figures [ convfracz ] and [ tquenchz ] only if the stellar mass range above the mass limit is more than half of the entire stellar mass extent of the bin .
the values ( solid lines ) and their associated uncertainties ( dashed lines ) are obtained by performing the procedure described in the previous sections in overlapping redshift bins defined such that @xmath237 where @xmath238 is the width of the redshift bin and @xmath71 its center .
this means we span larger volumes at higher redshift , modulating the decrease in sample density ( malmquist bias ) and retaining sufficient sample statistics .
it is also close to a constant bin in cosmic time .
the x - axis of both figures is scaled such that the width of the redshift bins is constant and is shown as the horizontal error bar .
we include only galaxies in a stellar mass complete sample for each redshift bin .
in addition to the 3d - hst based constraints , we add constraints at @xmath2 , obtained using the same method to ensure homogeneity . the observational sample is drawn from sdss and the mock sample from the redshift zero snapshot of the @xcite model .
we describe the details of how those datasets are processed in appendix [ app_sdss ] . for this sample
we restrict to stellar masses above @xmath234 to avoid including low mass galaxies with large @xmath239 corrections .
the evolution of @xmath223 as seen in figure [ convfrac_mhalo ] is now clearly visible over the large redshift range probed by 3d - hst .
the fraction of environmentally quenched satellite galaxies is a function of @xmath240 , @xmath134 and redshift . at fixed redshift @xmath223 is higher for higher mass galaxies and at fixed stellar mass it is higher in more massive haloes .
more notably , the redshift evolution follows a decreasing trend with increasing redshift such that at @xmath241 the excess of quenching of satellite galaxies becomes more marginal ( at least for massive galaxies ) as first predicted by @xcite using halo accretion models .
several observational works reached a similar conclusion .
@xcite , @xcite , and @xcite found a significant build - up of the faint end of the red sequence ( of passive galaxies ) in cluster environments from @xmath13 toward lower redshift .
this implies an increase in the fraction of quenched satellites with decreasing redshift for low mass galaxies .
recently , @xcite found that the environmental quenching efficiency tends to zero at @xmath242 , although their analysis is only based on local overdensity and does not separate centrals and satellites . with the 3d - hst dataset we can not rule out that satellite quenching is still efficient for lower mass satellites at @xmath243 ; deeper samples are required to robustly assess the satellite quenching efficiency at @xmath244 . moving to the present day universe ( sdss data )
does not significantly affect the fraction of environmentally quenched satellites despite the age of the universe nearly doubling compared to the lowest redshift probed by the 3d - hst sample .
the redshift dependence of the quenching timescale originates from the combination of the evolution of @xmath223 and the distributions of infall times for satellite galaxies .
the redshift evolution of @xmath223 in the high halo mass bin is well matched by the halo assembly history ( at lower redshift they have been satellites on average for more time ) and therefore @xmath225 is mostly independent of redshift .
however , for lower mass galaxies a mild redshift evolution of @xmath225 might be present .
however the slope is much shallower than the ageing of the universe .
for this reason , going to higher redshift , @xmath225 approaches the hubble time and the satellite quenching efficiency decreases . despite the large uncertainty on the quenching times at low halo mass
, their redshift evolution appears to be largely independent of halo mass .
this means that the halo mass dependence of the conversion fractions may be mostly driven by an increase in the time spent as satellites in more massive haloes . at @xmath2
a more significant difference is found between the quenching times in the two halo mass bins . in the next section
we discuss which mechanism can produce these observational signatures .
there is a growing consensus that the evolution of central galaxies is regulated by the balance between cosmological accretion , star formation and gas ejection processes in a so - called `` equilibrium growth model '' ( e.g. * ? ? ?
the reservoir of cold gas in each galaxy is replenished by accretion , and will fuel star formation . as the rate of cosmological accretion
is correlated with the mass of the halo , this regulates mass growth via star formation . as a result the eventual stellar mass is also tightly correlated with halo mass , driving a tight relation between star formation rate and stellar mass for normal star forming galaxies ( the main sequence - ms - of star forming galaxies , e.g. * ? ? ?
when galaxies fall into a more massive halo the accretion of new gas from the cosmic web is expected to cease : such gas will instead be accreted ( and shock heated ) when it reaches the parent halo @xcite .
more recently @xcite estimate that this process occurs at a minimum halo mass @xmath245 , which is largely independent of redshift .
this roughly corresponds to the minimum halo mass at which satellites are detected in the 3d - hst survey ( see figure [ mhalodistr ] ) .
there are several additional ways in which a satellite galaxy s gas and stellar content can be modified through interaction with its environment , including stripping of the hot or cold gas , and tidal interactions among galaxies or with the halo potential itself .
an important combined effect is to remove ( partially or completely ) the gas reservoir leading to the quenching of star formation .
however , as pointed out by @xcite , and @xcite , it might not be necessary to invoke these mechanisms of environmental quenching to be effective .
the high sfr typical of galaxies at high redshift , combined with outflows , can lead to exhaustion of the gas reservoir in the absence of cosmological accretion .
our approach to link the conversion fractions to the distributions of time spent as satellite is based on the assumption that a galaxy starts to experience satellite specific processes at the time of its first infall into a larger halo and , in particular , that the cosmological accretion is shut off at that time .
we now examine whether a pure exhaustion of the gas reservoir can explain the quenching times we observe , or whether additional gas - removal mechanisms are required .
first we appeal to the similarity of quenching times in the two halo mass bins shown in figure [ tquenchz ] to support the pure gas exhaustion scenario .
other than at @xmath2 , the derived quenching times are indeed consistent within the uncertainties , therefore the main quenching mechanism has to be largely independent of halo mass .
ram pressure stripping is often invoked as the main quenching mechanism for satellite galaxies in low redshift clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
its efficiency is a function of the intracluster medium ( icm ) density and the velocity of galaxies in the halo .
more massive haloes have a denser icm and satellites move faster through it which exerts a stronger dynamical pressure on the gas leading to faster stripping ( and shorter quenching times ) in more massive haloes @xcite
. our 3d - hst dataset does not extend to the extreme high mass end of the halo mass function in which ram pressure effects have been clearly observed ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and so the lack of significantly shorter quenching times in the higher halo mass bin is consistent with the lack of stripping , and indeed of any strong halo - mass dependent gas - stripping process
. however @xcite find a small halo mass dependence of the quenching times comparing their geec2 group sample ( @xmath246 ) to the gclass cluster sample ( @xmath247 ) .
these evidences might indicate that dynamical stripping can play a minor role in more massive haloes even at @xmath218 . at @xmath2 instead ,
thanks to the large area covered by the sdss dataset , a number of very massive haloes are included in the higher halo mass bin .
this , combined with the presence of hot and dense icm in massive haloes in the local universe might be sufficient to explain the shorter quenching times in the high halo mass bin .
@xcite , and @xcite found quenching timescales which are possibly shorter in massive clusters of galaxies ( @xmath8 gyr ) .
however , a quantitative comparison is hampered by the different definitions of the quenching timescale .
having ruled out gaseous stripping as the main driver of satellite quenching in the range of halo mass commonly probed by our samples ( @xmath4 ) , we now concentrate on how the gas exhaustion scenario can explain the observed values of @xmath225 . to explain the quenching times at @xmath2 , @xcite presented a model dubbed the `` delayed then rapid '' quenching scenario , shown in the top panel of figure [ quenchmodels ] .
this model assumes that @xmath225 can be divided into two phases . during the first phase ,
usually called the `` delay time '' ( @xmath248 ) , the star formation activity of satellites on average follows the ms of central galaxies .
after this phase the star formation rate drops rapidly and satellite galaxies become passive on a short timescale called the `` fading time '' .
@xcite estimated an exponential fading with a characteristic timescale @xmath249 gyr which depends on stellar mass at @xmath2 . at @xmath250 , @xcite ; and @xcite estimated the fading time to be @xmath251 gyr , by identifying a `` transition '' population of galaxies likely to be transitioning from a star forming to a passive phase .
these values suggest little redshift evolution of the fading timescale with cosmic time .
@xcite developed a physical interpretation of this model .
these authors assumed that the long delay times are only possible if the satellite galaxy has maintained a multi - phase reservoir which can cool onto the galaxy and replenish the star forming gas ( typically molecular ) at roughly the same rate as the gas is lost to star formation ( and potentially outflows ) .
a constant molecular gas reservoir produces a nearly constant sfr according to the kennicutt - schmidt relation @xcite .
then the eventual depletion of this cold gas results in the rapid fading phase .
an alternative scenario would be that satellite galaxies retain only their molecular gas reservoirs after infall . in this case , if we assume a constant efficiency for star formation we should expect a star formation history which immediately departs from the ms , declining exponentially as the molecular gas is exhausted ( `` slow quenching '' model shown in the bottom panel of figure [ quenchmodels ] ) . by using our data
we directly test those two toy models .
we use the star formation rates ( @xmath252 ) for 3d - hst galaxies presented in the @xcite catalogue . by limiting to galaxies in the redshift range @xmath253 , stellar mass range @xmath254 and a maximum offset below the main sequence of 0.5 dex
, we make sure that the sfr estimates are reliable and , for 91% of the objects , are obtained from _ spitzer _
24@xmath255 m observations combined with a uv monochromatic luminosity to take into account both dust obscured and unobscured star formation . for the remaining 9% sfr estimates are from an sed fitting procedure ( see * ? ? ?
* ; * ? ? ?
there is growing evidence of curvature in the ms , which becomes shallower at higher stellar mass .
@xcite , @xcite , and @xcite interpreted this as a decline in star formation efficiency caused by the growth of bulges or bars in massive galaxies . to study the effects of environment above the internal processes driving the star formation efficiency at fixed stellar mass ,
we convert the sfr into an offset from this curved ms : @xmath256 using the @xcite parametrization of the ms from @xcite . in order to test the two models we again resort to the mock sample . for each central galaxy in the mocks we assign a random offset from the main sequence obtained from a pure sample ( @xmath194 ) of observed centrals : @xmath257 . for satellite galaxies , instead ,
their @xmath258 is a function of their time spent as satellites ( @xmath226 ) as follows : @xmath259 where @xmath225 and @xmath260 are the total quenching time and the delay time respectively , and @xmath261 is the characteristic timescale of the exponential fading phase . the latter is computed for each galaxy independently such that the sfr drops 1 dex below the ms in ( @xmath262 ) gyr . as we already computed @xmath225 ,
the only free parameter remaining in this family of models is @xmath260 .
we define the `` slow quenching '' model for @xmath263 , and `` delayed then rapid '' those where @xmath264 .
figure [ msoffset ] shows the distributions of @xmath258 for 3d - hst satellites in two stellar mass bins , obtained as usual by weighting all galaxies by @xmath157 , and for the two models obtained from the mock sample in the same way .
the histograms are normalized to the total number of 3d - hst satellites in the same stellar mass bin ( including uvj passive galaxies ) .
we stress that this comparison is meaningful because our models include the cross - talk between centrals and satellites . in the `` delayed then rapid '' scenario ,
the value of the delay time that best reproduces the observed data is @xmath265 gyr for the @xmath266 ( @xmath267 ) stellar mass bins respectively .
this means the average satellite fades with an @xmath268-folding timescale of @xmath269 gyr .
our values are consistent with those from @xcite at @xmath2 and other independent estimates at high - z .
@xcite performed a statistical identification of central and satellite galaxies in the ultravista and 3d - hst fields and found that the onset of satellite quenching occurs 1.5 - 2 gyr later than that of central galaxies at fixed number density .
these values are in good agreement with the delay times estimated in our work .
conversely the `` slow quenching '' model predicts too many galaxies below the main sequence but which are not uvj passive ( `` transition '' galaxies ) .
the fraction of 3d - hst satellites for which @xmath270 is 65% ( 46% ) , which compares to 67% ( 47% ) for the `` delayed then rapid '' model ; instead it drops to 52% ( 39% ) for the `` slow quenching '' model .
we tested that the distributions of @xmath258 and the estimated fading times are not biased by inaccurate uv+ir sfr for agn candidates in the sample . because the candels fields have uniform coverage of deep _
spitzer_/irac observations , we remove the sources selected by the irac color - color criteria presented in @xcite .
we find that neither the @xmath258 distributions , nor the fading time estimates change appreciably . in conclusion
the fading of the star formation activity must be a relatively rapid phenomenon which follows a long phase where satellite galaxies have a sfr which is indistinguishable from that of centrals .
this is further supported by the evidence that the passive and star forming populations are well separated in color and sfr and that the `` green valley '' in between them is sparsely populated across different environments @xcite .
finally , we discuss the implications of the quenching times on the gas content of satellites at the time of infall .
because satellite galaxies are not thought to accrete gas after infall , their continued star formation occurs at the expense of gas previously bound to the galaxy .
as previously discussed , @xcite explain the fading phase by the depletion of molecular gas .
the depletion time of molecular gas ( @xmath271 ) has been derived by several authors @xcite .
there is general consensus that this timescale ( which is an @xmath268-folding time ) is @xmath272 gyr at @xmath2 and is @xmath273 gyr at @xmath191 .
moreover it is independent of stellar mass . in this framework
we might expect fading times shorter than ( or similar to ) @xmath271 , where shorter fading times are possible where a fraction of the gas is lost to outflows .
our fading times are indeed somewhat shorter than the molecular gas depletion times , consistent with this picture , but with a mass dependence which suggests a mass - dependent outflow rate . in figure [ tquench_tfade ]
we show the ratio of the delay time to the fading time as a function of redshift in bins of stellar mass . assuming that the fading phase is driven by depletion of molecular gas ( in absence of further replenishment ) , this ratio informs us about the relative time spent refuelling the galaxy to keep it on the ms ( from a gas reservoir initially in a warmer phase ) to the time spent depleting the molecular gas .
we note that the delay time is estimated via the quenching time ( which is a function of stellar mass and redshift ) while the fading time is computed only for 2 stellar mass bins at @xmath253 , with @xmath2 fading timescales taken from @xcite .
errors in figure [ tquench_tfade ] propagate only the errors on the total quenching time . for all stellar mass bins this ratio is above unity , which we interpret to mean that gas in a non - molecular phase is required to supply fuel for star formation , and
this gas is likely to exceed the molecular gas in mass .
the longer delay times at @xmath2 suggest that a smaller fraction of the gas mass is in molecular form .
this model also implies that a significant fraction of the final stellar mass of satellite galaxies is built - up during the satellite phase .
a multi - phase gas reservoir is observed in the local universe in the form of ionized , atomic , and molecular hydrogen .
atomic hydrogen cools , replenishing the molecular gas reservoir which is depleted by star formation . in the local universe , using the scaling relations derived from the _ herschel _ reference sample @xcite , the observed mass of atomic hydrogen is found to be 2 - 3 times larger than the amount of molecular hydrogen for our most massive stellar mass bin .
this ratio increases to @xmath274 for the lower mass objects , although with a large uncertainty .
these numbers are consistent with the picture that much , if not all , of the reservoir required to maintain the satellite on the ms during the delay phase at @xmath275 can be ( initially ) in an atomic phase .
it is also plausible that much of the gas reservoir bound to higher redshift galaxies is contained in a non - molecular form , and that this can be retained and used for star formation when the galaxies become satellites . assuming that outflows are not only active during the fading phase but rather during the entire quenching time , the mass in the multi - phase gas reservoir needs to be even larger , although it is not straightforward to constrain by how much . in conclusion
our work supports a `` delayed then rapid '' quenching scenario for satellite galaxies regulated by star formation , depletion and cooling of a multi - phase gas reservoir .
in this work , we have characterized the environment of galaxies in the 3d - hst survey at @xmath276 .
we used the projected density within fixed apertures coupled with a newly developed method for edge corrections to obtain a definitive measurement of the environment in five well studied deep - fields : goods - s , cosmos , uds , aegis , goods - n . using a recent semi - analytic model of galaxy formation ,
we have assigned physical quantities describing the properties of dark matter haloes to observed galaxies .
our results can be summarized as follows : 1 . the 3d - hst deep fields host galaxies in a wide range of environments , from underdense regions to relatively massive clusters .
this large variety is accurately quantified thanks to a homogeneous coverage of high quality redshifts provided by the 3d - hst grism observations .
2 . as described in @xcite , a calibration of density into physically motivated quantities ( e.g. halo mass , central / satellite status ) requires a mock catalogue tailored to match the properties of the 3d - hst survey .
we developed such a catalogue and performed a careful match to the observational sample .
as a result , each 3d - hst galaxy is assigned a probability that it is a central or satellite galaxy with an associated probability distribution function of halo mass for each type .
this bayesian approach naturally takes into account sources of contamination in the matching process .
we publicly release our calibrated environment catalogue to the community .
3 . the 3d - hst sample provides us with a unique dataset to study the processes governing environmental quenching from @xmath277 to the present day over a wide range of halo mass .
as no galaxy has a perfectly defined environment , a bayesian analysis including forward modelling of the mock catalogue allows us to recover `` pure '' passive fractions of central and satellite populations .
we also estimated robust and realistic uncertainties through a monte carlo error propagation scheme that takes into account the use of probabilistic quantities .
4 . by computing conversion fractions (
i.e. the excess of quenched satellite galaxies compared to central galaxies at the same epoch and stellar mass ) @xcite , we find that satellite galaxies are efficiently environmentally quenched in haloes of any mass up to @xmath278 . above these redshifts
the fraction of passive satellites is roughly consistent with that of central galaxies .
5 . under the assumption that the earliest satellites to be accreted become passive first
, we derive environmental quenching timescales .
these are long ( @xmath279 gyr at @xmath280 ; 5 - 7 gyr at @xmath2 ) and longer at lower stellar mass . as they become comparable to the hubble time by @xmath281 , effective environmental quenching of satellites is not possible at earlier times .
more remarkably , their halo mass dependence is negligible . by assuming that cosmological accretion stops
when a galaxy becomes a satellite , we were able to interpret these evidences in a `` gas exhaustion '' scenario ( i.e. the overconsumption model of * ? ? ? * ) where quenching happens because satellite galaxies eventually run out of their fuel which sustains further star formation .
we tested two toy models of satellite quenching : the `` delayed then rapid '' quenching scenario proposed by @xcite and a continuous `` slow quenching '' from the time of first infall . by comparing the observed sfr distribution for 3d - hst satellites to the predictions of these toy models we found that the scenario that best reproduces the data at @xmath282 is `` delayed then rapid '' . consistently with the results of @xcite at @xmath2 , we find that the fading of the star formation activity is a relatively rapid phenomenon ( @xmath283 gyr , lower at higher mass ) which follows a long phase where satellite galaxies have a sfr which is indistinguishable from that of centrals .
7 . by linking the fading to the depletion of molecular gas
we conclude that the `` delayed then rapid '' scenario is best explained , even at high redshift , by the presence of a significant multi - phase reservoir which can cool onto the galaxy and replenish the star forming gas at roughly the same rate as the gas is turned into stars .
this analysis of satellite quenching is only one of many possible analyses that can be performed with the environmental catalogue built in this work . in the future ,
the advent of the _ james webb space telescope , wfirst _ and _ euclid _ space missions , as well as highly multiplexed spectroscopic instruments from the ground ( e.g. , _ moons _ at vlt ; _ pfs _ at subaru ) , will provide excellent redshift estimates for fainter objects over a much larger area , to which similar techniques to calibrate environment can be applied .
this , in combination with deeper scaling relations for the atomic and molecular gas components from the _ square kilometer array _ and _ alma _ will revolutionize measurements to constrain how galaxies evolve and quench as a function of their environment .
it is a pleasure to thank nikhil arora for his help in a preliminary examination of the local sdss data , to alessandro boselli , gabriella de lucia , sean mcgee , greg rudnick , and andrew wetzel for useful discussion , to michael balogh and alexis finoguenov for providing catalogs in useful format , and to stefano zibetti and tams budavri for the work in helping deriving densities for the sdss sample as presented in @xcite , and @xcite .
we thank the anonymous referee for his / her comments that improved the quality of the manuscript .
mfossati and djw acknowledge the support of the deutsche forschungsgemeinschaft via projects , and .
jccc acknowledges the support of the deutsche zentrum fr luft- und raumfahrt ( dlr ) via project i d 50or1513 .
kew gratefully acknowledge support by nasa through hubble fellowship grant # hf2 - 51368 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa .
this work is based on observations taken by the 3d - hst treasury program ( go 12177 and 12328 ) with the nasa / esa hst , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 . based on observations made with eso telescopes at the la silla or paranal observatories under programme id(s ) 175.a-0839(b ) , 175.a-0839(d ) , 175.a-0839(i ) , 175.a-0839(j ) , 175.a-0839(h ) , 175.a-0839(f ) , 092.a-0091 , 093.a-0079 , 093.a-0187 , 094.a-0217 , 094.a-0287 , 095.a-0047 , 095.a-0109 , 096.a-0093 , and 096.a-0025 .
the mosdef data were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m . keck foundation .
we 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 .
we acknowledge all the teams and observatories that provided datasets included in the photometric and spectroscopic catalogues used in this work .
funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the u.s .
department of energy office of science .
the sdss - iii web site is http://www.sdss3.org/. sdss - iii is managed by the astrophysical research consortium for the participating institutions of the sdss - iii collaboration including the university of arizona , the brazilian participation group , brookhaven national laboratory , carnegie mellon university , university of florida , the french participation group , the german participation group , harvard university , the instituto de astrofisica de canarias , the michigan state / notre dame / jina participation group , johns hopkins university , lawrence berkeley national laboratory , max planck institute for astrophysics , max planck institute for extraterrestrial physics , new mexico state university , new york university , ohio state university , pennsylvania state university , university of portsmouth , princeton university , the spanish participation group , university of tokyo , university of utah , vanderbilt university , university of virginia , university of washington , and yale university .
natexlab#1#1 , c. , mazure , a. , ilbert , o. , et al .
2005 , , 443 , 805 , h. , allende prieto , c. , an , d. , et al .
2011 , , 193 , 29 , s. , cristiani , s. , moscardini , l. , et al .
1999 , , 310 , 540 , s. , vandame , b. , benoist , c. , et al .
2001 , , 379 , 740 , m. l. n. , willner , s. p. , fazio , g. g. , et al .
2013 , , 769 , 80 , i. k. , balogh , m. l. , bower , r. g. , et al .
2006 , , 373 , 469 , m. , eke , v. , miller , c. , et al .
2004 , , 348 , 1355 , m. l. , morris , s. l. , yee , h. k. c. , carlberg , r. g. , & ellingson , e. 1997 , , 488 , l75 , m. l. , navarro , j. f. , & morris , s. l. 2000 , , 540 , 113 , m. l. , mcgee , s. l. , mok , a. , et al .
2014 , , 443 , 2679 . 2016 , , 456 , 4364 , e. f. , wolf , c. , meisenheimer , k. , et al .
2004 , , 608 , 752 , r. , wake , d. a. , brammer , g. b. , et al .
2016 , , 822 , 30 , r. , hudelot , p. , mccracken , h. j. , et al . 2012 , , 545 , a23 , m. r. , & moustakas , j. 2009 , , 47 , 159 , m. r. , & roweis , s. 2007 , , 133 , 734 , a. , boissier , s. , cortese , l. , & gavazzi , g. 2008 , , 674 , 742 , a. , cortese , l. , boquien , m. , et al .
2014 , , 564 , a66 , a. , & gavazzi , g. 2006 , , 118 , 517 . 2014 , , 22 , 74 , a. , voyer , e. , boissier , s. , et al .
2014 , , 570 , a69 , a. , cuillandre , j. c. , fossati , m. , et al .
2016 , , 587 , a68 , g. b. , van dokkum , p. g. , & coppi , p. 2008
, , 686 , 1503 , g. b. , van dokkum , p. g. , franx , m. , et al .
2012 , , 200 , 13 , j. , charlot , s. , white , s. d. m. , et al .
2004 , , 351 , 1151 , g. , & charlot , s. 2003 , , 344 , 1000 , g. , & valtonen , m. 1990 , , 350 , 89 , d. , armus , l. , bohlin , r. c. , et al .
2000 , , 533 , 682 , c. n. , van dokkum , p. g. , urry , c. m. , et al .
2010 , , 189 , 270 , g. 2003 , , 115 , 763 , c. , gunn , j. e. , & white , m. 2009 , , 699 , 486 , m. c. , newman , j. a. , croton , d. j. , et al .
2006 , , 370 , 198 , d. j. , springel , v. , white , s. d. m. , et al .
2006 , , 365 , 11 , o. , granett , b. r. , branchini , e. , et al .
2014 , , 565 , a67 , b. , mobasher , b. , sobral , d. , et al .
2016 , arxiv e - prints , arxiv:1605.03182 , g. , & blaizot , j. 2007 , , 375 , 2 , g. , boylan - kolchin , m. , benson , a. j. , fontanot , f. , & monaco , p. 2010
, , 406 , 1533 , g. , weinmann , s. , poggianti , b. m. , aragn - salamanca , a. , & zaritsky , d. 2012 , , 423 , 1277 , g. , poggianti , b. m. , aragn - salamanca , a. , et al .
2007 , , 374 , 809 , a. , & birnboim , y. 2006 , , 368 , 2 , j. l. , koekemoer , a. m. , brusa , m. , et al .
2012 , , 748 , 142 , a. 1980 , , 236 , 351 , t. , hildebrandt , h. , lerchster , m. , et al .
2009 , , 493 , 1197 , g. , finoguenov , a. , tanaka , m. , et al .
2013 , , 765 , 117 , g. , popesso , p. , finoguenov , a. , et al .
2016 , , 455 , 2839 , j. , & thomas , d. 2015 , , 451 , 660 , a. , guzzo , l. , hasinger , g. , et al .
2007 , , 172 , 182 , a. , watson , m. g. , tanaka , m. , et al .
2010 , , 403 , 2063 , a. , tanaka , m. , cooper , m. , et al .
2015 , , 576 , a130 , f. , springel , v. , angulo , r. e. , & henriques , b. 2012 , , 426 , 2335 , m. , fumagalli , m. , boselli , a. , et al .
2016 , , 455 , 2028 , m. , gavazzi , g. , boselli , a. , & fumagalli , m. 2012 , , 544 , a128 , m. , wilman , d. j. , fontanot , f. , et al .
2015 , , 446 , 2582 , m. , van dokkum , p. g. , frster schreiber , n. m. , et al .
2008 , , 688 , 770 , m. , fossati , m. , hau , g. k. t. , et al .
2014 , , 445 , 4335 , h. , kosugi , g. , akiyama , m. , et al .
2008 , , 176 , 1 , a. , & bell , e. f. 2009 , , 185 , 253 , g. , boselli , a. , mayer , l. , et al .
2001 , , 563 , l23 , g. , fumagalli , m. , cucciati , o. , & boselli , a. 2010 , , 517 , a73 , g. , fumagalli , m. , fossati , m. , et al .
2013 , , 553 , a89 , g. , consolandi , g. , dotti , m. , et al .
2015 , , 580 , a116 , e. , van dokkum , p. g. , herrera , d. , et al .
2006 , , 162 , 1 , j. e. , simpson , c. , rawlings , s. , read , a. m. , & watson , m. 2007 , , 381 , 1369 , r. , tacconi , l. j. , lutz , d. , et al .
2015 , , 800 , 20 , r. , cimatti , a. , daddi , e. , et al .
2003 , , 592 , 721 , n. a. , kocevski , d. d. , faber , s. m. , et al .
2011 , , 197 , 35 , j. e. , & gott , iii , j. r. 1972 , , 176 , 1 , q. , white , s. , angulo , r. e. , et al .
2013 , , 428 , 1351 , q. , white , s. , boylan - kolchin , m. , et al .
2011 , , 413 , 101 , y. , ferguson , h. c. , giavalisco , m. , et al .
2013 , , 207 , 24 , m. r. , schaye , j. , & jeeson - daniel , a. 2012 , , 419 , 2133 , c. p. , pereira , m. j. , smith , g. p. , et al .
2015 , , 806 , 101 , b. m. b. , white , s. d. m. , thomas , p. a. , et al .
2015 , , 451 , 2663 . 2016 , arxiv e - prints , arxiv:1611.02286 , m. , de lucia , g. , wilman , d. , et al .
2014 , , 444 , 2938 , p.
f. , cox , t. j. , kere , d. , & hernquist , l. 2008 , , 175 , 390 , b .- c . , wang , w .- h .
, hsieh , c .- c .
, et al . 2012 , , 203 , 23 , l .- t . , salvato , m. , nandra , k. , et al .
2014 , , 796 , 60 , o. , arnouts , s. , mccracken , h. j. , et al .
2006 , , 457 , 841 , o. , capak , p. , salvato , m. , et al .
2009 , , 690 , 1236 , t. , feigelson , e. d. , akritas , m. g. , & babu , g. j. 1990 , , 364 , 104 , w. , sluka , c. , schindler , s. , ferrari , c. , & ziegler , b. 2009 , , 499 , 87 , g. , white , s. d. m. , heckman , t. m. , et al .
2004 , , 353 , 713 , g. , heckman , t. m. , tremonti , c. , et al .
2003 , , 346 , 1055 , j. d. p. , abramson , a. , & bravo - alfaro , h. 2015 , , 150 , 59 , jr .
, r. c. 1998 , , 36 , 189 . 1998 , , 498 , 541 , c. , lilly , s. j. , kova , k. , et al .
2013 , , 769 , 24 , t. , yamada , t. , akiyama , m. , et al .
2004 , , 350 , 1005 , a. m. , faber , s. m. , ferguson , h. c. , et al .
2011 , , 197 , 36 , k. , lilly , s. j. , cucciati , o. , et al .
2010 , , 708 , 505 , k. , lilly , s. j. , knobel , c. , et al .
2014 , , 438 , 717 , m. , van dokkum , p. g. , labb , i. , et al .
2009 , , 700 , 221 , m. , shapley , a. e. , reddy , n. a. , et al .
2015 , , 218 , 15 , j. , cimatti , a. , zamorani , g. , et al .
2009 , , 504 , 331 , j. , cimatti , a. , daddi , e. , et al .
2013 , , 549 , a63 , r. b. , tinsley , b. m. , & caldwell , c. n. 1980 , , 237 , 692 , o. , tasca , l. a. m. , cassata , p. , et al .
2015 , , 576 , a79 , b. d. , brandt , w. n. , alexander , d. m. , et al .
2005 , , 161 , 21 , s. j. , carollo , c. m. , pipino , a. , renzini , a. , & peng , y. 2013 , , 772 , 119 , s. j. , le fvre , o. , renzini , a. , et al .
2007 , , 172 , 70 , d. , van dokkum , p. g. , frster schreiber , n. m. , et al .
2009 , , 701 , 1765 , c. , moore , b. , mayer , l. , et al .
2005 , , 364 , 607 , h. j. , milvang - jensen , b. , dunlop , j. , et al .
2012 , , 544 , a156 , s. l. , balogh , m. l. , bower , r. g. , font , a. s. , & mccarthy , i. g. 2009 , , 400 , 937 , s. l. , bower , r. g. , & balogh , m. l. 2014 , , 442 , l105 , i. s. , steidel , c. c. , epps , h. w. , et al .
2012 , in , vol . 8446 , ground - based and airborne instrumentation for astronomy iv , 84460j , j. t. , simard , l. , palmer , m. , ellison , s. l. , & patton , d. r. 2014 , , 210 , 3 , j. t. , saglia , r. p. , bender , r. , et al .
2015 , , 804 , l4 , p. , busarello , g. , dopita , m. a. , et al .
2013 , , 429 , 1747 , d. 1983 , , 264 , 24 , a. , balogh , m. l. , mcgee , s. l. , et al .
2013 , , 431 , 1090 . 2014 , , 438 , 3070 , i. g. , brammer , g. b. , van dokkum , p. g. , et al .
2016 , , 225 , 27 , b. , lake , g. , & katz , n. 1998 , , 495 , 139 , e. , barmby , p. , rigopoulou , d. , et al .
2003 , , 403 , 493 , s. i. , croton , d. j. , skibba , r. a. , et al .
2012 , , 419 , 2670 , a. , wilson , g. , yee , h. k. c. , et al .
2012 , , 746 , 188 , a. , marchesini , d. , stefanon , m. , et al .
2013 , , 777 , 18 , a. , van der burg , r. f. j. , mcgee , s. l. , et al .
2014 , , 796 , 65 , e. j. , van dokkum , p. g. , frster schreiber , n. m. , et al .
2016 , , 828 , 27 , k. g. , weiner , b. j. , faber , s. m. , et al .
2007 , , 660 , l43 , jr . , a. 1974 , , 194 , 1 , j. b. 1974 , , 27 , 21 , a. , vulcani , b. , poggianti , b. m. , et al .
2016 , , 816 , l25 , c. , momcheva , i. , willmer , c. n. a. , et al .
2010 , , 716 , 1503 , y .- j . , lilly , s. j. , renzini , a. , & carollo , m. 2012 , , 757 , 4 , y .- j . ,
lilly , s. j. , kova , k. , et al . 2010 , , 721 , 193 , s. , aldering , g. , goldhaber , g. , et al . 1999 , , 517 , 565 , s. , wilman , d. j. , zibetti , s. , & budavri , t. 2014 , , 438 , 2233 , ade , p. a. r. , aghanim , n. , et al .
2014 , , 571 , a16 , b. m. , bridges , t. j. , komiyama , y. , et al .
2004 , , 601 , 197 , r. f. , williams , r. j. , franx , m. , & hildebrandt , h. 2012 , , 744 , 88 , e. , & hensler , g. 2005 , , 433 , 875 , g. , von der linden , a. , pell , r. , et al .
2009 , , 700 , 1559 , a. , kauffmann , g. , wang , j. , et al .
2011 , , 415 , 61 , s. , castellano , m. , pentericci , l. , et al .
2009 , , 501 , 865 , d. b. , salvato , m. , aussel , h. , et al .
2007 , , 172 , 86 , k. , urry , c. m. , simmons , b. d. , et al .
2014 , , 440 , 889 , m. 1959 , , 129 , 243 , n. , aussel , h. , benson , a. , et al .
2007 , , 172 , 150 , r. , bender , r. , agudo berbel , a. , et al .
2013 , the messenger , 151 , 21 , g. m. , croton , d. j. , skibba , r. a. , et al .
2013 , , 433 , 3314 , j. , & rees , m. j. 1998 , , 331 , l1 , r. e. , whitaker , k. e. , momcheva , i. g. , et al .
2014 , , 214 , 24 , j. m. , manrique , a. , garca - gmez , c. , et al .
2001 , , 548 , 97 , v. , white , s. d. m. , tormen , g. , & kauffmann , g. 2001 , , 328 , 726 , v. , white , s. d. m. , jenkins , a. , et al .
2005 , , 435 , 629 , m. , donahue , m. , & voit , g. m. 2007 , , 671 , 190 , l. j. , neri , r. , genzel , r. , et al .
2013 , , 768 , 74 , t. , dekel , a. , oesch , p. , et al .
2014 , , 789 , 164 , m. , finoguenov , a. , & ueda , y. 2010 , , 716 , l152 , y. , scoville , n. , murayama , t. , et al .
2007 , , 172 , 9 , l. a. m. , le fevre , o. , ribeiro , b. , et al .
2016 , arxiv e - prints , arxiv:1602.01842 , e. n. , franx , m. , van dokkum , p. g. , et al .
2009 , , 183 , 295 , s. , & bryan , g. l. 2010 , , 709 , 1203 , d. , castellano , m. , fontana , a. , & giallongo , e. 2007 , , 463 , 853 , f. c. , aquino , d. , yang , x. , et al .
2008 , , 387 , 79 , r. f. j. , muzzin , a. , hoekstra , h. , et al .
2013 , , 557 , a15 , b. , cayatte , v. , balkowski , c. , & duschl , w. j. 2001 , , 561 , 708 , b. , huchtmeier , w. , & van driel , w. 2005 , , 439 , 921 , j. , de lucia , g. , kitzbichler , m. g. , & white , s. d. m. 2008 , , 384 , 1301 , b. j. , phillips , a. c. , faber , s. m. , et al .
2005 , , 620 , 595 , a. r. , tinker , j. l. , & conroy , c. 2012 , , 424 , 232 , a. r. , tinker , j. l. , conroy , c. , & van den bosch , f. c. 2013 , , 432 , 336 , k. e. , labb , i. , van dokkum , p. g. , et al .
2011 , , 735 , 86 , k. e. , franx , m. , leja , j. , et al .
2014 , , 795 , 104 , s. d. m. , & frenk , c. s. 1991 , , 379 , 52 , s. d. m. , & rees , m. j. 1978 , , 183 , 341 , r. j. , quadri , r. f. , franx , m. , van dokkum , p. , & labb , i. 2009 , , 691 , 1879 , d. j. , balogh , m. l. , bower , r. g. , et al . 2005 , , 358 , 71 , d. j. , zibetti , s. , & budavri , t. 2010 , , 406 , 1701 , e. , frster schreiber , n. m. , wuyts , s. , et al .
2015 , , 799 , 209 , m. , yoshida , m. , komiyama , y. , et al .
2010 , , 140 , 1814 , d. g. , adelman , j. , anderson , jr .
, j. e. , et al .
2000 , , 120 , 1579
the goods - s field is part of a larger field known as the extended chandra deep field south ( ecdfs , * ? ? ? * ) .
this field has been covered by the multiwavelength survey by yale - chile ( musyc , * ? ? ? * ) in 32 broad and medium bands from the optical to the medium infrared wavelengths .
the broadband data originates from various sources @xcite and a consistent reduction and analysis is performed by the musyc team @xcite .
the source extraction is performed on a deep combined image of three optical filters ( @xmath284 ) and reaches a depth of @xmath285 mag .
stars are removed from the catalogue by using the ` star_flag ` parameter . in order to select galaxies in a consistent way as for 3d - hst
we need deep observations in a filter with a central wavelength as close as possible to that of wfc3/f140w ( @xmath286 ) . however , the near infrared observations from musyc are shallow and only reach a depth of @xmath287 mag .
we therefore match the musyc catalogue with the taiwan ecdfs near - infrared survey ( tenis , * ? ? ?
this survey provides deep @xmath34 and @xmath288 images of the ecdfs area with limiting magnitudes of 24.5 and 23.9 respectively .
hereafter , where sky coordinates matching between different catalogues is required we select the closest match within a 1 arcsec radius .
the comparison of @xmath34 band magnitudes from the two surveys for sources above the sensitivity limit of the musyc data shows a remarkable agreement .
we then match the musyc and 3d - hst / goods - s catalogue , again by sky coordinates . using the galaxies that are present in both surveys we fit a linear function between @xmath25 and @xmath289 magnitudes .
given the significant overlap between the filters we neglect color terms in the fit .
the best bisector fit @xcite is @xmath290 .
then we use this function to generate @xmath25 magnitudes for all objects in the musyc+tenis catalogue .
we evaluate the depth of the resulting catalogue by inspecting the histogram of the number counts in the @xmath25 band .
since this is obtained from deep @xmath289 data ( 24.5 mag ) , the limiting factor will be the depth of the @xmath284 selection band of musyc .
the number counts increase linearly in log space up to @xmath291 and we therefore adopt this value for the selection . since this limit is brighter than the one we set for the primary sample in 3d - hst , a fraction of the neighbours are missed .
we correct for this by assigning to each musyc galaxy in eq .
[ eqsigma ] a weight @xmath292 that corresponds to the ratio of the cumulative luminosity function at @xmath293 and @xmath38 mag from the deeper 3d - hst catalogue .
the most recent calculation of photometric redshifts in ecdfs is presented by @xcite .
these authors combined the musyc photometry with tenis and hst / candels @xcite in the goods - s area .
we match our catalog with their catalog based on musyc i d and we find a match for each source .
@xcite also present a compilation of spectroscopic redshifts available in the literature which we use whenever available .
photo - zs are computed using _ lephare _ @xcite and their accuracy depends primarily on the availability and depth of multiwavelength photometry .
the goods - s area has deep hst coverage from candels , but those galaxies are already present in our primary 3d - hst catalog .
therefore we are primarily interested in sources outside the candels/3d - hst area . in the ecdfs footprint which is not covered by hst
more than 30 photometric bands are available and photo - zs are quite accurate : @xmath294 for galaxies with @xmath295 respectively .
those values degrade where continuum spectral features ( e.g. balmer break ) are redshifted outside the range observed with medium band filters ( @xmath243 ) , although low number statistics hampers a robust determination of the photometric redshift quality .
stellar masses are computed using the photometric data and the redshift information by choosing the same grid of templates used by @xcite for the 3d - hst fields .
we assess the quality of the stellar masses by comparing to those from @xcite , where musyc and 3d - hst overlap and we find a median offset of 0.01 dex and a scatter of 0.15 dex .
the scatter is driven by the scatter in photometric redshifts in the two catalogs . as a last step we remove from this catalog galaxies in the 3d - hst footprint that satisfy the selection criteria for the primary environment sample , to obtain a pure catalogue that we use only for the edge corrections .
the entire cosmos 2deg@xmath296 field has been observed in 30 photometric bands from uv to medium infrared ( including several medium bands ) @xcite .
photometric redshifts are computed using _ lephare _ and are presented by @xcite for sources with @xmath297 mag .
we include spectroscopic redshifts from zcosmos - bright @xcite where available .
the photometric redshift uncertainty is evaluated by @xcite comparing photo - z to spec - z and is @xmath298 for galaxies with @xmath299 respectively .
the latter value must be taken with caution as it is calibrated using a small number of objects .
we remake this comparison by using 3d - hst spec - z and grism - z as a reference ( restricting our analysis to the 3d - hst / cosmos field ) .
we divide the sample into bright ( @xmath300 mag ) and faint ( @xmath301 mag ) for @xmath302 and irrespective of magnitude for @xmath303 .
we note that for faint sources the effective magnitude limit is that of the grism redshift extraction @xmath0 mag and the comparison is limited by the degraded accuracy of grism redshifts for faint sources with no emission line detection ( see figure [ fig1 ] ) .
the redshift accuracy is : @xmath304 for the bright ( faint ) sample at low redshift and @xmath305 for the high redshift sample .
those values are consistent with the determination by @xcite and the reduced accuracy at high redshift is due to the lack of narrow bands in the nir . to overcome this limitation the newfirm medium band survey ( nmbs , *
* ) observed the cosmos field with 5 medium band filters in the @xmath34 and @xmath35 bands and a broadband filter in @xmath36 . as a result
the accuracy of photometric redshifts is significantly improved ( see section 5 in * ? ? ? * ) and we use those photo - z where they are available .
deep @xmath34 band magnitudes are provided by the ultravista survey @xcite . after matching their catalog via sky
coordinates we generate synthetic @xmath25 magnitudes as described in the previous section and using the best fit : @xmath306 .
the depth of our catalog is limited by the depth of the @xmath307 selection band from @xcite .
the number counts increase linearly in log space until @xmath308 and we therefore adopt this value for the selection limit .
as for the musyc catalog this limit is brighter than the one we set for the primary sample in 3d - hst therefore we assign to each galaxy in eq .
[ eqsigma ] a weight ( @xmath309 ) .
stars are removed from the catalog by using the ` type ` flag from @xcite .
we compute stellar masses as described in the previous section .
the agreement with stellar masses from @xcite is remarkable , with a median offset of 0.02 dex and a scatter of 0.20 dex .
lastly we remove the primary 3d - hst sources from this edge correction sample .
the 3d - hst uds field is part of a larger field known as ukidss uds .
this field features deep near infrared @xmath34 , @xmath35 , and @xmath36 observations with the ukidss telescope ( almaini et al . in prep ) complemented by optical and medium infrared data @xcite .
the uds / dr8 catalog selection is performed in @xmath36 band and the completeness limit is @xmath310 mag . as for the previous fields we exclude stars and we compute synthetic @xmath25 magnitudes using the best fit relation : @xmath311 .
the depth of our catalog matches the limiting magnitude for the primary 3d - hst sample , thus we do not apply any statistical weight for the ukidss uds galaxies when computing the density .
photometric redshifts ( w. hartley private comm . ) have a typical accuracy of @xmath312 due to the lack of narrow- or medium band photometry in this field . as for the other fields we compute stellar masses using the fast code and we find a good agreement with the values from @xcite for the 3d - hst / uds field with an offset of -0.03 dex and a scatter of 0.22 dex . again
the last step is to remove the 3d - hst primary sources via positional matching with the @xcite catalogue .
halo mass is the parameter which most easily allows the interpretation of environmental effects across cosmic time .
it also allows for easier and less biased comparisons across different works .
moreover it can be directly linked to models ( either semi - analytic or hydrodynamical ) allowing a better understanding of which physical processes are most relevant at different halo masses .
density , on the other hand , depends on the depth ( and to some extent the observing strategy ) of each survey .
detailed and quantitative comparisons are also made difficult by different approaches to density ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however it is a parameter directly obtained from the observed redshift space coordinates of the population of galaxies under investigation . in this respect
it is less sensitive to the quality and uncertainties in the calibration of halo mass . in this appendix
we derive the passive fraction of galaxies in two bins of density and compare them to those obtained in figure [ passfrac_mhalo ] using halo mass .
the observed fractions of passive centrals and satellites in bins of @xmath176 and density contrast @xmath313 are given by @xmath314 where ty refers to a given type ( centrals or satellites ) , @xmath179 is 1 if a galaxy is uvj passive and 0 otherwise , @xmath180 is 1 of a galaxy is in the stellar mass bin and 0 otherwise , @xmath315 is 1 if a galaxy is in the density bin and 0 otherwise , and @xmath181 is the probability that a galaxy is of a given type . in this equation
the only uncertain property for each object is its central / satellite status , while the cross talk between multiple density bins is not present ( as it was for halo mass ) .
we therefore perform a simpler decontamination procedure .
for each density , stellar mass and redshift bin , we assign to real centrals in the mocks a probability of being passive equal to the passive fraction of the pure sample of observed central galaxies @xmath316 , while the passive fraction of satellites @xmath317 is a free parameter
. then we use equation [ eqpassive_dens ] to compute the observed passive fractions for mock galaxies ( therefore contaminating the `` pure '' values ) .
we solve for @xmath317 by maximising the likelihood that the contaminated passive fractions for mock galaxies match the observed passive fractions ( jointly for centrals and satellites ) .
this procedure is repeated 500 times in a monte carlo fashion in order to propagate the uncertainties on the datapoints to the `` pure '' ( decontaminated ) passive fractions .
the decontaminated values of the passive fraction for centrals and satellites shown in figure [ passfrac_dens ] are qualitatively similar to those obtained in bins of halo mass in the same redshift slices ( see figure [ passfrac_mhalo ] ) .
we conclude that the dependence of environmental quenching when binned on local density is similar to that in bins of halo mass , where density is a more directly observed quantity .
in this appendix we illustrate the results of the fitting process described in section [ sec_decontamination ] for a single redshift bin .
figure [ corner_plot_example ] presents the constraints on the model parameters ( marginalised over the nuisance parameter @xmath318 ) for a single stellar mass and redshift bin .
the panels along the diagonal show the marginalised posterior distributions for each for the three parameters @xmath319 .
the red solid lines show the median value of each parameter , and the black dashed lines show the @xmath53 confidence intervals .
the off - diagonal panels show the marginalised posterior distributions for a pair of model parameters .
the black contours show the @xmath53 , @xmath320 , and @xmath321 confidence intervals .
the fits for the other stellar mass bins give qualitatively similar results .
figure [ psat_model_example ] shows the median value ( thick blue lines ) of @xmath206 as a function of log halo mass and @xmath53 confidence intervals in each stellar mass bin . despite the significant covariance of the model parameters ,
the shape of the passive fraction models for satellites is well determined .
the horizontal black lines show the halo mass range that includes @xmath322 of the satellites in each stellar mass bin .
the average passive fractions in the two halo mass bins above and below @xmath209 , presented as the thick blue lines in figure [ passfrac_mhalo ] , are shown for each stellar mass bin in figure [ psat_model_example ] by the black points .
we add in figure [ psat_model_example ] an additional test of the result presented in section [ sec_tquench_mhalo ] that the quenching time is largely independent of halo mass .
we compute a single quenching time per stellar mass bin without binning the data in halo mass
. then we compute which fraction of mock galaxies have @xmath323 as a function of halo mass .
this is converted in a probability of being passive as a function of halo mass which we show as solid red lines ( with @xmath53 confidence intervals as dashed lines ) in figure [ psat_model_example ] .
the agreement with the best fit values of @xmath206 is remarkable in most of the stellar mass bins , further supporting the result of a quenching timescale that is independent of halo mass .
the @xmath2 points in figures [ convfracz ] and [ tquenchz ] are obtained from a sample of galaxies in the local universe selected from the sdss @xcite survey .
we use the data from the sdss dr8 database @xcite cross correlated with an updated version of the multi - scale density catalog from @xcite ( with densities computed according to equation [ eqsigma ] ; updated dr8 catalog as used by * ? ? ?
* ; * ? ? ?
sdss dr8 includes 5 color @xmath324 imaging of 14555 square degrees .
the spectroscopic part of the survey provides redshifts for 77% of objects brighter than a limit of @xmath325 across 8032 square degrees .
our sample is derived from the spectroscopic database .
luminosities are computed by k - correcting and adding the distance modulus to the petrosian r - band magnitude .
k - corrections are performed using the k - correct idl tool @xcite .
we select as primary galaxies those with @xmath326 mag and @xmath327 .
in contrast to the method we use at high redshift the sample of neighbours ( galaxies used to calculate the density in equation [ eqsigma ] ) is restricted to @xmath328 mag .
this ensures a volume limited sample for the neighbours in this redshift range , while for the primary galaxies we correct for volume incompleteness using @xmath329 corrections .
the primary sample numbers @xmath330 galaxies .
stellar masses and star formation rates are obtained from the jhu - mpa catalogues updated to dr7 @xcite . for this work we use the density computed on a fixed scale of 1 mpc , with a velocity cut of @xmath331 .
this scale is larger than what we use in the 3d - hst sample in order to take into account the growth of structure with cosmic time .
we stress that our results do not significantly depend on the scale chosen because the halo mass calibration is performed self consistently and we only compare calibrated quantities across the two samples .
we have further computed stellar mass ranks for each primary galaxy in the adaptive aperture as described in section [ sec_censat ] .
one limitation of the sdss spectroscopic strategy is that not all the spectroscopic targets can be actually observed because two fibers can not be placed closer than @xmath332 on the sky and each patch of the sky is only observed once ( although with small overlaps between adjacent spectroscopic plates ) . as a result
the spectroscopic catalogue does not contain all the sources detected in the imaging .
spectroscopic incompleteness is taken into account in the computation of the densities as described by @xcite , and we further consider it when we match to the mock galaxy sample .
passive galaxies are selected using the specific star formation rate ( @xmath333 ) as a tracer . for consistency with previous studies ( e.g. * ? ? ?
* ) we define passive galaxies those with @xmath334 .
we note that this corresponds to a @xmath335 dex offset from the main sequence of star forming galaxies at @xmath2 , which is consistent with the division of uvj star forming from uvj passive galaxies adopted in section [ sec_passfrac_mhalo ] .
we generate a model galaxy sample that matches the stellar mass and density distributions of the sdss observational catalogue .
to do so we take the sam from @xcite at the @xmath2 snapshot of the millennium simulation . in this case
we do not use lightcones but a three dimensional box because of the large area covered by sdss and the single redshift bin .
densities are computed by projecting one of the axes of the box into a redshift axis as described in @xcite .
we set an aperture size of 1 mpc , a velocity cut @xmath331 , and we compute densities according to equation [ eqsigma ]
. the model sample does not suffer from spectroscopic incompleteness ; on the other hand the distribution of @xmath336-band magnitudes does not match perfectly the one obtained from the observations . to overcome both those issues at once we employ a method that iterates on the magnitude limits for the primary and the neighbours samples until the number density and the density distribution of the selected sample match the observational data . before doing that we need to derive the total number of photometric galaxies in the sdss dr8 footprint ( more precisely in the area followed up by spectroscopy ) that would have been observed if fiber collisions were not a limitation .
we query the sdss database for the number of galaxies in the spectroscopic database and the number of galaxies in the photometric database that would satisfy the criteria for spectroscopic follow - up .
the ratio of those values is 0.769 .
therefore the number density of mock galaxies needs to be @xmath337 where @xmath338 is the number density of primary galaxies in our observational catalogue once we account for @xmath329 corrections .
the absolute magnitude cuts we set in the models using this iterative method are @xmath339 mag and @xmath340 mag for the primary and the neighbour samples respectively .
we note that these cuts are up to 1 mag deeper than those used in the sdss sample .
this difference arises in a non perfect match of the @xmath336-band luminosity function , while stellar mass functions are better matched between the sam and the sdss data .
figure [ mstar_dens_compare_sdss ] shows that , with this choice of magnitude limits , both the density and the stellar mass distributions are well matched .
this is a critical step to trust our bayesian approach to halo mass and central / satellite status . as a last step we assign to each sdss galaxy ( and to model galaxies )
a probability that it is central ( @xmath149 ) or satellite ( @xmath150 ) and the halo mass pdfs @xmath163 and @xmath164 as described in section [ sec_halomass ] .
figure [ threeplots_mhalo_sdss ] shows the passive fraction for centrals and satellites , conversion fractions and satellite quenching timescales derived for the sdss sample as described in sections [ sec_passfrac_mhalo ] , [ sec_convfrac_mhalo ] , and [ sec_tquench_mhalo ] .
section [ sec_redshiftevo ] contains the scientific discussion of these results in the context of the evolution of satellite quenching efficiency and timescales from @xmath2 to @xmath277 .
the environmental properties of 3d - hst galaxies are made available at http://dx.doi.org/10.5281/zenodo.168056 .
conditional halo mass pdfs given that each galaxy is a central or a satellite and covering the range @xmath341 with 100 uniform bins are also available as separate tables in the same repository .
table [ envtableexample ] gives an example of the quantities provided in the catalog and the description of the columns follows : * \(1 ) 3d - hst field * \(2 ) 3d - hst photometric i d from @xcite * \(3 ) 3d - hst spectroscopic ( grism ) i d from @xcite * \(4 ) fraction of the 0.75 mpc aperture in the photometric catalogue * \(5 ) density of galaxies in an aperture of 0.75 mpc radius ( see eq . [ eqsigma ] ) * \(6 ) overdensity of galaxies in an aperture of 0.75 mpc radius ( see eq .
[ eqoverdens ] ) * \(7 ) stellar mass rank in the adaptive aperture * \(8 ) and ( 9 ) probability that the galaxy is a central or a satellite * ( 10 ) , ( 11 ) , and ( 12 ) @xmath342 , @xmath343 , and @xmath344 percentile of the log halo mass cumulative pdf given that the galaxy is a central * ( 13 ) , ( 14 ) , and ( 15 ) @xmath342 , @xmath343 , and @xmath344 percentile of the log halo mass cumulative pdf given that the galaxy is a satellite | we make publicly available a catalog of calibrated environmental measures for galaxies in the five 3d - hst / candels deep fields . leveraging the spectroscopic and grism redshifts from the 3d - hst survey , multi wavelength photometry from candels , and wider field public data for edge corrections ,
we derive densities in fixed apertures to characterize the environment of galaxies brighter than @xmath0 mag in the redshift range @xmath1 . by linking observed galaxies to a mock sample ,
selected to reproduce the 3d - hst sample selection and redshift accuracy , each 3d - hst galaxy is assigned a probability density function of the host halo mass , and a probability that is a central or a satellite galaxy .
the same procedure is applied to a @xmath2 sample selected from sdss .
we compute the fraction of passive central and satellite galaxies as a function of stellar and halo mass , and redshift , and then derive the fraction of galaxies that were quenched by environment specific processes . using the mock sample , we estimate that the timescale for satellite quenching is @xmath3 gyr ; longer at lower stellar mass or lower redshift , but remarkably independent of halo mass .
this indicates that , in the range of environments commonly found within the 3d - hst sample ( @xmath4 ) , satellites are quenched by exhaustion of their gas reservoir in absence of cosmological accretion .
we find that the quenching times can be separated into a delay phase during which satellite galaxies behave similarly to centrals at fixed stellar mass , and a phase where the star formation rate drops rapidly ( @xmath5 gyr ) , as shown previously at @xmath2 .
we conclude that this scenario requires satellite galaxies to retain a large reservoir of multi - phase gas upon accretion , even at high redshift , and that this gas sustains star formation for the long quenching times observed . |
in the theory of nonlinear systems , the logistic map @xmath0 with @xmath1 , @xmath2 $ ] , and @xmath3 , is well - known to provide one of the simplest example of what is referred to as a _ chaotic _ system . that is , a system which , for some range of its parameters , possesses at least one bounded orbit @xmath4 such that ( i ) @xmath5 is not asymptotically periodic and ( ii ) the lyapunov exponent , defined in the context of eq.([log ] ) by the usual relation @xmath6 is greater than zero @xcite . for the logistic map at @xmath7 , specifically , @xmath8 is positive and equals @xmath9 for almost all @xmath10
. this value can be easily calculated by noting that the logistic map is conjugate to the tent map or by having recourse to the fact that the dynamical system defined by eq.([log ] ) is ergodic for this particular value of @xmath11 @xcite . in this latter case ,
one is justified to calculate analytically the lyapunov exponent as an ensemble average @xmath12 the function @xmath13 \label{inv1}\ ] ] being the invariant density of the map satisfying @xmath14 , where @xmath15 is the preimage of an arbitrary subset @xmath16 of the unit interval @xcite . in this paper , we shall focus on another property of the logistic map at @xmath17 , namely the existence of the following closed - form solution @xmath18 where @xmath19 .
the importance of this formula relies evidently on the fact that @xmath20 can be evaluated directly , for any initial point @xmath21 , without actually computing the intermediate values @xmath22 . in the remaining of this work
, we follow the steps used in the derivation of the above formula to construct similar expressions involving general terms of the form @xmath23 . in doing so
, we shall see that a whole new family of polynomial maps can be defined which , essentially , generalize the logistic map by preserving the density eq.([inv1 ] ) , and whose lyapunov exponents can be calculated easily . other properties of these newly defined dynamical maps are also studied .
in particular , we shall point out some similarities between the maps and the tschebysheff polynomials , and prove , finally , that the family of maps as a whole is closed under the composition of functions .
the explicit solution of the logistic map at the particular value @xmath7 can be derived by substituting in @xmath24 the change of coordinates @xmath25 , valid for @xmath2 $ ] , and by using the trigonometric identity @xmath26 in order to obtain @xmath27 .
this is equivalent to the map @xmath28 , which has the explicit solution @xmath29 hence , the complete solution in terms of the coordinate @xmath30 must correspond to eq.([clo ] ) . following these steps
, one can imagine to construct a map having the solution @xmath31 by expressing @xmath32 in terms of @xmath33 with the identity @xmath34 in this case , one gets @xmath35 as a possible dynamical map on @xmath36 $ ] having the required solution . more generally , one can use the recurrence formula @xmath37-\sin [ ( n-2)\theta ] \label{rec}\ ] ] to define a whole set of maps which express @xmath38 in terms of @xmath39 .
this set , the _ sine functions set _ , is defined specifically as @xmath40 where @xmath41 ^ 2 $ ] , and @xmath42 note that the intermediate functions @xmath43 are expressed as such in order to verify eq.([rec ] ) with the variable change @xmath44 .
c @xmath45 + + + + + + table 1 .
first five members of @xmath46 . using these definitions
, we calculate the first five functions of @xmath46 listed in table 1 . obviously , by construction of @xmath46 , @xmath47 is the logistic equation itself with parameter @xmath7 . from a more general perspective
, it can also be seen that the maps @xmath45 are degree @xmath48 polynomials whose leading coefficient , i.e. , the coefficient of the highest degree term , is equal to @xmath49 in absolute value .
these two results , satisfied by any function @xmath45 , is proven more formally in ref.@xcite .
note that the latter property allows us to extend the similarity with the logistic map by parameterizing the functions of @xmath46 in the following manner @xmath50 with @xmath51 .
we call the set of functions @xmath52 the _ family of _ @xmath45 , which can be characterized numerically by bifurcation diagrams and lyapunov spectrums such as the ones shown in figure 1 .
many of the interesting properties of the logistic map at @xmath7 can be investigated more intuitively by making explicit the fact that eq.([smap ] ) is equivalent to a shift map @xmath53 on the binary expression of @xmath54 @xcite .
indeed , if we express @xmath54 as a binary number @xmath55 then applying eq.([smap ] ) to @xmath54 is equivalent to shifting all the bits of @xmath54 to the left and dropping the integer part . in other words , @xmath56 where @xmath57 for @xmath58 , and @xmath59 .
not surprisingly , the same is true for the maps @xmath45 , since eq.([smap ] ) was the guideline in defining the family @xmath46 .
however , in the case of @xmath45 , the shift map to consider takes effect on @xmath54 written in base @xmath48 .
this follows from the following result which generalizes effectively the solution of eqs.([clo ] ) and ( [ smap ] ) .
_ theorem 1 .
_ let @xmath60 be the orbit of @xmath21 under @xmath61 .
if we write @xmath25 , then we have that @xmath62 where , as usual , @xmath19 .
we omit the proof of this theorem as it follows directly from the next lemma .
_ lemma 1_. consider @xmath45 as defined previously .
we have that @xmath63 _ proof : _ the result is obvious for @xmath64 and @xmath65 .
suppose eq.([sin1 ] ) true for @xmath66 and @xmath67 , that is to say @xmath68(\sin ^2\theta ) \\
& = & \sin ^2[(n-1)\theta ] , \end{aligned}\ ] ] and @xmath69 $ ] . then , for @xmath70@xmath71 ^ 2 \\ & = & [ 2\cos \theta \sin ( ( n-1)\theta ) -\sin ( ( n-2)\theta ) ] ^2 \\ & = & \sin ^2(n\theta ) , \end{aligned}\ ] ] where we have used the identity ( [ rec]).@xmath72 note that we could have proceeded to a similar generalization of the logistic map using cosine functions instead of sine functions , while preserving its shift property .
one possible way of achieving this is to define the _ cosine functions set _
@xmath73 where @xmath74 ^ 2 $ ] , and @xmath75 contrary to the @xmath43 s , the functions @xmath76 have the interesting property that they are polynomials of degree @xmath48 .
in fact , the set @xmath77 coincides with the set of tschebysheff polynomials on the unit interval @xcite , the latter set satisfying the exact same recurrence formula as eq.([rec1 ] ) .
we thus have that @xmath78 must constitute a set of orthogonal polynomials , i.e. , @xmath79 for all integers @xmath48 and @xmath80 , where @xmath81 is the delta - kronecker function .
this fact can be further proved using the property @xmath82 , well - known to be satisfied by the tschebysheff functions .
note that @xmath83 is also a set of orthogonal functions ; its members satisfy indeed the relation @xmath84 . in the remaining of this work
, we shall restrain our study to the set @xmath46 , since the maps @xmath45 are directly related to @xmath85 by the expression @xmath86 hence , as far as their dynamics are concerned , the functions @xmath45 and @xmath87 are totally equivalent .
the analysis of the chaoticity properties of a map @xmath88 is greatly simplified by studying conjugate maps of @xmath88 which are obtained by applying a global change of variables .
recall that two maps @xmath89 and @xmath90 are _ conjugate _ if there exists a homeomorphism , i.e. , a bijective and continuous map @xmath91 such that @xmath92 .
the function @xmath93 is called a _
conjugacy_. in the context of @xmath94 , a possible conjugate function of @xmath45 can be constructed as follows .
let @xmath95 be a piecewise linear function ( a generalized tent map ) defined on subintervals @xmath96 $ ] of @xmath97 $ ] by setting @xmath98 with @xmath99 .
_ theorem 2_. @xmath45 is conjugate to @xmath95 with conjugacy @xmath100 .
_ proof _ : first note that @xmath101 is both continuous and bijective on the interval @xmath97 $ ] .
now , on the first hand we have that @xmath102 .
on the other hand , @xmath103 thus , we have proved that @xmath104 for all integers
@xmath48.@xmath105 figure 2 depicts the graphs of @xmath45 and the corresponding @xmath95 for @xmath106 .
( @xmath107 is only the tent map since , as we mentioned , @xmath47 is the logistic map . ) in each case , note that @xmath45 is a `` smooth '' version of @xmath108 .
_ remark_. the functions @xmath85 are pairwise conjugate to the functions @xmath108 with conjugacy map @xmath109 .
evidently , since the conjugacy is an equivalence relation , this has the consequence that any @xmath87 is conjugate to @xmath45 .
in this section , we evaluate the invariant densities of the maps @xmath45 and their lyapunov exponents based on the conjugates found in the previous section . as one would expect from the similarity of the graphs of @xmath110 and @xmath111 in figure 2 , the functions @xmath95 should exhibit a unique and constant invariant density @xmath112,\ ] ] just as it is the case for the tent map .
this can be proven using adler and bowen s results on markov transformations @xcite , ( see also @xcite ) . at this point
, the invariant density of the functions @xmath45 can be found using the transformation formula @xmath113 where @xmath114 so as to obtain @xmath115 the unicity of @xmath116 is assured by the fact that @xmath117 is also unique , which means that both functions @xmath45 and @xmath95 must be ergodic @xcite . now , to evaluate the lyapunov exponent of @xmath45 we may use the fact that @xmath118 for almost all @xmath119 $ ] to infer that @xmath120 almost everywhere in the case of @xmath95 .
accordingly , since lyapunov exponents are invariant under smooth and differentiable coordinate transformations @xcite , we have the following theorem .
( a more extensive proof of this result , which takes care of the pathological points where @xmath121 is not defined , is contained in ref.@xcite . ) _ theorem 3 . _
the lyapunov exponent of @xmath45 is @xmath122 almost everywhere ( with respect to the invariant measure @xmath116 ) .
the above theorem shows that the members of @xmath46 are non - conjugate to each other simply because they possess different lyapunov exponents .
it also shows that @xmath123 , and consequently the set of tschebysheff polynomials , are sets of _ chaotic _ maps .
indeed , @xmath124 for @xmath125 , and by using the shift property of @xmath45 we can choose @xmath126 , with @xmath54 irrational , to build an orbit that is not asymptotically periodic .
another way to convince ourselves that all the polynomials in @xmath46 have chaotic orbits is to use the celebrated result `` period-3 implies chaos '' @xcite , and find an initial point @xmath21 of period 3 for each @xmath45 .
for instance , for a @xmath127 let @xmath126 where @xmath128 again , using the shift map property , we must have @xmath129 we thus extended the chaoticity properties of the logistic map to an infinite family of polynomials .
to complete the study of the properties of @xmath46 , we now deduce that it is an abelian monoid with respect to the composition of functions ( @xmath130 ) .
a monoid , precisely , is a non - empty set @xmath131 together with a binary associative operation , say @xmath132 , such that @xmath133 for @xmath134 .
there must also be an element @xmath135 , called the identity element , for which @xmath136 for all @xmath137 .
moreover , a monoid is called abelian if the binary operation is commutative @xcite . in @xmath46 ,
the identity element is @xmath138 . also ,
the composition of function is clearly associative .
now , to prove that @xmath46 is indeed abelian monoid , we verify that it is closed under composition and that this composition is commutative , a condition that is not verified in the case of composition of general functions .
however , before we do so , we present next a new expression of @xmath45 on the unit interval .
_ lemma 2 .
_ for all @xmath45 and @xmath119 $ ] , @xmath139 _ proof : _ let @xmath119 $ ] .
there exists a @xmath140 $ ] such that @xmath141 , and thus @xmath142 .
now , from lemma 1 we have @xmath143.@xmath72 _ theorem 4 . _ ( monoid property ) let @xmath144 and @xmath145 be any positive integers .
we have that @xmath146.\ ] ] _ proof : _ for @xmath144 and @xmath145 given , consider @xmath147 and @xmath148 .
then , for any @xmath119 $ ] , we obtain from lemma 2 @xmath149 \nonumber \\ & = & \sin ^2[n_1\arcsin ( \sin ( n_2\arcsin \sqrt{x } ) ) ] \nonumber \\ & = & \sin ^2(n_1n_2\arcsin \sqrt{x } ) \nonumber \\ & = & s[n_1n_2](x).\end{aligned}\ ] ] obviously , @xmath150(x)=s[n_2n_1](x)$ ] , so the composition is commutative.@xmath72 as a direct consequence of the monoid property , @xmath151-periodic points of a certain polynomial @xmath45 can be looked at as fixed points of the function @xmath152 where @xmath153 .
furthermore , a polynomial @xmath45 of very high degree can be computed easily by decomposing its expression using lower degree polynomial of the family @xmath46 .
explicitly , consider @xmath154 .
we say that @xmath45 is a _ prime element _ of @xmath46 if @xmath48 is a prime number . using this definition
, we have as a result of theorem 4 and the fundamental theorem of arithmetic that any polynomial @xmath45 must possess a unique decomposition in prime elements of @xmath46 .
to conclude , note that our study of the sine functions , written in the form @xmath155 , have been restricted to positive integers @xmath48 . in a similar manner , it could be interesting to investigate functions of the type @xmath156 with @xmath157 real .
one observation about this extra generalization is that , as for @xmath158 , the function @xmath159 does not exhibit chaotic properties for @xmath160 .
the function @xmath161 , for example , is conjugate to @xmath162 , and has all of its orbits attracted to @xmath163 . yet ,
this is not surprising since the lyapunov exponent of this map must be @xmath164 .
this brings us to conjecture that @xmath165 must admit chaotic behavior if and only if @xmath166 , considering that the lyapunov exponents of @xmath159 should be @xmath167 . a complete proof of this result ,
however , can not be given here using the same symbolic dynamic approach used for @xmath45 , for the simple reason that the expression of a point in `` base @xmath48 '' makes sense only if @xmath48 is an integer greater than @xmath168 .
v.p . would like to thank a. mingarelli for helpful discussions .
this work was supported in part by the national sciences and engineering research council of canada ( nserc ) through the es a scolarship program .
this definition of chaos is not necessarily equivalent to the more widely held definition based on topological transitivity and sensitivity to initial conditions ( see , e.g. , refs.@xcite ) .
however , the conditions of chaoticity given in the text , which are used throughout this work , are more easy to verify in practice , and nevertheless capture most of the essense of what is understood as chaos . | a family of non - conjugate chaotic maps generalizing the well - known logistic function is defined , and some of its basic properties studied .
a simple formula for the lyapunov exponents of all the maps contained in this family is given based on the construction of conjugacies .
moreover , it is shown that , despite the dissimilarity of their polynomial expressions , all the maps possess the same invariant density . other algebraic properties of the family , which shows some relationship with the set of tschebysheff polynomials , are also investigated . ' '' '' * simple iterated maps , such as the baker map and the logistic map , are the subject of constant fascination .
part of the interest for these systems is linked to the fact that they provide an easy and pedagogical way to understand how complex and chaotic behavior can arise from simple dynamical models .
even more remarkable , yet , is the fact that studies of low - dimensional maps have proven to be fruitful in understanding the basic mechanisms responsible for the appearance of chaos in a large class of dynamical systems ( e.g. , differential flows , high - dimensional maps ) .
one paradigm example of such mechanisms is the so - called period - doubling cascades of fixed points , encountered qualitatively in many physical systems of interest . in this paper
, we enlarge the set of maps known to be chaotic by presenting a generalization of the logistic map .
the generalization , more precisely , enables us to construct an infinite number of one - dimensional maps which are chaotic in the sense that they all have positive lyapunov exponents , and possess at least one orbit that is not asymptotically periodic . * ' '' '' |
in 1995 , the karmen experiment at the rutherford appleton laboratory reported an anomaly in the time distribution of the charged and neutral current events induced by neutrinos from @xmath4 and @xmath5 decays at rest @xcite .
this was ascribed to the production of a new particle , denoted @xmath6 , in the anomalous pion decay @xmath7 with a small branching ratio in the range @xmath8 depending on the lifetime of @xmath6 .
the particle must be neutral since it passes through over 7 m of steel shielding .
the time - of - flight to the detector is @xmath9 , implying that the particle moves non - relativistically with velocity @xmath10 ms@xmath11 .
this requires its mass to be 33.9 mev and its kinetic energy to be @xmath12 kev .
of the pion
muon mass difference . ]
the observed energy in the detector is @xmath13 mev , which must therefore come from the decay of @xmath6 . since
1995 the karmen experiment has been upgraded to significantly reduce the cosmic ray background .
it has recently been reported that the time anomaly persists in the new data @xcite . a time - of - flight likelihood analysis adopting the hypothesis that it is due to a decaying particle as described above has a negative natural log - likelihood ratio of 9 , i.e. less than 1 in @xmath14 chance of being a statistical fluctuation .
the significance is thus sufficiently high that we are motivated to reexamine its physical origin .
there have already been several proposals to explain the karmen anomaly @xcite . in ref.@xcite , the authors considered in detail the possibility that @xmath6 is a neutrino and concluded that a @xmath15 doublet neutrino was excluded by existing data .
this was further reinforced by the subsequent improvement @xcite in the experimental upper limit on the branching ratio , @xmath16 versus the minimum value of @xmath17 required in the doublet neutrino interpretation .
however a sterile neutrino interpretation was found to be consistent , within strict limits on the mixing parameters ( see also ref.@xcite ) , although this may still be in conflict with astrophysical and cosmological constraints @xcite . in ref.@xcite
a solution was proposed based on the anomalous muon decay @xmath18 , where @xmath6 is taken to be a scalar boson of mass 103.9 mev .
however this implies too large a value for the energy released in the @xmath6 decay and the required branching ratio is also constrained by the recent bound br@xmath19 ( 90% c.l . )
thus it is necessary to add to the model 2 other scalar bosons into which @xmath6 can cascade decay in order to dilute the energy @xcite .
this model appears viable but is somewhat baroque . in ref.@xcite
, a supersymmetric solution was considered .
the @xmath6 particle was interpreted as a photino ( or zino ) and the anomalous pion decay @xmath20 was assumed to proceed via the r - parity violating operator @xmath21 .
the same operator then enables the photino to decay radiatively as @xmath22 via a one - loop diagram with a @xmath23 quark and @xmath24 squark in the loop .
however the expected peak at 17 mev has not been reported in the new data @xcite on the energy spectrum of the anomalous events , so a 2-body decay for the @xmath6 particle seems disfavoured
. therefore this model @xcite may not be viable in its present form .
we present below the necessary extension to produce a 3-body decay for such a light neutralino . in the superpotential to be unnaturally small , @xmath25 mev .
moreover this scenario implies a mev mass @xmath26 which is definitively ruled out by cosmological and astrophysical arguments @xcite . ]
we consider the lightest neutralino in supersymmetry , @xmath27 , to be the hypothetical @xmath6 particle , with mass @xmath28 mev .
this will also be the lightest supersymmetric particle ( lsp ) in our model .
since @xmath6 is effectively stable on collider time - scales ( this is quantified below ) our model will experimentally look very similar to the mssm .
now in a gut - inspired mssm , @xmath29 , and assuming this relation requires @xmath30 gev from current lep data @xcite .
thus in order to obtain a very light neutralino we must consider @xmath0 and @xmath1 to be independent parameters .
a small @xmath1 implies at least one light chargino which is excluded by experiment , while a small @xmath0 implies that the lsp will be dominantly bino .
we will quantify this below and determine regions in the @xmath31 parameter space consistent with all experimental limits .
the solutions indeed turn out to be dominantly bino with a small higgsino contribution .
furthermore we invoke 2 non - zero r - parity violating operators .
the pion decay @xmath32 proceeds through the operator @xmath33 and the leading order feynman diagrams are shown in fig .
[ fig : pion ] .
the neutralino is assumed to decay as @xmath34 through either @xmath35 or @xmath36 .
note that this is the only kinematically accessible tree - level 3-body visible decay for such a light lsp .
.,width=377 ] in fig .
[ fig : karmensol ] we show the values of the branching ratio for @xmath37 and lifetimes @xmath38 which are compatible with the karmen data @xcite . in order to determine the required range of the couplings @xmath39 , @xmath40 in our model which are consistent with the solutions in fig .
[ fig : karmensol ] , we must first determine the pion branching ratio in terms of the supersymmetric parameters .
the partial width as computed from the diagrams in fig .
[ fig : pion ] is @xmath41 where @xmath42 and @xmath43 denote the pion and the muon masses , @xmath44 denote the corresponding scalar fermion masses , @xmath45 are the first generation _ current _ quark masses , and @xmath46 is the charged pion decay constant .
the constants @xmath47 refer to the neutralino coupling and are given in table [ tab : picp ] both for the general case and for the limiting cases of either a pure bino or photino neutralino .
the phase space factor is given by @xmath48 [ m^2_\pi-(m_\mu - m_{\tilde\chi^0_1})^2])^{1/2}/(2m_\pi)$ ] .
in the appendix we give some details of how this result is obtained .
-particle lifetime .
the hashed area denotes the experimental upper bound ( [ psi]).,width=377 ] .neutralino coupling coefficients for the pion decay . [ cols="^,^,^,^ " , ] [ tab : photon ] with the higher luminosities expected at the b - factories kek - b and babar , a few events may be expected .
the standard model ( sm ) cross section at this energy is 2.3 fb , corresponding to @xmath49 events at kek - b and @xmath50 events at babar .
the statistical uncertainty still exceeds the signal rate so we do not expect any sensitivity to a light neutralino . at the nlc
we expect a substantially higher number of events .
the sm cross section for the same cuts is about 0.35 pb for 3 neutrinos @xcite corresponding to @xmath51 events , with a small statistical error of @xmath52 events .
thus this can provide a test of our model . in our model , @xmath54 , therefore the decay @xmath55 is kinematically accessible , and
the @xmath56 can be considered to be effectively massless , just like a neutrino .
the lsp decays outside the lep detectors , thus the process @xmath57 will contribute to the invisible width of the @xmath53 .
the current measurement of the invisible @xmath53 width translated into the number of light neutrino species is @xcite @xmath58 so we must require that @xmath59 , where the rhs refers to one neutrino species only
. a pure bino lsp does not couple to the @xmath53 at tree - level .
the dominant contribution will thus come from the higgsino admixtures of the lsp , @xmath60 , in the notation of ref.@xcite .
this enters with the fourth power in the decay rate @xmath57 .
the higgsino has equal strength coupling to the @xmath53 compared to a neutrino , thus yielding the constraint @xmath61 we shall see below that it is straightforward to find regions which satisfy this in @xmath31 parameter space .
it is important to establish whether it is indeed possible to have a neutralino lsp with @xmath62 mev within the mssm .
to this end we have scanned the mssm parameter space @xmath31 with independent @xmath63 , for a neutralino in the mass range @xmath64 this leads to the neutralino iso - mass curves shown in fig .
[ fig : mssmsols ] , taking @xmath65 gev and 2 representative values of @xmath3 .
we have not been able to find any solutions with @xmath66 . in order to obtain such a light neutralino some fine - tuning
is required in the mssm parameters , of about a few parts in @xmath67 for @xmath68 and a few parts in @xmath69 for @xmath70 @xcite .
the fine - tuning is reduced for larger @xmath1 and @xmath2 and small @xmath0 because a light neutralino can then be generated by the see - saw mechanism ; it is reduced for large values of @xmath3 because in the limit @xmath71 there is a zero mass eigenvalue for @xmath72 @xcite .
we have checked that the higgsino contribution always satisfies the bound ( [ higgs ] ) . in order to avoid an observable light chargino we require @xmath73 gev , which eliminates the region below the hashed lines in fig .
[ fig : mssmsols ] for the specified values of @xmath3 .
the lsp is indeed dominantly bino along the solution curves in fig .
[ fig : mssmsols ] .
the second lightest neutralino , @xmath74 , is dominantly wino for @xmath75 gev , while for larger values it is mainly higgsino . for @xmath76 gev , @xmath77 gev , and for @xmath78
gev , @xmath79 gev .
parameter space giving a @xmath62 mev neutralino for @xmath65 gev and 2 representative values of @xmath3 .
the width of the lines is 0.01 mev . below the hashed lines the chargino mass is less than 150 gev
the dotted lines have @xmath80 and the solid lines have @xmath81.,width=302 ] any new physics which couples to the sm can give contributions to the electroweak precision observables via radiative corrections .
the effect of the new physics on vacuum polarization diagrams , the so called oblique radiative corrections , is usually parameterised using either the @xmath82 and @xmath83 parameters of ref.@xcite or the @xmath84 , @xmath85 and @xmath86 parameters of ref.@xcite .
the calculation of these parameters is based on an expansion in @xmath87 , where @xmath88 is the momentum scale of the gauge boson propagator , typically @xmath89 or smaller and @xmath90 is the scale of the new physics , assumed to be well above @xmath89 .
if however there are new _ light _ particles in the spectrum , as in the present case , these approximations are typically insufficient and one must in general also calculate the box or vertex corrections @xcite .
this full calculation is however beyond the scope of our analysis and is not attempted here .
there is one exception however and that is the ratio of the charged to neutral current neutrino
electron / muon scattering events the @xmath91-parameter .
this is defined at @xmath92 and the expansion is thus trivial . in the following
we limit ourselves to a calculation of the contribution to the @xmath91 parameter from the full set of charginos and neutralinos .
the radiative correction to the @xmath91 parameter , @xmath93 , is given by the @xmath94 and @xmath95 self energies with zero momentum flow @xmath96 the dominant sm contributions to @xmath93 arise from the ( heavy ) top quark and the higgs boson .
assuming the mass of the latter to be @xmath97 , and subtracting the sm contributions we are left with the experimental limit on new physics at the @xmath98 level of @xcite @xmath99 we have calculated the contribution to @xmath93 , which we denote @xmath100 , from all charginos and neutralinos for given parameter points @xmath101 .
we find full agreement with the results given in ref.@xcite .
we then determine @xmath100 along the solution curves given in fig .
[ fig : mssmsols ] .
the dotted lines indicate solutions for which @xmath80 , while the solid lines indicate solutions for which @xmath102 . thus there is no conflict at least with the experimental constraint on the @xmath91 parameter .
massive particles are expected to come into thermal equilibrium in the early universe and their relic abundance is essentially the equilibrium value at ` freeze - out ' when their self - annihilation rate drops below the hubble expansion rate . for the neutralino under consideration ,
the self - annihilation cross - section is ( s - wave ) suppressed @xcite so the surviving abundance is rather high : @xmath103 this energy density will be subsequently released when the neutralinos decay and this has the potential to disrupt standard cosmology , in particular primordial nucleosynthesis @xcite .
specifically , since the neutralinos will be non - relativistic during nucleosynthesis , the hubble expansion rate will be speeded up , while the electromagnetic energy generated through the subsequent decays will dilute the nucleon - to - photon ratio , resulting in an increased helium-4 abundance @xcite .
the decay electrons will also compton scatter the thermal background photons to energies high enough to directly alter the abundance of e.g. deuterium through photodissociation @xcite .
the observationally inferred primordial abundances thus enable stringent bounds to be placed on the relic abundance of the decaying particle as a function of its lifetime . for the above abundance ( [ cosmoabund ] ) , the decay lifetime is required to be less than a few thousand seconds in order that the primordial d / h ratio is not reduced below its conservative lower limit of @xmath104 , and further required to be less than a few hundred seconds in order that the primordial @xmath105he mass fraction not exceed its conservative upper limit of 25% @xcite .
thus the cosmological lifetime bound is essentially the same as the one derived earlier ( [ psi4 ] ) from experimental considerations .
very light neutralinos can also be produced through nucleon bremsstrahlung and @xmath106 annihilation in supernovae such as sn 1987a . since the squark /
selectron masses are now restricted to be above @xmath107 @xcite , the neutralinos can not be trapped in the supernova core by scatterings on nuclei or electrons , so will escape freely . the energy lost through this process can be comparable to the neutrino luminosity so may result in significant shortening of the @xmath108 burst . the neutralino luminosity can be decreased by increasing the sfermion mass but it has been shown that consistency with observations of sn 1987a is not possible for any sfermion mass less than @xmath109 tev @xcite .
this constraint is evaded if the neutralino is unstable due to r - parity violation , as in the present case .
however there are then further constraints on the energy released in the decays . given the experimental upper bound ( [ psi4 ] ) as well as the cosmological upper bound on the lifetime
, the decays would have occurred _ within _ the progenitor star .
moreover the lower bound ( [ bounds ] ) on the lifetime implies that the neutralinos can not have decayed within the supernova core .
the electromagnetic energy released in the decays would have been thermalised leading to distortions of the lightcurve .
however the neutralinos under consideration here have a mass which is of the same order as the core temperature @xcite so one must reconsider their production rate in order to quantify this potentially important constraint .
in 1997 , the hera collaborations reported an anomaly in their high @xmath110 data @xcite whose most likely explanation was in terms of r - parity violation @xcite .
however , this required a significant @xmath111 operator .
while this is not excluded by our model it is a completely distinct possibility .
together the two operators might contribute to the decay @xmath112 .
however this does not lead to a significant new bound since the bound on the coupling @xmath39 is already so strict @xcite .
the trilinear r - parity violating couplings we have introduced also generate majorana masses for neutrinos through one - loop self - energy diagrams @xcite : @xmath113 where @xmath114 is a colour factor , and @xmath115 is the left - right mixing term in the sfermion sector . the question then arises whether our model can account also for the superkamiokande data suggesting oscillation of atmospheric @xmath116 into @xmath117 @xcite .
the results indicate that the neutrinos mix almost maximally and that they are nearly mass - degenerate , @xmath118 ev@xmath119 .
if neutrino masses are hierarchical then the natural interpretation is that one of the neutrinos ( presumably the @xmath117 ) has a mass of @xmath120 ev ( although the possibility of a close mass - degeneracy for a heavier @xmath116@xmath117 pair is not excluded ) . now
the sfermion left - right mixing is not well determined , however within a given framework , e.g. supergravity - inspired models , approximate relations such as @xmath121 arise .
thus @xmath39 , the coupling responsible in our model for the pion decaying to the neutralino , generates a mass @xmath122 which is too small to be of phenomenological interest .
the couplings @xmath40 responsible for neutralino decay also generate rather small masses : @xmath123 thus the absolute scale of the masses seems too low to explain the data on atmospheric neutrinos .
however if other r - parity violating couplings are also present , it may well be possible to generate a neutrino mass pattern consistent with the observations , both of atmospheric and solar neutrinos @xcite .
experimentally , our model largely looks like the mssm with non - universal gaugino masses and with a very light lsp .
thus most future tests of the mssm also apply to our model , e.g. chargino pair production
. a specific test would be to identify a very light lsp for example via neutralino pair production @xcite . at an @xmath106 collider
one can study the process @xmath124 where @xmath74 subsequently decays visibly @xcite . in fig .
[ fig : chi2chi1 ] we show the cross section evaluated along our mssm solution curves for both lep2 ( @xmath125 ) and the nlc ( @xmath126 )
. this should be directly observable , provided it is kinematically accessible , i.e. @xmath127 . for the solutions in fig .
[ fig : mssmsols ] .
the solid lines correspond to 0.1 pb@xmath1281 pb , the dashed lines to 10 [email protected] pb , the dotted lines to 1 fb@xmath12810 fb , and the dot - dashed lines to @xmath1291 fb.,title="fig:",scaledwidth=48.0% ] for the solutions in fig . [ fig : mssmsols ] .
the solid lines correspond to 0.1 pb@xmath1281 pb , the dashed lines to 10 [email protected] pb , the dotted lines to 1 fb@xmath12810 fb , and the dot - dashed lines to @xmath1291 fb.,title="fig:",scaledwidth=48.0% ] experimentally the main difference between r - parity violation with a long - lived neutralino lsp and the mssm is the possibility of resonant sparticle production .
the value of @xmath130 is too small for the observation of resonant slepton production at hadron colliders @xcite .
however values of @xmath131 should allow a test for resonant second or third generation sneutrino production at @xmath106-colliders for masses upto @xmath132 @xcite .
one can also test for the first generation via the mechanism described in ref.@xcite .
a further upgrade of the karmen detector may allow a better resolution of the decay of the @xmath6 particle , in particular the angular distribution of the decay products . for reference
we show in fig .
[ fig : angledist ] the differential decay rate of the lsp in our model as a function of the angle between the two final state electrons .
the karmen time anomaly is particularly intriguing because contrary to several other reported @xmath133 effects in the literature , its significance has not diminished with improved statistics , nor has it been explained away as a systematic effect .
in fact the anomaly persists in the karmen-2 data , which has a much reduced background @xcite , with the same characteristics as in the karmen-1 data @xcite .
it would appear that there is no independent experiment which has the sensitivity to reproduce this result .
in particular although the lsnd experiment also studies pions and muons decaying at rest , it lacks the distinctive time - structure of the beam in the karmen experiment necessary to isolate the anomaly .
since karmen-2 acquires only of @xmath134 anomaly events per year of running , it is clear that a definitive resolution of the problem will have to await an upgraded detector with tracking capability .
phenomenological models for the anomaly as due to the production and decay of a new particle are very tightly constrained .
the only viable proposals at present are a singlet neutrino decaying through its large mixing with the @xmath117 @xcite , or a neutralino decaying through violation of r - parity @xcite which we have extended and investigated in detail .
an important lesson from our investigation is that contrary to popular belief a neutralino lighter than even the pion is _ not _ excluded by present accelerator data unless a gut relation between gaugino masses is assumed . whether the karmen anomaly is indeed the first evidence for such a particle is a matter for future experiments to decide .
* note added : * while this manuscript was under review , the e815 ( nutev ) experiment at fermilab reported a search for a 33.9 mev neutral particle produced in pion decay decaying to a partially electomagnetic state such as @xmath135 or @xmath136 ( j.a .
_ , hep - ex/9912062 ) .
no evidence was found for such a particle but the lifetimes probed ( @xmath137 s ) are much smaller than the lower limits ( [ bounds ] ) on the neutralino lifetime in our model so there are no implications .
we note however that the exclusion of such short lifetimes is relevant in the context of the constraints from sn 1987a on the decaying particle hypothesis @xcite .
these constraints have been investigated further in two other recent reports ( i. goldman , r. mohapatra and s. nussinov , hep - ph/9912465 , m. kachelriess , hep - ph/0001160 ) which conclude that our model is excluded by the observations of sn 1987a .
we reserve judgement on this issue for the reasons mentioned earlier .
the rate of the pion decay @xmath138 can be calculated using chiral perturbation theory . to do this we need an effective lagrangian for the 4-fermion interaction @xmath139 , @xmath140 , @xmath56 and @xmath5 with the sfermion degrees of freedom integrated out ( analogous to using the fermi lagrangian in the sm calculation ) .
this gives @xmath141 where we have also fierz - reordered the lagrangian ( and neglected some tensor - tensor interaction terms which can not contribute to the pion decay rate ) .
the matrix element of the axial - vector current between the pion and the vacuum is @xmath142 where @xmath143 and @xmath144 are isospin indices . using eq .
( [ chiral ] ) we obtain @xmath145 contracting this with the pion 4-momentum and using the dirac equation for the up and down quarks yields @xmath146 where @xmath147 are thus the current quark masses . the amplitude for the decay ( [ piondec2 ] ) is then , @xmath148 this gives the decay rate quoted in eq .
( [ pionbr ] ) . the additional contribution to the decay rate @xmath149 given in section [ sec : limits ]
can be calculated in the same way .
s. dawson , nucl .
b * 261 * , 297 ( 1985 ) ; j. butterworth and h. dreiner , nucl .
b * 397 * , 3 ( 1993 ) ; h. dreiner and p. morawitz , nucl .
b * 428 * , 31 ( 1994 ) ; e. baltz and p. gondolo , phys .
d * 57 * , 2969 ( 1998 ) .
j. ellis , k.a .
olive , s. sarkar and d.w .
sciama , phys .
b * 215 * , 404 ( 1988 ) ; j.a .
grifols and e. masso , phys .
b 220 , 591 ( 1989 ) , nucl . phys .
b * 331 * , 244 ( 1990 ) ; k. lau , phys .
d * 47 * , 1087 ( 1993 ) .
d. choudhury and s. raychaudhuri , phys .
b * 401 * , 54 ( 1997 ) ; h. dreiner and p. morawitz , nucl .
b * 503 * , 55 ( 1997 ) ; g. altarelli , j. ellis , g.f .
giudice , s. lola and m.l .
mangano , nucl .
b * 506 * , 3 ( 1997 ) ; j. kalinowski , r. ruckl , h. spiesberger and p.m. zerwas , z. phys .
c * 74 * , 595 ( 1997 ) .
h. dreiner and s. lola , in _
@xmath106 collisions at tev energies _ , hamburg ( 1995 ) , report desy-96 - 123-d ; j. erler , j.l .
feng and n. polonsky , phys .
78 * , 3063 ( 1997 ) ; j. kalinowski , r. ruckl , h. spiesberger and p.m. zerwas , phys
b * 406 * , 314 ( 1997 ) . | we interpret the karmen time anomaly as being due to the production of a ( dominantly bino ) neutralino with mass 33.9 mev , which is the lightest supersymmetric particle but decays into 3 leptons through the violation of r - parity . for independent gaugino masses
@xmath0 and @xmath1 we find regions in the ( @xmath0 , @xmath1 , @xmath2 , @xmath3 ) parameter space where such a light neutralino is consistent with all experiments . future tests of this hypothesis are outlined . |
the study of metals on semiconductors dates back to the nineteenth century and has seen a vigorous recent revival due to tremendous interest in nanotechnology . our ability to manipulate atoms ,
placing them at will on different surface sites to create exotic artificial structures , has led to further investigations @xcite of electronic and transport properties of free and supported nanowires .
the placement of trivalent atoms ( al , ga and in ) on si(001 ) can lead to the formation of low - dimensional structures , exhibiting significant new electronic properties .
one can easily compute the electronic properties of free standing nanowires .
for this , one must first compute the total energy and determine the possible stable structures .
such calculations have indeed been carried out for nanowires @xcite consisting of a wide variety of atoms , _
e.g. _ , k , al , cu , ni , au and si .
a general finding is that a zigzag structure in the form of an equilateral triangle is the most stable @xcite .
this can be understood as arising primarily due to the maximization of coordination number for each atom in a quasi 1d structure .
another structure which also has a local minimum , but not terribly stable , is a wide angle isosceles triangle which somehow is reminiscent of the bulk environment .
for example , si which is a four fold coordinated in the bulk ( tetrahedral angle @xmath1 ) shows @xcite a local minimum at an angle of @xmath2 . in general , free standing nanowires tend to be metallic ( have bands crossing the fermi level ) but these nanowires in practice are to be supported .
silicon is the most widely used substrate for practical applications and the low index surfaces , si(001 ) is the surface of choice . with the downward spiral toward nano devices , it is desirable to study the metallic properties at the lowest possible coverages .
it is in this context that the study of metals like al , ga and in at submonolayer coverages on si(001 ) take on the added importance .
the interaction of metal nanowires with substrate can significantly alter the electronic properties .
low - energy - electron - diffraction ( leed ) , auger electron spectroscopy ( aes ) and scanning tunneling microscopy ( stm ) studies have provided much informations about the interaction of al , ga and in overlayers on the si(001 ) surface at different coverages and at different temperatures @xcite .
ide et al .
@xcite observed 2@xmath32 , 2@xmath33 , 4@xmath35 , and 1@xmath37 structures depending on the coverages less than 1 monolayer ( ml ) of al on si(001 ) and the substrate temperature in their leed and aes experiments .
however , the 2@xmath32 and 2@xmath33 phases were observed around 0.5 and 1/3 ml coverages respectively at low temperatures .
sakamoto and kawanami @xcite performed reflection high energy electron diffraction experiments and established the existence of phases with various symmetries ( 2 @xmath3 3 , 2 @xmath35 , 2@xmath32 and 8@xmath31 ) for ga coverages less than 1 ml and at temperatures @xmath4 350@xmath5 c. later , bourguignon et al .
@xcite examined the evolution of 1ml of ga on si(001 ) with leed and observed all the above mentioned phases and in addition they also observed 1@xmath32 phase at 1 ml coverage .
nogami et al .
@xcite did stm studies and observed 2@xmath32 at and below 0.5 ml coverage of al , ga and in on si(001 ) .
knall et al .
@xcite performed experiment for in adsorption on si(001 ) using rhed , leed , aes and stm and observed transition from 2@xmath32 to 2@xmath31 structure of in as coverage increased .
all the experimental results @xcite point to the fact that at low temperatures al , ga and in form 2@xmath32 structure on the si(001 ) surface at 0.5 ml coverage . at higher metal coverage ,
nagomi et al .
@xcite observed 3d island formation of al on si(001 ) around 1 ml coverage while ordered structures were observed for ga and in on si(001 ) at 1 ml coverage without any island formation @xcite .
in particular , a 1@xmath32 structure was observed for ga , and 2@xmath32 , 2@xmath31 structures were observed for in on si(001 ) . since ga does not form islands at 1ml coverage and it has been argued that al forming islands at 1 ml coverage is unusual , we present a complete theoretical study for al / si(001 ) at 1 ml coverage .
note that several experimental results were explained @xcite in terms of metal dimers oriented parallel to the si dimer rows _
i.e. _ metal dimers were assumed to be directed perpendicular to the si dimers in the reconstructed surface .
this orientation of metal dimers is named as ` ortho ' dimers .
electronic structural calculations were done by batra @xcite for al on si(001 ) at 0.5 ml coverage and he explained the experimentally observed patterns in terms of these ` ortho ' dimers .
later , northrup et al .
@xcite did theoretical calculations on the same system and showed that the orientation of metal dimers parallel to the si dimers was most favorable .
these dimers orientated parallel to the si dimers are called as ` para ' dimers . a series of experiments @xcite performed recently for the structural configuration of metal dimers on the si(001 ) surface indeed confirm the findings of northrup et al .
@xcite at 0.5 ml coverage . to the best of our knowledge
, there is hardly any detailed theoretical study at higher coverages ( 1 ml ) for these systems . in this paper , we put significant effort to perform zero temperature electronic structural calculations for both 0.5 ml and 1 ml coverages of al , ga and in on the si(001 ) surface and compare our results with the experimental and theoretical findings where available .
we attempt to put the work in perspective and fill the holes by presenting new results .
in particular , we find that our calculations at 0.5 ml give results in accord with northrup s predictions and also with the recent experimental results obtained by sakama et al @xcite .
more significantly , we obtain an orientational transition ( ` para ' to ` ortho ' ) for al dimers as we go from 0.5 ml to 1 ml coverage
. our calculations also reveal that the reconstruction of si(001 ) is completely lifted at 1 ml coverage of al .
we find that in and ga interacts differently with the surface ; for these metals the reconstruction is not lifted at 1 ml coverage .
one finds the staggered mixed ( ` ortho ' and ` para ' ) dimers to be the lowest energy structure for in while the mixed dimers are non - staggered for ga .
the important conclusion is that the ` ortho ' dimers , which are energetically unfavorable at 0.5 ml coverage , become viable at higher coverage facilitating ` para ' to ` ortho ' transition as a function of coverage .
also , a previously reported semi - metallic phase of supported zigzag al nanowire is found to be stable against a dimerizing peierls transition unlike @xmath0 .
the soliton @xcite formation in the supported nanowire is argued to be improbable .
the paper is organized as follows .
the calculational parameters are given in sec .
ii and the results and discussions are presented in sec .
iii followed by a summary of our findings in sec .
first principle total energy calculations were carried out within density functional theory at zero temperature using the vasp code @xcite .
the wave functions are expressed by plane waves with the cutoff energy @xmath6 ev . the brillouin zone ( bz )
integrations are performed by using the monkhorst - pack scheme with 4@xmath34@xmath31 @xmath7-point meshes for 2@xmath32 primitive cells .
ions are represented by ultra - soft vanderbilt type pseudopotentials and results for fully relaxed atomic structures are obtained using the generalized gradient approximation ( gga ) .
the preconditioned conjugate gradient method is used for the wave function optimization and the conjugate gradient method for ionic relaxation .
the si(001 ) surface is represented by a repeated slab geometry .
each slab contains five si atomic planes with hydrogen atoms passivating the si atoms at the bottom of the slab .
consecutive slabs are separated by a vacuum space of 9 .
the si atoms on the top four layer of the slab are allowed to relax while those in the bottom layer of the slab and the passivating hydrogen atoms are kept fixed to simulate the bulk like termination . the convergence with respect to the energy cutoff and the number of @xmath7 points for similar structures has been established @xcite .
we have performed total energy calculations using the above parameters for both the ideal and the reconstructed si ( 001 ) surfaces .
for the ideal case , the si atoms on the surface are arranged in a square pattern of side 3.84 . for the reconstructed surface
the si atoms rearrange in such a way that the top layer si atoms form dimer rows along the @xmath8 [ 1@xmath90 ] direction but the atoms move along the @xmath10 [ 110 ] direction to form dimers . here
we consider 2@xmath32 supercell for our calculations . at the 0.5 ml coverage of al , ga and in we start only with the reconstructed surface and at 1 ml coverage we start with both the ideal and reconstructed surfaces . here
we consider various structural arrangements of metal atoms on the si(001 ) surface .
the @xmath11 primitive cell contains two metal atoms which corresponds to 0.5 ml coverage .
the results for al / si(001 ) are elaborated here and the results for ga and in are briefly discussed .
two metal atoms in the cell may be placed in various ways on the surface .
we consider several configurations to identify energetically the most favorable configuration .
the configurations considered here are shown in the fig .
1 . 2 cells closed by thick borders are shown for different configurations of metal atoms at 0.5 ml coverage .
the unfilled circles represent the top layer si atoms and the filled circles represent the metal atoms . ] the metallic dimers configurations on reconstructed si(001 ) shown in figs .
1(a ) , 1(b ) and 1(c ) were considered by batra @xcite .
he found that the configuration in fig .
1(b ) was the most stable one followed by the configuration in fig .
1(c ) . the configurations in figs .
1(c ) and 1(f ) were considered by northrup et al @xcite and they concluded that configuration in 1(f ) was energetically the most favorable one . the arrangement in 1(c ) is referred to as the ` ortho ' orientation of metal dimers and that in 1(f ) is the ` para ' orientation of metal dimers as described earlier by the northrup et al @xcite .
in addition we consider two more configurations and they are shown in figs .
1(d ) and 1(e ) respectively .
the relative energies and the distance between the neighbor atoms for all the configurations are given in table i. all the energies are measured with respect to the reference energy which at 0.5 ml coverage of metal atoms correspond to the configuration shown in fig .
for calculating the reference energy , all the si atoms are held fixed and the metal atoms are allowed to relax along the [ 001 ] direction .
the reference energies for the cases of al , ga , and in are -131.625 ev , -130.714 ev , and -130.255 ev . respectively .
it is clear from table i that the configuration shown in 1(f ) _ i.e. _ , al dimer placed between the si dimer rows with ` para ' orientation has the lowest energy and hence the most favorable configuration .
our results support the calculations first made by northrup et al @xcite and later by brocks et al@xcite .
we also observe that the next most favorable configuration is that shown in 1(e ) which also has ` para ' orientation of al dimer but the dimer is placed within a si dimer row .
however , the energy difference between these two configurations with ` para ' orientation of metal dimers is @xmath4 1 ev . this is because the dimer within the si row [ 1(e ) ] has to reside at much higher position compared to that placed between the dimer rows [ 1(f ) ] and
hence has a weaker binding .
we also note that the energy difference between the ` para ' and ` ortho ' orientation of dimers is even larger , the ` para ' being more stable @xmath4 1.3 ev .
to calculate the potential energy variation of the system as a function of the al dimer orientation , we place the al dimer between the si dimer rows at a fixed height and rotate about the [ 001 ] direction .
the result is shown in fig .
3 . this shows two minima at ` ortho ' and ` para ' orientations respectively and there is a small barrier as one goes from ` ortho ' to ` para ' orientation .
therefore , the al dimers take ` para ' orientation which is strongly more favorable compared to the ` ortho ' orientation .
this result also agrees with the observation in a recent experiment by sakama et al @xcite .
we note that though the si dimers are stretched by 0.26 the reconstruction of the si surface is not lifted at 0.5 ml of coverage .
we have also done similar calculations for ga and in adsorption on the reconstructed si(001 ) surface and found that the ` para ' orientation of ga and in dimers are also favored with relative energies -3.29 ev and -2.90 ev respectively
. the results for ga and in also agree with recent experimental results obtained by sakama et al @xcite .
the surface pattern at 0.5 ml coverage is 2@xmath32 .
it is also clear that all the patterns observed at low temperatures and below 0.5 ml coverage of al , ga and in on the reconstructed si(001 ) can be explained by the ` para ' dimers as proposed by northrup @xcite , but contrary to the earlier proposal by batra @xcite .
2 reconstructed cell as the al dimer is rotated from ` ortho ' to ` para ' orientation .
the al dimer is placed between the si dimer rows of the reconstructed si(001 ) surface . ]
one focus of batra s work @xcite , however , was to find a stable metallic configuration for al .
he reported a zigzag nano - structure shown in fig .
1(b ) which he labeled as @xmath12 , as being stable with a local minimum in energy .
our present calculations support this finding except that we get @xmath13 in contrast to batra s value of @xmath14 .
this overlayer structure of al atoms is reminiscent of a uniform bond @xmath0 structure with the important difference that there is no direct al - al bond .
the nearest neighbor al - al distance ( @xmath4 4 ) in the overlayer is much longer than the bulk al - al distance ( @xmath4 2.8 ) and they see " each other only through the delocalized electrons in the @xmath15-orbitals . in @xmath0 , c - c atoms are not only coupled through @xmath16-electrons but they also form strong uniform @xmath17-bonds leading to a semi - metallic structure .
this structure becomes semi - conducting upon undergoing a bond alternation peierls like transition to double ( d ) , single ( s ) , ds , ds ... repeating pattern of bonds .
it supports the soliton @xcite formation by creating a defective structure of the type ..
ds , ds , sd , sd .. by having two single bonds adjacent to each other .
the al overlayer structure@xcite has two electrons of each al atom tied to a surface si atom in an @xmath18 like fashion while the third electron of al is in a free electron like state along the nanowire ( @xmath8-direction ) . here
x represents a surface si atom which nominally has a single electron in a dangling @xmath15-orbital available for bonding with al .
interatomic distance between al and surface si atoms , d(al - x ) @xmath19 2.4 , implies a strong bond .
the delocalized @xmath15 electron ( one per al atom ) creates a one dimensional semi - metallic system . naturally , under such situation metallic wires may be realized and it will be conducting under small bias conditions .
but then the standard peierls distortion may come into play depending upon the strength of the surface bonds . within the error of our calculations ( estimated at @xmath4 7 mev )
we did not find any lowering of the total energy upon dimerization , unlike @xmath20 .
in fact we noted that a dimerization of al atoms in the zigzag structure by an amount @xmath21 raised the total energy by @xmath4 0.1 ev . thus the surface bonds formed by supported nanowires can lock the metallic structure in place .
there is no possibility of a soliton formation in the al overlayer structure because of the large al - al distance .
however , it will be interesting to look for soliton behavior in other overlayers .
we consider the adsorption of four metal atoms on the si(001 ) surface of a 2@xmath32 unit cell which corresponds to 1 ml coverage .
results for al , ga , and in on si(001 ) are presented here . various possible configurations considered are shown in fig .
calculations are done for both the ideal and the reconstructed si surfaces to find energetically the most favorable configuration .
the 2@xmath32 unit cells are bordered by thick lines in fig .
4 . the configurations considereed include para - para and ortho - ortho metal dimers on reconstructed ( figs .
4(a ) - 4(d ) ) and ideal ( figs .
4(e ) - 4(h ) ) on si(001 ) surface .
in addition , we also consider mixed dimers ( ortho and para ) on reconstructed surfaces ( figs .
4(i ) - 4(l ) ) .
mixed dimers on ideal si surface were not stable and readily caused the silicon surface to reconstruct . 2 cells closed by thick borders are shown for different configurations metal atoms at 1 ml coverage .
the unfilled circles represent the top layer si atoms and the filled atoms represent the metal atoms ( al or ga or in ) . ]
the relative energies for the relaxed structures of al , ga and in on the reconstructed surfaces are given in table ii for four possible configurations and the same on the ideal surfaces are shown in table iii .
the energies for the configurations with mixed dimers are given in table iv .
the reference energies at 1 ml coverage of metal atoms are obtained for the configuration shown in fig .
2(b ) where all si atoms are held fixed the metal atoms are allowed to relax along the ( 001 ) direction .
the reference energies for the cases of al , ga and in are -137.963 ev , -136.485 ev and -136.044 ev respectively .
a general conclusion is that with increasing coverage beyond 0.5 ml , ` ortho ' dimers become part of the low energy configurations leading towards a ` para ' to ` ortho ' transition at 1 ml coverage . for in , mixed staggered dimers ( ` ortho ' and ` para ' ) give the lowest energy configuration . for ga ,
mixed dimers are non - staggered , while for al ` para ' to ` ortho ' transition of dimers is complete .
this is an important conclusion because at or below 0.5 ml ` ortho ' dimers were not favorable .
also , the puzzle that ` ortho ' and ` para ' dimers are both seen is resolved due to the preponderance of ` ortho ' dimers at increasing coverages .
let us first consider the structure consisting of al dimers with ` para ' orientation as it was the most favorable structure at 0.5 ml coverage .
two configurations are possible with ` para ' orientation of al dimers : ( i ) the configurations shown in figs . 4(a ) and 4(e ) on the reconstructed and ideal surfaces respectively and ( ii ) the configurations shown in figs .
4(b ) and 4(f ) on the reconstructed and ideal surfaces respectively .
after complete relaxation , we note that the final low energy structure is independent of the initial surface pattern ( reconstructed /ideal ) .
when we start from an ideal surface , the si atoms dimerize and the low energy structure is the one with reconstructed si surface .
same final structure is achieved when we start form the reconstructed surface .
also note that the relaxed system with the staggered dimers configuration in 4(b ) consisting of parallel dimers on the reconstructed surface is less favorable compared to that on the same reconstructed surface with the configuration in 4(a ) . comparing 4(b ) and 4(f ) we find that the final energies of the systems differ by a large amount : staggered dimers on the ideal surface are least favorable .
next we consider the configurations in 4(c ) and 4(d ) .
these structures are obtained by just rotating the al dimers in 4(a ) and 4(b ) by 90 degree ( _ i.e. _ , dimers now take ` ortho ' orientation ) . from table
ii it is clear that these configurations are more favorable compared to those in 4(a ) and 4(b ) .
we find that after complete relaxation , the total energies and the structures of these two configurations become independent of the initial surface pattern .
in addition , we consider the configurations consisting of mixed dimers ( ` ortho ' and ` para ' ) as shown in figs .
4(i ) - 4(l ) .
similar configurations on ideal surface are unstable , the si surface reconstructs itself and eventually leads to the configurations 4(i ) , 4(j ) and 4(k ) .
again for al , the configurations 4(j ) and 4(l ) are unstable . from table
iv we see that the configuration with mixed al dimers shown in fig .
4(i ) has a deep local minumum .
this configuration is more favorable compared to the configurations consisting of only ` para ' orientation of al dimers ( i.e. , figs .
4(a ) and 4(f ) ) .
however , comparing the energies for configurations 4(i ) and 4(g ) we find that the lowest energy structure is the one shown in fig .
4(g ) where the reconstruction is completely lifted .
all the si bonds are saturated here . since the reconstruction is completely lifted at 1 ml coverage , we conclude that the breaking of si dimer bonds is initiated just beyond 0.5 ml of coverage .
most interestingly we note that an orientational transition for the al dimers takes place as one goes from 0.5 ml coverage to 1 ml coverage : the al dimers change their orientation from ` para ' to ` ortho ' after breaking the si dimer bonds .
this is also supported by the charge density plot for the lowest energy configurations at 0.5 and 1 ml coverage as shown in figs .
5(a ) and 5(b ) respectively . earlier
, it was argued that at 0.5 ml coverage all si bonds are saturated and therefore beyond 0.5 ml coverage al atoms may reside on top of each other and form 3d island .
one experimental study observed the 3d island formation beyond 0.5 ml of al coverage @xcite .
our calculation , however , suggests that for al an orientational transition can take place at 1 ml coverage and the surface structure becomes regular having a 1@xmath22 pattern where all bonds of si and al are saturated due to removal of the reconstruction of the si surface .
this leads us to speculate that further experiments on al may seek out 1@xmath32 pattern at 1 ml coverage .
furthermore , at coverages between 0.5 ml and 1 ml both the ` para ' and ` ortho ' al dimers may coexist .
this also clarifies why experiments at times have reported both ` para ' & ` ortho ' dimers because precise coverage is a difficult parameter to quantify .
now we consider the ga adsorption at 1 ml coverage . if we start with the reconstructed surface we find that the configuration in fig .
4(i ) with mixed dimers leads to the lowest energy structure where the reconstruction of the si surface is not lifted during the relaxation process .
the next best configuration is shown in fig .
4(g ) and consists of only ` ortho ' dimers .
the structure remains ideal even after complete ionic relaxation .
the configurations in figs .
4(j ) and 4(l ) are unstable . comparing final energies for various configurations on ideal and reconstructed surfaces
we observe that the configuration in 4(i ) is energetically the most favorable structure .
experiments indicate observing the 1@xmath32 structure shown in 4(g ) @xcite . here
the reconstruction of the si surface is lifted , ga dimers prefer ` ortho ' orientation leading to a 1@xmath32 patten .
our calculations suggest that mixed dimers ( fig .
4(i ) ) and ` ortho ' dimers ( fig .
4(g ) ) as two contenders for the energy favored structures . for the case of in we find that energetically the most favorable configuration is given in fig .
this configuration corresponds to a 2@xmath32 surface pattern .
the configurations in figs .
4(i ) and 4(l ) are unstable .
more significantly , we note that the reconstruction of the surface is not lifted .
we also find that the configurations consisting of ` para ' dimers ( i.e. , figs . 4(a ) and a(b ) ) are degenerate and energetically next favorable structures .
the configurations in figs .
4(a ) and 4(b ) correspond to the 2@xmath31 and 2@xmath32 surface structures respectively .
experimental results indicate the existence of both the 2@xmath32 and 2@xmath31 surface structures near 1 ml coverage of in .
the most and the next favorable structures from our calculations are consistent with the experimentally observed 2@xmath32 and 2@xmath31 surface structures respectively . to understand the bonding of ga and in atoms on si(001 )
, the charge density distributions are plotted on the plane of metal atoms for the most favorable configurations at 1ml coverage of ga and in and they are shown in figs .
6(a ) and 6(b ) respectively .
we observe that to a first approximation metal atoms display strong intra - plane bondings amongst themselves leaving the substrate in reconstructed state .
the charge distribution in fig .
6(b ) also helps us to understand why the configuration 4(i ) is unstable for in . in being larger in size ,
the in atoms in ` ortho ' orientation of configuration 4(i ) are pushed away from each other to have a suitable coupling among in atoms and thus favoring the configuration 4(j ) with ( mixed ) staggered dimers .
it is worth mentioning that one usually ends up with semi - metallic or non - metallic nanowires when supported on substrates as we saw above for the group iii metals on si(001 ) .
very recently , there have been several important studies @xcite of monovalent au chains on various high indexed si surfaces dealing with metallicity , peierls distortion , charge density instability , fractional band fillings , and possibility of a spin charge separation due to the luttinger - liquid phase .
in fact , himpsel and his group showed that au nanowire supported on si(553 ) is stable and it shows metallic behavior @xcite .
we have done extensive total energy calculations for al , ga and in adsorption on the si(001 ) surface at 0.5 and 1 ml coverages .
the ` para ' orientation of dimers is the most stable orientation at 0.5 ml coverage for al , ga and in , which supports the results of northrup et al and also agrees with the recent experimental observations .
all the patterns observed in various experiments at low temperature and below 0.5 ml coverages can be explained by the ` para ' orientation of metal dimers .
an interesting zigzag semi - metallic structure similar to @xmath0 is formed at 0.5 ml but is not the lowest energy structure .
soliton formation in overlayers is a fundamental topic for further study . at 1 ml coverage
a tendency towards the ` ortho ' orientation of dimers is noted .
thus we predict an orientational transition from ` para ' to ` ortho ' as we go from 0.5 ml to 1 ml coverage for al and possibly ga on the si surface .
in addition , the reconstruction of the si surface is completely lifted at 1 ml coverage of al .
these results lead us to conclude that the surface patterns at 0.5 ml coverage can be explained in terms of ` para ' orientation of metal dimers and those at intermediate coverages between 0.5 and 1 ml may be explained by the co - existence of ` para ' and ` ortho ' orientation of dimers . for in mixed `
para ' and ` ortho ' dimers give the lowest energy configuration that exhibit 2@xmath32 surface pattern and the transition is not complete .
p. a. dowben , surf .
rep . * 40 * , 151 ( 2000 ) .
d. s. portal , e. atracho , j. junquera , p. ordenjon , j. m. soler , phys .
lett . * 83 * , 3884 ( 1999 ) .
d. s. portal , e. atracho , j. junquera , a. garcia , and j. m. soler , surf .
sci , * 482 - 485 * , 1261 ( 2001 ) .
j. a. torres , e. tosatti , a. d. coros , f. ercolessi , j. j. kohanoff , f. d. di tolla , and j. m. soler , surf .
sci . * 426 * , l441 ( 1999 ) .
m. okamoto and k. takayanagi , phys .
b * 60 * , 7808 ( 1999 ) .
h. hakkinen , r. n. barnett , u. landman , j. phys .
b * 103 * , 8814 ( 1999 ) ; h. hakkinen , r. n. barnett , a. g. scherbakov , u. landman , j. phys .
chem b * 104 * , 9063 ( 2000 ) .
p. sen , s. ciraci , a. budlum , and i. p. batra , phys .
b * 64 * , 195420 ( 2001 ) . i. p. batra , t. ciani , d. boddupalli , and l. soberano , _ technical preceedings of the 2003 nanotechnology conference and trade show , shan francisco , usa _ , * 2 * , 206 ( 2003 ) .
t. ide , t nishimori , and t. ichinokawa , surf .
sci . * 209 * , 335 ( 1989 ) .
t. sakamoto and h. kawanami , surf .
sci . * 111 * , 177 ( 1981 ) b. bourguignon , k. l. carleton , and s. r. leone , surf .
sci . * 204 * , 455 ( 1988 ) .
j. nogami , a. a. baski , and c. f. quate , phys .
b * 44 * , 1415 ( 1991 ) .
j. nogami , sang - il park , and c. f. quate , appl .
. lett . * 53 * , 2086 ( 1988 ) .
j. knall , j. -e .
sundgren , g. v. hansson , and j. e. greene , surf .
sci . * 166 * , 512 ( 1986 ) .
a. a. baski , j. nogami , and c. f. quate , phy .
b * 43 * , 9316 ( 1991 ) .
a. a. baski , j. nogami , and c. f. quate , j. vac .
technol . * a 8 * , 245 ( 1990 ) ; * a 9 * , 1946 ( 1991 ) .
h. sakama , ken - ichi murakami , k. nishikata , and a. kawazu , phys .
b * 48 * , 5278 ( 1993 ) .
h. sakama , ken - ichi murakami , k. nishikata , and a. kawazu , phys .
b * 50 * , 14977 ( 1994 ) .
h. sakama , ken - ichi murakami , k. nishikata , and a. kawazu , phys .
b * 53 * , 1080 ( 1996 ) .
h. sakama , a. kawazu , t. sueyoshi , t. sato , and m. iwatsuki , phys . rev .
b * 54 * , 8756 ( 1996 ) .
i. p. batra , phys .
lett . * 63 * , 1704 ( 1989 ) .
j. e. northrup , m. c. schabel , c. j. karlsson , and r. i. g. uhrberg , phys .
b * 44 * , 13799 ( 1991 ) . by j.
r. schrieffer in _ ` highlights of condensed matter theory : proceedings of the international school of physics , enrico fermi , course lxxxix varenna , italy , june 28 - july 16 , 1983 ' _ ; edited by bassani , fumi and tosi .
series : enrico fermi international summar school ser . published by elsevier ( 1985 ) .
p. sen , i. p. batra , s. sivananthan , and c. h. grein , phys .
b , * 68 * , 045314 ( 2003 ) .
g. brocks , p. j. kelly , and r. car , phys .
lett . * 70 * , 2786 ( 1993 ) .
i. p. batra , phys .
b * 42 * , 9162 ( 1990 ) .
g. kresse and j. hafner , phys .
b * 47 * , r558 ( 1993 ) ; g. kresse and j. furtmuller , _ ibid . _
* 54 * , 11169 ( 1996 ) .
j. n. crain , a. kirakosian , k. n. altmann , c. bromberger , s. c. erwin , j. l. mcchesney , j. -l .
lin , and f. j. himpsel , phys .
lett . * 90 * , 176805 ( 2003 ) . f. j. himpsel , k. n. altmann , r. bennewitz , j. n. crain , a. kirakosian , j. -l .
lin , j. l. mcchesney , j. phys .
matt . * 13 * , 11097 ( 2001 ) .
i. k. rabinson , p. a. bennett , and f. j. himpsel , phys .
88 * , 096104 ( 2002 ) .
k. n. altmann , j. n. crain , a. kirakosian , j. -l .
lin , d. y. petrovykh , and f. j. himpsel , phys .
b * 64 * , 035406 ( 2001 ) .
j. n. crain , j. l. mcchesney , fan zheng , m. c. gallagher , p. c. snijders , m. bissen , c. gundelach , s. c. erwin , and f. j. himpsel , _ ` gold chain with tailored electronic states ' _ , ( unpublished , 2003 ) . | extensive electronic structure calculations are performed to obtain the stable geometries of metals like al , ga and in on the si(001 ) surface at 0.5 ml and 1 ml coverages .
our results coupled with previous theoretical findings explain the recent experimental data in a comprehensive fashion . at low coverages , as shown by previous works , ` para ' dimers give the lowest energy structure . with increasing coverage beyond 0.5 ml , ` ortho ' dimers become part of low energy configurations leading toward a ` para ' to ` ortho ' transition at 1 ml coverage . for in mixed staggered dimers ( ` ortho ' and ` para ' ) give the lowest energy configuration . for ga ,
mixed dimers are non - staggered , while for al ` para ' to ` ortho ' transition of dimers is complete . thus at intermediate coverages between 0.5 and 1 ml , the ` ortho ' and ` para ' dimers may coexist on the surface . consequently , this may be an explanation of the fact that the experimental observations can be successfully interpreted using either orientation .
a supported zigzag structure at 0.5 ml , which resembles @xmath0 , does not undergo a dimerization transition , and hence stays semi - metallic .
also , unlike @xmath0 the soliton formation is ruled out for this structure .
-0.25 in -0.2 in 0.2 in |
flavor - changing neutral current ( fcnc ) processes are forbidden at the tree level in the standard model ( sm ) , but are induced by loop or box diagrams .
if non - sm particles participate in the loop or box diagrams , their amplitudes may interfere with the sm amplitudes .
this makes fcnc processes an ideal place to search for new physics .
cleo first observed and measured the radiative penguin decay @xmath61@xcite , which constrains the magnitude of the effective wilson coefficient of the electromagnetic penguin operator , @xmath62 .
this provides the most stringent indirect limit on the charged higgs mass range@xcite .
however , it can not constrain the phase of @xmath63 , which is essential to obtain definitive evidence of new physics since @xmath63 is positive in the sm while it can be negative in non - sm physics@xcite . in particular , @xmath63 is constrained to be negative in the minimal sugra model@xcite .
the electroweak penguin decay @xmath0 is promising from this point of view since the coefficients @xmath64 and @xmath65 can be completely determined by measuring the dilepton invariant mass distributions , forward - backward charge asymmetry of the dilepton together with the @xmath61 decay rate@xcite .
the branching fractions predicted within the framework of the standard model are listed in tables [ tab : brpred ] and [ tab : brpredinc ] [ [ ref : ali][ref : inclusive - ks ] ] .
although several groups @xcite@xcite@xcite have searched for exclusive @xmath66 and @xmath67 decays is referred as @xmath68 hereafter . ] and cleo@xcite has searched for inclusive @xmath0 decays , no evidence has been observed . .branching fractions for @xmath66 and @xmath69
decays predicted in the standard model .
[ cols= " < , < , < , < " , ] [ tab : results ]
we have observed evidence for the electroweak penguin decay @xmath4 .
the preliminary branching fraction for this decay mode is , @xmath5 .
this value is consistent with theoretical predictions @xcite @xcite @xcite .
the errors are dominated by statistics , which can be improved in the coming runs .
we report 90% confidence level upper limits of branching fractions for the following exclusive and inclusive @xmath70 decays .
@xmath6 , + @xmath71 , + @xmath72 , + @xmath9 , + @xmath10 .
these values are close to the sm predictions .
the results reported here are preliminary .
we wish to thank the kekb accelerator group for the excellent operation of the kekb accelerator .
we acknowledge support from the ministry of education , culture , sports , science , and technology of japan and the japan society for the promotion of science ; the australian research council and the australian department of industry , science and resources ; the department of science and technology of india ; the bk21 program of the ministry of education of korea and the chep src program of the korea science and engineering foundation ; the polish state committee for scientific research under contract no.2p03b 17017 ; the ministry of science and technology of russian federation ; the national science council and the ministry of education of taiwan ; the japan - taiwan cooperative program of the interchange association ; and the u.s .
department of energy .
a.i . would like to acknowledge support from research fellowships of the japan society for the promotion of science for young scientists .
t. besmer , c. greub , t. hurth .
cern - th-2001 - 136 , butp-01 - 12 , zu - th-15 - 01 , hep - ph/0105292 ; m. ciuchini , g. degrassi , p.gambino and g. f. giudice , ; c. bobeth , m. misiak and j. urban , ; f. borzumati , c. greub , t. hurth and d. wyler , . m.h.r . khan_et al . _ ,
nucl . inst
a413 * 201 ( 1998 ) ; t. sumiyoshi _ et al .
a433 * 385 ( 1999 ) ; t. iijima _
et al . _ ,
proceedings of the 7th international conference on instrumentation for colliding beam physics , hamamatsu , japan , nov .
1519 , 1999 . | we report a search for the flavor - changing neutral current decay @xmath0 using a 29.5 fb@xmath1 data sample accumulated at the @xmath2 resonance with belle detector at the kekb @xmath3 storage ring . we observe evidence for @xmath4 and report the preliminary branching fraction of @xmath5 .
we also set the following 90% confidence level upper limits on the branching fractions for exclusive and inclusive decays , @xmath6 , + @xmath7 , + @xmath8 , + @xmath9 , + @xmath10 .
# 1#2#3#4#1 * # 2 * , # 3 ( # 4 ) k. abe@xmath11 , k. abe@xmath12 , r. abe@xmath13 , i. adachi@xmath11 , byoung sup ahn@xmath14 , h. aihara@xmath15 , m. akatsu@xmath16 , k. asai@xmath17 , m. asai@xmath18 , y. asano@xmath19 , t. aso@xmath20 , v. aulchenko@xmath21 , t. aushev@xmath22 , a. m. bakich@xmath23 , e. banas@xmath24 , s. behari@xmath11 , p. k. behera@xmath25 , d. beiline@xmath21 , a. bondar@xmath21 , a. bozek@xmath24 , t. e. browder@xmath26 , b. c. k. casey@xmath26 , p. chang@xmath27 , y. chao@xmath27 , k .- f .
chen@xmath27 , b. g. cheon@xmath28 , r. chistov@xmath22 , s .- k .
choi@xmath29 , y. choi@xmath28 , l. y. dong@xmath30 , j. dragic@xmath31 , a. drutskoy@xmath22 , s. eidelman@xmath21 , v. eiges@xmath22 , y. enari@xmath16 , c. w. everton@xmath31 , f. fang@xmath26 , h. fujii@xmath11 , c. fukunaga@xmath32 , m. fukushima@xmath33 , a. garmash@xmath34 , a. gordon@xmath31 , k. gotow@xmath35 , h. guler@xmath26 , r. guo@xmath36 , j. haba@xmath11 , h. hamasaki@xmath11 , k. hanagaki@xmath37 , f. handa@xmath38 , k. hara@xmath39 , t. hara@xmath39 , n. c. hastings@xmath31 , h. hayashii@xmath17 , m. hazumi@xmath39 , e. m. heenan@xmath31 , y. higasino@xmath16 , i. higuchi@xmath38 , t. higuchi@xmath15 , t. hirai@xmath40 , h. hirano@xmath41 , t. hojo@xmath39 , t. hokuue@xmath16 , y. hoshi@xmath12 , k. hoshina@xmath41 , s. r. hou@xmath27 , w .- s .
hou@xmath27 , s .- c .
hsu@xmath27 , h .- c .
huang@xmath27 , y. igarashi@xmath11 , t. iijima@xmath11 , h. ikeda@xmath11 , k. ikeda@xmath17 , k. inami@xmath16 , a. ishikawa@xmath16 , h. ishino@xmath40 , r. itoh@xmath11 , g. iwai@xmath13 , h. iwasaki@xmath11 , y. iwasaki@xmath11 , d. j. jackson@xmath39 , p. jalocha@xmath24 , h. k. jang@xmath42 , m. jones@xmath26 , r. kagan@xmath22 , h. kakuno@xmath40 , j. kaneko@xmath40 , j. h. kang@xmath43 , j. s. kang@xmath14 , p. kapusta@xmath24 , n. katayama@xmath11 , h. kawai@xmath44 , h. kawai@xmath15 , y. kawakami@xmath16 , n. kawamura@xmath45 , t. kawasaki@xmath13 , h. kichimi@xmath11 , d. w. kim@xmath28 , heejong kim@xmath43 , h. j. kim@xmath43 , hyunwoo kim@xmath14 , s. k. kim@xmath42 , t. h. kim@xmath43 , k. kinoshita@xmath46 , s. kobayashi@xmath47 , s. koishi@xmath40 , h. konishi@xmath41 , k. korotushenko@xmath37 , p. krokovny@xmath21 , r. kulasiri@xmath46 , s. kumar@xmath48 , t. kuniya@xmath47 , e. kurihara@xmath44 , a. kuzmin@xmath21 , y .- j .
kwon@xmath43 , j. s. lange@xmath49 , s. h. lee@xmath42 , c. leonidopoulos@xmath37 , y .- s .
lin@xmath27 , d. liventsev@xmath22 , r .- s .
lu@xmath27 , d. marlow@xmath37 , t. matsubara@xmath15 , s. matsui@xmath16 , s. matsumoto@xmath50 , t. matsumoto@xmath16 , y. mikami@xmath38 , k. misono@xmath16 , k. miyabayashi@xmath17 , h. miyake@xmath39 , h. miyata@xmath13 , l. c. moffitt@xmath31 , g. r. moloney@xmath31 , g. f. moorhead@xmath31 , n. morgan@xmath35 , s. mori@xmath19 , t. mori@xmath50 , a. murakami@xmath47 , t. nagamine@xmath38 , y. nagasaka@xmath18 , y. nagashima@xmath39 , t. nakadaira@xmath15 , t. nakamura@xmath40 , e. nakano@xmath51 , m. nakao@xmath11 , h. nakazawa@xmath50 , j. w. nam@xmath28 , z. natkaniec@xmath24 , k. neichi@xmath12 , s. nishida@xmath52 , o. nitoh@xmath41 , s. noguchi@xmath17 , t. nozaki@xmath11 , s. ogawa@xmath53 , t. ohshima@xmath16 , y. ohshima@xmath40 , t. okabe@xmath16 , t. okazaki@xmath17 , s. okuno@xmath54 , s. l. olsen@xmath26 , h. ozaki@xmath11 , p. pakhlov@xmath22 , h. palka@xmath24 , c. s. park@xmath42 , c. w. park@xmath14 , h. park@xmath55 , l. s. peak@xmath23 , m. peters@xmath26 , l. e. piilonen@xmath35 , e. prebys@xmath37 , j. l. rodriguez@xmath26 , n. root@xmath21 , m. rozanska@xmath24 , k. rybicki@xmath24 , j. ryuko@xmath39 , h. sagawa@xmath11 , y. sakai@xmath11 , h. sakamoto@xmath52 , m. satapathy@xmath25 , a. satpathy@xmath56 , s. schrenk@xmath46 , s. semenov@xmath22 , k. senyo@xmath16 , y. settai@xmath50 , m. e. sevior@xmath31 , h. shibuya@xmath53 , b. shwartz@xmath21 , a. sidorov@xmath21 , s. stani@xmath19 , a. sugi@xmath16 , a. sugiyama@xmath16 , k. sumisawa@xmath11 , t. sumiyoshi@xmath11 , j .- i .
suzuki@xmath11 , k. suzuki@xmath44 , s. suzuki@xmath57 , s. y. suzuki@xmath11 , s. k. swain@xmath26 , h. tajima@xmath15 , t. takahashi@xmath51 , f. takasaki@xmath11 , m. takita@xmath39 , k. tamai@xmath11 , n. tamura@xmath13 , j. tanaka@xmath15 , m. tanaka@xmath11 , y. tanaka@xmath58 , g. n. taylor@xmath31 , y. teramoto@xmath51 , m. tomoto@xmath11 , t. tomura@xmath15 , s. n. tovey@xmath31 , k. trabelsi@xmath26 , t. tsuboyama@xmath11 , t. tsukamoto@xmath11 , s. uehara@xmath11 , k. ueno@xmath27 , y. unno@xmath44 , s. uno@xmath11 , y. ushiroda@xmath11 , s. e. vahsen@xmath37 , k. e. varvell@xmath23 , c. c. wang@xmath27 , c. h. wang@xmath59 , j. g. wang@xmath35 , m .- z .
wang@xmath27 , y. watanabe@xmath40 , e. won@xmath42 , b. d. yabsley@xmath11 , y. yamada@xmath11 , m. yamaga@xmath38 , a. yamaguchi@xmath38 , h. yamamoto@xmath26 , t. yamanaka@xmath39 , y. yamashita@xmath60 , m. yamauchi@xmath11 , s. yanaka@xmath40 , m. yokoyama@xmath15 , k. yoshida@xmath16 , y. yusa@xmath38 , h. yuta@xmath45 , c. c. zhang@xmath30 , j. zhang@xmath19 , h. w. zhao@xmath11 , y. zheng@xmath26 , v. zhilich@xmath21 , and d. ontar@xmath19 @xmath45aomori university , aomori + @xmath21budker institute of nuclear physics ,
novosibirsk + @xmath44chiba university , chiba + @xmath50chuo university , tokyo + @xmath46university of cincinnati , cincinnati oh + @xmath49university of frankfurt , frankfurt + @xmath29gyeongsang national university , chinju + @xmath26university of hawaii , honolulu hi + @xmath11high energy accelerator research organization ( kek ) , tsukuba + @xmath18hiroshima institute of technology , hiroshima + @xmath33institute for cosmic ray research , university of tokyo , tokyo + @xmath30institute of high energy physics , chinese academy of sciences , beijing + @xmath22institute for theoretical and experimental physics , moscow + @xmath54kanagawa university , yokohama + @xmath14korea university , seoul + @xmath52kyoto university , kyoto + @xmath55kyungpook national university , taegu + @xmath31university of melbourne , victoria + @xmath58nagasaki institute of applied science , nagasaki + @xmath16nagoya university , nagoya + @xmath17nara women s university , nara + @xmath36national kaohsiung normal university , kaohsiung + @xmath59national lien - ho institute of technology , miao li + @xmath27national taiwan university , taipei + @xmath24h .
niewodniczanski institute of nuclear physics , krakow + @xmath60nihon dental college , niigata + @xmath13niigata university , niigata + @xmath51osaka city university , osaka + @xmath39osaka university , osaka + @xmath48panjab university , chandigarh + @xmath37princeton university , princeton nj + @xmath47saga university , saga + @xmath42seoul national university , seoul + @xmath28sungkyunkwan university , suwon + @xmath23university of sydney , sydney nsw + @xmath53toho university , funabashi + @xmath12tohoku gakuin university , tagajo + @xmath38tohoku university , sendai + @xmath15university of tokyo , tokyo + @xmath40tokyo institute of technology , tokyo + @xmath32tokyo metropolitan university , tokyo + @xmath41tokyo university of agriculture and technology , tokyo + @xmath20toyama national college of maritime technology , toyama + @xmath19university of tsukuba , tsukuba + @xmath25utkal university , bhubaneswer + @xmath35virginia polytechnic institute and state university , blacksburg va + @xmath57yokkaichi university , yokkaichi + @xmath43yonsei university , seoul + |
assuming a uniform distribution of matter on large scales , the observed data of high - redshift type ia supernovae(snia)@xcite point to @xmath2-dominated flat friedmann - robertson - walker ( frw ) models .
the darkness of the snia is reduced to accelerating expansion of the universe due to a positive @xmath2 term .
these @xmath2-dominated frw models are consistent also with the observed data of temperature anisotropy in the cosmic microwave background ( cmb ) radiation@xcite , except for the low - multipole components@xcite .
moreover the observed correlation between the cmb and large - scale structure supports these @xmath2-dominated models , which can generate anisotropies due to the first - order(linear ) isw effect@xcite . on the other hand , alternative inhomogeneous models that can explain the snia data without introducing a cosmological constant @xmath2 have been independently proposed by c@xmath5l@xmath6rier@xcite , goodwin et al.@xcite and tomita@xcite and subsequently studied by several authors@xcite .
it turned out that some inhomogeneous cosmological models with an inner large - scale underdense region ( which we called _ a local void _ in our previous works ) with a small hubble constant ( @xmath7 ) in the outer flat region can also explain the cmb data@xcite as well as the snia data . in these models ,
the cosmological copernican principle is violated since we need to live near the center of an underdense region .
however , recent observational studies such as the baryon acoustic oscillations ( bao)@xcite , the kinematic sunyaev - zeldovich effect either from clusters @xcite or reionized regions@xcite put stringent constraints on these anti - copernican models . as a result , models with a local void on 300 mpc scales seem to be ruled out . at the moment , we need to consider inhomogeneous models with a local void on gpc scales so that the constraints from bao at epochs of @xmath8 may be avoided .
recently several gpc void models have been studied by clifton et al.@xcite and garca - bellido and haugb@xmath9lle@xcite . in this paper
we study the isw effect in flat frw models with an inner underdense region on gpc scales based on previous results@xcite .
then we compare it with the isw effect in the concordant flat frw model with a cosmological constant @xmath2 .
as we shall show , the isw effect will be an excellent discriminator between our anti - copernican models and the standard concordant copernican model . in 2 , we present our inhomogeneous cosmological model with inner underdense regions and in 3 we derive analytic formulae for calculating the isw effect in the inner and outer regions and we discuss the property of temperature anisotropy due to the isw effect in our models and the concordant model . 4 is dedicated to concluding remarks . in
what follows , we use the units of @xmath10 . for spatial coordinates
, we use latin subscripts running from 1 to 3 .
our inhomogeneous anti - copernican models without a cosmological constant @xmath2 consist of two inner underdense regions ( i and ii ) and an outer flat region ( iii ) .
the former regions are described by negatively curved frw models ( @xmath11 and @xmath12 ) and the outer region is by the einstein - de sitter model ( eds)(@xmath13 ) . here in these regions we use homogeneous models locally because the isw effect can be treated only in homogeneous models at present .
we assume that these regions are connected by two infinitesimally thin walls at redshifts @xmath0 and @xmath14 corresponding to the boundary between i and ii and the boundary between ii and iii , respectively .
the latter redshift value @xmath14 corresponds to a @xmath15 gpc radius of the spherical underdense region .
the hubble constants @xmath16 and @xmath17 in these regions satisfy a relation @xmath18 . here
we consider the following two cases : @xmath19 where @xmath20 stands for the value necessary for the observed cmb anisotropies in the eds model , @xmath21 in case 2 is the standard value in the local measurement , and the case 1 with smaller @xmath22 and @xmath23 is taken so as to consider a stringent observational condition which is given by the kinematic sunyaev - zeldovich effect@xcite .
if we regard the outer region as the background , the inner region can be interpreted as a local inhomogeneity and has an optical influence on the temperature of cmb radiation .
if the observer is exactly at the center , the influence is isotropic , but if he is off - center , it brings dipole , quadrupole and the other multipole anisotropies .
these anisotropies have already been analyzed and discussed in previous papers@xcite . in what follows ,
we study the isw effect due to small - scale density perturbations ( of a simple spherical top - hat type ) in the three regions in the inhomogeneous model . in spherical coordinates @xmath24 ) ,
the background metric of a constantly negatively curved spacetime in the inner regions i and ii containing pressureless matter with a matter density @xmath25 is given by @xmath26 and @xmath27 = 6/(\cosh \eta -1 ) , \end{aligned}\ ] ] where @xmath28 and @xmath29 are the cosmic time and the conformal time .
prime @xmath30 represents @xmath31 and @xmath32 is a constant .
the hubble parameter , the density parameter and the redshift are @xmath33 where @xmath34 is i or ii , @xmath35 is the present value of the conformal time @xmath29 and for @xmath36 we have @xmath37 and @xmath38 .
the constant @xmath32 for region i or ii is given by @xmath39 where @xmath40 .
in the outer region iii containing pressureless matter , the space - time metric is @xmath41,\ ] ] where @xmath42 .
the hubble parameter , the density parameter and the redshift are @xmath43 , and @xmath44 . here for @xmath45
we have as @xmath46 . for comparison
, we consider a concordant flat frw model with a cosmological constant @xmath2 . the metric ( [ eq : m1 ] ) is @xmath47 and the scale factor satisfies @xmath48 , where @xmath49 and @xmath50 are the energy density of matter and that of a cosmological constant @xmath2 , respectively . as the model parameter of the concordant flat-@xmath2 model ,
we adopt @xmath51 and @xmath52 km / s / mpc .
now we consider growing mode of density perturbations and the integrated sachs - wolfe effect in the inner and outer regions , separately . the first - order gauge - invariant growing density perturbations @xmath53 and the gauge - invariant potential perturbation ( in the growing mode ) @xmath54@xcite are expressed as @xmath55 with @xmath56 where @xmath53 corresponds to the density perturbations in the comoving synchronous gauge , @xmath57 is equal to the potential perturbations @xmath58 in the longitudinal or poisson gauge , and @xmath59 is the potential function given as an arbitrary function of spatial coordinates .
the expression of @xmath60 was derived from the solution shown by lifshits and khalatinikov@xcite .
@xmath61 is the laplacian of @xmath59 in the space @xmath62 , that is , @xmath63 , where @xmath64 is covariant derivatives in the three dimensional space with @xmath65 .
so we obtain from eqs.([eq : m1a ] ) and ( [ eq : r1 ] ) @xmath66 the first - order gauge - invariant temperature fluctuation due to the linear isw effect is expressed as @xmath67 in the poisson gauge , where the prime is @xmath68 and @xmath69 is the affine parameter along the light path . @xmath70 and @xmath71 are the emitter s and observer s values at the decoupling and present epochs , respectively . the matter density contrast for a top - hat type spherical void.,width=302 ] in this paper we consider a simple spherical top - hat type of compensated density perturbation following our previous paper@xcite , whose spatial size is much smaller than the horizon size . the spatial variation of the density perturbation is schematically shown in fig.[fig : rs2 ] .
we consider the cmb photon paths passing through the center of the spherical perturbation .
when the epoch in the center of the perturbation is @xmath29 , the integral of @xmath58 along the light path reduces approximately to @xmath72 \int^{\lambda_e}_{\lambda_o } d\lambda \ f / c,\ ] ] where @xmath34 is i or ii , @xmath73 and a constant @xmath74 are the central values of @xmath75 and @xmath61 , respectively , and the subscript @xmath76 denotes the present epoch . from integration of @xmath59 for the above top - hat type perturbations derived in the previous paper@xcite , we obtain @xmath77 w_1 ( y ) \cr & \times & \bigl(\epsilon_{\alpha m 0}/\epsilon_{i m 0}\bigr)_c \bigl(h_{\alpha 0}/h_{i0}\bigr)^3 , \end{aligned}\ ] ] where @xmath78 is defined as @xmath79 and @xmath80 . for the value of @xmath81 , we adopt @xmath82 as an example .
the first - order isw effect does not appear and the second - order isw effcet is the lowest one .
the second - order temperature fluctuations were derived in our previous paper@xcite and expressed as @xmath83 where @xmath84 , \ ( \zeta_1 + 9\ \zeta_2 ) ' = - ( 39/700 ) \eta$ ] for the eds model , @xmath85 is the radius of inhomogeneities ( cf .
[ fig : rs2 ] ) , and the central value of the density perturbation @xmath86 is related to a constant @xmath74 as @xmath87 using eq.([eq : r9 ] ) , the temperature fluctuations are expressed as @xmath88 the temperature fluctuations are negative definite .
they are not exactly observed fluctuations , because their observed values should be the difference from the average value @xmath89 of the sum of the second - order temperature fluctuations which is caused by all possible primordial density perturbations and renormalized into the background temperature .
this average value is derived , taking account of power spectrum of density perturbations , in the procedure shown in a separate paper@xcite .
so the above @xmath90 should be here used to show the order of magnitude of second - order isw effect , but for the perturbations with large amplitudes , the above second - order fluctuations are regarded approximately as observed values , as the mean value can be neglected .
the deformation of the walls brings the complicated perturbations inside the walls and their neighborhoods , as was studied by one of the present authors through the analysis of the junction condition@xcite .
they include not only density perturbations , but also gravitational - wave and rotational perturbations .
gravitational - wave perturbations propagate , but their amplitudes are very small and the contribution to density perturbations is negligible , because of the small coupling between them .
moreover the density and rotational perturbations caused by the perturbed walls do not propagate in the present dust matter models and are constrained in the just neighborhoods of the walls . in the most part of the i ,
ii and iii regions , therefore , we see density perturbations which are independent of the wall motions and were caused primordially due to the common origin .
their amplitudes in the three regions were nearly equal at the early stages , but the present amplitudes became different , because they had different growth rates .
here we neglect the above complicated perturbations inside the walls and in their narrow neighborhoods .
then the three density perturbations @xmath91 and @xmath92 included in the two equations ( [ eq : r6 ] ) and ( [ eq : r10 ] ) are related as follows .
first we assume that @xmath93 and @xmath94 should be equal at early epochs of equal densities with the redshifts @xmath95 and @xmath96 , i.e. @xmath97 , where @xmath98 . the present densities @xmath99 and @xmath100 are related as @xmath101 and @xmath102 . then @xmath95 and @xmath103 are related as @xmath104^{1/3 } = [ ( \omega_{i0}/\omega_{iii0 } ) ( h_{i0}/h_{iii0})^2]^{1/3 } , \cr ( 1+z_3)/(1+z_2 ) & = & [ \rho_{ii0}/\rho_{iii0}]^{1/3 } = [ ( \omega_{ii0}/\omega_{iii0 } ) ( h_{ii0}/h_{iii0})^2]^{1/3}. \end{aligned}\ ] ] taking account of the growth rates , we obtain @xmath105 , @xmath106 and @xmath107 .
accordingly , we obtain @xmath108 where @xmath109 and @xmath60 are calculated using eqs.([eq : m2 ] ) and ( [ eq : r2 ] ) . here
we set @xmath110 .
then we have @xmath111 are @xmath112 and @xmath113 in cases 1 and 2 , respectively . for comparison ,
we show the first - order and second - order temperature fluctuations in the flat-@xmath2 models with @xmath114 , which were derived as @xmath115 and @xmath116 in our previous paper@xcite .
they are expressed as @xmath117_0 \bigl({a'\over a}\bigr)_0^{-3 } \bigl [ { a'\over a}+\bigl({a '' \over a } - 3\bigl({a'\over a}\bigr)^2\bigr ) p ' \bigr]\ \bigl({\epsilon_{m0}\over \epsilon_{im0}}\bigr)_c
\bigl({h_0\over h_{i0}}\bigr)^3 w_1(y ) , \end{aligned}\ ] ] and @xmath118_0 ^ 2 \bigl({a'\over a}\bigr)_0^{-3 } ( \zeta_1 + 9\zeta_2 ) ' \bigl({(\epsilon_{m0})_c^2 \over ( \epsilon_{m0i})_c}\bigr ) \bigl({h_0\over h_{i0}}\bigr)^3 w_2(y ) , \end{aligned}\ ] ] where @xmath119 and @xmath120 are first - order density perturbation and its central value at present epoch , and @xmath121 and @xmath122 are auxiliary quantities used in the previous paper@xcite and their definitions are shown in appendix . in the following ,
we consider the behaviors of fluctuations due to the isw effect in our inhomogeneous model in comparison with those in the concordant flat-@xmath2 model . to do so
, we show the amplitudes of temperature anisotropy due to the isw effect from a spherical compensating void / cluster with a given comoving radius and the density contrast , represented by the five quantities @xmath123 and @xmath124 .
note that the first - order quantities @xmath125 , @xmath126 and @xmath127 do not depend on @xmath128 and @xmath120 , while the second - order ones @xmath129 and @xmath124 are proportional to @xmath128 and @xmath120 , respectively .
here @xmath74 denotes the values at the centers of the voids / clusters .
we assume that @xmath130 , so that the cosmological situation in the neighborhood of our observer in the flat-@xmath2 model may be equal to that in the inner region i of the model . in fig .
[ fig : void0 ] , we show the behaviors of @xmath131 and @xmath132 in the interval @xmath133 , in cases 1 and 2 , respectively . in fig .
[ fig : void2 ] , in a similar manner , we show the behaviors of @xmath131 and @xmath132 in the interval @xmath134 , in cases 1 and 2 , respectively . in these figures we adopted @xmath135 ,
for which @xmath132 is comparable with @xmath125 and @xmath126 , and @xmath127 is smaller than @xmath136 for @xmath137 . for comparison , @xmath127 and @xmath136
are also shown in them . here
the second - order quantities are multiplied by @xmath138 , because they are negative definite and we use only positive quantities in the figures . from these figures
, we can see that @xmath139 and @xmath127 are comparable in the regions i and ii , though their behaviors are different .
it is found that @xmath140 is smaller than @xmath141 , though both quantities are of second - order .
this reflects the strong dependence on the hubble constants ( cf .
eq.([eq : r10 ] ) and eq.([eq : r12 ] ) ) and the @xmath2-dependence of the second - order isw effect which was studied in the previous paper@xcite . as a result
we find the following common features in cases 1 and 2 from these figures .
\(1 ) in the inner regions i and ii , @xmath131 and @xmath127 are comparable , irrespective of @xmath142 and @xmath120 , though @xmath125 and @xmath126 in case 1 seem to be larger by a factor @xmath143 than @xmath125 and @xmath126 .
\(2 ) in the outer region near the wall of @xmath144 , @xmath141 is roughly comparable with @xmath127 for @xmath145 and it is smaller or larger than @xmath127 for @xmath146 or @xmath147 , respectively .
far outside the wall , @xmath141 is larger than @xmath127 .
\(3 ) since @xmath148 is @xmath149 for perturbations with a size @xmath150mpc ( @xmath151 km / s / mpc ) , @xmath129 is extremely large or negligible compared with @xmath127 for perturbations with @xmath152 or @xmath153mpc , respectively .
\(4 ) first - order quantities @xmath131 and @xmath127 are positive / negative for a cluster / void with @xmath128 and @xmath120 , respectively , while second - order quantities @xmath141 and @xmath124 are negative definite . therefore , in the concordant model , the expected amplitude is larger for voids than clusters .
such an asymmetry is not expected in the outer region in our inhomogeneous model .
\(5 ) at epochs of @xmath154 for @xmath155 ,
latexmath:[$|\theta^{(2)}_\lambda| < flat-@xmath2 models have different signs for a cluster / void with a density contrast @xmath157 . for @xmath158 , however , the second - order isw effect is dominant also in the flat-@xmath2 model .
therefore , the temperature fluctuations in the flat-@xmath2 models for @xmath159 is negative definite as in the outer region of our inhomogeneous models .
cc the @xmath109-dependence of first and second order temperature fluctuations in case 1(left ) and in case 2 ( right ) for photons passing through the center of a compensated spherical void at @xmath160 .
solid curves denote @xmath161 and @xmath126 and the curve @xmath162 denotes @xmath129 .
the curves @xmath163 and @xmath74 denote @xmath127 and @xmath124 , respectively .
the dotted vertical lines denote the boundaries at @xmath0 and @xmath164 .
we adopted @xmath135 , for which @xmath132 is comparable with @xmath125 and @xmath126.,width=415 ] the @xmath109-dependence of first and second order temperature fluctuations in case 1(left ) and in case 2 ( right ) for photons passing through the center of a compensated spherical void at @xmath160 .
solid curves denote @xmath161 and @xmath126 and the curve @xmath162 denotes @xmath129 .
the curves @xmath163 and @xmath74 denote @xmath127 and @xmath124 , respectively .
the dotted vertical lines denote the boundaries at @xmath0 and @xmath164 .
we adopted @xmath135 , for which @xmath132 is comparable with @xmath125 and @xmath126.,width=415 ] cc the @xmath109-dependence of first and second order temperature fluctuations in case 1(left ) and in case 2 ( right ) for photons passing through the center of a compensated spherical void at @xmath165 .
solid curves denote @xmath161 and @xmath126 and the curve @xmath162 denotes @xmath129 .
the curves @xmath163 and @xmath74 denote @xmath127 and @xmath124 , respectively .
the dotted vertical lines denote the boundaries at @xmath0 and @xmath164 .
we adopted @xmath135 , for which @xmath132 is comparable with @xmath125 and @xmath126.,width=415 ] the @xmath109-dependence of first and second order temperature fluctuations in case 1(left ) and in case 2 ( right ) for photons passing through the center of a compensated spherical void at @xmath165 .
solid curves denote @xmath161 and @xmath126 and the curve @xmath162 denotes @xmath129 . the curves @xmath163 and @xmath74 denote @xmath127 and @xmath124 , respectively .
the dotted vertical lines denote the boundaries at @xmath0 and @xmath164 .
we adopted @xmath135 , for which @xmath132 is comparable with @xmath125 and @xmath126.,width=415 ]
in this paper we studied the first - order and second - order isw effect in our anti - copernican inhomogeneous model with underdense regions in comparison with that in the concordant flat-@xmath2 model .
we found that a distinct feature appears at the outer region at moderate redshifts . in the concordant model ,
the expected amplitude of temperature anisotropy is larger for voids than clusters whereas such an asymmetry is not expected in the outer region in our inhomogeneous model .
we showed , moreover , that the first - order isw effect in the inner regions of our models is comparable with that in the flat-@xmath2 model , and that the second - order isw effect in the outer region depends on the amplitude @xmath157 of density perturbations .
the isw effect due to perturbations on scales larger than @xmath166 mpc with a density contrast @xmath167 in the outer region is negligible . on the other hand , in the inner region
, no isw effect appears due to perturbations on scales larger than the radius of the inner region . in our inhomogeneous model with underdense regions ,
the isw effect does not contribute to the low - multipole components of cmb anisotropies@xcite in accord with the assertion proposed by hunt and sarker@xcite , while in the flat-@xmath2 model , the contribution from the isw effect due to large - scale linear perturbations is significant . the observed correlation between the cmb sky with the large - scale structure is usually interpreted as the evidence of the cosmological constant @xmath2 , which causes the first - order isw effect@xcite .
recently , moreover , hot and cold spots on the cmb sky associated with super - structures ( with @xmath168 ) in sdss luminous red galaxy catalog were measured by granett et al.@xcite and the consistency with the isw effect in the flat-@xmath2 models was shown .
it should be noted , however , that they may be brought in principle by the first and second - order isw effect also in our models with underdense regions , as we showed in the present paper .
therefore , the observational evidence for the existence of small - scale isw effect for light paths through clusters of galaxies , superclusters and supervoids may support not only the flat-@xmath2 model , but also our models with underdense regions . in order to make a clear distinction between the two models ,
it is better to compare the overall amplitudes of temperature fluctuations associated with a void(negative density ) with those associated with a cluster(positive density ) .
as we have seen , for @xmath169 ( which depends on @xmath157 ) , the amplitudes for quasi - linear voids are larger than those for quasi - linear clusters in the concordant model due to the second - order effect , while such an asymmetry can not be expected in our inhomogeneous model since there is no first - order effect in the outer region@xcite .
@xmath171 satisfies @xmath172 and its solution is expressed as @xmath173^{1/2 } \int^{{a}}_0 d \tilde{a } \tilde{a}^{3/2}[\omega_{m0}+\omega_{\lambda 0 } \tilde{a}^3]^{-1/2 } + { 2 \over 3\omega_{m0}}{a } , \cr \eta & = & \int^{{a}}_0 d\tilde{a } \tilde{a}^{-1/2}[\omega_{m0}+\omega_{\lambda 0 } \tilde{a}^3]^{-1/2}.\end{aligned}\ ] ] the functions @xmath174 and @xmath175 are defined as @xmath176 \big\},\end{aligned}\ ] ] where @xmath177 is satisfies @xmath178.\ ] ] 1 b.p .
schmidt et al . , astrophys .
j. * 507*,46 ( 1998 ) .
a.g . riess et al . , astron .
j. * 116*,1009 ( 1998 ) .
a.g . riess et al .
j. * 118*,2668 ( 2000 ) .
s. perlmutter et al . , astrophys .
j. * 517*,565 ( 1999 ) .
bennet et al . , astrophys .
* 148 * , 1 ( 2003 ) .
spergel et al . , astrophys .
* 148 * , 175 ( 2003 ) .
a. de oliveira - costa et al . , phys . rev . * d69 * , 063516 ( 2004 ) .
contaldi et al .
, j. cosmol .
* 7 * , 2 ( 2003 ) .
s. boughn and r. crittenden , nature * 427*,45 ( 2004 ) r.g .
crittenden and n. turok , phys .
lett . * 76 * , 575 ( 1966 ) .
granett , m.c .
neyrinck and i. szapudi , arxiv:0805.3695 .
c@xmath179l@xmath179rier , astron . astrophys . * 353*,63 ( 2000 ) ; astro - ph/9907206 ; m.n .
c@xmath179l@xmath179rier , new adv . phys . *
1*,29 ( 2007 ) .
goodwin , a.j .
barber , j. griffin and l.i .
onuora , astro - ph/9906187 . k. tomita , astrophys .
j. * 529 * , 26 ( 2000 ) ; astro - ph/9905287 .
k. tomita , astrophys .
j. * 529 * , 38 ( 2000 ) ; astro - ph/9906027 .
k. tomita , mnras * 326 * , 287 ( 2001 ) k. tomita , prog .
phys . * 106 * , 929 ( 2001 ) .
h. alnes , m. amarzguioui and o. gron , phys .
rev.*d73 * , 083519 ( 2006 ) ; h. alnes and m. amarzguioui , phys .
rev.*d75 * , 023506 ( 2007 ) .
t. biswas , r. mansouri and a. notari , j. cosmol .
* 12 * , 017 ( 2007 ) .
s. alexander , t. biswas , a. notari and d. vaid , arxiv:0712.0370 .
h. iguchi , t. nakamura and k. nakao , prog .
. phys . * 108 * , 809 ( 2002 ) .
a. blanchard et al . , astron
. astrophys . * 412 * , 35 ( 2003 ) .
p. hunt and s. sarkar , phys
. rev . * d76 * , 123504 ( 2007 ) ; arxiv:0706.2443 .
d.j . eisenstein et al . , astrophys .
j. * 633*,560 ( 2005 ) .
seo and d.j .
eisenstein , astrophys .
j. * 598*,720 ( 2003 ) .
percival et al . , mnras * 381*,1053 ( 2007 ) .
percival et al . , astrophys .
j. * 657*,51 ( 2007 ) .
percival et al . , astrophys .
j. * 657*,645 ( 2007 ) .
e. gaztaaga , r. miquel and e. snchez , arxiv:0808.1921 .
j.p . zibin , a. moss and d. scott , arxiv:0809.3761 .
j. garca - bellido and t. haugb@xmath180lle , arxiv:0807.1326 . r.r .
caldwell and a. stebbins , arxiv:0711.3459 .
t. clifton , p.g .
ferreira and k. land , arxiv:0807.0443
. j. garca - bellido and t. haugb@xmath180lle , j. cosmol .
. phys . * 4 * , 3 ( 2008 ) .
k. tomita , phys .
* d71 * , 083504 ( 2005 ) .
k. tomita , phys .
rev . * d72 * , 043526 ( 2005 )
. k. tomita , phys .
* d72 * , 103506 ( 2005 ) .
inoue and j. silk , astrophys .
j. * 648 * , 23 ( 2006 ) .
inoue and j. silk , astrophys .
j. * 664 * , 650 ( 2007 ) .
k. tomita and k.t .
inoue , phys .
rev.*d77 * , 103522 ( 2008 ) .
k. tomita , astrophys .
j. * 584 * , 580 ( 2003 ) .
moffat , j. cosmol .
astropart . phys . *
10 * , 12 ( 2005 ) .
h. alnes and m. amarzguioui , phys .
rev.*d74 * , 103520 ( 2006 ) .
bardeen , phys .
rev.*d22 * , 1882 ( 1980 ) .
lifshitz and i.m .
khalatnikov , adv .
* 12 * , 185 ( 1963 ) .
k. tomita , phys .
rev.*d77 * , 103521 ( 2008 ) .
k. tomita , phys .
rev.*d70 * , 123502 ( 2004 ) .
n. sakai , k.t .
inoue , phys . rev .
* d78 * , 063510 ( 2008 ) . | recent observational data of supernovae indicate that we may live in an underdense region , which challenges the copernican principle .
we show that the integrated sachs - wolfe ( isw ) effect is an excellent discriminator between anti - copernican inhomogeneous models and the standard copernican models . as a reference model
, we consider an anti - copernican inhomogeneous model that consists of two inner negatively curved underdense regions and an outer flat einstein - de sitter region .
we assume that these regions are connected by two thin - walls at redshifts @xmath0 and @xmath1 . in the inner two regions , the first - order isw effect is dominant and comparable to that in the concordant flat-@xmath2 models . in the outer einstein - de sitter region ,
the first - order isw effect vanishes but the second - order isw effect plays a dominant role , while the first - order isw effect is dominant in the flat-@xmath2 models at moderate redshifts .
this difference can discrimate the anti - copernican models from the concordant flat-@xmath2 model . at high redshits , the second - order isw effect is dominant both in our inhomogeneous model and the concordant model . in the outer region ,
moreover , the isw effect due to large - scale density perturbations with a present matter density contrast @xmath3 is negligible , while the effect due to small - scale density perturbations ( such as clusters of galaxies , superclusters and voids ) with @xmath4 would generate anisotropies which are larger than those generated by the isw effect in the concordant model . |
in recent times , the notions of persistence and the associated persistence exponent has become one of the active topics of research in non - equilibrium physics@xcite . in general , the persistence probability @xmath2 is the probability that a stochastic variable @xmath23 remains above or below a certain arbitrary value ( say , its initial value ) for a time interval @xmath24 $ ] .
the idea of persistence is particularly relevant in coarsening systems where @xmath2 has the natural interpretation of being the fraction of space in the system which remains in the same phase , starting from a random initial configuration . in these systems ( as in some other systems also ) @xmath2 typically decays as a power - law at large @xmath13 , with an exponent that is non - trivial to compute and does not appear to be simply related to other known exponents that characterize the process .
the coarsening dynamics of the one dimensional @xmath1-state potts model at zero temperature is one of the few cases where the persistence exponent @xmath4 is known exactly .
the solution was provided by derrida _ et.al _ through a mapping of the process to a coagulation model in steady state@xcite .
it was shown that at late times @xmath13 , well beyond the time scale of equilibration , the fraction of persistent spins left in a finite system of linear dimension @xmath25 scale as @xmath26 , where @xmath4 is given by the non - trivial expression @xmath27^{2}. \label{eq : eq1}\ ] ] for times @xmath28 , it follows that @xmath29 .
an interesting question in this context is about the spatial distribution of spins which are persistent up to any given time @xmath13 .
clearly , in a many - body process like the time evolution of potts model , the probability that a given spin is persistent is closely linked to the state of other spins .
this inter - dependence of spins is crucial in that it makes the time evolution at any single site strongly non - markovian , which makes the computation of the persistence exponent highly non - trivial .
a study of the spatial aspects of the persistence problem is therefore important from the point of view developing an intuitive understanding of the phenomenon , and also illuminates the interplay between persistence decay and the underlying domain coarsening process . the spatial distribution of persistent sites and its time evolution
have been studied previously through numerical simulations in one dimensional diffusion equation@xcite , @xmath1-state potts model@xcite , two dimensional ising model@xcite and one dimensional ising model with parallel dynamics@xcite .
an analytic study using a rate equation approach under the independent interval approximation(iia ) has been carried out for one dimensional @xmath30 model which is closely related to 1d ising model@xcite .
it is now generally understood from physical arguments and simulations that for a coarsening process in @xmath31 dimensions where characteristic length scale has the power - law growth @xmath32 , the set of persistent sites forms a fractal structure with fractal dimension @xmath33 over length scales @xmath34@xcite .
the distribution is homogeneous beyond this length scale .
furthermore , since @xmath35 , the distribution is expected to be homogeneous over all length scales if @xmath36 .
this has important consequences for systems like the potts model where @xmath37 changes with the potts state @xmath1 .
in particular , for potts model in @xmath38 , bray and odonoghue@xcite argued that a transition from fractal to homogeneous distribution occurs as @xmath37 crosses @xmath39 .
this transition is also marked by an abrupt change in the dynamical exponent characterizing the separation between persistent sites .
the characteristic length scale was conjectured to have the unusual dynamical scaling form @xmath40 with @xmath41 .
this conjecture based on physical arguments was supported by numerical simulations@xcite . in this paper , we use a mean - field approach to address the problem of spatial distribution of persistent sites in @xmath1-state potts model . the essential idea behind this approach is as follows .
it is well - known that the @xmath0 coarsening dynamics of the @xmath1-state potts model can be mapped to a reaction - diffusion process . in this process
, the interfaces between different species of potts spins are represented by diffusing particles @xmath42 , which annihilate or coagulate upon meeting with @xmath1-dependent probabilities . in the mean field approach , these diffusing particles are treated as homogeneously distributed , with ( time - dependent ) density equal to the average density in the original reaction - diffusion problem . this approach has been discussed in some earlier works as a heuristic argument @xcite and as a toy model for persistence @xcite .
we argue that this approach yields a lower bound for the persistence exponent in the potts model .
we then construct an artificial model which is devoid of spatial correlations among diffusing particles , but with persistence exponent tuned to be exactly equal to the potts model value .
we refer to this model as the modified mean - field approximation ( mmfa ) and use this approximation to study the spatial distribution of persistent sites in @xmath1-state potts model .
we outline our main results at this point . within the mmfa
, we show analytically that the correlation length for the spatial distribution of persistent spins scales as @xmath43 and the equal time pair correlation function @xmath44 ( defined as the probability that the spin at origin and the point @xmath45 are persistent at time @xmath13 ) has the scaling form @xmath46 for any value of @xmath37 .
this shows that the persistent spins have a fractal distribution with @xmath47 over length scales @xmath48 when @xmath49 , but @xmath50 when @xmath51 .
we find that the characteristic length scale of the spatial distribution of persistent spins has the unusual scaling form @xmath40 where @xmath15=max(@xmath16 ) , in agreement with the conjecture in @xcite .
the empty interval distribution itself has the scaling form @xmath52 $ ] , where @xmath53 for @xmath8 and decays exponentially with @xmath22 when @xmath10 . the exponents @xmath54 and @xmath18 depends on @xmath37 through the relations @xmath55 and @xmath18=max[@xmath19 ] , where @xmath57 is the heaviside step function .
we support these results with numerical simulations .
the paper is arranged as follows . in the next section ,
we outline the mean field approach . in section iii
, we introduce the mmfa and compute the pair correlation and empty interval distribution of persistent sites to characterize their spatial distribution .
these predictions are compared with the results of numerical simulations in the @xmath1-state potts model in sec .
we summarize our results and present our conclusions in sec . v.
in the zero temperature coarsening dynamics of @xmath1-state potts model in @xmath38 , the interfacial points between different species of potts spins perform independent random walks on the lattice and annihilate each other with probability @xmath58 ( @xmath59 ) , or coagulate with probability @xmath59 . in the process , persistent sites are ` wiped out ' , and the surviving random walkers build up spatial correlations among themselves .
the distribution of intervals between the surviving random walkers at any ( sufficiently late ) time @xmath13 is described by a ( stationary ) scaling function which is known exactly for all values of @xmath1@xcite .
the average density @xmath60 at time @xmath13 decays as @xmath61 asymptotically@xcite . the essential idea behind the mean field ( mf )
approximation is to treat the random walkers as forming a homogeneous background of average density @xmath60 , as far as the persistent sites are concerned .
we define the persistence probability @xmath2 as the probability that the site at the origin is unvisited by any walker till time @xmath13 .
then , the probability that the site the probability that the site is visited by a walker between time @xmath13 and @xmath62 is @xmath63 .
let @xmath64 be the probability that the site at origin is visited by a walker for the first time at time @xmath13 , whose initial position was @xmath22 at @xmath65 . within the mf approximation ,
any walker would survive with probability @xmath60 , and the probability that it will make a first visit to origin at time @xmath13 is given by @xmath66@xcite .
it follows that @xmath67 .
we now integrate @xmath64 over all initial positions @xmath22 and multiply by the probability that the origin is persistent at time @xmath13 , which is simply @xmath2 .
so we find @xmath68 where @xmath69 is the smoluchowski constant@xcite in @xmath38 . after substituting for @xmath70 and @xmath60 , we find @xmath71 , so that @xmath72 , and @xmath73 is the persistence exponent within the mf model@xcite .
it is interesting to compare the mean field prediction for @xmath37 with the exact value of the exponent . for @xmath74 , @xmath75 , while the exact value from eq.[eq : eq1 ] is @xmath76 . for the @xmath77 case , the mf model predicts @xmath78 , which is to be compared with the exact value @xmath79 . upon extending the comparison to the entire range of values of @xmath1 ,
it is clear that the mean field treatment consistently under - estimates the value of @xmath37 .
we now argue that @xmath80 is , in fact , a lower bound for @xmath4 . in the mean field approach discussed so far , it is assumed that the random walkers disappear from the lattice at random at such a rate so that their average density falls as @xmath60 .
the actual reaction - diffusion process is quite different , because only walkers which come very close to another walker are likely to be removed . clearly in regions of space
where walkers come close to each other , they are likely to visit the same site again and again .
this effect is much more within the mean field approach , where the walkers actually pass through each other , possibly several times before disappearing .
thus , it is plausible that for the same average density of walkers , a larger number of persistent sites will be visited in the actual reaction - diffusion model , compared to its mean field analogue .
since this is true for all times , the average density of persistent sites in mean field theory will be higher than the same in the actual potts model dynamics .
this would naturally imply that @xmath81 interestingly , we show now that the mean field argument presented above yields the correct value for the persistence exponent for @xmath30 model in @xmath82 .
this is not surprising , since for this model , the upper critical dimension is @xmath83 , and the mean field treatment becomes exact above this dimension .
it can be shown that in @xmath82 , the probability that a random walker starting at an an arbitrary point crosses a circle of radius @xmath84 around the origin for the first time at @xmath13 is given by the expression@xcite @xmath85 which is the analogue of smoluchowski constant in @xmath82 .
the asymptotic particle density decay for @xmath30 model in @xmath82 has the form @xmath86@xcite . upon extending the mf arguments presented previously
, we find that @xmath87 after substituting for @xmath60 and @xmath88 , and taking the limit @xmath89 , we find that @xmath90 so that @xmath91 in @xmath82 .
this result is exact , as has been shown by a rigorous field - theoretic calculation@xcite .
we thus observe that while the mean field approach , in general , gives only a lower bound for the persistence exponent , it correctly identifies the essential features that brings about this power - law decay , ie .
, the diffusive motion of interfacial points and the @xmath92 decay in their over - all density . in the following sections , we use a slightly modified version of this treatment to study the spatial distribution of persistent sites in @xmath1-state potts model
our purpose is to use the mean field approach to study the spatial distribution of persistent spins in the @xmath1-state potts model , and in particular , to understand the transition from fractal to homogeneous distribution as @xmath37 crosses @xmath39 .
however , it may be noted that in the mean field approximation to the dynamics of the potts model , the largest value of @xmath37 ( attained at @xmath77 ) is @xmath93 which is less than the transition value @xmath39 .
this problem is circumvented by defining an artificial model where we define the diffusing particles as non - interacting random walkers , who can pass through each other .
the model also allows for multiple occupancy of lattice sites .
the dynamics consists of random walkers being picked at random and taken out of the lattice at a time - dependent rate , which is tuned to produce power - law decay @xmath94 with any arbitrary @xmath95 . from the arguments presented in the previous section ,
it is obvious that this would be the case if the average density were to decay as @xmath96 asymptotically . by construction , this model is devoid of spatial correlations among reacting particles ( ie .
, it is still mean field ) but @xmath95 is now arbitrary . if we now choose @xmath97 , this model is an approximation to the @xmath1-state potts model , with the simplifying feature that the spatial correlation between interfacial points is now absent .
we shall refer to this model as the modified mean field approximation ( mmfa ) for the original potts model . in this section
, we compute the equal time pair correlation function for persistent spins in @xmath1-state potts model under the mmfa .
we define @xmath5 as the probability that both the site at origin and the site at @xmath98 are persistent at time @xmath13 .
our purpose is to compute @xmath5 for various @xmath99 .
the generalization of eq.[eq : eq3 ] to this case is @xmath100 where @xmath101 is the probability that a particle with initial position @xmath102 will make a first visit to the origin at time @xmath13 , _ without ever crossing @xmath99 in the interval [ 0:t]_. the factor 2 in front takes into account the probability that either of the sites could be reached by one of the diffusing particles . unlike the first case , @xmath101 is now different for @xmath103 and @xmath104 . for @xmath103 ,
the constraint of no crossing at @xmath99 is irrelevant for the computation of @xmath101 , since to reach @xmath99 , the particle would have to cross the origin first .
so @xmath105 simply for @xmath103 , and so @xmath106 for @xmath107 , this is no longer true , and @xmath101 needs to be computed separately .
the quantity that we need here is @xmath108 , the probability that a diffusing particle whose position at @xmath65 is @xmath22 , will reach the origin for the first time at @xmath13 , without ever crossing the point @xmath99 in between . then @xmath109 . to find @xmath108 ,
let us use the following standard method .
consider a random walk starting from @xmath110 at @xmath65 with absorbing barriers at @xmath111 and @xmath99 .
if the probability distribution of the position @xmath112 of the walker at time @xmath13 is @xmath113 , then @xmath114 the expression for @xmath113 is known exactly , and the asymptotic form at large @xmath13@xcite is @xmath115 from which , we find @xmath116 we note that for @xmath117 , the @xmath118 mode is the dominant term in the sum , and this gives @xmath119 + smaller terms that vanish as @xmath120 . clearly in this limit , we recover the @xmath121 term , as we should .
it then follows that @xmath122 where @xmath123 with @xmath124 .
after substitution of eq.[eq : eq7 ] and eq.[eq : eq11 ] in eq .
[ eq : eq6 ] , we find = 2.3 in @xmath125 \label{eq : eq13}\ ] ] which admits a scaling solution of the form @xmath126 and the scaling function @xmath127 is given by the following expression .
@xmath128 \label{eq : eq15}\ ] ] let us now consider the limiting behaviour of the scaling function for @xmath8 and @xmath10 . in the first case , it is clear from fig.1 that @xmath129 , and so @xmath130 , which implies that @xmath7 as @xmath131 . in the opposite extreme @xmath132 as @xmath133 , and so @xmath134 , which means that @xmath127 approaches a constant value in this limit .
it is further clear that , from the definition of the scaling function as given by the expression eq.[eq : eq14 ] , this constant is unity , since we expect @xmath135 as @xmath136 . for convenience of later calculations
, we approximate the scaling function as @xmath137 where @xmath84 is a number , of order unity . we see that under the mmfa , the pair correlation function has a scaling form which is same for all values of @xmath37 , with power - law decay @xmath138 for short distances @xmath48 . as is well - known , power law decay of pair correlation function points to the underlying scale invariance of the spatial distribution of the persistent spins .
this is characteristic of a fractal distribution under some circumstances . to see this ,
let us first define @xmath139 , which is the probability of finding a persistent spin at a distance @xmath99 from another persistent spin .
now , the integral @xmath140 is the total number of persistent spins within a radius @xmath141 of a persistent spin .
clearly , from the scaling form described above , @xmath142 for @xmath143 , where @xmath144max@xmath145 . for @xmath146 , we find that @xmath147 , which is simply a homogeneous distribution .
thus , if we look at length scales @xmath143 , there is a fractal structure when @xmath49 .
however , when @xmath51 , this scale - invariant structure is replaced by a few isolated sites , whose number does not grow with the length scale of observation .
clearly , the spatial distribution of persistent spins undergoes a transition as @xmath37 crosses @xmath39 . indeed ,
if we consider time scales beyond equilibration time @xmath148 , for @xmath49 the total number of persistent spins left in the system scales as @xmath149 , whereas for @xmath51 , there are only a finite number of persistent spins left .
this important difference is not adequately reflected in the pair correlation function , which has the same scaling form for all values of @xmath37 . in the next section
, we study another quantity to characterize the spatial distribution which undergoes a rather significant change in its scaling properties across the transition .
this quantity is the empty interval distribution , which is one of the standard tools in the study of one - dimensional reaction - diffusion processes .
an empty interval , in our convention , is the separation between two consecutive persistent sites .
the empty interval distribution ( eid ) @xmath11 is defined as the number of such intervals of length @xmath12 at time @xmath13 . for convenience
, we also divide this quantity with the system size @xmath150 so that @xmath11 satisfies the following normalization conditions .
@xmath151 computing the eid directly , even under the mean field approximation , is non - trivial .
instead we shall compute it from the pair correlation function using the independent interval approximation ( iia ) , where the lengths of successive empty intervals are considered as independent random variables .
the iia has been a valuable tool in the study of one - dimensional problems , and has been successfully applied to study spatial distribution of persistent spins in @xmath30 model . under the iia ,
the relation between @xmath11 and @xmath5 is @xmath152 it is convenient to express this relation in terms of the laplace transforms @xmath153 and @xmath154 , where @xmath155 was defined in the previous section . under these transformations ,
eq.[eq : eq18 ] maybe expressed in the form @xmath156 from the scaling form for @xmath5 given by eq.[eq : eq14 ] , we find that @xmath157 where @xmath158 , and @xmath159 the lower limit is put as @xmath160 instead of zero to take care of possible small argument divergence in the scaling function .
let us first consider the case where @xmath161 : in this case the scaling function @xmath127 is integrable , so we put the lower limit in the previous equation as zero .
using eq.[eq : eq16 ] , we find that @xmath162 where @xmath163 is the incomplete gamma function . after substituting in eq.[eq : eq19 ] and eq .
[ eq : eq20 ] , and taking the @xmath164 limit ( keeping @xmath1 fixed ) , we find @xmath165 \label{eq : eq22}\ ] ] it follows that @xmath166 so that @xmath167 \label{eq : eq24}\ ] ] which has the same form as eq.[eq : eq22 ] , and @xmath168dx$ ] . after integrating by parts
, we find @xmath169 \label{eq : eq25}\ ] ] where @xmath170 and @xmath171 .
we assume that @xmath172 is integrable , so that @xmath173 and @xmath174
. finally we have @xmath175 now we may try to deduce the behaviour of the function @xmath172 at large and small arguments from its laplace transform , given by the previous equation . to find the behaviour of @xmath172 near @xmath176 ,
we use the standard formula@xcite
@xmath177 where @xmath178 is the l.t of @xmath179 . now , for large @xmath1 , @xmath180 , so that @xmath181 as @xmath182 .
it follows that @xmath183 from which we find @xmath184 and , after using the relation @xmath185 , @xmath186 the behaviour of @xmath172 at large @xmath22 , one has to look for the singularities of @xmath187 in eq .
[ eq : eq26 ] .
if @xmath187 has a singularity of the form @xmath188 , then , upon inversion of the l.t , it follows that @xmath189 as @xmath133@xcite . in order to find the singularity , we plotted the denominator of eq .
[ eq : eq26 ] against its argument ( fig .
we find that the function crosses zero at one point in the negative @xmath1 axis . by careful numerical analysis
using bisection method , we have determined this crossing point to be at @xmath190 , where the numerical constant @xmath191 for @xmath192 and @xmath193 for @xmath194 .
this implies that the leading term in the decay of @xmath172 at large @xmath22 is exponential , ie .
, @xmath195 as @xmath196 , with a possible power - law prefactor .
consequently , the scaling function @xmath197 also has similar exponential decay at large @xmath22 .
@xmath198 = 2.3 in we also determine the characteristic length scales of the distribution using the scaling form for @xmath11 .
these may be defined as the ratios of moments of the distribution . @xmath199 where @xmath200 are the moments of @xmath11 .
clearly , @xmath201 by definition , whereas all higher order length scales @xmath202 which follows from the dynamic scaling form given by eq.[eq : eq23 ] , [ eq : eq29 ] and [ eq : eq30 ] for @xmath11
. we now continue our study of empty interval distribution for the case where @xmath51 .
although the mmfa allows us to study arbitrarily large values of @xmath37 , we restrict ourselves to the regime @xmath203 , since our basic aim is to study the persistence in @xmath1-state potts model where @xmath204 .
furthermore , for @xmath77 where @xmath205 , @xmath11 can be found exactly@xcite and is known to be a pure exponential . for @xmath206 , the scaling function @xmath127 has a non - integrable @xmath207 singularity near @xmath176 ( we do not study explicitly the logarithmic singularity occurring for @xmath208 ) .
we integrate by parts and find let us now define @xmath210 . for @xmath164 and finite @xmath211 , we have @xmath212 after substitution in eq.[eq : eq20 ] we find @xmath213 \label{eq : eq35}\ ] ] we define the constant @xmath214 , in terms of which @xmath215 .
now we substitute in eq.[eq : eq19 ] and find @xmath216 \label{eq : eq36}\ ] ] upon inversion of the l.t , we find that @xmath217 \label{eq : eq36a}\ ] ] where @xmath218 \label{eq : eq36b}\ ] ] the scaling function has a rather unnatural @xmath219-function singularity at the origin .
however , a more careful analysis show that for any finite ( but still large ) time @xmath13 , the divergence at origin is only power - law , but with a different exponent than the previous case ( @xmath49 ) .
we start with the expression given by eq.[eq : eq20 ] and eq.[eq : eq33 ] for @xmath220 .
after keeping the leading finite @xmath13 correction , we find that @xmath221 for purposes that will be clear later , let us define @xmath222 so that @xmath223 .
we also define the associated laplace transform @xmath224 .
the laplace transforms are related through @xmath225 using the expression eq.[eq : eq19 ] for @xmath226 , we find that @xmath227^{2 } } \label{eq : eq39}\ ] ] where @xmath228 , and is given by the expression @xmath229 \label{eq : eq40}\ ] ] after substitution in eq.[eq : eq38 ] and taking the limit @xmath164 , we find @xmath230^{2 } } \label{eq : eq41}\ ] ] which gives the scaling forms @xmath231 where @xmath232 by definition .
the laplace transform of the scaling function @xmath233 is @xmath234^{2 } } \label{eq : eq44}\ ] ] we notice that if the finite @xmath13 correction term is not included , @xmath235 is finite , and in that case , the small argument divergence of @xmath233 will be sharper than any power - law .
this is what is reflected in the appearance of the @xmath219-function in eq.[eq : eq38 ] .
however , when this term is included , the multiplying factor has to be @xmath236 in order to make the resulting expression finite as @xmath237 .
this implies that the small @xmath22 divergence for @xmath233 has the power - law form @xmath238 for small @xmath22 . from eq .
[ eq : eq45 ] , we find a similar power - law divergence in @xmath239 also .
@xmath240 in the large @xmath22 limit , @xmath239 becomes time independent , and decays exponentially with @xmath22 as in eq .
[ eq : eq36b ] , ie .
, @xmath241 the characteristic length scales are easy to compute . from the scaling form
, it follows that all the characteristic lengths have identical asymptotic scaling behaviour .
@xmath242 the difference in the asymptotic scaling behaviour of the characteristic length scale as @xmath37 crosses @xmath39 may be seen as a competition between two length scales , the diffusive length scale @xmath243 which gives the mean separation between two random walkers , and the persistence scale @xmath244 which is the mean separation between two persistent spins .
the characteristic length scale is dominated by the larger of the two , ie . , we may write @xmath245 where @xmath246 is defined through the dynamical scaling form for @xmath11 . @xmath247 the scaling function @xmath248 when @xmath49 and @xmath249 when @xmath250 . in general , the small argument behaviour of @xmath21 has the power - law form @xmath251 where the exponents @xmath54 and @xmath18 are given by @xmath252 ,
\label{eq : eq51}\ ] ] and @xmath57 is the heaviside step function . for large @xmath22 , the scaling function is time - independent and decays exponentially with @xmath22 .
we also note from the scaling form that over small distances @xmath253 , @xmath254 where @xmath18 is given by eq.[eq : eq51 ] .
we studied the quantities @xmath5 and @xmath11 numerically by simulating the kinetics of @xmath1-state potts model with random initial conditions .
the time evolution of spin configurations via glauber dynamics is implemented using the mapping of this dynamics to a reaction - diffusion problem , as mentioned in the introduction . in this procedure , a set of diffusing particles @xmath42 are initially distributed at random on the lattice with a certain average initial density @xmath255 ( which we fix as @xmath39 ) .
when two diffusing particles meet , they annihilate each other or coagulate with probability @xmath58 and @xmath59 respectively .
we count one mc step in the simulation after the position of every particle in the lattice has been updated once .
persistent spins(sites ) at any time @xmath13 are those sites which have not been touched by a random walker till that time .
all the simulations were done on a lattice with @xmath256 sites , and the results were averaged over 100 different starting configurations . in order to check the different dynamic scaling behaviour for @xmath257 and @xmath51 , we did our simulations for three different values of @xmath1 - 2,5 and 10 . for later reference , we note that from eq.[eq : eq1 ] , the corresponding values of the persistence exponent are @xmath258 and @xmath259 . in figs . 3 - 5 , and later in figs . 7 - 10 , we have employed logarithmic binning of the data in intervals of size @xmath260 ( @xmath261 ) . since the statistical noise was considerable . however , for all exponent measurements , we have used only the bare(not binned ) data . in fig .
3 - 5 , we have plotted the scaling function @xmath127 for the pair correlation function @xmath5 against the scaling variable @xmath262 for three @xmath1-values , @xmath263 and 10 .
we find excellent scaling collapse for all three values of @xmath1 , which is in agreement with the dynamic scaling picture provided by the mmfa in eq.[eq : eq14 ] . in the figures , we find power - law decay of @xmath127 for small @xmath22 , with a sharp cross - over to the flat long - distance behaviour , which is also in agreement with the assumption we made in eq.[eq : eq16 ] .
we also note that the constant @xmath84 introduced in eq.[eq : eq16 ] is in fact very close to 1 .
= 1.8 in = 1.8 in = 1.8 in in fig . 6
, we plot the characteristic length scale @xmath264 against time @xmath13 for @xmath263 and 10 and measure the dynamical exponent @xmath15 .
the observed slopes of the lines are systematically higher than the theoretical prediction in eq.[eq : eq50 ] by around 0.05 , while the statistical error in all the three cases was only @xmath265 or smaller .
the observed deviation could be possibly due to the fact that the asymptotic behaviour is not fully reflected over the time scales which we used .
the number of persistent spins left in the system falls rapidly with time for high values of @xmath1 , and so we were forced to restrict ourselves to times @xmath266 .
in fact , even for @xmath74 case , previous simulations over longer time scales have shown the presence of an additive power - law correction to the asymptotic scaling behaviour@xcite .
= 1.8 in = 1.8 in in fig .
7 - 9 , we check the dynamic scaling form eq.[eq : eq51 ] for @xmath11 against the scaling variable @xmath267 for three values of @xmath12,5 and 10 .
we find that for @xmath74 , excellent scaling collapse is obtained with @xmath268 ( fig .
7 ) . for small @xmath22
, we find power - law decay of the scaling function , which crosses over to fast exponential decay at large @xmath22 . for higher values of @xmath1 ( where @xmath269 ) , we find that with @xmath270 , we find very good scaling collapse for @xmath10 . but for @xmath8
, we find systematic deviation from scaling collapse , which was also observed earlier by bray and odonoghue@xcite .
this observation supports the theoretical prediction based on mmfa , and shows that in this regime , the scaling function has explicit time dependence . to show this more clearly , and to verify the predicted time - dependence , we plotted the quantity @xmath271 against @xmath12 for three widely spaced values of @xmath13 for all three @xmath1 values studied in fig.9 .
we see that in all three cases , a simple power - law decay with @xmath12 is observed for @xmath253 , thus validating eq.[eq : eq52 ] .
the measurement of the exponent @xmath18 gives values in reasonable agreement with theoretical prediction , although for @xmath272 , the statistical error is significant .
= 1.8 in = 1.8 in = 1.8 in
in this paper , we have studied the spatial aspects of persistence in one dimensional @xmath1-state potts model using a mean field approximation .
we have computed the pair correlation function for persistent spins under this approximation , and used it to compute the empty interval distribution under the independent interval approximation .
we find dynamical scaling behaviour in both these quantities .
the time dependence of the characteristic length scale and the behaviour of the scaling function was found in both cases .
we showed analytically within the mean field approximation the transition from fractal to homogeneous distribution of persistent spins as the persistence exponent crosses @xmath39 .
we support our results by numerical simulations in the kinetic @xmath1-state potts model .
this research was supported in part by a grant ( dmr 0088451 ) from the u.s national science foundation .
the author would like to thank p. ray for a critical reading of the manuscript and suggestions for improvement . for a review , s. n. majumdar , curr .
( india ) * 77 * , 370 ( 1999 ) ( e - print cond - mat/9907407 ) .
b. derrida , v. hakim and v. pasquier , phys .
lett . * 75 * , 751 ( 1995 ) ; j. stat . phys . * 85 * , 763 ( 1996 ) .
d. h. zanette , phys .
e. * 55 * , 2462 ( 1997 ) .
g. manoj and p. ray , e - print cond - mat/9901130 .
g. manoj and p. ray , j. phys .
a * 33 , * l109 ( 2000 ) .
g. manoj and p. ray , j. phys .
a * 33 * , 5489 ( 2000 ) .
a. j. bray and s. j. odonoghue , phys .
e * 62 * , 3366 ( 2000 ) .
s. jain and h. flynn , j. phys .
a * 33 * , 8383 ( 2000 ) .
g. i. menon , p. ray and p. shukla , phys .
e * 64 * , 046102 ( 2001 ) .
g. manoj and p. ray , phys .
e * 62 * , 7755 ( 2000 ) .
j. cardy , j. phys .
a * 28 * , l19 ( 1995 ) .
s. j. odonoghue and a. j. bray , phys .
e * 64 * , 041105 ( 2001 ) .
b. derrida and r. zeitak , phys .
e * 54 * , 2513 ( 1996 ) .
w. feller , _ introduction to probability theory and applications _
( wiley , new york , 1966 ) vol . 1 .
m. von smoluchowski , z. phys . chem . * 92 * , 129 ( 1917 ) .
d. c. torney and h. m. mcconnell , proc .
( london ) ser .
a * 387 * , 147 ( 1983 ) .
b. p. lee , j. phys .
a * 27 * , 2633 ( 1994 ) .
m. g. smith , _ laplace transform theory _ ( van nostrand and company , london , 1966 ) . | we study the persistence properties of the @xmath0 coarsening dynamics of one dimensional @xmath1-state potts model using a modified mean - field approximation ( mmfa ) . in this approximation
, the spatial correlations between the interfaces separating spins with different potts states is ignored , but the correct time dependence of the mean density @xmath2 of persistent spins is imposed . for this model , it is known that @xmath2 follows a power - law decay with time , @xmath3 where @xmath4 is the @xmath1-dependent persistence exponent .
we study the spatial structure of the persistent region within the mmfa .
we show that the persistent site pair correlation function @xmath5 has the scaling form @xmath6 for all values of the persistence exponent @xmath4 .
the scaling function has the limiting behaviour @xmath7 ( @xmath8 ) and @xmath9 ( @xmath10 ) .
we then show within the independent interval approximation ( iia ) that the distribution @xmath11 of separation @xmath12 between two consecutive persistent spins at time @xmath13 has the asymptotic scaling form @xmath14 where the dynamical exponent has the form @xmath15=max(@xmath16 ) .
the behaviour of the scaling function for large and small values of the arguments is found analytically .
we find that for small separations @xmath17 where @xmath18=max(@xmath19 ) , while for large separations @xmath20 , @xmath21 decays exponentially with @xmath22 .
the unusual dynamical scaling form and the behaviour of the scaling function is supported by numerical simulations . 2 |
the study of spin - polarized transport across interfaces is a subject of long history @xcite .
the recent advent of semiconductor - based electronic devices at the nanoscale has revived the interest in transferring , controlling and detecting spin currents .
this research area has been termed _
spintronics _ due to the exciting possibility of future , successful spin - based electronic technology @xcite .
nevertheless , spintronics is interesting as well for fundamental physics , both experimentally and theoretically , as its basic constituent the spin is of quantum nature only .
the most simple building block of spintronic transport systems is probably the magnetic tunnel junction .
it comprises two ferromagnetic electrodes sandwiching a paramagnetic layer .
vertical transport , where current flows across the interfaces , is characterized by the tunneling magnetoresistance ( tmr ) , which measures the relative change in the junction resistance when the contacts magnetizations is changed from parallel to antiparallel alignment @xcite .
the tmr is also important in investigating the properties of spintronic resonant tunneling diodes .
remarkably , such devices have been recently built in all - semiconductor heterostructures , taking advantage of diluted magnetic semiconductors made of iii - v @xcite ( hole - like transport ) and ii - vi compounds @xcite ( electron - like transport ) . when the size of the paramagnetic island in a magnetic tunnel junction becomes comparable to the carrier fermi wavelength , the system behaves effectively as zero - dimensional .
then , quantum effects arise from the quasi - localized nature of electrons and from the phase - coherent transport .
the ultimate miniaturization limit is just a single resonant level coupled to a fermi sea of itinerant electrons , which may be regarded as an artificial realization of the quantum _ impurity problem _ @xcite .
extensive studies of the impurity problem have been performed in semiconductor _ quantum dots _
@xcite , where the island ( an electron droplet ) is formed by means of a constriction in a two - dimensional electron gas .
both the discrete energy levels and the tunneling couplings may be tuned almost at will .
very recently , a few experimental works have begun to deal with spin polarized leads @xcite . in this paper
, we consider spintronic transport through quantum dots with strong correlations .
it is well known that the electron - electron interaction interaction plays a dominant role in the low temperature transport through quantum dots @xcite . in the coulomb blockade regime
, the electron dynamics can be described in terms of single - electron tunneling plus mean - field charging effects . between two coulomb blockade peaks ,
transport is blockaded and the electron number in the dot does not fluctuate . when the electron number is odd , the topmost resonant level is singly occupied .
however , when temperature approaches the energy scale @xmath0 , the spin of the localized electron becomes screened by an antiferromagnetic interaction with the conduction band electrons .
the resulting strong correlations arise from the interplay of higher order tunneling processes and the on - site coulomb interaction . as a consequence ,
the many - body state in the @xmath1 limit is a singlet formed between the quantum impurity and the continuum electrons .
the impact in the transport properties of the system is strong ; e.g. , the temperature - dependent conductance results in a universal function of @xmath2 , achieving the quantum limit ( @xmath3 per spin ) at @xmath1 .
this is the celebrated kondo effect @xcite in quantum dots @xcite , which was observed several years ago @xcite .
now , an important condition for the kondo effect to take place is the degeneracy ( between spins up and down ) of the ground state of the dot .
such spin degeneracy may be broken with an external magnetic field ( zeeman splitting ) and is well understood @xcite .
but will the kondo effect be preserved when the spin transfer across the tunnel barrier is spin - dependent ?
how will the conventional picture of the kondo resonance in quantum dots be affected within a spin - polarized medium ? to answer these questions various theoretical groups have lately contributed . although these works differ in some predictions due to the range of applicability of the distinct approaches used therein and their limitations , the analysis of the problem with numerical renormalization group @xcite have led to the conclusion that the kondo state is robust enough ( though with a lower @xmath4 ) against nonzero spin polarizations in the leads when particle - hole symmetry is not broken and real charge fluctuations are completely suppressed @xcite . note that particle - hole asymmetry may be induced in the dot with nearby electric gates , shifting the resonant level away from the symmetric point ( see below ) @xcite .
we predict that this would give rise to a sharp decrease of the linear - response conductance .
since the phenomenon is absent when the magnetizations of the electrodes point along opposite directions ( antiparallel alignment ) , we propose @xcite the tmr as a possible experimental signature of this _ spintronic kondo effect_. another important element of many theories of spintronic transport is the description of intrinsic `` spin relaxation '' mechanisms that allow for nonequilibrium spin populations to relax .
long spin coherence times @xmath5 have been reported in semiconductor quantum wells @xcite and dots @xcite .
the effect of spin relaxation is known to reduce the tmr for a coulomb - blockaded quantum dot @xcite and it leads to a suppression of the fano factor ( shot noise ) in the antiparallel configuration @xcite . at lower temperatures ( @xmath6 ) , spin decoherence causes the destruction of the kondo effect due the failure of the formation of the many - body singlet state .
one could also think about more coherent `` spin - flip '' process , e.g. , arising from the potential spin - orbit coupling which causes the rotation of electron spin in the dot ( this effect of spin - orbit coupling for the localized electron should be distinguished from that of the spin - orbit coupling of conduction electrons , which due to time - reversal symmetry has no influence on the kondo effect @xcite ) . more specifically , when the amplitude of spin flip scattering rate is larger than the kondo temperature , @xmath7 , the density of states ( dos ) at the impurity site is expected to develop a splitting and thus a decrease of the linear conductance .
this prediction has been confirmed with equation - of - motion technique @xcite , slave - boson mean - field theory @xcite and numerical renormalization group @xcite .
therefore , spin - flip processes , coherent or incoherent , tend to suppress the kondo effect .
we model the quantum dot as a single discrete level with energy @xmath8 containing an unpaired spin-@xmath9 electron with @xmath10 and charging energy @xmath11 .
therefore , the dot is an electronic impurity tunnelling coupled to continuum electrons with a model hamiltonian given by the anderson hamiltonian : @xmath12 where ( see fig . [ scheme ] ) @xmath13 are written in terms of the creation and annihilation operators in the dot @xmath14 ( the occupation number is defined as @xmath15 ) and in the leads @xmath16 , with @xmath17 the wavevector and @xmath18 labeling left ( @xmath19 ) and right ( @xmath20 ) reservoirs .
tunneling of electrons from the dot to the leads is described by the hopping parameter @xmath21 . in @xmath22
included is an internal spin - flip process with rate @xmath23 @xcite .
notice that in this framework the spin - flip process is purely coherent , and precisely speaking it does not account for incoherent spin relaxation processes .
it may originate either from the transverse component of an applied magnetic field or from a tunable spin - orbit coupling of the rashba - dresselhaus type _ in the dot _
@xcite ( see the introduction and compare with ref .
since the spin - flip processes , coherent or incoherent , have similar influence on kondo effect , we leave the term in @xmath24 in @xmath25 phenomenological .
what is important here is that @xmath24 lifts the degeneracy of the discrete level and that it can not be eliminated with a unitary transformation since the lead magnetizations already mark a privileged spin direction . for @xmath26 and @xmath27 , the su(2 ) symmetry
is broken and the spin symmetry of the problem is u(1 ) whereas in the presence of both spin flip scattering and ferromagnetic electrodes , the u(1 ) spin symmetry is explicitly broken .
due to coupling to the leads , the electron in the dot becomes quasilocalized with a escaping rate related to the hybridization broadening , @xmath28 .
this is the imaginary part of the hopping self - energy , which is spin - dependent because tunneling is spin - dependent .
this can be achieved by coupling the dot to ferromagnetic leads .
we take constant tunneling coefficients @xmath29 and equal tunnel barriers ( symmetric couplings : @xmath30 ) . in the wide - band limit ,
the energy dependence of @xmath31 is unimportant ( which is a good approximation for low voltages ) .
moreover , we assume that the degree of spin polarization on lead @xmath18 is given by @xmath32 notice that eq .
( [ eq_p ] ) is already a gross simplification as it might well be that @xmath33 has little to do with the real magnetization of the reservoir .
in fact , various definitions for @xmath34 are possible depending on the experiment @xcite .
in addition , we neglect proximity effects such as stray fields coming from the ferromagnets and consider that the bandwidth @xmath35 is spin independent .
[ we prefer not to delve into details since already the simple form of eq .
( [ eq_p ] ) gives rise to nontrivial effects which can be directly measured ] .
we consider collinear magnetizations , both in parallel ( p ) and antiparallel ( ap ) configurations . with the approximations discussed above , we have for the p case ( @xmath36 ) @xmath37 and @xmath38 , where @xmath39 , whereas the ap case ( @xmath40 ) yields @xmath41 and @xmath42 .
let us first provide an intuitive picture of the influence of spin - polarized transport in the kondo resonance at @xmath26 . we take @xmath43 . for p alignment in the fully polarized case ( @xmath44 )
, the singlet state can not form due to the lack of spin down electrons .
hence , we expect a decrease of the kondo temperature ( @xmath4 is roughly the binding energy of the singlet state ) with increasing @xmath34 . in the ap configuration ( @xmath45 ) , however , the kondo effect survives since an spin up ( down ) localized electron may be screened by the right ( left ) electrode .
of course , the conductance would be zero unless a vanishingly small @xmath24 is allowed to come into play .
now , in the p case there may arise an exchange field @xcite acting on the dot as an effective zeeman splitting @xcite .
is the kondo effect robust against this exchange field ?
the answer is yes !
when the gate voltage is tuned in such a way that @xmath46 ( the symmetric anderson model ) , charge fluctuations become suppressed . only when particle - hole symmetry is broken ( @xmath47 ) do we find a splitting in the kondo peak of the local dos .
we briefly review now the different theoretical methods employed to solve eq .
( [ eq_h ] ) .
the equation - of - motion technique @xcite is useful to study nonequilibrium situations ( for finite bias ) at relatively `` high '' temperatures ( @xmath48 ) .
although it reproduces qualitatively the dos peaks , it fails to describe properly the strong coupling regime , where kondo physics completely quenches the impurity spin . when applied to our problem
, it predicts the exchange field induced splitting but not its disappearance at @xmath46 . on the other hand , slave - boson mean - field theory @xcite correctly accounts for the fermi - liquid fixed point of the kondo problem at @xmath1 . as a result , it is only valid when the particle - hole symmetry is not broken ( no splitting ) .
the noncrossing approximation @xcite is another slave - boson based approach and offers a consistent picture of the kondo effect at @xmath0 .
however , it does not take into account vertex corrections and produces spurious peaks at @xmath49 in the presence of a magnetic field .
finally , a numerical renormalization group calculation @xcite encompasses the whole regime but remains valid only at equilibrium . in the following , we report results using the interpolative @xmath11-finite perturbation theory @xcite since it gives a good description of the dynamical properties of eq .
( [ eq_h ] ) for a wide range of parameters .
it can describes both , the kondo regime and the mixed - valence regime ( where the dot level is close to @xmath49 , @xmath50 ) .
however , in this approach the width of the kondo resonance decreases algebraically instead of having an exponentional decay .
then , we elaborate as well on a numerical renormalization group analysis , which leads to _ nonperturbative _ results for all the regimes listed above .
as indicated above , the kondo resonance for a quantum dot coupled to two ferromagnets with parallel magnetizations , splits away from the symmetric case ( @xmath51 ) where charge fluctuations are important .
figure [ fig2 ] shows our results using the interpolative @xmath11-finite perturbation theory including magnetic leads .
the dos for the symmetric anderson model is plotted in fig .
[ fig2](a ) for the unpolarized case @xmath52 and for nonzero polarization @xmath53 in the p configuration . for unpolarized leads
, the dos shows the usual kondo resonance reaching the _ unitary limit _ and two broad peaks at @xmath54 corresponding to the two mean - field ( electron - like and hole - like ) peaks .
the main effect of the polarized reservoirs is to make the kondo resonance narrower but keeping the same dos height at @xmath49 ; i.e. , for @xmath46 the lead magnetizations preserve the unitary limit .
the physical scenario changes dramatically when charge fluctuations are important as in the mixed valence regime . the solid line in fig .
[ fig2](b ) corresponds to @xmath55 for @xmath52 .
the dos displays a strongly renormalized level by the charge fluctuations close to @xmath49 with no evidence of the mean - field peaks . for a finite spin polarization @xmath53 , fig .
[ fig2](b ) depicts both the spin up @xmath56 and the spin down @xmath57 contributions to the local dos @xmath58 .
here , @xmath59 , moves toward positive ( negative ) frequencies . as a result , @xmath60 shows a splitting at low frequencies and the quantum occupations per spin change appreciably : @xmath61 and @xmath62 . this demonstrates the sensitivity of the spintronic kondo effect to variations of the external gate voltage . at @xmath1 with @xmath63 .
solid line for @xmath52 , and dashed line for @xmath53 ( @xmath64 ) .
( b ) density of states for the asymmetric anderson model corresponding to the mixed valence regime , @xmath55 at @xmath1 .
the nonmagnetic case corresponds to the solid line .
spin up and down density of states for @xmath53 are shown with the dashed lines.,width=321 ] we now use a numerical renormalization group calculation to investigate both the linear conductance and the splitting in the total dos as a function of the gate voltage @xmath65 and the polarization of the leads @xmath34 . figure [ nrg](a ) shows the splitting @xmath66 of the kondo peak as a function of the gate voltage @xcite .
it increases linearly from zero as the gate moves away from the symmetric point @xmath46 . in terms of the lead polarization @xmath34 ( the arrangement is parallel )
, @xmath66 is linear as well [ see fig . [ nrg](b ) ] . of the kondo peak as a function of @xmath67 for @xmath68 and @xmath69 ( @xmath35 is the continuum bandwidth ) .
( b ) @xmath66 versus @xmath34 for @xmath70 and @xmath71 .
( c ) tmr versus @xmath34 for @xmath70 and @xmath69 .
( d ) tmr as a function of @xmath67 for @xmath68 and @xmath69 . in all cases
@xmath72.,width=321 ] in fig .
[ nrg](d ) we plot the tmr defined as @xmath73 where @xmath74 ( @xmath75 ) is the linear conductance in the p ( ap ) case . for the symmetric anderson model , the kondo effect survives even for a finite value of polarization @xmath76 .
then , at @xmath46 we find that @xmath77 , in excellent agreement with the numerical result . away from the symmetric point ,
i.e. , @xmath78 , @xmath79 gets strongly suppressed as @xmath34 increases .
then , the system exhibits a strong negative tmr [ see fig . [
nrg](c ) ] . as a result
, we predict a _ sharp _ peak of the tmr by varying the gate potential [ see fig .
[ nrg](d ) ] .
the origin of this peak is exclusively due to the particularities of the spintronic kondo effect .
we have so far discussed the case of a quantum dot symmetrically coupled to the leads , i.e. , @xmath80 .
as we have seen , the p configuration leads to nontrivial effects in the transport properties of the dot for @xmath81 since @xmath82 and @xmath83 are not equally coupled to the leads whereas for the ap configuration both @xmath82 and @xmath83 are renormalized by the kondo correlations in the same manner .
this scenario is modified when we take an asymmetric quantum dot , @xmath84 . in this case , both configurations ( p and ap ) give rise to a split dos since @xmath85 is coupled to the leads with @xmath86 unlike @xmath87 ( with @xmath88 ) . in general , there will be splitting provided @xmath89
the shot noise are the dynamical fluctuations of the current ( current - current correlations ) that appear in electric conductors due to the quantization of the charge .
research on shot noise in mesoscopic physics has developed into a fruitful area of research @xcite .
nevertheless , there have hitherto been very few attempts to investigate shot noise in kondo impurities @xcite . here , we consider the the noise power ( the fourier transform of the time correlator of the electric current ) at zero frequency : @xmath90 \ , , \label{noise1}\ ] ] where @xmath91 describes the fluctuations of the current away from its average value @xmath92
. we shall work at @xmath1 so that the current will fluctuate due to quantum fluctuations only ( we disregard thermal fluctuations ) . within slave - boson mean - field theory ,
the shot noise in a two - terminal geometry is shown @xcite to have the well known expression @xmath93 , i.e. , the conventional result for the partition noise but with _ renormalized _
transmissions @xmath94 .
this is valid as long as we restrict ourselves to the fermi - liquid fixed point of the kondo problem .
versus applied dc voltage @xmath95 of a quantum dot in the kondo regime attached to ferromagnetic leads and in the presence of spin flip scattering with rate @xmath24 .
voltage is units of the kondo temperature @xmath96 .
we take @xmath97 and @xmath98 .
( a ) lead polarizations are @xmath99 .
( b ) @xmath26 and @xmath100 ( parallel alignment).,width=321 ] it is customary to define the fano factor : @xmath101 since we are dealing with a two - terminal system , we have dropped the lead indices .
now , for a classical conductor with no correlations , the fano factor equals 1 ( poissonian limit ) .
deviations of this limit are usually due to the application of pauli principle or to the effect of strong electron - electron interactions as those giving rise to the kondo effect . in fig .
[ noise](a ) , we plot the influence of spin flips in @xmath102 .
the polarization in the leads is taken as @xmath99 . at low bias
, @xmath102 behaves as @xmath103 @xcite . since for @xmath26
the kondo resonance achieves the unitary limit at zero bias , the fano factor is completely suppressed down to zero . as @xmath24 increases , spin flips induce decoherence in the correlated motion of the electrons which leads to the singlet formation .
hence , the transmission at @xmath49 departs from its unitary limit and , as a consequence , @xmath102 increases at zero bias ( @xmath104 ) . for larger voltages ,
notice that at @xmath26 we recover the limit @xmath105 of a double - barrier resonant system @xcite .
for @xmath106 the behavior of @xmath102 at larger @xmath95 depends on the particular two - peak structure of the transmission @xcite . in fig .
[ noise](b ) , we calculate the fano factor for @xmath26 and different values of the lead magnetizations ( p configuration , @xmath100 ) .
we find that @xmath102 increases with the polarization @xmath34 at a given voltage bias ( except at @xmath104 ) .
this is caused by the suppression of the kondo effect for @xmath107 . for ap alignments ( not shown here ) , @xmath102 increases less rapidly due to the independence of @xmath4 on the lead polarization .
notice that in slave - boson mean - field theories the fluctuations of the boson field are neglected .
however , we do not expect large deviations from the results reported here when @xmath108 .
the boson fluctuations will evidently become important as temperature approaches @xmath4 .
we have demonstrated that rich physics appears when the formation of the kondo state in a quantum dot competes with the presence of spin - polarized tunneling currents and spin - flip processes .
we discuss now possible extensions of the theory that could dramatically alter the effects exposed above .
we are confident that the field will still offer unexpected results and that , consequently , future calculations and experiments will be full of rewards .
we have thus far considered metallic free - electron ferromagnets as the injecting and receiving contacts . in reality ,
transition - metal electrodes are described by exchange hamiltonians . in these models , it is assumed that electrical conduction is carried by itinerant @xmath109-electrons while ( insulator ) magnetism is caused by a different group : localized @xmath110-electrons .
interaction between free electrons and localized moments gives rise to _ electron - magnon _ coupling @xcite .
it has been suggested that magnon - assisted tunneling in magnetic tunnel junctions may lower the tmr as a function of the bias voltage @xcite , giving rise to a zero - bias anomaly in nonlinear current voltage characteristics @xcite ( see , e.g. , ref .
@xcite for a more detailed review on the subject ) .
the peak width is given by the energy involved in the spin excitations , which is of the order of the curie temperature ( @xmath111 ) of the metal .
as @xmath112 , one would naively expect that the kondo effect will be always destroyed by emission and absorption of magnons via spin - flip processes . within the tunneling hamiltonian formalism
, we replace @xmath113 in eq .
( [ eq_h ] ) with @xmath114 where the first term describes the conduction band electrons , the second term is the heisenberg interaction between localized moments @xmath115 at neighboring sites @xmath116 and @xmath117 of lead @xmath18 and the third term is the interaction between a localized moment at site @xmath116 and an itinerant electron with creation operator @xmath118 . after applying the holstein - primakoff transformation ,
the ferromagnet low - energy spin excitations can be written in terms of a bosonic collective bath ( magnons ) , each carrying a magnetic moment @xmath119 .
thus , electron - magnon interaction @xmath120 at the interface induces spin mixing @xcite : @xmath121\,,\ ] ] where the @xmath122 is a renormalized coupling constant ( which is taken as momentum independent for simplicity ) and @xmath123 is the magnon creation operator . in eq .
( [ eq_hem ] ) , we have written down only the spin - flip part of the electron - magnon interaction .
it will lead to nontrivial correlations when combined with the hamiltonian @xmath22 of a coulomb - blockade quantum dot in eq .
( [ eq_h ] ) .
in fact , @xmath124 involves spin - flip inelastic transitions ( the term proportional to @xmath24 in eq .
( [ eq_hr ] ) is simply elastic ) , inducing decoherence in the kondo resonance by means of emission ( dominant at low @xmath125 ) and absorption of magnons of energy @xmath126 .
( at low temperatures , @xmath127 depends approximately on @xmath128 in a quadratic way , @xmath129 ) .
furthermore , subtle out of equilibrium effects such as interlayer exchange interaction @xcite may arise as well .
we should mention that magnon excitations may also act as a _ dissipative _ bath .
since the standard spectral density for a three - dimensional ( cubic ) ferromagnet goes as @xmath130 @xcite , we would deal with a _ subohmic _ bath .
( the spectrum of the bath is cut off by a maximum magnon frequency due , e.g. , to an anisotropy energy ) . for comparison ,
the dos of magnons in antiferromagnetic two - dimensional systems varies as @xmath131 , which amounts to an _ ohmic _ bath .
double - quantum - dot ( dqd ) systems have recently attracted much attention since that they form the simplest artificial systems showing molecule - like correlations at the nanoscale . as a consequence
, a dqd has been proposed as a basic constituent of a solid - state quantum computer @xcite .
is the interdot exchange coupling between the spin of the dots .
( a ) parallel configuration of the leads polarization .
dashed arrows correspond to the majority spins and solid arrows are for minority ones .
( b ) antiparallel alignment of the leads polarization.,width=321 ] two quantum dots can be coupled either in series or in parallel , allowing for tunneling and capacitive couplings in between . as far as kondo physics
is concerned , a dqd may be regarded as an artificial realization of the _ two - impurity kondo problem _
it consists of two kondo impurities with spin @xmath132 and @xmath133 interacting via an antiferromagnetic ( af ) exchange coupling , @xmath134 : @xmath135 where @xmath136 creates an electron at dot 1 ( 2 ) with spin @xmath137 ( @xmath138 ) .
it has been shown that the ratio @xmath139 determines the ground state of the system .
in particular , when @xmath140 the two dots are locked into a antiferromagnetic singlet state whereas for @xmath141 each dot forms its own kondo state with continuum electrons in the leads .
the critical value at which the transition from the kondo state ( ks ) to the af phase takes place can be obtained by comparing their ground state energies .
thus , the critical point depends on the kondo temperature for each dot ( @xmath142 ) as follows @xmath143 for a symmetrically coupled dqd with a common gate @xmath144 one has @xmath145 .
in general for @xmath146 we have @xmath147 .
since @xmath148 , and @xmath149 depend exponentially on the tunneling couplings and the level positions , a small asymmetry between these parameters induces a huge change in the ratio @xmath150 .
let us consider a parallel dqd ( see fig .
[ dqd ] ) connected to two ferromagnetic leads .
the kondo temperature for the dot @xmath151 depends on the configuration of the polarization of the leads ( parallel or antiparallel ) . to simplify , we take the same polarization for the leads @xmath152 ( p alignment ) and identical dots @xmath153 . now
the transition from ks@xmath154af singlet state is achieved more easily by increasing @xmath34 .
we keep the antiferromagnetic coupling fixed @xmath155 and vary @xmath34 . in this way , @xmath156 becomes smaller leading to a weaker kondo effect . here
, the width of the zero - bias anomaly ( zba ) decreases with @xmath34 and the conductance reaches the unitary limit as in the case of a single dot @xcite . by further increasing @xmath34 ,
the af coupling is much stronger than the kondo scale leading to the transition @xmath157 when @xmath158 . in dqd systems with a strong interdot coulomb interaction , the total charge allowed in both dots at the same time is just one electron . as a consequence ,
there are four ground states with the same energy , namely , @xmath159 . quantum fluctuations between these states due to coupling to the leads yield , in the low temperature limit , a highly correlated state with su(4 ) symmetry @xcite .
note that these fluctuations do not involve only spin flips in the dqd ( spin kondo effect ) but also flips in the _ orbital _
sector @xcite . to describe these new processes
, we define the _ pseudospin _ as a fictitious spin that points along @xmath160(@xmath161)@xmath162 when the electron lies at the dot 1(2 ) .
then , it is shown that the spin kondo state becomes intermingled with a pseudospin kondo state , giving rise to a complete entanglement between the spin and charge degrees of freedom .
the spin - pseudospin entanglement develops from the correlated tunneling that involves a flip of the spin _ and _ the pseudospin of the dqd system at the same time .
technically , it arises from the schrieffer - wolff transformation @xcite which maps the dqd hamiltonian into an effective exchange coupling between the localized spin and pseudospin and the conduction electrons . in the dqd hamiltonian ,
one replaces @xmath22 in eq .
( [ eq_h ] ) with the hamiltonian of two dots plus a charging energy @xmath163 between them : @xmath164 @xmath165 denotes the occupation number on dot @xmath116 .
we denote with @xmath166 the pseudospin operator in the dqd system .
the resulting kondo hamiltonian formally reads : @xmath167 where @xmath168 is the field operator with @xmath169 $ ] a spinor in the representation of even and odd channels of the lead operators @xcite . in eq .
( [ eq_su4 ] ) @xmath170 is a coupling constant which goes to the strongly fixed point in the flow diagram .
we recall that the su(4 ) kondo state takes place provided there are two conduction channels ( described by the matrix @xmath171 ) .
the latter equation explains the entanglement between the spin and the orbital electronic degrees of freedom .
the transport properties of a su(4 ) kondo state strongly differ from the conventional su(2 )
kondo state , both at equilibrium and out of equilibrium @xcite . first , the kondo temperature inferred from eq .
( [ eq_su4 ] ) is largely enhanced ( around 200 times ) as compared to @xmath4 of a spin kondo system @xcite .
this means that the differential conductance peak ( which mimics the dqd density of states ) becomes greatly broadened in a transition from the su(2 ) to the su(4 ) kondo physics .
such a transition can be tuned with a magnetic flux in an aharonov - bohm interferometer with one dot at each arm @xcite .
second , the kondo resonance is no longer peaked at @xmath49 but at @xmath172 to fulfill the friedel - langreth sum rule @xcite . how would spintronic transport modify a su(4 ) kondo resonance ?
in the presence of ferromagnetic leads and away from the particle - hole symmetry point , we expect the spin part of the kondo state to slowly vanish with increasing lead polarization .
experimentally , one would see a splitting of the kondo resonance into _ three _ peaks .
the centered peak would still correspond to the pseudospin kondo state , which is not sensitive to the magnetization at the leads . at the same time , @xmath173 decreases but the linear conductance would increase since the su(2 ) kondo resonance associated to the orbital kondo effect peaks at @xmath49 again . to further destroy the pseudospin kondo state ,
two possibilities emerge from an analogy with the spin case .
first , one allows for tunneling between the dots , which breaks the fourfold degeneracy favoring the formation of a bonding ( symmetric ) state between the dots .
then , interdot tunneling acts as an external zeeman splitting in the spin sector .
second , one could consider asymmetric couplings of the dqd to the leads ; e.g. , the dqd system may be coupled strongly to the left lead , @xmath174 . this way , we regard the leads as _ pseudospin polarized _ much like the spin - dependent tunneling due to ferromagnetic leads in the spin case
. we may even define the pseudospin polarization for each spin species as @xmath175 [ cf .
( [ eq_p ] ) ] . at this point ,
more calculations are needed to further exploit this analogy between spins and pseudospins , which may give rise to a unifying picture of the influence of real and pseudo - spin polarized leads in the transport through quantum - dot structures .
we have investigated the spintronic properties of a quantum dot in the kondo regime .
we have considered a dot attached to ferromagnetic leads and in the presence of intradot spin flip scattering . using both perturbation theory in the on - site interaction and the nrg method ,
we have shown that the kondo effect is not necessarily suppressed by the spin polarizations of the leads : for the symmetric anderson model , where charge fluctuations are completely suppressed , the kondo effect is robust even for finite polarizations . for the asymmetric anderson model ,
the kondo peak does split into two .
this is due to the presence or absence of particle - hole symmetry . in the presence of particle - hole symmetry , the kondo peak at the fermi level
remains unsplit even at finite polarizations and the linear conductance achieves the unitary limit .
this remains true as long as only spin fluctuations are present in the qd . on the contrary ,
when particle - hole symmetry is absent , the conductance is suppressed due to the visible splitting of the kondo peak .
since the kondo resonance is mostly unaltered for antiparallel magnetizations , we have calculated the tmr in the kondo , mixed - valence , and empty - level regimes .
the tmr shows a characteristic behavior for each of them .
in addition , we have shown that the tmr is strongly affected in the presence of spin flip processes .
we have studied the form of the shot noise when charge fluctuations are completely suppressed .
we have shown that the fano factor approaches the poissonian limit when the spin flip scattering rate is of the order of the kondo temperature .
for parallel arrangements , the fano factor enhances with increasing lead polarizations .
moreover , we have suggested and discussed possible new advances in this field such as the influence of magnons in the kondo state of a quantum dot , the effect of spin - polarized currents on double - quantum - dot systems mimicking the two - impurity problem and on more exotic kondo states with higher symmetry .
the physics addressed in this paper is realistic and can be visible within the scope of present techniques as the energies we treat are within the kondo scale . in particular , a change has been detected in the resistivity of a kondo alloy due to spin - polarized currents @xcite .
furthermore , it is already possible to attach ferromagnetic leads to a carbon nanotube @xcite , and a carbon - nanotube quantum dot has been shown to display kondo physics below an unusually high temperature @xcite . finally , a quantum dot coupled to ferromagnetic electrodes has been proposed as a promising candidate for spin injection devices , and studied experimentally both in the coulomb blockade regime @xcite and in the kondo regime@xcite .
we thank r. aguado , m. bttiker , t. costi , l. glazman , k. kang , k. le hur , g. platero , p. simon , and e.v .
sukhorukov for valuable comments and discussions on this subject .
d.s . and r.l .
were supported by the the eu rtn under contract no .
hprn - ct-2000 - 00144 , nanoscale dynamics , and by the spanish mecd .
m .- s.c .
acknowledges supports from the skore - a program , and from the essc at postech .
url # 1`#1`urlprefix p.m. tedrow and r. meservey , phys .
rev . lett . *
26 * , 192 ( 1971 ) .
m. julliere , phys .
a * 54 * , 225 ( 1975 ) .
for a recent review , see _ semiconductor spintronics and quantum computation _ , edited by d.d .
awschalom , d. loss , and n. samarth ( springer , berlin , 2002 ) .
prinz , science * 282 * , 1660 ( 1998 ) .
et al . _ ,
* 73 * , 363 ( 1998 ) .
et al . _ , physica b ( amsterdam ) * 320 * 396 ( 1998 ) . t. hayashi , m. tanaka , and a. asamitsu , j. appl . phys . *
87 * , 4673 ( 2000 ) .
a. slobodskyy _
et al . _ ,
. lett . * 90 * , 246601 ( 2003 ) . g.d .
mahan , _ many - body physics _ ,
( plenum press , new york , 1983 ) , sec . 1.4 .
et al . _ , in _ mesoscopic electron transport _ , edited by l.l . sohn , l.p .
kouwenhoven , and g. schn ( kluwer series e345 , 1997 ) .
p. hawrylak _
et al . _ ,
b. * 59 * , 2801 ( 1999 ) .
m. keller _
et al . _ ,
b * 64 * , 033302 ( 2001 )
hewson , _ the kondo problem to heavy fermions _ ( cambridge university press , cambridge , 1993 ) .
glazman and m.e .
rakh , pisma zh .
. fiz . * 47 * , 378 ( 1988 ) [ jetp lett . * 47 * , 452 ( 1988 ) ] .
ng and p.a .
lee , phys .
lett . * 61 * , 1768 ( 1988 ) .
l. kouwenhoven and l. glazman , phys .
world * 14 * , 33 ( 2001 ) .
d. goldhaber - gordon _ et al .
_ , nature ( london ) * 391 * , 156 ( 1998 ) ; s.m .
cronenwett , _ et al .
_ , science * 281 * , 540 ( 1998 ) ; j. schmid _ et al .
_ , physica b ( amsterdam ) * 256258 * , 182 ( 1998 ) .
n. sergueev _ et al .
b * 65 * , 165303 ( 2002 ) . p.
zhang _ et al .
lett . * 89 * , 286803 ( 2002 ) .
b.r . buka and s. lipiski , phys .
b * 67 * , 024404 ( 2003 ) .
r. lpez and d. snchez , phys .
lett . * 90 * , 116602 ( 2003 ) .
j. martinek _ et al .
lett . * 91 * , 127203 ( 2003 ) .
b. dong , h.l .
cui , s.y .
liu , and x.l .
lei , j. phys . : condens .
matter * 15 * , 8435 ( 2003 ) .
j. martinek _ et al .
. lett . * 91 * , 247202 ( 2003 ) .
choi , r. lpez and d. snchez , phys .
lett . * 92 * , 056601 ( 2004 ) .
r. l and z - r .
liu , cond - mat/0210350 ( unpublished ) ( 2001 ) .
j. ma , b. dong , x.l .
lei , cond - mat/0212645 ( unpublished ) ( 2001 ) .
j. ma and x.l .
lei , cond - mat/0309520 ( unpublished ) ( 2003 ) .
y. tanaka and n. kawakami , cond - mat/0406490 ( unpublished ) ( 2004 ) .
in the recent preprint j. martinek _ et al .
_ , cond - mat/0406323 ( unpublished ) ( 2004 ) , the gate - controlled splitting investigated in ref .
@xcite is studied in the presence of structured dos , adding details to the overall picture .
et al . _ ,
science * 281 * , 1284 ( 1997 ) .
r. hanson _
et al . _ ,
* 91 * , 196802 ( 2003 ) .
w. rudziski and j. barna , phys .
b * 64 * , 085318 ( 2001 )
et al . _ , cond - mat/0209263 ( unpublished ) .
y. meir and n. s. wingreen , phys .
b * 50 * , r4947 ( 1994 ) .
i.i . mazin , phys .
* 83 * , 1427 ( 1999 ) .
ng , phys .
lett . * 76 * , 487 ( 1996 ) .
p. coleman , phys . rev .
b * 29 * , 3035 ( 1984 ) ; d.m . newns and n. read , adv .
* 36 * , 799 ( 1987 ) .
bickers , rev .
phys . * 59 * , 847 ( 1987 ) ; n.s .
wingreen and y. meir , phys .
b * 49 * , 11040 ( 1990 ) .
wilson , rev .
phys . * 47 * , 773 ( 1975 ) .
krishna - murthy _
et al . _ ,
b * 21 * , 1003 ( 1980 ) ; t.a .
et al . _ ,
j. phys . : condens .
matter * 6 * , 2519 ( 1994 ) .
w. hofstetter , phys .
lett . * 85 * , 1508 ( 2000 ) .
h. kajueter and g. kotliar phys .
lett . * 77 * , 131 ( 1996 ) . for a review ,
m. blanter and m. bttiker , phys . rep . *
336 * , 1 ( 2000 ) .
y. meir and a. golub , phys .
lett . * 88 * , 116802 ( 2002 ) . b. dong and x.l .
lei , j. phys . : condens .
matter * 14 * , 4963 ( 2002 ) .
y. avishai , a. golub , and a.d .
zaikin , phys .
b * 67 * , 041301(r ) ( 2003 ) . t. aono , a. golub , and y. avishai , phys . rev .
b * 68 * , 045312 ( 2003 ) .
r. lpez , r. aguado , and g. platero , phys .
b * 69 * , 235305 ( 2004 ) .
d. snchez and r. lpez , cond - mat/0403485 ( unpublished ) .
hertz and k. aoi , phys .
b * 8 * , 3252 ( 1973 )
. s. zhang _
* 79 * , 3744 ( 1997 ) .
a.m. bratkovsky , appl .
. lett . * 72 * , 2334 ( 1998 ) .
lett . * 74 * , 3273 ( 1997 ) .
e. y. tsymbal , o. mryasov , and p. r. leclair , j. phys . : condens .
matter * 15 * , r109 ( 2003 ) . c. heide _
et al . _ ,
b * 59 * , 4287 ( 1999 ) . c. jayaprakash , h.r .
krishna - murthy and j.w .
wilkins , phys .
* 47 * , 737 ( 1981 ) .
a. georges and y. meir , phys .
lett . * 82 * , 3508 ( 1999 ) .
r. lpez , r. aguado , and g. platero phys .
* 89 * , 136802 ( 2002 ) .
p. simon , r. lpez , and y. oreg , cond - mat/0404540 ( unpublished , 2004 ) . n.j .
craig _ et al_. , science * 304 * 565 ( 2004 ) .
lett . * 90 * , 026602 ( 2003 ) .
g. zarnd _
et al . _ , solid state comm .
* 126 * , 463 ( 2003 ) ; k. le hur and p. simon phys .
b * 67 * , 201308 ( 2003 ) .
r. lpez _ et al .
_ , cond - mat/0402343 ( unpublished ) ( 2004 ) .
d. boese _ et al .
b * 66 * , 125315 ( 2002 ) .
t. taniyama _
et al . _ ,
* 90 * , 016601 ( 2003 ) .
k. tsukagoshi _ et al .
_ , nature ( london ) * 401 * , 572 ( 1999 ) .
j. nygrd _ et al .
_ , nature ( london ) * 408 * , 342 ( 2000 ) .
deshmukh and d.c .
ralph , phys .
* 89 * , 266803 ( 2002 ) a. n. pasupathy , r. c. bialczak , j. martinek , j. e. grose , l. a. k. doney , p. l. mceuen , and d. c. ralph , science * 306 * , 86 ( 2004 ) ; j. nygard , w. koehl , n. mason , l. dicarlo , and c. marcus , cond - mat/0410467 ( unpublished ) . | we investigate the spin - dependent transport properties of quantum - dot based structures where kondo correlations dominate the electronic dynamics .
the coupling to ferromagnetic leads with parallel magnetizations is known to give rise to nontrivial effects in the local density of states of a single quantum dot .
we show that this influence strongly depends on whether charge fluctuations are present or absent in the dot .
this result is confirmed with numerical renormalization group calculations and perturbation theory in the on - site interaction . in the fermi - liquid fixed point , we determine the correlations of the electric current at zero temperature ( shot noise ) and demonstrate that the fano factor is suppressed below the poissonian limit for the symmetric point of the anderson hamiltonian even for nonzero lead magnetizations .
we discuss possible avenues of future research in this field : coupling to the low energy excitations of the ferromagnets ( magnons ) , extension to double quantum dot systems with interdot antiferromagnetic interaction and effect of spin - polarized currents on higher symmetry kondo states such as su(4 ) . |
* ( fop ) * optimal and sub - optimal algorithms for the special case of solving * ospmp * for simo mac / miso bc have been proposed in @xcite .
the optimal algorithm is much more complex than algorithm o as it involves several inner and outer iterations .
the difference between algorithm o and the sub - optimal algorithm are as follows .
1 ) the sub - optimal algorithm works for simo mac , avoiding the calculation of beamforming matrices , while algorithm o works for mimo cases ; 2 ) to find the encoding / decoding order , after obtain the optimal solution of * spmp * with fixed @xmath82 in step 1 , the sub - optimal algorithm in @xcite needs to solve an equation to obtain a weight vector @xmath83_{,l=1, ... ,l}$ ] * * * * @xmath84s@xmath82 is directly given by the polite water - filling levels .
same as the algorithm in @xcite , it is possible for algorithm o to cycle through a finite number of orders . in this case
, we can simply choose the best one from the finite number of orders .
for mac and bc , it is observed that the reason of non - convergence is that the corresponding boundary point can not be achieved without time - sharing .
algorithms to solve * ofop * and * ospmp * for mimo bc have been proposed in @xcite where the problems are converted to the weighted sum - rate maximization problem .
the algorithms need to repeatedly solve a weighted sum - rate maximization problem for weight vectors searched by the ellipsoid algorithm .
the weighted sum - rate maximization problem is solved by the steepest ascent algorithm . for * ospmp * , additional search for the total power is also needed .
in contrast , algorithm o solves the problems directly in one sequence of iterations , replaces the steepest ascent algorithm with sinr based algorithms or polite water - filling , only searches weight vector suggested by the polite water - filling level , resulting in much lower complexity .
simulations are used to verify the performance of the proposed algorithms.dpc and sic are employed for interference cancellation .
block fading channel is assumed and the channel matrices are independently generated by @xmath85 , where @xmath86 has zero - mean i.i.d .
gaussian entries with unit variance ; and @xmath87 is a positive number and the value is set as @xmath88 except for fig .
[ fig : fig3 ] and fig .
[ fig : fig4 ] . in , each simulation is averaged over 100 random channel realizations____the best solution among many solutions obtained by running the algorithm for many times with different initial points and with the encoding / decoding order obtained by algorithm o. for the plots with iteration numbers , we show rates or power after @xmath89 iterations / rounds , where the last 0.5 iteration / round is the forward link update .
we demonstrate the superior convergence speed of algorithm pr and pr1 .
sum power versus iteration number is shown in fig . [
fig : fig3 ] for the itree network of fig .
[ fig : itree ] and in fig .
[ fig : fig4 ] for a 3-user interference channel .
each node has four antennas . in the upper sub - plot of fig .
[ fig : fig3 ] , we consider the moderate interference case , where we set @xmath90 . in the lower sub - plot of fig .
[ fig : fig3 ] , we consider strong interference case , where we set @xmath91 for the interfering links , and @xmath92 for other @xmath93 s .
it is not surprising that algorithms pr and pr1 have faster convergence speed because polite water - filling exploits the structure of the problem . in the upper sub - plot of fig .
[ fig : fig4 ] , we set @xmath94 . in the lower sub - plot , we consider a strong interference channel , i.e. , @xmath95 , and @xmath96 .
again , algorithm pr1 converges faster than algorithm b. since the problem is non - convex , the algorithms may converge to different stationary points . but
it can be observed that all the stationary points achieve similar performance . ` ' ` ' set as @xmath97/15 $ ] , where @xmath98 is the total required rate , with still achieves near - optimal performance , while with the meb order performs worse than that .
not showing is thatthe target rates are set as @xmath99/16 $ ] .
the encoding / decoding order is partially fixed and is the same as that in example [ exa : pseudobm ] of section [ sub : orderoptimization ] .
for the pseudo mac formed by link 2 and link 3 and the pseudo bc formed by link 4 and link 5 , the fixed order that @xmath23 is decoded after @xmath22 and @xmath100 is decoded after @xmath101 , and its improved order obtained by algorithm o are applied .
improved order obtained by algorithm o we illustrate the convergence behavior of the distributed optimization with local csi .
[ fig : fig6 ] plots the total transmit power and the minimum rate of the users achieved by algorithm prd versus the number of training rounds for a 3-user interference channel with .
whenwhen the general mimo one - hop interference networks named b - mac networks with gaussian input and any valid coupling matrices are considered .
we design algorithms for maximizing the minimum of weighted rates under sum power constraints and for minimizing sum power under rate constraints .
they can be used in in two kinds of algorithms are designed .
the first kind takes advantage of existing sinr optimization algorithms by finding simple and optimal mappings between the achievable rate region and the sinr region .
the mappings can be used for many other optimization problems .
the second kind takes advantage of the .
both centralized and distributed algorithms are designed .
a. liu , y. liu , h. xiang , and w. luo , `` duality , polite water - filling , and optimization for mimo b - mac interference networks and itree networks , '' _ submitted to ieee trans .
info . theory _ ,
apr . 2010 .
[ online ] .
available : http://arxiv4.library.cornell.edu/abs/1004.2484 m. maddah - ali , a. motahari , and a. khandani , `` communication over mimo x channels : interference alignment , decomposition , and performance analysis , '' _ ieee transactions on information theory _
, vol . 54 , no . 8 , pp .
34573470 , aug .
2008 .
chang , l. tassiulas , and f. rashid - farrokhi , `` joint transmitter receiver diversity for efficient space division multiaccess , '' _ ieee transactions on wireless communications _
, vol . 1 , no . 1 ,
1627 , jan 2002 .
m. schubert and h. boche , `` solution of the multiuser downlink beamforming problem with individual sinr constraints , '' _ ieee transactions on vehicular technology _ , vol .
53 , no . 1 ,
1828 , jan . 2004 .
f. fung , w. yu , and t. j. lim , `` precoding for the multiantenna downlink : multiuser snr gap and optimal user ordering , '' _ ieee transactions on communications _ , vol .
55 , no . 1 ,
188 197 , jan .
2007 .
n. jindal , w. rhee , s. vishwanath , s. jafar , and a. goldsmith , `` sum power iterative water - filling for multi - antenna gaussian broadcast channels , '' _ ieee trans .
inform . theory _
51 , no . 4 , pp .
15701580 , april 2005 .
m. varanasi and t. guess , `` optimum decision feedback multiuser equalization with successive decoding achieves the total capacity of the gaussian multiple - access channel , '' in _ proc .
thirty - first asilomar conference on signals , systems and computers _ , vol . 2 , 1997 ,
. 14051409 .
a. liu , a. sabharwal , y. liu , h. xiang , and w. luo , `` distributed mimo network optimization based on local message passing and duality , '' _ in proc .
47th annu .
allerton conf .
monticello , illinois , usa _ , 2009 . | general mimo interference networks , named b - mac networks , which is a combination of multiple interfering broadcast channels ( bc ) and multiaccess channels ( mac ) . two related optimization problems , maximizing the minimum of weighted rates under a sum - power constraint and minimizing the sum - power under rate constraints , are considered .
the first approach takes advantage of existing efficient algorithms for sinr problems by building a bridge between rate and sinr through the design of optimal mappings between them so that the problems can be converted to sinr constraint problems .
the approach can be applied to other optimization problems as well .
the second approach employs polite water - filling , which is the optimal network version of water - filling that we recently found .
it replaces almost all generic optimization algorithms currently used for networks and reduces the complexity while demonstrating superior performance even in non - convex cases .
both centralized and distributed algorithms are designed and the performance is analyzed in addition to numeric examples . , mimo , interference network , [ [ section ] ] he optimization under rate constraints for general multiple - input multiple - output ( mimo ) interference networks , where each transmitter may send data to multiple receivers and each receiver may collect data from multiple transmitters .
consequently , the network is a combination of multiple interfering broadcast channels ( bc ) and multiaccess channels ( mac).@xcite@xcite as special cases we assume gaussian input and that each interference is either completely cancelled or treated as noise .
a wide range of interference cancellation is allowed , from no cancellation to any cancellation specified by a valid binary _ coupling matrix _ of the data links .
for example , simple linear receivers , dirty paper coding ( dpc ) @xcite at transmitters , and/or successive interference cancellation ( sic ) at receivers may be employed .
two optimization problems are considered for .
* fop * which maximizes the minimum of scaled rates of all links , where the scale factors are the inverse of the target rates . a****solving the ( * spmp * ) under the rate constraints to the design of the in some cases such as cooperative cellular networks ,
it is possible to obtain global csi if the base stations are allowed to exchange csi , making the relevant . in ad hoc or large networks
, we have to design distributed optimization algorithms with local csi . *
* * * @xcitewhich means that if a set of sinrs is achievable in the forward links , then the same sinrs can be achieved in the reverse links when the set of transmit and receive beamforming vectors are fixed .
thus , optimizing the transmit vectors of the forward links is equivalent to the much simpler problem of optimizing the receive vectors in the reverse links . * * considering interference cancellation and encoding / decoding order , the * fop * and * spmp * for mimo bc / mac have been completely solved in @xcite by converting them to convex weighted sum - rate maximization problems for mac .
steepest ascent algorithm for the weighted sum - rate maximization a high complexity algorithm that can find the optimal encoding / decoding order for miso bc / simo mac is proposed in @xcite that needs several inner and outer iterations .
a heuristic low - complexity algorithm in @xcite finds the near - optimal encoding / decoding order for * spmp * by observing that the optimal solution of * spmp * must be the optimal solution of some weighted sum - rate maximization problem , in which the weight vector can be found and used to determine the decoding order .
in summary , the * fop * and * spmp * for mimo b - mac networks have been open problems .
the contribution of the paper is as follows .
* _ rate - sinr _ _ conversion _ : one of the difficulties of solving the problems is the joint optimization of beamforming matrices of all links . one approach is to decompose a link to multiple single - input single - output ( siso ) streams and optimize the beamforming vectors through sinr duality , if a bridge between rate and sinr can be built to determine the optimal number of streams and rate / power allocation among the streams . in section [ sec : rate - sinr ] , we show that any pareto rate point of an achievable rate region can be mapped to a pareto sinr point of the achievable sinr region through two optimal and simple mappings that produce equal rate and equal power streams respectively .
the significance of this result is that it offers a method to convert the * _ sinr based algorithms _ : using the above result , we take advantage of existing algorithms for sinr problems to solve * * * * * another approach is to directly solve for the beamforming matrices . for the convex problem of mimo mac , steepest ascent algorithm
is used except for the special case of sum - rate optimal points , where iterative water - filling can be employed @xcite .
the b - mac network problems are non - convex in general and thus , better algorithms , like water - filling , than the steepest ascent algorithm is highly desirable .
however , directly applying traditional water - filling is far from optimal @xcite . in @xcite
, we recently found the long sought optimal network version of water - filling , polite water - filling , which is the optimal input structure of any pareto rate point , not only the sum - rate optimal point , of the achievable region of a mimo b - mac network .
this network version of water - filling is polite because it optimally balances between reducing interference to others and maximizing a link s own rate .
the superiority of the polite water - filling is demonstrated for weighted sum - rate maximization in @xcite and the superiority is because it is hard not to obtain good results when the optimal input structure is imposed to the solution at each iteration . in section [ sub : itree ] , using polite water - filling , we design an algorithm to monotonically improve the output of the sinr based algorithms for itree networks defined later , if the output does not satisfy the kkt condition . furthermore , in section [ sub : pr - pr1 ] , purely polite water - filling based algorithms are designed that have faster convergence speed . * _ distributed algorithm _ : in a network , it is highly desirable to use distributed algorithms .
the polite water - filling based algorithm is well suited for distributed implementation , which is shown in section [ sub : distributed - implementation ] , where each node only needs to estimate / exchange the local csi but the performance of each iteration is the same as that of the centralized algorithm .
* _ _ optimization of encoding and decoding order__s : another difficulty is to find the optimal encoding / decoding order when interference cancellation techniques like dpc / sic are employed .
again , polite water - filling proves useful in section [ sub : orderoptimization ] because the water - filling levels of the links can be used to identify the optimal encoding / decoding order for bc / mac and pseudo - bc / mac defined later .
the rest of the paper is organized as follows .
section [ sec : system model ] defines the achievable rate region and formulates the problems .
section [ sec : preliminary ] summarizes the preliminaries on sinr duality and polite water - filling .
section [ sec : algorithms ] presents the efficient centralized and distributed algorithms .
the performance of the algorithms is verified by simulation in section [ sec : simulation - results ] .
the conclusion is given in section [ sec : conclusion ] .
[ [ section-1 ] ] let @xmath0 and @xmath1 denote the virtual transmitter and receiver of link @xmath2 equipped with @xmath3 transmit antennas and @xmath4 receive antennas respectively .
the received signal at @xmath1 is @xmath5 where @xmath6 is the transmit signal of link @xmath7 and is assumed to be circularly symmetric complex gaussian ; @xmath8 is the channel matrix between @xmath9 and @xmath1 ; and @xmath10 is a circularly symmetric complex gaussian noise vector with zero mean and identity covariance matrix . to handle a wide range of interference cancellation possibilities ,
we define a coupling matrix @xmath11 as a function of the interference cancellation scheme @xcite .
it specifies whether interference is completely cancelled or treated as noise : if @xmath12 , after interference cancellation , still causes interference to @xmath13 , @xmath14 and otherwise , @xmath15 .
for example , if the virtual transmitters ( receivers ) of several links are associated with the same physical transmitter ( receiver ) , interference cancellation techniques such as dirty paper coding ( successive decoding and cancellation ) can be applied at this physical transmitter ( receiver ) to improve the performance .
the coupling matrices valid for the results of this paper are those for which there exists a transmission and receiving scheme such that each signal is decoded and possibly cancelled by no more than one receiver .
possible extension to the han - kobayashi scheme , where a common message is decoded by more than one receiver , is discussed in @xcite .
we give some examples of valid coupling matrices . for a bc ( mac ) employing dpc ( sic ) where the @xmath16 link is the @xmath16 one to be encoded ( decoded ) , the coupling matrix is given by @xmath17 and @xmath18 . in fig .
[ fig : sysfig1 ] , we give an example of a b - mac network employing dpc and sic . when no data is transmitted over link 4 and 5 , the following @xmath19 are valid coupling matrices for link @xmath20 under the corresponding encoding and decoding orders : _
@xmath21 is encoded after @xmath22 and @xmath22 is decoded after @xmath23 ; _
b_. @xmath22 is encoded after @xmath21 and @xmath22 is decoded after @xmath23 ; _
c_. @xmath21 is encoded after @xmath22 and @xmath23 is decoded after @xmath22 ; _ d_. there is no interference cancellation.@xmath24 , & \:\mathbf{\phi}^{b}=\left[\begin{array}{ccc } 0 & 1 & 1\\ 0 & 0 & 0\\ 1 & 1 & 0\end{array}\right],\\ \mathbf{\phi}^{c}=\left[\begin{array}{ccc } 0 & 0 & 1\\ 1 & 0 & 1\\ 1 & 0 & 0\end{array}\right ] , & \:\mathbf{\phi}^{d}=\left[\begin{array}{ccc } 0 & 1 & 1\\ 1 & 0 & 1\\ 1 & 1 & 0\end{array}\right].\end{aligned}\ ] ] note that when dpc and sic are combined , an interference may not be fully cancelled under a specific encoding and decoding order .
such case can not be described by the coupling matrix of 0 s and 1 s defined above .
but a valid coupling matrix can serve for an upper or lower bound .
see more discussion in @xcite . if not explicitly stated otherwise , achievable regions in this paper refer to the following .
note that @xmath25 by definition .
the interference - plus - noise covariance matrix of the @xmath26 link is @xmath27 where @xmath28 is the covariance matrix of @xmath12 .
we denote all the covariance matrices as @xmath29 then the achievable mutual information ( rate ) of link @xmath2 is given by a function of @xmath30 and @xmath31@xmath32 the _ achievable rate region _ with _ _ a fixed coupling matrix @xmath31 _ _ and sum power constraint @xmath33 is defined as @xmath34 a bigger achievable rate region can be defined by the convex closure of @xmath35 , where @xmath36 is a set of valid coupling matrices .
for example , if dpc and/or sic are employed , @xmath36 can be a set of valid coupling matrices corresponding to various valid encoding and/or decoding orders .
the algorithms rely on the duality between the forward and reverse links of a b - mac network .
the reverse links are obtained by reversing the transmission direction and replacing the channel matrices by their conjugate transposes . .
we use the notation @xmath37 to denote the corresponding terms in the reverse links .
for example , in the reverse links of the b - mac network in fig .
[ fig : sysfig1 ] , @xmath38 ( @xmath39 ) becomes the receiver ( transmitter ) , and @xmath40 is decoded after @xmath41 and @xmath42 is encoded after @xmath43 , if in the forward links , @xmath21 is encoded after @xmath22 and @xmath22 is decoded after @xmath23 .
the interference - plus - noise covariance matrix of reverse link @xmath2 is @xmath44 and the rate of reverse link @xmath2 is given by @xmath45 * ( fop ) * coupling matrix @xmath31 * * * * for the special case of dpc and sic , the optimal _ _ coupling matrix _ _
@xmath31 , or equivalently , the optimal encoding and/or decoding order of * * is partially solved in section * * * * although we focus on the sum power and white noise in this paper for simplicity , the results can be directly applied to a much larger class of problems with a single linear constraint @xmath46 in * fop * ( or objective function @xmath47 in * spmp * ) and/or colored noise with covariance @xmath48=\mathbf{w}_{l}$ ] , which includes the weighted sum power minimization problem in @xcite as a special case . only variable changes @xmath49 and
@xmath50 are needed , where @xmath51 and @xmath52 are positive definite for meaningful cases or @xmath53 will result in infinite power and rate with probability one . ] .
the single linear constraint appears in lagrange functions for problems with multiple linear constraints @xcite , and thus , the results in this paper serve as the basis to solve them @xcite . special cases of multiple linear constraints include individual power constraints , per - antenna power constraints , interference constraints in cognitive radios , etc .. [ [ section-2 ] ] the algorithms are based on sinr duality , e.g. , @xcite , rate duality , and polite water - filling developed earlier @xcite . they are reviewed below .
[ [ section-3 ] ] @xmath54 and @xmath55 @xmath56 . [
[ section-4 ] ] .
the reverse link input covariance matrices are obtained by the following transformation .
let @xmath57 be a decomposition of @xmath30 .
compute the mmse - sic receive vectors @xmath58 from ( [ eq : mmserev1 g ] ) and the reverse transmit powers @xmath59 from ( [ eq : qpower ] ) . the _ covariance transformation _ from @xmath30 to @xmath60 is @xmath61 we give the rate duality under a linear constraint @xmath46 and/or colored noise with covariance @xmath48=\mathbf{w}_{l}$ ] . the covariance transformation for this case
is also calculated from the mmse receive beams and power allocation that makes sinrs of the forward and reverse links equalis that the identity noise covariance in @xmath62 is replaced by @xmath53 and the all - one vector @xmath63 in ( [ eq : qpower ] ) is replaced by the vector @xmath64_{m=1, ...
,m_{l},l=1, ... ,l}$ ]
. for convenience , let @xmath65,\sum_{l=1}^{l}\textrm{tr}\left(\mathbf{\sigma}_{l}\hat{\mathbf{w}}_{l}\right)\leq p_{t},\left[\mathbf{w}_{l}\right]\right),\label{eq : net - color - linear - constraint}\ ] ] denote a network where the channel matrices are @xmath66 $ ] ; the input covariance matrices must satisfy the linear constraint @xmath46 ; and the covariance matrix of the noise at the receiver of link @xmath2 is @xmath53 .
[ thm : linear - color - dual]the dual of the network ( [ eq : net - color - linear - constraint ] ) is@xmath67,\sum_{l=1}^{l}\textrm{tr}\left(\hat{\mathbf{\sigma}}_{l}\mathbf{w}_{l}\right)\leq p_{t},\left[\hat{\mathbf{w}}_{l}\right]\right)\label{eq : net - forward - color - dual}\ ] ] in the sense that 1 ) they have the same achievable rate region ; 2 ) if @xmath30 achieves certain rates and satisfies the linear constraint in network ( [ eq : net - color - linear - constraint ] ) , its covariance transformation @xmath60 achieves better rates in network ( [ eq : net - forward - color - dual ] ) under the linear constraint @xmath68 . _
_ its covariance transformation @xmath60 as . @xmath69s@xmath70@xmath71 the input covariance matrix @xmath72 is said to @xmath73 satisfies the structure of water - filling over @xmath74 , i.e.,____@xmath72 s [ thm : wfst]the input covariance matrices @xmath30 of a pareto rate point of the achievable region and its covariance transformation @xmath60 the polite water - filling structure .
the following theorem proved in states that @xmath72 having the polite water - filling structure suffices for @xmath75 to have the polite water - filling structure even at a non - pareto rate point .
[ thm : fequrgwf]if one input covariance matrix @xmath72 has the polite water - filling structure while other @xmath76 @xmath77 are fixed , so does its covariance transformation @xmath75 , i.e. , @xmath78 satisfies the structure of water - filling over the reverse equivalent channel @xmath79 .
further more , @xmath75 can be expressed as @xmath80 where @xmath81 is the polite water - filling level in ( [ eq : wffar ] ) . |
the ` standard model ' of extragalactic classical double radio sources by scheuer ( 1974 ) explains these sources as twin jets emerging from the active galactic nucleus ( agn ) and impinging on the surrounding gas .
the compact , high surface brightness regions or hot spots at the end of the jets are interpreted as the sites of the interaction between the jets and the environment . after passing through the hot spot region
, the jet material inflates the radio lobes or cocoon which is observed as diffuse emission in between the hot spots and the source core ( e.g. muxlow & garrington 1991 and references therein ) .
this picture forms the basis of virtually all more recent attempts at modeling the dynamics and radio emission properties of powerful radio galaxies and radio - loud quasars ( begelman & cioffi 1989 , falle 1991 , daly 1994 , nath 1995 , kaiser & alexander 1997 , chyy 1997 , kaiser et al .
1997 , blundell et al .
1999 ) .
these models for the evolution of individual classical doubles or frii - type objects ( fanaroff & riley 1974 ) can be used to study the cosmological evolution of the population as a whole .
one of the more important trends is the apparent decrease of the mean linear size of the radio structures with increasing redshift ( see neeser et al .
1995 and references therein ) .
two recent attempts in fully explaining the linear size redshift anti - correlation were presented by blundell et al .
( 1999 ) and kaiser & alexander ( 1999a ) .
blundell et al .
argue that a specific form of pre - aging of the relativistic particle population in the hot spots in connection with the lower flux limit of complete observed samples causes this anti - correlation ( the so - called youth - redshift degeneracy , blundell et al . 1999 , blundell & rawlings 1999 ) .
alternatively , kaiser & alexander propose that the apparent smaller sizes of radio sources at high redshift could be caused by a denser environment of these objects at high redshift . clearly , the evolution or non - evolution of the radio source environments is of great interest to decide this and other important cosmological questions .
various methods have been employed in determining the properties of the radio source environments .
x - ray observations of the hot gas around radio sources in clusters yield direct estimates of the gas density ( crawford et al . 1999 , hardcastle & worrall 1999 ) .
however , such studies are at present confined to objects at low redshifts .
furthermore , the agn ( e.g. sambruna et al . 1999 ) and the large - scale radio structure contribute to the x - ray emission ( brunetti et al . 1999 , kaiser & alexander 1999b , sect .
[ sec : disc ] ) .
the properties of the density distribution inferred from x - ray observations may therefore be considerably influenced by the presence of the radio source itself .
faraday rotation of the polarisation angle of the synchrotron emission and the related depolarisation can be used to determine the gas density of the material surrounding the radio lobes ( e.g. garrington et al .
1988 , laing 1988 ) .
unfortunately , this method does not provide a direct measure of the gas density as the rotation measure also depends on the strength of the magnetic field in the source environment .
usually it is not possible to break this degeneracy because the strength of the magnetic field is only poorly known .
constraints on the density of the radio source environments also come from optical or infrared galaxy counts around the host galaxies ( e.g. hill & lilly 1991 ) .
it is not straightforward to decide whether galaxies in the field of the radio source host are associated or chance background objects . resolving the ambiguity
would ideally require the spectroscopic measurement of the redshifts of all objects in question .
this is very time consuming . in any case
, the method provides only indirect constraints on the gas density around the radio source as this has to be inferred from comparison with low redshift clusters or groups of similar richness .
the ages of powerful radio sources can in principle be determined from their radio spectrum ( e.g. alexander & leahy 1987 ) .
the various energy losses of the relativistic particles depend on time and so the shape of the radio spectrum contains an encoded history of the source . in practice the time - dependence of the energy losses complicates the estimation of the spectral age because different parts of the source have different ages .
even when radio maps at various frequencies which fully resolve the radio lobes are used , it is difficult to disentangle the effects of the various loss mechanisms and possible bulk backflow of the cocoon material along the jet ( e.g. rudnick et al .
1994 ) .
the model developed in this paper aims at tracing the individual evolution of parts of the cocoon and thereby providing more accurate estimates for the source age . at the same time
the model also constrains the density in the source environment and other parameters like the energy transport rate of the jets .
it is solely based on radio observations which are available for a large number of objects , even at high redshift .
this model may therefore provide an important step in determining the cosmological evolution of the frii radio source population . in sect .
[ sec : spec ] i show that diffusion of relativistic particles in the cocoons of frii - type objects should not significantly change the energy distribution of particles . in sect . [
sec : mod ] the dynamical model of kaiser & alexander ( 1997 , hereafter ka ) and its extension by kaiser et al .
( 1997 , hereafter kda ) to include synchrotron emission are briefly summarised . a 3-dimensional model of the synchrotron emissivity of the cocoon based on this analysis
is constructed in sects .
[ sec : spat ] and [ sec : back ] .
methods for comparing the model predictions with observations are developed in sect .
[ sec : compa ] .
the degeneracy of model parameters resulting from the comparison method is also discussed here .
the model is then applied to three frii - type radio sources , cygnus a , 3c 219 and 3c 215 , in sect . [
sec : appl ] .
the results are discussed in sect .
[ sec : disc ] .
the main conclusions are summarised in sect .
[ sec : conc ] . throughout this paper
i use @xmath1 km s@xmath2 mpc@xmath2 and @xmath3 .
the age of extragalactic radio sources can be estimated by the use of spectral aging arguments .
this method relies on the determination of the break in the radio spectrum caused by the time - dependent energy losses of the relativistic electrons within the cocoon ( e.g. alexander & leahy 1987 ) . at frequencies higher than the break frequency
, @xmath4 , the spectrum significantly steepens due to the radiative energy losses of the electrons .
the spectral age of a electron population , @xmath5 , with observed break frequency @xmath6 within a magnetic field of strength @xmath7 which does not vary in time is proportional to @xmath8 . despite the implication of source dynamics that the magnetic field does vary in time ,
the constant field relation is usually used in determining the spectral age of a given source . in general
the strength of the magnetic field will decrease while the source expands and so @xmath9 will be an overestimate of the true source age , @xmath10 . however , because of variations of the magnetic field in the cocoon and the backflow of gas in this region , @xmath5 is usually found to be lower than the dynamical ages inferred from the advance speed of the cocoon by various methods ( alexander & leahy 1987 ) . the model presented in the following can be viewed as an extension of the spectral aging formalism . within the cocoon it traces the evolution of the local magnetic field in time .
this allows the accurate determination of the energy distribution function of the relativistic particle population and thus the emitted spectrum at a given location along the cocoon .
it is therefore not surprising that the model predicts an older age for cygnus a ( see sect .
[ sec : compa ] ) than the classical spectral index analysis of carilli et al .
the combination of a dynamical model with the accurate treatment of the local evolution within the cocoon of the radio emission properties implies that the spectral and dynamical ages are identical . in the context of spectral aging methods
other processes changing the energy distribution of the relativistic electrons and thus invalidating the age estimates have been put forward .
these are summarised and discussed by blundell & rawlings ( 2000 ) .
they show that most of these processes are rather inefficient and will not strongly influence spectral aging methods or the model discussed here .
however , blundell & rawlings ( 2000 ) claim that the anomalous diffusion mechanism of rechester & rosenbluth ( 1978 ) can lead to very fast diffusion of relativistic particles through the tangled magnetic field in the cocoon .
so much so that the radio spectra observed at any point along the cocoon essentially arise from the electrons accelerated by the jet shock at the hot spot within the last @xmath11 years of the observation .
this would imply that no information on the source age can be derived from the spatial properties of the observed emission .
support for a universal particle energy spectrum may come from the interpretation of observations of cygnus a by katz - stone et al .
( 1993 ) , rudnick et al .
( 1994 ) and katz - stone & rudnick ( 1994 ) .
they find that a single , non - power law spectrum shifted in frequency by the local strength of the magnetic field is emitted by all parts of the lobes of cygnus a. note however , that the spectral shape they find may be caused by the free - free absorption in our own galaxy of the emission of cygnus a at low frequencies ( carilli et al .
1991 . in this case , the universal spectrum is simply explained by shifting an aging energy spectrum of relativistic particles which is not a single power law .
alternatively , changes in the strength of the magnetic field along a given line of sight may also introduce additional curvature in the observed spectrum ( rudnick et al . 1994 ) .
the mentioned interpretation of the radio observations of cygnus a in itself is therefore no proof of efficient diffusion acting in the lobes of radio galaxies . in the appendix
i show that anomalous diffusion is probably much less effective in the cocoon plasma as previously thought .
in general we do not observe any signature for diffusion losses of the cocoons of radio sources and so diffusion will not alter the distribution of relativistic particles within the cocoon .
this allows us to use the spatial distribution of the synchrotron radio emission of frii sources to infer their age .
the model developed in the following can be viewed as an extension to the classical spectral aging methods in that it takes into account the evolution of the magnetic field in the lobe .
in this sect .
i briefly summarise the dynamical and radiative model that form the basis for the extended treatment presented in this paper . following this
the prescription for the spatial distribution of the synchrotron emission within the radio lobes is developed .
the large scale structure of radio galaxies and radio - loud quasars of type frii is formed by twin jets emanating from the central agn buried inside the nucleus of the host galaxy .
the jets propagate into opposite directions from the core of the source .
they end in strong jet shocks and , after passing through these shock , the jet material inflates the cocoon surrounding the jets .
the cocoon is overpressured with respect to the ambient medium and therefore drives strong bow shocks into this material .
falle ( 1991 ) and ka showed that the expansion of the bow shock and the cocoon should be self - similar which is supported by observations ( e.g. leahy & williams 1984 , leahy et al .
1989 , black 1992 ) . in these models
the density distribution of the material the radio source is expanding into is approximated by a power law , @xmath12 , where @xmath13 is the radial distance from the source centre and @xmath14 is the core radius of the density distribution .
x - ray observations of groups and clusters of galaxies show that the density of the hot gas in these structures is often distributed according to ( e.g. sarazin 1988 ) @xmath15 ^{3\beta ' /2}}. \label{king}\ ] ] outside a few core radii , @xmath16 , the power law assumed above with @xmath17 provides an adequate fit to eq .
( [ beta ] ) . even for smaller distances
@xmath13 good power law approximations can be found by adjusting @xmath18 and @xmath19 ( e.g. kaiser & alexander 1999a ) . in the model of ka
it is also assumed that the rate at which energy is transported along each jet , @xmath20 , is constant and that the jets are in pressure equilibrium with their own cocoon . the very high sound speed within the cocoon results in a practically uniform pressure within this region apart from the tip of the cocoon .
the pressure in this ` hot spot ' region , named for the very strong radio emission originating in the shock at the end of the jets , is somewhat higher as the cocoon material injected by the jets at these points is not yet in pressure equilibrium with the rest of the cocoon . in the following i
will concentrate on only one jet and the half of the cocoon it is contained in . from ka
i take the expressions for the evolution of the uniform cocoon pressure , @xmath21 and that of the physical length of the jet , @xmath22 here , @xmath23 is the ratio of specific heats of the gas surrounding the radio source , @xmath24 is the age of the jet flow and @xmath25 is a dimensionless constant .
the ratio , @xmath26 , of the pressure in the hot spot region and @xmath27 is constant in the model of ka .
kda extend the model of ka to include the synchrotron emission of the cocoon .
this is done by splitting up the cocoon into small volume elements , @xmath28 , the evolution of which is then followed individually . by assuming that these elements are injected by the jet into the hot spot region at time @xmath29 during a short time interval @xmath30 and then become part of the cocoon , kda find @xmath31 where @xmath32 is the ratio of specific heats for the cocoon material and @xmath33 $ ] . because the expansion of the cocoon is self - similar we can set for the volume of the total volume of the cocoon , following the notation of ka , @xmath34 , where @xmath35 is a dimensionless constant and depends on the geometric shape of the cocoon . in order to ensure self - consistency the integration of eq .
( [ element ] ) over the injection time from @xmath36 to @xmath37 must be equal to the total cocoon volume @xmath38 .
substituting eq .
( [ length ] ) for @xmath39 then yields @xmath40 ^{1/\left ( 5-\beta \right ) } .\ ] ] note , that this expression for @xmath25 is different from the one given by ka . in their analysis
they used the expression for conservation of energy for the entire cocoon @xmath41 = q_{\rm{o}}dt,\ ] ] where @xmath42 is the volume of the hot spot region with pressure @xmath43 .
ka then used the simplifying assumption that the cocoon has a cylindrical geometry , the expansion of which is governed along the jet axis by @xmath43 while its growth perpendicular to this direction is driven by @xmath44 .
this implies @xmath45 , where @xmath46 is the ratio of the length of one jet and the full width of the associated lobe halfway down the jet .
kaiser & alexander ( 1999b ) subsequently derived empirical fitting formulae for @xmath47 as functions of @xmath0 and @xmath46 from an analysis of the flow of shocked gas between the bow shock and the cocoon .
their results showed that the original approximation tends to overestimate the value of @xmath47 . in the following
i use a generalised empirical fitting formula based on their result and further calculations with additional values of @xmath0 , @xmath48 in order to satisfy eqs .
( [ constant ] ) and ( [ energy ] ) i now generalise the approach of ka by setting @xmath49 . because of the self - similar expansion of the cocoon @xmath50 is a constant and will depend on the ratio @xmath26 .
the strict separation of the cocoon material into the hot spot region and the rest of the cocoon with two distinctly different values of the respective pressure within these regions is of course artificial . in the cocoons of real frii
objects the transition of jet material from hot spot to cocoon will be accomplished in a continuous hydrodynamical flow along a pressure gradient much smoother than the sudden change from @xmath43 to @xmath27 described here .
however , a detailed model of this flow is beyond the scope of this paper and for simplicity i will use the assumption of a strict spatial separation in the following . as the two cocoon regions are in physical contact with each other
, @xmath50 may become negative as the expansion of one region may influence the evolution of the other .
substituting eqs .
( [ pressure ] ) and ( [ length ] ) into eq .
( [ energy ] ) gives @xmath51 the cocoon volume elements @xmath28 are filled with a magnetised plasma and a population of relativistic electrons accelerated at the shock terminating the jet flow at the hot spot .
they therefore emit synchrotron radio radiation . in optically thin conditions
the monochromatic luminosity due to this process can be calculated by folding the emissivity of single electrons with their energy distribution ( e.g. shu 1991 ) .
following kda i assume that the initial energy distribution of the relativistic electrons as they leave the acceleration region of the hot spot follows a power law with exponent @xmath52 between @xmath53 and @xmath54 .
the electrons are subject to energy losses due to the adiabatic expansion of @xmath28 , the emission of synchrotron radiation and inverse compton scattering of the cmb . for a given volume element at time @xmath24 that was injected into the cocoon at time @xmath29 these losses result in an energy distribution ( see kda ) @xmath55 with @xmath56 + \frac{u_{\rm{cmb}}\left ( z\right ) } { c_{7}}\left [ 1-\left ( \frac{t_{\rm{i}}}{t}\right ) ^{c_{7}}\right ] \right\ } , \end{aligned}\ ] ] where @xmath57 is the energy density of the cmb radiation field at the source redshift @xmath58 and @xmath59 is the rest mass of an electron . here
i have used @xmath60 and @xmath61 .
the normalisation of the energy spectrum , @xmath62 , is given by integrating the initial power law distribution over the entire energy range @xmath63 ^{-1},\end{aligned}\ ] ] or , for @xmath64 , @xmath65 ^{-1 } , \label{normal}\ ] ] where @xmath66 is the total energy density of the relativistic particle distribution . at time
@xmath29 i set @xmath67 and for simplicity @xmath68 $ ] . in the following
i assume that the minimum energy condition ( e.g. miley 1980 ) is initially fulfilled in each volume element @xmath69 and therefore @xmath70 . from eqs .
( [ pressure ] ) and ( [ length ] ) it follows that @xmath71 and with the assumption of completely tangled magnetic fields i find @xmath72 ( see also ka ) . with this
the set of equations describing the radio synchrotron emissivity , @xmath73 , of a given volume element injected into the cocoon at time @xmath74 only depends on the present value of the pressure in the cocoon , @xmath75 , and the age of the radio source , @xmath10 .
so far the cocoon volume elements @xmath28 were only characterised by their injection time into the cocoon , @xmath76 . from the analysis above it
is not possible to decide where they are located spatially in the cocoon . from the above analysis it is clear that the radio spectrum emitted by a given cocoon volume element depends on the ` energy loss history ' of this part of the cocoon .
adiabatic losses only change the normalisation of the emitted spectrum while its slope at a given frequency is governed by the radiative loss processes of the relativistic electrons . in the model described above the strength of the magnetic field which determines the magnitude of synchrotron losses
is tied to the pressure in the cocoon , @xmath27 .
the value of @xmath27 in turn depends on the energy transport rate of the jet , @xmath20 , and a combination of parameters describing the density distribution of the gas the cocoon is expanding into , @xmath77 . in the analytical scenario presented here
the volume elements are the building blocks of the cocoon and the variation of the radio surface brightness of the cocoon of an frii along its major axis can therefore potentially provide information on the properties of the source environment .
for this i now identify the cocoon volume elements @xmath28 with infinitesimally thin cylindrical slices with their radius , @xmath78 , perpendicular to the jet axis .
a similar approach is used in the model of chyy ( 1997 ) . however , in this model the dynamical evolution and the radio luminosity of the entire cocoon depend on the geometrical shape and evolution of the volume elements . in the model of ka and kda
this is not the case and the identification of the @xmath28 with a specific geometrical shape does not alter their results .
the volume of a thin cylindrical slice of the cocoon is given by @xmath79 , where @xmath80 is the very small thickness of the slice along the jet axis .
the radius of the slices , @xmath81 , depends on their position along the jet axis , @xmath82 , which i define in units of @xmath39 , the length of one half of the entire cocoon ( see sect .
[ sec : dyn ] ) .
the outer edges of the slices form the cocoon boundary or contact discontinuity .
most frii sources have cocoons of a relatively undistorted , ellipsoidal shape ( e.g. leahy et al . 1989 ) .
i therefore parameterise the cocoon boundary as @xmath83 where @xmath84 , @xmath85 and @xmath86 can be determined from radio observations of the cocoon . following the observational results of leahy & williams ( 1984 ) and leahy et al .
( 1989 ) ka and kda used the aspect ratio , @xmath46 , to characterise the geometrical shape of the cocoons of frii sources .
this ratio is defined as the length of one side of the cocoon measured from the radio core to the cocoon tip divided by its width measured half - way along this line . using this definition
it is straightforward to express @xmath87 in terms of @xmath46 as @xmath88^{-\alpha
_ 2}. \label{alpha_exp}\ ] ] the dimensionless volume constant @xmath89 now becomes @xmath90 where @xmath91 is the complete beta - function . tab .
[ tab : coeff ] shows typical values for the dimensionless constants in the model .
ccccc @xmath92 & @xmath89 & @xmath93 & @xmath94 & @xmath95 + + @xmath96 $ ] & @xmath97 $ ] & @xmath98 $ ] & @xmath99 $ ] & @xmath100 $ ] + the cylindrical slices are injected into the cocoon at a time @xmath101 . for simplicity
i assume that the slices remain and thus move within the cocoon as entities afterwards .
in other words , i neglect any mixing of material between slices and i also assume that the geometrical shape of the slices does not deviate from the initial thin cylinders . numerical simulations ( e.g. falle 1994 ) show that the gas flow in the cocoon is rather turbulent , at least close to the hot spot region .
it is likely that large - scale turbulent mixing in the cocoon leads to large distortions of the projected cocoon shape as seen in radio observations . in this case , the regular cocoon shape described by eq .
( [ shape ] ) will be a poor representation of the ` true ' cocoon shape .
for such distorted sources it is unlikely that the model presented here will provide a good description .
however , the simple picture of cylindrical slices may still represent the ` average ' behaviour of the gas flow in more regularly shaped cocoons rather well .
the model is designed to constrain source and environment parameters using mainly the gradient of the radio surface brightness along the cocoon .
problems with this simplified model will therefore arise if the relativistic electrons in the cocoon are distributed efficiently by diffusion . in sect .
[ sec : disc ] i show that the diffusion of relativistic particles is unlikely to change their distribution on large scales .
it is therefore reasonable to assume that the relativistic particles are effectively tied to the cocoon slice they were originally injected into .
numerical simulations of the large scale structure of frii sources strongly suggest that a backflow of material along the jet axis is established within the cocoon ( e.g. norman et al .
the model describing the source dynamics predicts the growth of the cocoon to be self - similar and therefore the backflow within the cocoon should be self - similar as well .
this suggests that the position of a slice of cocoon material injected into the cocoon at time @xmath74 is given by @xmath102 , where @xmath103 governs the velocity of the backflow at a given position along the cocoon . in order for the model to be self - consistent ,
all @xmath104 have to add up to the total volume of the cocoon , @xmath105 . using eqs .
( [ pressure ] ) , ( [ element ] ) , ( [ constant ] ) , ( [ alpha_exp ] ) and ( [ c3_exp ] ) and replacing @xmath106 by @xmath107 an implicit eq . for @xmath103 can be found from this integration ; @xmath108 / \left ( \alpha _ 1 \alpha _ 3 \right ) \right ) .
\nonumber \label{alpha3}\end{aligned}\ ] ] note that this expression requires @xmath109 and gives @xmath110 . for the values of the shape parameters used in sect .
[ sec : appl ] ( @xmath111 and @xmath112 ) and @xmath113 i find @xmath114 .
note that @xmath115 for @xmath116 .
the backflow velocity within the cocoon is given by @xmath117 where a dot denotes a time derivative .
in the rest frame of the host galaxy the backflow is observed to flow in the direction of the source core for @xmath118 . for @xmath119
the cocoon material is stationary in this frame and for positive values the backflow is strictly speaking not a ` backflow ' but trailing after the advancing hot spot .
the backflow is fastest just behind the hot spot and decelerates along the cocoon .
the deceleration implies a pressure gradient along the cocoon which may seriously violate the assumption made for the dynamical model of a constant pressure throughout the cocoon away from the hot spots . to estimate the magnitude of the pressure gradient i use euler s equation @xmath120 the density of the cocoon material , @xmath121 ,
is given by @xmath122 where i assumed that the entire energy in the jet is transported in the form of kinetic energy of the flow with a bulk velocity corresponding to the lorentz factor @xmath123 .
with this , it is straightforward with the help of eq .
( [ back ] ) to integrate eq .
( [ euler ] ) which yields the pressure along the cocoon as a function of @xmath124 ; @xmath125 } \nonumber \\ & \times & \left ( \frac{\dot{l}_{\rm j}}{c } \right ) ^2 l^{2+c_4/\alpha _
3 \left(1 - \gamma _ { \rm c } \right ) } , \label{check}\end{aligned}\ ] ] where @xmath126 and i have used the condition @xmath127 . the mean advance speed of the cocoon of an frii source , @xmath128 , is inferred from observations of lobe asymmetries to be in the range from 0.05 c ( scheuer 1995 ) to 0.1 c ( arshakian & longair 2000 ) .
this is in good agreement with the predictions of the dynamical model used here ( see ka ) .
the ratio @xmath47 is usually of order 5 ( see eq .
[ empirical ] ) and so even for only mildly relativistic bulk flow in the jet ( @xmath129 ) and large gradients in the backflow velocity , e.g. @xmath130 , eq .
( [ check ] ) predicts @xmath131 .
note that for @xmath132 the lower limit of this ratio increases to 0.82 . from this
i conclude that the existence of a pressure gradient along the jet within the cocoon is required to decelerate the backflow of the cocoon material .
however , the pressure varies only by a factor of a few at most along the entire length of the cocoon .
this implies that within this limit the model is self - consistent .
the model presented here depends on a large number of parameters .
i assume that each volume element only contains the relativistic particle population and therefore @xmath133 .
the other model parameters can be roughly grouped into * geometrical parameters : the constants describing the cocoon shape @xmath85 , @xmath86 and @xmath46 and the orientation of the jet axis with respect to the line - of - sight ( los ) , @xmath134 , * properties of the initial energy distribution of the relativistic electrons and/or positrons : the slope of the distribution , @xmath135 , and the high energy cut - off , @xmath136 , * properties of the source and its environment : the pressure within the cocoon , @xmath137 , the age of the source , @xmath10 , and the slope of the power law density distribution of the environment , @xmath0 . in the following
i discuss the methods employed in comparing the model predictions with observations . using these methods , i
will then show that because of the nature of the model several degeneracies between model parameters exist . eliminating these
requires further assumptions to be made but also reduces the complexity of the parameter estimation .
a radio map of an frii source is composed of pixels which contain information on the monochromatic radio surface brightness of the source , @xmath138 , at a given position @xmath139 projected onto the plane of the sky . here
i use the _ projected _ distance along the jet axis , @xmath140 , and perpendicular to the jet , @xmath141 .
both are measured in units of the projected length of the cocoon , @xmath142 . in optically thin conditions
generally present in radio source lobes this projection corresponds to a los integral of the synchrotron emissivity , @xmath143 , through the 3-dimensional source at each pixel location . following the analysis in sects .
[ sec : dyn ] and [ sec : syn ] , @xmath143 is only a function of the unprojected dimensionless distance from the core measured along the jet axis , @xmath124 , for a given set of the eight source and environment parameters .
using the model described above , it is therefore possible to construct ` virtual radio maps ' for a given set of model parameters by projecting the 3-dimensional model onto the plane of the sky . since a radio source is in general viewed with its jet axis at an angle @xmath134 to the los , the value of @xmath124 , and therefore that of @xmath144 , changes along the path of the los integral . also , the distance from the core to the tip of the cocoon measured on an observed radio map , @xmath142 , is not equal to the physical size of the cocoon , @xmath145 .
this , of course , also implies that the aspect ratio of the cocoon , @xmath146 , measured in the observed map is smaller than the ` real ' value @xmath147 , i.e. radio sources appear ` fatter ' than they really are .
all these projection effects have to be taken into account when comparing virtual radio maps resulting from these models to observations .
the surface brightness is calculated at a frequency @xmath148 which is given by the observing frequency in the frame of the observer , @xmath149 , and the redshift of the source , i.e. @xmath150 .
finally , to account for cosmological effects the result of the los integration must be multiplied by @xmath151 , where @xmath152 is the luminosity distance to the source .
the model map can be calculated at arbitrary resolution . however , before comparing the result with the observed map it must be convolved with the beam of the radio telescope used for the observations .
this was done assuming a 2-dimensional gaussian shape for the telescope beam . once a virtual map is compiled for a set of model parameters this map
can be compared pixel by pixel to a map resulting from observations of an frii radio source using a @xmath153-technique , @xmath154 ^2}{\sigma _ { \nu}^2}.\ ] ] here , @xmath155 is the measured monochromatic surface brightness with rms error @xmath156 and @xmath157 is the model prediction .
it is then possible to find the best - fitting model by varying the model parameters and thereby minimising the resulting @xmath158-difference between virtual and observed map .
the minimisation routine uses a n - dimensional downhill simplex method ( press et al .
1992 ) , where n is the number of model parameters to be fitted . the minimisation can be done separately for the two halves of each cocoon since the model describes one jet and the associated half of the cocoon .
although in principle the minimisation can be done with one observed map at a single observing frequency , the constraints on the model parameters are improved by using two maps at two different frequencies . in this case
the @xmath153-differences resulting from the two maps are simply added together . in principle
this is equivalent to compiling spectral index maps from two observed maps and comparing these with the model predictions .
however , using the two maps directly has the advantage that pixels below the rms limit in one map but not in the other are not entirely lost for the fitting procedure .
furthermore , information on the absolute surface brightness in a given location is not contained in a spectral index map .
the model would have to be normalised ` by hand ' and the various possibilities to do this would lead to ambiguities in the estimation of model parameters . in sec . [
sec : appl ] i use two individual maps at two frequency for each source . using the best - fitting model parameters @xmath159 and @xmath10
, the energy transport rate of the jet , @xmath160 , and the parameter combination @xmath161 describing the density distribution in the source environment can be calculated from eqs .
( [ pressure ] ) and ( [ length ] ) .
preliminary results using this comparison technique applied to radio observations of cygnus a was presented in kaiser ( 2000 ) . in sect .
[ sec : appl ] this analysis is extended to also include 3c 219 and 3c 215 .
the 2-dimensional comparison method described in the previous section requires a ray - tracing algorithm for the projection of the 3-dimensional model . for many maps of radio
sources the number of pixels are so large that this method can become computationally very expensive .
furthermore , for maps of lower resolution the cocoon may not be resolved in the direction perpendicular to the jet axis .
in such maps the surface brightness gradient along the jet axis can be extracted by taking a cut through the map along the line connecting the source core and the radio hot spot in the cocoon on one side .
this yields a 1-dimensional curve of surface brightness as a function of projected distance from the source core , @xmath162 .
the off - axis pixels will not add much information .
the function @xmath163 can then be compared with the model predictions and a best - fitting model may be found using two maps at two observing frequencies as outlined in the 2-dimensional case .
this 1-dimensional comparison involves a much smaller number of pixels for which a model prediction must be calculated than the 2-dimensional method .
lcc + & @xmath164 & @xmath165 + & @xmath166 & @xmath167 + & @xmath168 & @xmath169 + & @xmath46 & @xmath170 + & @xmath85 & @xmath171 + [ -2.25ex]fixed parameters & @xmath86 & @xmath172 + & @xmath135 & @xmath173 + & @xmath136 & @xmath174 + & @xmath0 & @xmath175 + & @xmath176 & @xmath177 + + & @xmath178 & @xmath179 + & @xmath180 & @xmath181 + derived parameters & @xmath182 & @xmath183 + & @xmath184 & @xmath185 + & @xmath146 & @xmath171 + & @xmath103 & @xmath186 + to study the influence of the individual model parameters on the surface brightness distribution , i define a fiducial model with a set of fixed model parameters given in tab . [
tab : modpara ] .
these model parameters imply a magnetic field just behind the hot spot of 13 nt or 130 @xmath187 g .
the highest frequency of the emitted synchrotron spectrum is then @xmath188 ghz which is close to the usually assumed high frequency cut - off in the spectrum of radio galaxies in minimum energy arguments ( e.g. miley 1980 ) . the density parameter @xmath189 kg m@xmath190 which corresponds to a central density of @xmath191 kg m@xmath192 or 0.3 particles per cm@xmath192 if @xmath194 kpc .
i also assume that the jet of the sources is 100 kpc long and is viewed at an angle of @xmath195 to the line of sight .
this implies that it would be observed to have a length of @xmath196 kpc corresponding to 31.2 " at a redshift @xmath197 . for a measured @xmath198
the aspect ratio of the cocoon is @xmath199 for the assumed viewing angle . for simplicity and ease of comparison
i use the 1-dimensional model which only predicts the surface brightness distribution along the jet axis ( see previous sect . ) .
the variation of the model predictions with varying parameters for the 2-dimensional case are essentially similar but the differences between maps are more difficult to visualise . the solid line in fig .
[ fig : pccomp ] shows @xmath200 at @xmath201 ghz assuming a pixel size of 0.3``@xmath2020.3 '' , corresponding to 0.8 [email protected] kpc for @xmath197 , appropriate for a telescope beam of 1.2 " fwhm . for simplicity and in contrast to the 2-dimensional comparison method the averaging effects of the observing beam which extends over four pixels on the curve was not taken into account . in any case , in this sect .
the only interest is in gross trends of the model predictions when the model parameters are varied and the effective smoothing of the beam on the already rather smooth curve is small .
a continuous curve is shown , since the pixel size is small compared to the scale of the plot .
the length of the cocoon corresponds to more than 100 pixels . as pointed out in sect .
[ sec : syn ] , the emission of the hot spot is not modeled in the approach presented here . the cocoon surface brightness shown in this and the following figures is caused by the cocoon material only after it has passed through the hot spot region .
the emission from the hot spot will in general dominate the total emission from the end of the cocoon at @xmath203 . in a comparison with observed maps the predictions of the model
can therefore not be used in this region .
also shown in fig .
[ fig : pccomp ] are the results for the same model with a higher and lower cocoon pressure . in the model the strength of the magnetic field and the energy density of the relativistic particles in the cocoon is tied to the cocoon pressure .
a higher pressure therefore causes a higher peak of the surface brightness distribution .
however , the increased synchrotron energy losses of the relativistic electrons also lead to a stronger gradient of the distribution towards the core of the source , i.e. the older parts of the cocoon . in fig .
[ fig : tcomp ] the effects of the source age on the distribution of @xmath204 is shown .
the population of relativistic particles at a given relative distance , @xmath124 , from the core of the source has spent more time in the cocoon in an old source compared to a young source .
this implies stronger energy losses and the peak of the surface brightness distribution is therefore located closer to the hot spot in an old source .
furthermore , in an observation with a given detectable threshold of @xmath200 a larger fraction of the cocoon of a young source will be visible than of an older source . the same source viewed at different angles of the jet axis to the los , as shown in fig .
[ fig : avcomp ] , results in a scaling of the surface brightness similar to the effects of changing the cocoon pressure .
note however , that the overall normalisation of @xmath205 depends sensitively on the value of the cocoon pressure .
the effects of a higher pressure also include a steepening of @xmath204 starting from the peak of this function towards the source core ( fig .
[ fig : pccomp ] ) .
this is not seen for variations of the viewing angle ( fig .
[ fig : avcomp ] ) . in practice
therefore , the best - fitting values for the cocoon pressure is set mostly by the overall surface brightness of the cocoon while the viewing angle is mainly determined by the behaviour of @xmath206 close to the core .
[ fig : multicomp ] summarises the effects of the remaining model parameters on the surface brightness distribution . changing the aspect ratio of the cocoon , @xmath46 , results in a scaling of @xmath204 very similar to the effects of a changing viewing angle , @xmath134 .
decreasing @xmath86 leads to the end of the cocoon to become more blunt which in turn implies a lower value of @xmath103 , i.e. a slower backflow within the cocoon ( eqs . [ alpha3 ] and [ back ] ) .
this is analogous to an older particle population at a given distance from the hot spot and the resulting surface brightness distribution is similar to that of an older source .
changing the second shape parameter @xmath85 has similar effects as that of a variation of @xmath86 and is therefore not shown .
the main effects of changing the slope , @xmath135 , and the high energy cut - off of the initial energy distribution of the relativistic electrons , @xmath136 , is a change in the scaling of @xmath205 ( see fig .
[ fig : multicomp ] ) .
this is caused by the dependence of the normalisation of the energy spectrum on @xmath135 and @xmath136 , eq .
( [ normal ] ) .
both scaling effects are similar to the effects of a variation of the viewing angle , @xmath134 .
the effect of a steeper energy spectrum also causes an off - set in the distribution of the spectral index as a function of @xmath124 as shown in fig .
[ fig : speccomp ] .
however , in the cocoon region closest to the source centre and least influenced by the hot spot emission , this effect is small compared to that of a variation of the viewing angle , @xmath134 .
finally , a different value of the slope of the gas density distribution in the source environment , @xmath0 , has a negligible effect on @xmath204 .
the relevant curves for @xmath207 and @xmath208 are indistinguishable from that for @xmath113 shown in figs .
[ fig : pccomp ] to [ fig : multicomp ] .
both , @xmath85 and @xmath86 , control how pointed the shape of the cocoon is .
the ends of the cocoon are dominated by the emission from the hot spots which are not part of the model .
furthermore , if the jet direction is not stable over the lifetime of the source , they may at times advance ahead of the rest of the cocoon which distorts the cocoon shape .
this has been referred to as the ` dentist drill ' effect ( scheuer 1982 ) . in the previous sect .
( see also eq .
[ alpha3 ] ) it was already shown that changing @xmath85 and/or @xmath86 results in a change of the profile and magnitude of the backflow in the cocoon .
this in turn causes changes to the surface brightness profiles similar to changes of the source age . for these reasons the model can not provide strong constraints on either of these parameters , at least not independent of the source age
therefore i set in the following @xmath209 and @xmath210 .
the surface brightness distribution predicted by the model is almost independent of the slope of the density profile of the external medium , @xmath0 .
i therefore set @xmath113 without influencing the model results significantly . note that because of eq .
( [ alpha3 ] ) this then also implies a fixed value for @xmath211 .
this may seem a rather significant restriction of the model but as was shown in the previous sect .
@xmath85 or @xmath86 , through their influence on @xmath103 , and the age of the source , @xmath10 , are degenerate model parameters . in the absence of geometrical constraints on @xmath85 or @xmath86 , setting them to reasonable values allows the determination of the source age from the fitting method .
the ages derived from the model will therefore always depend somewhat on the choice for the source geometry .
the slope of the initial energy distribution of the relativistic particles , @xmath135 , mainly influences the spectral index distribution in the cocoon .
observations suggest that @xmath212 and the model is therefore restricted to values in this range .
it follows that the model is rather insensitive to the exact value of the high energy cut - off of the distribution , @xmath136 .
a value of @xmath213 much smaller than the @xmath174 used in the fiducial model will , in addition to a change of the overall scaling , cause the emission region to shorten along the jet axis at a given observing frequency .
however , because the highest frequency of the synchrotron radiation of the cocoon , @xmath214 , depends on @xmath215 a small value of the high energy cut - off also implies a significantly smaller @xmath214 .
values substantially below @xmath216 ghz , which is assumed , here are unlikely in view of observations .
i therefore set @xmath217 .
as was pointed out above , the effects of varying the initial slope of the energy distribution of the relativistic particles , @xmath135 , are small in general . in sect .
[ sec : appl ] the 1 and 2-dimensional comparison methods are applied to observations of three frii - type radio sources . it is shown there that in almost all cases the best - fitting model parameters for the 2-dimensional method require @xmath135 to be close to 2 . to prevent the degeneracy between @xmath135 and @xmath134 to influence the model fits in the 1-dimensional method which involves fewer degrees of freedom , i set @xmath64 in this case .
finally , the degeneracy between @xmath46 , the width of the cocoon , and the viewing angle , @xmath134 , is resolved by determining the projected cocoon width @xmath146 from the observed map and use @xmath218 in the model calculations . using these additional assumptions ,
the number of model parameters which are fitted by comparison to the observations decreases to four : the pressure in the cocoon , @xmath159 , the source age , @xmath10 , the viewing angle , @xmath134 , and the initial slope of the energy distribution of the relativistic particles , @xmath135 . in the case of the 1-dimensional comparison method @xmath135
is fixed to a value of 2 .
to test the model predictions for the source environment against direct x - ray observations over a range of different viewing angles , the radio data of three different frii - type objects are used : the narrow - line radio galaxy cygnus a , the broad - line radio galaxy 3c 219 and the radio - loud quasar 3c 215 . according to orientation - based unification schemes of the various sub - classes of radio - loud agn ( e.g. barthel 1989 ) , the viewing angle , @xmath134 , of cygnus a
should be greater than those of 3c 219 and 3c 215 .
furthermore , 3c 219 and 3c 215 were selected because of their rather irregular radio lobe structure .
the model is based on a very regular geometrical shape of the cocoon , eq .
( [ shape ] ) , and using 3c 219 and 3c 215 it is possible to estimate to what extent this restriction limits the applicability of the model .
note , that results of the model fitting for cygnus a are presented in [ fig : least ] and [ fig : heast ] only at one frequency .
however , the model fits are always obtained for all sources using two maps at two different frequencies . .
[ fig : compa ] for this source at @xmath219 i used raw data from the vla archive at 1.8 ghz and 5 ghz .
the lower frequency observations were made in august / september 1987 in a - array and the 5 ghz data were obtained in january 1984 in b - array .
a detailed analysis of these and other observations of cygnus a can be found in carilli et al .
standard calibration and self - calibration was performed using the software package aips .
this resulted in two radio maps with comparable angular resolution of 1.3 `` at the two observing frequencies .
the clean components were restored in maps with an individual pixel size of 0.3''@xmath2020.3 " .
the map at 1.8 ghz with the hot spot emission removed ( see below ) is shown in the upper panel of fig .
[ fig : compa ] .
the rms noise in the maps is 0.03 jy beam@xmath2 at 1.8 ghz and 0.006 jy beam@xmath2 at 5 ghz .
all pixels in the maps with a surface brightness below 5@xmath220 were discarded for the 2-dimensional comparison method . in the case of the 1-dimensional method
only pixels below 3-@xmath220 were neglected .
the surface brightness distribution along the jet axis , @xmath200 , for the 1-dimensional method was obtained for both radio lobes along a cut from the core of the source to the eastern and western hot spot respectively . since the hot spot emission is not modeled , it has to be removed from the maps and the 1-dimensional cuts . in the maps an aperture centered on the surface brightness peak in each lobe and with a radius of 2.6
" corresponding to twice the beam width is removed ( see fig . [
fig : compa ] ) . in the western lobe
the bright secondary hot spot ( carilli et al . 1991 ) and the bright ridge connecting this hot spot with the main one are also removed . for the 1-dimensional method ,
the distance of the hot spot to the edge of the lobe , @xmath221 , was estimated as the distance of the maximum of @xmath204 to the last point at which this function has a value above 3@xmath220 in the direction away from the source core . to remove the contribution of the hot spots to @xmath200
, all pixels within @xmath222 of the edge of the lobe were neglected in the following 1-dimensional comparison process .
the spatial resolution of the radio maps of cygnus a is comparatively high .
the 1.3 `` angular resolution corresponds to a spatial resolution of about 1.9 kpc . for many sources , particularly at high redshift , maps of such high quality are not available . in order to estimate the effects of a lower angular and therefore lower spatial resolution
, i also convolved the two maps of cygnus a with a gaussian beam of 5 '' fwhm . in the case of the 2-dimensional method the radius of the aperture used to remove the hot spot emission
was fixed to 5 " . for the 1-dimensional method ,
the surface brightness distribution along the jet axis was extracted from these lower resolution maps in the same way as in the higher resolution case .
note that the emission of the hot spots is smeared out over a larger area in the low resolution maps . to avoid any bias from the enhanced emission of the hot spot region i used the higher value of @xmath222 obtained from the low resolution maps in both , the high and low resolution , 1-dimensional comparison .
cllcccc & & & @xmath223 & @xmath224 & @xmath225 & @xmath135 + + & & 2-d & @xmath226 & @xmath227 & @xmath228 & @xmath229 + & [ -2.25ex]5 " resolution & 1-d & @xmath230 & @xmath231 & @xmath232 & + [ -2.25ex]cygnus a , eastern lobe & & 2-d & @xmath233 & @xmath234 & @xmath235 & @xmath236 + & [ -2.25ex]1.3 " resolution & 1-d & @xmath237 & @xmath238 & @xmath239 & + & & 2-d & @xmath240 & @xmath241 & @xmath242 & @xmath243 + & [ -2.25ex]5 " resolution & 1-d & @xmath244 & @xmath245 & @xmath246 & + [ -2.25ex]cygnus a , western lobe & & 2-d & @xmath247 & @xmath248 & @xmath249 & @xmath250 + & [ -2.25ex]1.3 " resolution & 1-d & @xmath251 & @xmath252 & @xmath253 & + & & 2-d & @xmath254 & @xmath255 & @xmath256 & @xmath257 + [ -2.25ex]3c 219 & [ -2.25ex]4.3 " resolution & 1-d & @xmath258 & @xmath259 & @xmath260 & + & & 2-d & @xmath261 & @xmath262 & @xmath263 & @xmath264 + [ -2.25ex]3c 215 & [ -2.25ex]1.9 " resolution & 1-d & @xmath265 & @xmath266 & @xmath267 & + the eastern lobe of cygnus a is covered by 46.6 independent telescope beams along the jet while in the widest part there are 22.2 beams across . in the lower resolution maps these numbers decrease to 12.9 and 6.1 , respectively .
note , that only a fraction of the lobes has an observed surface brightness above the rms limits , i.e. they do not extend all the way from the hot spots to the core in the observations .
this implies that the model fits are based on regions covering less area than the theoretical extend of the cocoons . for both resolutions
i find an axial ratio , @xmath146 , of 2.1 at the point where the lobe is widest from the 2-dimensional maps .
the length of the lobe , @xmath142 , is 64.4 `` for the low resolution map and 60.6 '' for the high resolution map .
the prediction of the best - fitting model in comparison with the observations is shown in the lower panels of figs .
[ fig : compa ] and [ fig : compac ] for the 2-dimensional method and in figs .
[ fig : least ] and [ fig : heast ] for the 1-dimensional method at 1.8 ghz .
the 5 ghz data is not shown for the 1-dimensional method but is similar to the result at 1.8 ghz .
the parameters of the best - fitting models are given in table [ tab : res ] .
the errors on these and for all the following model fits are estimated using the boot - strap method ( e.g. press et al .
it is not possible to calculate error estimates using the @xmath153-values derived in the minimisation procedure directly as the values of the surface brightness in neighboring pixels are not independent .
roughly 2000 data sets were created by drawing data with replacement from the original set . the same minimisation procedure as in the original model fitting
was then applied to them and the error given in the table is the 1@xmath220 limit on the respective model parameters . even after subtracting the contribution of the hot spot
the deviations of the model from the observations are large .
the fact that the model fit is poorer at 5 ghz is mainly caused by the smaller rms noise of the observed 5 ghz map .
as expected , the model can not fit structures which appear as discrete surface brightness enhancements in the maps .
this is particularly clear in the case of the bright arc seen just behind the hot spot in radio maps of the eastern lobe ( see figs .
[ fig : compa ] and [ fig : compac ] ) which also causes the secondary peak in @xmath204 at about @xmath268 ( see fig . [
fig : heast ] ) .
although convolving the maps with a larger beam ` draws ' some flux from the hot spot into the arc mentioned above , the results for the two different resolutions are very similar for both comparison methods . at both frequencies
the observed maps show a concentration of the radio emission towards the centres of both lobes .
this region is also extended a long way along the jet axis , particularly in the western lobe ( see following section ) . in figs .
[ fig : compa ] and [ fig : compac ] it is clear that the model can not fit this concentration properly .
the very smooth lobes of the model are ` fatter ' than the observed lobes further away from the hot spots and do not extend as far back as the observations indicate .
this is particularly striking in the western lobe at 5 ghz ( see fig .
[ fig : compac ] ) .
this clearly illustrates the limitations of the model assumption of a smooth shape of the cocoon and a regular backflow within the cocoon .
the large uncertainty of the viewing angle , @xmath134 , is caused by the model mainly depending on @xmath269 which changes only by a factor 1.2 within the estimated errors .
this is comparable to the uncertainties of the other model parameters .
cllcccc & & & @xmath225 & @xmath270 & @xmath271 & @xmath272 + + & & 2-d & @xmath228 & @xmath273 & @xmath274 & @xmath275 + & [ -2.25ex]5 " resolution & 1-d & @xmath232 & @xmath276 & @xmath277 & @xmath278 + [ -2.25ex]cygnus a , eastern lobe & & 2-d & @xmath235 & @xmath279 & @xmath280 & @xmath281 + & [ -2.25ex]1.3 " resolution & 1-d & @xmath239 & @xmath282 & @xmath283 & @xmath284 + & & 2-d & @xmath242 & @xmath285 & @xmath286 & @xmath287 + & [ -2.25ex]5 " resolution & 1-d & @xmath246 & @xmath288 & @xmath289 & @xmath290 + [ -2.25ex]cygnus a , western lobe & & 2-d & @xmath249 & @xmath291 & @xmath292 & @xmath293 + & [ -2.25ex]1.3 " resolution & 1-d & @xmath253 & @xmath294 & @xmath295 & @xmath296 + cygnus a & observed & & @xmath297 & & @xmath298 & @xmath299 + & & 2-d & @xmath256 & @xmath300 & @xmath301 & @xmath302 + [ -2.25ex]3c 219 & [ -2.25ex]4.3 " resolution & 1-d & @xmath260 & @xmath303 & @xmath304 & @xmath305 + 3c 219 & observed & & @xmath306 & & @xmath307 & @xmath308 + & & 2-d & @xmath263 & @xmath309 & @xmath310 & @xmath311 + [ -2.25ex]3c 215 & [ -2.25ex]1.9 " resolution & 1-d & @xmath267 & @xmath312 & @xmath313 & @xmath314 + 3c 215 & observed & & @xmath306 & & @xmath315 & @xmath316 + using eqs . ( [ pressure ] ) and ( [ length ] ) the power of the jet , @xmath160 , and the central value of the density distribution of the gas surrounding cygnus a , @xmath317 , are calculated .
the results are given in table [ tab : deriv ] . here
i assume that the core radius of the environmental density distribution in eq .
( [ king ] ) is given by @xmath318 kpc .
the viewing angle @xmath134 is inferred from the flux ratio of the jet to the counter - jet in the two lobes of cygnus a ( hardcastle et al .
this assumes that the two jets are identical and that the observed flux ratio is entirely due to relativistic beaming effects .
furthermore , a constant bulk velocity within the jets , @xmath319 , is assumed and set to 0.62 c. variations of @xmath320 across the source and asymmetries between the two jet sides will significantly influence the estimate for @xmath134 .
however , the model is consistent with the estimate given by hardcastle et al . ( 1999 ) .
the observed central density of source environment given in table [ tab : deriv ] is derived from x - ray observations rosat of the hot gas surrounding cygnus a hardcastle & worrall ( 2000 ) . for this , the prescription of birkinshaw & worrall ( 1993 ) for the conversion of central surface brightness to central proton density was used . the core radius ,
@xmath16 was estimated by hardcastle & worrall ( 2000 ) from the x - ray observations and i use their value , @xmath318 kpc , in converting from the density parameter @xmath321 given by the model to @xmath317 .
the value thus found from the best - fitting model agrees within the error with the x - ray observations . the western lobe of cygnus a is covered by 53.5 independent beams along the jet and 23.2 beams at the widest point perpendicular to the jet .
for the lower resolution maps i find 14.6 and 6.3 beams , respectively .
similar to the eastern lobe the observed emission does not extend all the way from the hot spot to the core and so the model fit is based on a smaller area . from the 2-dimensional maps i find @xmath322 for the widest part of the lobe and the length of the lobe is 73.0 `` and 69.5 '' for the low and high resolution case respectively .
the best - fitting models for the two different resolutions again agree well .
the models yield an age for the western lobe somewhat higher than that of its eastern counterpart ( see table [ tab : res ] ) which is mainly caused by its greater length
. however , the pressure within the cocoon is remarkably similar in both lobes .
this implies also good agreement between the estimates for the jet power and the density of the source environment between the two sides of cygnus a ( table [ tab : deriv ] ) . for 3c 219 at @xmath323 vla maps at 1.5 ghz in b - array and 4.9 ghz in c - array
were used .
the observations were taken in october and december 1998 by dennett - thorpe et al .
( in preparation ) who also performed standard reduction on the data set using aips .
the resolution of the resulting maps is roughly 4.3 `` and the rms noise is @xmath324 jy beam@xmath2 at 1.5 ghz and @xmath325 jansky beam@xmath2 at 4.9 ghz . the northern lobe of 3c 219 has a rather irregular shape and no clear hot spot ( clarke et al . 1992 ) .
this leads to large ambiguities in the determination of its length or the geometrical parameters needed for the model presented here .
i therefore only used the southern lobe which has a length of 40.4 '' and an aspect ratio at its widest point of @xmath326 .
the lobe is covered by 9.4 independent beams along the jet axis and by 5.9 beams perpendicular to it .
the jet and counter - jet in 3c 219 are unusually bright and so the jet emission was removed from the maps of the southern lobe . for the 1-dimensional comparison method ,
i extracted @xmath204 along a line off - set by 2 `` to the south of the line connecting the core of the source with the hot spot of the southern lobe .
this avoids contamination of the surface brightness distribution by the jet emission .
the hot spot in the southern lobe of 3c 219 is somewhat set back from the edge of the lobe .
an aperture with a radius of 4.3 '' , i.e. the width of the telescope beam , centered on the surface brightness peak was removed from the maps . for the 1-dimensional comparison , only values of @xmath204 core - wards of the hot spot were used .
the parameters of the best - fitting model are given in table [ tab : res ] .
the uncertainties of the model parameters is comparable to those found for the two lobes of cygnus a. the angle to the los of 3c 219 is found to be smaller than that of cygnus a. since 3c 219 is a broad line radio galaxy this is in the expected sense , but the value of @xmath134 predicted by the model is about double that inferred from the flux ratio of the jet and counter - jet . as mentioned above , the jets of 3c 219 are unusually bright and this may reflect some enhanced disruption of the jet flow by turbulence or even a complete restart of the jets in this source ( clarke & burns 1991 ) . the latter possibility has let schoenmakers et al . ( 2000 ) to include this source among their examples of double - double radio galaxies ( ddrg ) .
the morphology of these sources strongly suggests restarting jets ( kaiser et al .
2000 ) . the large jet to counter - jet flux ratio of 3c 219 may therefore be caused by effects other than relativistic beaming .
the best - fitting value of @xmath327 is consistent with orientation - based unification schemes .
the central density of the gas surrounding 3c 219 , @xmath317 , predicted by the model is considerably smaller than that inferred from x - ray observations .
to derive @xmath317 i used @xmath328 kpc ( hardcastle & worrall 1999 ) .
this discrepancy will be discussed in sect .
[ sec : disc ] .
this source is a radio - loud quasar at @xmath329 with very irregular morphology ( bridle et al .
i obtained raw observational data from the vla archive at 1.5 ghz in a - array and 4.9 ghz in b - array .
the 1.5 ghz observations were carried out by miley in may 1986 while the 4.9 ghz observations were taken by hough in december 1987 . again standard reduction with aips
was performed on the data and resulted in two maps with an angular resolution of 1.9 `` .
the maps were restored using a pixel size of 0.3''@xmath2020.3 `` and the rms noise is @xmath330 jansky beam@xmath2 at 1.5 ghz and @xmath331 jansky beam@xmath2 at 4.9 ghz .
the southern half of 3c 215 is very distorted with the jet bending in various places with enhanced surface brightness ( bridle et al .
this part of the source is not consistent with a regular frii - type lobe morphology and resembles in its outer regions an fri - type structure .
therefore no attempt was made to apply the model to the southern part of the source .
the northern lobe is more regular , however , the hot spot here is weak and the lobe widens considerably close to the core in a north - eastern direction .
the lobe has a length of 26.6 '' and its aspect ratio @xmath146 at the point where the width of the lobe is greatest is 1.2 .
the hot spot in the northern lobe is not located at the very edge of the lobe similar to the southern lobe of 3c 219 .
an aperture centered on the surface brightness peak with a radius 3.8 corresponding to the size of two telescope beams was removed from the map . for the 1-dimensional comparison the surface brightness distribution was extracted along a line off - set by about 2 " to the east from the core - hot spot direction to avoid emission from the jet . again
only pixels core - wards from the hot spot were used in the 1-dimensional case . the northern lobe of 3c 215 is covered by 14.0 independent beams along the jet and 11.7 beams at the widest point perpendicular to the jet .
parameters of the best - fitting model and the derived properties of the environment of 3c 215 are given in tables [ tab : res ] and [ tab : deriv ] .
the uncertainties for the model parameters are considerably larger for this source than for the two previous ones .
the viewing angle to the jet axis , @xmath134 , is smaller than for cygnus a or or 3c 219 .
this is again consistent with the predictions of unification schemes as 3c 215 is a quasar .
the smaller observed value is again inferred from the flux ratio of the jet and counter - jet of 3c 215 ( bridle et al .
1994 ) . similarly to 3c 219
this ratio may be increased in 3c 215 because of the distorted morphology of its large scale radio structure .
the southern jet does not seem to be embedded in a cocoon and it is therefore unlikely that the two jets are intrinsically identical .
the flux ratio probably reflects physical processes other than purely relativistic beaming . for the determination of @xmath317 i used @xmath332 kpc from hardcastle & worrall ( 1999 ) .
the density of the gaseous environment of 3c 215 is predicted to be much lower than inferred from x - ray observations .
discussion of this point is deferred to sect .
[ sec : disc ] .
in general it is expected that an increase in the amount of information available to a given model fitting procedure should decrease the uncertainty in the determination of the model parameters . comparing the results of the 1 and 2-dimensional methods for the lobes of cygnus
a one would therefore expect that the parameter uncertainties decrease for the higher resolution maps .
this is indeed the case for the 2-dimensional method , particularly for the western lobe .
however , the 1-dimensional method shows the opposite . here
the uncertainties are larger for the higher resolution maps .
this is caused by the attempt to fit a very smooth model for the surface brightness to observational data which shows considerably greater local variations than the model .
the fact that the model is unable to fit the local surface brightness structure observed was already noted in the previous sect .. in the 2-dimensional case the off - axis parts of the cocoons provide additional information and the influence of local structure is therefore to some extent averaged out in the fitting procedure . in other words ,
the 2-dimensional method is able to make use of the larger amount of information in higher resolution maps .
the 1-dimensional method is restricted to a cut through the lobe . here
, the averaging effect of a larger telescope beam provides for a smoother surface density profile and the model fits the data better . in the case of radio maps of low resolution which do not or only barely resolve the lobes perpendicular to the jets along most of their lengths , the 2-dimensional maps add little information to 1-dimensional cuts along the lobes . in these cases the additional model parameter @xmath135 in the 2-dimensional method presented here allows the model to fit the data with a large range of parameter combinations .
the uncertainties of the model parameters are then larger than for the 1-dimensional method .
this effect can be seen for 3c 215 and , to a lesser extent , for the low resolution maps of the western lobe of cygnus a. it is of course possible to fix the value of @xmath135 for the 2-dimensional method as well but this does not significantly improve the constraints on the model parameters compared to the 1-dimensional method . which method is the best to use for a particular set of radio maps depends on the quality of the maps . if the radio lobes are well resolved along to the jet axis as well as perpendicular to it then the 2-dimensional methods will provide better constraints . however , it is computationally expensive . in the case of poorer resolution
the 2-dimensional method will not add anything to the results obtained from a 1-dimensional comparison .
for heavily distorted lobe structures both methods will fail but the 1-dimensional method may still provide order of magnitude estimates if the lobes are not entirely dominated by bright localised structure
. orientation - based unification schemes attempt to explain radio galaxies and radio - loud quasars as essentially the same type of objects albeit viewed at different angles to the jet axis ( barthel 1989 ) .
the broad line radio galaxies may then be identified as the low redshift analoga to radio - loud quasars .
these unification schemes imply that the viewing angle , @xmath134 , to the jet axis is greater than about @xmath333 for radio galaxies and smaller for quasars .
tests of this hypothesis are often inconclusive as the determination of @xmath134 is difficult in practice .
as described above , the viewing angle @xmath134 may be determined from radio observations alone . however , since the model depends mostly on @xmath334 , the uncertainties are large particularly for @xmath335 .
fortunately , the other model parameters and the source and environment properties inferred from these do not depend strongly on @xmath134 and its error in this range .
this agrees with the results of wan & daly ( 1998 ) who study the effects of the viewing angle on the determination of a variety of source properties in great detail .
the source orientation can only be determined accurately if the 2-dimensional comparison method is used on radio maps well resolved perpendicular to the jet axis .
this agrees with the findings of wan & daly ( 1998 ) who use a different source model . despite the large uncertainties in the model results presented here it is interesting to note that for the three sources studied the best - fitting values of @xmath134 are smaller for the quasar ( 3c 215 ) and the broad line radio galaxy ( 3c 219 ) than for the radio galaxy ( cygnus a ) . as i showed above , the model parameters of the best - fitting models
can be used to infer the density parameter , @xmath321 , describing the density distribution in the environment of frii objects .
it is impossible using just this model to separate this into the central density , @xmath317 , and the core radius , @xmath16 .
furthermore , the model is insensitive to the slope , @xmath0 , of the external density profile ( see sect . [
sec : degen ] ) . to infer central densities , values for the core radius and @xmath0 must be taken from other observations . in table
[ tab : deriv ] @xmath317 for the three example sources is shown for the core radii given by hardcastle & worrall ( 1999 ) and @xmath113 .
as noted in sect .
[ sec : appl ] , the central densities found for 3c 219 and 3c 215 are inconsistent with those derived from the x - ray observations .
in several studies it was found that the pressure of the frii source environment derived from x - ray observations apparently exceeds the pressure inside the radio lobes ( e.g. hardcastle & worrall 2000 and references therein ) .
the discrepancy in density found here is essentially the same phenomenon in the framework of isothermal density distributions for the source environments as described by the @xmath0-model , eq .
( [ beta ] ) .
all recent models for the evolution frii sources , including the one discussed here , are based on the assumption that the cocoon is overpressured with respect to its surroundings .
the x - ray observations seem to contradict this assumption .
can the discrepancies be resolved ? in the model described here
i approximate the density distribution in the source environment , assumed to follow the @xmath0-model , by power laws . for cocoons extending well beyond the core radius the exponent of the power law
is given by @xmath336 .
in the case of cygnus a this is justified as both lobes extend over more than 100 kpc while @xmath318 kpc . in the analysis @xmath113
was used which is identical with the result of hardcastle , @xmath337 . for 3c 219 and 3c 215 @xmath338 from the x - ray observations and in both sources the radio
lobes only extend to about @xmath339 .
formally the model can not be applied to these two sources because the underlying dynamical model is valid only for @xmath340 or @xmath341 . however , as the cocoons in both sources do not extend much beyond the core radius , the density distribution in the environments of 3c 219 and 3c 215 are well represented by a power law with exponent @xmath113 . in these cases
@xmath342 found from the model is not identical to the central density of the @xmath0-model . from eq .
( [ beta ] ) it can be seen that the latter is given by @xmath343 , where @xmath344 is the length of the radio lobe in units of @xmath16 .
this correction is insufficient to make the model results consistent with the x - ray observations of 3c 219 and 3c 215 .
x - ray observations of the hot gaseous environment of agn are influenced by the presence of the active galaxy .
before the environment properties can be extracted , the bright x - ray emission of the agn itself appearing as a point source must be carefully removed . in the case of radio - loud objects the large scale structure caused by the jets
can also alter the x - ray emission of its surroundings .
the magnitude of these effects is difficult to estimate if the x - ray observations do not fully resolve the scale of the cocoon .
the hot spots at the end of the jets are strong sources of inverse compton scattered x - ray photons .
when resolved , these inverse compton scattered photons are found to distort the x - ray contours of the extended emission ( cygnus a , carilli et al .
1994 ; 3c 295 , harris et al .
the more extended cocoon material itself may also act as a scatterer of cmb photons or of agn emission ( brunetti et al .
1999 ) . finally , the bow shock surrounding the cocoon compresses and heats the gas in the source environment .
kaiser & alexander ( 1999b ) give an estimate for the expected x - ray luminosity from this shocked layer of gas , @xmath345 _ { \nu _ 2}^{\nu _ 1},\end{aligned}\ ] ] where two typographical errors are corrected .
expressions for @xmath346 , @xmath347 and @xmath348 may be found in kaiser & alexander ( 1999b ) .
the square brackets mean the difference of the exponential function at the limits of the observing band @xmath349 and @xmath350 in the rest frame of the source .
the resulting x - ray luminosity can be converted to a rosat count rate using the internet version of pimms . for the best - fitting model parameters for cygnus
a this implies that roughly 2% of all counts of the rosat observations presented by hardcastle & worrall ( 2000 ) attributed to the entire extended emission come from the layer of shocked gas in between bow shock and cocoon of this source . for 3c 219 this region around the southern lobe alone
contributes about 10% of the relevant counts .
the contribution of the northern lobe of 3c 215 is negligible ( @xmath351% ) .
however , in this case at least three very compact radio emission regions along the distorted jets are detected which may be powerful inverse compton sources ( bridle et al .
only x - ray observations resolving the scale of the radio structure in this and other sources will allow us to determine the contribution to the total x - ray emission from such compact regions and hot spots . even without detailed observations
it is clear that powerful radio sources can contribute significantly to the extended x - ray emission in the central part of their environments .
helsdon & ponman ( 2000 ) show for the case of loose groups of galaxies that such an overestimate at the centre of an x - ray surface brightness profile may lead to overestimates of the core radius and also of @xmath352 when fitted with a @xmath0-model . using these overestimated values of @xmath16 and @xmath0 then yields values for @xmath353 which are too low compared to the ` real ' central density .
an exact analysis of the magnitude of this effect is beyond the scope of this paper . in any case
, this effect may explain the discrepancies found between the two methods of estimating the central densities and pressures of the frii environments .
in this paper an analytical model for the surface brightness distribution of the cocoons of frii - type radio sources is developed .
the model is based on the self - similar model of kaiser & alexander ( 1997 , ka ) and its extension by kaiser et al .
( 1997 , kda ) .
the cocoon is split into small volume elements , the temporal evolution of which is traced individually .
these elements are then identified with cylindrical slices which are rotationally symmetric about the jet axis .
the bulk backflow of these slices is determined self - consistently from the constraints on the cocoon shape .
thus a 3-dimensional model of the synchrotron emissivity is constructed . projecting this along the los , the model yields surface brightness profiles in one or two dimensions in dependence of several model parameters .
it is shown that a number of degeneracies prevent constraints to be placed on all parameters .
comparatively robust estimates can be found for the cocoon pressure , @xmath159 , the source age , @xmath10 , the angle of the jet to the los , @xmath134 , and , if the 2-dimensional comparison method is used , the initial slope of the energy distribution of the relativistic particles , @xmath135 .
the model may be viewed as an extension of the classical spectral aging methods .
however , it accurately takes into account the loss history of the relativistic particles and the evolution of the magnetic field in the cocoon .
the age estimates derived from the model are therefore accurate and not only lower limits .
it is also shown that diffusion of relativistic particles can not significantly distort the energy spectrum of the relativistic particles in the cocoon . in sect .
[ sec : appl ] the model predictions were compared with observational data for the three sources cygnus a , 3c 219 and 3c 215 .
the four free model parameters are constrained to within at least a factor 2 .
the uncertainties of the cocoon pressure are remarkably small , less than 50% , while the source ages are constrained by the model to within 80% or better .
the errors on the viewing angle are large as the model mainly depends on @xmath334 which changes in the range of orientation angles found here ( 31@xmath354 to 90@xmath354 ) by only 52% .
the uncertainties in the model parameters translate to errors in the determination of the jet power , @xmath160 , and the density parameter , @xmath321 . however , the resulting uncertainties still allow at least order of magnitude estimates for these quantities .
the model results do not depend strongly on the resolution of the radio observations used in the fitting procedure .
the results for both lobes of cygnus a derived at 1.3 `` and 5 '' resolution are consistent with each other within the errors .
however , the uncertainties in the parameter determination increase significantly with decreasing resolution . for radio maps which barely resolve the lobes perpendicular to the jet axis
, the 2-dimensional method adds little in terms of constraints on model parameters to the 1-dimensional method . from the examples studied here it is not possible to give a firm lower limit for the resolution of maps used to fit with the model . however , the parameter uncertainties calculated here suggest that at least 4 independent beams along the jet axis are required . as was pointed out in sect . [
sec : back ] , distorted lobe structures may indicate large turbulent flows within the cocoon .
as the model is based on the assumption of largely non - turbulent backflow it can not be used for such irregular sources .
the model can be used to infer the viewing angle of frii - type sources to the los . despite the large uncertainties in this determination ,
the three sources modeled here show the dependence of @xmath134 on the spectral type of the host galaxies expected from orientation - based unification schemes .
the central density of the source environments found using this model is consistent with x - ray observations in the case of cygnus a. in the case of 3c 219 and 3c 215 the estimates are too low .
these discrepancies may stem from overestimates from x - ray observations of the core radius , @xmath16 , and slope , @xmath0 , of the density distributions .
they are caused by the contribution of the large scale structure , i.e. cocoons and hot spots , of the radio sources to the x - ray emission .
the , at present , insufficient spatial resolution of x - ray maps prevents the removal of these contributions before the determination of @xmath16 and @xmath0 .
the model allows the simultaneous determination of a number of key parameters of extragalactic radio sources of type frii . together with the 1 and 2dimensional comparison methods , it is straightforward to analyse large numbers of objects drawn from complete samples of extragalactic radio sources .
the quality of the radio data needed is only moderate and will already be available for many of the objects in question
. however , the large scale structure of the objects should be reasonably regular in the sense that large scale turbulent motion within the cocoon is unlikely . within these limits
the method can provide order of magnitude estimates for the properties of the sources and their environments . this may help to answer some of the questions related to the cosmological evolution of the frii population and unification schemes of radio - loud agn .
it is a pleasure to thank p. n. best , j. dennett - thorpe , h. j. hardcastle and m. lacy for valuable discussions .
i would also like to thank p. n. best and a. p. schoenmakers for help with the reduction of the radio data and j. dennett - thorpe for providing the maps of 3c 219 .
many thanks to the anonymous referee for valuable comments .
63 alexander p. , leahy j.p .
, 1987 , mnras 225 , 1 arshakian t.g . ,
longair m.s . , 2000 ,
mnras 311 , 846 barthel p.d . , 1989 ,
apj 336 , 606 begelman m.c . ,
cioffi d.f . , 1989 , apj 345 , l21 birkinshaw m. , worrall d.m . , 1993 ,
apj 412 , 568 black a.r.s . , 1992 , _ jets , wings and hotspots the complex nature of powerful radio galaxies _ ,
thesis , university of cambridge blundell k.m . , rawlings s. , 1999 , nat 399 , 330 blundell k.m . , rawlings s. , 2000 , aj 119 , 1111 blundell k.m . , rawlings s. , willott c.j . , 1999 , aj 117 , 677 blundell k.m . , rawlings s. , willott c.j . , kassim n.e .
, perley r.a . , 2000a , in j. biretta and j. p. leahy ( eds . ) , lifecycles of radio galaxies .
new astronomy reviews , astro - ph/9910158 blundell k.m . ,
kassim n.e . ,
perley r.a .
, 2000b , in r. d. ekers et al .
( ed . ) , the universe at low radio frequencies .
asp conference series , astro - ph/0004005 bridle a.h .
, hough d.h . ,
lonsdale c.j . , burns j.o .
, laing r.a . , 1994 , aj 108 , 766 brunetti g. , comastri a. , setti g. , feretti l. , 1999 , a&a 342 , 57 carilli c.l . , perley r.a .
, dreher j.w . , leahy j.p . , 1991 , apj 383 , 554 carilli c.l . , perley r.a . ,
harris d.e . , 1994 ,
mnras 270 , 173 chuvilgin l.g . ,
ptuskin v.s . , 1993 ,
a&a 279 , 278 chyy k.t . , 1997 ,
mnras 289 , 355 clarke d.a .
, burns j.o . , 1991 , apj 369 , 308 clarke d.a . ,
bridle a.h . , burns j.o .
, perley r.a .
, norman m.l . , 1992 , aj 385 , 173 crawford c.s . , lehmann i. , fabian a.c . , bremer m.n . ,
hasinger g. , 1999 , mnras 308 , 1159 daly r.a .
, 1994 , apj 426 , 38 duffy p. , kirk j.g . , gallant y.a .
, dendy r.o . , 1995 , a&a 302 , l21 falle s.a.e.g . , 1991 ,
mnras 250 , 581 falle s.a.e.g .
, 1994 , mnras 269 , 607 fanaroff b.l .
, riley j.m . , 1974 , mnras 167 , 31 garrington s.t . ,
leahy j.p . , conway r.g . , laing r.a . , 1988 , nat 331 , 147 hardcastle m.j . , worrall d.m . , 1999 , mnras 309 , 969 hardcastle m.j . , worrall d.m . , 2000 , mnras : submitted .
hardcastle m.j .
, alexander p. , pooley g.g .
, riley j.m . , 1999 , mnras 304 , 135 d.e . , nulsen p.e.j .
, ponman t.j .
, 2000 , apj 530 , l81 helsdon s.f .
, ponman t.j . , 2000 , mnras : submitted , astro - ph/0002051 hill g.j . , lilly s.j . , 1991 , apj 367 , 1 hughes p.a . , miller l. , 1991 , in p. a. hughes ( ed . ) , beams and jets in astrophysics .
cambridge university press , p. 52
kaiser c.r . , 2000 , in j. biretta and j. p. leahy ( eds . ) , lifecycles of radio galaxies .
new astronomy reviews , astro - ph/9908261 kaiser c.r . , alexander p. , 1997
, mnras 286 , 215 , [ ka ] kaiser c.r . , alexander p. , 1999a
, mnras 302 , 515 kaiser c.r . , alexander p. , 1999b
, mnras 305 , 707 kaiser c.r . , dennett - thorpe j. , alexander p. , 1997
, mnras 292 , 723 , [ kda ] kaiser c.r . , schoenmakers a.p . , rttgering h.j.a . ,
2000 , mnras 315 , 381 katz - stone d.m .
, rudnick l. , 1994 , apj 426 , 116 katz - stone d.m . , rudnick l. , anderson m.c . , 1993 ,
apj 407 , 549 laing r.a . , 1980 ,
mnras 193 , 439 laing r.a . , 1988 ,
nat 331 , 149 leahy j.p .
, williams a.g .
, 1984 , mnras 210 , 929 leahy j.p .
, muxlow t.w.b .
, stephens p.w . , 1989 ,
mnras 239 , 401 miley g. , 1980 , ara&a 18 , 165 muxlow t.w.b .
, garrington s.t . , 1991 , in p. a. hughes ( ed . ) ,
beams and jets in astrophysics .
cambridge university press , p. 52
nath b.b . , 1995 ,
mnras 274 , 208 neeser m.j .
, eales s.a . , duncan - green j. , leahy j.p . ,
rawlings s. , 1995 , apj 451 , 76 norman m.l . , smarr l. , winkler k.h.a . , smith m.d . , 1982 ,
a&a 113 , 285 press w.h . , teukolsky s.a .
, vetterling w.t . ,
flannery b.p . , 1992 , _ numerical recipes .
second edition .
_ , cambridge university press , cambridge , uk . rechester a.b . , rosenbluth m.n . , 1978 , phys
40 , 38 roland j. , hanisch r.j . , pelletier g. , 1990 , a&a 231 , 327 rudnick l. , katz - stone d.m . , anderson m.c . , 1994 , apj supp .
90 , 955 sambruna r.m . , eracleous m. , mushotzky r.f . , 1999 ,
apj 526 , 60 sarazin c.l . , 1988 , _ x - ray emission from clusters of galaxies _ , cambridge university press , cambridge . scheuer p.a.g . , 1974 ,
mnras 166 , 513 scheuer p.a.g . , 1982 ,
in d. s. heeschen and c. m. wade ( eds . ) , extragalactic radio sources .
reidel , p. 163
scheuer p.a.g .
, 1995 , mnras 277 , 331 schoenmakers a.p .
, de bruyn a.g .
, rttgering h.j.a . , van der laan h. , kaiser c.r . , 2000 ,
mnras 315 , 371 shu f.h . , 1991 , _ the physics of astrophysics .
vol.1 : radiation _ , university science books , mill valley wan l. , daly r.a . , 1998 , apj 499 , 614
in the model it will be assumed that the magnetic field within the cocoon is completely tangled on a scale @xmath355 much smaller than the size of the cocoon , @xmath356 . in this case
the relativistic electrons can diffuse quickly within ` patches ' of coherent magnetic field of a size @xmath355 .
however , it is difficult for them to escape their patch as this would involve diffusion perpendicular to field lines . in the presence of turbulent motion within the plasma underlying the magnetic field ,
patches can be stretched out and in this case the diffusion of charged particles into an adjacent patch becomes more likely .
duffy et al . ( 1995 ) calculate how far a given electrons has to travel along a field line in its original patch before the patch becomes so stretched out that it can escape .
note here that this treatment is valid only in the quasi - linear regime for which the relative amplitude of irregularities of the magnetic field is much smaller than the ratio of the turbulent correlation lengths perpendicular and parallel to the local magnetic field .
the efficiency of this diffusion then depends crucially on whether the electron is able to travel this distance ballistically or whether it must diffuse along the field line .
the more efficient ballistic regime requires that where @xmath358 is the gyro - radius of the particle in the field , @xmath359 is the diffusion coefficient perpendicular to the field and @xmath360 is the time it takes the particle to escape the patch . for a relativistic electron moving at speed @xmath361 and corresponding lorentz factor
@xmath362 following chuvilgin & ptuskin ( 1993 ) @xmath363 with @xmath364 the bohm diffusion coefficient and @xmath365 , where @xmath366 is the rate of collision of the particle and @xmath367 is its gyro - frequency .
for the assumption of blundell & rawlings ( 2000 ) that the coherence length of the magnetic field is roughly 10 kpc in all directions , the inequality ( [ ineql ] ) yields @xmath368 for a magnetic field strength of 130 @xmath187 g , appropriate for the fiducial model of sect .
[ sec : degen ] , and an electron with a lorentz factor 1000 .
this implies that the time between collisions of this electron must be greater than 2 myr , i.e. the mean free path of the electron is of order 600 kpc .
this is clearly unphysical in the case of @xmath369 kpc as relativistic particles would then simply escape the cocoon very quickly .
the relativistic electrons in the cocoon must therefore diffuse along the magnetic field lines in between jumps from one patch of coherent magnetic field to another .
the expression of duffy et al .
( 1995 ) for the rms diffusion length after a time @xmath10 , @xmath344 , in this case can be approximated by assuming that the mean free path of the electron is less than the coherence length of the magnetic field , i.e. less than 10 kpc , i find for the same magnetic field and lorentz factor @xmath371 kpc if @xmath372 myr .
a significant mixing of relativistic particles along the cocoon due to anomalous diffusion is therefore unlikely .
of course , it may be argued that as the above analysis only applies to the quasi - linear regime , the diffusion timescale in a highly turbulent flow may be much shorter .
consider such a flow to be present in the cocoons of frii - type radio sources . in this case
, the relativistic electrons may diffuse through the observed lobes within a time short compared to the age of the source .
however , since diffusion is a stochastic process and the geometry of the cocoon is elongated , most of the particles will leave the cocoon sideways and will be lost to the surrounding gas before traveling large distances along the cocoon .
it is likely that the diffusion time for the relativistic particles depends on their energy and so the diffusion losses , if present , will significantly change the emission spectrum which is not observed ( e.g. roland et al
even in the case of efficient energy independent diffusion the observed radio lobes should show diffuse edges in low frequency radio maps .
again this is not observed ( e.g. roland et al .
1990 , blundell et al .
2000a , b ) . from the above it is clear that in the presence of very efficient diffusion some special confinement mechanism for the relativistic particles in the cocoon preventing their escape sideways is needed .
this may be provided for by the compression and shearing of the tangled magnetic field at the edges of the cocoon .
this process will align the magnetic field close to the cocoon edge with this surface and therefore act as a kind of magnetic bottle .
the order thus introduced in the originally tangled magnetic field due to this process leads to an enhanced polarisation of the emitted radiation in this region ( e.g. laing 1980 ) .
large volume compression ratios are ruled out as the sound speed in the cocoon is high ( e.g. ka ) .
however , for a conservative estimate consider a volume compression ratio of 10 in the sheet of compressed material at the edge of the cocoon of a given source .
this already implies that the maximum theoretical value of polarisation of the synchrotron emission should be observed at the edge of the radio lobes ( hughes & miller 1991 ) . for adiabatic compression of a tangled magnetic field the strength of the field increases by a factor of roughly 4.6 .
the rate of collisions of a given particle is probably increased as well as the irregularities in the magnetic field are also compressed .
however , the case of @xmath373 provides for a lower limit of @xmath374 and so with @xmath375 it is clear that the diffusion coefficient within the compressed region perpendicular to the field lines decreases by a factor 25 at most compared to the inner cocoon .
this implies that in a given time a relativistic particle diffuses a 5 times shorter distance in the compressed boundary layer compared to the inner regions of the cocoon . to significantly influence the time the particle remains in the cocoon this compressed layer must at the very least occupy 20% of the width of the cocoon which is unlikely .
note , that although the magnetic field is compressed in the boundary layer and therefore aligned with the cocoon surface some field lines may still be perpendicular to this surface . along these field lines diffusion will be even faster allowing many electrons to escape the cocoon .
from this i conclude that diffusion , even if it is highly effective in the inner cocoon and there exists a compressed boundary layer along the edges of the cocoon , will not alter the distribution of relativistic particles within the cocoon .
this allows us to use the spatial distribution of the synchrotron radio emission of frii sources to infer their age .
the model developed in the following can be viewed as an extension to the classical spectral aging methods in that it takes into account the evolution of the magnetic field in the lobe | a 3-dimensional model of the synchrotron emissivity of the cocoons of powerful extragalactic radio sources of type frii is constructed .
bulk backflow and energy losses of the relativistic electrons , radiative and adiabatic , are self - consistently taken into account .
thus the model is an extension of spectral aging methods including the underlying source dynamics into the age estimates .
the discrepancies between spectral ages and dynamical ages arising from earlier methods are resolved .
it is also shown that diffusion of relativistic particles within the cocoon is unlikely to significantly change the particle spectrum and thus the emitted radio spectrum .
projection along the line of sight yields 1 or 2-dimensional surface brightness distributions which can be compared with observations . from the model parameters constraints on the source age
, the density of the source environment and the angle to the line of sight can be derived .
application of the method to cygnus a , 3c 219 and 3c 215 show that the method provides robust estimates for the model parameters for sources with comparatively regularly shaped radio lobes .
the resolution of the radio maps required is only moderate . within the large uncertainties for the orientation angle ,
the three example sources are found to be consistent with orientation - based unification schemes of radio - loud agn . in the case of cygnus a the gas density of the environment is found to agree with independent x - ray measurements . for 3c 219 and 3c 215
the densities derived from the model are apparently too low .
it is suggested that these discrepancies are caused by overestimates of slope and core radius of @xmath0-models for the gas density distribution from x - ray observations for clusters hosting powerful radio sources . |
low mass x ray binaries ( lmxbs ) are binary systems in which a compact object either a neutron star or a black hole
accretes matter from a low mass companion star .
many of these systems are found to be transient , so called soft x
ray transients ( sxts ) .
several hundreds have been found in our own galaxy ( see ) and many of the x ray point sources in other galaxies are likely to be lmxbs as well ( @xcite ) . however , some of these x ray sources in other galaxies have luminosities in excess of the eddington limits of both neutron star and ten solar mass black hole ( bh ) lmxbs ( cf . ) .
these systems are thought to contain intermediate mass bhs ( @xcite ) or stellar mass bhs where the emission is anisotropic ( @xcite ) or indeed super eddington ( @xcite ) .
determining the distance to lmxbs is important , e.g. to determine the peak or quiescent luminosity and to determine if the compact object in lmxbs receives an ( asymmetric ) kick velocity at birth or not .
typically , a large distance to the galactic plane implies the occurrence of a kick , unless the systems at large z distances were formed in the halo . in this paper
we collect distance estimates from the literature and discuss possible systematic trends related to the used distance estimation methods ( section 2 ) . in section 3
we discuss the implications for the galactic distribution of ( bh ) lmxbs and the galactic population of ultra luminous x ray sources ( ulxs ) .
for most bh sxts it is not feasible to obtain a trigonometric parallax measurement . instead the distance generally is determined by comparing the derived absolute @xmath4band magnitude with the ( dereddened ) apparent magnitude , taking into account a possible contribution from residual accretion ( the distance derived using this method is sometimes called a photometric parallax ) .
a first guess of the distance can be obtained by assuming that the absolute magnitude is that of a main sequence star of the observed spectral type , after determining the best
fit spectral type from the data , e.g. via the optimal extraction technique ( see @xcite for a description of this technique ) .
we call this method @xmath5 . however , ideally , the radius , spectral type , and luminosity class are determined directly from the data .
we denote this method @xmath6 .
the observations give the orbital period , p , the radial velocity of the donor star , k@xmath7 and , by comparing the donor star spectrum with templates that are doppler broadened , the rotational velocity of the donor ( v@xmath8 ) .
the inclination can be estimated from modelling the ellipsoidal variations in the light curve . from kepler s third law and the assumptions that the donor fills its roche lobe and
is in co rotation with the orbit , all system parameters can be determined as follows .
the mass function and inclination relate the masses of the two components , while a combination of kepler s law and the equation for the volume of the roche lobe gives @xmath9^\frac{1}{3}$ ] ( e.g. @xcite ; q is defined here as the mass of the secondary divided by the mass of the primary ) , giving an independent estimate of q thus providing the system parameters .
the radius can be estimated from @xmath10 . alternatively ,
if @xmath11 or the inclination are not known , a quite good estimate of the radius can be obtained using the fact that the mean density of a roche lobe filling star depends only on the orbital period ( @xcite ) .
the radius of the donor star can be estimated via @xmath12 ( where @xmath13 is in hours , @xmath14 in solar masses , and @xmath15 in units of solar radii ) . in order to estimate the absolute magnitude one either uses the surface brightness for the observed spectral type or colour , as given e.g. by @xcite and or uses the determined radius and effective temperature , together with an appropriate bolometric correction .
alternatively , one uses the fact that for given surface brightness the observed flux scales with the angular diameter ( 2 r / d ) , which together with the radius gives the distance .
however , there are ( small ) differences in surface brightness for different luminosity classes of the same spectral type . in case of gro
j165540 and grs 1915 + 105 , @xcite and @xcite , respectively determined limits on the distance from the observed proper motion for receding and approaching blobs , assuming the jet ejections are intrinsically symmetric and noting that the maximum velocity of the ejections is the speed of light .
we list this method as method @xmath16 .
the distance can also be estimated using the interstellar absorption properties of the source .
there are different ways to do this . for several interstellar absorption lines and diffuse interstellar bands
the observed equivalent width correlates with colour excess ( cf . ) .
the colour excess can be converted in a distance estimate ( e.g. using the calibration of @xcite ) .
it is also possible to compare the observed extinction to that of ( ob ) stars in the observed field for which the distance is known or can be derived ( e.g. ) . the distance can also be constrained by using high resolution spectroscopic observations of the interstellar absorption lines to trace individual gas clouds and their velocities . assuming that the apparent velocities of the different gas clouds are projected galactic rotation velocities , a lower limit on the distance of the object can be found ( cf .
we refer to distances derived using the interstellar absorption properties as method @xmath17 .
finally , we note here that various other methods to estimate the distance to bh sxts have been proposed and used .
we do not discuss these in detail but we merely mention some of them : @xcite used the normalisation of a model describing the accretion disc flux , used the transition between the high / soft and low / hard x ray states , and @xcite proposed to use the normalisation of the radio x ray correlation in the low / hard state as a distance indicator . in table
[ disbhc ] we show the distance estimates in the literature based on method @xmath5 , @xmath6 , @xmath16 , and @xmath17 for sources for which the best estimate of the mass of the compact object is above the upper limit of the neutron star mass of @xmath18 m@xmath19 ( @xcite ; @xcite ) .
below , we briefly discuss the sources listed in table [ disbhc ] for which disparate values for the distance exist in the recent literature . [ cols="<,^,^,^,^,^,^,^ " , ] @xmath20 see the discussion and references in @xcite references : [ 1 ] @xcite ; [ 2 ] ; [ 3 ] @xcite ; [ 4 ] @xcite [ 5 ] @xcite ; [ 6 ] ; [ 7 ] @xcite ; [ 8 ] @xcite ; [ 9 ] ; [ 10 ] @xcite ; [ 11 ] @xcite ; [ 12 ] ; [ 13 ] @xcite ; [ 14 ] @xcite ; [ 15 ] hynes et al .
( 2004 ) ; [ 16 ] @xcite ; [ 17 ] @xcite ; [ 18 ] @xcite ; [ 19 ] @xcite ; [ 20 ] @xcite ; [ 21 ] ; [ 22 ] @xcite ; [ 23 ] @xcite ; [ 24 ] @xcite it has been suggested that the burst flux could be anisotropic ( ) . however , given the good agreement between the globular cluster distances and the distances to the lmxbs in these globular clusters derived from the photospheric radius expansion burst properties ( ) that seems unlikely .
furthermore , @xcite showed that considering the burst peak fluxes of 61 photospheric radius expansion bursts in the atoll source 4u 172834 the degree of anisotropy in the burst emission in less than 2 per cent .
two out of the eight neutron star systems studied by have a photospheric radius expansion peak burst luminosity that is lower than that of the other six .
this lower peak luminosity is consistent with the eddington luminosity limit for hydrogen rather then helium rich material for a neutron star mass of 1.4 m@xmath19 .
so , for some of the neutron star systems in our list we could have overestimated the distance by a factor @xmath21 .
if we take a 1:4 ratio as was found by this would affect 8 systems in our sample . as was noted by the two sources
whose peak luminosity is consistent with hydrogen accretion have an orbital period characteristic for normal lmxbs , rather than ultra compact x ray binaries with periods less than @xmath221 hour ( which necessarily accrete hydrogen poor material e.g. @xcite ) . because the fraction of ultra compact x ray binaries in globular clusters might be higher than in the field , the ratio between the number of sources with a hydrogen rich , low eddington burst luminosity and the number of sources with a helium rich
, high eddington burst luminosity could be higher .
it is unlikely , however , that the distances to _ all _ neutron star lmxbs in our sample have been overestimated .
e.g. in the long period lmxbs ks 173126 and 4u 163653 the photospheric radius expansion burst was consistent with a helium rich explosion ( @xcite , @xcite , respectively , see also @xcite ) .
using the data set compiled above , leaving the distances derived using method @xmath6 as they are , keeping the discrepancy in the distance estimate between method @xmath6 and the radius expansion burst method in mind , we investigate the galactic distribution of neutron star and bh lmxbs and the peak luminosity of bhs .
@xcite investigated the galactic z distribution of neutron star lmxbs and concluded that neutron stars should receive an asymmetric kick at birth from the fact that the rms value of the z distribution was @xmath23 kpc .
we obtain an rms z values of 1025 pc and 1125 pc for persistent and transient neutron star lmxbs , respectively ( we round rms z values to the nearest 25 pc ) . in this
we have excluded xte j2123058 since with a z value of -12.4 kpc it would dominate the outcome and @xcite show that the systemic velocity is consistent with it being a halo source . it can be argued on the basis of the large z values that cyg x2 and exo 0748676 are also halo sources although this is much less clear . e.g. @xcite argue that if cyg x2 is at a distance of @xmath2211.6 kpc it could have originated in the galactic plane . however , when we also exclude cyg x2 and exo 0748676 we find an rms z
value for the persistent and transient neutron star lmxbs of 700 pc and 850 pc , respectively .
@xcite found an rms z value of @xmath22500 pc when they excluded cyg x2 and exo 0748676 from their sample .
since we have used 3.8@xmath24 erg s@xmath25 for the eddington luminosity for the radius expansion peak luminosity for all neutron stars , the rms z value we derived for neutron star lmxbs corresponds to an upper limit .
if we use 2.0@xmath24 erg s@xmath25 as the photospheric radius expansion peak burst luminosity we find an rms z value of 775 pc and 850 pc for persistent and transient neutron star lmxbs , respectively ( here we only excluded xte j2123085 , if we also exclude cyg x2 and exo 0748676 we find 550 and 650 pc for persistent and transient neutron star lmxbs , respectively ) .
i.e. our findings confirm the result of @xcite . in a follow up paper , @xcite investigated the differences between the rms values of the neutron star and bh lmxbs z distributions .
they found that the rms value for the bhs was substantially lower than that of the neutron star z distribution ( more than a factor 2 ) . using the distances of the bhs given in table [ disbhc ]
we now find an rms value of @xmath22625 pc ( we took a distance of 2.5 kpc for gro j0422 + 32 ; we excluded the likely halo object xte j1118 + 480 ; @xcite [ although note that some of the evidence leading to the suggestion that xte j1118 + 480 is a halo object was based on the low rms z value @xcite found for bh lmxbs ] ) , i.e. close to the upper limit we find for the neutron star systems .
increasing the distance of xte j1550564 with a factor 3 does not significantly increase the rms value of the bh population .
the main reason for the difference between the findings of @xcite and our findings is that the distance estimates for most bhs have gone up .
the conclusion drawn by @xcite that bhs receive a significantly smaller kick than neutron stars is no longer tenable .
we plotted the z values for transient and persistent neutron star lmxbs ( open diamonds and squares , respectively ) and bhs ( black dots ) using the distances from table [ disbhc ] and [ nsdist ] as a function of their projected distance to the galactic centre in figure [ yztot ] .
it is interesting that the rms value for the projected distance to the galactic centre for the neutron stars and bh is 4.8 kpc and 7.0 kpc , respectively ( excluding the [ likely ] halo sources ) . because the galactic potential in the z direction decreases with increasing galactocentric radius ( e.g. @xcite ) the larger scale height of bhs
can partly be due to this effect , rather than a kick velocity .
e.g. for neutron star systems @xcite estimate an rms value of the z distribution near the sun of 650 pc .
furthermore , the symmetric kick velocity ( imparted due to mass loss in the supernova ) scales with the companion mass and the mass lost in the supernova ( e.g. ) both can be ( much ) larger in the case of bh systems , but it scales inversely with the total mass of the remaining binary .
lastly , for large asymmetric kick velocities the binary is more likely disrupted in case of a neutron star than a bh .
a detailed investigation of all the possible explanations for the high rms value for the bhs is needed in order to draw firm conclusions and we defer this to a later paper . , where @xmath26 and @xmath27 are the cartesian coordinates in the galactic plane ( see figure [ xytot ] ) .
the location of the sun at an assumed distance of 8.5 kpc is indicated with an open circle , and the location of the galactic centre is indicated with a cross .
four neutron star systems fall off the figure ( exo 0748676 , z=-3.1 kpc , xte j2123058 , z=-12.4 kpc , xte j1710281 , [email protected] kpc , and cyg x2 , d@xmath2815 kpc ) . in this figure
we use the neutron star distances derived assuming the eddington peak burst luminosity was 3.8@xmath24 erg s@xmath25.,width=302 ] after @xcite we also compare the distribution of the l and b coordinates of the neutron star and bh lmxbs .
we first compare the l and b coordinates of neutron star lmxbs in table [ nsdist ] with those of the bhs listed in table [ disbhc ] . a k s test shows that the probability that the distributions are the same is 18 per cent for the l coordinate and 52 per cent for the b coordinate ( the k
s d value is 0.33 and 0.24 , respectively . here and below : the effective number of data point is always larger than 10 so that the probabilities we quote are quite accurate , cf .
@xcite ) . to investigate this further and to minimise selection effects
, we plot the l and b coordinates of _ all _ neutron star systems , e.g. systems where a burst was found , and pulsars listed in ( solid line histogram in figure [ ldist ] ) .
we compare this with the bh sources listed in table [ disbhc ] _ and _ the bh candidates for which no mass estimate based on a radial velocity study exist ( systems in table 3 in @xcite ; the dashed line histogram in figure [ ldist ] ) .
a one dimensional k s test shows that the probability that the neutron star and bh distributions in l are drawn from the same distribution is 37 per cent whereas the probability that b values are drawn from the same distribution is 90 per cent ( the k
s d value is 0.19 and 0.12 , respectively ) .
hence , using a larger sample of neutron stars and bhs makes the probability that the neutron star and bh distributions are the same larger .
the probabilities we derive are higher than those derived by @xcite ; a difference we attribute to the fact that @xcite had fewer bh systems in their study .
we conclude that there is no evidence for a difference in the l and b distributions of neutron star and bh lmxbs .
interestingly , there seems to be an excess of 12 out of a total of 22 bh and an excess of 19 out of a total of 45 neutron star systems with -30@xmath29 over systems with 0@xmath3 .
there is an excess of 5 out of 35 bh and 13 out of 53 neutron star systems with -10@xmath30 and 0@xmath31 ( see figure [ ldist ] ) .
the expected mean number of systems with a positive / negative l of a symmetric distribution around l=0@xmath32 is @xmath33 , with @xmath34 the total number of systems .
the variance on this number is @xmath35 . comparing this with the observed number of systems yields significances of the observed asymmetries of 2.8@xmath36 for the neutron star distribution in l , 3.0@xmath36 for the bh distribution in l , and 3.8@xmath36 for the combined neutron star and bh distribution in l. the combined significance for the neutron star and bh distribution in b is 1.9@xmath36 only , so we do not discuss this apparent asymmetry in b further .
the asymmetry in the l
distribution is not an obvious observational selection effect .
for instance , the bepposax wfc monitoring the galactic centre region had a field of view of @xmath37 centred on the galactic centre ( @xcite ) .
the dust maps from @xcite do not show a large dust asymmetry around the galactic centre .
it seems that the neutron star and bh lmxb distribution in l is symmetric around l=-5@xmath32 .
possibly this asymmetry is related to the bar structure in the inner part of our galaxy .
it is known that the bar causes asymmetries in the stellar and gas distributions around ( l , b ) = ( 0,0 ) ( e.g. see the recent review by @xcite ) .
furthermore , @xcite found that the distribution of lmxbs follows the stellar mass distribution , and hence not the spiral arm structure .
using the distances listed in table [ disbhc ] and [ nsdist ] we also plot in figure [ xytot ] the galactic x y distribution of the neutron stars for which a photospheric radius expansion burst was observed and the bhs for which a dynamical mass estimate has been derived .
the galactic spiral arm structure according to the model described in @xcite has been overplotted .
it is clear that few lmxbs at the other side of the galactic centre have been detected , especially if the eddington limit for hydrogen rich material for a 1.4 m@xmath19 neutron star is applicable to the peak burst flux of most photospheric radius expansion bursts ( see also ) .
there also seems to be a paucity of nearby neutron star lxmbs ( this was also noted by @xcite ) . in order to try to quantify this we performed a two dimensional kolmogorov
smirnov ( k s ) test ( @xcite ) to test whether the neutron star and bh populations are the same .
the k s test gives a d of 0.45 and a probability that the two populations are the same of 3.8 per cent .
if we decrease the distance of all neutron star systems by a factor @xmath38 the probability that the distributions are the same increases only slightly to 5.2 per cent ( d=0.44 ) .
it is possible that the discrepancy in the spatial distribution of the neutron stars and bhs is a selection effect .
since we only included bhs for which a mass limit larger than 3@xmath39 has been derived , the optical counterpart must have been detected in quiescence , hence this favours nearby systems .
the bh distances are important for the maximum observed bh luminosity .
we found a somewhat larger distance for v404 cyg than often used previously .
@xcite found an upper limit on the distance of v404 cyg of 3.7 kpc assuming the peak outburst source luminosity was eddington limited taking their 90 per cent confidence upper limit to the mass of the bh of 15 m@xmath19 .
the distance of 4.0@xmath0 kpc we find shows that the maximum observed luminosity exceeds the eddington luminosity for a 10 m@xmath19 bh ( the limit often quoted to decide whether a source is a ulx or not ) .
the distance derived for xte j1550564 may have been underestimated by as much as a factor 3 : its distance could be as large as 3@xmath40 kpc , although the uncertainties are large .
for this distance the outburst peak luminosity in the 220 kev band alone would be 7@xmath41 erg @xmath42 s@xmath25 ( taking the flux from @xcite ) . for 4u 154347 , sax j1819.32525 and grs 1915 + 105 , systems for which the distance was determined by method @xmath6 , @xmath6 , and @xmath16 , respectively , it was noticed earlier that the outburst peak luminosities are super
eddington for the best
fit bh masses ( cf . ;
@xcite show that the peak outburst of gro j100945 most likely also exceeded the eddington luminosity for a for @xmath43 bh ( even if the distance is 5.7 instead of the 9 kpc they favoured ) .
it seems therefore likely that there are several sources in our own galaxy which we would classify as transient ulxs had we observed them in other galaxies . for these sources
a mass determination has shown that they are not intermediate mass bhs but rather 515 m@xmath19 bhs .
furthermore , the fact that lmxb sources ( an old population ) seem to be capable of producing super eddington luminosities could also help explain the presence of ulxs in elliptical galaxies .
assuming that most ulxs are stellar mass bhs the fact that many ulxs are found a few arc seconds away from the young stellar clusters could be caused by kick velocities received at bh formation .
if super eddington luminosities were to be explained by effects of `` mild beaming '' ( cf .
@xcite ) then one would naively expect the inclinations of the sources with the highest luminosities to be lowest . in figure [ incli ]
we plot the inclination distribution for 13 bh sxts in figure [ incli ] ( the inclinations for gx 3394 and xte j1859 + 226 are not well known and those sources have therefore not been included ; inclinations from @xcite ) .
although the amount of sources is low we see that the inclination distribution of the bh ulx sources is not skewed to low inclinations but instead clusters around the fiducial 60@xmath32 point .
we did not include xte j1550564 ( @xmath44 ) and gro j0422 + 32 ( @xmath45 ) as ulxs .
we studied the distance estimates for bh sxts in detail .
we note that the uncertainties in the distance estimates due to e.g. uncertainties in the spectral type ( i.e. temperature ) of the companion star are probably large . comparing the distances derived for the neutron stars cyg x2 and xte j2123058 using the distance estimation method @xmath6 , which is used for most bh sxts , with the photospheric radius expansion burst method , we find that the latter gives larger distances .
this could mean that for some reason method @xmath6 systematically underestimates the distance .
possibly this is related to an erroneous spectral classification of the companion star caused by its fast rotation .
if this is indeed the case this would have important consequences for the reported difference in quiescent x ray luminosities of bh and neutron star sxts , for the bh sxt peak outburst luminosities , for the bh masses , and for the galactic distribution of bh lmxbs .
we find that the distance towards xte j1550564 may have been underestimated by as much as a factor 3 because the interstellar extinction could have been overestimated in that case .
as was noticed before , several bh sxts observed in our galaxy would be classified as ulxs had we observed them in another galaxy ( at least 5 but perhaps even 7 out of the 15 dynamically confirmed bh sxts ) .
this suggests that many ( transient ) ulxs in other galaxies could well be stellar mass bhs .
a re evaluation of the distance to the galactic plane of neutron star and bh lmxbs shows that there is no longer evidence for a smaller rms value of the z distribution for bh systems .
such a difference had been interpreted as evidence for the absence of asymmetric kicks during bh formation . however , before firm conclusions can be drawn about the similarities or differences between neutron stars and bh kicks the details of the formation and galactic distribution have to be investigated .
finally , we found that the l distribution of galactic lmxbs is asymmetric around l=0@xmath32 .
support for this work was provided by nasa through chandra postdoctoral fellowship grant number pf340027 awarded by the chandra x ray center , which is operated by the smithsonian astrophysical observatory for nasa under contract nas839073 .
gn is supported by pparc .
the authors would like to thank the referee for his / her valuable comments which helped improve the manuscript considerably .
we would like to thank hans jakob grimm , jeff mcclintock , jim pringle , and frank verbunt for useful discussions and duncan galloway for sharing results before publication .
this research made use of results provided by the asm / rxte teams at mit and at the rxte sof and gof at nasa s gsfc .
the research has made extensive use of nasa s astrophysics data system . , j. , 2001 , in exploring the gamma - ray universe .
proceedings of the fourth integral workshop , 4 - 8 september 2000 , alicante , spain .
editor : b. battrick , scientific editors : a. gimenez , v. reglero & c. winkler .
esa sp-459 , noordwijk : esa publications division , isbn 92 - 9092 - 677 - 5 , 2001 , 459 , 463 , j. e. , remillard , r. a. , 2004 , in x - ray binaries , cambridge astrophysics series , cambridge , ma : cambridge university press , c1995 , edited by lewin , walter h.g . ; van paradijs , jan ; van den heuvel , edward p.j . | we investigated the reported distances of galactic black hole ( bh ) and neutron star low mass x
ray binaries ( lmxbs ) . comparing the distances derived for the neutron stars cyg x2 and xte j2123058 using the observed eddington limited photospheric radius expansion bursts with the distances derived using the observed radius and effective temperature of the companion star we find that the latter are smaller by approximately a factor of 1.52 .
the latter method is often employed to determine the distance to bh lmxbs .
a possible explanation for this discrepancy is that the stellar absorption lines in fast rotating companion stars are different from those in the slowly rotating template stars as was found before for early type stars .
this could lead to a systematic mis
classification of the spectral type of the companion star , which in turn would yield a systematic error in the distance .
further , we derive a distance of 4.0@xmath0 kpc for v404 cyg , using parameters available in the literature . the interstellar extinction seems to have been overestimated for xte j1550564 and possibly for two other bh sources ( h 170525 , and gs 2000 + 25 ) as well . as a result hereof the distance to xte j1550564 may have been underestimated by as much as a factor three .
we find that , using the new distances for xte j1550564 and v404 cyg , the maximum outburst luminosity for at least 5 but perhaps even 7 of the 15 bh soft x
ray transients exceeds the eddington luminosity for a @xmath1 bh showing that these systems would be classified as ultra luminous x ray sources had we observed them in other galaxies .
this renders support for the idea that many ultra luminous x ray sources are stellar
mass rather than intermediate
mass bhs .
we find that the rms value of the distance to the galactic plane for bhs is consistent with that of neutron star lmxbs .
this suggests that bhs could also receive a kick velocity during their formation although this has to be investigated in more detail .
we find that the galactic neutron star and bh l and b distributions are consistent with being the same .
the neutron star and bh distribution is asymmetric in l with an excess of systems between -30@xmath2 over systems with 0@xmath3 .
stars : stars : black holes stars : neutron stars x - rays : stars |
transition disks trace a key step in the formation of planetary systems , intermediate between gas - rich protoplanetary disks and debris disks , where primordial gas is cleared away , leaving only remnant dust .
these disks are observationally identified by weak mid - ir emission ( at @xmath115 @xmath5 m ) relative to the taurus median spectral energy distribution ( najita et al .
2007 , i.e. the median sed of primordial disks in the young ( @xmath62 myr ) taurus star - forming region ) .
while numerous physical processes may be responsible for the depletion of gas and dust in transition disks , cleared gaps in particular may be an indicator of a planet or brown dwarf companion in the midst of formation .
thus , these disk systems have been key targets for direct imaging searches for planets .
transition disks tend to be found in young star - forming regions @xmath7100 pc from earth .
only recently have imaging techniques been available to probe the inner portions of these disks , where planets are likely to form . in the last few years
, there have been a number of notable discoveries of companions in transition disks with cleared gaps . via the interferometric sparse - aperture masking
( henceforth sam ) technique @xcite , @xcite found lkca 15b , a candidate protoplanet embedded within the disk of a @xmath12 myr solar analog .
@xcite detected a 0.1 - 0.4 m@xmath8 companion in the disk around the herbig ae / be star hd 142527 .
this binary companion was recently confirmed and shown to be accreting via direct imaging @xcite .
hd 142527 has a particularly wide gap in its disk ( @xmath9 - 120 au , fukugawa et al .
2006 , rameau et al .
2012 , casassus et al .
2012 ) , which may be explained by a stellar companion on an eccentric orbit .
finally , @xcite reported an as yet unconfirmed candidate protoplanet detection in the disk of the herbig ae / be star hd 100546 .
the herbig ae / be star hd 169142 possesses a nearly face - on transition disk , and has been studied in detail both spectroscopically ( meeus et al .
2001 , van boekel et al . 2005 ) , and through imaging ( habart et al . 2006 , kuhn et al .
2001 , hales et al .
2006 , grady et al .
2007 , fukugawa et al .
2010 , honda et al .
2012 , marias et al . 2011 ) .
recently , @xcite detected a well - resolved annular gap from 40 - 70 au via polarized light imaging .
thus , like lkca 15 , hd 142527 , and hd 100546 , it is an excellent candidate to possess a planetary mass or brown dwarf companion in the midst of formation .
adopted stellar parameters are presented in table [ tab : properties ] .
in general , we adopt the stellar parameters from table 1 of quanz et al . 2013b . an accurate age estimate is particularly important for estimating candidate companion properties ; hd 169142 has been assigned a fairly wide range of ages .
guimares et al .
2006 assign a comparatively young age of 1 - 5 myr based in hd 169142 s hr diagram position , while blondel @xmath10 djie 2006 claim a similar mass and age as @xmath11 pic for hd 169142 .
grady et al .
2007 find a comoving companion 9.3 separated from hd 169142 .
this object is a 130 mas separation weak - line t tauri binary , thus they assign an age of 6@xmath12 myr for the entire system . for our analysis here , we thus adopt an age range of 3 - 12 myr .
sylvester et al .
1996 derive a photometric distance of 145 pc to hd 169142 , based on an a5 spectral type ; blondell & djie 2006 update this to 151 pc based on a spectral type of a7v .
as derived spectral type will strongly affect the photometric distance , we adopt a distance range of 145@xmath1310 pc here to account for any uncertainty in spectral typing .
first epoch observations were conducted using the novel annular groove phase mask ( henceforth agpm ) vector vortex coronagraph @xcite with the naco camera at the vlt @xcite .
the agpm coronagraph uses an annular groove phase mask to redirect on - axis starlight out of the pupil @xcite . in this manner
, the agpm coronagraph enables l band contrasts of @[email protected] mag at inner working angles down to 0.09@xmath2 @xcite .
hd 169142 was observed in l from 02:47 ut to 04:54 ut on 14 july 2013 , covering nearly an hour both before and after transit .
the derotator was turned off to enable azimuthal differential imaging ( adi ) techniques .
seeing varied from 1 - 2@xmath2 over the observation .
we used a base exposure time of 0.25 s with 120 coadds and obtained a total of 1 hour 15 minutes on - sky exposure time .
sky frames were interspersed after every 20 frames and 0.05 s unsaturated exposures were taken at the beginning of the sequence .
care was taken to keep the primary star centered under the agpm .
observation details are presented in table [ tab : observations ] .
data were flat - fielded and bad - pixel corrected using dome flats and darks taken as part of eso standard calibrations .
numerous dust spots are apparent on the agpm , thus , slight misalignments between science and sky frames can produce significant cosmetic errors .
thus , we carefully selected the sky frame which minimized cosmetic errors for each science frame .
eleven science frames ( out of 126 total science frames ) which showed significant cosmetic errors even after sky subtraction were discarded .
each frame was centroided on the raw data using the idl routine mpfit2dpeak . after basic data reduction and centroiding on the center of the coronagraphic mask using the idl mpfit2dpeak routine
, we analyze the data using 3 independent principal component analysis ( pca ) pipelines ( following the algorithms of soummer et al .
2012 , amara & quanz 2012 ) .
all three pipelines yielded similar results
we report here the results using the publically available pipeline of dimitri mawet ( http://www.sc.eso.org/@xmath1dmawet/dimitrimawet/idl_pca_pipeline.html ) .
followup observations were conducted using the 585 actuator 1000 hz adaptive secondary ao system ( magao ) at the 6.5 m magellan clay telescope ( close et al .
2013 ) . with 378 corrected modes ,
the magao system is one of the highest sampled ao systems on a large telescope ( @xmath75 m ) and has demonstrated high spatial resolutions ( 20 - 30 mas ) in the visible ( as short as @xmath14 = 0.6 @xmath5 m , close et al .
magao feeds both a visible ( visao , 0.5 - 1 @xmath5 m ) and infrared camera ( clio2 , 1 - 5 @xmath5 m ) simultaneously .
we obtained a total of 5 datasets for hd 169142 with magao from 8 - 15 april 2014 .
observing details are presented in table [ tab : observations ] .
four of these datasets were obtained with the clio2 narrow camera including two h band observations , one k@xmath4 band observation , and one 3.9 @xmath5 m 5@xmath15 filter observation .
we also obtained a deep z dataset using the visao camera .
the derotator was turned off to enable angular differential imaging ( adi ) techniques .
all magao data was reduced in the standard manner ( flat - field correction , sky - subtraction , bad - pixel correction ) .
the h , k@xmath4 , and z datasets remained unsaturated in the core , thus we align our images using the unsaturated core .
we then process our data using the independent pca - based pipelines of cois rodigas and males ( see e.g. rodigas et al .
2014 and males et al .
both pipelines yielded similar results . at 3.9 @xmath5 m , we acquired data for hd 169142 as well as for a psf star of similar magnitude .
we achieved very high strehl ratios for both stars in dry , photometric conditions .
seeing remained around 0.4 - 0.6@xmath2 for the entire night .
after image registration , we built a psf from the nearby star images and subtracted the psf from each of our magao / clio2 images of hd 169142 , then rotated to place north up / east left .
our best subtraction is for a psf scaling of 115@xmath15 , which removes the airy rings quite well .
dark pixels in the core are due to saturation / non - linearity .
in july 2013 , we detected a faint , pointlike feature in our l vortex dataset at @xmath10@xmath3 pa , with a separation of @xmath10.11 " .
this pointlike feature was independently imaged in june 2013 by reggiani et al .
multiband followup observations were conducted using magellan ao in april 2014 .
images from all epochs are presented in figs .
[ fig : image1 ] and [ fig : image2 ] .
astrometry and photometry are presented in table [ tab : properties ] . the correct rotation
to place north up and east left was verified at each epoch . for the july 2013 vortex data ,
a dataset on hd 142527 was acquired on the same run . the bright hd 142527 disk ( see e.g. fukagawa et al . 2006 , rameau et al .
2012 ) is correctly derotated by the pipeline . for the april 2014 magao data ,
the correct derotation has been validated using images of the trapezium cluster .
l photometry and astrometry was derived for the pointlike feature by inserting scaled psf images into the raw data and selecting the fluxes and positions that best match the actual feature .
best astrometry and photometry is presented in table [ tab : properties ] .
assuming a point source , our best subtraction yielded @[email protected] for this feature , at a separation of [email protected]@xmath2 and pa of 0@xmath1314@xmath3 . given that the inner working angle of the vortex coronagraph is 0.09@xmath2 @xcite , this pointlike feature is marginally resolved at best . to estimate the s / n level of our detection , we calculated the mean and standard deviation in a circular aperture with diameter of @xmath14/d centered on the pointlike feature and in the 6 additional resolution elements available at this separation and then calculated s / n from equation 9 of @xcite ( appropriate for speckle - dominated regions ) : @xmath16 where @xmath17 is the mean within the aperture containing the feature , @xmath18 is the mean of the remaining 6 resolution elements , @xmath19 is the standard deviation of the remaining 6 resolution elements , and @xmath20 is the number of resolution elements not containing the feature .
we find s / n@xmath16 for the detected feature , thus this detection is unlikely to be a false positive a result which is further bolstered by the independent detection of the same feature by reggiani et al .
initially assuming purely photospheric emission , we estimate mass using monte carlo methods to account for the range of possible ages and uncertainties in photometry for this system .
an ensemble of 10@xmath21 possible ages are drawn from a uniform distribution running from 3 - 12 myr .
an ensemble of 10@xmath21 absolute magnitudes are simulated , assuming gaussian errors on photometry and distance ( 0.3 mag error in @xmath0mag , 10 pc error in distance ) .
we then interpolate with age and single band absolute magnitude to find the best mass for the companion from the dusty and cond models of @xcite and @xcite .
mass range histograms are presented in fig .
2 . adopting a 3 - 12 myr age range at a distance of 145 pc and the dusty models
, this candidate would correspond to a relatively high mass 60 - 80 m@xmath22 brown dwarf .
no counterpart to the l detection was found in the april 2014 magao followup .
our non - coronagraphic 3.9 @xmath5 m followup imaging was too shallow to retrieve the pointlike feature ( @xmath0mag=5.6 at 0.11 ) .
however , a 60 - 80 m@xmath22 object at these ages should have been quite bright in the near - ir ( h band absolute magnitude @xmath67 ) and have been easily detected given our h and k@xmath23 sensitivity . indeed , our sensitivity is sufficient to detect a @xmath17 - 15 m@xmath22 object at this separation and age range .
such a high - mass companion should open up a very wide gap in the disk , inconsistent with the ring at similar radii observed polarimetrically by quanz et al .
the lack of a near - ir counterpart to the l detection signifies that we are not observing the photosphere of a substellar object here but may be a potential disk feature its extremely red colors indicate that it is extraordinarily unlikely to be a background . we performed radiative transfer calculations to investigate whether or not a passively heated disk feature can reproduce the observed point - like source .
we used the best fit disk model parameters from maaskant et al .
2013 to describe an axisymmetric disk . then to simulate a clump we increased the pressure scale height in the disk .
the scale height was increased as a 2d gaussian centered on the observed point source with an fwhm of 7au and a peak perturbation of 30% .
we used 0.1@xmath5 m size astronomical silicates for the dust opacity to get an upper limit on the scattered light perturbations .
larger grains would scatter preferentially forward , perpendicular to the line of sight , which would decrease the contrast between the clump and the central star .
then we measured the contrast on the calculated l images .
we find that a `` dust clump '' at the position of our candidate would have a contrast of @xmath117 mag ( 1@xmath2410@xmath25 ) between the star and the clump , much fainter than the detected candidate .
this result is not surprising , given the distance of the pointlike feature from the star .
assuming we observe only thermal dust emission from an irradiated clump , if the clump peaks at l , it would have t@xmath1850 k. thus , adopting this as the maximum temperature of a potential dust clump , an irradiated clump with this temperature would be at a distance of the order of 0.6 au ( considering temperature and radius of the primary from quanz et al .
2013b ) , inconsistent with the @xmath710 au separation of this detection .
tentatively , we consider this candidate to be a disk feature heated by an as yet unknown mechanism .
we also marginally detected a candidate companion in h and ks with a separation of @xmath10.18 and a pa of @xmath133@xmath3 , but at low significance ( s / n=3 - 5 , using the same procedure as for the pointlike l feature ) .
these data were taken at two different nods , with very different looking psfs in the two nods , limiting the precision of our photometry to @xmath130.5 mag .
thus , while this point source was detected in two different bands on 3 nights ( with s / n@xmath15 in the first dataset , but only s / n@xmath13 in the additional datasets ) , it requires confirmation at an additional epoch , with a more stable psf , to verify common proper motion and rule out imaging and reduction artifacts .
if real and purely photospheric , this candidate would be a 8 - 15 m@xmath22 planet / substellar object ( dusty and cond models ) .
no corresponding l counterpart was found in the july 2013 l vortex dataset .
if purely photospheric , the l counterpart to this object is expected to have an absolute magnitude @xmath78.5 mag ( dusty or cond models ) , corresponding to an apparent magnitude of @xmath714.3 mag .
this is considerably below our achieved contrast for this dataset at this separation ( limiting magnitude of @xmath112.7 , from insertion and retrieval of psf images into the raw data ) .
we also do not see a counterpart to this point source in our deep magao / visao z dataset , down to a contrast of 10@xmath26 .
if real , the very red colors of this object are inconsistent with background stars with spectral types earlier than mid to late m.
the known geometry of this disk limits the potential masses of companions associated with imaged asymmetries .
quanz et al .
2013b image an inner dust ring peaking at 25 au and an annular gap from 40 - 70 au .
the l detection lies within the 25 au dust ring and may be connected to ongoing companion formation in the disk .
if associated with a massive object , it may have had some role in clearing out the region inside this ring .
our tentative h / k detection does not lie within the 40 - 70 au gap but is coincident with the 25 au dust ring .
this suggests that any substellar / planetary counterpart to this tentative detection can not be very massive , as a massive object ( @xmath270.55 m@xmath22 should completely clear away dust from its immediate environment ( kley @xmath10 nelson , 2012 ) .
if massive objects are associated with these detections , they would be some of the closest companions imaged to date . among
confirmed directly imaged exoplanets , only @xmath11 pic b , hr 8799d , and hr 8799e have comparably small projected separations ( lagrange et al .
2010 , marois et al . 2008 , 2010 ) .
some disk asymmetries detected via sparse aperture masking ( sam ) have been described due to forward scattering off an inclined disk ( e.g. cieza et al .
2013 , olofsson et al .
2013 ) . however , hd 169142 is nearly face on ( quanz et al .
2013b ) , so forward scattering can not have produced any observable structure in this case .
our l detection is also not well - described by dust features heated only by the central star ; it is simply too far from the star to be heated enough to produce the observed emission .
however , bright disk - related features have been observed in other transition disks .
for instance , kraus et al .
2013 detect bright asymmetries via sparse aperture masking in h , k , and l band within the 15 - 40 au inner gap in the pre - transitional disk v1247 orionis .
these asymmetries are not well - fit by a companion model ; these authors attribute these features to spiral features or accretion streams due to the interaction of the dust disk with the substellar bodies inside the gap responsible for its clearing .
kraus & ireland 2012 detect a protoplanet candidate ( also via sparse aperture masking ) around lkca 15 . while their detection is well fit by a single point source at k , the l detection appears extended and not entirely coincident with the k detection .
they interpret the k detection as the protoplanet itself and the l detection as circumplanetary material potentially shocked and heated by jets from the protoplanet . in the case of hd 169142
, the possible energy source for the observed l disk feature remains unknown .
as previously noted , the lack of a h / k counterpart rules out photospheric emission and the l detection is too strong to be described by passive heating from the star .
if due to jets or accretion onto a forming planetary companion , we would expect a h@xmath28 counterpart to the l detection ( close et al .
. however , follette et al . in prep have observed hd 169142 with magao in h@xmath28 and do not detect an h@xmath28 counterpart to the l disk feature .
the l detection may be related to emission line regions inside the hd 169142 disk .
habart et al .
2006 found strong 3.3 @xmath5 m pah emission ( from a c - h stretching feature of neutral pahs ) in the hd 169142 disk out to 0.3 ( using high - resolution ao - supported naco long - slit spectroscopy and along a slit oriented north to south , i.e. intersecting our detection ) .
maaskant et al .
2014 also find strong pah emission ( however , primarily in ionized features ) in the hd 169142 disk , likely stemming from the inner disk gap .
a weak blue leak ( on order 1@xmath15 ) in the naco l filter might allow us sensitivity to pah features in the disk .
if our l feature is indeed due to pah emission , future 3.3 @xmath5 m follow - up observations should easily retrieve this feature .
we report the detection of a faint pointlike structure in the hd 169142 transition disk , also independently detected by reggiani et al . 2014 .
this structure has @xmath0mag(l)@xmath16.4 , at a separation of @xmath10.11@xmath2 and pa of @xmath10@xmath3 .
given its lack of an h or k@xmath4 counterpart despite its relative brightness , this object can not be due to the photosphere of a substellar or planetary mass companion and must instead be a disk feature .
however , the observed l detection is too strong to be described by passive heating from the star and ongoing accretion is ruled out by the magao h@xmath28 non - detection of follette et al . in prep .
while the location of the l pointlike feature right within the disk gap strongly suggests it may be connected to ongoing planet - formation ( i.e. whatever process cleared out the gap ) , the energy source fueling this feature without producing a corresponding near - ir counterpart still remains mysterious .
hopefully , future multi - wavelength followup observations will elucidate the source of this feature and perhaps in the process refine our understanding of ongoing planet - formation in this disk .
amara , a. , & quanz , s. p. 2012
, , 427 , 948 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 2002 , , 382 , 563 biller , b. , lacour , s. , juhsz , a. , et al .
2012 , , 753 , l38 blondel , p. f. c. , & djie , h. r. e. t. a. 2006 , , 456 , 1045 casassus , s. , perez , m. , s. , jordn , a. , et al .
2012 , , 754 , l31 chabrier , g. , baraffe , i. , allard , f. , & hauschildt , p. 2000 , , 542 ,
464 cieza , l. a. , lacour , s. , schreiber , m. r. , et al .
2013 , , 762 , l12 close , l. m. , males , j. r. , morzinski , k. , et al . 2013 , , 774 , 94 close , l. m. , follette , k. b. , males , j. r. , et al .
2014 , , 781 , l30 fukagawa , m. , tamura , m. , itoh , y. , et al .
2006 , , 636 , l153 fukagawa , m. , tamura , m. , itoh , y. , et al .
2010 , , 62 , 347 grady , c. a. , schneider , g. , hamaguchi , k. , et al .
2007 , , 665 , 1391 guimares , m. m. , alencar , s. h. p. , corradi , w. j. b. , & vieira , s. l. a. 2006 , , 457 , 581 habart , e. , natta , a. , testi , l. , & carbillet , m. 2006 , , 449 , 1067 hales , a. s. , gledhill , t. m. , barlow , m. j. , & lowe , k. t. e. 2006 , , 365 , 1348 hg , e. , fabricius , c. , makarov , v. v. , et al .
2000 , , 357 , 367 honda , m. , maaskant , k. , okamoto , y. k. , et al .
2012 , , 752 , 143 kley , w. , & nelson , r. p. 2012
, , 50 , 211 kraus , a. l. , & ireland , m. j. 2012 , , 745 , 5 kraus , s. , ireland , m. j. , sitko , m. l. , et al . 2013 , , 768 , 80 kuhn , j. r. , potter , d. , & parise , b. 2001 , , 553 , l189 lacour , s. , tuthill , p. , amico , p. , et al .
2011 , , 532 , a72 lagrange , a .- m . ,
bonnefoy , m. , chauvin , g. , et al .
2010 , science , 329 , 57 lenzen , r. , hartung , m. , brandner , w. , et al . 2003 , , 4841 , 944 maaskant , k. m. , honda , m. , waters , l. b. f. m. , et al .
2013 , , 555 , a64 maaskant , k. m. , min , m. , waters , l. b. f. m. , & tielens , a. g. g. m. 2014 , , 563 , a78 males , j. r. , close , l. m. , morzinski , k. m. , et al .
2014 , , 786 , 32 malfait , k. , bogaert , e. , & waelkens , c. 1998 , , 331 , 211 marias , n. , telesco , c. m. , fisher , r. s. , & packham , c. 2011 , , 737 , 57 marois , c. , macintosh , b. , barman , t. , et al .
2008 , science , 322 , 1348 marois , c. , zuckerman , b. , konopacky , q. m. , macintosh , b. , & barman , t. 2010 , , 468 , 1080 mawet , d. , riaud , p. , absil , o. , & surdej , j. 2005 , , 633 , 1191 mawet , d. , absil , o. , delacroix , c. , et al .
2013 , , 552 , l13 mawet , d. , milli , j. , wahhaj , z. , et al .
2014 , accepted to apj , arxiv:1407.2247 meeus , g. , waters , l. b. f. m. , bouwman , j. , et al .
2001 , , 365 , 476 najita , j. r. , strom , s. e. , & muzerolle , j. 2007 , , 378 , 369 olofsson , j. , benisty , m. , le bouquin , j .- b .
, et al .
2013 , , 552 , a4 rameau , j. , chauvin , g. , lagrange , a .- m . , et al .
2012 , , 546 , a24 rodigas , t. j. , debes , j. h. , hinz , p. m. , et al .
2014 , , 783 , 21 rousset , g. , lacombe , f. , puget , p. , et al . 2003 , , 4839 , 140 quanz , s. p. , amara , a. , meyer , m. r. , et al . 2013 , , 766 , l1 quanz , s. p. , avenhaus , h. , buenzli , e. , et al .
2013 , , 766 , l2 soummer , r. , pueyo , l. , & larkin , j. 2012 , , 755 , l28 sylvester , r. j. , skinner , c. j. , barlow , m. j. , & mannings , v. 1996 , , 279 , 915 van boekel , r. , min , m. , waters , l. b. f. m. , et al .
2005 , , 437 , 189 distance & + age & + proper motion ( @xmath29 , @xmath30 ) & + separation : 14 july 2013 ut & & [email protected]@xmath2 ( @xmath116 au ) + position angle : 14 july 2013 ut & & 0@xmath1314@xmath3 + @xmath31 & & [email protected] + @xmath32 ( mag ) & 6.91 & + @xmath33 ( mag ) & 6.41 & + @xmath34 ( mag ) & 5.64 & [email protected] + @xmath35 ( mag ) & -0.2 & [email protected] + spectral type & a7v@xmath36 & + estimated mass & 1.65 m@xmath37 & + lcccc vlt naco vortex & 14 july 2013 & l & 75 min & 135.2 + magao clio-2 narrow & 9 april 2014 & h & 46 min & 178.5 + magao clio-2 narrow & 10 april 2014 & h & 30.4 min & 172.9 + magao clio-2 narrow & 12 april 2014 & k@xmath4 & 70 min & 180.1 + magao clio-2 narrow & 15 april 2014 & 3.9@xmath5 m & 60 min & 164.1 + magao visao narrow & 15 april 2014 & z@xmath38 & 82.4 & 176.8 + [email protected] , at a separation of [email protected]@xmath2 and pa of 0@xmath1314@xmath3 .
right : 3.9 @xmath5 m april 2014 psf - subtracted magao clio-2 image .
dark pixels in the core are due to saturation / non - linearity .
the april 2014 magao image reached shallower contrasts than the original vortex image , hence the faint point - like feature is not retrieved .
our noncoronagraphic dataset was too shallow to retrieve the candidate ( @xmath0mag=5.6 at 0.11 ) .
[ fig : image1],width=480 ] | we report the detection of a faint pointlike feature possibly related to ongoing planet - formation in the disk of the transition disk star hd 169142 .
the pointlike feature has a @xmath0mag(l)@xmath16.4 , at a separation of @xmath10.11@xmath2 and pa@xmath10@xmath3 .
given its lack of an h or k@xmath4 counterpart despite its relative brightness , this candidate can not be explained by purely photospheric emission and must be a disk feature heated by an as yet unknown source .
its extremely red colors make it highly unlikely to be a background object , but future multi - wavelength followup is necessary for confirmation and characterization of this feature . |
the spin hall effect ( she ) has been attracting a great deal of interest due to its fundamental as well as technological importance . the intrinsic she originates from the berry phase of the multiband bloch function , so that the spin hall conductivity ( shc ) takes a material - specific value that is independent of impurity scattering in the low resistivity regime @xcite . in transition metals ,
the intrinsic shc takes a considerably large value @xcite , because of the effective aharonov - bohm ( ab ) phase " induced by @xmath6-orbital degrees of freedom and the atomic spin - orbit interaction ( soi ) @xcite .
intrinsic she in sr@xmath7ruo@xmath8 had been calculated in ref .
@xcite based on the @xmath9 tight - binding model . by the same mechanism , she in transition metals had been studied based on the realistic multiorbital tight - binding model in ref . @xcite and the first - principles calculations @xcite .
the obtained shcs agree well with recent experimental observations , in both the magnitude and sign @xcite .
in addition to the intrinsic mechanism , she is also induced by impurity scattering with the aid of soi ; this phenomenon is known as the extrinsic she @xcite .
it consists of a skew scattering term ( @xmath10 ) and a side jump term ( @xmath11 ) ; this fact suggests that considerable large shc can be realized due to the skew scattering term in the very low resistivity metals , beyond the material - specific intrinsic shc .
for this reason , study of the extrinsic she is significant for realizing excellent spintronics device to produce or detect the spin current .
the extrinsic she due to nonmagnetic impurities or vacancies had been studied in refs .
@xcite , by developing the theory of the extrinsic anomalous hall effect ( ahe ) @xcite .
the obtained shc is , however , very small unless the enhancement factor for the soi due to the multiband effect is taken into consideration .
for this reason , it is not easy to estimate the magnitudes of @xmath12 and @xmath13 due to nonmagnetic impurities . in transition metals
, there is no apparent experimental evidence for a large extrinsic she due to natural randomness .
dilute magnetic impurities such as fe and ce have been found to induce a large extrinsic ahe in simple metals @xcite .
the hall angles in aufe @xcite and rare - earth doped metals @xcite reach [email protected] under high magnetic fields .
since the she has a similar mechanism to that of the ahe , one might expect that a giant she can be realized by magnetic impurities .
however , theoretical analysis of the extrinsic she in a simple @xmath15-electron metal due to magnetic impurities has not been performed previously . in this paper , we study the extrinsic she based on a single - impurity anderson model for ce and yb atoms , both of which are typical magnetic impurities .
nonperturbative effect of the large soi for @xmath16-electrons ( @xmath17 k ) is correctly accounted for .
due to skew scattering , the obtained spin hall angle @xmath18 , which is the ratio between the transverse spin hall current and the longitudinal electric current , exceeds @xmath19 for @xmath5{$\;\stackrel{>}{\sim}\;$}}0.03 $ ] , where @xmath2 is the phase shift for @xmath3 partial wave ; it is more than 10 times greater than that in nonmagnetic metals such as pt @xcite .
since the spin hall angle represents the efficiency of the creating spin current , giant spin hall current is expected to emerge in simple metals by introducing dilute magnetic impurities .
moreover , the extrinsic shc can take a large value well above the kondo temperature @xmath20 , in contrast to the anomalous hall conductivity ( ahc ) .
the origin of the giant she is found to be the large soi and angular momentum in the rare - earth impurities , which had not been treated appropriately in previous studies of the extrinsic she .
in this study , we use the single - impurity anderson model for ce and yb atoms . in a ce@xmath21 ion with a @xmath22 configuration ,
the @xmath23 level is about 3000 k higher than the @xmath24 level due to the strong atomic soi @xcite .
therefore , we consider only @xmath24 states for the ce - impurity . in the same way
, we consider only @xmath23 states in yb@xmath21 ion with @xmath25 configuration . here
, we introduce the following single - impurity anderson model for ce@xmath26yb atoms with both @xmath6- and @xmath16-orbitals @xcite : @xmath27 where we have employed the electron ( hole ) picture for the ce ( yb ) impurity .
@xmath28 is the creation operator of a conduction electron with spin @xmath29 .
@xmath30 is the creation operator of a @xmath16-electron with total angular momentum @xmath31 and @xmath32-component @xmath33 for ce@xmath21 ( yb@xmath21 ) .
@xmath34 is the creation operator of a @xmath6-electron with angular momentum @xmath35 @xcite .
@xmath36 is the energy for the conduction - electrons , and @xmath37 ( @xmath38 ) is the localized @xmath39-level energy . @xmath40 and @xmath41 are the mixing potentials , which are given by v^f_m & = & v_f _
m a^m_m y^m_3 ( ) , + v^d_m & = & v_d y^m_2 ( ) , where @xmath42 is the clebsch - gordan ( c - g ) coefficient and @xmath43 is the spherical harmonic function .
we will show that the phase factor in @xmath44 and @xmath45 are indispensable to realize the she . for @xmath24 , @xmath46 @xcite .
note that the soi was neglected in the study of the ahe in ref .
@xcite . in the present study
, we neglect the crystalline electric field of the @xmath16-orbitals , since it is small due to the small radius of the @xmath16-orbital wave function @xcite .
we put @xmath47 hereafter . to discuss the scattering problem , it is useful to derive an effective hamiltonian for the conduction electrons by integrating out the @xmath16 and @xmath6 electrons in eq .
( [ eq : ham ] ) .
the obtained hamiltonian is given by @xcite h_c&= & _ _ c^_c _ + _ , , j^d_, c^ _ c_ + _ , j^f_, c^_c_ , [ eq : ham - re ] where j^d_, & = & 1-e^d_m v^d_m v^d _ m + & = & 4j_d _ m y^m_2 ( ) ^ , + j^f_,k & = & 1-e^f_m v^f_m v^f _ m + & = & 4j_f _ m m m a^m_m a^m_m y^m_3 ( ) ^ , and @xmath48 and @xmath49 . here , @xmath50 ; @xmath51 is the @xmath16-electron self - energy due to the coulomb interaction @xmath52 @xcite . according to the scaling theory @xcite
, @xmath53 approaches the fermi level as the temperature decreases due to the kondo effect : @xmath54 is strongly enhanced near @xmath55 , and below @xmath20 , @xmath56 due to strong resonant scattering , where @xmath57 is the density of state of the conduction band per spin .
we also assume that @xmath58 since @xmath59 @xcite .
in this section , we study the scattering problem . in the both models given by eq .
( [ eq : ham ] ) and eq . ( [ eq : ham - re ] ) , the @xmath60-matrix due to the @xmath61-@xmath16 resonant scattering is equivalent ; it is given by t^f_, = j^f_, + _ _
1,_1 j^f_,_1_1 g^0__1 t^f__1_1, , [ eq : tmat ] where its diagrammatic expression is shown in fig .
[ fig : tmat ] ( a ) , @xmath62 is the number of @xmath63-points , and @xmath64 .
( a ) diagrammatic expression for the @xmath60-matrix due to @xmath61-@xmath16 resonant scattering . ( b ) the diagram that vanishes after @xmath65 -summation . ]
we note that the term containing both @xmath66 and @xmath67 given in fig .
[ fig : tmat ] ( b ) vanishes identically due to the orthogonality of spherical harmonic functions .
the solution of eq .
( [ eq : tmat ] ) for @xmath68 and @xmath69 is simply given by t^f_,()= 3 j_f 11-j_f g ( ) , where we have used the relations @xmath70 and @xmath71 .
@xmath72 is the local green function .
assuming approximate particle - hole symmetry near @xmath73 , we put @xmath74 .
then , the quasiparticle damping rate in the @xmath60-matrix approximation is given by _
f = -n_imp i m t^fr_,(0 ) = 3 n_imp n(0 ) j_f^21 + ( n(0 ) j_f)^2 , [ eq : gamma - t ] where @xmath75 is the impurity concentration . note that eq .
( [ eq : gamma - t ] ) is exact for @xmath76 . in this section
, we study the skew scattering term using linear response theory .
initially , we consider the case @xmath77 and @xmath78 , where the born approximation is valid . in analogy to @xcite ,
the skew scattering term is given by ^ss_sh & = & - n_imp _
+ & & |g^r_(0)|^2 \ { t^f(2)r_,(0 ) j^d _ + c.c } , [ eq : ss - born1 ] where @xmath79 is the electron charge .
its diagrammatic expression is shown in fig .
[ fig : ss - born ] ( a ) .
diagrammatic expressions for shc induced by skew scattering ( a ) within the lowest order ( extended born approximation ) contribution , and ( b ) in the @xmath60-matrix approximation with full order diagrams .
, title="fig : " ] diagrammatic expressions for shc induced by skew scattering ( a ) within the lowest order ( extended born approximation ) contribution , and ( b ) in the @xmath60-matrix approximation with full order diagrams .
, title="fig : " ] here , t^f(2)r_,(0 ) & = & _ _ 1,_1 j^f_,_1_1 g^r_(0 ) j^f__1_1, + & = & g^r(0 ) j^f_, j_f is the second order term of the @xmath60-matrix ; the first order term in @xmath80 does not contribute to @xmath81 up to the first order term in @xmath82 @xcite .
note that any diagram that contains the part shown in fig .
[ fig : tmat ] ( b ) vanishes identically .
the retarded green function is given by @xmath83 , where @xmath84 represents the quasiparticle damping rate .
here , we put = _ f + _ 0 , where @xmath85 is the damping rate due to nonmagnetic scattering , such as @xmath61-@xmath6 scattering ( @xmath86 ) and the scattering due to disorders .
the charge current is given by @xmath87 , where @xmath88 .
the spin current is then given by @xmath89 .
first , we consider the angular integration in eq .
( [ eq : ss - born1 ] ) , which is given by _ 2 _ , [ eq : average1 ] where @xmath90 , and @xmath91 denotes the average over the fermi level , which is defined as @xmath92 . since @xmath93 , angular integration such as @xmath94 appears in eq .
( [ eq : average1 ] ) .
this integral is finite only when @xmath95 .
therefore , the interference of the @xmath96 and @xmath3 partial waves is essential for skew scattering @xcite . after performing the angular integrations ,
( [ eq : average1 ] ) is given as eq .
( [ eq : average1 ] ) = i 47 j_d j_f , using the relations @xmath97 for small @xmath98 , and @xmath99 , eq . ( [ eq : ss - born1 ] ) is transformed into _ sh^ss & = & - n_imp j_d j^2_f n(0 ) + & & _ , ( -_k ) ( -_k ) .
[ eq : ss - born2 ] since @xmath100 in the born approximation , @xmath101 is given by ^ss_sh = - j_d k^4_f .
[ eq : ss - born ] now , we derive the skew scattering term using the @xmath60-matrix approximation , which gives the exact result for @xmath76 . in this case , @xmath102 in eq .
( [ eq : ss - born1 ] ) is replaced with the full @xmath60-matrix @xmath103 , where the first order term in @xmath104 has been dropped .
the diagrammatic expression for @xmath101 is shown in fig .
[ fig : ss - born ] ( b ) .
the angular integration in the @xmath60-matrix approximation can be performed as & & re _ 2 _ = 47 j_d j^2_f n(0)1 + ( n(0 ) j_f ) ^2 .
thus , @xmath101 in the @xmath60-matirx approximation is given by @xmath105 times eq .
( [ eq : ss - born2 ] ) .
therefore , the expression for the skew scattering term in eq .
( [ eq : ss - born ] ) is valid beyond the born approximation , by considering @xmath106 given by eq .
( [ eq : gamma - t ] ) . here
, we discuss the temperature dependence of @xmath101 . from eq .
( [ eq : ss - born ] ) , @xmath101 is proportional to @xmath107 when @xmath108 . since @xmath109 for @xmath110 @xcite
, @xmath101 decreases as the temperature decreases .
in contrast , @xmath111 for @xmath112 , which increases at low temperatures if @xmath85 is constant . in this section ,
we briefly discuss the side jump term @xmath113 . in the @xmath60-matrix approximation , @xmath113 is given by @xcite _ sh^sj & = & -e2 _ , j^f_,k_x _
k_y |g^r ( 0)|^2 + & & , whose diagrammatic expression is shown in fig .
[ fig : sj - born ] .
diagrammatic expression for the side jump term in the @xmath60-matrix approximation .
both two diagrams give the same contribution .
note that @xmath82 is not necessary for the side jump . ]
in contrast to the skew scattering term , @xmath82 is not necessary for the side jump term . as reported in ref .
@xcite , the large anomalous velocity @xmath114 , which is not perpendicular to the fermi surface , arises from the @xmath63-derivative of the phase factor in the @xmath61-@xmath16 mixing potential as follows : v^f_m k_x v^a_m = -i ( m-2 ) k_yk_x^2 + k_y^2 v^f_m .
the final expression for the side jump term in the @xmath60-matrix approximation is given by _
sh^sj = -e2 43 k_f _ f , [ eq : sj - tmat ] where we have used the relations @xmath115 and @xmath116 . by using eqs .
( [ eq : ss - born ] ) and ( [ eq : sj - tmat ] ) , the ratio @xmath117 is given by @xmath118 . therefore , @xmath113 will exceed @xmath101 in dirty metals . however , the skew scattering contribution will be dominant in simple metals with magnetic impurities for @xmath119{$\;\stackrel{<}{\sim}\;$}}0.01 \
( \rho_0 { \raisebox{-0.75ex}[-1.5ex]{$\;\stackrel{<}{\sim}\;$}}10\mu \omega$]cm ) @xcite .
in section [ ss - term ] , we have studied the shc due to skew scattering by using the green function method . in this section
, we discuss the origin of the skew scattering mechanism based on the boltzmann transport theory . for this purpose , we study a simplified two - orbital model with @xmath120 , assuming the strong crystalline electric field . in this model , @xmath121 . since @xmath122 approximately , the second order term of @xmath123 is simply given as @xmath124 . in the boltzmann transport theory ,
the spin hall resistivity due to skew scattering is @xmath125 , where @xmath126 represents the scattering probability , which is proportional to @xmath127 due to fermi s golden rule in the present model . according to ref .
@xcite , skew scattering occurs when the scattering probability includes an assymetric component @xmath128 . in the present model ,
@xmath129 arises from the interference of the @xmath16 and @xmath6 scattering channel , @xmath130 .
in fact , @xmath131 in this model since @xmath67 contains the term @xmath132^{\ast } \propto { \rm e}^{\pm 2i ( \phi_{k}-\phi_{k ' } ) } $ ] . in summary ,
a conduction electron with @xmath133 hybridizes with @xmath134 state due to the strong soi , and therefore the spin - dependent skew scattering probability @xmath135 arises from the interference of the @xmath16 and @xmath6 angular momenta .
thus , the origin of the she due to skew scattering mechanism is well understood based on the simplified two - orbital model .
first , we estimate the magnitude of the spin hall angle @xmath136 .
the longitudinal conductivity is given by @xmath137 , where @xmath138 is the density of the conduction electrons .
then , the spin hall angle is given by _
h = 8 ^ 27 j_d n(0 ) _ f. [ eq : hallangle ] since @xmath139 @xcite , eq .
( [ eq : hallangle ] ) can be rewritten as @xmath140 if @xmath141 . in the case where @xmath108
, @xmath142 reaches @xmath143 for @xmath144 @xcite . therefore , giant she will be realized in metals with rare - earth impurities .
recently , giant she was observed in fept / au devices @xcite .
it was reported that the large spin hall angle in au is @xmath145 . since the intrinsic she in au is small due to the fermi surface being mainly composed of @xmath15-electrons @xcite , the observed giant she may be explained by the skew scattering mechanism .
very recently , the extrinsic she in aufe metal has been studied with the first - principles band structure calculations independently of our study @xcite .
second , we estimate the magnitude of the skew scattering term for @xmath146 , where the spin hall angle is @xmath147 .
then , @xmath148 from the above discussion . in the case where the resistivity is @xmath149 cm , we obtain @xmath150@xmath151 if we put the length of unit cell @xmath152 , which is the case for pt and au .
the obtained skew scattering term is about @xmath153 times larger than the shc observed in pt : @xmath154@xmath151 .
since the resistivity in pt and au is @xmath155 cm and @xmath156 cm , respectively , the skew scattering term due to magnetic impurities should dominate the intrinsic term .
in contrast , the skew scattering due to natural randomness " has not been observed in the ahe in transition metals , even in @xmath155 cm @xcite .
therefore , the orbital degrees of freedom of the impurity is indispensable for the giant extrinsic she .
finally , we compare the magnitude of @xmath81 and @xmath157 . if we put @xmath158 and @xmath159 in eq .
( [ eq : sj - tmat ] ) , @xmath160@xmath151 for @xmath152 .
therefore , the skew scattering term will be comparable in magnitude to the side jump term when @xmath161 cm .
thus , the skew scattering term is dominant over the side jump term when @xmath162 cm .
in summary , we have studied the extrinsic she due to magnetic impurities , where ce and yb impurities have been discussed as a typical case . the obtained expression for the skew scattering term
shows that the spin hall angle reaches @xmath4 , which is more than 10 times larger than that in pt .
therefore , we propose that the extrinsic she due to rare - earth impurities is useful as a method for creating large spin current efficiently .
the present study also suggests that the giant she will arise from other rare - earth atoms as well as 4@xmath6 and 5@xmath6 atoms that possesses strong soi . as in the case for the intrinsic she , orbital degrees of freedom are essential for generating a huge extrinsic she . according to eq .
( [ eq : hallangle ] ) , giant spin hall angle always appears when @xmath163 .
thus , the kondo resonance , @xmath164 , is not a necessary condition for the giant she . in contrast
, the ahc ( @xmath165 ) is enhanced by the kondo effect since it is proportional to the uniform susceptibility . in a similar way to the present study , we can obtain @xmath101 and @xmath113 for @xmath23 , which is the case for yb - impurities . in this case ,
@xmath166 , and @xmath106 is given by @xmath167 times eq .
( [ eq : gamma - t ] ) .
the obtained @xmath101 and @xmath113 are given by @xmath168 times eq .
( [ eq : ss - born ] ) and eq .
( [ eq : sj - tmat ] ) , respectively .
we are grateful to d. s. hirashima and j. goryo for fruitful discussions . | we investigate the extrinsic spin hall effect in the electron gas model due to magnetic impurities , by focusing on ce- and yb - impurities . in the dilute limit
, the skew scattering term dominates the side jump term . for ce - impurities , the spin hall angle @xmath0 due to skew scattering
is given by @xmath1 , where @xmath2 is the phase shift for @xmath3 partial wave .
since @xmath0 reaches @xmath4 if @xmath5{$\;\stackrel{>}{\sim}\;$}}0.03 $ ] , considerably large spin hall effect is expected to emerge in metals with rare - earth impurities .
the present study provides the highly efficient way to generate a spin current . |
the past two decades have seen substantial improvements in our understanding of the demographics of nuclear supermassive black holes ( bhs ) , predominantly driven by the availability of bh mass measurements for nearby galaxies from dynamical techniques .
the advent of the _ hubble space telescope _ has enabled a few dozen dynamical bh mass measurements , which reveal that supermassive bhs are ubiquitous components of bulge - dominated galaxy centers ( e.g. , * ? ? ? * ) .
furthermore , there appears to be a remarkably tight correlation between bh mass and bulge properties , including velocity dispersion ( the @xmath5 relation ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , luminosity ( the @xmath7relation , e.g. , * ? ? ?
* ) , and mass ( e.g. , * ? ? ?
the apparently tiny scatter in the @xmath5 relation has led to a common view that the growth of bhs is intimately connected to the growth of the surrounding galaxy ( see discussion in * ? ? ?
* ) . in this paper ,
we explore the demographics of a new sample of dynamical bh masses measured using observations of circumnuclear masers ( kuo , c. y. et al . in prep . )
. apart from our own galactic center , where individual stars are used as test particles to weigh the bh ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , the most precise bh masses are derived using water megamaser spots in a keplerian circumnuclear disk ( e.g. , * ? ? ?
such megamasers are very close to an ideal dynamical tracer of the central mass at the galaxy center , as they delineate a disk with a well - measured keplerian rotation curve within fractions of a pc from the center .
they are detectable when the disk is aligned within a few degrees of the line of sight .
masers detected in edge - on nuclear disks have characteristic triple - peaked spectral profiles , making the disk inclination angle well - known in these cases @xcite .
bh mass measurements based on the keplerian rotation curves of circumnuclear megamaser disks are not prone to the same systematic uncertainties as the dynamical methods based on optical observations . in the case of gas dynamical methods , it can be very difficult to determine the level of non - virial gas support ( e.g. , due to turbulence ) but neglecting these extra sources of support can lead to significant bias in the final result ( e.g. , * ? ? ?
stellar dynamical methods are sensitive to assumptions about galaxy anisotropy and radial variation in dark matter fraction that are not well constrained observationally in general ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . the cleanest and best - studied megamaser galaxy is the nearby seyfert galaxy ngc 4258 .
observations of centripetal acceleration from the frequency change of the maser components at the systemic velocity of the system , combined with vlba images of the megamaser disk , allow for the most precise extragalactic angular diameter distance measurement @xcite .
other observations provide detailed constraints on the thickness , velocity structure , and toroidal magnetic field in an accretion disk ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for a long time it proved difficult to detect additional well - defined keplerian megamaser disks , ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but the situation has changed dramatically with the advent of the green bank telescope ( gbt ) of the nrao . thanks to the increased sensitivity of the gbt with a wide - band spectrometer , more megamaser candidates have been discovered @xcite . of the @xmath8 extragalactic megamasers , approximately half
were discovered with the gbt and @xmath9 show evidence of originating in a disk .
the primary goal of ongoing megamaser searches is to find additional disk masers that can be used for angular diameter distance measurements in the megamaser cosmology project ( mcp ; * ? ? ?
* ; * ? ? ?
* ) . with a sufficient number of angular diameter distances to galaxies in the hubble flow
, the mcp aims to determine a precise independent hubble constant to a few percent accuracy @xcite .
an important outcome of the mcp is that the keplerian megamaser disks also provide a precise measurement of the enclosed mass within a few tenths of a pc from the center of these spiral galaxies , dominated by a bh at the center , a valuable resource in its own right .
the bh masses for seven maser galaxies are presented in kuo et al .
( in preparation ) .
the present paper examines the demographics of the new megamaser galaxies , focusing on the shape and scatter of bh - bulge relations for these galaxies . to date
, the vast majority of dynamical bh masses have been obtained in massive elliptical and s0 galaxies ( fig .
1 ) . there are only five targets with bh mass measurements @xmath10 @xmath2 ( one of those being the milky way ) and there are only 11 spiral galaxies with hubble type of sa or later in the sample of 67 galaxies used in the most recent calibration of the @xmath5 relation ( * ? ? ?
1 ) . the disproportionate representation of massive galaxies is easily understood .
local elliptical galaxies are relatively dust - free , and the bhs are massive enough that we can resolve their gravitational spheres of influence at distances of tens of mpc .
spiral galaxies tend to be dusty and star - forming , complicating dynamical measurements .
our knowledge of bh demographics at low masses is rudimentary , such that the shape and scatter in bh - bulge scaling relations as well as the bh occupation fraction in spiral galaxies remain uncertain ( e.g. , * ? ? ?
the main focus of this work is to exploit the new sample of reliable bh masses afforded by the new maser studies to explore bh demographics for bhs with @xmath0@xmath11 @xmath2 in spiral galaxies .
the paper starts with a description of the sample ( 2 ) .
we then describe the observations and data reduction ( 3 ) and the spectroscopic ( 4 ) and imaging ( 5 ) analysis . in
6 we present the primary result of the paper , namely the location of the maser galaxies in the @xmath5 plane , while in 7 the scalings between @xmath0 and bulge luminosity / mass are briefly discussed .
the impact of our results on our knowledge of bh demographics , particularly as probed with active galaxies , is discussed in 8 , and we summarize and conclude in 9 .
throughout we assume the following cosmological parameters to calculate distances : @xmath12 km s@xmath13 mpc@xmath13 , @xmath14 , and @xmath15 .
over the years there have been a number of searches for megamaser galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . by and large ,
the surveys have targeted known obscured active galaxies , with a detection rate of @xmath16 .
once -0 mm detected in a single - dish survey , megamasers that show disk characteristics , e.g. , both systemic and high - velocity doppler components , and that are bright enough , are subsequently imaged with very long baseline interferometry ( vlbi ) to map the positions and velocities of individual maser spots .
the map determines whether a maser is in a clean dynamical system or has more complicated structures such as outflows ( e.g. , * ? ? ?
* ) . in this paper we present velocity dispersions both for the kuo et al .
sample and two other objects , ngc 3393 @xcite and ic 2560 @xcite . in the latter case , while it is assumed that the systemic masers arise from the inner edge of the masing disk , without vlbi mapping the enclosed mass is uncertain by factors of a few .
we expect better data will yield a concrete mass .
the vlbi ( and thus mass constraints ) for the remaining seven objects are presented for the first time in c - y .
kuo et al . in preparation .
the details of the bh mass fitting are described in kuo et al . but are reviewed here briefly for completeness .
the bh masses are derived by fitting a keplerian rotation curve to the positions and velocities of the maser spots .
the distance is fixed using an assumed value of @xmath17 and the inclination of the maser disk is assumed to be @xmath18 .
all of the megamaser disks for which we have measured an inclination are within @xmath19 of being perfectly edge - on , so the inclination contributes less than a @xmath20 error to the measurement of the black hole mass . in ngc 4388
we could not measure a disk inclination , but even if the disk is @xmath21 from edge - on , the contribution to the error in the bh mass would be only @xmath22 , which is comparable to the bh mass error caused by the distance uncertainty ( @xmath23 ) .
ongoing monitoring of the most promising targets will allow the measurement of accelerations in the systemic features .
joint modeling of @xmath0 and distance for these targets will lead to a tighter constraint on each ( e.g. , * ? ? ?
it is interesting to note that the bh masses are strongly clustered around @xmath11 @xmath2 .
we suspect that this is a simple consequence of picking random galaxies from the local active - galaxy mass function . without accounting for incompleteness
, the observed active mass function is strongly peaked at a mass @xmath11 @xmath2(e.g .
* ) . at higher mass , there are very few systems radiating at an appreciable fraction of their eddington limit and we are woefully incomplete at lower mass @xcite . while we do not sample the full range of bh masses in the local universe with this sample , our selection is unbiased with respect to bulge mass and velocity dispersion and thus does not invalidate our results .
the megamaser galaxies are found predominantly in early - to - mid type spiral galaxies , ranging in hubble type from s0 ( e.g. , ngc 1194 ) to sb ( e.g. , ngc 4388 ) .
table 1 includes the rc3 morphological types @xcite and @xmath24-band luminosities and colors from hyperleda @xcite . as noted above
, the parent sample consists predominantly of known active galaxies , ranging in distance from 18 to 150 mpc .
only ugc 3789 did not have a documented active galactic nucleus ( agn ) when the circumnuclear maser was discovered , although is it a narrow - line agn . with the exception of ngc 4388 , which has a weak broad h@xmath25 line @xcite ,
the active galaxies are classified as narrow - line ( obscured ) agns from the optical spectra .
furthermore , those with x - ray spectra have large column densities .
most are compton thick .
the only exceptions are ngc 4388 , which has been observed to undergo large variations in column density over time scales of hours @xcite , and ngc 6264 , ngc 6323 , and ugc 3789 , which have not yet been observed in the x - rays .
the attempted _ xmm - newton_observations of the former two were destroyed by flaring .
-0 mm the primary observations discussed here were obtained using the dupont 2.5 m telescope at las campanas observatory in the south , and the apache point observatory ( apo ) 3.5 m telescope in the north . in the appendix
we describe an auxiliary data set that we use to cross - check our velocity dispersion measurements . at the dupont telescope
, we used the boller & chivens ( b&c ) spectrograph with the 600 lines mm@xmath13 grating and the 2 slit , which yielded an instrumental resolution of @xmath26 km s@xmath13 and a wavelength coverage of @xmath27 , covering the 4000 break , the g - band at 4304 , h@xmath28+[]@xmath29 , the @xmath30@xmath31 triplet , and ( in most cases ) h@xmath25 .
we observed at two slit positions ( table 1 ) for the majority of the galaxies , both along the major and minor axis , at the lowest possible airmass to mitigate differential refraction .
we observed the northern targets with apo using the dual imaging spectrograph ( dis ) , with a @xmath32 slit , a 1200 lines mm@xmath13 grating , a central wavelength of 4820 , a spectral range of 4250 - 5400 and a dispersion of @xmath33 km s@xmath13 , as measured from the arc lines .
these observations were performed at the galaxy major axis only .
we also present one calcium triplet @xmath34 ( cat ) measurement , based on a red setting with a central wavelength of 8540 (8000 - 9100 ) and comparable instrumental resolution .
the reduction of both the dis and b&c data proceeded in a similar fashion .
flat - fielding , bias - subtraction , and wavelength calibration were performed within iraf . in the case of the b&c data ,
flux calibration and telluric correction was also performed within iraf , using ltt 3218 , ltt 3864 , ltt 6248 , hilt 600 , ltt 3218 , feige 56 , and cd 32 as flux calibrator stars in order to match the airmass and time of observation .
flux calibration and telluric correction of the dis data used idl routines as described by @xcite . in this case
we used feige 34 as the flux calibrator . in the appendix
we compare the [ ] luminosities between two independent data sets and demonstrate that our flux calibration is good to @xmath35 . the proper physical aperture used to calibrate the @xmath5 relation remains a matter of debate ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and at present robust @xmath36 measurements are only available for a small fraction of the maser sample .
thus , we have chosen to extract spectra with apertures ranging from the resolution limit of the observations ( @xmath37 ) to the s / n limit of the spectra ( @xmath38 ) .
typically the signal - to - noise ( s / n ) ratio is optimized with apertures of @xmath39 .
the resulting dispersion in @xmath6 values is included in the error budget , and is @xmath40 .
we use direct - pixel fitting ( e.g. , * ? ? ?
* ) to measure stellar velocity dispersions .
the stellar templates are broadened and fitted in pixel rather than fourier space . -0.05 in -0 mm this method is computationally expensive relative to fourier methods ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but by no means prohibitive by modern standards .
direct pixel fitting provides many benefits over fourier techniques ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for one thing , it is easier to incorporate noise arrays directly . of particular importance to active galaxies , arbitrary regions of the spectrum ( such as narrow emission lines ) may be masked in a straightforward manner without introducing spurious signals into the final results .
the details of the code used here are described in @xcite and @xcite . in short ,
each spectrum is modeled as the linear combination of template stars @xmath41 that are shifted to zero velocity , broadened by a gaussian , @xmath42 , with width @xmath43 and diluted by a constant [ or power - law if necessary ; @xmath44 .
finally , the entire model is multiplied by a polynomial [ @xmath45 ; typically third order ] to account for intrinsic spectral differences , reddening , and errors in flux calibration : @xmath46 + c(\lambda ) \}\ ] ] the nonlinear levenberg - marquardt algorithm , as implemented by mpfit in idl @xcite , is used to minimize @xmath47 in pixel space between the galaxy and the model .
we have a library of template stars observed with the b&c spectrograph . for the fitting we use the following stars : hd 36634 ( g6 iii ) , hd 129505 ( g8 iiii ) , hd 126571 ( k0 iii ) ,
hip 70788 ( k3 iii ) , and hd 131903 ( a9 v ) .
for the apo observations , our spectral library is much more limited , and we have only one reliable velocity template star .
thus , we utilize a large library of high - s / n , high - spectral resolution ( @xmath48 km s@xmath13 ) stellar templates from @xcite .
specifically , we use the following stars : hd 26574 ( f2 ii - iii ) , hd 107950 ( g6 iii ) , hd 18322 ( k1 iii ) , and hd 131507 ( k4 iii ) . in order to translate the measured dispersion to the true dispersion , we first measure the instrumental dispersion of the dis spectra using arc lines ( 52 km s@xmath13 ) , which yields a quadrature difference of @xmath49 km s@xmath13between the dis and valdes spectra .
we also measure the dispersion of the dis velocity template using a valdes star of the same spectral type and find @xmath50 km s@xmath13 .
the consistency of the two methods is reassuring .
thus , the final dispersion is simply @xmath51 .
note that because we use stellar templates observed with an identical instrumental set - up , we need apply no corrections to the b&c spectra . given our spectral coverage , the most robust spectral features with high enough equivalent width ( ew ) for dispersion measurements are either the @xmath30 triplet or the g - band .
we avoid the ca h+k features because they are highly sensitive to spectral type , occur in a spectral region with a strong continuum discontinuity , and may be contaminated by interstellar absorption ( see review in * ? ? ? * ) .
the @xmath30 features are in many cases contaminated by emission from [ ] @xmath52 and [ ] @xmath53 .
furthermore , differences in the @xmath25/fe ratios between the template stars ( from the solar neighborhood ) and external galaxies will lead to systematic errors in @xmath6 measurements from this region @xcite .
for these reasons , we focus on the g - band region for the velocity dispersion measurements .
we have multiple slit positions for most of the dupont spectra . in table 2
we present measurements for each slit position with the @xmath54 aperture , but we also present the weighted mean of the two , and use that value in all subsequent work .
example fits are shown in figure 2 while the dispersions are presented in table 2 .
there are literature measurements available for a limited number of targets in our sample .
direct comparison can be tricky ; the dispersions may be highly aperture dependent ( e.g. , * ? ? ?
* ; * ? ? ?
nevertheless , in general we find decent agreement with the literature values , as summarized in figure 3 .
two targets ( ngc 2273 , ngc 4388 ) appear in the recent atlas of stellar velocity dispersions measured for the palomar spectroscopic survey of nearby galaxies @xcite , taken with a comparable @xmath55 aperture @xcite .
the two sets of observations agree within the uncertainties . for ngc 2273
we find @xmath56 km s@xmath13 and ho et al .
present @xmath57 km s@xmath13 , while for ngc 4388 the values are @xmath58 km s@xmath13 along the major axis and @xmath59 km s@xmath13 along the minor axis . in the case of ngc 4388 ,
note that our observation is only marginally spectrally resolved . however , our sdss measurement , presented in the appendix , is @xmath60 , giving us some confidence in this result .
other literature comparisons for these two targets are presented in @xcite .
two other galaxies ( ngc 3393 and ic 2560 ) were observed by @xcite with a @xmath61 slit .
again , the measurements agree within the quoted errors . in the case of ic 2560 , we find @xmath62 km s@xmath13 along the major axis compared to their @xmath63 km s@xmath13 , while for ngc 3393 we find @xmath64 km s@xmath13 along the major axis and @xmath65 km s@xmath13 along the minor axis , as compared to @xmath66 km s@xmath13 .
ngc 3393 is also presented by @xcite measured within a @xmath67 extraction .
they find @xmath68 km s@xmath13 , which is marginally consistent with our minor - axis measurement .
one convenient by - product of our @xmath6 measurement technique is a pure emission - line spectrum that has been corrected for underlying absorption lines ( e.g. , h@xmath28 ) . since the narrow emission lines ( particularly [ ] ) represent one of the only methods for estimating bolometric luminosities for these obscured objects , we describe the measurements here .
the fitting routines are described in @xcite and @xcite .
briefly , we fit the [ ] @xmath69 lines with sums of up to four gaussians .
the two lines are restricted to have identical shapes and the velocity difference and relative strengths are fixed to laboratory values ( 3.1 in the latter case ) .
the h@xmath28 line is fitted independently with the sum of two gaussians ( which is always completely sufficient given the typical signal - to - noise ratio , s / n , in this line ) .
we also calculate the full - width at half maximum ( fwhm ) of the summation of all components in the [ ] model .
the velocity offsets between individual components are small compared to the total line width , so the fwhm is well - defined . to bracket systematic uncertainties
, we measure the line width in the spectra from the 4and 10 apertures and the error bar represents the difference in those two widths . for the majority of the b&c spectra , as well as the dis observation of ngc 6323 , we also model the h@xmath25 + [ ] complex .
ideally we would use measurements of the [ ] line width as a model to deblend h@xmath25 from [ ] .
( e.g. , * ? ? ?
* ; * ? ? ?
however , in this case we simply describe each narrow line with the same combination of gaussians , fix the relative strengths of the [ ] @xmath70 lines to 1:2.96 , and fix the relative wavelengths of all three lines to their laboratory values . in the case of ngc 4388 ,
we further include a broad component fixed to have the same fwhm as reported by @xcite , but we are not sensitive to the presence of weak broad lines in the other objects .
we derive an intrinsic reddening correction for [ ] using the h@xmath28 and h@xmath25 fluxes when available , but do not apply it to the fluxes .
all the line measurements are summarized in table 3 .
we are primarily interested in the [ ] luminosities as an isotropic indicator of the agn luminosity . estimating robust bolometric luminosities in these objects
is complicated since megamaser systems are found preferentially in obscured active galaxies ( e.g. , * ? ? ?
* ) . among the systems with circumnuclear disks , @xmath71
are compton thick ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus , there are only a few available luminosity diagnostics that are thought to be isotropic and immune from significant absorption .
one is the strong and ubiquitous [ ] line that has been used both to find agns and to determine their intrinsic luminosities ( e.g. , * ? ? ?
* ; * ? ? ?
* ) although concerns about reddening remain ( e.g. , * ? ? ?
* ; * ? ? ?
the shape and equivalent width ( ew ) of the fe k@xmath25 line at 6.7 kev are also sensitive to the intrinsic luminosity and inclination of the obscuring material ( e.g. , * ? ? ?
in a few cases , detailed modeling of the fe k@xmath25 lines have been performed , which yields an independent estimate of the bolometric luminosities @xcite .
finally , @xcite argue for the use of the [ ] @xmath72 emission line as a higher - fidelity tracer of agn bolometric luminosity ( see also * ? ? ?
* ; * ? ? ?
using the [ ] bolometric correction presented in @xcite , we derive bolometric luminosities for the entire sample using our measurements of @xmath73}}$ ] ( table 3 ) .
since this bolometric correction was derived without correction for internal reddening , we do not deredden either , but tabulate the proper correction for reference and comparison with the other indicators .
the estimated eddington ratios range from @xmath74@xmath75 . including internal reddening would change the values by 0.6 dex at most .
there are only two objects with estimated bolometric luminosities from the fe k@xmath25 line ( excluding ic 2560 for the moment ) , also shown in table 3 .
there are also three cases with [ ] measurements in the literature ; we have converted to bolometric luminosity using the calibration of j. rigby et al . in preparation .
as expected , @xmath76 ( [ ] ) tends to be higher than @xmath76 ( [ ] ) , but the number of objects is too small to draw definite conclusions about the average magnitude of this effect , or how directly it is related to our internal reddening estimates . the presence or absence of young stellar populations in the bulge regions of these galaxies provide important clues as to their growth and evolution ( 5.3 ) .
the spectra provide useful constraints on stellar ages , although the modeling is complicated by the narrow emission lines that litter the spectrum .
we do not perform full stellar population synthesis modeling . instead , we take a few representative sets of @xcite models and perform a least - squares fit to the galaxy continua , fixing the stellar velocity dispersion to that derived above .
we adopt the padova isochrones @xcite and a @xcite initial mass function .
we take both single - age stellar population models and exponentially decaying models with a decay time @xmath77 gyr . for simplicity ,
we use two models with ages of one and 10 gyr respectively , although we experiment with models having ages of @xmath78 and 5 gyr and also one - half of solar metallicity .
we adopt a single reddening screen for each galaxy , using the reddening law of @xcite .
see example fits in figure 4 .
the fraction of light in young stars is tabulated in table 2 . of the six galaxies for which our spectra include the 4000 break , we see that only two show evidence for a significant contribution from intermediate - age populations ( ngc 2960 and ic 2560 ) .
ngc 3393 and ngc 6264 may have a very minor young component , but presumably higher spatial resolution would be needed to isolate it .
we can compare these results with the more detailed stellar population fits of @xcite , whose sample also includes ic 2560 , ngc 4388 , and ngc 3393 . in ngc 3393
we both agree that the stellar populations are old ; cid fernandes et al .
find that @xmath79 of the luminosity in this source come from stars with an age @xmath80 myr .
they find a substantial @xmath81 contribution from young stars in ic 2560 , consistent with our findings .
the only ambiguous case is ngc 4388 .
they find that @xmath82 of the luminosity comes from this young component , while we do not find compelling evidence for a young component .
in fact , ngc 4388 is quite difficult to model , both because it is extincted and because the emission lines have very high equivalent widths .
those galaxies ( ngc 2273 , ugc 3789 , and ngc 6323 ) that we observed with apo do not have the blue coverage needed to constrain the stellar populations . in the case of ngc 2273 , young stars associated with an inner ring are well - documented @xcite .
unfortunately , we do not have much information for the other two .
the bulge color of ngc 6323 , @xmath83 ( see below ) , is typical for an early - type spiral galaxy @xcite , suggesting that either there is not much star formation and/or that there is a considerable quantity of dust .
overall we find evidence for ongoing or recent star formation in roughly half of the sample .
as discussed briefly in 2 , the velocity dispersion profiles of late - type spirals can be complicated by kinematically cold , disk - like components as well as bars or other non - axisymmetries . as a result
, it is impossible to pick a single , well - motivated aperture size , and the long - slit data presented here are not deep enough to derive robust rotation curves .
we can say that the measured dispersions do not change by more than @xmath84 over the range of radii we probe , and this spread in values is included in the error budget .
robust @xmath36 measurements are needed to define the optimal radius of extraction for each galaxy .
along similar lines , it is possible that the slit orientation relative to that of the bar might impact @xmath6(e.g .
* ; * ? ? ?
however , for the objects with both minor and major axis spectra , we see at most a 40% difference in the two measurements .
ideally , we would like to obtain integral - field spectroscopy for the objects to settle this ambiguity .
an additional concern is that differing stellar populations and metallicities lead to bias due to template mismatch .
while our code includes a mix of spectral types , subtle systematics from region to region remain ( see an example in * ? ? ?
when we calculate errors in @xmath6 , we include the differences between the nominal dispersions and those measured from the 5100 - 5400 region [ excluding the @xmath30 lines themselves since variations in @xmath25/fe tend to bias the dispersion measurements high ( e.g. , * ? ? ?
often , spiral galaxies contain bars , ovals , rings , dust lanes , nuclear spirals , and nuclear star clusters in their centers ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
furthermore , in galaxies with young stellar populations , the mass - to - light ratio ( @xmath85 ) may vary significantly both within a given galaxy and from object to object . to properly decompose such light profiles
requires both deep imaging on large scales to quantify the disk component and high resolution nuclear imaging , ideally in a red band , to reveal the complex nuclear morphology .
we first tried to perform detailed profile decompositions using two - micron all - sky survey ( 2mass ) images @xcite .
as shown in figure 5 , these images are too shallow to detect large - scale disks and rings .
furthermore , the resolution ( @xmath86 ) is insufficient to uncover nuclear features . a combination of @xmath87band observations with _ hst _ and from the ground is needed to fully model the entire galaxy . as an initial exploration we have settled on a compromise .
there are five galaxies that fall within the sloan digital sky survey footprint .
we have performed two - dimensional image decomposition on the @xmath88 images of these five galaxies , which allows us to measure a magnitude and a color for their bulge components .
we then use the relations of @xcite to convert the @xmath89 color into @xmath90 and thereby stellar mass .
we perform image decomposition in two dimensions using the program @xmath91
two - dimensional fitting allows flexibility in the ellipticities and position angles of different components that can help break degeneracies in fitting multi - component galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) as reviewed in detail in @xcite .
-0.05 in -0 mm @xmath91 and related programs create model galaxies as the sum of ellipsoids with parametrized surface brightness profiles that are convolved with the point - spread - function ( psf ) of the image and then compared with the data .
best - fit parameters are those that minimize the differences between data and model . in general
, we model the galaxies with @xcite functions : @xmath92 \right\},\ ] ] where @xmath36 is the effective ( half - light ) radius , @xmath93 is the intensity at @xmath36 , @xmath94 is the srsic index , and @xmath95 is chosen such that @xmath96 -0.1 in -0 mm an exponential disk has @xmath97 while @xmath98 corresponds to the well - known @xcite profile .
we adopt the psf models provided by the sdss , and note that our analysis is not very sensitive to the detailed psf shape . unlike typical two - dimensional modeling ,
@xmath91 now incorporates coordinate rotations and fourier modes ( departures from axisymmetric modeling ) that are particularly useful in handling dust - lanes , bars , and spiral structure @xcite . as described above , in addition to the structural measurements we also need color information to estimate the stellar masses of the galaxies .
we first fit the @xmath99band images to extract structural information and then apply the same model to the @xmath100band images , allowing only the total flux to vary for each component . in figure 6
we show the fits to the @xmath99band images .
we caution that the measurements are preliminary in the absence of high - resolution imaging .
therefore , we present only the `` bulge '' luminosities and colors , and reserve more detailed discussion of the galaxy structure ( e.g. , srsic index ) to future work . for the moment
, we simply note that the bulge colors provide an additional crude handle on the stellar populations in these galaxies ( table 2 ) . in the case of ngc 1194 and ngc 6323
, the @xmath89 color is typical of sa galaxies @xcite , ngc 2960 has the color of an sbc galaxy , and ngc 4388 and ngc 6264 are the bluest , with colors typical of scd galaxies .
-0 mm -0 mm clearly , errors in the psf shape and sky level add scatter to the photometric decompositions .
we expect that these uncertainties are @xmath101 mag .
therefore , we are dominated by systematic errors . more specifically , because of prominent dust lanes in many galaxies , combined with the limited spatial resolution of the sdss , we do not have a strong constraint on the srsic index of the bulge , which in turn contributes large errors to the effective radii as well as in the ratio of bulge - to - total luminosity ( b / t ) .
currently , our error bars represent the spread in derived luminosities resulting from a range of different models , which give some indication of the uncertainties due to model assumptions ( e.g. , * ? ? ?
after running these models , we report the one with the lowest value of reduced @xmath47 and assign parameter uncertainties based on the spread across the models .
let us take as an example the galaxy ngc 1194 .
in this case we have fit a model with the bulge srsic index fixed to @xmath98 and the disk fixed to @xmath97 .
as the next level of complexity we have allowed each component to vary freely , yielding @xmath102 for the bulge and @xmath103 for the `` disk '' . finally , we have run two models with different parameterizations of the dust lane that have @xmath104 for the bulge and disk respectively . in all cases
the fits would be deemed acceptable , but using a range of models allows us to quantify the dependence of the bulge parameters on our model assumptions . in this case , the resulting bulge magnitude varies by @xmath105 mag between the model with @xmath98 and those including dust .
these systematic uncertainties exceed by far those resulting from sky and psf uncertainties , and are the ones we report here . in this subsection
we discuss the morphological properties of the maser sample .
in particular , we address the bulge morphology .
first consider ngc 1194 ( fig .
this galaxy has a massive bulge component and is classified as an s0 .
both in the radial profile and the image , it is clear that the bulge in ngc 1194 has smooth isophotes and a large ratio of bulge - to - total ( b / t ) luminosities .
much more typical of the megamaser disk galaxies is ngc 6264 , which has a strong bar ( fig .
other barred galaxies include ngc 2273 , ngc 3393 , ngc 4388 , and ic 2560 .
roughly speaking , the bar fraction in the maser sample is consistent with the @xmath106 seen in the spiral galaxy population ( e.g. , * ? ? ?
* ; * ? ? ?
finally , we examine ngc 2273 ( fig .
7 ) , also a barred spiral galaxy . in addition , ngc 2273 contains two outer rings and a centrally concentrated bulge component @xcite . zooming in with _ hst_/wfpc2 and nicmos reveals that the galaxy center consists of an inner ring surrounding a nuclear spiral ( e.g. , * ? ? ?
* ; * ? ? ?
this spiral component is very different in shape , kinematics and stellar populations from a classical bulge .
the latter is understood to be a dynamically hot stellar component with an old stellar population , effectively an elliptical galaxy with a surrounding disk ( see discussion in * ? ? ? * ) .
in contrast , the center of ngc 2273 is a disk with ongoing star formation fueled by a reservoir of molecular gas ( e.g. , * ? ? ? * ) that is likely fed by the larger - scale bar .
furthermore , there is a distinct drop in @xmath6 within the inner @xmath107 where the kinematics are dominated by rotation , presumably because the young bright stars trace the motion of the gas in which they formed @xcite .
it has been clear for a long time that the `` bulges '' of spiral galaxies are often disk - like .
the first clues were kinematic , with some bulge - like components displaying very high ratios of rotational velocity to velocity dispersion ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
_ hst _ imaging reveals disk - like and flattened central isophotes ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , bars , ovals , nuclear spirals and rings in many bulge - like components ( e.g. , * ? ? ?
* ; * ? ? ?
* ) as well as ongoing star formation @xcite .
to distinguish these systems from `` classical '' bulges , they are referred to as `` pseudobulges '' .
it is thought that secular , internal processes ( e.g. , bars or spiral arms ) are responsible for building pseudobulge mass , in contrast to the rapid merging processes , i.e. violent relaxation , thought to build elliptical galaxies and classical bulges .
the literature is reviewed in detail in @xcite . throughout
, we will use the term `` bulge '' to refer to any luminous , centrally concentrated galaxy center , where classical bulges are of the elliptical type and pseudobulges are essentially disks .
different authors use different ( overlapping ) criteria to identify pseudobulges .
@xcite start with a sample of morphologically selected pseudobulges , meaning that a nuclear bar , spiral , or ring is detected within the bulge region ( see also * ? ? ?
they show that the distribution of srsic indices in local bulges is bimodal , and suggest that pseudobulges are galaxies with @xmath108 ( for a definition of the srsicindex , see 5.1 ) .
these authors also show that the average b / t of pseudobulges ( 0.16 ) is lower than that of classical bulges ( 0.4 ) with a large spread .
@xcite , on the other hand , advocates use of the @xcite relation as a discriminator , since pseudobulges tend to have lower central surface brightnesses at a fixed radius ( see also * ? ? ?
* ; * ? ? ?
* ) . finally , while classical bulges are typified by old stellar populations , pseudobulges tend to have ongoing star formation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
since deriving robust velocity measurements is beyond the scope of this paper , and in the absence of more robust structural information , we rely on morphology and stellar population properties at the present time .
the nearest , well - studied targets in our sample ( ngc 4388 , and ngc 2273 ) probably contain pseudobulges . in the case of ngc 2273 this classification is based on both the young stellar populations and the rings and nuclear disk .
ngc 4388 is less certain , but there is clear evidence for recent star formation and dust .
we suspect that ngc 6264 contains a pseudobulge , given its morphological similarities with ngc 2273 ( namely the outer ring and inner bar ) and the evidence for young stars .
the same goes for ngc 3393 and ic 2560 , which each contain an outer ring , a bar , and an inner ring .
on the other hand , ngc 1194 , with both evolved stellar populations and a large bulge , probably contains a classical bulge .
ngc 2960 has some of the clearest evidence for ongoing star formation , and so we tentatively put it into the pseudobulge category .
finally , we remain agnostic about ngc 6323 , which is one of the most distant targets .
thus , of the nine targets we consider , at least seven likely contain pseudobulges .
-0 mm -0 mm
in figure 8 we present the location of the megamaser galaxies in the @xmath5 plane .
the maser galaxies do not follow the extrapolation of the @xmath5 relation defined by the elliptical galaxies . instead , they scatter towards smaller bh masses at a given velocity dispersion .
quantitatively , taking @xmath109 @xmath0@xmath110log(@xmath0 ) @xmath111 log[@xmath112(@xmath6 ) ] , where log[@xmath112(@xmath6 ) ] is the expected @xmath0 given @xmath6 , we find @xmath113@xmath114@xmath115 dex .
there are many hints in the literature that the @xmath5 relation does not extend to low - mass and late - type galaxies in a straightforward manner ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , the precision bh masses afforded by the maser galaxies make a much stronger case .
the @xmath5 relation is not universal .
neither the shape nor the scatter of the elliptical galaxy @xmath5 relation provides a good description of the maser galaxies in this plane .
we now add the maser galaxies to the larger sample of local galaxies with dynamical bh masses to show that indeed a single , low - scatter power - law does not provide an adequate description of all galaxies in the @xmath5 plane . for convenience and to facilitate comparison with previous work , we assume a power law for all fits , although that form may not provide the best description of the sample as a whole .
we fit not only the full sample , but also divide the galaxies by morphological type . for ease of comparison , we adopt the divisions of @xcite . in essence , the galaxies are divided into two primary morphological bins : elliptical and spiral galaxies . for completeness we perform fits assuming that s0 galaxies belong to each subgroup . we adopt nomenclature from gltekin et al .. `` elliptical '' refers only to the elliptical galaxies , while `` early - type '' refers to elliptical and s0 galaxies .
`` late - type '' galaxies are spiral galaxies excluding s0s while `` nonelliptical '' includes all spiral galaxies ( s0 and later ) .
all fits are shown in table 4 and figure 9 .
in addition to these two divisions , we also consider pseudobulges by themselves .
early work suggested that pseudobulges ( see 5.3 ) obey the @xmath5 relation of classical bulges based on a very small sample of galaxies @xcite . further study , based on larger samples , has pointed to differences both in the slope and scatter of bh - bulge relations in pseudobulges @xcite , but the results are not conclusive , predominantly because the available number of dynamical bh masses in pseudobulges is so small .
the maser sample mitigates this problem . as a fiducial sample ,
we adopt the pseudobulges tabulated by gltekin et al .. however , both ngc 3245 and ngc 4342 were classified as classical bulges by @xcite and @xcite respectively .
thus , we have performed a second fit ( pseudo2 ) excluding these two galaxies .
while the slopes differ between the two fits ( although not significantly ) , the intrinsic scatter does not change . to facilitate direct comparison , we adopt the fitting procedure of gltekin et al .. as is commonly done , we fit the relation : log ( @xmath0/@xmath2)@xmath116 .
we use a generalized maximum likelihood analysis that is designed to incorporate intrinsic scatter in a natural manner . for simplicity , we assume a gaussian distribution in both the measurement errors and the intrinsic scatter ( directly investigated by gltekin et al . ) . for a set of observed points ( @xmath117 ) ,
we maximize the total likelihood : @xmath118 in the presence of measurement errors , if the likelihood of measuring a mass @xmath119 for a true mass @xmath112 is @xmath120 , then for a given observation the likelihood is : @xmath121 we assume both @xmath122 and @xmath123 to have a log - normal form .
upper limits are included , where the mass is taken to be below the upper limit with a gaussian probability given by the measurement uncertainty ( see details in appendix a of gltekin et al . ) .
their inclusion does not change the fits significantly . since errors in the independent variable
are not naturally included , we run 2000 monte carlo simulations for each fit in which the dispersion values are drawn from a gaussian distribution around the measured value , assuming symmetrical errors in the logarithm . the reported error is a quadrature sum of the error from the fit and the width of the final distribution of fitted values from the monte carlo runs . as a sanity check , we also employ a @xmath47 fitting technique adapted from @xcite . in this formalism ,
intrinsic scatter is included as an additional term such that the final reduced @xmath47@xmath124 .
while we only present results from the maximum likelihood treatment , we note that the fits are consistent between the two methods in all cases . the formal
fitting quantifies the visual impression of figure 8 . at low bh mass ( @xmath0@xmath125 @xmath2 ) , galaxies begin to deviate significantly from the @xmath5 relation of elliptical galaxies .
the zeropoint is lower by @xmath126 dex and the scatter is larger by @xmath127 dex .
it is worth noting that , in agreement with gltekin et al . , the fits to the pseudobulge galaxies are in no way distinguishable from fits to either the late - type or nonelliptical subsamples .
our data demonstrate that the power - law relationship between @xmath0 and @xmath6 is not universal .
specifically , we find a population of bhs that apparently have not fully grown to the present day @xmath5 relation of elliptical galaxies .
our measurement techniques may not yield @xmath6 values that are directly comparable with those in elliptical galaxies .
there is firstly the question of whether we are contaminated by disk light . from the sdss imaging
, it is quite clear that the bulge completely dominates the light within 5 . as an additional sanity check , in figure 10_a _
we look for correlations between the galaxy axial ratio ( as proxy for the inclination ) and deviations from the @xmath5relation , @xmath109 @xmath0 .
there is no compelling evidence for a correlation between the two ( kendall s tau value of @xmath128 with a probability @xmath129 that no correlation is present ) , suggesting that our @xmath6 measurements are not strongly biased by rotation in the central regions .
we still worry about aperture biases in our @xmath6 measurements , because we do not fully sample the radial behavior of @xmath6 .
we have measured @xmath6 within apertures ranging from 2 to 10 , and the values only vary by @xmath130 within this range .
to reiterate , we are not reporting @xmath6 within @xmath131 at this time because we do not wish to fold the ( large ) uncertainties in the structural measurements into our dispersion measurements . in any case
, the radial variations in @xmath6 are insignificant on the scales we can currently measure . on the other hand , as ngc 2273 demonstrates , two - dimensional @xmath6 profiles with higher spatial resolution may well reveal a decline in @xmath6 within some galaxy centers .
of course , there is the much more serious concern that these galaxies do contain small classical bulges .
the recent paper by @xcite presents direct evidence that isolating the classical bulge component in composite systems can bring later - type galaxies in line with the local elliptical galaxy @xmath5 relation .
we hope to investigate this concern directly in future work .
using the subsample of galaxies with sdss data , we are able to investigate the correlation between @xmath0 and @xmath132 . we first show the @xmath7 relation in figure 11_a _ , along with the best - fit relation from @xcite .
the @xmath133band luminosities are calculated from the @xmath134-band magnitudes and @xmath89 colors , using a grid of average galaxy colors derived from the @xcite templates .
we caution that the sample is still small .
furthermore , we are only spatially resolving the inner @xmath135 pc of these galaxies , while the nuclear spiral in ngc 2273 is @xmath136 pc . at the same time
, dust extinction is clearly significant in all of these systems , and while @xmath91 now includes dust lanes , the uncertainties are substantial ( e.g. , b / t can vary by nearly a factor of two depending on our treatment ) . with these caveats in mind ,
we find that the location and scatter of the megamaser galaxy bulges analyzed so far is consistent with that of the early - type inactive sample .
we see a slight tendency for the megamaser galaxies to be overluminous at a fixed bh mass . given that these are spiral galaxies , with bulge colors that are significantly bluer than typical elliptical galaxies , the offset may be attributed to stellar population differences .
thus , in figure 11_b _ we also show the @xmath137 relation , where the bulge @xmath90is inferred from the bulge @xmath89 color ( table 2 ) using the relation of @xcite .
we find that these maser galaxies are consistent with the fiducial @xmath137 relation . on the other hand
, we find no evidence that @xmath0 correlates with total galaxy luminosity or mass @xcite .
0.2 in if the @xmath5 relation does not hold for these objects , while @xmath137 does , the galaxies ought to be outliers in the @xcite relation , which we briefly investigate now . in principle
, examining the location of the maser galaxy bulges in the faber - jackson relation is an additional tool for evaluating bulge morphology . as seen in figure 12 , these five galaxies are consistent with the faber - jackson relation seen in the @xcite sample .
we have argued that a large fraction of the galaxies in our sample contain pseudobulges ( with the probable exceptions of ngc 1194 and ngc 6323 ) .
the literature on the faber - jackson relation in pseudobulges is limited .
@xcite find that while the faber - jackson relation of elliptical galaxies defines an upper envelope for the pseudobulges , they tend to scatter toward low values of @xmath6 at a fixed @xmath138band luminosity ( see also * ? ? ?
* ; * ? ? ?
the general finding that pseudobulges are more diffuse at a given luminosity ( e.g. , * ? ? ?
* ) is in the same sense . on the other hand ,
@xcite report the opposite , namely large @xmath6 at a fixed stellar mass .
finally @xcite report no difference in the @xmath139band faber - jackson relation of late - type spirals , but they are comparing only to early - type spirals , not elliptical galaxies . we await a larger maser sample to draw definitive conclusions .
we have demonstrated that there is a population of bhs that scatter below the @xmath5 relation defined for elliptical galaxies . in this section
we explore some of the ramifications of a mass - dependent @xmath5 relation both for our understanding of bh demographics and for our ability to infer bh masses from @xmath6 in low - mass galaxies .
-0 mm -0 mm -0 mm it is common practice in the literature to take the observed distribution of bulge velocity dispersions or magnitudes and infer the local bh mass function ( e.g. , * ? ? ?
* ; * ? ? ?
if either the scatter or the shape of these relations change at low mass , then the shape and amplitude of the inferred bh mass function for masses @xmath125 @xmath2 will also change . increased scatter will tend to broaden the overall distribution , thus increasing the relative density of objects at high and low mass .
the putative offset and change of slope , in contrast , decrease the overall mass density for a fixed distribution in @xmath6 . at present
it is unclear whether the mass functions inferred from bulge luminosity will change .
for this reason , and because @xmath6 distributions are not well - measured in the regime of interest , we do not present a new bh mass function , but merely note that current estimates of the space densities of bhs with @xmath140@xmath0/@xmath2@xmath141 are probably overestimates .
this is relevant , for instance , to projected tidal disruption rates in present and upcoming time - domain surveys ( e.g. , * ? ? ?
currently , the most direct means to estimate bh masses in active systems comes from reverberation mapping , in which a size scale for the broad - line region is derived from the lag between continuum variability and the corresponding variability from the photoionized broad - line region gas .
combining the broad - line region size with the width of an emission line yields a `` virial '' bh mass to within a scaling factor @xmath142 that depends on the kinematics and structure of the broad - line region .
eventually we hope to measure @xmath142 in individual agns and we are making good progress ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
-0 mm for the moment , an average @xmath142 is derived as the bh mass zeropoint that brings ensembles of low - redshift , low - mass agns into agreement with the @xmath5 ( e .. g , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or the @xmath7@xcite relations .
if the @xmath5 relation is intrinsically different for low - mass or late - type systems , then @xmath142 must change as well , since the majority of the systems used to derive @xmath142 have low masses and late types @xcite . in order to illustrate the sense of the shift ,
we derive @xmath142 here based on our new @xmath5 relation for either the nonelliptical ( including s0 galaxies ) or the late - type ( excluding s0 ) galaxies . note that we are ignoring the large uncertainties in the measured slope and zeropoint , simply as illustration .
our procedure is very simple .
we fix the slope of the relation as derived from one of the fits above , and then perform a least - squares fit for the zeropoint of the active galaxies using the unscaled virial products ( e.g. , @xmath143@xmath144/@xmath145 ) as tabulated by @xcite . starting with the @xmath5 relations ( table 4 )
we derive @xmath146 using the nonelliptical and late - type fit respectively .
the most recent calibration of @xmath142 , from woo et al . , is @xmath147 .
thus , we infer that @xmath142 may be a factor of @xmath148 lower than the nominal value used in the literature ( fig .
13 ) . in certain contexts
a difference of this size matters .
most notably , studies looking at potential evolution in the @xmath5 or @xmath137 relations with cosmic time uniformly adopt the @xmath142 value from onken et al .
@xcite . however , the @xcite sample is clearly composed predominantly of late - type galaxies . if they were to adopt a lower value of @xmath142 , then the evidence for redshift evolution , which they report to be a factor of @xmath149 at @xmath150 , could disappear @xcite .
we note also that the mild offset that is apparent in the @xmath5relation of the lowest - mass agns ( e.g. , * ? ? ?
* ; * ? ? ?
* ) would also disappear with this revised scaling . on the other hand , scaling relations at these low masses are particularly complicated , given that we have no direct @xmath0 measurements @xcite .
@xcite find a very well - defined relation between bh mass and bulge luminosity for active galaxies with reverberation mapping . with the proper imaging in hand , we will attempt the same exercise using the @xmath137relation of the maser galaxies .
we wish to make one more point related to active galaxies in particular .
two recent papers have found that _ total _
, rather than bulge , luminosity correlates more tightly with @xmath0 in active systems and that the relation does not evolve with cosmic time out to @xmath151 @xcite .
the favored explanation is that secular processes are growing the bulge components from the galaxy disks such that the correlation with total galaxy luminosity is preserved . in figure 11_a
_ we show the bennert et al .
relation between _ total _ galaxy luminosity and bh mass measured for local reverberation - mapped sources ( _ dashed line _ ) . in addition , we include a number of nearby reverberation - mapped galaxies that both bennert and @xcite exclude from their samples ( ngc 3227 , ngc 3516 , ngc 4051 , ngc 4151 , ngc 5548 , ngc 7469 ) . unlike the bennert objects , neither the ngc galaxies nor the maser galaxies obey a correlation between total galaxy luminosity and @xmath0 .
indeed , the lack of correlation with total galaxy mass in inactive samples was an original motivation to focus on bulge properties ( e.g. , * ? ? ? * , kormendy , j. et al . in preparation ) .
what we do not understand at the moment is why , for certain subsets of the reverberation - mapped and higher - redshift agn samples , total galaxy luminosity appears to provide a tighter correlation with @xmath0 .
we are tempted to blame uncertainties in the zeropoint of the agn masses ( see above ) .
however , we do not yet have a full explanation for this apparent discrepancy .
ideally , we would like to use our results to draw more general conclusions about the population of bhs with @xmath0@xmath152 @xmath2that live predominantly in spiral galaxies .
however , we must also consider the possibility that the maser sources represent a biased sample .
first of all , the selection process may be biased by the presence of nuclear activity .
were we dealing with distant quasars , we might be selecting on bh mass , while in local inactive systems we select on galaxy properties .
however , these agns are heavily obscured , and at nearly every wavelength stellar light dominates the galaxy luminosity on large spatial scales .
thus , the masers were selected first by virtue of their galaxy luminosity and then based on emission - line ratios and we do not expect a large selection bias based on nuclear activity .
a more serious concern is that we do not see a representative snapshot of the galaxy population when focusing on accreting bhs .
active bhs have not necessarily reached their final masses , and thus may be growing towards the inactive @xmath5 relation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the bhs presented here are radiating at @xmath40 of their eddington limits , and thus would need @xmath153 gyr to grow onto the inactive relation .
this is somewhat long compared to the @xmath11 year lifetimes typically estimated for quasars ( e.g. , * ? ? ?
typical growth times for pseudobulges , in contrast , are probably many gyr ( e.g. , * ? ? ? * ) .
therefore , the notion that these particular bhs knew to accrete material at this moment in order to move towards the inactive relation seems somewhat contrived . in addition , we plot @xmath114 against eddington ratio in figure 10_b_. there is no correlation between these two quantities ( kendall s @xmath128 , probability of no correlation is @xmath129 ) .
we see no evidence that the maser galaxies provide a biased tracer of bh - bulge scaling relations .
we consider the host galaxy properties of a new sample of active galaxies with megamaser disks .
fitting of the maser spots yields bh masses with @xmath154 precision ( c. y. kuo et al . in preparation ) .
bh mass measurements in the maser galaxies more than double the sample of galaxies with dynamical bh masses of @xmath0@xmath155 @xmath2 , as well as increasing significantly the number of spiral galaxies with dynamical bh mass measurements .
our primary results are as follows : 1 .
there is a population of spiral galaxies , probed by the maser measurements , with bhs that systematically lie below the @xmath5 relation of massive elliptical galaxies . as a result
, the @xmath5 relation fits to later - type and lower - mass galaxies display both a larger scatter and lower zeropoint than fits to elliptical galaxies alone ( 6 ) .
there is no universal power - law @xmath5 relation .
2 . from our limited data set , the galaxies appear to obey scalings between @xmath0 and bulge luminosity or mass . on the other hand @xmath0does
not correlate strongly with total galaxy luminosity for the maser galaxies ( 7 ) .
3 . to rule out definitively the presence of small classical bulges buried in the bars , dust , and young stellar populations found at the centers of these galaxies , we require spatially resolved kinematics and high - resolution imaging , which will fully characterize the complicated bulge regions of these galaxies . 4 .
changes in the @xmath5 relation at low mass translate directly into uncertainties in the local bh mass function for @xmath0@xmath156 @xmath2 .
furthermore , they imply ambiguity in the overall scaling of bh masses in active galaxies based on reverberation mapping .
we now indulge in speculation about the physical ramifications of our results .
generic arguments about bh self - regulation , taken at face value , seem to predict a simple power - law relationship between @xmath0and @xmath6 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , although the shape and scatter of the relation may depend on @xmath0(e.g .
* ; * ? ? ?
* ; * ? ? ?
this relation is established when the bhs grow large enough that their accretion power can overcome the galaxy potential . on the other hand ,
the maser galaxies have systematically lower bh masses at a fixed @xmath6 . perhaps
because of the quiescent history of these spiral galaxies , there has never been a major feeding episode in the past history of their bhs .
if low - mass galaxies do not feed their bhs very effectively , then the bh never approaches the `` limiting '' value set by the galaxy potential @xcite . presumably the same processes that make a galaxy into an elliptical or spiral ( merging history ) also determine the bh feeding efficiency . in this scenario ,
we expect a measurable difference in scatter between elliptical and spiral galaxies with @xmath0@xmath152 @xmath2 .
we note that ongoing adaptive - optics assisted integral - field unit observations are exploring the low - mass regime for elliptical galaxies with improved statistics @xcite , which will allow us to perform this test .
alternatively , @xcite proposes that bh - bulge scaling relations are tightened at high mass due to multiple minor mergers , which , by the central limit theorem , yield a tight @xmath5 relation at the present day ( see also * ? ? ?
* ; * ? ? ?
* ) . in this picture
the only important parameter is number of mergers , rather than detailed galaxy morphology .
here we might expect to find increased scatter in both the elliptical and spiral galaxies at @xmath0@xmath125 @xmath2 .
it is interesting to note , in passing , that one might also expect globular cluster frequency to correlate more tightly with @xmath0 than bulge luminosity @xcite .
furthermore , we might also expect this process to lower @xmath0/@xmath132 over cosmic time ( e.g. , * ? ? ?
this paper represents an initial exploration of scaling relations for the maser galaxies , but there are many experiments we would like to perform . as noted above , further observations are required to determine whether or not there are small classical bulges embedded in all of these galaxies , whose @xmath6 and @xmath157 correlate tightly with @xmath0 @xcite .
integral - field spectroscopy would reveal the @xmath6 profile and allow us to determine whether there is a transition scale at which we recover a tight correlation between @xmath0and @xmath6 .
multi - color photometry , ideally with _
hst _ imaging of the nuclei , will allow us to decompose the light profile into its constituent parts , measure stellar masses for each , and then determine whether the bh mass correlates best with any particular combination of components .
there are other potentially interesting bh - bulge scaling relation projections to investigate as well .
for instance , it may be that @xmath0 correlates more tightly with the dynamical mass than @xmath6or stellar mass ( e.g. , * ? ? ? * ; * ? ? ?
there is also a suggestion in the literature that @xmath0 actually correlates with the dark matter halo mass ( e.g. , * ? ? ?
* ) , although there are clearly complications in spiral galaxies ( e.g. , * ? ? ?
at present only a couple of the galaxies have resolved rotation curves , but we will obtain observations of the entire sample with the evla in order to investigate both the gas content and the @xmath0-@xmath158 correlation in these galaxies . beyond the present work
there are now many suggestions that the @xmath5 relation may not be the unbroken , low - scatter power - law originally proposed ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . with the larger samples available to date ,
the scatter has increased by @xmath159 dex @xcite . at the same time , new analysis incorporating the impact of dark matter at large radius @xcite has shown that the bh masses in the most massive ellipticals may be systematically underestimated by a factor of @xmath160 .
finally , new dynamical models incorporating triaxiality may well increase the scatter in bh mass at a fixed bulge mass as well ( e.g. , * ? ? ?
given that there is no universal @xmath5 relation , it may be time to revist the role the bhs play in galaxy evolution . + 0.5truein 1.8 in -0 mm in addition to the b&c data discussed in the primary text , we obtained auxiliary spectra for some of the southern targets in one magellan / baade night .
we used the inamori - magellan areal camera & spectrograph ( imacs ; * ? ? ? * ) with the 1200 l mm@xmath13 grating .
the primary setting had a tilt of 17.15 with a resulting central wavelength of 4540 , a spectral range of 3800 - 5300 , and a dispersion of @xmath161 km s@xmath13 .
we also obtained cat observations for two targets , with a tilt of 32.8 , a resulting central wavelength of 8340 , and a spectral range of 7500 - 9000 .
the observing conditions were not optimal ; it was only four days from full moon , and the typical seeing ranged from 07 to 14 over the course of the night .
imacs consists of an array of four by two ccds with considerable tilt in the spatial direction .
therefore we used the cosmos code developed by g. oemler , d. kelson , and g. walth to reduce long - slit imacs data .
the cosmos code has a variety of benefits over standard long - slit reductions of the imacs data .
for one thing , because the size of the chip gaps is built into the code , we are able to solve for a global ( rather than a chip - by - chip ) wavelength solution in a straightforward manner .
furthermore , the code contains an optical model for the instrument and thus does a better job of modeling curvature in the spatial direction .
we perform bias - subtraction , flat - fielding , dispersion - correction , and rectification within cosmos and then use standard iraf routines for tracing and extracting one - dimensional spectra , including sky subtraction . finally , we use idl routines as described by @xcite to perform flux calibration and telluric absorption corrections .
in addition to the imacs spectra , three targets ( ngc 1194 , ngc 4388 , and ngc 6264 ) have spectra in the sdss database .
these have a spectral resolution of @xmath162 km s@xmath13(e.g .
* ) and completely cover the g - band region .
we use valdes template stars for the sdss spectra as well as the dis spectra .
the velocity dispersion measurements proceeded as above , except that our library of template stars taken with imacs is limited to g and k giants .
thus we use a large library of template stars provided by the elodie catalog @xcite .
these stars were observed with @xmath163 or @xmath164 km s@xmath13 resolution . in all cases
we adopt the following stars as our templates : hd025457 ( f5 v ) , hd148856 ( g8 iii ) , hd185351 ( k0 iii ) , and hd124897 ( k1.5 iii ) .
we fit our internal velocity dispersion templates with the elodie stars in order to determine the relative broadening between the two systems . in the case of imacs
the spectral resolution varies significantly as a function of wavelength , but in the range @xmath165 , fits to the imacs stars with the elodie templates yield a relative dispersion of @xmath166 km s@xmath13 . as an additional consistency check ,
we take our measured dispersions from the internal velocity stars and compare them to those measured with the elodie stars in the identical wavelength range , where the latter are corrected for the resolution difference in quadrature .
the result of that test is shown in figure 14 .
the agreement is completely satisfactory , with a fractional difference of @xmath167 .
we are reassured both that there is no significant systematic offset between the two sets of measurements and that the scatter is considerably smaller than any other source of uncertainty .
overall , we find satisfactory agreement between the imacs and b&c measurements ( table a5 ) .
the one exception is the galaxy ngc 3393 . in this case ,
our imacs data were not taken along the major axis , but rather at pa=12@xmath168 .
since the imacs measurement is nominally consistent with our minor axis measurement , we suspect this pa difference may at least partially explain the discrepancy .
furthermore , we are able to test our flux calibration by comparing the fluxes in [ ] between the two sets of observations
. since we can match approximately the slit size and position angle , any discrepancies should predominantly reflect differences in flux calibration .
it is reassuring that our measurements agree within @xmath169 in all cases but ngc 2960 , where the [ ] line falls on a noisy part of the imacs chip . ,
a. , hare , t. , bigelow , b. c. , & osip , d. j. 2006 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
6269 , society of photo - optical instrumentation engineers ( spie ) conference series , k. , richstone , d. o. , gebhardt , k. , lauer , t. r. , tremaine , s. , aller , m. c. , bender , r. , dressler , a. , faber , s. m. , filippenko , a. v. , green , r. , ho , l. c. , kormendy , j. , magorrian , j. , pinkney , j. , & siopis , c. 2009 , , 698 , 198 , p. t. , greenhill , l. j. , moran , j. m. , lovell , j. e. j. , kuiper , t. b. h. , jauncey , d. l. , cameron , l. b. , gmez , j. f. , garca - mir , c. , moll , e. , de gregorio - monsalvo , i. , & jimnez - bailn , e. 2006 , , 638 , 100 | the black hole ( bh)-bulge correlations have greatly influenced the last decade of effort to understand galaxy evolution .
current knowledge of these correlations is limited predominantly to high bh masses ( @xmath0@xmath1 @xmath2 ) that can be measured using direct stellar , gas , and maser kinematics .
these objects , however , do not represent the demographics of more typical @xmath3 galaxies .
this study transcends prior limitations to probe bhs that are an order of magnitude lower in mass , using bh mass measurements derived from the dynamics of h@xmath4o megamasers in circumnuclear disks .
the masers trace the keplerian rotation of circumnuclear molecular disks starting at radii of a few tenths of a pc from the central bh .
modeling of the rotation curves , presented by kuo et al .
( 2010 ) , yields bh masses with exquisite precision .
we present stellar velocity dispersion measurements for a sample of nine megamaser disk galaxies based on long - slit observations using the b&c spectrograph on the dupont telescope and the dis spectrograph on the 3.5 m telescope at apache point .
we also perform bulge - to - disk decomposition of a subset of five of these galaxies with sdss imaging .
the maser galaxies as a group fall below the @xmath5 relation defined by elliptical galaxies .
we show , now with very precise bh mass measurements , that the low - scatter power - law relation between @xmath0 and @xmath6seen in elliptical galaxies is not universal .
the elliptical galaxy @xmath5 relation can not be used to derive the bh mass function at low mass or the zeropoint for active bh masses .
the processes ( perhaps bh self - regulation or minor merging ) that operate at higher mass have not effectively established an @xmath5 relation in this low - mass regime . |
the big bang cosmology has been greatly successful in explaining the evolution of the universe . however , despite its success
, it falls short in describing underlying nature of some fundamental physics problems .
for example , three of the leading unsolved mysteries include cosmic inflation , dark matter and dark energy , whose underlying descriptions remain yet unknown . in attempting to solve
some ( all ) of the problems , cosmologists and gravitational physicists have sought out the possible modifications of the einstein gravity on the cosmic scales . among many scenarios , theories that include an additional scalar field
provide compelling candidates for alternative theories of gravity
. in general , the modifications may introduce new gravitational degrees of freedom in addition to the metric tensor and can be in principle described by a scalar - tensor theory of gravity @xcite .
however , in any non - degenerate theory that time derivative of fundamental dynamical variable is higher than second order , there exist a linear instability , or ostrogradski s instability .
such instabilities whose hamiltonian is not bounded from below lead to negative norm states or negative energy states which in higher derivative theories are terminologically called ghost - like . besides , theories with ghostlike degrees of freedom provide inconsistencies with the experimental tests . to avoid the presence of the ostrogradski instability , the euler - lagrange equations have to be at most second - order .
until recently , it turns out that very successful models of modified gravity can be described in terms of a class of horndeski s scalar - tensor theory @xcite . in spite of the existence of the derivative interactions ,
the horndeski s theory is known as the most general scalar - tensor theory in four dimensions with one scalar degree of freedom whose equations of motion are kept up to second order in time and spatial derivatives and therefore they are deprived of the ostrogradski s instability .
the horndeski s theory has been investigated so far for various cosmological purposes , e.g. dark energy @xcite , screening mechanisms @xcite and also inflation @xcite . in attempting to generalizations
, we may consider a scenario that the scalar field is directly coupled to the matter sector . in this case , matter does not follow geodesics associated with the gravitational metric @xmath2 but instead with another metric @xmath3 that in the simplest situation they are related via @xmath4 which is known as the conformal transformation ( or coupling ) @xcite . here
it is worth noting that the matter frame metric @xmath3 can be constructed by the purely gravitational one @xmath2 and only one scalar field @xmath0 , but not by the derivatives of the scalar field itself .
the gravity and matter frames , @xmath2 and @xmath3 , are often referred to as the einstein and jordan frames , respectively .
as the most general vistas of the scalar - tensor theory , the matter frame metric which is conformally constructed from the gravitational metric and the scalar field itself can also be further generalized by adding the derivatives of the scalar field . since the higher derivative terms in the equations of motion give rise to ghost - like instabilities associated with the ostrogradski s theorem , the simplest viable case is to keep only the first order derivatives of the scalar field , _
viz_. ; @xmath5 where @xmath6 is the covariant derivative of the scalar field associated with the gravity frame @xmath2 and @xmath7 .
note that the above relation is often called the disformal transformation . in the present work we will restrict ourself to the simpler argument in which the disformally - transformed gravitational sector belongs to a subset of the horndeski class @xcite , namely ; @xmath8 for @xmath9 , eq([disfort ] )
reduces to the ordinary conformal transformation and its consequences have been well - established , at least when it depends on @xmath0 only , while when @xmath10 , @xmath11 features the pure disformal transformation .
regarding the scalar - tensor theory with the disformal coupling , there are plenty of compelling applications in the cosmological problems .
for instance , the authors of ref.@xcite showed that the disformal transformations are very useful to devise models of the dark sector .
in particular , inflationary models have also been investigated in contexts of the disformal coupling @xcite . however , another crucial issue for successful models of inflation is the ( pre)reheating mechanism . in this work ,
we anticipate to investigate this mechanism for an inflationary model with the presence of a disformal coupling between two scalar fields .
the reheating and preheating mechanisms for inflationary models in which the couplings between inflaton and matter fields are induced from conformal transformation have been studied in @xcite .
the reheating process in inflationary scenario with disformal coupling between inflaton and other scalar field has been discussed in @xcite .
the preheating process due to the direct kinetic coupling between inflaton and other scalar field has been investigated in @xcite .
the paper is organized as follows : in sec.([s2 ] ) , we discuss the evolution of the background system and then we quantify inflationary trajectories by specifying a particular choice of free functions , @xmath12 and @xmath11 . in sec.([sec3 ] ) , we will investigate the particle production due to parametric resonances in a model of inflation in which the disformally - transformed gravitational sector belongs to a subset of the horndeski class eq.([disfort ] ) . for this model of inflation , the scalar field @xmath0 ( which is the inflaton field ) naturally coupled to another scalar field @xmath1 induced by the disformal transformation .
we summarize our findings in the last section .
we will here start by deriving the set of transformations relating background fields of a disformally coupled theory .
the full action for our model is described by the following action ( in planck units ) : @xmath13 - \int d^{4}x\sqrt{-{\tilde g}}\big[\frac{1}{2}{\tilde g}^{\mu\nu}\chi_{,\mu}\chi_{,\nu } + v(\chi)\big]\,.\end{aligned}\ ] ] bearing in mind that the model can be seen as a scalar - tensor theory in the einstein frame coupled to matter fields that propagate on a physical metric denoted by @xmath14 .
the relation between the two - frame metrices is given by the disformal transformation , defined in eq.([disfort ] ) .
the transformation features how the physical metric for the matter in the action for this theory is disformally related to the gravitational metric .
the stress energy momentum tensors in the jordan frame is defined by @xmath15 for which we can commonly impose a perfect fluid description by defining a jordan frame energy density , @xmath16 , four - velocity field @xmath17 , and pressure ,
@xmath18 : @xmath19 in the einstein frame , we also have @xmath20 using the relations in eq.([jf ] ) and eq.([ef ] ) , a map between the two objects can readily be derived to obtain : @xmath21 where the disformal scalar @xmath22 given below in eq.([gammag ] ) parameterizes the relative contribution of the disformal factor . using a disformal relation to the relatively simple form of eq.([disfort ] )
, we can rewrite the action in terms of @xmath2 to yield @xcite @xmath23\,,\end{aligned}\ ] ] where the arguments of @xmath24 and @xmath25 are understood , and for convenience we have defined the parameter @xmath22 as @xmath26 in order to obtain the field equations , we perform the variation of the action ( [ pureg ] ) with respect to the metric @xmath2 to yield @xmath27 where @xmath28 is the usual energy - momentum tensor for a minimally - coupled scalar field and @xmath29 involves the cross terms due to the disformal coupling : @xmath30 and @xmath31g_{\mu\nu } + \frac{c}{\gamma}\chi_{,\mu}\chi_{,\nu } - 2\gamma d\left(\phi_{,\sigma}\chi^{,\sigma}\right)\chi_{,(\mu}\phi_{,\nu)}\nonumber\\&&+ \left[\gamma c \left(\frac{1}{2}g^{\alpha\beta}\chi_{,\alpha}\chi_{,\beta } + cv\right ) + \frac{\gamma^{3}d^{2}}{2c}\left(\phi_{,\sigma}\chi^{,\sigma}\right)^{2}\right]\phi_{,\mu}\phi_{,\nu}\,,\end{aligned}\ ] ] and @xmath32 is the trace of @xmath29 . here
we obtain the equation of motion for a field @xmath0 as @xmath33 and @xmath34 is given by a rather complicated expression : @xmath35 - \frac{1}{2c}\big[c't^{(\chi ) } + d't^{(\chi)\mu\nu}\phi_{,\mu}\phi_{,\nu}\big]\right)\,,\end{aligned}\ ] ] where primes denote derivatives with respect to the field @xmath0 .
likewise , varying the action with respect to the field @xmath1 we also obtain its equation of motion : @xmath36\left(\phi_{,\sigma}\chi^{,\sigma}\right)\nonumber\\&- & \gamma^{2}\frac{d}{c}\left[\phi^{,\mu}\phi^{,\nu}\nabla_{\mu}\chi_{,\nu } + \left(\phi_{,\sigma}\chi^{,\sigma}\right)\box\phi\right ] + \gamma^{4}\frac{d^{2}}{c^{2 } } \left(\phi_{,\sigma}\chi^{,\sigma}\right)\phi^{,\mu}\phi^{,\nu}\nabla_{\mu}\phi_{,\nu } = 0\,.\end{aligned}\ ] ] according to ref.@xcite , we will consider models where the couplings are given by @xmath37 here , we have used four parameters , @xmath38 and @xmath39 to describe our coupling functions .
however , an interesting special case of this parametrisation is when @xmath40 and the couplings become constants , @xmath41 and @xmath42 . inspired by string theory , the scale @xmath43 can be identified with the ( inverse ) tension of the 3-brane ( @xmath44 ) whose an associated disformal mass scale @xmath45 is given by @xmath46 @xcite . for model of inflation
, the authors of ref.@xcite discovered that fields with sub - planckian initial conditions can drive large amounts of inflation in the presence of a large enough disformal coupling .
here they used @xmath40 , @xmath47 and @xmath48 which corresponds to a mass scale of @xmath49 .
since @xmath22 takes an extremely large value , and remains fairly constant ( still much greater than unity ) until the end of inflation , it is claimed in @xcite that physics of preheating after inflation can be influenced by disformal coupling for this choice of the parameters .
however , we will see in the following section that the disformal coupling terms can drive parametric resonance in the preheating process as long as the term @xmath50 is not much smaller than unity although @xmath51 .
what we are interested in for the present work is to investigate the preheating process after inflation .
we will first assume that the spacetime and the inflaton field @xmath0 give a classical background and another scalar field @xmath1 is treated as a quantum field on that background .
we specialize the klein - gordon equations for the fields @xmath0 and @xmath1 to a cosmological background and find respectively @xmath52 \frac{c'}{c } - \frac{1}{2}\left(\gamma^{2 } - 1\right)\rho_{\chi}\frac{d'}{d}\ , , \end{aligned}\ ] ] and @xmath53\dot{\phi}\dot{\chi } \,,\end{aligned}\ ] ] where @xmath54 here , the energy density @xmath55 and pressure @xmath56 of the field @xmath1 are respectively given by @xmath57 from the einstein equation given in the previous section , we obtain the friedmann equation : @xmath58 where @xmath59 is the hubble parameter . in order to examine whether the disformal coupling can influence preheating process
, we study the evolution of the inflaton field @xmath0 and @xmath22 during inflationary stage by firstly assuming that the matter field @xmath1 has no contribution to the dynamics of inflation . in the case where the contribution from the field @xmath1 can be neglected and @xmath60 , eq.([kgphi ] )
becomes @xmath61 under the slow - roll evolution , it follows from eq.([h2 ] ) that @xmath62 and therefore the above equation becomes @xmath63 since @xmath64 is required for slow - roll evolution , the above equation implies that @xmath65 during inflation . from eq.([gamma ] ) , we see that @xmath66 becomes infinite in the limit when @xmath67 .
thus we have either @xmath68 or @xmath69 through out the evolution of the universe .
nevertheless , @xmath69 is not possible when @xmath70 oscillates around zero during preheating , and hence we consider only the case @xmath71 .
substituting @xmath70 from eq.([slowroll ] ) into eq.([gamma ] ) , we find that @xmath22 is larger when @xmath72 gets closer to unity .
according to the observational bound @xmath73 @xcite , @xmath22 will significantly larger than unity if @xmath74 .
the upper bound on @xmath22 is set by @xmath75 to ensure that the contribution from @xmath1 on dynamics of inflation can be neglected . in the following ,
we quantify the inflationary trajectory in two separate cases .
since @xmath70 is nearly constant during inflation , the simplest situation in which @xmath22 is large until the end of inflation is that a ratio @xmath77 is constant .
it follows from eq.([dc ] ) that the ratio @xmath77 is constant when @xmath78 , which includes the case of constant @xmath12 and @xmath11 if @xmath79 .
since the universe stops acceleration , i.e. , inflation ends , when @xmath80 , we find that @xmath70 drops by factor @xmath81 at the end of inflation . here
, @xmath82 is the value of the inflaton field at the end of inflation .
hence , we expect that at the end of inflation , the term @xmath83 is in the same order of magnitude with that during inflation .
we will see that this term can contribute to the parametric resonance process .
in addition to the case of a constant @xmath77 , a large @xmath22 at the end of inflation can also be obtained if @xmath77 increases when @xmath0 decreases .
since @xmath0 decreases near the end of inflation , the behavior of @xmath77 implies that although @xmath51 during inflation , @xmath22 can become larger when the value of @xmath0 significantly drops around the end of inflation .
this situation corresponds to the case where @xmath85 , @xmath86 .
after the end of inflation , @xmath70 oscillates with decreasing amplitude . to estimate the amplitude of @xmath22 and @xmath87 during preheating period
, we examine eq.([kgphia ] ) numerically by setting @xmath88 and setting @xmath89 as well as @xmath90 at the initial time of inflation .
the plots of @xmath22 and @xmath91 are shown in fig.([fig : evo ] ) the upper panals show the evolution of @xmath92 after inflation for the cases where @xmath79 ( left panal ) and @xmath93 ( right panal ) . the lower left panal shows the evolution of @xmath91 for the case where @xmath94 .
the lower right panal shows the evolution of @xmath22 . in this panal
, the lines `` 1 '' and `` 2 '' represent the cases where @xmath95 and @xmath94 , respectively . in the plots for the case @xmath76 , we set @xmath96 and @xmath97 , while we set @xmath98 and @xmath99 for the case @xmath100 . ]
supposing that the process of parametric resonance is at the stage in which the backreaction of the created particle can be neglected .
however , as already mentioned in ref.@xcite , the backreaction of the quantum field @xmath1 to the dynamics of the inflaton field will be relevant if its occupation numbers have grown sufficiently and then can inhibit the resonance particle production .
we will also leave this interesting topic for our future investigation .
another word of saying , we are interested in an epoch at which the inflaton is dominant , so that the evolution equation for @xmath0 during preheating is also given by eq.([kgphia ] ) .
it is trivial to figure out the solution of eq.([kgphia ] ) during preheating .
simply , we use for a power - law evolution of a scale factor @xmath101 and equation of motion becomes : @xmath102 the general solution of the effective equation of @xmath0 can be basically expressed in terms of the bessel functions as @xmath103\,,\end{aligned}\ ] ] where @xmath104 and @xmath105 are constants depending on the initial conditions at the end of inflation , and @xmath106 are bessel functions of order @xmath107 , with @xmath108 .
it is well known that the second term of eq.([sophi ] ) diverges in the limit @xmath109 . therefore , keeping the first term is good enough for the estimate purpose .
the physical solution to eq.([kgphiat ] ) is then reduced to : @xmath110 for a large argument expansion of fractional bessel functions such that @xmath109 , the physical solution can be approximately by a cosinusoidal function , see ref.@xcite : @xmath111 here we can choose a constant @xmath104 by considering the oscillatory behavior which starts just at the end of inflation .
as already mentioned above , the field value at the end of inflation is approximately given by @xmath112 .
following ref.@xcite , the energy and pressure densities associated to the physical solution ( [ sophi ] ) are given , after averaging over several oscillations , by @xmath113 = \frac{1}{2}m^{2}_{\phi}z^{2}\nonumber\\ & & p_{\phi } \approx \left<\frac{1}{2}{\dot \phi}^{2 } + \frac{1}{2}m^{2}_{\phi}\phi^{2}\right > \approx \frac{1}{2}m^{2}_{\phi}z^{2}\left[\left<\sin^{2}(m_{\phi}t - 3p\pi/4)\right > - \left<\cos^{2}(m_{\phi}t - 3p\pi/4)\right>\right ] = 0\,,\end{aligned}\ ] ] with @xmath114 .
since the averaged pressure is fairly negligible @xmath115 , eq([omega ] ) implies that @xmath116 with @xmath117 .
using the above approximations , the physical solution is finally expressed as @xmath118 as given in eq.([sophif ] ) .
the value of the scalar field here is measured in units of @xmath119 and time is measured in units of @xmath120 .
[ plotphi ] ] typically , it would be of great interesting for evaluating the hubble constant during the first oscillation . as shown in fig.([plotphi ] ) , during the first period of oscillation , the amplitudes of the field @xmath121 drops to around @xmath122 of the reduced planck mass , @xmath119 .
we may also expect during this early phase of oscillation that the field s kinetic energy is roughly equal to its potential energy , and hence we can estimate the energy density of the field to be @xmath123 . this approximation allows us to further estimate the hubble parameter and we find that the hubble rate would then be @xmath124 .
notice that the estimate for @xmath125 is consistent with the results found in ref.@xcite for completely different scenarios .
we next consider the dynamics of the quantum field @xmath1 .
since both c and d are in general functions of a scalar field , we first consider the fourth term of eq([kgchi ] ) and write for an coefficient of @xmath126 @xmath127 + \frac{c'\dot{\phi}}{c}\left(\frac{d{\dot{\phi}}^{2}}{c}\right ) - \frac{d'\dot{\phi}}{d}\left(\frac{d{\dot{\phi}}^{2}}{c}\right)\right)\nonumber\\&= & \frac{\gamma^{2}}{2}\left(\frac{d}{dt}\big[\frac{d{\dot{\phi}}^{2}}{c}\big ] + \frac{c'\dot{\phi}}{c}\left(\frac{\gamma^{2}-1}{\gamma^{2}}\right ) - \frac{d'\dot{\phi}}{d}\left(\frac{\gamma^{2 } -1}{\gamma^{2}}\right)\right)\nonumber\\&=&\frac{\gamma^{2}}{2}\frac{d}{dt}\big[\frac{d{\dot{\phi}}^{2}}{c}\big ] + \frac{c'\dot{\phi}}{2c}\left(\gamma^{2}-1\right ) - \frac{d'\dot{\phi}}{2d}\left(\gamma^{2 } -1\right)\,.\end{aligned}\ ] ] after substituting back into eq.([kgchi ] ) , we find @xmath128\dot{\chi}+\frac{c}{\gamma^{2}}v'(\chi ) = -\frac{c'}{c}\dot{\phi}\dot{\chi } \,.\end{aligned}\ ] ] the above expression can be further simplified by considering @xmath129\dot{\chi}+\frac{c'}{c}\dot{\phi}\dot{\chi } = \frac{d}{dt}\left(\log\big[\frac{ca^{3}}{\sqrt{1-\frac{d}{c}\dot{\phi}^{2}}}\big]\right)\dot{\chi}\,.\end{aligned}\ ] ] by defining a new parameter : @xmath130 \,,\end{aligned}\ ] ] and substituting the above relations eq.([kgchigeng]-[newa ] ) into eq.([kgchigen ] ) , we obtain @xmath131 where @xmath132 is defined as @xmath133 expanding the scalar fields @xmath1 in terms of the heisenberg representation as @xmath134 where @xmath135 and @xmath136 are annihilation and creation operators , we find that @xmath137 obeys the following equation of motion : @xmath138 fourier transforming this equation and rescaling the field by @xmath139 yields @xmath140 setting the potential @xmath141 as @xmath142 , eq.([ymodegen ] ) becomes @xmath143\right\ } y_{k}\ , , \label{dccont0}\end{aligned}\ ] ] where @xmath144 and @xmath145 . using @xmath146 and @xmath147 which implies @xmath148 during the preheating , the above equation becomes @xmath149 y_{k } \approx 0\ , , \label{dccont}\end{aligned}\ ] ] where we have neglected small contributions coming from terms like @xmath150 and @xmath151 . it is noticed that , however , @xmath152 is always less than unity although @xmath153 due to slow - roll evolution during inflation .
this quantity can increase in time but can not be larger than unity after inflation because @xmath22 becomes infinite when this quantity equals to one .
note that the form of this equation is very similar to that of ref.@xcite in the context of the derivative coupling between the inflaton and scalar fields .
surprisingly , preheating in our model may be qualitatively different from the cases considered in various literatures .
this is because the field has an approximatedly effective squared mass for @xmath154 given by @xmath155 by substituting the approximated solution given in eq.([sophif ] ) into the above expression , we get @xmath156 where we have again neglected small contributions coming from terms like @xmath150 and @xmath151 .
likewise , it is noticed that the @xmath157-dependent terms only contribute to the higher orders and can also be neglected in our analysis .
it is also worth noting that for @xmath158 the leading - order terms stem from the conformal coupling , while the higher ones come from the disformal coupling . in the situation where @xmath159 , the dominant terms in the above expression are @xmath160\,\frac{\phi_{e}}{m_{\phi}t}\sin(m_{\phi}t)\ , .
\label{meffr1}\end{aligned}\ ] ] however , for the case with @xmath161 , we instead have @xmath162 it follows from the above equation that if @xmath163 , the second line of eq.([meffr2 ] ) can be neglected . for simplicity
, we will assume that @xmath164 in the following consideration . in terms of @xmath165 , eq.([dccont ] )
becomes @xmath166 where a time - dependent frequency of modes @xmath167 is given by @xmath168 certainly , eq.([mat ] ) describes an oscillatory behavior with a periodically changing frequency @xmath169 . in case of @xmath170 , the physical momentum
@xmath171 coincides with @xmath172 for minkowski space such that @xmath173 .
the periodicity of eq.([mat ] ) may in principle lead to the parametric resonance for modes with certain values of @xmath174 . in order to quantify the parametric resonance behavior in our model , we consider two - separate cases as follows : in this case , @xmath176 is given by eq.([meffr1 ] ) .
the periodicity of eq.([mat ] ) for this case may lead to the parametric resonance for modes with certain values of @xmath174 . in order to examine this behavior , we will introduce a new variable , @xmath177 , defined by @xmath178 . in the minkowski space for which we neglect the expansion of the universe taking @xmath170 and use a simple trigonometric identity , eq.([mat ] ) becomes the standard mathieu equation : @xmath179 where @xmath180 regarding the mathieu equation , there are instability bands in which the modes @xmath167 grow exponentially with the growth index @xmath181 . as it was well known that these instability bands do depend on the parameters @xmath182 and @xmath183 . in order to ensure that the particle production is efficient
, the mathieu equation must satisfy the broad resonance regions for which the conditions @xmath184 where @xmath185 , and @xmath186 should be satisfied . nevertheless , in general
, the parameter @xmath183 can also depend on time and then decreases with time .
therefore it must take a large enough initial value . in the present analysis
, we discover that there are two cases for this model to initially provide large values for @xmath183 : ( i ) @xmath163 and ( ii ) @xmath187 . hence ,
for the disformal coupling model , the parametric resonance can proceed efficiently to reheat the universe .
this result is different from that obtained in the other derivatively - coupled model @xcite .
we now turn to the case when @xmath189 and assume that @xmath164 . again in order to examine the periodicity behavior of eq.([mat ] ) , we will introduce a new variable , @xmath177 , relating to @xmath190 via @xmath191 , and eq.([mat ] ) can be recast in the form of the standard mathieu equation : @xmath192 where @xmath193 for this case , there is no need for @xmath43 to be large since the broad resonance can occur when @xmath163 .
however , it decreases faster than that of our preceeding model , so that the broad resonance tends to disappear more quickly .
contrary to ref.@xcite , if @xmath194 , the amplitude of the driving force appearing in eq.([mathe13 ] ) can not be large enough for efficient parametric resonance due to the bound of @xmath195 from the observational constraints on inflation .
let us summarize our investigation by first comparing our results with some selected inflationary models .
the first model we anticipate to compare with is the study of preheating in derivatively - coupled inflation models @xcite . in this model , parameters take the form @xmath196 with @xmath197 being a vacuum expectation value of the inflaton field . as mentioned in ref.@xcite ,
an initial value of @xmath183 is more efficient for longer wavelength modes and lighter fields .
notice from eq.([ndc ] ) that a value of @xmath183 is at most of order one since @xmath198 .
if this is the case , any resonance proceeds close to the end of instability bands , and is ineffective if the field are very heavy , @xmath199 .
moreover , in a standard scalar chaotic inflation @xcite , the parameters take the form @xmath200 the result is very similar to that of our disformally - coupleld scenario and naturally gives large values for @xmath183 , initially .
however , it decreases faster than that of our model , so that the broad resonance tends to disappear more quickly . in this work
, we investigate the preheating mechanism in a disformally - coupled inflationary model where the scalar field @xmath0 ( which is the inflaton field ) naturally coupled to another matter field @xmath1 induced by the disformal transformation . in the present scenario ,
the emergence of novel derivative interactions mixing the kinetic terms of the fields is natural .
we start by deriving the evolution of the background system and then quantify how the disformal coupling influences inflationary trajectory .
we examine the particle production due to parametric resonances in the models and find in minkowski space that the stage of parametric resonances can be described by the standard mathieu equation .
interestingly in our analysis , we discover that broad resonances can be typically achieved and are potentially efficient causing the particle number density in this process to exponentially increase .
however , there are some limitations in the present work - for example , since the above presentation is based on the assumption that the backreaction is negligible , it will be of interest to take into account the effect of backreaction for more realistic model ; the effect of the expansion of the universe has to be also included .
due to initially large value of the parameter @xmath183 , the field may pass through many instability bands .
this behavior leads to a stochastic change in the growth index @xmath201 .
therefore , the system as a stochastic resonance @xcite is of great interest for our analysis . a thorough and comprehensive analysis towards eq.([mathe ] ) and eq.([mathe2 ] )
is numerically required , see e.g. @xcite , and we leave it for a future project .
the work of k. karwan is financially supported by the thailand research fund ( trf ) with grant no .
rsa5780053 and p. channuie thanks the thailand research fund ( trf ) for financial supports ( grant no .
trg5780143 ) .
a. de felice , t. kobayashi and s. tsujikawa , `` effective gravitational couplings for cosmological perturbations in the most general scalar - tensor theories with second - order field equations , '' phys .
b * 706 * , 123 ( 2011 ) m. p. decross , d. i. kaiser , a. prabhu , c. prescod - weinstein and e. i. sfakianakis , `` preheating after multifield inflation with nonminimal couplings , i : covariant formalism and attractor behavior , '' arxiv:1510.08553 [ astro-ph.co ] . | in this work , we investigate the preheating mechanism in a disformally - coupled inflationary model where the scalar field @xmath0 ( which is the inflaton field ) naturally coupled to another matter field @xmath1 induced by the disformal transformation . in the present scenario ,
the emergence of novel derivative interactions mixing the kinetic terms of the fields is natural .
we start by deriving the evolution of the background system and then quantify how the disformal coupling influences inflationary trajectory .
we examine the particle production due to parametric resonances in the models and find in minkowski space that the stage of parametric resonances can be described by the standard mathieu equation .
interestingly in our analysis , we discover that broad resonances can be typically achieved and are potentially efficient causing the particle number density in this process to exponentially increase .
+ pacs numbers : 98.80.-k , 04.50.kd , 95.30.sf |
cspbcl@xmath0 crystal was investigated as a typical material undergoing structural phase transitions above room temperature ( rt ) @xcite ; the crystal is perovskite in the temperature range above 320 k , transforms to tetragonal at 320 k , and becomes orthorhombic or monoclinic at 310 k. the phase transitions are considered to be ferroelastic , related to zone - boundary phonons of m@xmath0 and r@xmath1 @xcite . on the other hand phase transition below rt
was explored in the measurement of doped cspbcl@xmath0 with the electron - paramagnetic - resonance ( epr ) method , so that the results implied that phase transition takes place at about 180 k @xcite .
however the temperature of phase transition estimated from the epr experiments is 185 @xcite and 176 k @xcite , and slightly different with each other .
the discrepancy could come from the doping effect on the phase transition , and therefore it seems more appropriate to examine the non - doped crystal for determining the phase - transition point .
the crystal structure below rt has been studied recently with the x - ray diffraction @xcite ; the study concluded that the crystal is orthorhombic below rt and implied that structural transition does not take place below rt .
thus phase transition below rt is not conclusive from the structural examination .
the result implies that if phase transition does take place below rt , it can be second - order transition that is not associated with any crystallographic transition .
another recent study on phase transition at low temperatures has been performed by focusing on optical absorption tail and photoluminescence ( pl ) of cspbcl@xmath0 .
since the crystal is a semiconductor with direct bandgap @xcite , the electronic excited states are generated under photoexcitation .
the absorption tail and pl properties suddenly change around 170 k @xcite ; the results suggest phase transition below rt . to explore the phase transition , photoacoustic spectroscopy ( pas ) is a suitable method because pas enables us to observe self - induced thermoelastic ultrasonic wave under photoexcitation .
in fact phase transitions have been often studied by pas and detected successfully @xcite ; even second - order transition were observed . in the present study , it is aimed to clarify the properties of observed pa signals and furthermore to test the existence of phase transition below rt .
experimental details are described in sec .
experimental results are presented and the wavelet analysis is applied to the pa wave in sec
. [ results ] .
the wavelet analysis has the advantage of resolving the superimposed waves into the several waves corresponding to each frequency component ; for example , the electric noise mixed in the experimental data is apparently separated from pa signals . after clarifying pa - signal properties , the existence of phase transitions below rt is discussed in sec . [ discussion ] .
cspbcl@xmath0 crystals were grown with the bridgman technique . to purify 99.999% pbcl@xmath2 powder , the single crystal was previously grown , and 99.99% cscl powder and the pbcl@xmath2 single crystal were mixed at equimolar amount and sealed in a quartz ampoule at about 10@xmath3 torr .
the ampoule was set in a bridgman furnace and was moved from the point of 970 k to that of 720 k at about 20 k / day . though the melting temperature of cspbcl@xmath0 is about 880 k , the ampoule was kept below the melting point for several days to anneal the specimen .
finally it was cooled from 720 k to rt for a day .
the as - grown ingot of cspbcl@xmath0 was aggregate of single crystals .
the crystal was cleaved in @xmath4 plane perpendicular to @xmath5 axis ; the crystals can be cleaved only in @xmath4 plane .
typical dimension of the cleaved specimen was 5@xmath65@xmath61 mm@xmath7 along @xmath8 , @xmath9 and @xmath5 axes respectively .
both cleaved @xmath4 planes were used for pa measurement ; one was the plane injected by incident photons and the other was firmly attached to a piezoelectric transducer ( pzt ) with conductive organic paste .
experimental configuration is schematically drawn in fig .
[ fig1 ] .
two type of pzt were used ; the pzt of 3.99-mhz resonant frequency was a plate of 5@xmath65@xmath61 mm@xmath7 and that of 1.97-mhz resonant frequency was a 10-mm diameter and 1-mm thickness disc .
one of the cleaved planes was attached to a pzt as shown in fig .
curie temperatures of the pzts were 603 k and therefore the pzts do not undergo any phase transition in the temperature range of 150300 k at which the present measurement was performed .
the crystal and pzt were set in a he - flow cryostat .
acoustic - wave ringing effect in a pzt was checked ; the effect is compared with the pa wave propagated from cspbcl@xmath0 in the next section .
the pa signals were induced by the third harmonic generation of nd : yag laser ; the laser pulses were 3.49 ev and 5-ns width , and the reputation was 10 hz . since cspbcl@xmath0 crystals have the fundamental absorption edge at 3.0 ev @xcite , the incident light is absorbed due to the electronic interband transition , except for the surface reflection .
the absorption takes place in the thin surface layer because of the absorption length less than 1 @xmath10 m and the sample thickness of 1 mm . in this configuration the pa signal generated in the surface layer propagates through the crystal and is detected by the pzt on the back surface .
thus the detected pa wave is bulk wave propagated through the crystal ; the measured pa wave does not only propagate along @xmath5 axis but also includes the off - axis wave which travels along @xmath8- or @xmath9-axis and the multiple reflection in the crystal .
the intensity of incident laser light was about 200 @xmath10j / pulse , and the incident light is loosely focused at the spot size of 1-mm diameter on the sample surface ; the sample was not damaged by the laser - light irradiation .
the pa wave was measured by an oscilloscope without any preamplifier .
different time - range data were measured by varying the coupling resistance to the oscilloscope .
the amplitude and phase of the pa signals were picked up by a two - phase lockin amplifier at the cycle of incident pulse .
the reference input to the lockin detector was taken from the transistor - transistor - logic pulses synchronized with the laser pulses . in the setup
the lockin detector pick up the signals of the same frequency with the reference ; the amplitude is proportional to the peak intensity because the lockin detector automatically tunes the phase of sine curve to the peak position of pa signal . measuring the pa signals by the lockin ,
the coupling resistance is required high ; the pa wave is therefore broadened to ms - order , so that the lockin technique is applicable to the present measurement . at the end the relation of the pulsed pa measurement with the conventional pa measurement is made mention of ; pulsed pa signals are generally stronger than the conventional pa signals induced by chopper - modulated laser light because the laser pulses have higher photon density @xcite . in the conventional pa measurement
using cw laser light of 3.81 ev and a chopper , the amplitude and phase of pa signals of cspbcl@xmath0 show a profile similar to those of pulsed pa signals ( see fig .
[ fig4 ] ) .
the pa signals were just weaker than the pulsed pa signals .
consequently there is not any special advantage to employ the conventional pa technique here .
figure [ fig2 ] shows detected pa wave at 296 k ( curve ( a ) ) in a @xmath10s range .
the wave was induced by 3.49-ev pulsed laser light of 5-ns width at @xmath11 @xmath10s .
the pa wave continues to oscillate prominently for more than 80 @xmath10s . for comparison with the pa wave ,
pzt ringing ( curve ( b ) ) was also displayed to a similar scale ; it was measured under direct photoirradiation onto the pzt attached to nothing but lead wires .
the ringing in the pzt is directly stimulated due to the laser pulse , independent of photon energy of laser pulses .
the envelope of rapid component decays exponentially at about 10 @xmath10s and the slow component weakly survives up to 100 @xmath10s . as seen in fig .
[ fig2 ] , the decay of pa wave is different from the pzt ringing .
that is , the ringing effect on the detected pa wave is not dominant .
probably the oscillatory structure of pa wave mainly comes from long lasting multiple reflection of the thermoelastic wave in the crystal . besides the pzt ringing such as curve ( b ) in fig .
[ fig2 ] becomes comparable to the pa waves only under direct and 1.5-times more intense excitation .
thus , though the influence of ringing is in principal included in the measured pa wave , the contribution is quite small .
it is also shown in the analysis of fig .
[ fig3 ] . figure [ fig3](a ) presents the pa wave at 297 k , which was detected by a pzt of resonant frequency of 1.97 mhz .
solid line in fig .
[ fig3](a ) represents the measured pa wave , and dashed line shows the incident laser pulse of 3.49 ev .
to analyse this real time oscillation , we choose to use the wavelet method . in the present analysis , the fourth b - spline wavelet @xmath12 in the inset of fig .
[ fig3](b ) is selected from many wavelets ; the @xmath12 is a finite wave packet composed of third - order polynomials and corresponds mainly to a narrow frequency band via fourier transformation @xcite . first the measured pa wave in fig .
[ fig3](a ) is approximated by the fourth b spline .
a good approximation is displayed as @xmath13 in fig .
[ fig3](b ) .
the pa wave @xmath13 is decomposed as follows : @xmath14 where @xmath15 , and @xmath16 are coefficients @xcite .
the @xmath17 is well resolved by @xmath18 ( @xmath19 ) .
figure [ fig3](b ) displays the result of wavelet decomposition . @xmath20 and
@xmath21 make small contribution ; the @xmath21 represents electric noise at 25 @xmath10s which appears without pa signal and stems from the trigger signal for measurement .
@xmath22 is apparently predominant , and @xmath23 and @xmath24 contribute to some extent .
though @xmath25 also contributes slightly , it almost corresponds to the pzt ringing observed in this time range .
@xmath23 , @xmath22 and @xmath24 are consequently regarded as the contributors responsible for the pa wave .
@xmath22 corresponds to the frequency band at 1.5 mhz . on the other hand @xmath25
does to the band at 0.4 mhz ; the frequency is distinctive from the pzt resonant frequency of 1.97 mhz , that is , the ringing just stems from multiple reflection of acoustic wave in the pzt .
thus , by removing the electric noise and ringing components , the pa component in time domain is found to be superposition of a few frequency bands . at the end of pa - wave analysis
, it is to be noted that pa wave is almost invariant in 150300 k. therefore pa wave at 296 k was analysed here .
pa - signal amplitude ( solid circle ) and phase ( dot ) are plotted against temperature ( 150250 k ) in fig .
they were measured , by using 10-hz and 3.49-ev laser pulses and a pzt of resonant frequency of 1.97 mhz , with the raising and lowering rates of 1.2 k / min .
the incident light was parallel to @xmath5 axis .
the amplitude and phase measured with decreasing temperature indicate a prominent peak and dip at 198 k respectively while those with increasing temperature have a similar peak and dip at 231 k. another prominent change in the phase , that is , the leap by two degrees is observed at 177 k in the data on cooling .
arrows indicate the lower edge of phase change .
the origins of changes are discussed in sec .
[ discussion ] .
tens of jumped data points appear for both amplitude and phase only in the data on cooling .
the irregular data presumably originate from unexpected cracking of the crystal on cooling because the irregularity is far less on heating .
besides it appears very differently on second cooling or in another sample .
however the prominent dips and peaks are reproduced ; it is therefore relevant that they indicate intrinsic changes in the crystal . concerning with the cracking , there are a few remarks : ( i ) the pa measurement was begun at rt and kept to 150 k. before measuring pa signals on heating , the sample
was cooled to about 120 k. it was due to the thermal controller .
( ii ) the excess cooling probably induces further cracking .
in fact , the amplitude of pa signals on heating is about half on cooling ( see fig . [ fig4 ] ) .
( iii ) in the present setup of fig .
[ fig1 ] , pa wave travels through the crystal .
the cracking consequently decreases measured pa signals .
the changes of pa - signal amplitude and phase in fig .
[ fig4 ] is discussed in this section . as shown in figs .
[ fig2 ] and [ fig3 ] , pa wave consists of thermoelatic ultrasonic components .
therefore the amplitude and phase detected by a two - phase lockin amplifier correspond to those of thermoelatic wave .
the amplitude is proportional to the peak intensity in two - phase lockin equipment .
the phase generally depends on the amplitude .
figure [ fig4 ] shows the amplitude ( closed circles ) and phase ( dots ) of thermoelastic wave vs temperature .
the intensity has tens of microvolts at 150250 k while the phase increases as temperature becomes lower .
the gradual changes come from the thermal deformation of crystal . the changes at 198 and 231 k in fig .
[ fig4 ] are probably due to those of propagation of bulk thermoelastic wave ; the change seems different from gradual thermal deformation of the crystal and from the change of the pzt which has curie point at 603 k. the changes are therefore ascribed to a phase transition at the temperatures .
indeed they show similar profiles except for irregular data .
the transition has large hysteresis between cooling and heating .
a theoretical study on photoacoustic detection of phase transition @xcite showed that the profiles of amplitude and phase of pa signals are similar to those at a second - order phase transition .
it is therefore likely that the transition at 198 and 231 k is second - order .
arrows in fig .
[ fig4 ] indicate the lower edge of phase and correspond to the transition points estimated from the indication by @xcite . according to the calculation
, the amplitude decreases to zero at a first - order transition because of ultrasonic attenuation of structural transition . as noted in sec .
[ intro ] , phase transitions below rt in cspbcl@xmath0 have been hard to be detected .
this is because structural study by x - ray diffraction is ineffective in the crystal @xcite .
the result is consistent with the assignment that the type of transition at 198 and 231 k is second - order .
in addition another phenomenological analysis for second - order phase transition @xcite derives the leap of sound velocity and the increase of absorption coefficient of ultrasonic wave at second - order phase - transition point when lattice distortion couples with the square of order parameter ; it is often called landau - khalatnikov mechanism .
the dips of amplitude in fig .
[ fig4 ] are explained by the increase in absorption coefficient of mhz - thermoelastic wave in the crystal .
on the other hand another possibility for the type of transition still exists .
the study employing x - ray diffraction indicated that the crystal , keeping the crystallographic group , varies the length of unit cell along @xmath8 and @xmath5 axes while it hardly changes the length along @xmath9 axis @xcite .
it is thus possible that slight displacement of lattice takes place in the unit cell ; the slight change might not be well discriminated in the x - ray - diffraction study .
recently order - disorder transition was reported at 193 k by raman scattering @xcite ; the assignment is consistent with the possibility of slight structural change .
consequently there are two possibilities at present ; the type of transition at 198 and 231 k is due to second - order or the slight structural transition in the unit cell . the leap of phase at 177 k in fig . [ fig4 ] appears on cooling and is sharper than the change at 198 k. it is improbable that the sudden change at 177 k comes from the adiabatic deformation of the crystal .
that is , it also suggests the existence of a phase transition at the temperature .
the transition is most likely different from that at 198 k because the profiles are inconsistent with each other .
the leap at 177 k is possibly due to a second - order phase transition though the type is not conclusive just like that at 198 k. the leap at 177 k is observed only on cooling .
the reason is presumably as follows : as remarked at the end of sec .
[ results ] , the amplitude of pa signal on heating is two - times weaker than that on cooling , because of cracking .
it is consequently harder to detect the slight change of pa signal on heating . indeed
the leap at 177 k on cooling is the same order with the noisy fluctuation of signals .
thus the leap on heating may be hidden .
the cracking perhaps connects to the ferroelastic nature at low temperatures .
ferroelastic domains are typically several micrometers in various materials .
the boundary of domains prevents pa wave from propagating for long distance up to 1 mm . on the other hand
pl properties mainly depend on the local structure within several tens of nanometers .
indeed pl shows hysteresis on cooling and heating .
the explanation for the absence of pa - signal leap on heating is compatible with the change of pl properties . in conclusion , pa signals in cspbcl@xmath0 crystals have been studied below rt ; it has been shown that the leap of the pa signals appears at 177 and 198 k on cooling and at 231 k on heating .
though the profiles imply second - order phase transition , the slight structural transition in the unit cell is not excluded .
thus the types are not yet conclusive at present
. however phase transitions below rt are confirmed by examining pa signals .
the transition at 177 k is probably responsible for the leap of epr signals @xcite and for the sudden change of optical and pl properties @xcite .
i would like to acknowledge dr .
k. yoshino ( miyazaki univ . ) for introducing me to pas , and also thank mr .
t. kobayashi for the growth of cspbcl@xmath0 crystals and prof .
t. hayashi ( kyoto univ . ) for support to the study .
this work was partly supported by grant - in - aid for research fellow of the japan society for the promotion of science .
etxebarria , j. , uriarte , s. , fernndez , j. , tello , m.j . and gmez - cuevas a. ( 1984 ) .
piezoelectric photoacoustic detection of first- and second - order phase transitions .
_ j. phys .
c. : solid state phys .
_ , * 17 * , 6601 .
heidrich , k. , schfer , w. , schreiber , m. , schtig , j. , grandke t. and stolz , h.j .
electronic structure , photoemission spectra , and vacuum - ultraviolet optical spectra of cspbcl@xmath0 and cspbbr@xmath0 . _ phys .
b _ , * 24 * , 5642 . and a pzt .
the sample was photoexcited on one cleaved plane and pa signals were detected on the other by the pzt attached with conductive organic paste .
dotted lines in the sample stand for the pa waves schematically .
the piezoelectric voltage was measured along the thickness direction .
this figure is drawn for the pzt of 5@xmath65@xmath61 mm@xmath7 .
more details are described in sec .
[ exp ] . ] at 296 k with 3.49-ev laser pulses of 5-ns width .
incident laser pulses reach onto the surface at 12.5 @xmath10s .
the pzt used in the measurement has the resonance at 3.99 mhz .
( b ) ringing at 296 k in a pzt of 3.99-mhz resonance ; it was induced by direct photoinjection onto the pzt .
the incident light was 2.33-ev laser pulses of 5-ns width .
3.49-ev laser pulses also induce the same shape of ringing . the ringing effect on the detected pa wave
is described in the text .
both spectra are presented to a similar scale . ] at 297 k ( solid line ) ; it is induced by 5-ns - width laser pulses of 3.49 ev . dashed line : the position at which the laser pulses reach onto the sample surface
. the dashed line is vertically offset .
( b ) wavelet analysis with the fourth b - spline wavelet @xmath12 ; the inset shows the @xmath12 which is center symmetrical .
@xmath13 denotes a good approximation for the pa wave in fig .
[ fig3](a ) .
@xmath18 ( @xmath19 ) resolve @xmath17 ; they are vertically offset and displayed to the same scale with @xmath17 . ]
the resonant frequency of pzt was 1.97 mhz .
the increasing and decreasing rates of temperatures were 1.2 k / min .
arrows indicate the lower edge of phase change and are indicative of phase transitions ; the discussion on phase transitions is given in sec .
[ discussion ] . ] | phase transitions in cspbcl@xmath0 crystals have been explored by photoacoustic ( pa ) method in a wide temperature range of 150300 k. the pa signals , induced by ns - pulsed laser excitation of electronic states and detected with a piezoelectric transducer , are thermoelastic ultrasonic waves lasting for hundreds of microseconds . the wavelet analysis of the pa wave has clarified that the thermoelastic wave includes several mhz components .
the pa signals have been made use of in order to examine phase transitions at low temperatures in the crystal .
changes of pa signals indicate that two phase transitions take place at low temperatures around 200 k. the types of phase transitions are discussed taking into recent experimental results . _
keywords : _ photoacoustic spectroscopy ; ultrasonic wave ; phase transitions ; cspbcl@xmath0 |
the study of rare kaon decays has a glorious history . at the level of @xmath2 in branching ratio ,
parity violation was first `` discovered '' as the @xmath3-@xmath4 puzzle@xcite , cp violation was discovered at the level of @xmath5@xcite , and the gim mechanism@xcite was suggested by the strong suppression of the decay @xmath6 , or the absence of the flavor changing neutral currents ( fcnc) the decay mode was subsequently found at the level of @xmath7 .
further searches for rare kaon decays at lower branching ratio undoubtably involve rich physics and may yet offer another surprise e.g. they are likely to elucidate the origin of cp violation , and provide some clues for physics beyond the standard model ( sm ) . in this talk , the decay modes via fcnc with the expected branching ratios of less than @xmath8 are discussed . since exotic decays violating the lepton flavor conservation law have been discussed by the previous speaker@xcite , the focus of this talk is on the rare kaon decays which are allowed in the sm .
typical leading feynman diagrams of these decays are shown in fig .
1 , where @xmath9 and @xmath10 at the top lines are for @xmath11 and @xmath12 decays , respectively , and @xmath13 indicates the electron or the muon . because of the gim mechanism , the lowest order diagrams come from second order weak interactions with a @xmath10-type quark in the loop diagrams , in which the top - quark contributions dominate because of the mass .
this makes these decay modes very sensitive to @xmath14 , the least constrained coupling - constant of the top and down quarks in the cabibbo - kobayashi - maskawa ( ckm ) matrix .
due to one of the unitarity conditions of the ckm matrix @xmath15 and with the approximation @xmath16 , @xmath14 and @xmath17 form two sides of a unitarity triangle as shown in fig .
the height of the triangle @xmath18 is an indication of `` direct '' cp violation , or cp violation through the decay amplitude . for the decay @xmath0 ,
the branching ratio is roughly proportional to @xmath19 and for @xmath1 to @xmath20 .
these measurements alone can determine the unitarity triangle , and when they are combined with measurements from @xmath21-decays it will provide over - constrained information that is sensitive to a presence of supersymmetry and other physics beyond the sm@xcite .
if the final states include a charged lepton pair , there are additional contributions from diagrams with virtual photons that usually prevent clear interpretations of measurements@xcite .
the decays @xmath22 have a `` direct '' cp violating component , which is expected to occur , if it were the only component , at a branching ratio of @xmath24 .
there are two other contributions , however , at the same level to this process ; one arises from the mixing of the cp even state in @xmath25 , and the other from two - virtual - photon intermediate states that conserve cp . at present , there are theoretical ambiguities in the estimations of these contributions , which may eventually be sorted out by measurements of @xmath26 and other radiative decays . to make the matter a bit more complicated , there is a physical background coming from the radiative decay @xmath27 , which has the identical event topology . with the tightest cuts ,
the background level is estimated to be still at a @xmath28 level for @xmath29@xcite .
the situation is similar in the case of @xmath30 .
the present upper limits ( 90 % c.l . ) are @xmath31@xcite and @xmath32@xcite .
the decay @xmath33 is very similar to @xmath6 , which is sensitive to the @xmath34 parameter@xcite , but it is further suppressed by the helicity mechanism .
this decay is sensitive to pseudo - scalar interactions coming from physics beyond the sm .
four events from the decay @xmath33 have been reported recently which correspond to a branching ratio , @xmath35@xcite , being consistent with the unitarity bound expected from the long - distance contributions .
in the sm calculation of the decay @xmath0 , the dominant contribution comes from second order loop diagrams with a virtual top quark ( fig . 1 ) .
the hadronic matrix element can be extracted from the decay @xmath36 and theoretical uncertainties in the calculation due to long distance contributions and other effects are small@xcite .
the signature of the decay @xmath0 is a single pion with no other observable decay - products .
definitive observation of this signal requires that all possible backgrounds are suppressed well below the signal level .
major background sources are : a muon from the copious decay @xmath37 ( @xmath38 ) which is misidentified as a pion , a pion from the decay @xmath39 ( @xmath40 ) when two photons from the @xmath41 decay are unobserved , a beam pion scattered by the target into the detector , and charge exchange reactions of @xmath12 s which result in decays @xmath42 , where @xmath13 ( @xmath43 or @xmath44 ) is undetected . in order to suppress the backgrounds , good particle identification , efficient photon veto , and detection of incoming particles are essential . the e787 experiment at brookhaven national laboratory ( bnl ) as shown in fig . 3
is designed to effectively distinguish these backgrounds from the signal .
kaons of about 700 mev / c at a rate of @xmath45 per 1.6-s spill are detected and identified by a erenkov counter and hodoscopes , degraded by beo and stopped in an active target , primarily consisting of 413 5-mm square scintillating fibers .
the momentum ( @xmath46 ) , kinetic energy ( @xmath47 ) and range ( @xmath48 ) of decay products are measured using the target , a central drift chamber , 21 layers of 1.9-cm thick plastic scintillator ( range stack ) and two layers of straw chambers , all contained in a 1-t magnetic field . the @xmath49 decay sequence of the decay products in the range stack scintillator is observed by 500-mhz transient digitizers for particle identification .
photons are detected by a @xmath50-sr calorimeter consisting of a 14-radiation - length - thick lead / scintillator barrel detector , 13.4-radiation - length - thick end caps of csi crystals , and a 3.5-radiation - length - thick lead - glass erenkov counter which also works as an active beam degrader .
the e787 experiment reported an observation of one clean event at a branching ratio @xmath51 using the data sample taken in 1995@xcite . since then , additional data samples taken in 1996 and 1997 , together with those taken in 1995 , have been analyzed with improved algorithms , which have resulted in less non - gaussian tails in @xmath46 , @xmath48 and @xmath47 measurements , and a @xmath5230 % higher acceptance than that in ref.@xcite .
also , in the 19967 runs , lowering the incident @xmath12 beam momentum resulted in a larger fraction of kaons stopping in the target , which reduced accidental hits originated from nuclear reactions in the beam degrader .
the higher proton intensity at the production target compensated the kaon yield loss at lower momentum . in order to avoid a possible bias in the analysis , the signal region
is kept untouched until the background estimations as well as optimization of acceptance have been done . in the background study , the data sample after applying all the final cuts except two orthogonal ( uncorrelated ) cut groups to be studied , e.g. the kinematical cuts and those related to the observation of the decay sequence @xmath49 in the stopping counter for the estimation of the @xmath37 background , is used to obtain the suppression factor for each cut group .
the correlation between the two cut groups , which may invalidate the method , is studied by varying the cuts being studied or by enhancing certain types of background events .
the total background for the entire 19951997 exposure with the final analysis cuts is estimated to be @xmath53 events .
the acceptance for @xmath0 , @xmath54 is calculated based on data and monte carlo calculations .
the largest uncertainty comes from the uncertainty in pion - nucleus interaction .
the measurement of the branching ratio for @xmath39 within a few % of ref.@xcite confirms the acceptance calculation .
analysis of the full data sample as shown in fig .
4 has yielded only the single event previously reported .
based on the acceptance @xmath55 and the total exposure of @xmath56 kaons , the new branching ratio is @xmath57@xcite .
this provides a constraint , @xmath58 .
the goal of a new experiment e949 at bnl is to improve the sensitivity by an order of magnitude . since the ags in the rhic era will be used for two hours a day to feed heavy ions into the rhic ring , the rest of 22 hours can be used for the high energy program .
the operation is expected to provide a more stable and longer running period . exploiting a higher beam intensity ,
the incident @xmath12 momentum can be further lowered to reduce accidental coincidence .
the photon veto capability will be improved by additional lead / scintillator layers in the barrel region and by additional active degrader in the beam region . in the region below the @xmath39 peak at 205 mev / c , where nuclear interactions result in a large momentum - tail ,
the additional photon veto capability may suppress the background by more than an order of magnitude as low as to the signal level , doubling the phase space in the search . the decay @xmath1 violates the cp conservation law through decay amplitude .
this process is expected to occur at @xmath59 @xcite .
the contribution from cp mixing in @xmath60 is expected to be around @xmath61@xcite .
since no charged leptons are involved in this decay , the process is free from virtual photon contributions and clean for the study of the origin of cp violation .
the theoretical ambiguity is only @xmath62 except those in the ckm matrix elements .
conversely , the observation of this decay mode uniquely determines @xmath18 or @xmath63 in the wolfenstein parametrization .
the goal of the experiment is to determine @xmath18 with a 1015 % accuracy , which corresponds to a single event sensitivity of @xmath64 .
the present upper limit of this decay is @xmath65@xcite .
the decay @xmath1 is a three - body decay that involves only neutral particles .
the signature of this decay is two photons from the @xmath41 and no other activity in the detector . in this decay mode ,
available kinematical parameters are very limited ; relatively easy ones to measure are the positions and energies of @xmath66-rays . also , the direction and position of the @xmath60 beam can be limited by tightly collimating the beam at a cost of beam intensity .
the decay vertex can be reconstructed from these two constraints with the assumption of the pion mass .
the kek experiment@xcite and the fnal approach@xcite are classified in this category .
the bnl experiment e926@xcite attempts to measure more kinematic values ; the directions of @xmath66-rays and the momentum of the @xmath60 . measurements of @xmath66-ray directions allow full reconstruction of the @xmath41 kinematics without the assumption of the @xmath41 mass , providing more constraints with redundancy , which is necessary to suppress the background to the level well below the signal . the time - of - flight ( tof ) of a @xmath60 between the production target and the decay vertex
can be measured for low momentum @xmath60 s if the incident proton beam is bunched .
this measurement allows calculation of the missing mass and kinematical reconstruction in the center of mass system , which is effective to eliminate backgrounds from two - body decays .
the major decay modes of @xmath60 are @xmath67 and @xmath68 .
suppression of most backgrounds is achieved by high - efficiency hermetic photon and charged - particle detector system surrounding the decay volume , and kinematical constraints . in the bnl e926 experiment , a low energy @xmath60 beam with an average momentum around 650
mev / c is produced by irradiating a target with a 24-gev proton beam from the ags and extracted at 40@xmath69 with respect to the incident beam .
the proton beam is bunched to form a @xmath70200 ps wide bucket at a rate of 25 mhz .
about 16 % of @xmath60 s decay in the 4-m long decay volume , which is evacuated to a level of @xmath71 torr to suppress the background from neutron - induced @xmath41 production .
the decay region is surrounded by a charged particle veto system and a photon veto system of 18-radiation - length - thick lead / scintillator sandwiches in the barrel region .
the two photons from the @xmath41 decay are converted into pairs of a positron and an electron in a 2-radiation - length preradiator next to the vacuum region for the measurement of the directions of the @xmath66-rays .
the preradiator consists of sandwiches of 2-mm thick scintillator , a copper plate as a mechanical support and radiator , and a tracking chamber .
this is followed by an 18-radiation - length calorimeter for the measurement of @xmath66-ray energies and for vetoing additional photons .
the estimates of sensitivity for @xmath72 are tightly coupled to the cuts required for background suppression , particularly for the @xmath73 and @xmath74 backgrounds .
an acceptance of @xmath75 % for the case s / n=0.5 comes from the combination of factors ; 0.58 for the fiducial region and usable kaon momentum region , 0.33 for the solid angle , 0.5 for the efficiency of the preradiator , and the remaining factor for the @xmath41 mass cut and other cuts . assuming three years of running with @xmath76 kaons per 2-s beam spill ( 50 % duty factor ) , the expected number of events is 65 .
the toughest background is the cp violating @xmath73 decay when two photons are undetected .
backgrounds from @xmath73 arise when two photons are detected in the forward detector and the other two are undetected anywhere .
these backgrounds can be classified into two categories ; even pairing when two photons come from the same @xmath41 , and odd pairing when two photons come from different @xmath41 s .
events in the even pairing category form a two - body - decay peak in the momentum spectrum of @xmath41 in the cm system , while the two @xmath66-rays in the odd pairing category do not reproduce the @xmath41 mass . the above requirements and photon veto
essentially suppress @xmath73 backgrounds to 0.2 of the signal level .
similarly , @xmath74 backgrounds can be suppressed to 0.1 of the expected signal .
backgrounds from other @xmath60 decay modes are estimated to be less than 0.1 of the signal . because of the large angle extraction , the cross sections for producing @xmath77 s are small and they completely decay before reaching the decay volume .
backgrounds could arise from @xmath77 s produced by halo neutrons and @xmath60 s , but they are estimated to be negligible .
neutrons with @xmath78 mev / c can react on the remaining atoms in the vacuum region to produce @xmath41 s . this background level is again estimated to be 0.01 .
accidental backgrounds are caused by beam halo neutrons and @xmath66-rays which create a @xmath41 signal in the preradiator .
the background level is estimated to be 0.02 of the expected signal .
the bnl e926 was proposed in 1997 but the group is still waiting for full funding . in the present scenario ,
the detector construction is expected to start in 2002 .
the kek experiment has partially been approved , but the sensitivity goal is only around the sm level the new japan hadron facility is expected to improve the sensitivity .
studies of rare kaon decays have contributed to the discoveries of several symmetry violations .
the decays @xmath0 and @xmath1 are expected to `` complete '' the measurement of the ckm matrix in the next decade independently from the @xmath21-decay system , and to elucidate the origin of cp violation .
lee and c.n .
yang , phys . rev . * 104 * , 254 ( 1956 )
christenson @xmath79 , phys .
* 13 * , 138 ( 1964 ) .
glashow , j. iliopoulos and l. maiani , phys .
d2 * , 1285 ( 1970 ) .
k. jungmann , these proceedings .
buras , these proceedings ; g. buchalla and a.j .
buras , phys . rev . *
d54 * , 6782 ( 1996 ) .
y. nir and m.p .
worah , phys . lett . *
b423 * , 319 ( 1998 ) .
littenberg and g. valencia , ann .
sci . * 43 * , 729 ( 1993 ) .
l. wolfenstein , phys .
* 51 * , 1945 ( 1983 ) .
greenlee , phys . rev .
* d42 * , 3724 ( 1992 ) .
j. whitemore for ktev , to be published in the proceedings of kaon99 conference at fermilab , 1999 .
ktev collaboration , hep - ex/0001006 ( 2000 ) .
bnl e871 collaboration , phys .
* 81 * , 4309 ( 1998 ) .
bnl e787 collaboration , phys .
lett . * 79 * , 2204 ( 1997 ) .
particle data group , review of particle physics , z. phys . * c3 * , 1 ( 1998 ) .
bnl e787 collaboration , phys .
lett . * 84 * , 3768 ( 2000 ) .
littenberg , phys . rev . *
d39 * , 3322 ( 1989 ) .
e799-ii / ktev collaboration , phys . rev . * d61 * , 072006 ( 2000 ) .
kek e391a collaboration , kek internal 96 - 13 ( 1996 ) .
kami collaboration , expression of interest , hep - ex/9709026 ( 1997 ) .
bnl e926 collaboration , ags proposal e926 ( 1997 ) . | rare kaon decays via flavor changing neutral currents are discussed in the context of the ckm unitarity triangle with a particular interest in the rare kaon decays @xmath0 and @xmath1 .
new results and the status of these experiments are reported . |
recent experiments by the weizmann group have stirred much attention to the phase that electrons acquire when they traverse a quantum dot ( qd ) embedded in the arm of a aharonov - bohm interferometer.@xcite these fascinating measurements call for a deeper understanding of electron transport through a strongly interacting object . if many electrons say @xmath2 populate a qd , the transmission phase @xmath3 of the tunneling electron displays a much studied@xcite universal behavior : first @xmath3 increases by @xmath0 through the conductance peak and then it lapses in the coulomb valley . here
we focus on the relatively less studied few - electron regime , where the phase evolution depends on @xmath4 . in the samples studied in ref .
the phase remains constant in the @xmath5 valley , independently from qd tunings , whereas it lapses in the coulomb blockaded devices of ref . .
for @xmath6 the phase is not reproducible and shows both increments and lapses , likely due to qd shape and exchange effects.@xcite in spite of several explanations of the few - electron scenario@xcite a clear picture is still missing .
many works introduced major simplifications , like spinless electrons,@xcite one - dimensional qds,@xcite simplified models for coulomb interaction,@xcite or other _ ad hoc _
assumptions.@xcite even the interpretation of the simplest @xmath7 transition is controversial , being variably attributed to the occupation of either the same@xcite or a different@xcite orbital from that of the first electron , to the role of excited doorway channels,@xcite to electron crystallization.@xcite in this paper we compute @xmath3 by fully including exchange and correlation effects .
the theory is based on the application of the friedel sum rule ( fsr ) as generalized in ref . to a multiorbital interacting qd an exact zero - temperature result , @xmath8 .
since the fsr holds for non degenerate ground states ( gss ) only,@xcite it may be applied to either singlet kondo gss at zero field , @xmath9 , or non - degenerate gss in the coulomb blockade regime ( @xmath10 for singlets and @xmath11 for doublets and higher - spin states ) .
it is worth recalling that , in the conductance valleys with odd @xmath4 at @xmath8 and @xmath9 , the qd is always in the kondo regime . in order to reach the coulomb blockade regime by keeping @xmath8
, one needs to apply the field to the aharonov - bohm ring to destroy dot - lead kondo correlations , hence removing qd spin degeneracy.@xcite this condition is reached when @xmath12 , with @xmath13 being the bohr magneton and @xmath14 the qd level width . in the experiments the temperature is very low ( @xmath15 mk ) and
the field is relatively weak ( @xmath16 mt ) .
therefore , in order to compare measurements with the theoretical predictions reported in this paper , tiny widths @xmath14 are required , which are actually smaller than the values presently reported.@xcite as anticipated above , the theory relies on the application of the fsr .
the key idea is that the phase variation @xmath17 for the addition of one electron to the qd is given by integrating the qd spectral density @xmath18 between two consecutive conductance valleys ( with @xmath19 ) .
this is evaluated exactly for an isolated qd via full configuration interaction ( ci ) calculations@xcite with @xmath20 .
the simulations agree with the coulomb - blockade results of ref . :
( i ) the phase evolution for @xmath21 strongly depends on the dot s shape and density .
( ii ) @xmath3 never lapses in the @xmath5 valley , independently from dot parameters .
additionaly , ( iii ) in the coulomb ( kondo ) regime the increment @xmath22 through the conductance peak is significantly smaller than @xmath0 ( @xmath1 ) as a consequence of strong coulomb correlation .
the theory fails to reproduce the coulomb blockade results of ref . .
the reason of this discrepancy is presently unclear , but it is likely related to the significant departure from the condition ruling the range of applicability of the theory , i.e. , @xmath12 at @xmath8 . in order to treat the coulomb blockade regime even at @xmath9 , one could compute the thermal green function of the fully correlated system at @xmath23 , with @xmath24 being the kondo temperature .
we leave for future work this alternative route , which would immediately provide the transmission phase @xmath3 without invoking the fsr.@xcite the structure of this paper is as follows . in sec . [
s : fsr ] we set the theoretical model and illustrate the usage of the fsr to compute the transmission phase variation @xmath22 between consecutive conductance valleys .
after briefly recalling the full ci method in sec .
[ s : ci ] , we address in sec .
[ s : sp ] the evolution of the phase in the absence of coulomb interaction ( but taking into account the spin degree of freedom ) .
we eventually consider the fully interacting case in sec .
[ s : corr ] .
the number sequence points to the consecutive filling of six electrons in a non - interacting picture .
( c ) transmission phase @xmath3 and phase shifts @xmath25 vs @xmath4 for the filling sequence plotted in ( b).,width=288 ] we model the experimental setup of ref . as in fig .
[ fig1](a ) .
electrons flow along @xmath26 from source ( s ) to drain ( d ) , tunneling through a qd of elliptical shape .
the ellipse is the generic low - energy form for a two - dimensional shallow , gate - defined potential,@xcite since the lowest - order non - vanishing terms of its series expansion are quadratic .
the scattering matrix @xmath27 is diagonal in the spin index @xmath28 in both kondo and coulomb blockade regimes .
in fact , in the kondo regime ( @xmath9 ) , no elastic spin - flip occurs.@xcite on the other hand , in the coulomb blockade regime only one @xmath28 channel is active at time for a given energy since @xmath29 removes spin degeneracy .
additionaly , we assume@xcite mirror reflection symmetry in the @xmath30 plane placed in the qd center , at @xmath31 [ fig .
[ fig1](a ) ] , hence the stationary scattering states @xmath32 , eigenstates of @xmath27 , are either even ( @xmath33 ) or odd ( @xmath34 ) with respect to @xmath35 reflection : for @xmath36 , @xmath37 and @xmath38 .\ ] ] here the even and odd outgoing waves are phase shifted by @xmath39 and @xmath40 , respectively , @xmath41 is the wave vector , @xmath42 is the qd nominal longitudinal axis .
the left and right travelling states @xmath43 are superpositions of even and odd states : @xmath44.\ ] ] inside the qd , @xmath45 , electrons experience two - body coulomb interactions in addition to the confinement potential [ cf .
eq . ] .
we first generalize the results of ref . for spinless ,
non interacting electrons by including spins .
the transmission amplitude @xmath46 for travelling states @xmath43 is : @xmath47 each time @xmath48 appearing in eq .
( [ eq : t ] ) changes sign , due to a variation of either @xmath39 or @xmath40 as a new electron tunnels into the qd , then a lapse of @xmath0 occurs for the transmission phase @xmath3 . since this happens when @xmath49 [ cf .
, the lapse is located in the conductance valley.@xcite we then include all many - body correlations by connecting the phase shift @xmath50 per channel @xmath51 ( @xmath52 labels the parity ) to the exact spectral density @xmath53 accumulated at the qd via the fsr [ cf .
( 20 ) of ref . ] : @xmath54 here @xmath53 is the density displaced at the qd by the electron in the scattering state @xmath51 tunneling at the energy @xmath55 fixed by the chemical potential @xmath56 .
we mimic the action of the plunger gate in the linear response regime by varying the value of @xmath57 with respect to the qd energy levels . in practice , to use the fsr we integrate it over the energy window between two consecutive coulomb valleys with @xmath4 and @xmath58 electrons in the qd , respectively.@xcite whereas this procedure provides the information on the total phase variation only , @xmath59 , it allows to compute @xmath60 from the interacting hamiltonian of the _ isolated _ dot , @xmath61 . this key result is based on the conservation of the total number of scattering plus qd states both in the presence and absence of the qd in the arm of the interferometer.@xcite the ci evaluation of @xmath62 relies on the formula @xmath63 where @xmath64 is the exact interacting gs of the isolated qd with @xmath4 electrons of energy @xmath65 , @xmath66 , and @xmath67 creates an electron with spin @xmath28 in the orbital of given parity @xmath68 and further specified by the set of quantum numbers @xmath69 .
equation follows from eq .
( 21 ) of ref . , which was inferred by connecting the phase shift to the delay time spent by the electron wave packet in the qd .
this delay is obtained by integrating the wave function square modulus over both time and space . by orthogonality of qd orbitals ,
only terms diagonal in @xmath69 indices survive in the formula .
we eventually link the transmission phase variation @xmath22 to @xmath70 through eq .
( [ eq : t ] ) . in the coulomb blockade regime
only one scattering channel @xmath51 is active at time between two consecutive valleys with respectively @xmath4 and @xmath58 electrons .
the active channel is univocally determined by the total spins and parities of @xmath64 and @xmath71 , as obtained by ci . on the other hand , in the kondo regime time - reversal invariance ( recall that @xmath9 ) implies that @xmath72 , evaluated as half the coulomb blockade value given by eq .
( [ eq : exemplum ] ) . in this way
we regain at once the result@xcite that @xmath73 for the addition of the first electron .
in fact , the term on the right hand side of eq .
( [ eq : exemplum ] ) is trivially one when @xmath74 .
the interacting hamiltonian of the isolated qd is @xmath75 where the single particle ( sp ) term is @xmath76 here @xmath77 , @xmath78 is the dielectric constant , @xmath79 is the electron effective mass , @xmath80 is the gyromagnetic factor , and the qd confinement frequencies in the @xmath26 and @xmath81 directions , @xmath82 and @xmath83 , have characteristic lengths @xmath84^{1/2}$ ] and @xmath85^{1/2}$ ] , respectively ( the ratio @xmath86 is related to the ellipse eccentrity ) . in eq .
( [ eq : hi ] ) the weak @xmath29 does not affect orbital degrees of freedom . to wholly include in our theory coulomb correlation
, we solve numerically the few - body problem of eq .
( [ eq : hi ] ) by means of the full ci method ( also known as exact diagonalization , for details see ref . ) .
the ci few - body gs @xmath64 is essentially a linear combination of the slater determinants @xmath87 , @xmath88 with the unknown @xmath89s being the output of the calculation . here
the determinants @xmath87 are obtained by filling in all possible ways with @xmath4 electrons the @xmath90 lowest - energy sp orbitals ( two - fold spin degenerate at @xmath9 ) , eigenstates of the sp hamiltonian . in the fock space of these slater determinants
@xmath61 is a large sparse matrix , that we exactly diagonalize by means of the parallel code _
donrodrigo_,@xcite eventually obtaining the coefficients @xmath89 of eq . .
the diagonalization proceeds in each hilbert space sector labeled by @xmath4 , the total spin , and the total parity of the few - body wave function .
after we have obtained the gss @xmath64 and @xmath71 , we evaluate @xmath60 via eq .
( [ eq : exemplum ] ) , and eventually infer @xmath22 as explained in sec .
[ s : fsr ] . in the ci calculations reported in sec .
[ s : corr ] we used @xmath91 and diagonalized matrices of maximum linear size 2.25 @xmath92 . the relative error for
the energy was less than @xmath93 for @xmath94 .
to illustrate the effect of the inclusion of the spin degree of freedom in the calculation of the transmission phase @xmath3 let us consider for the time being only the sp hamiltonian , @xmath95 , and neglect coulomb interaction .
the gs is a slater determinant with the lowest @xmath4 spin - orbitals filled , @xmath96 ( @xmath97 is the vacuum ) .
the _ aufbau _ filling sequence for @xmath98 is depicted in fig .
[ fig1](b ) for @xmath99 .
the first two electrons occupy the @xmath100 orbital with opposite spin , then the 3rd and 4th electrons fill in the @xmath101 orbital , which is shifted in energy from the @xmath102 orbital by @xmath103 .
note that the first electron entering a new sp level is always @xmath104 , due to the effect of @xmath29 .
the evaluation of the phase shift @xmath70 at each electron addition is straightforward , since in eq .
( [ eq : exemplum ] ) only one addendum gives a non - zero contribution to @xmath105 exactly one as a new spin - orbital @xmath106 is occupied ; the other ones vanish due to the orthogonality of the states . therefore in fig .
[ fig1](b ) one has @xmath107 for the sequence @xmath108 , @xmath109 , @xmath110 , @xmath109 , @xmath111 , @xmath112 of six consecutive electron additions , with the @xmath101 ( @xmath102 ) orbital even ( odd ) under @xmath35 .
the evolution of @xmath3 for the filling sequence of fig . [ fig1](b ) is shown in fig . [ fig1](c ) . both increments and lapses of @xmath3
are derived through eq .
( [ eq : t ] ) ( lapse locations in the conductance valleys with fixed @xmath4 are arbitrary ) .
a remarkable feature of fig . [ fig1](c ) is that @xmath3 increases by @xmath0 in both transitions @xmath113 and @xmath114 , since the first two electrons occupy the same @xmath100 orbital with opposite spin .
this is fundamentally different from the spinless case,@xcite where a total increase of @xmath3 by @xmath115 between @xmath74 and @xmath116 occurs only if the two electrons occupy orbitals of different parities .
two lapses of @xmath0 occur for @xmath3 in the blockaded regions with @xmath117 and @xmath118 [ fig .
[ fig1](c ) ] as the phases @xmath119 and @xmath120 increase more than @xmath0 , respectively . the whole pattern of @xmath3 in fig .
[ fig1](c ) up to five electrons coincides with fig . 4a of ref . ,
provided that one interprets the @xmath121 smooth transition as a phase lapse ( in sec .
[ s : corr ] we consider an alternative interpretation ) .
this agreement is surprising , since in the experiment sp levels have a small energy separation ( @xmath122 mev ) , if compared to characteristic coulomb energies ( @xmath123 mev),@xcite and therefore one would expect significant differences from the non - interacting model of fig .
[ fig1 ] . on the other hand
, the @xmath3-evolution would be basically the same as in fig .
[ fig1 ] if the interacting qd ground state were well approximated by a single slater determinant , as in hartee - fock theory where coulomb interaction is included as a mean field .
when correlation effects beyond the mean - field level@xcite are relevant , we expect that @xmath124 , as suggested in ref . . since the ci ground states
@xmath64 are superpositions of the slater determinants @xmath87 [ cf .
after expansion on this basis many cross terms give no contribution to eq .
( [ eq : exemplum ] ) : the stronger the correlation , the larger the number of slater determinants , the smaller @xmath60 .
this seems to be the case in ref . for @xmath125 in the @xmath7 transition of fig .
4b and @xmath125 for @xmath126 in fig .
5 . this could even be the case for @xmath127 for @xmath128 in fig .
4a of ref . ,
if one excludes the possibility of a phase lapse .
such interpretation is alternative to the one suggested in the previous section . in the @xmath129 space , in units of @xmath0 , for the @xmath7 transition in the coulomb blockade regime .
the grey code goes from 0 ( white ) to 1 ( black ) .
the value in the kondo regime is obtained by dividing @xmath22 by two.,width=336 ] to assess the impact of correlation in the ci results , we parameterize the electron density @xmath130 of the circular dot ( @xmath131 ) via the dimensionless radius @xmath132 of the circle whose area is equal to the area per electron , @xmath133 $ ] , where @xmath134 is the effective bohr radius and @xmath130 is estimated as in ref . .
we next focus on the evolution of @xmath3 as a function of both @xmath132 and ellipse anisotropy ratio @xmath86 .
the first electron addition in the coulomb ( kondo ) regime always gives @xmath135 ( @xmath73 ) .
then @xmath3 remains constant in the @xmath5 valley , independently from the values of either @xmath132 or @xmath86 [ cf .
[ figlast ] ] .
the second electron addition is analyzed in fig .
[ figcontour ] , plotting in the @xmath129 space the contour map of @xmath22 for @xmath7 .
here we vary the anisotropy ratio @xmath86 by keeping the ellipse area @xmath136 fixed , so the density remains constant .
the contour lines of fig .
[ figcontour ] provide the value of @xmath22 in the coulomb blockade regime , whereas its kondo counterpart may be simply obtained by dividing @xmath22 by two ( cf .
[ figlast ] ) .
as @xmath132 increases , @xmath22 monotonously decreases , since correlation effects become stronger at lower density , as the coulomb term in @xmath61 [ eq . ( [ eq : hi ] ) ] overcomes the sp term .
a similar trend occurs by increasing @xmath86 , since a stronger anisotropy effectively lowers the dimensionality of the system , again enforcing correlation effects.@xcite note that by overlapping the experimental value @xmath125 with the plot of fig .
[ figcontour ] we find @xmath137 for @xmath94 .
this value corresponds to @xmath138 mev for gaas , which is comparable to the experimental estimate of 0.5 mev.@xcite from the analysis of ci data of fig .
[ figcontour ] we find that the orbital parities of both the first and second electrons are always even , independently from @xmath132 and @xmath86 .
this prediction agrees with the wigner - mattis theorem : the two - electron gs is always a singlet and the orbital part of its wave function is nodeless.@xcite this result conflicts with other explanations,@xcite and it is expected to hold even in the presence of disorder and/or more complicated potentials .
vs @xmath4 for different dot anisotropies @xmath86 .
solid ( dashed ) lines refer to the coulomb blockade ( kondo ) regime .
we used gaas parameters and @xmath139 mev . from top to bottom : @xmath140 ( 1.5)@xmath141 , ( 1.15)@xmath141 , ( 1.05)@xmath141 , 1 , 1.05 , 1.15 , 1.5.,width=288 ] in fig . [ figlast ] we follow the evolution of @xmath3 up to five electrons for a significant range of qd anisotropies . at the experimental density ( @xmath139 mev ) a slight variation of @xmath86 is sufficient to alter the phase behavior for @xmath6 .
indeed , the relative differences between the values of @xmath86 for the 3rd , 4th , and 5th panels are as small as 5% .
therefore , @xmath3 is sensitive to fluctuations of the experimental qd parameters , as reported in ref . .
the occurrence of alternative scenarios in fig .
[ figlast ] is another signature of correlation .
in fact , several excited states lie very close in energy to the gs , as it is the case in the crossover to electron crystallization.@xcite hence a small deformation of the qd shape easily induces a crossing between states of different symmetry .
we here highlight only the most relevant features of a rich zoology , focusing on coulomb blockade results ( solid lines in fig .
[ figlast ] ) . in a circular
qd at such low density ( 4th panel of fig .
[ figlast ] ) the three - electron gs is a spin quadruplet as an effect of correlation.@xcite because the two - electron gs is a singlet , the transition @xmath142 is spin blockaded , i.e. , @xmath143 without any lapse as @xmath144 .
a slight deformation of the qd ( 3rd and 5th panels ) changes the @xmath118 gs into a doublet , lifting the spin blockade ( @xmath145 between @xmath116 and @xmath118 ) .
the @xmath146 gs is a more robust triplet , since the spin polarization is due to hund s rule an open shell effect.@xcite however , a stronger deformation of the qd ( 2nd and 6th panels ) breaks the orbital degeneracy of the sp levels of the 2nd shell inducing a transition to a singlet gs . at such anisotropy ratios singlets and doublets
typically alternate for even and odd electron numbers , respectively .
a further increase of the deformation ( 1st and 7th panels ) changes the filling sequence of higher - energy orbitals . in fig .
[ figlast ] we also plot @xmath3 in the kondo regime ( dashed lines ) for those cases such that the qd spin is totally screened by the cloud of opposite - spin tunneling electrons.@xcite this excludes high - spin gss other than singlets an doublets occuring in the 3rd , 4th , and 5th panels .
the hallmark of correlation is that @xmath22 is a fraction of @xmath0 and @xmath1 in the coulomb blockade and kondo regimes , respectively [ e.g. , compare the sp phase evolution of fig .
[ fig1](c ) with its correlated counterpart in the 2nd panel of fig . [ figlast ] ] .
in conclusion , we highlighted the role of exchange and correlation in the transmission phase of a few - electron quantum dot .
our findings are relevant for transport experiments through strongly interacting nano objects , including molecules and carbon - based nanostructures .
35 m. avinum - kalish , m. heiblum , o. zarchin , d. mahalu , and v. umansky , nature ( london ) * 436 , * 529 ( 2005 ) .
m. zaffalon , a. bid , m. heiblum , d. mahalu , and v. umansky , phys .
* 100 , * 226601 ( 2008 ) .
r. schuster , e. buks , m. heiblum , d. mahalu , v. umansky , and h. shtrikman , nature ( london ) * 385 , * 417 ( 1997 ) .
g. hackenbroich , phys . rep .
* 343 , * 463 ( 2001 ) . c. karrasch , t. hecht , a. weichselbaum , y. oreg , j. von delft , and v. meden , phys . rev .
* 98 , * 186802 ( 2007 ) .
t. hecht , a. weichselbaum , y. oreg , and j. von delft , phys .
b * 80 , * 115330 ( 2009 ) .
v. meden and f. marquardt , phys .
* 96 , * 146801 ( 2006 ) .
d. i. golosov and y. gefen , phys . rev .
b * 74 , * 205316 ( 2006 ) . o. entin - wohlman , a. aharony , y. imry , and y. levinson , europhys
* 50 , * 354 ( 2000 ) .
a. bertoni and g. goldoni , phys .
b * 75 , * 235318 ( 2007 ) . p. g. silvestrov and y. imry , new j. phys .
* 9 , * 125 ( 2007 ) .
s. a. gurvitz , phys .
b * 77 , * 201302 ( 2008 ) .
a. yahalom and r. englman , phys .
b * 74 , * 115328 ( 2006 )
. l. o. baksmaty , c. yannouleas , and u. landman , phys .
* 101 , * 136803 ( 2008 ) .
m. rontani , phys .
* 97 , * 076801 ( 2006 ) .
d. c. langreth , phys . rev .
* 150 , * 516 ( 1966 ) . t. k. ng and p. a. lee , phys .
* 61 , * 1768 ( 1988 ) .
m. rontani , c. cavazzoni , d. bellucci , and g. goldoni , j. chem . phys .
* 124 , * 124102 ( 2006 ) .
the mirror symmetry is dispensable here , since the eigenstates of @xmath27 in the generic case are obtained by a simple rotation of the @xmath147 basis [ m. pustilnik and l. i. glazman , phys .
* 87 , * 216601 ( 2001 ) ] .
lee , phys .
* 82 , * 2358 ( 1999 ) .
t. taniguchi and m. bttiker , phys .
b * 60 , * 13814 ( 1999 ) . for a direct evaluation of @xmath148
see m. olea and b. r. buka , phys .
b * 75 , * 125301 ( 2007 ) .
website : http://www.s3.infm.it/donrodrigo . here
we disregard the relaxation of self - consistent orbitals at the hartree - fock level , which may decrease @xmath60 from its unit value . on the other hand ,
our ci results for @xmath60 are exact , including both the mean - field relaxation as well as the effect of correlations beyond mean - field .
the latter are dominant at large values of @xmath132 . c. p. garca , v. pellegrini , a. pinczuk , m. rontani , g. goldoni , e. molinari , b. s. dennis , l. n. pfeiffer , and k. w. west , phys .
* 95 , * 266806 ( 2005 ) .
electrons are expected to crystallize in one - dimensional quantum dots embedded in semiconducting carbon nanotubes [ a. secchi and m. rontani , arxiv.0908.2092 ] .
d. c. mattis , _ the theory of magnetism _
( harper , 1965 , new york ) . s. m. reimann and m. manninen , rev .
* 74 , * 1283 ( 2002 ) .
s. kalliakos , m. rontani , v. pellegrini , c. p. garcia , a. pinczuk , g. goldoni , e. molinari , l. n. pfeiffer , and k. w. west , nature phys .
* 4 , * 467 ( 2008 ) .
a. singha , v. pellegrini , a. pinczuk , l. n. pfeiffer , k. w. west , and m. rontani , phys .
* 104 , * 246802 ( 2010 ) .
a. c. hewson , _ the kondo problem to heavy fermions _
( cambridge university press , 1993 , cambridge ) . | the transmission phase through a quantum dot with few electrons shows a complex , non - universal behavior . here
we combine configuration - interaction calculations treating rigorously coulomb interaction and the friedel sum rule to provide a rationale for the experimental findings .
the phase evolution for more than two electrons is found to strongly depend on dot s shape and electron density , whereas from one to two the phase never lapses . in the coulomb ( kondo )
regime the phase shifts are significant fractions of @xmath0 ( @xmath1 ) for the second and subsequent charge addition if the dot is strongly correlated .
these results are explained by the proper inclusion in the theory of coulomb interaction , spin , and orbital degrees of freedom .
= 1 |
the ejection of two electrons caused by absorbtion of a single photon is one of the fundamental few - body processes in atomic physics .
the process is called double photoionization or , less frequently , photo - double - ionization .
the characteristic feature of this process is that it proceeds exclusively through the electron - electron interaction .
because of this , double photoionization has long been used as a testing ground for understanding of the electron correlation phenomena .
the traditional system for studying double photoionization is the helium atom , for which numerous experimental and theoretical investigations have been performed during the last four decades ( for a recent review , see ref .
most widely studied is the ratio of the double - to - single photoionization cross sections @xmath1 as a function of the energy of the incoming photon @xmath2 .
early calculations of this ratio were mainly concentrated either at the near - threshold region of the photon energies ( @xmath3 79 ev for helium , with @xmath4 being the double ionization energy ) , where wannier theory is shown to be applicable @xcite , or at the asymptotical nonrelativistic region @xmath5 @xcite ( where @xmath6 is the electron mass ) .
the intermediate region of the photon energies ( @xmath7 200 ev for helium , where @xmath8 reaches its maximum ) turned out to be much more difficult for an accurate theoretical treatment .
reliable theoretical predictions in this region were obtained only in the end of 1990th by means of sophisticated many - body techniques ( notably , the close - coupling methods ) @xcite . since the helium atom is an essentially nonrelativistic system , any relativistic effects in its double photoionization
are considered to be of little importance at the present level of experimental precision . however , the recent experiments on the double @xmath9-shell photoionization of moderately heavy atoms up to silver @xcite demonstrated significant enhancements ( of about factor of five for silver ) of the cross section as compared with results of nonrelativistic calculations . a natural candidate for explaining
this enhancement would be relativity , which obviously can not be disregarded when dealing with the deeply bound electron states in silver .
because of the relative isolation of the @xmath9-shell electrons from the outer electrons , the double @xmath9-shell ionization of a heavy atom is often compared to the double photoionization of the corresponding helium - like ion .
however , an accurate theoretical treatment of double @xmath9-shell ionization should include both the relativistic effects on the inner - shell electrons and the electron - correlation effects induced by the outer shells .
such a calculation is rather difficult and has not been performed so far .
in contrast to many - electron atoms , the helium - like ion is a relatively simple system for which an _ ab initio _ description is feasible .
so far , there have not been any direct measurements of the double photoionization of helium - like ions .
however , with the advent of new powerful light sources , such as the free - electron laser ( flash ) in hamburg , the linear coherent light source ( lcls ) at stanford , and the x - ray free electron laser ( xfel ) at hamburg , the experimental study of various photoabsorbtion processes is going to be possible for a great variety of ions in different charge states @xcite .
measurements of the double photoionization in moderately heavy helium - like ions would allow us to effectively test our understanding of the _ relativistic _ electron - correlation effects .
an alternative approach to the investigation of the double photoionization of highly charged ions is to study it in the inverse kinematics , through the radiative double electron capture .
for bare oxygen , such measurement was recently accomplished in ref .
the experimental upper bound on this process for bare uranium was reported in ref .
@xcite .
the goal of the present investigation is to perform an _ ab initio
_ relativistic calculation of the double photoionization of a helium - like atom .
the electron - electron correlation ( both on the initial and the final states ) will be taken into account to the leading order of the perturbation theory .
to the given order of perturbation theory , the treatment is _ exact _ , i.e. , includes all multipoles of the absorbed photon , the retardation in the electron - electron interaction as well as the interaction of the electrons with the nucleus without any expansion in the binding field .
the treatment is gauge invariant , both with respect to the gauge used for the absorbed photon and the gauge of the electron - electron interaction .
the higher - order electron - correlation effects omitted are estimated by comparing the present numerical results with the experimental data available for helium .
the remaining paper is organized as follows . in sec .
[ sec:1 ] we present a short summary of the relativistic formulas for the single photoionization , which forms the basis for our treatment of the double photoionization . in sec .
[ sec:2 ] we describe the qed theory of the double photoionization to the lowest relevant order of the perturbation theory . sec .
[ sec:3 ] presents details of the calculation .
numerical results are presented and discussed in sec .
[ sec:4 ] . the relativistic units ( @xmath10 )
are used throughout this paper .
in this section we summarize the relativistic formulas for the single photoionization , as they build the basis for the description of the double photoionization
. differential cross section of the photoionization of a one - electron atom is @xmath11 where @xmath12 is the fine - structure constant , @xmath2 is the energy of the absorbed photon , @xmath13 is the initial ( bound - state ) energy , @xmath14 and @xmath15 are the energy and the momentum of the emitted electron , respectively , @xmath16 .
the amplitude of the process @xmath17 is given by @xmath18 where @xmath19 denotes the initial dirac bound state with the relativistic angular quantum number @xmath20 and its projection @xmath21 , and @xmath22 is the wave function of the emitted electron .
the general relativistic expression for the photon absorption operator @xmath23 is @xmath24 where @xmath25 is a three - component vector of the dirac matrices , @xmath26 is the polarization vector of the absorbed photon , @xmath27 is the photon momentum , @xmath28 , and @xmath29 is the gauge parameter . in our treatment ,
all electron states are the eigenfunctions of the dirac hamiltonian , so the gauge - dependent part of @xmath23 vanishes identically .
however , we keep the general gauge of the absorbed photon in actual calculations in order to check the numerical procedure .
the wave function @xmath30 in eq .
( [ 1a ] ) is the one - electron continuum dirac state with the asymptotic momentum @xmath15 , helicity @xmath31 , and the `` @xmath32 '' asymptotic behaviour ( i.e. , the plane wave modified by the coulomb logarithmic phase plus the incoming spherical wave ) @xcite @xmath33 } \nonumber \\ & + f(\theta,\phi)\,\frac{e^{-i[pr+\eta\,\ln 2pr]}}{r}\,\end{aligned}\ ] ] as @xmath34 , where @xmath35 is the sommerfeld parameter .
the free - electron 4-spinors @xmath36 @xcite are normalized by the condition @xmath37 .
the explicit expression of the electron wave function is @xmath38 where @xmath39 are the scattering phases @xcite , @xmath40 , @xmath41 , and @xmath42 are dirac continuum states with given relativistic angular momentum @xmath43 and the momentum projection @xmath44 , normalized on the energy scale @xcite .
taking into account that @xmath45 , the single differential cross section of the photoionization of a hydrogen - like atom is written as @xmath46 where the energy of the emitted electron is fixed by the energy conservation , @xmath47 . in our analysis of the double photoionization , we will need the cross section of the single photoionization of a _ helium - like _ atom . in the independent - particle approximation ,
the wave function of the initial two - electron state is @xmath48 where @xmath49 for the equivalent electrons and @xmath50 otherwise . averaging the cross section over the momentum projection of the initial state @xmath51 , employing the explicit expression for the continuum dirac state ( [ 3 ] ) , and integrating over the angles over the emitted electron
, we obtain the total photoionization cross section of a helium - like atom as @xmath52\,,\end{aligned}\ ] ] where @xmath53 and @xmath54 and we assumed that the energy of the photon is sufficient to ionize any of the two initial - state electrons .
if this is not the case , only one of the two terms in the brackets should be retained .
the electron correlation effects on @xmath55 omitted in the independent - particle approximation are well studied in the literature . in particular , the leading nonrelativistic term of the @xmath56 expansion of the high - energy asymptotics of @xmath55 was obtained in ref .
. however , since we are presently interested in the ratio of the double - to - single photoionization cross sections @xmath57 , we prefer to treat both these cross sections on the same footing , i.e. , to the leading nonvanishing order of perturbation theory with respect to the electron - electron interaction .
and @xmath58 denote the bound electron states , @xmath59 and @xmath60 are the continuum electron states , @xmath61 denotes the incoming photon .
double lines denote electrons propagating in the binding nuclear field .
[ fig : dpi ] ] according to the general rules of quantum field theory @xcite , the differential cross section of the double photoionization is @xmath62 where @xmath63 and @xmath64 are the energy and the momentum of the two final - state electrons , respectively , @xmath13 is the energy of the initial bound state , and @xmath65 is the amplitude of the process . to the leading order of perturbation theory , the amplitude of the process is represented by the two feynman diagrams shown in fig . [ fig : dpi ] , where the antisymmetrization of the initial and the final two - electron states is assumed .
the initial - state wave function is the same as for the single photoionization [ see eq .
( [ 4 ] ) ] and the initial - state energy is @xmath66 .
the wave function of the final state is @xmath67 note that in the present approach , we use the wave functions that do not include any electron - correlation effects ( apart from the antisymmetrization ) ; the electron - electron interaction enters explicitly into the amplitude of the process [ see eq .
( [ e2 ] ) below ] . in the nonrelativistic case ,
this approach has been successfully applied to the double photoionization by many authors , notably in refs .
although the formal parameter of the perturbative expansion is @xmath56 , the leading - term approximation was shown to adequately describe the double photoionization even for helium .
our present treatment is aimed primarily at the helium - like ions , for which the perturbative expansion converges much faster than for helium .
the triple differential cross section of the double photoionization obtained from eq .
( [ e222 ] ) is @xmath68 where the energy of the second electron is fixed by the energy conservation , @xmath69 .
the energy distribution of the emitted electrons can be conveniently parameterized by the fractional energy sharing parameter @xmath70 , @xmath71 where @xmath72 is the threshold value of the photon energy ( the double ionization energy ) . in our approximation , @xmath73 .
let us now consider the _ single differential cross section _ @xmath74 . substituting the explicit expression for the continuum wave function ( [ 3 ] ) into eq .
( [ e123 ] ) , integrating over the angles , and summing over @xmath75 and @xmath76 , we arrive at @xmath77 where the amplitude contains only the spherical - wave dirac continuum states .
the _ total cross section _ is obtained as an integral of the single differential cross section over a half of the energy sharing interval @xmath78 the other half of the energy interval corresponds to interchanging the first and the second electron , which is already accounted for by the wave function .
the general expression for the transition amplitude of the double photoionization process can be obtained by the two - time green s function method @xcite , @xmath79 the first term in the brackets corresponds to the electron - electron interaction modifying the final - state electron wave function ( the left graph in fig .
[ fig : dpi ] ) and the second one , the initial - state electron wave function ( the right graph in fig .
[ fig : dpi ] ) .
the summation over @xmath80 and @xmath81 corresponds to the permutation of the initial and final electrons , @xmath82 or @xmath83 , @xmath84 or @xmath85 , and @xmath86 and @xmath87 are the permutations sign .
the summation over @xmath88 in eq .
( [ e2 ] ) runs over the complete dirac spectrum , @xmath89 , and @xmath90 is the electron - electron interaction operator , @xmath91 where @xmath92 is the photon propagator , @xmath93 are the dirac matrices . in the feynman gauge , the electron - electron interaction takes the form @xmath94 where @xmath95\ , \frac{1-e^{i|\omega|x_{12}}}{x_{12}}\,.\end{aligned}\ ] ]
the final formulas for the single and double photoionization cross sections are presented in appendices [ app:1 ] and [ app:2 ] .
the calculation of the single photoionization cross section is straightforward , in contrast to that of the double photoionization .
the major difficulty in the numerical evaluation of the double photoionization is the summation over the complete spectrum of the dirac equation . in this work
we use the approach based on the analytical representation of the dirac coulomb green function , which is represented by an infinite sum over the partial waves .
the numerical approach was developed in the previous works @xcite , where it was used for calculating the qed and electron - electron interaction corrections to the radiative recombination of electrons with highly charged ions . for a given value of the relativistic angular momentum quantum number @xmath43 , the radial part of the dirac coulomb green function
is represented in terms of the two - component solutions of the radial dirac equation regular at the origin @xmath96 and the infinity @xmath97 , @xmath98 where @xmath99 denotes the energy argument of the green function , @xmath100 and @xmath101 are the radial arguments , and @xmath102 is the step function . for the point coulomb potential ,
the regular and irregular solutions of the radial dirac equation are expressed analytically in terms of the whittaker functions of the first and second kind ( for explicit formulas see , e.g. , ref .
@xcite ) . when the energy argument @xmath99 is real and greater than the electron rest mass @xmath103 , the dirac green function is a complex multi - valued function
so , care should be taken in this case to choose the correct branch ( i.e. , the sign of the imaginary part ) of the green function .
the branch of the green function is fixed by the sign of the infinitesimal imaginary addition @xmath104 in the energy denominators of eq .
( [ e2 ] ) . specifically , the energy argument of the green function in the first term in the brackets of eq .
( [ e2 ] ) , @xmath105 , is greater than the electron mass . for positive - energy intermediate states
, the energy denominator takes the form @xmath106 , which implies that the energy argument of the green function has a small positive imaginary addition and , therefore , the branch cut of the green function @xmath107 should be approached from above .
a serious problem arises in the numerical evaluation of the radial integrals for the left graph in fig .
[ fig : dpi ] ( with the electron correlation modifying the final - state wave function ) . in this case
, the continuum - state dirac wave function has to be integrated together with the dirac green function with the energy argument @xmath103 and the spherical bessel function .
all three functions are strongly oscillating and slowly decreasing for large radial arguments .
it is practically impossible to accurately evaluate such integral by a straightforward numerical integration . in this work ,
we use the method of the complex - plane rotation of the integration contour , which was previously applied by us to the evaluation of the free - free transition integrals in the bremsstrahlung @xcite .
we now consider the evaluation of the problematic radial integrals in more details .
let us introduce the radius of the atom @xmath108 , which is defined as the smallest distance at which all the bound - state electron wave functions vanish . in the inner region @xmath109 , evaluation of each partial - wave expansion term of the amplitude [ eq .
( [ e2 ] ) ] involves a three - dimensional radial integration . in the outer region @xmath110
, however , all the integrals involving the bound - state wave functions reach their asymptotical values and only an one - dimensional integral of the free - free type needs to be evaluated .
the general form of such integral is @xmath111 where @xmath112 is the spherical bessel function originating from the photon propagator in the electron - electron interaction , @xmath113 is the radial component of the dirac continuum wave function @xmath114 , and @xmath115 is the radial component of the irregular solution of the dirac equation originating from the green function @xmath116 .
the energy conservation requires that @xmath117 , which leads to the inequality @xmath118 , where @xmath119 are the electron momenta associated with the corresponding energies .
we now analytically continue the integrand of eq .
( [ e5 ] ) into the complex @xmath120 plane and take into account the asymptotical behaviour of individual functions for large values of @xmath121 , @xmath122\ , , \\
\phi_{\kappa_1}^{{\infty}}(e_1,r ) & \
\sim \exp\left [ p_1\ , \mbox{\rm im}(r)\right]\ , , \\
j_l(k_3 r ) & \
\sim \exp\left[k_3\,\left|\mbox{\rm im}(r)\right|\right]\,,\end{aligned}\ ] ] where only the leading exponential behaviour is kept .
we observe that , rotating the integration contour in the integral ( [ e5 ] ) into the lower complex half - plane @xmath123 , we transform the strongly oscillating integrand into an exponentially decreasing one , which falls off as @xmath124 $ ] for large @xmath120 .
after the rotation of the contour , the integral can be easily evaluated numerically up to a desired precision . after the radial integrals are successfully evaluated , the remaining problem is the summation of the partial - wave expansions .
altogether there are five partial - wave expansions to be delt with : the expansions of the both final electron wave functions , the green function , the wave function of the absorbed photon , and the photon propagator of the electron - electron interaction .
after the angular momentum selection rules are taken into account , only two partial - wave expansions of the two final - state electrons remain unbound .
the corresponding expansion parameters are @xmath125 and @xmath126 in eq .
( [ e124 ] ) .
the convergence of the resulting double partial - wave expansion is good for the photon energies near the threshold @xmath127 and a non - symmetric energy sharing @xmath128 but gradually deteriorates when @xmath2 increases and @xmath70 approaches @xmath129 . in the most difficult case considered here , @xmath130 and @xmath131 ,
up to @xmath132 partial waves were included into the calculation .
the truncated partial wave expansion was extrapolated into infinity by using the @xmath133 resummation algorithm ( see , e.g. , ref .
@xcite ) . in order to check the numerical procedure
, we evaluated the nonrelativistic limit of our calculations . the easiest way to do
this is to decrease the value of the fine structure constant ( or , equivalently , to increase the value of the speed of light ) by a large factor .
however , a straightforward implementation of this scheme leads to numerical instabilities , since the relativistic operators involve products of the upper and the lower components of the dirac wave function and the lower component vanishes in the nonrelativistic limit . in order to avoid this problem
, we express the lower component in terms of the upper component as @xmath134 which is valid to the leading order in @xmath135 .
after this substitution , our relativistic code yielded a stable numerical limit when the fine structure constant was decreased by orders of magnitude , thus giving us the nonrelativistic limit .
we start this section with a discussion of the total cross section of the double photoionization . as explained previously
, our approach accounts for the electron correlation to the first order of perturbation theory . within this approximation
we perform a rigorous relativistic treatment , without any further simplifications .
this approach is expected to yield accurate results for heavy helium - like ions , whereas for light ions , the precision of the method gradually deteriorates with decrease of the nuclear charge . for helium ,
the relativistic effects are weak but the electron correlation is strong , so that in this case our approach predictably yields worse results than the modern nonrelativistic methods .
however , we will use the helium case for estimating the higher - order electron correlation effects omitted , having in mind that in helium - like ions these effects are suppressed by the inverse power of the nuclear charge . in fig .
[ fig : nr ] , we compare the nonrelativistic limit of our calculations of the ratio of the double - to - single photoionization cross section @xmath136 in helium - like ions with the experimental data available for helium and with the theoretical high - energy limits . in order to facilitate the analysis of the results , we plot the scaled ratio @xmath137 as a function of the ratio @xmath138 .
firstly , we confirm the known statement @xcite that the _ nonrelativistic limit _ of the scaled ratio @xmath139 is well described by a universal function that does not depend on the nuclear charge number
. we will show later that this universal scaling is strongly violated by the relativistic effects .
[ fig : nr ] shows that the numerical results calculated for different nuclear charges are practically indistinguishable from each other in the nonrelativistic limit .
secondly , we observe good agreement between our nonrelativistic results and the leading term of the @xmath56 expansion of the asymptotics @xcite , @xmath140 .
the deviation of our numerical results from the experimental data @xcite is consistent with the deviation of the perturbative asymptotical value @xmath141 from the fully correlated result @xmath142 @xcite .
we , therefore , estimate the higher - order electron - correlation effects omitted in the present treatment to be about @xmath143 for the ratio @xmath8 . ; the dashed ( blue ) line , for @xmath144 ; the dashed - dotted ( red ) line , for @xmath145 .
( the three lines are practically indistinguishable on the picture . )
the triangles , squares , and diamonds represent the experimental results for helium obtained in refs .
@xcite , respectively .
the dotted line shows the nonrelativistic high - energy limit calculated to the leading order of the perturbation theory @xcite .
the dash - dot - dotted line shows the nonrelativistic asymptotical limit calculated with the fully correlated wave functions @xcite .
[ fig : nr ] ] we now turn to our relativistic calculations . fig .
[ fig : diff ] presents our numerical results obtained with a step - by - step inclusion of individual relativistic effects for several helium - like ions .
the four different treatments compared are ( i ) the nonrelativistic calculation ( dotted line ) , ( ii ) the calculation with the relativistic wave functions but with neglecting the retardation in the electron - electron interaction and all multipoles of the absorbed photon higher than the relativistic @xmath146 transition ( dash - dotted line ) , ( iii ) the calculation with the relativistic wave functions and the full retardation but without the higher multipoles of the absorbed photon ( dashed line ) , and ( iv ) the full relativistic treatment ( solid line ) .
we observe that for moderate photon energies , @xmath147 , the dominant relativistic effect is the retardation in the electron - electron interaction , whereas for high photon energies , @xmath148 , the effect of the higher multipoles of the absorbed photon becomes dominant .
altogether , the relativistic effects are large and change qualitatively the shape of the energy dependence of the ratio @xmath8 already for medium-@xmath0 ions .
in particular , the maximum of the curve located for helium at @xmath149 disappears for @xmath150 and @xmath8 becomes a monotonically growing function ( at least , up to the maximal photon energies accessible in our calculations ) , which is in contrast to the nonrelativistic case where @xmath8 gradually decreases to approach a constant high - energy limit from above .
it is interesting to note that similarly strong relativistic effects were recently reported for the nonelastic electron scattering from the hydrogen - like ions @xcite .
significant effects caused by the magnetic dipole and electric quadrupole transitions in the double photoionization of helium - like ions were previously reported in ref .
@xcite . in order to check our relativistic calculations , we employed different gauges for the absorbed photon and , separately , for the photon propagator of the electron - electron interaction .
we found that the gauge - dependent term in the photon absorbtion operator [ see eq .
( [ 1aa ] ) ] vanishes in the actual calculations .
independently , we demonstrated that our calculations with the electron - electron interaction operator in the feynman and the coulomb gauges [ see eqs .
( [ fey ] ) and ( [ cou ] ) ] yield the same results .
this was an important cross - check of the numerical procedure since the contributions of the two diagrams in fig .
[ fig : dpi ] are separately not gauge invariant . as an additional test , we calculated the contribution of the right graph in fig .
[ fig : dpi ] independently by a different numerical technique based on the @xmath151-spline representation of the dirac coulomb green function @xcite .
the results of our relativistic calculations for various helium - like ions are summarized in fig .
[ fig : full ] .
it can be readily seen that the relativistic effects strongly violate the nonrelativistic scaling rule stating that the function @xmath152 does not depend on the nuclear charge .
the relativistic enhancement of the ratio of the double - to - single photoionization cross sections can be conveniently parameterized as @xmath153\,,\end{aligned}\ ] ] where @xmath154 is the nonrelativistic limit of the ratio @xmath155 and @xmath156 is a smooth function of the nuclear charge @xmath0 and photon energy @xmath2 .
this function is plotted for several nuclear charges in fig .
[ fig : relenh ] .
we observe that the numerical values of @xmath156 lay within the interval of @xmath157 for the wide range of nuclear charge numbers and photon energies .
we also found that for @xmath158 , @xmath159 and is nearly @xmath0-independent .
in particular , for helium and 1 kev photon energy ( @xmath160 ) , the above formula predicts the relativistic enhancement of about 0.8% , which is negligible at the current level of experimental precision . the normalized energy distribution of the ejected electrons is shown in fig .
[ fig : sharing ] for several nuclear charges and photon energies .
again , we observe that the scaled energy distribution does not depend on the nuclear charge in the nonrelativistic limit and that this scaling is violated by the relativistic effects .
note that in the case of the energy distribution , the differential cross section has to be scaled by the factor of @xmath161 and not by @xmath162 as for the total cross section ; the additional factor is due to the fact that the integration interval in eq .
( [ eq : total ] ) is proportional to @xmath163 .
we also confirm the known statement @xcite that the highly asymmetric energy sharing between the two emitted electrons dominates at the high photon energies , with one of the electrons carrying away most of the photon energy . to the single photoionization cross section @xmath164 , for helium - like ions with different nuclear charges @xmath0 , as a function of the photon energy @xmath2 divided by the threshold photon energy of the double photoionization @xmath72 .
[ fig : full ] ] defined by eq .
( [ relenh ] ) , for different helium - like ions .
[ fig : relenh ] ]
in this paper , we presented a relativistic calculation of the double photoionization of helium - like atoms .
our approach is based on the partial - wave representation of the dirac continuum states and accounts for the retardation in the electron - electron interaction , the higher - order multipoles of the absorbed photon as well as the interaction of the electrons with the nucleus without any expansion in the binding field .
the electron - electron interaction is taken into account to the first order of perturbation theory .
the omitted higher - order electron correlation effects are estimated by comparing our numerical results for atomic helium with the experimental results and the available nonrelativistic theory .
the calculational results are shown to be gauge invariant both with respect of the gauge of the absorbed photon and the gauge of the electron - electron interaction .
our calculation shows that the relativistic effects become prominent in the double photoionization cross section already for medium-@xmath0 ions .
these effects change the shape of the energy dependence of the ratio of the double - to - single photoionization @xmath8 drastically .
in particular , the well - known constant high - energy asymptotic behaviour of @xmath8 in helium gives place to the monotonically growing behaviour in the case of helium - like targets with @xmath165 .
we note that the reported magnitude of the relativistic effects is insufficient to explain the large discrepancy between theory and experiment observed for the double @xmath9-shell photoionization in neutral atoms .
in particular , for the photon energy of 90 kev in silver @xmath166 , we obtain the relativistic enhancement of about 50% [ see eq .
( [ relenh ] ) ] , which is much smaller than the 4-fold disagreement between theory and experiment @xcite .
it should be mentioned that in the present investigation we do not consider the compton scattering mechanism of the total double photoionization , which is known to contribute significantly for high - energy photons @xcite .
this reaction channel is more difficult for a relativistic calculation than the photoabsorption considered in the present work . to a certain extent
, the photoabsorption and the compton mechanisms of the double photoionization can be distinguished in the experiment by measuring the momentum vector of the recoil ions @xcite .
the work reported in this paper was supported by the helmholtz gemeinschaft ( nachwuchsgruppe vh - ng-421 ) .
the computations were partly performed on the computer cluster of st .
petersburg state polytechnical university .
in order to perform integrations over the angular variables in the matrix element of the photon absorption operator , we fix the @xmath167-axis of our coordinate system to be directed along the photon momentum @xmath27 and assume that the polarization vector @xmath168 has the only nonvanishing spherical component @xmath169 ( @xmath170 ) . the result for the matrix element of the gauge - independent part of the photon absorption operator is @xmath171 where @xmath172 and the initial and final states are the dirac states with given relativistic angular momentum quantum number @xmath43 and the momentum projection @xmath44 .
the radial integrals @xmath173 are @xmath174\,,\end{aligned}\ ] ] where @xmath112 is the spherical bessel function , @xmath175 and @xmath176 are the upper and the lower radial components of the dirac wave function , respectively , and @xmath177 are the angular coefficients given by eqs .
( c7)-(c9 ) of ref .
the matrix element of the gauge - dependent part of the photon absorption operator can be evaluated as @xmath178\,,\end{aligned}\ ] ] where the angular coefficients @xmath179 are given by eq .
( c10 ) of ref .
@xcite and @xmath180 is the radial integral @xmath181\,.\end{aligned}\ ] ]
let us write the amplitude ( [ e2 ] ) as @xmath182 where @xmath183 corresponds to the part of eq .
( [ e2 ] ) with @xmath184 and @xmath185 and the remaining three terms are obtained by permutations .
the contributions due to the two terms in the braces of eq .
( [ e2 ] ) will be denoted by subscripts @xmath186 and @xmath151 , respectively , @xmath187 .
the calculation formulas take the simplest form in the case when the initial electron state is the ground state of the atom . in this case , the permutation over the @xmath188 and @xmath58 electrons yields just a combinatorial factor and all summations over the momentum projections can be evaluated in the closed form .
the result for the single differential cross section is @xmath189\biggr|^2\,.\end{aligned}\ ] ] the amplitude is @xmath190 , s. w. epp , j. r. c. lpez - urrutia , g. brenner , v. mckel , p. h. mokler , r. treusch , m. kuhlmann , m. v. yurkov , j. feldhaus , j. r. schneider , m. wellhfer , m. martins , w. wurth , and j. ullrich , phys .
lett . * 98 * , 183001 ( 2007 ) .
g. bednarz , d. sierpowski , t. sthlker , a. warczak , h. beyer , f. bosch , a. bruning - demian , h. bruning , x. cai , a. gumberidze , s. hagmann , c. kozhuharov , d. liesen , x. ma , p. mokler , a. muthig , z. stachura , and s. toleikis , nucl .
methods * b 205 * , 573 ( 2003 ) , 11th international conference on the physics of highly charged ions .
r. drner , t. vogt , v. mergel , h. khemliche , s. kravis , c. l. cocke , j. ullrich , m. unverzagt , l. spielberger , m. damrau , o. jagutzki , i. ali , b. weaver , k. ullmann , c. c. hsu , m. jung , e. p. kanter , b. sonntag , m. h. prior , e. rotenberg , j. denlinger , t. warwick , s. t. manson , and h. schmidt - bcking , phys . rev
. lett . * 76 * , 2654 ( 1996 ) .
l. spielberger , o. jagutzki , r. drner , j. ullrich , u. meyer , v. mergel , m. unverzagt , m. damrau , t. vogt , i. ali , k. khayyat , d. bahr , h. g. schmidt , r. frahm , and h. schmidt - bcking , phys .
. lett . * 74 * , 4615 ( 1995 ) . | a fully relativistic calculation of the double photoionization of helium - like atoms is presented .
the approach is based on the partial - wave representation of the dirac continuum states and accounts for the retardation in the electron - electron interaction as well as the higher - order multipoles of the absorbed photon .
the electron - electron interaction is taken into account to the leading order of perturbation theory .
the relativistic effects are shown to become prominent already for the medium-@xmath0 ions , changing the shape and the asymptotic behaviour of the photon energy dependence of the ratio of the double - to - single photoionization cross section . |
the hamiltonian formulation of classical mechanics assigns simultaneous , arbitrarily accurate , values for the canonically conjugate position and momentum to distinguishable particles .
indeed in classical mechanics these simultaneous values are regarded as properties of the particles themselves ; measurement simply reveals these values and need not , in principle , add uncertainty to their determination . in quantum mechanics the conjugate position and momentum are represented by noncommuting operators , @xmath0 and @xmath1 , with @xmath2=i\hbar$ ] .
it follows that there is no physical state for which position and momentum can take dispersion free values .
this is the content of the standard formulation of the uncertainty principle @xmath3 , where @xmath4 .
this does not mean , however , that simultaneous measurements of position and momentum are impossible , only that such simultaneous measurements can not be made arbitrarily accurate @xcite .
our objective is to show that the _ observed _ phase space , reconstructed from the classical stochastic measurement record of continuous joint measurements of position and momentum , corresponds to the classical phase space description with added noise . by a careful specification of the measurement model , including the relevant states of the apparatus , together with a markov assumption ,
we show that the observed classical stochastic measurement record corresponds to the conditional quantum averages @xmath5 and @xmath6 .
we obtain a stochastic schrdinger equation for the conditional state of the system with complex noise , together with classical stochastic equations for the variables corresponding to the observed measurement records .
this extends the results of bhattacharya et al .
@xcite where only a position measurement was considered .
other authors @xcite have considered a general type of stochastic schrdinger equation ( which may be interpreted as corresponding to a measurement ) in the framework of _ quantum state diffusion _ to examine dissipative chaos in open systems . a model for the simultaneous measurement of position and momentum was given long ago by arthurs and kelly @xcite .
this model consists of two meters which are allowed to interact instantaneously with the system .
the interaction couples one of the meters to position and the other to momentum , encoding the results of the measurement in the final states of the meters . projective measurements
are then made on each of the meter states separately . in more general terms
the arthurs - kelly model is an explicit example of the gelfand - naimark - segal theorem @xcite which enables us to write a positive joint probability density of two real variables ( the actual measurement outcomes ) in terms of an inner product on an extended hilbert space ( extended to include the two apparatus as well as the system to be measured ) .
it can also be considered as an example of a _ positive operator valued measure _ @xcite .
the arthurs - kelly model suggests a route to a phase space description of quantum mechanics based on the joint outcomes of simultaneous measurements of position and momentum .
however , to achieve a phase space description of the dynamics it is necessary to consider a time sequence of such measurements .
the arthurs - kelly model forces the conditional state of the system into a coherent state after a single measurement and is too strong for this purpose .
we generalize the arthurs - kelly model to allow for a weakening of the measurement . in this way
repeated measurements can take place without invariably reducing the system to a coherent state . by simultaneously increasing the number of measurements and weakening the strength of each measurement ,
the continuous limit is achieved .
we show that under continuous observation the evolution of the system state is described by an it stochastic schrdinger equation . unlike evolution via the ordinary schrdinger equation ,
continuous observation forces the conditional state of the system to remain localized .
hence the quantum mean of the phase space variables can be thought of as a trajectory .
it is in this light that we investigate quantum nonlinear dynamics under continuous observation .
the theory of continuous measurement in quantum mechanics was first developed by barchielli and co - workers @xcite ( see also @xcite ) and some of the results in this paper are contained in their work
. however our approach uses the arthurs - kelly measurement model as a starting point and follows the work on continuous position measurements by caves and one of us in @xcite . in this way our results have a clear physical interpretation . in section
ii we generalize the arthurs - kelly measurement model . in section
iii we derive the stochastic schrdinger equation for continuous measurement which is then investigated analytically in section iv and simulated numerically in section v for a chaotic system . finally , in section vi we discuss our results
following braunstein et al .
@xcite we rederive and generalize the result implicit in the work of arthurs and kelly @xcite .
the simultaneous measurement of position and momentum of a one dimensional quantum system is achieved by constructing an interaction governed by the hamiltonian @xmath7 which couples the system hamiltonian @xmath8 , with two detectors @xmath9 and @xmath10 .
the measurement model also requires that the detectors have been prepared in the initial states @xmath11 where @xmath12 and @xmath13 .
the parameter @xmath14 is called the _ squeezing parameter _ and biases the coupling so that one may obtain more information on either position or momentum .
the parameter @xmath15 will be used to weaken the measurement , decreasing the amount of information collected on both position and momentum . before the interaction the system state @xmath16 and the combined detector state @xmath17
are assumed to be uncorrelated with an initial combined density operator of @xmath18 . at @xmath19 the evolution operator for the interaction @xmath20\ ] ]
couples the system to the measurement apparatus , entangling the system and detector states .
after the interaction the probability of finding the detector positions in the small area @xmath21\times[x_2,x_2+dx_2]$ ] is @xmath22 where @xmath23 and @xmath24 are the traces over the system and detector states , respectively , @xmath25 is the probability density , and the _ resolution operator _
p_1p_2|d_1d_2\rangle \\ & = & \hat{d}(sx_1,\textstyle\frac{1}{s}x_2)\hat{\upsilon}_{\!\sigma}(0,0)\hat{d}(sx_1,\textstyle\frac{1}{s}x_2)^\dag.\end{aligned}\ ] ] the displacement operator @xmath27\ ] ] and @xmath28 we now define the annihilation operator @xmath29 which satisfies @xmath30=1\ ] ] and rewrite the displacement operator as @xmath31 where @xmath32 .
it can be easily shown that @xmath33 the coherent states @xcite are now defined by applying the displacement operator onto the vacuum state @xmath34 they satisfy the following relations @xmath35|\alpha+\beta\rangle \label{a4}\\ \langle\alpha|\beta\rangle & & = \exp\left(-\textstyle\frac{1}{2}|\alpha|^2-\textstyle\frac{1}{2}|\beta|^2+\alpha^*\beta\right ) \label{a5}\\ \pi^{-1}\int d^2\alpha&&|\alpha\rangle\langle\alpha|=1 . \label{a6}\end{aligned}\ ] ] now defining @xmath36 we can rewrite ( [ up00 ] ) as @xmath37 using the above relations and integrating we obtain @xmath38 where @xmath39 denotes normal ordering .
hence , @xmath40 if we now define @xmath41 then the resolution operator becomes @xmath42 : \\ & \equiv & ( 2\hbar)^{-\frac{1}{2}}\hat{\upsilon}_{\!\sigma}(\chi ) .
\label{res}\end{aligned}\ ] ] the probability of finding the detector positions in the small area @xmath43\times[\chi_2,\chi_2+d\chi_2]$ ] is now @xmath44 where @xmath45 and @xmath46:\end{aligned}\ ] ] is an _ effect density _ @xcite .
it can be easily shown that @xmath47 hence , defining the notion of a mean for this measurement process @xmath48 we find that @xmath49 or @xmath50 where @xmath51 is the quantum expectation . thus the readout variables @xmath52 and @xmath53 give , respectively , the position and momentum of the system with additional noise dependent on @xmath15 .
we obtain maximal information from the system when the variances are at a minimum .
that is , when @xmath54 . in this case
@xmath55 : \\ & = & \pi^{-1}|\chi\rangle\langle\chi|\end{aligned}\ ] ] and the probability density reduces to the husimi density or @xmath56 function @xmath57 suppose we take a measurement and obtain the outcome @xmath58 . as a consequence of the strength of this measurement
the system state collapses to @xmath59 however when @xmath60 the resolution operator has the asymptotic expansion @xmath61+o(\textstyle\frac{1}{\sigma^5 } ) .
\label{asympx}\end{aligned}\ ] ] using this result one can show that the system state conditioned on the measurement outcome @xmath58 has the expansion @xmath62-(\hat{a}^\dag-\chi'^*)(\hat{a}-\chi'),\hat{\rho}\bigr\}+o(\textstyle\frac{1}{\sigma^4})\ ] ] and thus , if @xmath15 is large enough , the process of measurement will have negligible effect on the system .
consider a sequence of phase space measurements governed by the hamiltonian @xmath63 where after each measurement the detectors are reset into the initial states given by equation ( [ dxi ] ) or ( [ dpi ] ) .
the assumption that the detectors are reset is equivalent to making a markov assumption for a single apparatus coupled to the system . by resetting the detector at each time step we ensure that no coherent memory of the system state survives in the states of the apparatus . following @xcite we will first derive the master equation for unconditional ( or nonselective ) evolution of the system density operator in the continuous limit @xmath64 , @xmath65 . by unconditional evolution
we mean that no account is taken of the measured results .
thus after each measurement occurs we ignore the result and average over all possible measurement outcomes . if we denote the system density operator immediately before the @xmath66-th measurement by @xmath67 then @xmath68 where @xmath69 for any operator @xmath70 it is possible to show that @xmath71\right]+o(\textstyle\frac{1}{\sigma^4}).\ ] ] hence we obtain @xmath72-\textstyle\frac{1}{{\delta t}\sigma^2}\left[\hat{a},\left[\hat{a}^\dag,\hat{\rho}(n{\delta t})\right]\right]+o({\delta
t})+o(\textstyle\frac{1}{\sigma^2})+o(\textstyle\frac{1}{{\delta t}\sigma^4}).\ ] ] by setting @xmath73 and taking the continuous limit @xmath64 , @xmath65 , with @xmath74 held constant , we obtain the master equation for unconditional evolution : @xmath75-\gamma\left[\hat{a},\left[\hat{a}^\dag,\hat{\rho}\right]\right ] \label{uncon}\\ & = & -\textstyle\frac{i}{\hbar}\left[\hat{h},\hat{\rho}\right]-\textstyle\frac{1}{2\hbar}\gamma_1\big[\hat{x},\big[\hat{x},\hat{\rho}\big]\big]-\textstyle\frac{1}{2\hbar}\gamma_2\big[\hat{p},\big[\hat{p},\hat{\rho}\big]\big ] \label{xpuncondequ}\end{aligned}\ ] ] where @xmath76 and @xmath77 . this equation has already been derived by barchielli et al .
@xcite , but their approach is somewhat different . by setting @xmath78 in ( [ xpuncondequ ] )
we obtain the unconditional master equation for continuous position measurements previously derived in @xcite .
we now wish to derive the conditional ( or selective ) master equation for the system density operator . in this case
the evolution of the system is conditioned on a history of measurement readouts @xmath79 where @xmath80 is the detector position for the @xmath66-th measurement .
hence , if @xmath67 is the system density operator immediately before the @xmath66-th measurement , then @xmath81^{-1}\hat{\upsilon}_{\!\sigma}(\chi(n{\delta t}))\hat{\rho}(n{\delta t}){\hat{\upsilon}_{\!\sigma}(\chi(n{\delta t}))}^\dag{\hat{u}}^\dag.\ ] ] to proceed we extend the definition of the readout variable by setting @xmath82 and introduce the new variable @xmath83 hence @xmath84 and @xmath85 using ( [ m1 ] ) and ( [ m2 ] ) we obtain @xmath86 where the subscript @xmath87 has been added to emphasize that the mean and variance are conditioned through @xmath16 on the entire history of measurement readouts . now letting @xmath88 we find that @xmath89 where we have set @xmath74 to be constant in anticipation of the continuous limit . hence for @xmath90 small enough we can approximate increments in the variable @xmath91 by @xmath92 where it is understood that the complex it increment @xmath93 is of order @xmath94 and satisfies @xmath95 , @xmath96 . in the continuous limit @xmath64 with @xmath73 constant
we have @xmath97 where @xmath98 is a complex wiener process @xcite and the it differential @xmath99 satisfies the algebra @xmath100 to simplify the following we will always replace @xmath101 by @xmath90 and set @xmath102 in anticipation of the above algebra in the continuous limit . using ( [ xincrement1 ] ) together with ( [ xincrement2 ] ) we obtain the following asymptotic expansion for the resolution operator ( [ res ] ) @xmath103 where @xmath104 and it is understood that @xmath105 , @xmath106 and @xmath107 . hence we find that @xmath108\right]\right)+o({\delta t}^3)\ ] ] and thus , using ( [ asympu ] ) we obtain @xmath109+\gamma^\frac{1}{2}\left\{\hat{{\cal a}}^\dag\delta \xi + \hat{{\cal a}}\delta\xi^*,\hat{\rho}\right\}-\gamma{\delta t}\left[\hat{{\cal a}},\left[\hat{{\cal a}}^\dag,\hat{\rho}\right]\right]+o({\delta t}^\frac{3}{2 } ) \\ & = & \hat{\rho}-\textstyle\frac{i{\delta t}}{\hbar}\left[\hat{h},\hat{\rho}\right]-\gamma{\delta t}\left[\hat{a},\left[\hat{a}^\dag,\hat{\rho}\right]\right]+\gamma^\frac{1}{2}{\cal h}[\hat{a}^\dag]\hat{\rho}\delta \xi+\gamma^\frac{1}{2}{\cal h}[\hat{a}]\hat{\rho}\delta\xi^*+o({\delta t}^\frac{3}{2})\end{aligned}\ ] ] where we have defined the superoperator @xmath110\hat{\rho}\equiv\left\{\hat{a}-{\,\text{tr}}(\hat{a}\hat{\rho}),\hat{\rho}\right\}.\ ] ] in the limit @xmath64 we obtain the master equation for conditional evolution : @xmath111dt-\gamma\left[\hat { a},\left[\hat{a}^\dag,\hat{\rho}(t)\right]\right]dt+\gamma^\frac{1}{2}{\cal h}[\hat{a}^\dag]\hat{\rho}(t)d\xi(t)+\gamma^\frac{1}{2}{\cal h}[\hat{a}]\hat{\rho}(t)d\xi(t)^*.\end{aligned}\ ] ] it is easy to see that upon averaging this stochastic differential equation we reproduce our original master equation for unconditional evolution ( [ uncon ] ) .
however , note that unlike the unconditional equation , this equation preserves the pure state property of @xmath16 .
one can easily prove this by showing @xmath112 under the assumption that @xmath113 , where @xmath114 . as a consequence of this fact , the above master equation has an analogue for pure state evolution in terms of a stochastic schrdinger equation : @xmath115 where @xmath116 . in terms of position and momentum variables
the master equation for conditional evolution reads as @xmath117dt-\textstyle\frac{1}{2\hbar}\gamma_1\big[\hat{x},\big[\hat{x},\hat{\rho}\big]\big]dt-\textstyle\frac{1}{2\hbar}\gamma_2\big[\hat{p},\big[\hat{p},\hat{\rho}\big]\big]dt+\hbar^{-\frac{1}{2}}{\gamma_1}^{\frac{1}{2}}{\cal h}[\hat{x}]\hat{\rho}dw_1+\hbar^{-\frac{1}{2}}{\gamma_2}^{\frac{1}{2}}{\cal h}[\hat{p}]\hat{\rho}dw_2 \label{xpcondequ}\ ] ] where @xmath118 and the readout variables @xmath52 and @xmath53 obey the following stochastic processes @xmath119 with @xmath120 by setting @xmath78 in ( [ xpcondequ ] ) we obtain the conditional master equation for continuous position measurements previously derived in @xcite .
note that @xmath52 and @xmath53 are charged by stationary white noise @xmath121 making their graph highly irregular .
it is thus better to represent the measured trajectory by @xmath122 and @xmath123 .
we will now investigate the effect of measurement on the system state by setting @xmath124 and defining @xmath125 where @xmath126 and @xmath127 are the expected variances in position and momentum , and @xmath128 is the expected covariance between position and momentum .
the average is taken over all possible measurement histories .
one can then derive the following set of coupled differential equations : @xmath129 the solutions of which are @xmath130 where @xmath131 , @xmath132 and @xmath133 .
hence the process of measurement induces the system state to collapse into a coherent state . if the measurement retrieves no information on momentum , i.e. @xmath78 , then @xmath134 and the growth in the momentum variance is unbounded .
similarly , if the measurement retrieves no information on position then the position variance grows unbounded .
however , when both position and momentum are measured simultaneously the system state is forced into a coherent state .
when @xmath135 we expect that for a suitable choice of @xmath136 , any spreading of the quantum wavepacket caused by nonlinearities in the hamiltonian will be counteracted by the measurement induced localization .
one might naively assume that if the measurement only retrieves information on position ( or momentum ) then the state will not localize .
however this is not always the case .
note that when @xmath137 in equation ( [ vp ] ) the momentum variance will initially decrease .
thus if the system dynamics is such that it increases the covariance between position and momentum , then the continuous measurement of position may also localize momentum .
for example , consider the hamiltonian describing free particle motion , @xmath138 .
when @xmath78 the variances and covariance satisfy @xmath139 although we could not solve these equations analytically , it is easy to see that all physical solutions asymptotically attract to the stable fixed point @xmath140 hence , measurement of position does not introduce a diffusion in momentum , and the state localizes ( this result has been derived previously in @xcite ) .
however , for free particle motion , if we only measure momentum the state does not localize .
the system dynamics accelerates the growth in the position variance .
see @xcite for more on localization .
we will now numerically investigate the solution of the stochastic schrdinger equation ( [ stochschro ] ) for the driven system @xmath141 the numerical method to solve this equation is simple . to take advantage of
the measurement induced localization we use a local moving number basis @xmath142 truncated at some finite value @xmath143 .
the stochastic terms are integrated using the first - order euler method while other terms are integrated by diagonalizing the position and momentum operators and using the split - operator formula .
we will first consider an integrable case when @xmath144 , @xmath145 , @xmath146 and @xmath147 .
the initial state was chosen to be a coherent state ( @xmath148 ) centered at @xmath149 when @xmath150 .
the husimi density of the initial state together with the contours of the hamiltonian are plotted in fig .
the husimi density of the evolved state ( @xmath151 ) together with the trajectories @xmath152 for different measurement schemes is plotted in figures 1(b)-(d ) .
the evolved state in fig .
1(b ) is the result when no measurements occur ( @xmath153 ) . in this case , nonlinearities in the hamiltonian cause the state to shear as it evolves , spreading it along the contours .
the trajectory has little meaning when @xmath153 . in fig .
1(c ) the evolved state is the result of continuous simultaneous measurement of position and momentum with @xmath154 and @xmath148 . in this case
the state has remained localized as it follows the contours .
the continuous measurement of position only ( @xmath78 ) when @xmath155 has also kept the state localized .
this is shown in fig .
the combined variance of position and momentum is plotted in fig .
we must emphasize that only when both position and momentum are measured together does the trajectory correspond via ( [ xa],[xb ] ) to the outcome of an actual measurement .
if only position is measured , only @xmath122 observed while @xmath123 is simply the result of a mathematical calculation .
now consider the chaotic case when @xmath156 , @xmath157 , @xmath145 , @xmath158 , @xmath159 and @xmath147 .
a poincar stroboscopic map with unit strobing frequency is plotted in fig .
for an initial state the same as above , the evolved state ( @xmath160 ) together with the trajectories for different measurement schemes is plotted in figures 2(b)-(d ) . when no measurement occurs ( fig .
2(b ) ) the chaotic action of stretching and folding spreads the state across the phase space .
however when we continuously measure position and momentum ( fig .
2(c ) ) the state remains localized and the trajectory resembles classical motion with noise .
this noise will vanish as we approach the classical limit @xmath161 . in fig .
3(a ) we have plotted the quantum trajectory for the same parameter values as above except with @xmath162 . in this case
the total variance remains below @xmath163 ( fig .
3(b ) ) and the noise is not visible .
the corresponding classical trajectory is plotted in grey and is only visible when it deviates from the quantum at @xmath164 .
the evolution of this system under the continuous measurement of position has already been studied by bhattacharya et al .
they also find that the measurement keeps the system state localized .
it is not surprising that this is also the case when only momentum is measured ( fig .
we have derived an it stochastic schrdinger equation ( [ stochschro ] ) describing the evolution of a quantum system under the continuous simultaneous measurement of position and momentum .
the outcome of this measurement is a classical stochastic record obeying ( [ calx ] ) . as a consequence of the measurement ,
the system state is forced to remain localized allowing a classical interpretation of the quantum mean of the phase space variables as the trajectory of the system .
this trajectory corresponds to the actual measured trajectory minus noise ( [ xa],[xb ] ) .
furthermore , the localization property allows a well - defined classical limit via ehrenfest s theorem .
indeed , for small @xmath165 , numerical results show that the quantum system approximately follows classical trajectories .
however a more complete theoretical understanding of the classical limit under continuous measurement is needed . 9 p. busch , int . j. theor
* 24 * , 63 ( 1985 ) .
t. bhattacharya , s. habib and k. jacobs , phys .
( to be published ) .
n. gisin and i.c .
percival , j. phys .
a : math . gen . * 25 * , 5677 ( 1992 ) .
n. gisin and i.c .
percival , j. phys . a : math
* 26 * , 2233 ( 1993 ) .
n. gisin and i.c .
percival , j. phys .
a : math . gen . * 26 * , 2245 ( 1993 ) .
spiller and j.f .
ralph , phys .
a * 194 * , 235 ( 1994 ) .
r. schack , t.a .
brun and i.c .
percival , j. phys .
a : math . gen .
* 28 * , 5401 ( 1995 ) .
brun , i.c .
percival and r. schack , j. phys .
a : math . gen . * 29 * , 2077 ( 1996 ) .
e. arthurs and j.l .
kelly , jr . , bell .
j. * 44 * , 725 ( 1965 ) .
conway , _ a course in functional analysis _
( springer - verlag , berlin , 1985 ) .
k. kraus , _ states , effects , and operations : fundamental notions of quantum theory _
( springer - verlag , berlin , 1983 ) . m.a .
nielsen and i.l .
chuang , _ quantum computation and quantum information _
( cambridge , 2000 ) .
a. barchielli , l. lanz and g.m .
prosperi , nuovo cimento b * 72 * , 79 ( 1982 ) .
a. barchielli , nuovo cimento b * 74 * , 113 ( 1983 ) .
a. barchielli , l. lanz and g.m .
prosperi , found .
* 13 * , 779 ( 1983 ) .
l. disi , phys .
a * 129 * , 419 ( 1988 ) .
belavkin and p. staszewski , phys .
a * 140 * , 359 ( 1989 ) .
caves and g.j .
milburn , phys .
rev . a * 36 * , 5543 ( 1987 ) . g.j .
milburn , quantum semiclass . opt . * 8 * , 269 ( 1996 ) .
braunstein , c.m .
caves and g.j .
milburn , phys .
a * 43 * , 1153 ( 1991 ) .
a. perelomov , _ generalized coherent states and their applications _
( springer - verlag , berlin , 1986 ) .
gardiner , _ handbook of stochastic methods _ ( springer - verlag , berlin , 1983 ) .
l. disi , phys .
a * 132 * , 233 ( 1988 ) .
percival , j. phys .
a : math . gen .
* 27 * , 1003 ( 1994 ) .
strunz , i.c .
percival , j. phys .
a : math . gen . * 31 * , 1801 ( 1998 ) .
percival , w.t .
strunz , j. phys .
a : math . gen . * 31 * , 1815 ( 1998 ) .
doherty , s.m .
tan , a.s .
parkins and d.f .
walls , phys .
a * 60 * , 2380 ( 1999 ) .
* fig . 1 .
( a ) * contours of @xmath166 and the initial state ( @xmath147 ) . *
( b - d ) * the trajectory and final state at @xmath151 when ( b ) @xmath153 , ( c ) @xmath154 and @xmath148 , and , ( d ) @xmath155 and @xmath78 . *
( e ) * the combined variance of position and momentum .
all quantities are dimensionless . + * fig .
2 . ( a ) * poincar map for @xmath167 and the initial state ( @xmath147 ) . *
( b - d ) * the trajectory and final state at @xmath160 when ( b ) @xmath153 , ( c ) @xmath154 and @xmath148 , and , ( d ) @xmath168 and @xmath169 . *
( e ) * the combined variance of position and momentum . + * fig .
* the quantum ( black ) and classical ( grey ) trajectories for the same hamiltonian as in fig .
2 except with @xmath162 . *
( b)*the combined variance of position and momentum . | classical dynamics is formulated as a hamiltonian flow on phase space , while quantum mechanics is formulated as a unitary dynamics in hilbert space .
these different formulations have made it difficult to directly compare quantum and classical nonlinear dynamics .
previous solutions have focussed on computing quantities associated with a statistical ensemble such as variance or entropy .
however a more direct comparison would compare classical predictions to the quantum for continuous simultaneous measurement of position and momentum of a single system . in this paper
we give a theory of such measurement and show that chaotic behaviour in classical systems can be reproduced by continuously measured quantum systems . |
the studies of spectral distribution in nuclei during the last decade has made available the best information to date regarding correlations in nuclear systems @xcite .
experimentally , @xmath1 reactions have provided precise information for spectroscopic factors of states close to the fermi energy @xcite in finite nuclei .
moreover , information on the tail of the spectral distribution at high missing energies and momenta , generated by short - range correlations ( src ) , has recently become available @xcite .
the observed spectroscopic factors have been reproduced for selected nuclei , such as @xmath2 and @xmath3 @xcite , and a particularly successful comparison between theory and experiment has been obtained for @xmath4 @xcite . in all these calculations ,
the effects of configuration mixing at low energy was taken properly into account .
besides , the effects of src have been investigated for different systems and are known to give a constant reduction of the spectroscopic strength of 10 - 15% , for the fragments close to the fermi energy @xcite . for the case of @xmath0 , there still exists a substantial disagreement between experiment @xcite and theory @xcite .
the latter has been able to explain only a part of the quenching of spectroscopic factors for proton knockout from the @xmath5-shell .
the results of refs .
@xcite show that the reasons for this discrepancy have to be looked for in terms of long - range correlations ( lrc ) and in particular in the couplings of single - particle ( sp ) motion to low - energy collective excitations .
in this paper we report about the work done along this line in refs .
@xcite in order to approach the above issues for @xmath0 .
[ sec:2 ] describes the formalism employed to describe the coupling of particle - hole ( ph ) , particle - particle ( pp ) and hole - hole ( hh ) collective excitation to the sp propagator .
this is done by embedding the technique of faddeev equations into the self - consistent green s function formalism @xcite .
the results of this calculation are given in sec .
[ sec:3 ] , where the effects of nuclear fragmentations and the random phase approximation ( rpa ) in the description of phonons are discussed .
these calculations show that a proper description of the experimental spectral strength requires an improvement of the spectrum , that goes beyond the rpa approach . a first step in this direction
is reported in sec .
[ sec:4 ] , where the bethe - salpeter equation ( bse ) is employed to pursue the improvement of the ph - rpa approach .
the double lines represent a dressed propagator and the wavy lines correspond to g - matrix interactions .
the first term is the brueckner - hartree - fock potential while the others represent the 2p1h/2h1p or higher contributions that are approximated through the faddeev tda / rpa equations .
[ fig : selfenergy ] , height=172 ] we consider the calculation of the sp green s function @xmath6 from which both the one - hole and one - particle spectral functions , for the removal and addition of a nucleon , can be extracted . in eq .
( [ eq : g1 ] ) , @xmath7 ( @xmath8 ) are the spectroscopic amplitudes for the excited states of a system with @xmath9 ( @xmath10 ) particles and the poles @xmath11 ( @xmath12 ) correspond to the excitation energies with respect to the @xmath13-body ground state .
the one - body green s function can be computed by solving the dyson equation @xmath14 where the irreducible self - energy @xmath15 acts as an effective , energy - dependent , potential .
the latter can be expanded in a feynman - dyson series @xcite in terms the exact propagator @xmath16 , which itself is a solution of eq .
( [ eq : dys ] ) . in this expansion
, @xmath17 can be represented as the sum of a one - body hartree - fock potential and terms that describe the coupling between the sp motion and more complex excitations @xcite .
this separation is depicted by the diagrams of fig .
[ fig : selfenergy ] , where the last two diagrams are expressed in terms of the three - line irreducible propagator @xmath18 which describes the propagation of at least 2h1p or 2p1h at the same time .
it is at the level of @xmath18 that the correlations involving interactions between different collective modes have to be included .
the scgf approach consist of first solving the self - energy and the dyson eq .
( [ eq : dys ] ) in terms of a mean - field ipm propagator @xmath19 .
the ( dressed ) solution @xmath16 is then used to evaluate the an improved self - energy , which then contains the effects of fragmentation .
the whole procedure is iterated until self - consistency is reached .
baym and kadanoff showed that a self - consistent solution of the above equation guarantees the fulfillment of the principal conservation laws @xcite . in the following we are interested in describing the coupling of sp motion to ph and pp(hh ) collective excitations of the system .
all the relevant information regarding the latter are included in the lehmann representations of the polarization propagator @xmath20 that describes the excited states in the system with @xmath13 particles , and of the two - particle propagator that is relevant for the @xmath21 excitations @xmath22 in the calculation of sec .
[ sec:3 ] , these quantities have been evaluated by solving the dressed tda / rpa ( dtda / drpa ) equations @xcite for the ph and pp case , respectively .
the fact that a dressed sp propagator is used brings in to the calculation the effects of the strength distribution for particle and hole fragments .
one must notice that the propagators ( [ eq : pi ] ) and ( [ eq : g2 ] ) are in fact the solution of more general bethe - salpeter equations , that also include correlations beyond the rpa level .
one of these is the mixing of many - phonons in the kernel of the bse .
a first calculation in this direction has been done for the ph propagator @xmath23 @xcite and is reported in sec .
[ sec:4 ] .
once one has obtained the ph ( [ eq : pi ] ) and pp(hh ) ( [ eq : g2 ] ) propagators , their coupling to the sp motion has to be accounted for in the nuclear self - energy .
to do this the 2p1h and 2h1p propagators of fig .
[ fig : selfenergy ] are computed by solving a set of faddeev equations @xcite .
the details of the formalism employed in this calculation have been discussed in ref .
@xcite and will not be shown here .
for the present discussion it is sufficient to note that this approach computes the motion of three - quasiparticle excitations in the same way it is normally done to solve the three - body problem .
this allows to include the effects of ph and pp(hh ) motion not only at the tda level but also with more collective rpa phonons and beyond .
such excitations are coupled to each other by the faddeev equations , generating an infinite series of diagrams like the one shown in fig .
[ fig : faddsum ] .
moreover , pauli correlations are properly taken into account at the 2p1h/2h1p level .
in the calculations described below , the dyson equation was solved within a model space consisting of harmonic oscillator sp states .
an oscillator parameter @xmath24 1.76 fm was chosen ( corresponding to @xmath25 13.4 mev ) and all the first four major shells ( from @xmath26 to @xmath27 ) plus the @xmath28 where included . inside this model space ,
the interaction used was a brueckner g - matrix @xcite derived from the bonn - c potential @xcite .
the results of refs .
@xcite , suggest that this model space is large enough to properly account for the low - energy collective states in which we are mainly interested in , the src being accounted for through the g - matrix effective interaction .
we note that energy dependence of the g - matrix also contributes to the normalization of the spectral strength . in ref .
@xcite , it was shown that this accounts for the extra depletion of spectroscopic factors that is induced by src , at least for the normally occupied shells in the ipm .
we mention that the solution of rpa equations may lead to collective instabilities whenever the iteration becomes too attractive .
a particular sensitivity to the strength of the g - matrix interaction was found in the calculation of @xcite , only for the lowest isoscalar @xmath29 solution .
for this reason it was chosen to artificially shift its eigenvalue to the experimental energy of first excited state , 6.05 mev , during the successive iterations .
note that the screening due the redistribution of the sp strength tends to reduce the instability problem in drpa .
moreover , the origin of the instability for this particular solution has been identified and corrected later , in ref .
@xcite .
ccccccccccccc shell & & tda & & rpa & & 1st itr . &
& 2nd itr . & & 3rd itr . & & 4th itr .
+ @xmath30 & & 0.775 & & 0.745 & & 0.775 & & 0.777 & & 0.774 & & 0.776 + + @xmath31 & & 0.766 & & 0.725 & & 0.725 & & 0.727 & & 0.722 & & 0.724 + & & & & & & 0.015 & & 0.027 & & 0.026 & & 0.026 + + + total occ .
& & 14.56 & & 14.56 & & 14.56 & & 14.57 & & 14.58 & & 14.63 + for an ipm ansatz the tda calculation is equivalent to the one of ref .
@xcite and yields the same result for the spectroscopic factors of the main particle and hole shells close to the fermi energy .
these correspond to 0.775 for @xmath32 and 0.766 for @xmath33 and are reported in table [ tab : spectfact ] .
the introduction of rpa correlations reduces these values and brings them down to 0.745 and 0.725 , respectively .
this result reduces the discrepancy with the experiment by about 4% and shows that collectivity beyond the tda level is relevant to explain the quenching of spectroscopic factors .
since the present formalism does not account for center - of - mass effects , the above quantities need to be increased by about 7% before they are compared with the experiment @xcite . for @xmath0 for angular momentum @xmath34 ( left ) and @xmath35 ( right ) . for the positive parity states , the solid bars correspond to results for @xmath36 and @xmath37 orbitals , while the thick lines refer to @xmath38 .
the top panels show the experimental values taken from @xcite .
the mid panels give the theoretical results for the self - consitent spectral function .
the bottom panels show the results obtained by repeating the 3@xmath39 iteration with a modified ph - drpa spectrum , in which the lowest eigenstates have been shifted to the corresponding experimental values .
[ fig : p+sd_hsf ] , height=384 ] the rpa results were iterated a few times , with the aim of studying the effects of fragmentation on the rpa phonons and , subsequently , on the spectral strength .
since we are looking at low - energy excitations , it is sufficient to keep track only of the most occupied fragments that appear close to the fermi energy in the ( dressed ) sp propagator , eq .
( [ eq : g1 ] ) , while the residual strength is collected in a effective pole @xcite .
as shown in table [ tab : spectfact ] , only a few iterations are required to reach convergence . the converged distribution of one - hole strength is compared in with the experimental one in fig .
[ fig : p+sd_hsf ] . the main difference between these results and
the one obtained by using an ipm input is the appearance of a second smaller @xmath33 fragment at -26.3 mev .
this peak arises in the first two iterations and appears to become stable in the last one , with a spectroscopic factor of 2.6% .
this can be interpreted as a peak that describes the fragments seen experimentally at slightly higher missing energy .
this is the first time that such a fragment is obtained in calculations of the spectral strength .
further insight into the appearance of this strength is discussed in sec .
[ sec : hsf_vs_o16 ] .
a second effect of including fragmentation in the construction of the rpa phonons is to increase the strength of the main hole peaks .
the @xmath32 strength increases from the 0.745 obtained with ipm input to 0.776 , essentially canceling the improvement gained by the introduction of rpa correlations over the tda ones .
the main peak of the @xmath33 remains at 0.722 but the appearance of the secondary fragment slightly increases the overall strength at low energy as well .
this behavior is due to the fact that redistribution of the strength ( and principally , the reduction of the main spectroscopic factors ) tends to screen the nuclear interaction , with respect to the ipm case .
obviously , this makes the disagreement with experiments a little worse and additional work is needed to resolve the disagreement with the data .
nevertheless it is clear that fragmentation is a relevant feature of nuclear systems and that it has to be properly taken into account . together with the main fragments , the dyson equation produces also a large number of solutions with small spectroscopic factors .
this strength extends down to about -130 mev for the one - hole spectral function and up to about 100 mev for the one - particle case .
this background represents the strength that is removed from the main peaks and shifted up to medium missing energies .
another 10% of strength is moved to very high energies due to src @xcite .
its location can not be explicitly calculated in the present approach but the effects on the reduction of spectroscopic factors at low energy is accounted through the energy dependence of the g - matrix . accordingly
the total number of nucleons deduced from the fully self - consistent spectral function is about a 10% lower than @xmath40 16 , as reported in table [ tab : spectfact ] .
we note that the present approach gives predictions also for the strength distribution of quasiparticle excitations .
although little experimental information is available for the addition of a nucleon ( as well as for the removal of a neutron ) .
recently , knockout and nucleon transfer reactions , employing nuclear probes , have been applied to measure the hole spectroscopic factors of both stable and exotic nuclei @xcite .
these tools can now reach better accuracies than in the past and may , in principle , be used to extract information on the one - particle spectral distribution . a deeper insight into the mechanisms that generate the fragmentation pattern can be gained by investigating directly the connection between the spectral function and some specific collective states . to clarify this point we repeated the third iteration using exactly the same input but without keeping the troublesome ph @xmath29 state at its experimental energy , it was discarded instead .
the resulting @xmath5 hole spectral function is shown in the lower - left panel of fig .
[ fig : p+sd_hsf ] . in this calculation
no breaking of the @xmath33 strength is obtained but only a single peak is found with a spectroscopic factor equal to 0.75 , which is the sum of the two fragments that are obtained when the @xmath42 state is taken into account .
this result can be interpreted by considering the @xmath33 fragments as a hole in the ground state and in the first excited @xmath29 state of the @xmath0 core , respectively .
the correct fragmentation pattern can be reproduced only when latter two levels are close enough , in energy , to one another .
the other two low - lying states of @xmath0 that may be of some relevance are the isoscalar @xmath43 and @xmath41 .
these excitations are reproduced reasonably well by rpa type calculations @xcite but are typically found at @xmath443 mev above the experimental results .
the lower panels of fig .
[ fig : p+sd_hsf ] show the results for the even parity spectral functions that are obtained when both the @xmath41 and the @xmath43 ph - drpa solutions are shifted to match their experimental values . in this case , a @xmath36 hole peak is obtained at a missing energy -17.7 mev , in agreement with experiment , and with a spectroscopic strength of 0.5% .
it should be noted that many _ weak _ fragments ( i.e. relative to the addition or knockout for a nucleon with a small spectroscopic factor ) are seen throughout the nuclear table of isotopes .
most of these have no direct explanation in terms of mean - field sp states and contain more complicated components , involving 2p1h/2h1p configurations and beyond .
jennings and escher have shown that different definitions of the overlap wave function can be given that completely describe the sp motion inside the nuclear medium .
although , each of them sample the pauli correlations in a different way , resulting in very different normalizations of the amplitudes of weak states @xcite .
this may have important consequences for the reliability of spectroscopic factors extracted from proton emission and pick - up / stripping reactions @xcite .
the above results suggest that the main impediment for further improvements of the description of the experimental spectral strength is associated with the deficiencies of the rpa ( drpa ) description of the excited states .
one important problem is the appearance of at most one collective phonon for a given @xmath45 combination while several low - lying isoscalar @xmath29 and @xmath46 excited states are observed at low energy in @xmath0 , as well as additional @xmath41 and @xmath43 states .
a possible way to proceed would be to first concentrate on an improved description of the collective phonons by extending the rpa to explicitly include the coupling to two - particle@xmath47two - hole ( 2p2h ) states .
such an extended rpa procedure has been applied in heavier nuclei with considerable success @xcite . in order to be relevant for @xmath0 ,
this approach requires an extension in which the coherence of the 2p2h states is included in the form of the presence of two - phonon excitations @xcite .
this correspond to adding two - phonon configurations to the kernel of the bse for the ph propagator ( [ eq : pi ] ) , as shown in fig .
[ fig:2pierpa ] @xcite .
we note that there has been a tremendous progress in recent years in the microscopic description of _ p _ shell nuclei using green s function monte carlo and no - core shell model methods @xcite .
a possible application of the no - core shell model to @xmath48 would properly include the 4p4h effects relevant for describing the spectrum of @xmath0 @xcite .
however the description of spectroscopic factors will require the construction of effective operators to include the effects of short - range correlations on these quantities whereas they are automatically included in the scgf method .
whence the interest in studying the spectrum of @xmath0 with the present method .
the formalism to include two - phonon configurations has been presented in ref .
@xcite , where it is referred as two - phonon extended rpa ( erpa ) . in this work
the ph - drpa equation has been solved first , using the self - consistent sp propagator derived in sec .
[ sec:3 ] .
the lowest drpa solutions for both the @xmath29 , @xmath41 and @xmath43 channels were shifted down to their relative experimental energies and then they were employed to generate the two - phonon contributions for the erpa calculation .
the spectrum obtained for @xmath0 is displayed in fig .
[ fig:2pierpa_3 g ] and table [ tab:2pierpa_3 g ] together the total ph strength @xmath49 of each state , @xmath50 table [ tab:2pierpa_3 g ] also reposts the relative occupation of ph and two - phonon admixtures in the wave function for each solution of the erpa equation .
the results yield an isoscalar @xmath29 state with a predominant ph character at @xmath4417 mev , as it was found in drpa .
although this solution is now characterized by a partial mixing with two - phonon configurations .
table [ tab:2pierpa_3 g ] shows that this is the result of mixing with the lowest solution in the same channel , which ends up at @xmath4411 mev .
the latter is predominately a two - phonon state .
it is also seen that in both cases the relevant configuration comes from the coupling of two @xmath29 phonons themselves . the calculations of ref .
@xcite have shown that the bulk of the 4p4h contributions to the first excited state of @xmath48 may come from the coupling of four different phonons with negative parity ( @xmath41 and @xmath43 ) .
however , the self - consistent role of coupling positive parity states was not considered in that work .
both this effect , the rpa correlations and the inclusion of the nuclear fragmentation allow for the at least partial inclusion of configurations beyond the 2p2h case even when no more than two - phonon coupling is considered , as in this paper .
a study of the importance of three- and four - phonon configurations will be addressed in the future .
the low - lying negative parity isoscalar states ( @xmath41 and @xmath43 ) are only slightly affected by two - phonon contributions and remain substantially above the experimental energy at 9.23 and 10.90 mev , respectively , and more correlations will be needed in order to lower their energy .
we note that these wave functions contain a two - phonon admixture obtained by coupling the low - lying @xmath29 excitation to either the @xmath41 or @xmath43 phonons .
according to the above interpretation of the first @xmath29 excited state this corresponds to the inclusion of 3p3h and beyond . at the same time two additional negative parity solutions , with two - phonon character , are found at higher energy in accordance with experiment .
cccccccccccccccc @xmath51 & & & & & & @xmath52 & @xmath53 & @xmath54 & @xmath55 & + @xmath56 & & @xmath57 & @xmath49 & & @xmath57 & @xmath49 & & @xmath58(% ) & @xmath59(% ) & & ( % ) & ( % ) & ( % ) & ( % ) & + @xmath43 & & & & & 13.37 & 0.148 & & 21 & 79 & & & & & 79 & + @xmath41 & & & & & 12.35 & 0.113 & & 16 & 84 & & & & 84 + + @xmath29 & & & & & 12.15 & 0.001 & & 1 & 99 & & 3 & 96 + @xmath60 & & & & & 12.14 & 0.007 & & 1 & 99 & & & 99 + @xmath46 & & & & & 12.12 & 0.008 & & 1 & 99 & & & 98 + + @xmath29 & & 16.62 & 0.717 & & 17.21 & 0.633 & & 88 & 12 & & 10 & 0.5 + + @xmath29 & & & & & 11.28 & 0.092 & & 12 & 88 & & 85 & 2 + + @xmath43 & & 11.19 & 0.720 & & 10.90 & 0.680 & & 94.1 & 5.9 & & & & & 5.8 & + @xmath41 & & 9.50 & 0.762 & & 9.23 & 0.735 & & 95.9 & 4.1 & & & & 4.0 + the two - phonon erpa approach also generates a triplet of states at about 12 mev , with quantum numbers @xmath29 , @xmath46 and @xmath60 .
a similar triplet is found experimentally at 12.05 , 11.52 and 11.10 mev , which also corresponds to twice the experimental energy of the first @xmath41 phonon .
the erpa solutions for this triplet are almost exclusively obtained by @xmath61 states and therefore have a 2p2h character , in accordance with refs .
further interaction in the pp and hh channels will be needed in order to reproduce the correct experimental splitting and experimental strength @xcite .
in the present paper the faddeev technique , combined with self - consistent green s function theory , has been applied to study 2h1p correlations at small missing energies for the nucleus of @xmath0 .
the application of the faddeev method allows for the first time the treatment of the coupling of ph and pp(hh ) collective modes within a rpa framework , while the effects of fragmentation are included through the dressing of the sp propagator .
the present approach allows to further identify the important role played by the low - lying excited states of @xmath0 .
these are seen to be essential to generate many of the fragments with small spectrocopic factors that are seen experimentally , like the @xmath38 and @xmath33 states of @xmath62 at 5.20 mev and @xmath449 mev . .
the main impediment in obtaining a good theoretical description of the single particle spectral function of @xmath0 has been identified in the poor description of the excitation spectrum , as obtained by solving the standard ( d)rpa equations .
we have improved on this by computing the effects of mixing of ph states with two - phonon configurations in the kernel of the bethe - salpeter equations . moreover ,
configurations beyond the 2p2h level were also partially included by means of improved rpa and self - consistency effects .
the results show that these contributions explain the formation of several excited states observed at low energy which are not obtained by rpa calculations .
still it appears that a full solution of the spectrum of @xmath0 with this method requires to consider the mixing of up to four phonons and the interaction in the pp and hh channels @xcite
. the inclusion of such effects will be the topic of future work .
one of us ( c.b . ) acknowledges useful discussions with j .- m .
sparenberg and b. jennings .
this work was supported in part by the u.s .
national science foundation under grants no .
phy-9900713 and phy-0140316 and in part by the natural sciences and engineering research council of canada ( nserc ) .
m. g. e. brand , k. allaart , and w. h. dickhoff , phys .
* b214 * , 483 ( 1988 ) ; m. g. e. brand , k. allaart , and w. h. dickhoff , nucl .
a509 * , 1 ( 1990 ) .
m. g. e. brand , g. a. rijsdijk , f. a. muller , k. allaart , and w. h. dickhoff , , 253 ( 1991 ) ; g. a. rijsdijk , k. allaart , and w. h. dickhoff , , 159 ( 1993 ) c. barbieri and w. h. dickhoff , in proc .
fifth workshop on _ electromagnetically induced two - hadron emission _ ,
p. grabmayr _ at al .
lund , published on cd , isbn : 91 - 631 - 1612-x , 2001 ) , p. 108 .
p. navrtil , j. p. vary , and b. r. barrett , phys .
c 62 , 054311 ( 2000 ) ; p. navrtil and b. r. barrett , phys .
c 57 , 3119 - 3128 ( 1998 ) ; p.
navrtil and w. e. ormand , los alamos archive : nucl - th/0305090 .
g. e. brown , and a. m. green , nucl
. phys . * 75 * , 401 ( 1966 ) ; w. c. haxton , and c. j. johnson , phys .
lett . * 65 * , 1325 ( 1990 ) ; e. k. warburton , b. a. brown , and d. j. millener , phys . lett . * b293 * , 7 ( 1992 ) . | the faddeev technique is employed to study the influence of both particle - particle and particle - hole phonons on the one - hole spectral function of @xmath0 .
the formalism includes the effects of nuclear fragmentation and accounts for collective excitations at a random phase approximation ( rpa ) level .
the latter are subsequently summed to all orders by the faddeev equations to obtain the nucleon self - energy .
the results indicate that the characteristics of hole fragmentation are related to the low - lying states of @xmath0 and an improvement of the description of this spectrum is required to understand the details of the experimental strength distribution .
a first calculation to improve on this includes the mixing of two - phonon configuration with particle - hole ones .
it is seen that this allows to understand the formation of excited states which can not be described by the rpa approach . |
the presence of an antiferromagnetic ( af ) phase rapidly replaced by an anomalous metallic phase upon doping is one of the prominent features of the phase diagram of high temperature superconducting cuprates @xcite .
indeed it was early suggested @xcite that the proximity to an insulating magnetic phase together with the nearly twodimensional structure of these materials could be responsible for the superconducting and anomalous normal - state ( i.e. non - fermi liquid ) properties of these systems via the creation of a quantum - disordered spin liquid ( resonating - valence - bond state ) . since
then a great deal of attention has been devoted to the fascinating interplay between magnetism and charge degrees of freedom , leading to various theoretical proposals and to different scenarios for the metal - insulator transition at low doping .
one scenario is based on the tendency of a magnetically ordered phase to segregate the additional holes due to doping @xcite .
the tendency to phase separation is then frustrated by the coulombic repusion between the segregated carriers @xcite , thus leading to formation of hole - rich domain walls separating af domains in the form of stripe textures . within this scenario , pseudogaps naturally arise in the underdoped phase of the materials as a consequence of stripe fluctuations and local pair formation @xcite .
obviously this will substantially affect the transition to the af insulating phase at low doping @xcite .
the transition between a magnetically ordered phase and a metallic state can also be strongly affected by extrinsic ingredients like the disorder induced by dopant ions ( sr or zn in the case of @xmath8 ) .
this may lead to the formation of local random magnetic moments giving rise to spin - glass ordering between the af insulating and the paramagnetic metallic ( pm ) phases @xcite .
the stripe and the spin - glass scenarios clearly illustrate the complicated and rich nature that the af - pm transition may take . however , other proposals are more directly related to magnetism and to the contiguity between the af and the metallic phase .
these proposals start from the seminal work of ref .
@xcite showing that the materials with half - filled cuo@xmath9 planes are suitably described in terms of a twodimensional quantum heisenberg af model with very small interplanar coupling .
following work @xcite suggested that , although the 2d model displays long - range order at @xmath10 , upon doping the added charges enhance the quantum spin fluctuations thus driving the system into a disordered state even at zero temperature ( quantum disordered phase ) @xcite .
the occurrence of an af quantum critical point ( qcp ) provides a natural framework to intepret the scaling behavior at low temperatures and low frequencies of the of the @xmath11-integrated magnetic susceptibility @xmath12 .
the susceptibility @xmath13 can be experimentally determined via magnetic resonance measurements of the nuclear copper spin - echo decay rate , @xmath14 and spin - lattice relaxation rate , @xmath15 @xcite .
it is seen that in a range of temperatures @xmath16 , depending on doping , @xmath17 @xcite . assuming the scaling relations @xcite , @xmath18 and @xmath19 , with @xmath20 a dynamic critical exponent , @xmath21 an anomalous dimension , @xmath22 the antiferromagnetic ordering wave - vector , @xmath23 the correlation length , @xmath24 a scaling function
, one gets @xmath25 @xcite .
the constancy of this ratio is thus interpreted as the signature of a critical behavior corresponding to a @xmath3 value of the dynamic index @xcite .
the strong critical spin fluctuations occurring at the af - qcp have also been claimed to provide a possible source of pairing and normal - state anomalies . despite the apparent simplicity of a scenario involving a direct transition between an insulating af phase and a pm phase ,
such a transition is far from being trivial and is still an open problem .
the main difficulty is in the lack of a microscopic model being able to smoothly interpolate between an insulating af ordered phase with gapped charge excitation and a pm phase . on the one hand ,
the insulating af phase is suitably represented by the heisenberg model and , in particular , the long - wavelength and low frequency behavior of spin fluctuations in the cuo@xmath9 planes can be well described by the quantum non - linear sigma model @xcite , which at zero temperature and in two spatial dimensions is characterized by the action @xmath26 here @xmath27 is a three - component vector field subject to the real - space constraint @xmath28 which describes the local staggered magnetization , and @xmath29 is a coupling constant . it was first argued in @xcite that this description could be carried over to the case of the lightly doped compounds .
in fact , one can think that in this case the @xmath27 field continues giving an effective description of the electron spins localized on the cu sites of the lattice , while the presence of itinerating holes provides a finite renormalization of the coupling constant .
the theory described by the action ( [ pura ] ) , besides having the correct symmetries , realizes a zero - temperature transition for a given critical value @xmath30 of the coupling ( see @xcite for the quantum version and @xcite for the classical theory ) thus providing a good model for the physical picture of the af - qcp @xcite . from this point of view the @xmath3 value of
the dynamic scaling exponent is a simple consequence of the lorentz invariance of ( [ pura ] ) .
however , it is clear that the non - linear @xmath31 model does not properly describe the low - energy charge excitations , which are characteristic of a metallic state possibly providing a damping mechanism for spin - waves ( see , e.g. , case b of ref .
@xcite ) . therefore the model in eq .
[ pura ] provides a good description of a transition between an insulating ( charge - gapped ) af ( afi ) and an insulating paramagnet ( pi ) . , but it appears to be inappropriate when spin - waves can decay in low - energy particle - hole excitations . on the other hand , attempts have also been made to provide a microscopic basis for the transition to a pm phase . in this regard two distint cases a and b depending on the shape of the fermi surface in the metal were investigated in ref .
@xcite within a spin - density - wave approach . in case
a the af wavevector @xmath32 ( @xmath33 is the lattice spacing ) was not able to connect points of the fermi surface , so that momentum and energy conservation did not allow the decay of spin - waves in particle - hole pairs . in this case
the transition was naturally characterized by a dynamic critical index @xmath34 . in case b , instead , there were ( `` hot '' ) points on the fermi surface connected by @xmath35 so that low - energy particle - hole pairs could be created by the decay of low - energy spin - waves .
however , in both cases , a and b , the phase with af long - range order is metallic and therefore the model does not suitably describe a transition between an afi and a pm model would suitably describe the system also in this case despite the metallic character of both the ordered and disordered phases . ] . the case a above shares some common features with a mixed spin - fermion model without a three - body ( i.e. kondo - like ) direct coupling between the local moments and
the fermion spins @xcite .
this model has been shown @xcite to display a @xmath10 transition between a metallic ( fermi liquid ) phase with spin commensurate long - range order and a metallic ( also fermi liquid ) quantum disordered phase .
owing to the absence of the three body coupling responsible for the direct spin - wave decay into particle - hole pairs , also the qcp of this model is characterized by @xmath34 . to the best of our knowledge
no simple microscopic modelization is presently available to describe a direct afi - pm transition .
of course the possibilities remain open that other extrinsic mechanisms ( stripes , disorder , ... ) provide intermediate steps between the afi and the pm phases .
alternatively , one may also envisage that an afi to a pi phase transition occurs first at lower doping , followed at slightly larger doping by a pi - pm transition between two spin - disordered phases .
together with the basic difficulty outlined above , the additional question arises concerning the observed @xmath34 critical behavior in the underdoped phase of the cuprates and the way this can be reconciled with the metallic character of this phase .
in particular it is believed that the fermi surface of the cuprates in the metallic phase contains `` hot '' points connected by the af wavevector
. therefore it should be possible for the spin - waves of arbitrarily low energy to decay into particle - hole pairs and get overdamped .
this would lead to a relaxational behavior contrasting with the @xmath34 ( i.e. propagating ) critical behavior of the spin - waves .
moreover , by accepting a @xmath36 critical behavior in the metallic phase , it remains to clarify how this may turn into the natural propagating behavior , which characterizes the spin excitations in the ordered side of the qcp .
this latter difficulty was also encountered and stressed by the authors of ref .
@xcite for their case b : within their mean - field description a @xmath36 behavior was obtained both in the disordered and the ordered af ( metallic ) phases .
this suggested that the simple formulation b of the mean - field sdw model was only applicable on the disordered side of ( and not too close to ) the af - qcp . in the present paper
we will not attempt to find a microscopic foundation to a direct afi - pm transition , but we will instead start from a semiphenomenological point of view by introducing a field - theoretical generalization of the non - linear @xmath31 model action ( [ pura ] ) to investigate a ) the condition to be fullfilled by a simple ( @xmath37-independent ) damping term in order to maintain a @xmath34 critical behavior ; b ) how in the @xmath38 case an additional crossover energy must behave in the ordered phase to provide a scale separating damped spin excitations at high energy and propagating spin excitations at low energy . the resulting theory will depend on the precise form of the effective damping term .
we first discuss the simpler case of the metallic phase on the @xmath39 side of the qcp . within a microscopic model of spins coupled to free itinerant holes , integrating out the holes degrees of freedom in the random phase approximation
gives a damping term of the form @xmath40 ( see _ e.g. _ @xcite ) .
however , the choice of the adequate low frequency form of @xmath41 should also take in consideration the fact that in underdoped cuprates a significant loss of low - frequency spectral weight is observed at low temperatures in the distribution of quasi - particles , a phenomenon often referred to as `` pseudogap '' ( see _ e.g. _ @xcite ) . the origin of this effect is a debated issue , and has been alternatively interpreted : as the result of spin - fermion scattering @xcite , as due to the intrinsecally non - fermi liquid nature of the system ( luttinger liquid ) @xcite , as a signal of the formation of preformed pairs for @xmath42 @xcite , as due to stripe charge fluctuations @xcite . in all cases ,
a precise computation of these effects implies solving a difficult self - consistency problem going well beyond the rpa approach .
we shall try here a rough , phenomenological approach , assuming that the damping coefficient @xmath43 can acquire at low frequencies an @xmath44-dependence @xmath45 , thus giving rise to an effective damping term of the form @xmath1 , with @xmath46 case corresponds to the case considered in ref .
@xcite within a large - n framework . ] .
_ _ _ _ once chosen a damping term @xmath47 we perform a renormalization group ( rg ) analysis of the resulting theory in the proximity of the qcp using the momentum - shell method and @xmath5-expansion .
we observe the existence of two fixed points , which we shall denote fp1 and fp2 , showing respectively a @xmath3 , undamped spin - wave behavior and a @xmath48 , dissipative behavior .
the point fp1 corresponds to the zero - temperature critical point of a pure quantum non - linear sigma model without any damping term , but it becomes unstable in a wide region of the parameters with respect to the dissipative fixed point fp2 .
there exists a range of values @xmath49 for which one obtains a @xmath50 dynamic scaling exponent in the proximity of the stable quantum critical point , even in the presence of a damping term . in other words , there are `` soft '' damping terms which do not destroy the @xmath3 , spin - wave behavior . here
@xmath51 is the dimension dependent lower bound for the exponents in the soft damping term , which we determine to be @xmath52 , where @xmath53 is the anomalous field dimension and @xmath54 is its value at the fp1 . it is worth noting that this result bears a resemblance to what is found in classical models with long - range interactions @xcite .
in particular the scaling exponent of the damping stems from the non - analytic form of @xmath47 , which is not reconstructed by the rg corrections ( for a discussion of this point in classical models see ref .
. therefore @xmath55 does not acquire singular corrections but for the anomalous field dimensions .
within the framework of an expansion in @xmath56 , where @xmath57 is the spatial dimension of the system and @xmath58 is the bare value of the dynamic scaling exponent in the vicinity of the stable fixed point , one would get @xmath59 , where the 1 or 2 subscripts in @xmath60 refer to fp1 or fp2 respectively . however , the extrapolation of results valid for @xmath61 to the physical region @xmath7 should be done with great care . in particular , using the numerical results for the critical exponents of the @xmath62 model in 2 + 1 dimensions , it appears that @xmath51 is very near to two and the range of `` soft '' @xmath63 values should be a very narrow one .
this means that within the here - considered model the appearance of a @xmath3 critical behavior in the cuprates can not be accounted for in the effective framework without assuming a value @xmath64 , _
i.e. _ , an almost linear decay of the quasi - particle spectral weight for low frequencies .
this spectral weight suppression is a rather severe condition to be obtained from the magnetic scattering itself , as also suggested by a direct perturbative calculation within a spin - fermion model @xcite , where density of states and vertex corrections to the fermion bubble produced only a minor change of the power - law dependence of the damping term .
it is worth noting that a substantial spectral weight reduction is however consistent with anderson s idea that a smooth connection between a charge - gapped insulator and a fermi - liquid metal is hardly conceivable . _
_ as far as the ordered phase is concerned , ( point ( b ) above ) it is necessary to take into account the presence of a non - zero staggered magnetization @xmath65 , which is accompanied by the presence of two quasi - particle bands separated by an energy @xmath66 .
this suggests @xcite to introduce a damping term having two distinct asymptotics : @xmath67 for @xmath68 and @xmath69 for @xmath70 , with @xmath71 in the ordered ( as well as in the disordered ) phase , but in a first approximation we ignore this dependence mainly for simplicity reasons , but also because in the insulating phase , where a gap for the charge excitations providing spin - wave decay is present , it would be unnatural to expect a relevant q - dependence . ]
the difficult point here is that the scale @xmath72 separating damped and undamped spin excitations vanishes at the transition .
this prevents using a conventional rg approach , and we will resort to simple scaling arguments and phenomenological assumptions . our analysis does not attain the same degree of reliability of a traditional landau - ginzburg - wilson theory and has a heuristic character .
we shall assume that @xmath73 has a power - law behavior , to be determined by matching the hydrodynamic behaviors on the two sides of the quantum transition .
for @xmath74 , _
i.e._for @xmath38 , we find that ( i ) @xmath73 should close at the microscopic " scale as @xmath75 for @xmath76 , where @xmath77 is a correlation length , and that ( ii ) the spin - wave velocity @xmath78 vanishes as @xmath79 , to be consistent with the presence of damping . _
_ the paper is organized as follows : in sec .
[ rengreq ] we derive the rg equations for the zero - temperature quantum non - linear sigma model with a damping term of generic form , following the scheme exposed in @xcite .
this part is rather technical and it can be skipped in a first reading without loosing the general meaning of the paper . in sec .
[ fixpanrgeq ] we perform the fixed point analysis of the rg equations for a damping term @xmath80 . here
we show the existence of the two fixed points mentioned above , study their respective regions of stability and compute the relevant critical exponents , with particular regard to the value of the dynamic scaling exponent @xmath81 . in sec .
[ ordphase ] we discuss damping in the ordered phase , introducing a phenomenological scale @xmath73 vanishing at the quantum critical point with a critical exponent that we determine by matching the hydrodynamic behaviors on the two sides of the transition . we also compute the critical behavior of the spin - wave velocity @xmath78 in the presence of damping . in sec .
[ conclusions ] we present some conclusive remarks and discuss the connection between our model and other physically interesting systems , like macroscopic quantum tunnelling and spin chains with long range interaction .
let us consider the euclidean action of the quantum non - linear sigma model @xcite at zero temperature in @xmath82 spatial dimensions , with a damping term @xmath41 of generic form and in the presence of a constant magnetic field @xmath83 : @xmath84 here @xmath27 is an @xmath85-component field satisfying the constraint @xmath86 in real space , @xmath87 , @xmath88 .
the magnetic field is oriented along the @xmath85-th direction .
the momentum and frequency integrations have been rescaled so that the adimensional spin velocity @xmath89 has bare value equal to one and the momentum integration is cut off at @xmath90 . with our rescaling , in ( [ azione1 ] )
the dimensionless coupling constant @xmath91 is related to the bare ( dimensional ) spin - wave velocity @xmath92 , the physical cut - off @xmath93 and the bare spin - stiffness constant @xmath94 by the relation @xmath95 @xcite . in the large @xmath96 (
here @xmath97 is the spin ) limit @xmath94 and @xmath92 are related to the coupling constant of the heisenberg model @xmath98 and the lattice spacing @xmath99 by @xmath100 and @xmath101 @xcite with @xmath102 .
@xmath103 is also adimensional .
action ( [ azione1 ] ) is formally equivalent to the hamiltonian of an anisotropic corresponding statistical mechanical model in @xmath104 dimensions .
the critical point @xmath105 of the classical model is here reinterpreted as a zero - temperature quantum critical point , separating the ordered and the disordered ground state of the system , which are realized , respectively , for @xmath106 and @xmath107 . in order to study the long - distance , low - frequency behavior of the @xmath27 field in the proximity of the quantum critical point we shall use the renormalization group and the @xmath5-expansion , following refs .
@xcite . in the broken symmetry phase we let @xmath108 , @xmath109 .
keeping only terms up to @xmath110 , action ( [ azione1 ] ) takes the form @xmath111 where the inverse propagator , given by @xmath112 depends on the damping term @xmath47 , and the measure term @xmath113 is @xmath114 in ( [ azione2 ] ) the term proportional to the extended phase - space volume @xmath115 comes from the perturbative evaluation of the jacobian @xmath116 in ( [ rsd ] ) @xmath117 is an arbitrary frequency cut - off which does not enter the final result , and @xmath118 .
let us denote the vertices appearing ( in the same order ) in ( [ azione2 ] ) as where the cut , double cut and triangle indicate respectively a @xmath119 , @xmath44 and @xmath41 insertion , and the dotted line stays for the first perturbative term of the jacobian .
let then @xmath120 , with @xmath121 we shall use the momentum - shell method to derive recursion relations _ la _
wilson @xcite , integrating out the @xmath122 with @xmath123 , @xmath124 and all @xmath44 , and rescaling @xmath125 hereby introducing a dynamic scaling exponent @xmath81 and a wave function renormalization @xmath126 .
we average with respect to the following propagator : @xmath127 the first perturbative term of the jacobian can be written as @xmath128 \int _ { k<1}{\,\mathrm{d}}\kappa \ , { \boldsymbol{\pi } } _ { { \boldsymbol{\kappa } } } \cdot { \boldsymbol{\pi } } _ { -{\boldsymbol{\kappa } } } \ ] ] and we will indicate it pictorially as _ _ after integration on the momentum shell the first term represents the modified first - loop jacobian contribution , while the second has to be absorbed in the mass renormalization and cancels the contribution from the second vertex in eq .
( [ azione2 ] ) ( see below ) , thus maintaining the theory massless at @xmath129 .
the usual diagrams renormalizing the @xmath130 and @xmath131 terms both give a @xmath132 contribution , with @xmath133 in order to renormalize @xmath134 and @xmath83 we have to consider the following contributions : @xmath135 { \boldsymbol{\pi } } _ { { \boldsymbol{\kappa } } _ { 1}}\cdot { \boldsymbol{\pi } } _ { -{\boldsymbol{\kappa } } _ { 1}}\\ \nonumber & \simeq & \frac{1}{2 } \int _ { k_{1}<e^{-l}}{\,\mathrm{d}}\kappa _ { 1}\left [ \int _ { e^{-l}<k_{2}<1}\!\!\frac{f(\omega _ { 1}+\omega _ { 2 } ) -f(\omega _ { 2})}{\delta _ { ( f)}(k_{2},\omega _ { 2})}{\,\mathrm{d}}\kappa _ { 2}+ \frac{n-1}{2}ghli_{(f ) } \right ] { \boldsymbol{\pi } } _ { { \boldsymbol{\kappa } } _ { 1}}\cdot { \boldsymbol{\pi } } _ { -{\boldsymbol{\kappa } } _ { 1 } } \end{aligned}\ ] ] the symmetry of ( [ azione1 ] ) requires that @xmath83 scales as @xmath136 .
since @xmath83 plays in ( [ azione2 ] ) the role of a mass , from one - loop mass renormalization we get that the field rescaling factor is @xmath137\ ] ] defining @xmath138 we finally get the rg equations : @xmath139 g\\ \label{rhatf } \frac{{\,\mathrm{d}}\hat f } { { \,\mathrm{d}}l}(\omega ) & = & \left ( 2-gi_{(f)}-z\omega \frac{{\,\mathrm{d}}}{{\,\mathrm{d}}\omega } \right ) \hat f ( \omega ) + gc_{(f)}(\omega ) \end{aligned}\ ] ] with @xmath140 equation ( [ rtildeg ] ) results as usual from field rescaling and renormalization of the @xmath130 term .
the term @xmath141 in eq .
( [ rhatf ] ) describes the frequency rescaling and @xmath142 is related to the contributions in eq .
( [ diagrammi ] ) .
let us now consider a damping term of the form @xmath143 , with @xmath144 a generic differentiable function behaving as @xmath145 for @xmath146 .
one observes that for @xmath147 and @xmath148 integrable in a neighborhood of @xmath149 the term @xmath150 is twice continuously differentiable , so that for @xmath151 it is possible to expand @xmath152 with @xmath153\ ] ] in order to prove ( [ sviluppo],[kappaf ] ) it is enough to rewrite ( [ conv0 ] ) as @xmath154 and to differentiate the expression twice with respect to @xmath155 .
@xmath156 is continuous but not differentiable in @xmath149 .
the content of eq .
( [ sviluppo ] ) is that _ the non analyticity of the damping term in @xmath149 is not reproduced under renormalization__. this implies that the @xmath41 term renormalizes only according to its bare dimensions and to the contribution of the field rescaling factor . on the other hand
, @xmath47 contributes to the renormalization of the @xmath131 term .
_ _ _ according to eq .
( [ kappaf ] ) , @xmath157 is essentially a measure of the non linearity of @xmath134 and is not strongly dependent on its explicit form . for @xmath80 , with @xmath2 and @xmath43 a damping coefficient , equation ( [ rhatf ] ) , together with ( [ sviluppo],[kappaf ] ) ,
gives @xmath158\gamma \\
\label{grt3 } \frac{{\,\mathrm{d}}}{{\,\mathrm{d}}l}\frac{1}{c^{2 } } & = & 2(1-z)\frac{1}{c^{2}}+ 2gk_{\alpha } ( c,\gamma ) \end{aligned}\ ] ] with _
@xmath159 _ in appendix [ compcritind ] the limit for @xmath160 of these expression is discussed ( see ( [ loop2],[conv2 ] ) ) . _ _
we shall now study the rg equations obtained in the previous section for the zero - temperature critical point of the quantum non - linear sigma model with a damping term @xmath161 , described by the action ( [ azione1 ] ) .
for the sake of clarity we rewrite here the rg equations ( [ rtildeg],[grt2],[grt3 ] ) : @xmath162g\\ \label{grt2 _ } \frac{{\,\mathrm{d}}\gamma } { { \,\mathrm{d}}l } & = & [ 2-\alpha z- gi_{\alpha } ( c,\gamma ) ] \gamma \\
\label{grt3 _ } \frac{{\,\mathrm{d}}}{{\,\mathrm{d}}l}\frac{1}{c^{2 } } & = & 2(1-z ) \frac{1}{c^{2}}+2gk_{\alpha } ( c,\gamma ) \end{aligned}\ ] ] where the integrals @xmath163 and @xmath164 were defined in ( [ loop],[conv ] ) .
the dynamic scaling exponent @xmath81 was introduced when choosing to rescale frequencies as @xmath165 , @xmath166 being the parameter of the renormalization group .
we look for fixed points of the rg transformation ( [ grt1_]-[grt3 _ ] ) .
one immediately finds two fixed points , characterized by 00.00 @xmath167 , for @xmath168 ; @xmath169i_{\alpha } ( c _ { * } , \gamma _ { * } ) } } } , \ , \ , \
, c^{\alpha } _ { * } \gamma _ { * } = u_{*},\ , \ , \ , z_{2}={{\displaystyle \frac{2n-2-d}{(n-2)\alpha + 1 } } } $ ] , + for @xmath170 ; where @xmath171 is the solution of @xmath172 with @xmath173 , @xmath174 .
strictly speaking , fp1 and fp2 are two _ lines _ of fixed points , since in both cases @xmath175 is arbitrary . in the case of fp2
either @xmath175 or @xmath176 can be arbitrarily chosen , and the remaining one is then determined through the constraint @xmath177 .
_ _ concerning eq .
( [ vincolo ] ) , note that , letting @xmath178 : 1 . a unique solution @xmath171 exists for @xmath179 ; 2 .
@xmath180 for @xmath181 ; 3 .
no solution exists for @xmath182 .
together with @xmath183 , this implies @xmath184 .
these are all consequences of the following properties : @xmath185 is a monotonically decreasing function of @xmath186 , tending to a finite value for @xmath187 and vanishing for @xmath188 ; @xmath189 is an increasing function , vanishing for @xmath187 and unbounded for @xmath188 .
the analogy with the statistical mechanical model shows that the perturbative expansion is in the parameter @xmath56 , where @xmath58 is the value of the dynamic scaling exponent obtained from purely dimensional considerations . for the @xmath190 case lorentz invariance imposes @xmath191 , while in the presence of a non - zero damping term @xmath192 with @xmath63 not too close to @xmath193 we expect the latter to be more relevant than @xmath131 , and @xmath194 .
letting then @xmath195 be the one - loop perturbative correction to the bare dynamic exponent @xmath58 and @xmath21 the field anomalous dimension , one finds in the two cases : 00.00 @xmath196 , @xmath197 , @xmath198 , @xmath199 , @xmath200 @xmath201 , @xmath202 , @xmath203 , + @xmath204 , @xmath205 , + with @xmath206 the solution of @xmath207 . the point fp2 exists for @xmath208 . for @xmath209 one gets the unphysical situation @xmath210 .
the additional condition @xmath211 amounts to @xmath212 , which is the condition of stability to order @xmath213 of fp2 with respect to fp1 .
more precisely , at @xmath214 fp2 merges into fp1 ( see below ) . for @xmath215 one
has @xmath216 ; keeping @xmath176 fixed this gives @xmath217 , as expected , since for fp2 one has @xmath218 at 0-th order in @xmath219 . for @xmath220 , ( [ loop ] ) shows that @xmath221 is finite , so that @xmath222 .
the same is true in the @xmath223 case , but some care must be exerted .
a direct computation using ( [ loop2],[conv2 ] ) leads to ambiguous results , since in this case the frequency wave - vector integration shell becomes singular : this is discussed more at length in appendix [ compcritind ] .
we just observe here that for @xmath160 , @xmath215 implies @xmath224 and one has @xmath225 .
one can send @xmath215 either with fixed @xmath226 , or with fixed @xmath227 , and in both cases the loop integral @xmath228 is finite at @xmath229 . in order to compute the critical exponent @xmath230 we linearize the rg equations ( [ grt1_]-[grt3 _ ] ) and find the eigenvalues of the linearized transformation ( see appendix [ compcritind ] for more detail ) : 00.00 @xmath231 , @xmath232 , @xmath233 for @xmath168 ; @xmath234 , @xmath232 , @xmath235 + for @xmath236 .
this shows that the point fp2 always has one attractive direction in its region of existence , while @xmath237 depends on @xmath238 .
( 0,0 ) ( -26,55)(-61,30 ) ( -57,88)(0,0)[bl]@xmath239 ( -43,88)(0,0)[bl]@xmath240 ( -26,88)(0,0)[bl]@xmath241 ( -73,54)(0,0)[bl]@xmath242 ( -65,48)(0,0)[bl]@xmath243 ( -56,42)(0,0)[bl]@xmath244 ( -20,25)(0,0)[bl]@xmath245 ( -38,18)(0,0)[bl]@xmath246 ( -50,12)(0,0)[bl]@xmath247 ( -95,46)(0,0)@xmath248 ( -45,-4)(0,0)@xmath0 we can now assess the stability of the two fixed points . in the @xmath249 plane
we distinguish the regions ( see fig .
[ stability ] for the @xmath250 heisenberg case ) : 00.00 @xmath251 ; @xmath252 i.e. @xmath253 . in region fp1 there exists only the corresponding fixed point , which is stable . in region
fp2 there appears a second , stable fixed point , while the point fp1 becomes unstable .
the crossing between the regions fp1 and fp2 takes place for @xmath254 . notice that on the border line one gets @xmath255 and @xmath256 , so that when moving from the region fp2 to fp1 the critical exponents @xmath21 and @xmath81 vary _ continuously__. _ _ _ finally we report in fig .
[ rgflow ] a numerical computation of the structure of the rg flow for typical values of the parameters . to this purpose
it is convenient to introduce the rescaled variables @xmath257 , @xmath258 , so that ( [ grt1_]-[grt3 _ ] ) take the form @xmath259\}\tilde g \\ \label{gra2 } \frac{{\,\mathrm{d}}\tilde\gamma } { { \,\mathrm{d}}l } & = & [ 2-\alpha -\tilde g ( \tilde i _ { \alpha } ( \tilde\gamma ) + \alpha \tilde k _ { \alpha } ( \tilde\gamma ) ) ] \tilde\gamma \\
\label{gra3 } \frac{{\,\mathrm{d}}}{{\,\mathrm{d}}l}\frac{1}{c^{2 } } & = & 2 [ 1-z+ \tilde g \tilde
k _ { \alpha } ( \tilde\gamma ) ] \frac{1}{c^{2 } } \end{aligned}\ ] ] where ( [ gra1 ] ) and ( [ gra2 ] ) are decoupled from ( [ gra3 ] ) and do not contain the dynamic exponent @xmath81 .
this means that in the new variables the rg trajectories project on the @xmath260 plane and the two lines of fixed points project on two corresponding fixed points .
this fact was not _ a priori _ obvious and greatly facilitates the analytical and qualitative study of the rg trajectories .
_ _ ( 0,0)(-97,45)(0,0)@xmath261(-45,-4)(0,0)@xmath262(-70,9 ) ( -70,9)(-2,-1)5(-58,66 ) ( -58,66)(-2,-1)5 we notice the interesting fact that there exists a range of values @xmath263 for which one obtains a @xmath50 dynamic scaling exponent in the proximity of the quantum critical point , even in the presence of a damping term .
in other words , there can be `` soft '' damping terms which do not spoil the spin - wave behavior of the fixed point . within the @xmath60-expansion for @xmath250 @xmath264
is given by @xmath265 ( cf .
fig . [ stability ] ) .
we shall argue about the extension of these results to the @xmath266 case in the next subsection .
the above analysis of the stability of fp1 vs. fp2 bears a strong similarity with the study of the @xmath267 theory with long range interactions @xcite . for small @xmath268 and @xmath219 one
has @xmath269 and @xmath270 and therefore @xmath271 , coherently with @xcite .
the two fixed points studied in @xcite correspond to our fp1 and fp2 ; in our approach they are selected by the value of @xmath58__. _ the continuity of @xmath21 for @xmath272 appears as a consequence of their relative stability , as in @xcite .
the computations here are considerably simpler , since first - loop calculations are sufficient to establish the general picture .
another difference is in the presence of the dynamic exponent @xmath81 , which was absent in the classical works @xcite . _ _ _ it is worth stressing the fact , noticed after eq .
( [ sviluppo ] ) , that the non - analyticity of the term @xmath273 is not reproduced under renormalization , at least at first loop . as a consequence @xmath274
does not get singular corrections by itself , but only scales according to its bare dimension and to the anomalous field dimension ) below refer to the coupling @xmath275 of the dissipative term in eq .
( [ azione1 ] ) . in the critical region @xmath105 and the behavior of @xmath55
coincides with the behavior of @xmath275 . ] .
the indication coming from @xcite is that this will be true also at higher orders .
this would imply that , as an exact result , @xmath43 scales in the critical region according to the exponent @xmath276 irrespectively from the approximation .
the value @xmath277 can then be obtained directly from the condition @xmath278 , giving @xmath279 which is in agreement with our one - loop analysis and is expected to remain valid at higher loop orders as well .
this exact relation can be used to settle the problem of the crossover from fp1 to fp2 .
this crossover takes place when @xmath256 , that is , using ( [ zeta ] ) , when @xmath280 , where @xmath281 is the anomalous field dimension for a given @xmath248 .
continuity of @xmath282 through the crossover implies @xmath283 with @xmath284 being the critical exponent of the sigma model with zero damping in @xmath104 . for twodimensional systems , like the cuprates , @xmath285 .
numerical estimates for the critical indices of the @xmath286 symmetric @xmath287 field theory obtained from summed perturbation series at fixed dimension @xmath288 ( see @xcite and table 25.4 of @xcite ) give @xmath289 and @xmath290 .
in this respect the @xmath291-expansion greatly overestimates @xmath292 and underestimates @xmath293 in @xmath266 ( see fig .
1 ) . this suggests that , as a consequence of this small value of @xmath292 , since the relation ( [ zeta ] ) is probably an exact relation , for @xmath294 near one the physically relevant line of fixed point would be fp2 .
it appears therefore that the @xmath3 behavior observed in the cuprates can be accounted for in this framework only in the hypothesis that some mechanism like the reduction of the spectral weight of low - frequency quasi - particle is so strong to effectively produce a @xmath273 damping term with @xmath63 very close to @xmath193 .
in the ordered phase the effect of holes in the small - doping limit of the cuprates acts as a perturbation of the @xmath10 antiferromagnetically ordered ground state .
the description of the interplay of spin and fermion degrees of freedom when approaching the qcp from the ordered side presents major difficulties .
the simple introduction in the effective sigma model of a @xmath295 term is not satisfactory , since this term would be relevant also in the proximity of the ordered critical point @xmath296 , and would completely destroy the spin - wave picture of the ordered ground state . in the ordered phase the appearance of a non - zero staggered magnetization @xmath297 is accompanied by the presence of two quasi - particle bands , separated by an energy @xmath73 ( for a mean field analysis see _
e.g. _ ref .
@xcite ) . as a consequence
, one would expect the decay of spin - wave excitations in particle - hole pairs to be disfavored for @xmath298 , and an effective damping be felt only for @xmath299 .
moreover , in an afi - pm transition @xmath73 should vanish at the qcp as well as @xmath297 does .
were this not the case , since @xmath300 in the disordered phase one could hardly match the critical behavior on the two sides of the transition .
we are not in a condition to directly link the scale @xmath73 to the staggered magnetization @xmath297 since the precise relation , if any , should emerge from the unknown microscopic dynamics of the quasiparticles . in this section
we try to get some insight in this difficult problem using simple scaling arguments .
we will conjecture for @xmath73 a scaling law with a critical exponent @xmath301 , computable in principle from the full theory including also the fermionic degrees of freedom , and we shall try to assess the value of @xmath301 in a self - consistent way , once the value of @xmath248 for the damping in the disordered phase is given . we emphasize that this procedure is not as reliable as the rg computation exposed in the preceding sections and has an heuristic character . _ _ a model damping term which could take into account the appearance of the scale @xmath73 is @xmath302 with @xmath71 , and @xmath303 .
we shall later assume that @xmath73 vanishes at the qcp , so that the crossover between the two regimes is continuously shifted while approaching the transition .
here we consider @xmath304 as a free parameter at the same level as @xmath305 and @xmath55 . assuming that changing the scale , @xmath47 approximately maintains the form ( [ model1 ] ) with varying values of @xmath305 and @xmath55 one finds from ( [ rhatf ] ) the characteristic exponents @xmath306 our phenomenological assumptions suggest that damping be relevant only at high frequencies . in this regime
the spin - wave velocity @xmath78 goes as @xmath307 with the running coupling constant @xmath308 and the dynamic exponent @xmath20 . in the neighborhood of a critical point , @xmath20 is determined by the condition that @xmath308 does not renormalize .
this corresponds to attributing all the renormalization of the @xmath309 term to the anomalous frequency scaling , @xmath310 .
the running coupling constant @xmath308 differs from the physical spin - wave velocity @xmath78 , which is measured at any scale @xmath311 using a fixed system of physical units : one has @xmath312 .
so , although @xmath308 does not renormalize by construction , when @xmath38 the physical velocity @xmath78 vanishes at the transition . in the case of the undamped sigma model the value @xmath34 is obtained as a consequence of lorentz invariance , so that there is no difference between @xmath308 and @xmath78 and the latter remains finite at the qcp .
however , the damping term breaks lorentz invariance and can drive @xmath78 to zero at the qcp . ] . on the other hand , at any given scale the relevant @xmath155-modes are characterized by the condition @xmath313 .
so , damping is relevant up to the scale @xmath314 where @xmath73 becomes of order @xmath315 . beyond this scale
the gap for spin - wave decay is visible and @xmath47 is irrelevant . if @xmath73 at the microscopic " scale @xmath316 had a finite value , eventually @xmath47 would be always irrelevant . the simplest way to take into account the physical condition that @xmath73 should vanish approaching
the transition is to assume that @xmath317 where @xmath318 and @xmath319 .
this condition encodes in a simplified , phenomenological way the complicated interaction between damping and antiferromagnetic order , mediated in the full theory by the fermionic degrees of freedom . taking into account ( [ initcond ] ) and
the scaling exponents ( [ gammarcritexp ] ) we get @xmath320 approximately at this scale damping becomes irrelevant and the critical behavior starts being characterized by @xmath34 .
this gives @xmath78 vanishing as @xmath321 for @xmath294 not too close to two ( specifically for @xmath322 ) by approaching the quantum transition from the disordered phase one observes , as already discussed , a dissipative behavior dominated by the fixed point fp2 , with the scaling law @xmath323 . deep into the broken symmetry phase one instead
has a propagating behavior , @xmath324 .
this should match the critical behavior at the scale @xmath325 , with @xmath326 the corresponding critical index .
continuity of the critical behavior in a neighborhood of the critical point imposes then @xmath327 , _
i.e. _ @xmath328 this is compatible with ( [ cswsimbars ] ) only for @xmath329 in terms of @xmath330 this condition on the critical behavior of the energy separation @xmath73 gives @xmath331 .
it is not unreasonable to expect that the damped sigma model we considered contains some of the physics of the quantum critical point of the underdoped layered cuprates . for this reason we performed a rg analysis directly on this model , trying to maintain sufficient generality in the form of the effective damping term , in particular introducing an exponent @xmath332 that could account for the observed pseudogap behavior . in the broken symmetry phase we tried to get some insight on the interplay between damping and af ordering using simple scaling arguments . in both
the ordered and the disordered phases , assuming continuity of the hydrodynamic behaviors at the quantum transition , we obtained that for @xmath74 ( where @xmath333 is expected to be very close to 2 ) it is not possible to account for the observed @xmath34 scaling law .
it was observed in ref .
@xcite that the @xmath34 scaling can be recovered at higher temperatures , when typical frequencies would become larger than the @xmath10 value of the damping term . however , experimentally it is the @xmath34 behavior which is observed at lower temperatures , while a relaxational @xmath36 behavior is observed at higher @xmath334 @xcite .
it was argued @xcite that this could be due to the neglected @xmath334-dependence of the damping term @xmath55 .
we propose in this paper that the same effect can be accounted for by assuming that the decay of the quasi - particle spectral weight for low frequencies produces a soft " damping term @xmath335 with @xmath336 .
the @xmath34 scaling would then be a consequence of an almost linear decay of the quasi - particle spectral weight .
we notice that this reduced weight is fully consistent with the observed occurrence of a pseudogap below a crossover temperature @xmath337 in the underdoped region of the phase diagram for the cuprates .
the origin of this pseudogap might arise from extrinsic mechanisms like local cooper - pair formation and/or stripe fluctuations .
alternatively a pseudogap could be the very consequence of the intrinsecally non fermi liquid nature of the system @xcite or of strong critical fluctuations occurring near the qcp .
in this last regard singular af fluctuations have been shown within a spin - fermion model to strongly modify the quasiparticles properties @xcite in the proximity of the `` hot''points of the fermi surface separated by the af wavevector .
since the particle - hole excitations around these points are responsible for the low - energy spin - wave decay , this naturally affects the damping of spin excitations , in a complicated interplay which deserves further investigation .
we notice that the fluctuation - induced spectral weight reduction for the fermionic excitations is a definite signature of a non - fermi liquid metallic state .
this is in agreement with general arguments put forward by anderson ( see _ e.g. _ , ref .
@xcite ) on the difficulty in continuously connecting a charge - gapped insulating phase with a normal fermi - liquid phase .
we remark , however , that anderson s arguments extend the presence of a non - fermi - liquid phase all over the metallic state as a consequence of low dimensionality . on the other hand , within the qcp scenario ,
the violation of the fermi - liquid properties is a consequence of critical fluctuations , which are only present around the critical point .
as soon as one moves away from this point , the fluctuations loose their critical character thus providing simple perturbative corrections to the fermi liquid state .
it is worth noting that the analysis carried out in this paper is not only related to the superconducting cuprates , but it has connections with other physically interesting problems , like the spin chains with long range interaction and macroscopic quantum tunnelling . by considering the model ( [ azione1 ] ) in @xmath338 ,
it is seen that the frequency axis is left as the only relevant `` direction '' ( cf .
the technical considerations presented in appendix [ compcritind ] ) . as a first consequence , all the critical behavior should be reformulated in terms of frequency by suitably rescaling of the critical indices by @xmath339 .
the frequency axis becomes then equivalent to a single space direction , via the dynamic critical index @xmath20 , thus establishing a connection with onedimensional classical models .
fourier transforming back to ( imaginary ) times the damping term in ( [ azione1 ] ) gives rise to a long range interaction of the type @xmath340 .
our results can be directly extended to the case @xmath341 .
this leads to the comparison with the onedimensional ( classical ) @xmath342-component spin model with long - range interaction considered in ref .
one can recognize that , once the spatial term is dropped in ( [ azione1 ] ) the connection is established through the coupling combinations @xmath275 and @xmath343 .
this allows to rewrite the rg equation for @xmath344 in the same form as eq .
( 7 ) of ref .
@xcite , with @xmath344 ( to be identified with @xmath334 ) having a fixed point ( fp2 in our model ) and @xmath345 .
whitin the @xmath338 case , @xmath346 deserves special attention because of the connection with macroscopic quantum tunnelling @xcite .
this connection is realized by observing that the @xmath347 non - linear @xmath31 model can be rewritten in terms of an angular degree of freedom and that the addition of the magnetic field produces a cos - like interaction ( sine - gordon model ) . in @xmath338
the model then describes a single degree of freedom with a kinetic ( @xmath348 ) term in a multivalley potential . in the presence of dissipation ,
a linear - in - frequency term can appear , leading again to long - range interactions on the ( imaginary ) time axis .
models of this kind were considered by chakravarty @xcite in the case of a double well potential and successively by schmid @xcite to describe a dissipative quantum particle in a onedimensional periodic potential . in particular , the only difference between our eq .
( [ azione1 ] ) at @xmath346 and this latter model is the non - periodicity of the dissipative term of ref .
@xcite . in the model of schmid , for vanishing @xmath103
( @xmath296 in his notation ) , a line of fixed points for @xmath282 ( corresponding to our @xmath275 ) was found , with a critical point @xmath349 separating a regime with relevant @xmath103 from one with irrelevant @xmath103 .
the long - range term introduced in @xcite is quadratic in the fields , so that the model for @xmath129 is free and @xmath282 ( i.e. @xmath275 ) does not renormalize non - linear @xmath31 model in @xmath7 for @xmath129 , where a singular point appears at a special value of the coupling ( see , for instance , @xcite ) ] .
a ( dual ) fixed point was also found for very large @xmath103 , where the model of ref .
@xcite was stated to be equivalent to the one of ref .
we observe that , differently from the model of ref .
@xcite our long - range term mantains the periodicity in the angular variable and provides an interaction .
this renormalizes the coupling , thus allowing @xmath275 to flow under renormalization even at @xmath129 .
one could conjecture that a new fixed point arises at finite @xmath103 .
an adequate analysis would require a rg study at finite @xmath103 going beyond the @xmath60-expansion and taking into account instantonic corrections .
this is an interesting open problem which deserves further investigation .
* acknowledgments : * we gratefully acknowledge interesting discussions with c. di castro , a. chubukov and i. kolokolov .
10 for a review on magnetism in the cuprates see : a ) d. c. johnston , _ handbook of magnetic materials _ ,
- 237 , edited by k. h. j. buschow ( north holland , amsterdam 1997 ) , and b ) c. berthier , m. h. julien , m. horvati , and y. berthier , j. de phys .
i * 6 * ( 1996 ) 2205 . for a recent review on magnetism in the cuprates with a particular attention to nmr - nqr experiments see : a. rigamonti , f. borsa and p. carretta , _ basic aspects and main results of nmr - nqr spectroscopies in high temperature superconductors _ , preprint 1998 .
p. w. anderson , science * 235 * ( 1987 ) 1196 .
v. j. emery , s. a. kivelson , and h. q. lin , phys .
* 64 * ( 1990 ) 475 ; m. marder , n. papanicolau , and g. c. psaltakis , phys .
b * 41 * ( 1990 ) 6920 . v. j. emery and s. a. kivelson , physica c * 209 * ( 1993 ) 597 . c. castellani , c. di castro , and m. grilli , z. phys .
b * 103 * ( 1997 ) 137 .
i. affleck and b. halperin , j. phys .
a * 29 * ( 1996 ) 2627 .
a. h. castro neto and d. hone , phys .
* 76 * ( 1996 ) 2165 .
j. h. cho , f. borsa , d. c. johnston , and d. r. torgeson , phys .
b * 46 * ( 1992 ) 3179 . f. c. chou _
* 75 * ( 1995 ) 2204 .
r. j. gooding , n. m. salem , and a. mailhot , phys .
b * 49 * ( 1994 ) 6067 , and references therein . i. ya .
korenblit , v. cherepanov , a. aharony , o. entin - wohlman , cond - mat/9709056 .
s. chakravarty , b. i. halperin , and d. r. nelson , phys .
b * 39 * ( 1989 ) 2344 .
s. sachdev and j. ye , phys .
* 69 * ( 1992 ) 2411 .
a. v. chubukov and s. sachdev , phys .
lett * 71 * ( 1993 ) 169 ; phys .
lett * 71 * ( 1993 ) 2680(e ) ; a. v. chubukov , s. sachdev , and j. ye , phys .
b * 49 * ( 1994 ) 11919 . a. sokol and d. pines , phys .
* 71 * ( 1993 ) 2813 .
a. sokol , in _ the los alamos symposium 1993 strongly correlated electronic materials _ , edited by k. s. bedell _ et al .
_ ( addison - wesley , reading , ma , 1994 ) , pp
. 490493 .
t. imai , c. p. slichter , k. yoshimura , and k. kosuge , phys .
* 70 * ( 1993 ) 1002 .
m. takigawa , phys .
b * 49 * ( 1994 ) 4158 .
v. barzykin , d. pines , a. sokol , and d. thelen , phys .
b * 49 * ( 1994 ) 1544 .
v. barzykin and d. pines , phys .
b * 52 * ( 1995 ) 13585 . a. v. chubukov , d. pines , and b. p. stojkovi , j. of phys .
* 8 * ( 1996 ) 10017 .
f. d. m. haldane , phys
. lett . * 93a * ( 1983 ) 464 ; phys .
* 50 * ( 1983 ) 1153 ; i. affleck , nucl . phys . * b257 * ( 1985 ) 397 .
a. m. polyakov , phys . lett . * 59b * ( 1975 ) 79 .
e. brzin and j. zinn - justin , phys .
b * 14 * ( 1977 ) 3110 .
j. zinn - justin , _ quantum field theory and critical phenomena _
( oxford university press , new york , 1993 ) .
s. sachdev , a. v. chubukov , and a. sokol , phys .
b * 51 * ( 1995 ) 14874 . b. i. shraiman and e. d. siggia , phys . rev . lett .
* 61 * , ( 1988 ) 467 ; phys .
b * 42 * ( 1990 ) 2485 .
s. sachdev , phys .
b * 49 * ( 1994 ) 6770 .
j. a. hertz , phys .
b * 14 * ( 1976 ) 1165 .
m. randeria and j .- c .
campuzano , in _ proceedings of the international school of physics enrico fermi " _ ( varenna , 1997 )
. a. v. chubukov , d. k. morr , and k. a. shakhnovich , phyl . mag .
b * 74 * ( 1996 ) 563 .
p. w. anderson , phys .
* 64 * , 1839 ( 1990 ) ; _ ibid . _
* 65 * , 2306 ( 1990 ) j. maly and k. levin , phys .
b * 55 * , r15657 ( 1996 ) . m. e. fisher , s. ma , and b. g. nickel , phys . rev . lett .
* 29 * ( 1972 ) 917 .
j. sak , phys .
b * 8 * ( 1973 ) 281 .
a. aharony , in _ phase transitions and critical phenomena _ , edited by c. domb and m. s. green ( academic press , 1976 ) , vol .
6 , p. 358 .
a. v. chubukov , phys .
b * 52 * ( 1995 ) r3840 . d. r. nelson and r. a. pelcovits , phys . rev .
b * 16 * ( 1977 ) 2191 .
j. c. le guillou and j. zinn - justin , phys .
* 39 * ( 1977 ) 95 ; phys .
b * 21 * ( 1980 ) 3976 .
p. monthoux and d. pines , phys .
b * 49 * ( 1994 ) 4261 .
r. hlubina and t. m. rice , phys .
b * 51 * ( 1995 ) 9253 .
g. baskaran and p. w. anderson , cond - mat/9706076 .
j. m. kosterlitz , phys .
* 37 * ( 1976 ) 1577 .
a. o. caldeira and a. j. legget , phys .
46 * , ( 1981 ) 211 .
the case @xmath350 and @xmath346 describes phase fluctuations in low dimensional superconductors .
rg analyses on models , which are slightly different from the one considered in the present paper have been carried out by s. chakravarty , _ et al .
b * 37 * , ( 1988 ) 3283 , v. j. emery and s. a. kivelson , phys
lett . * 74 * , ( 1995 ) 3253 , and , more recently , by m. v. feigelman and a. i. larkin , cond - mat/9803006 .
s. chakravarty , phys .
* 49 * ( 1982 ) 681 .
a. schmid , phys .
. lett . * 51 * ( 1983 ) 1506 .
for @xmath160 the expressions ( [ loop],[conv ] ) reduce to @xmath362 the same expressions can be obtained by a direct evaluation of ( [ loop0 ] , [ kappaf ] ) for @xmath223 .
in particular the original expression ( [ loop0 ] ) shows that @xmath363 is in fact an analytic function of @xmath364 and @xmath43 .
however , some problems arise when using expressions ( [ loop2 ] ) and ( [ conv2 ] ) to compute critical indices directly for @xmath365 instead of considering the @xmath366 limit of the general expressions given in section 3.1 . for @xmath365 in fact one
expands around the critical dimension @xmath367 and the factor @xmath368 coming from angular integration vanishes .
if the same computation is performed using field - theoretical methods , instead of wilson s integration on the momentum shell , no such problems arise : _
e.g. _ one has @xmath369 which correctly behaves as @xmath370 for @xmath371 , independently of the value of @xmath63 .
notice that the @xmath372 factor coming from @xmath373 is cancelled by the factor @xmath374 and the @xmath375 divergence arises from @xmath376 .
the problems with the wilson method are resolved if one makes a more symmetric choice of the integration shell , _
e.g. _ integrating over all the @xmath377 with @xmath378 ( which implies introducing a frequency cutoff ) . with this choice
@xmath163 is finite for @xmath224 .
we observed that @xmath379 computed with either choice of cutoffs have the same @xmath380 dependence , the only effect of the choice of the symmetric shell being that the integrals do not vanish for @xmath224 .
in particular , the expressions for @xmath381 , @xmath382 given in the text are valid for @xmath383 : as a matter of fact , they are computed from the values ( [ iaka ] ) , which do not contain the angular integration factor @xmath373 . | we investigate the behavior of the zero - temperature quantum non - linear sigma model in @xmath0 dimensions in the presence of a damping term of the form @xmath1 , with @xmath2 .
we find two fixed points : a spin - wave fixed point fp1 showing a dynamic scaling exponent @xmath3 and a dissipative fixed point fp2 with @xmath4 in the framework of the @xmath5-expansion it is seen that there is a range of values @xmath6 where the point fp1 is stable with respect to fp2 , so that the system realizes a @xmath3 quantum critical behavior even in the presence of a dissipative term .
however , reasonable arguments suggest that in @xmath7 this range is very narrow . in the broken symmetry phase
we discuss a phenomenological scaling approach , treating damping as a perturbation of the ordered ground state .
the relation of these results with the pseudogap effect observed in underdoped layered cuprates is discussed .
, and renormalization group , non linear sigma model , quantum critical point , cuprates .
75.10.jm , 75.50.ee , 64.60.ak , 74.72.-h |
there have been several radar observations of enhanced plasma and ion lines at the topside ionosphere during ground based heating experiments in which the transmitting frequency has been kept somewhat lower than the critical plasma frequency of the f2 layer maximum .
a proposed explanation for the observed topside plasma and ion lines is a nonlinear decay of large amplitude z mode waves , which have been mode converted from the o mode heater wave at the bottomside of the ionosphere , into langmuir and ion acoustic ( ia ) waves .
the observations have attracted much interest since both the ordinary ( o ) and fast extraordinary ( x ) electromagnetic waves are reflected at or below the critical region where @xmath0 , except at a very narrow angular range around the critical angle of incidence ( the spitze angle ) where the o mode wave is mode converted to z mode waves @xcite .
( throughout the manuscript , we use the standard notation @xmath1 , @xmath2 and @xmath3 @xcite where @xmath4 is the electron plasma frequency , @xmath5 is the electron gyrofrequency and @xmath6 is the transmitted frequency . ) enhanced topside plasma lines in the f2 region were first observed in heating experiments at arecibo by @xcite . during experiments where the eiscat uhf beam was directed near the spitze angle , @xcite observed enhancements of ion lines in the topside interaction f2 region where @xmath0 .
similar observations at eiscat were done in the e region by @xcite , where there were no clear dependence of the power of the topside echoes on the power of the plasma turbulence at the bottomside .
there are several observations at eiscat of topside echoes over an angular range of several degrees @xcite . in the experiment at eiscat by @xcite , which was performed with low duty cycle ( 200 ms every 10 s ) , topside excitation of ion and plasma lines were also observed in a angular range of many degrees from the spitze angle , including vertical incidence . in order for normally incident o wave to tunnel through the f peak , conversion of the o mode to the z mode should occur , and the following conditions are necessary .
first , that the transmitter frequency should be near the maximum plasma frequency of the f peak so that @xmath7 for the peak plasma frequency of the f layer , as discussed by @xcite and shown in the linear e - region simulations for a parabolic density profile @xcite .
second , the existence of the density / electric field fluctuations within the critical layer , which can occur due to the generation of electrostatic turbulence excited by the parametric instability or in the earlier phase of the nonlinear thermal self - focusing instability @xcite .
@xcite proposed that the topside plasma and ion lines could be due to resonant scatter of the hf pump wave into z mode waves on the small - scale magnetic field aligned striations , which in turn are created by the hf pump wave via the resonant instability , or , in the case of the low duty cycle experiments with pump frequencies near a multiple of the electron gyro frequency , due to a second - order four wave interaction @xcite .
rocket observations during heating experiments in arecibo @xcite revealed z mode waves and field aligned striations above the o mode reflection point , and evidence of generation of bernstein waves where the local plasma frequency matches an electron gyro harmonic . during experiments where the eiscat uhf beam was directed along the magnetic field lines , @xcite discovered spectra with an unusually large spectral width ( up to 200 khz ) and a mean frequency shifted by 200 - 300 khz away from the heating frequency in both the upshifted and downshifted channels .
this feature has been named the hf - induced outshifted line ( hfol ) .
the hfol spectra come from a height about 3 - 6 km above the reflection height .
@xcite propose that the hfol feature can be consistently explained if the upper region is the reflection layer of the z mode appearing due to the mode conversion at the reflection level of the injected o mode wave .
their scenario includes large - scale density depletions at the airy maxima of the z mode , the parametric decay of the z mode into electron bernstein and lower hybrid waves @xcite , and the production of suprathermal electrons by the electrostatic turbulence , which in turn excite langmuir waves inside the density depletions , yielding the hfol .
the interaction between the electromagnetic pump wave and the electrostatic turbulence have been modeled with zakharov - type equations @xcite and wave kinetic equations @xcite , to understand the wave spectra and radiation generation from electromagnetically driven langmuir turbulence @xcite . in this letter
we analyze the conversion of o mode waves into z mode waves and the generation of topside ionospheric turbulence numerically with a full - scale simulations of a generalized zakharov model @xcite based on a nested grid method @xcite .
we are using realistic length scales and ionospheric parameters relevant for the high - latitude facility eiscat near troms , norway , but similar parameters may be also found at high - latitude facilities such as haarp , alaska and hipas , near fairbanks , alaska .
in our one - dimensional simulation geometry , we assume that a large amplitude electromagnetic wave is injected vertically , along the @xmath8 axis , into the vertically stratified ionospheric layer . the mathematical model , derived in @xcite , is based on a separation of time scales , where the time - dependent quantities ( here denoted @xmath9 ) are separated into a slow and a high - frequency time scale , @xmath10 .
the high - frequency component is assumed to take the form @xmath11complex conjugate , where @xmath12 represents the slowly varying complex envelope of the high - frequency field , and @xmath13 is the frequency of the transmitted electromagnetic wave .
hence the time derivatives on the fast timescale will be transformed as @xmath14 when going from the @xmath15 variable to the @xmath12 variable .
the dynamics of the high - frequency electromagnetic waves and electron dynamics is coupled nonlinearly to the slow timescale electron dynamics via the ponderomotive force acting on the electrons , and is mediated to the ions via the electrostatic field on the slow timescale .
we first present the envelope equations on the fast time scale .
the transverse ( to the @xmath8 axis ) components of the electromagnetic field in the presence of a slowly varying electron density @xmath16 is governed by the electromagnetic wave equation @xmath17 @xmath18 where @xmath19 and @xmath20 is the transverse vector potential ( in coulomb gauge ) and electric field , respectively . here
@xmath21 is the speed of light in vacuum and @xmath22 is the electric vacuum permittivity .
the @xmath8 component of the electric field is governed by the maxwell equation @xmath23 the high - frequency electron dynamics is governed by the continuity and momentum equation @xmath24 and @xmath25 respectively , where @xmath26 is the geomagnetic field , @xmath27 is the electron collision frequency , @xmath28 is the magnitude of the electron charge , @xmath29 is the electron mass , @xmath30 is the electron thermal speed , @xmath31 is boltzmann s constant , @xmath32 is the electron temperature , and @xmath33 is the background electron number density .
we have assumed that the ions are immobile on the fast time scale due to their large mass compared to that of the electrons .
the slowly varying electron number density is separated as @xmath34 where @xmath35 represents the large - scale electron density profile due to ionization / recombination in the ionosphere , and @xmath36 is the slowly varying electron density fluctuations .
we regard the electrons as inertialess on the slow time scale , where we also assume quasi - neutrality @xmath37 .
the dynamics of the ions is governed by the continuity equation @xmath38 and , since the ions are assumed to be unmagnetized , the slow timescale plasma velocity is obtained from the ion momentum equation driven by the ponderomotive force @xmath39 where @xmath40^{1/2}$ ] is the ia speed , @xmath41 is the ion temperature , @xmath42 is the ion mass , and we have denoted @xmath43 and @xmath44 . in obtaining eq .
( [ eq_ia_mom ] ) , we have used the ion momentum equation @xmath45 together with the electron force balance equation @xmath46 where we have neglected effects of the geomagnetic field @xmath26 on the ponderomotive force ( the @xmath47-term ) since we will consider high - frequency waves with frequencies much higher than the electron gyro - frequency .
we next define the simulation setup and physical parameters used in the simulation ; the details of the numerical implementation is described in the appendix of @xcite .
the simulation code uses a one - dimensional geometry , along the @xmath8 axis .
our simulation box starts at an altitude of 200 km and ends at 400 km .
while @xmath19 and @xmath20 are represented on a grid with grid size 2 m everywhere , the rest of the quantities @xmath48 , @xmath49 , @xmath36 , @xmath50 and @xmath51 are resolved with a much denser grid of grid size 4 cm at the bottomside and topside interaction regions regions , @xmath52@xmath53 and @xmath54@xmath55 , in order to resolve small - scale structures due to electrostatic turbulence . outside these regions ,
all quantities are resolved on the 2-meter grid .
this nested grid procedure is used to avoid a severe courant - friedrich - levy ( cfl ) condition @xmath56 on the timestep if the electromagnetic field is resolved on the dense grid @xmath57 .
the ionospheric density profile will be assumed to have a gaussian shape of the form @xmath58 $ ] , where @xmath59 and @xmath60 are the maximum density and the altitude of the f peak , and @xmath61 is the density scale length .
the external magnetic field @xmath62 is set to @xmath63 t and is tilted @xmath64 rad ( @xmath65 ) to the vertical ( @xmath8 ) axis so that @xmath66 $ ] , which is the case at eiscat in troms .
we assume that the transmitter on ground has been switched on at @xmath67 .
the simulation starts at @xmath68 , when the electromagnetic wave has reached the altitude 200 km .
initially , all time - dependent fields are set to zero , and random density fluctuations of order @xmath69 are added to @xmath36 to seed the parametric instability in the plasma .
the electromagnetic wave is injected from the bottomside of the simulation box at 200 km , by setting the @xmath70 component of the electric field to 1 v / m on the upward propagating field ( see @xcite ) .
the transmitter frequency is set to @xmath71 , which is somewhat lower than the maximum f peak plasma frequency @xmath72 , calculated with @xmath73 . on the other hand , the chosen wave frequency is higher than the maximum z mode cutoff frequency @xmath74 calculated from @xmath75 with @xmath76 , @xmath77 , and @xmath78 . by this choice of pump frequency ,
the o mode is reflected by the overdense plasma layer , while z mode waves are allowed to propagate to the topside when electrostatic turbulence is generated due to the parametric instability .
we use oxygen ions so that @xmath79 where @xmath80 is the proton mass .
further , @xmath81 so that @xmath82 m / s and @xmath83 m / s , and we use @xmath84 , which is the electron number density at @xmath0 .
the electron collision frequency is set to @xmath85 and the effective ion `` collision frequency , '' due to landau damping , is set to @xmath86 , where we have neglected that the ion landau damping depends on the wavelength of the ia wave .
the value used here is approximately valid for ia waves with wavelengths of @xmath8712 m.
axis ) electric field in log scale and the density profile of the f layer ( solid line ) .
electrostatic turbulence generated due to the parametric instability at the bottom and top sides of the f layer : at @xmath88 , the o wave is transferred into z waves which tunnel through the layer , and at @xmath89 ms , the inverse conversion process yields o waves escaping the plasma . , width=321 ] the time evolution of the electric fields and ion density fluctuations are shown in figs . 1 and 2 from @xmath67 to @xmath90 .
figure 1 shows the spatial and temporal evolution of the perpendicular ( to the @xmath8 axis ) electric field amplitude @xmath91 .
after approximately @xmath92 , the electromagnetic wave reaches the bottomside interaction for the o mode at @xmath93 and for the x mode at a slighter lower altitude @xmath94 , after which a steady - state standing pattern quickly builds up . at @xmath95 , there is an efficient conversion of the incident o wave into the z mode due to rapid development of the density fluctuations ( discussed in connection to fig . 2 below ) .
the z mode waves propagate from the bottomside at @xmath96 to the topside interaction region for the o mode at @xmath97 . at @xmath98 ,
we observe excitation of o mode radiation at the topside at @xmath99 , and these o mode waves propagate away from the ionospheric profile on the topside . in fig .
2 , we have visualized the @xmath100 component of the high frequency electric field and the ion density fluctuations @xmath36 , in the vicinity of the topside and bottomside interaction regions for the o mode . in fig .
2(b ) , we see the arrival of the o mode at the bottom side at @xmath101 , after which a standing wave pattern @xcite is set up . at @xmath102 ,
the standing wave pattern breaks up into small - scale , large amplitude waves . at the same time , seen in fig .
2(d ) , large amplitude ion density fluctuations are excited first in the region of the maxima of @xmath103 at @xmath104 and @xmath105 ( see fig .
2 ) , and later spreading to lower altitudes .
this transition into wave turbulence is closely correlated with the excitation of z mode waves observed in fig .
1 . at @xmath106@xmath107 , we see in fig .
2(a ) the arrival of the @xmath108 mode at the topside ionosphere .
while some weaker waves propagate to higher altitudes , a band of large - amplitude electric field is set up at @xmath109 where the z mode wave meets the transformation point @xcite and converts into electrostatic waves . at @xmath110
later than @xmath111 the z mode wave decays into short wavelength langmuir and ia waves due to a parametric instability , followed by langmuir wave collapse and ion caviton formation . in fig 2(c ) , the ion fluctuations are seen at the topside in a narrow band at @xmath112 .
the large amplitude collapsed langmuir waves and the associated ion cavitons have sizes of a few tens of centimeters or less and a spacing of a meter or less , similar as in the study of @xcite .
the topside langmuir turbulence gives rise to the radiation of o mode waves seen in fig 1 at @xmath113 , in a process described by @xcite .
we have performed a simulation study of enhanced topside turbulence , where the transmitter frequency is kept slightly lower than the critical frequency of the f2 layer .
our simulation study reveals that the electromagnetic wave decays parametrically on the bottomside of the ionosphere , in which collapsing wavepackets and ion density depletions are created .
this is a three - wave decay of the o mode into one counter - propagating langmuir ( l ) wave and one forward scattered ia wave , obeying the matching conditions @xmath114 of the frequencies and @xmath115 of the wavenumbers , where @xmath116 , @xmath117 and @xmath118 are given by the o mode , magnetized langmuir , ia dispersion relation , respectively .
a wavelength of the o mode of the order 100 meters leads to wavelengths of the langmuir and ia waves of the order one meter @xcite .
the linear instability is followed by a rapid langmuir waves collapse and the generation of ion cavitons . in the nonlinear phase of the instability ,
there is a strong generation of z mode waves in the bottomside interaction region .
these large - amplitude z - mode waves are allowed to propagate into the denser plasma to the topside interaction region of the ionospheric layer , where they reach their transformation point and their amplitude increase as their group speed decreases and they are turning into electrostatic waves . at the topside ,
the z mode waves undergo a parametric instability into small - scale ( few tens of centimeters ) coupled electrostatic and ia waves , followed by langmuir wave collapse and the formation of ion cavitons . in this last step ,
langmuir turbulence generated o mode waves @xcite are propagating away from the plasma layer on the top side .
we note that the wave tunneling and the excitation of topside turbulence occur within a few milliseconds , and may explain the observed topside enhanced plasma and ion lines far away from the spitze region , in particular during low duty - cycle ionospheric interaction experiments where the impact of ionospheric irregularities are minimized @xcite .
dubois , d. f. , h. a. rose , and d. russell ( 1990 ) , excitation of strong langmuir turbulence in plasmas near critical density : application to hf heating of the ionosphere , _
j. geophys .
95_(a12 ) , 21,221 .
galeev , a. a. , r. z. sagdeev , v. d. shapiro , v. i. shevchenko ( 1977 ) langmuir turbulence and dissipation of high - frequency energy , _ sov .
jetp _ , _ 46 _ , 711720 .
ganguly , s. , and w. e. gordon ( 1983 ) , heater enhanced topside plasma line , _ geophys .
10_(10 ) , 977978 .
gelinas , l. j. , m. c. kelley , m. p. sulzer , e. mishin , m. j. starks ( 2003 ) , in situ observations during an hf heating experiment at arecibo : evidence for z - mode and electron cyclotron harmonic effects , _ j. geophys .
108_(a10 ) , 1382 , doi:10.1029/2003ja009922 .
ginzburg , v. l. , rasprostranenie elektromagnitnykh voln v plazme ( the propagation of electromagnetic waves in plasmas ) 2nd ed .
( moscow : nauka , 1967 ) [ translated into english ( oxford : pergamon press , 1970 ) ] .
gondarenko , n. a. , p. n. guzdar , s. l. ossakow , and p. a. bernhardt ( 2003 ) , linear mode conversion in inhomogeneous magnetized plasmas during ionospheric modification by hf radio waves , _ j. geophys
108_(a12 ) , 1470 , doi:10.1029/2003ja009985 .
gondarenko , n. a. , s. l. ossakow , and g. m. milikh ( 2006 ) , nonlinear evolution of thermal self - focusing instability in ionospheric modifications at high latitudes : aspect angle dependence , _ geophys .
_ , _ 33 _ , l16104 , doi:10.1029/2006gl025916 .
isham , b. , w. kofman , t. hagfors , j. nordling , bo thid , c. lahoz , and p. stubbe ( 1990 ) , new phenomena observed by eiscat during an rf ionospheric modification experiment , _ radio sci .
25_(3 ) , 251262 .
isham , b. , b. t. rietveld , t. hagfors , c. la hoz , e. mishin , w. kofman , t. b. leyser , and a. p. van eyken ( 1999 ) , aspect angle dependence of hf enhanced incoherent backscatter , _ adv .
space res . _ _
24_(8 ) , 10031006 .
mishin , e. , t. hagfors , and b. isham ( 2001 ) , a generation mechanism for topside enhanced incoherent backscatter during high frequency modification experiments in troms , _ geophys .
28_(3 ) , 479482 .
rietveld , m. t. , b. isham , t. grydeland , c. la hoz , t. b. leyser , f. honary , h. ueda , m. kosch , and t. hagfors ( 2002 ) , hf - pump - induced parametric instabilities in the auroral e - region , _ adv .
space res .
29_(9 ) , 13631368 . | we present a full - scale simulation study of a generalized zakharov model for the generation of the topside electrostatic turbulence due to the parametric instability during ionospheric heating experiments near the f region peak .
the nonlinear tunneling of electromagnetic waves through the ionospheric layer is attributed to multiple - stage parametric decay and mode - conversion processes . at the bottomside of the f region , electrostatic turbulence excited by the parametric instability results in the conversion of the ordinary ( o mode ) wave into a large amplitude extraordinary ( z mode ) wave tunneling through the f peak . at the topside interaction region
, the z mode undergoes parametric decay cascade process that results in the generation of the topside electrostatic turbulence and then conversion process yielding o waves that escape the plasma .
this study may explain the observed topside ionospheric turbulence during ground based ionospheric heating experiments . |
small wide - angle telescopes such as the berlin exoplanet search telescope ( best ; * ? ? ?
* ) offer a unique possibility of identifying stellar variability in a large number of stars . apart from giving us important information on the individual targets ,
these kinds of wide - field data also contain information on the statistical distribution of variable stars , which can be used not only to understand the structure and evolution of the galaxy , but also to analyze the expected false - alarm rates for transiting extrasolar planet searches @xcite . for space missions searching for extrasolar planets by photometric transits ,
knowledge of the variable stars in the target fields helps reduce the number of false alarms , thereby reducing the amount of follow - up observation needed .
in addition , the variable stars provide interesting targets for additional science .
the _ corot space mission , launched on 2006 december 27 , will observe a number of target fields for up to 150 days in order to perform asteroseismic studies of the target stars and to search for extrasolar planets by photometric transits .
the observation of the first 150 day `` long - run '' field ( lrc1 ) started in spring 2007 .
this field has been observed with best in order to identify variable stars .
_ corot is expected to obtain a precision of 0.66 ppm in 5 days of observation in the asteroseismic channel and 700 ppm in 1 hr of observation in the exoplanet channel @xcite .
this is an order of magnitude better than what can be obtained with best , because _ corot benefits from not having to observe through the atmosphere .
therefore , many of the transits that are expected to be detected with _
corot will not be visible in the best data , but despite this order - of - magnitude gap the best data will be highly valuable to the _ corot mission . _ _ _ _ _ in the exoplanet channel of the _ corot satellite , stars have a large point - spread function ( psf ) due to a prism installed in front of the detectors to obtain the colors of each star .
the color information will help to distinguish stellar activity from exoplanet transits .
in dense target fields , however , the resulting relatively large psfs of the stellar spectra can overlap .
prior knowledge of variables overlapping the psfs of interesting target stars will therefore be helpful to avoid misinterpretation .
_ in addition to searching exoplanetary transits , additional science programs will be performed in the _ corot fields on selected targets .
some of these additional programs are interested in variable stars , e.g. , binaries . to be able to place integration windows in the satellite field on such targets ,
their position has to be known before observing the field with _ corot .
a search for variable stars in the corot fields is therefore also helpful for the additional science program of the mission by detecting previously unknown variables in the field and confirming or rejecting the variability of the stars mentioned in variable star catalogs . _ _ a variability characterization like the one presented here of the target fields prior to _
corot observations is therefore important for two reasons : to find interesting variable stars to monitor further by preselecting windows and to make an analysis of the impact of the variable stars on detecting extrasolar planet transits with _ corot . _
_ the present paper is arranged as follows : section 4.2 describes the observations , section 4.3 summarizes the data reduction procedure , section 4.4 describes the data analysis , including search and detection of periodic variable stars , section 4.5 contains the variable star catalog of the observed field , and section 4.6 describes the analysis of the completeness of the data set .
concluding remarks are found in section 4.7 .
best is a small wide - angle telescope operated by the institut fr planetenforschung of the deutsches zentrum fr luft- und raumfahrt ( dlr ) and is located at the observatoire de haute - provence ( ohp ) , france .
it consists of a schmidt telescope with a 20 cm effective aperture , an @xmath10 focal ratio , and an ap-10 ccd camera , with 2048 @xmath11 2048 pixels of 14 @xmath12 m size @xcite .
the result is a 3.1@xmath7 @xmath11 3.1@xmath7 field of view ( fov ) with a pixel scale of 5@xmath95 .
this is comparable to _ corot , which has a pixel scale of 2@xmath93 and a fov for each of the four ccds of 1.3@xmath7 @xmath11 1.3@xmath7 . _ we present the results from a monitoring program of the _ corot field lrc1 in the galactic center direction centered at ( @xmath1,@xmath2)=(19@xmath300@xmath400@xmath50,@xmath600@xmath701@xmath855@xmath92 ) ( j2000.0 ) obtained from 2005 june 6 to september 7 .
the data set consists of a total of 35 nights of observations spread out over 98 nights in summer 2005 .
no data were obtained on nights with bad weather , around midsummer , or around full moon .
the observations were made without a filter in front of the ccd ( the spectral response of the ccd corresponds to a wide @xmath13 bandpass ) in order to detect as many photons as possible and to provide a maximum sampling rate .
an equal number of frames were obtained with 40 , 120 , and 240 s integration times in order to observe stars in a large magnitude range and to analyze the effect of crowding in the frames .
we only present data from the 240 s integration frames here , corresponding to 727 frames in total .
4.1 shows the 3.1@xmath7 @xmath11 3.1@xmath7 fov observed . _
the reduction of the data was performed as described by @xcite . based on a comparative test of three different photometry algorithms ( sextractor ,
momf @xcite and isis ) a data reduction pipeline was built @xcite in order to reduce noise originating from the stellar background , extending the data pipeline of best data , which was optimized for less crowded fields observed at its previous location at the thringer landessternwarte , germany ( rauer et al .
we outline the different steps in the pipeline below .
the frames are calibrated from night to night by standard procedures as described in @xcite .
hereafter , the reduction generally follows the method given in the isis package .
image subtraction is performed in order to remove both the sky and the stellar background .
this is done by subtracting a reference frame from all the frames , which removes the sky background and all the constant stars . when image subtraction is performed in the right way it will produce uncrowded frames where only the variable stars are present .
the noise will then be dominated by the photon noise from the target stars and from the stellar background .
before image subtraction can be made all the frames need to be transformed into the same ( @xmath14)-coordinate grid .
the frame with the best seeing from the middle of the observation run was selected as the template frame for this transformation .
then isis produced a reference frame by combining a number of best - seeing frames .
the reference frame was then subtracted from all the other frames with the isis parameters listed in table 4.1 .
this means that the reference frame is convolved with a convolution kernel in order to match all the stars in the frames .
having subtracted the reference frame from all the other frames , isis constructs a frame that is the mean of the absolute normalized deviations of all the subtracted frames .
this frame only contains nonconstant stars , as all the constant stars have been removed by the subtraction .
unfortunately , this only works for nonconstant stars that have a large amplitude , as is discussed below .
the isis package also includes a photometry routine that performs psf - weighted photometry on the variable stars detected , thereby producing light curves of the differential fluxes of the stars . for different reasons , discussed below , we did not use this routine .
we used the output of isis to distinguish between true periodic variable stars and stars that showed variability because of contamination from another nearby variable star .
there are three reasons why we do not use the photometry algorithm in isis .
the first is that isis only produces light curves of differential fluxes ; in order to convert these into magnitudes , photometry needs to also be performed on the reference frame in the same way as on the subtracted frames .
the second reason , which has been noted by @xcite and @xcite , is that the detection of variables in isis works best for variables with an amplitude of at least 0.1 mag , since small - amplitude variables tend to get washed out in the overall noise in the subtracted frames .
the last reason is that in order to do psf - weighted photometry , the stars need to have a well - defined psf ; this is not always true if the stars are undersampled .
we therefore preferred to perform simple unit - weight aperture photometry on the subtracted frames .
@xcite analyzed differences in performing psf - weighted and unit - weighted aperture photometry on the subtracted frames and concluded that psf - weighted photometry works best for the relatively faint stars , while unit - weighted aperture photometry works best for the relatively bright stars , but also that psf - weighted photometry gives rise to a constant error term in the light curves .
in addition , it is important that the transformation from differential fluxes to magnitudes is correct when evaluating the rms noise level obtained by different photometry algorithms and when comparing the noise level for ground - based transit surveys like best to that of space - based transit surveys like _ corot . by using the same unit - weighted aperture photometry algorithm on the subtracted frames , as well as on the reference frame ,
we are confident that this transformation is correct .
_ @xcite note that if they use an aperture radius below 7 pixels for stars with a fwhm of the order of 3 pixels , the amplitude of the light curves is artificially reduced . as the fwhms of the stars we analyze here are in the range from 1 to 3 pixels , depending on the seeing on different nights
, we use an aperture radius of 7 pixels .
we did try to reduce the aperture radius , which resulted not only in reduced amplitudes of the light curves but also reduced rms noise level in the light curves , as expected because of a decrease in the amount of background noise . in other words it might be possible to increase the precision by using smaller or nonconstant apertures , but then the amplitude and noise determination would not be as reliable . it is also noted here that although the psfs in the nonsubtracted frames may be small , there may be large non - gaussian residuals in the subtracted frames that require large apertures in order to ensure flux preservation .
it is important to note here that the aperture radius of 7 pixels is only used on the subtracted frames where all the constant stars are removed .
therefore , using an aperture radius of 7 pixels does not mean that we are not able to resolve stars separated by less than 7 pixels ( as long as they are not variables ) .
figure 4.2 shows the rms scatter in a single night ( june 29 ) of all the light curves as a function of stellar luminosity , with the identified periodic variable stars marked .
the figure shows that periodic variable stars are found roughly equally distributed over stellar luminosity .
the obtained precision is comparable to what has been obtained by other small wide - angle telescopes @xcite . as noted by @xcite
the light curves need to be calibrated for systematic changes in the zero points .
this is done by selecting a number of nonsaturated stars with low rms noise in the light curves and then calculating the mean change of the light curves .
we selected 2182 stars for calculating these zero - point changes and found that the corrections from frame to frame were below 0.02 mag , except for a few bad frames which contained clouds or airplane tracks .
even after the zero - point correction , residuals of systematic trends in the light curves caused by , e.g. , air mass , cloud cover , and temperature variations were still present in the data ( these residuals are sometimes referred to as `` red noise '' ) .
we therefore applied the algorithm introduced by @xcite in order to correct for systematic effects .
this clearly improved the rms noise level , in particular in the brightest stars . a detailed discussion of the photometric limitations of the best survey will be presented in a forthcoming paper .
the astrometric calibration was performed by matching the 1500 brightest stars in the stellar catalog obtained from the reference frame with the 2000 brightest stars ( magnitude limit 11.4 ) of the usno - a2.0 catalog @xcite .
the problem of matching the observed stars to stellar catalogs is not trivial .
the main reason for this is the large image scale of best ( 5@xmath95 pixel@xmath15 .
this means that stars in a crowded part of the frame will have multiple matches in the stellar catalogs .
the astrometric calibration was performed using match @xcite .
match finds a linear , quadratic , or cubic transformation between two stellar catalogs .
we used a cubic transformation to calculate the transformation from the ( @xmath14)-coordinates of the ccd to right ascension and declination .
we note that a linear transformation does not work because of the large fov . out of the brightest 1500 stars in the stellar catalog obtained from the reference frame it was possible to match 902 with a matching radius of 2@xmath16 degree .
from these 902 stars we calculated an absolute magnitude difference between the observations and the usno - a2.0 catalog @xmath13 band .
we note that we do not intend to perform a real absolute photometric calibration of the stars , since only relative magnitudes are needed to identify variables .
we used the variability index @xmath17 defined by @xcite to identify stars as suspected variable stars .
the index @xmath17 is defined as the normalized sum of the deviation of each data point ( or pair of data points with small separations in time ) in the light curve compared to the expected noise level for that star ( the expected noise level is calculated based on the signal - to - noise ratio of the star ) .
we considered data points to be pairs if they were separated by less than 0.03 day .
4.3 shows the index @xmath17 as a function of magnitude .
we classified all 9112 stars with an index @xmath17 greater than 2 as suspected variables , as all these stars show some kind of variability .
most of these stars are clearly variable stars , but either the period can not be properly determined because of an imperfect duty cycle or the stars are nonperiodic or multiperiodic variable stars . in this paper
we concentrate on the periodic variable stars .
all the stars classified as suspected variables were searched for periodicity with the method introduced by @xcite .
this method fits a set of periodic orthogonal polynomials to the observations and evaluates the quality of the fit with the use of an analysis - of - variance statistic .
all the suspected variable stars were fitted with two polynomials , and all objects with a fit quality better than 0.9 were examined visually . in the visual examination we selected stars as periodic variable stars if they showed a clear periodicity between 0.286 and 120 days .
we rejected light curves with periods close to 1 day , or harmonics of 1 day , because the periodicities in these light curves are likely to be caused by periodic changes in background level , temperature , air mass , etc .
we have identified 92 periodic variable stars in the data set ( tables 4.2 and 4.3 ) .
it is seen in fig .
4.1 that the periodic variable stars are detected all over the ccd , although the density depends on the stellar density in that part of the field .
the gcvs @xcite contains 60 variable stars in the field , 11 of which have periods less than 120 days .
we have identified all 11 stars in the best data set .
the phased light curves are shown in fig .
4.4 , and the corresponding parameters are given in table 4.3 . by looking at the light curves and comparing them with the index @xmath17 we were able to confirm variability in six cases ( g1 , g2 , g8 , g9 , g10 , and g11 ) .
none of the 11 stars were detected by the algorithm we had set up to search for periodicity .
g8 , g9 , and g11 are rr lyrae stars with such a characteristic light curve that they should have been detected by the algorithm . by going back in the analysis to see why these stars were not detected
, we see that they were all excluded during the visual inspection .
there are three reasons for this exclusion .
first , the method introduced by @xcite provides a period that is equal to half the true period .
this means that the phased light curves mainly show a falling trend .
second , the method introduced by schwarzenberg - czerny ( 1996 ) provides a period offset from the true period ( the period of g8 is off by 5% , g9 by 0.2% , and g11 by 33% ) , thereby making the phased light curve look a lot more noisy .
third , g8 and g11 have periods very close to 1 day .
g1 , g2 , and g10 do not show strongly periodic signals , and detection was not expected .
we were not able to confirm variability in five cases ( g3 , g4 , g5 , g6 , and g7 ) .
there are three possible reasons why we were not able to confirm variability in these three stars : ( 1 ) the identification as variable stars in the gcvs is wrong , ( 2 ) the stars are not variable anymore , and ( 3 ) the variability of the stars is below the noise level in the best data . in table 4.3
it is seen that all three stars should have amplitudes larger than 1 mag , and in fig .
4.4 it is seen that the observed variability in the light curves is well below 1 mag .
this rules out the third possibility .
we have divided the periodic variable stars identified by our search into two groups : periodic pulsating variables and eclipsing binaries .
the two groups each contain 43 stars . for most of the stars it is easy to make this classification , but
for some stars the classification may not be unique .
the light curves of the periodic variable stars are shown in figs 4.5 and 4.6 , and the parameters are listed in table 4.2 .
the detected periodic variable stars were cross - identified with the usno - a2.0 catalog . for each star we checked that the cross - identification was unique . if there were multiple or no identifications we did not assign a usno - a2.0 i d to the star .
the usno - a2.0 ids are given in table 4.2 .
we also looked up all the detected periodic variable stars in the simbad catalog but were only able to identify one of the periodic variable stars , i.e. , v66 = v1225 aql .
some of the stars show problems with blending .
this means that the stars next to the variable stars also show variability . for each case where we had two or more stars showing the same kind of variability within the radius of the aperture ( 7 pixels ) we have only identified one of the stars as a periodic variable .
it is difficult to verify from the light curves alone which star is the true variable , because a faint star next to a bright variable star can show larger relative variability than the true bright variable star . in order to overcome this problem we identified all the variable stars that were affected by blending in the variability frame of our isis pipeline ( the frame that was the mean of the absolute normalized deviations of all the subtracted frames ) , because this frame will only show truly variable stars
this approach only works well if the relative amplitudes in the variable stars are high .
a few stars , such as v20 , v31 , and v50 , seem to show some additional variability apart from the periodic signal . by inspecting the light curves we found that the additional variability originated from a few nights where the noise level was unusually high .
it is therefore most unlikely that this additional variability is of stellar origin .
forty of the 92 identified periodic variable stars are in the lrc1 fov of one of the two _ corot ccds dedicated to the planet program .
these are v3 , v4 , v6 , v9 , v11 , v12 , v13 , v15 , v21 , v23 , v25 , v26 , v27 , v32 , v38 , v39 , v40 , v41 , v42 , v43 , v44 , v45 , v47 , v49 , v50 , v53 , v54 , v58 , v59 , v60 , v61 , v65 , v69 , v78 , v83 , v84 , v85 , v86 , g2 , and g9 .
some of the remaining identified periodic variable stars are expected to be in the fov of some of the future pointings . _ we have identified all the periodic variable stars in the exodat catalog @xcite .
the exodat catalog is part of the preparation of _ corot and has been made in order to select which stars to observe with @xmath18 , taking into account spectral type , luminosity class , variability , and contamination ( overlapping psf on the focal plane ) .
it contains , among other things , @xmath19 photometry of the field observed with the 2.5 m isaac newton telescope at la palma .
_ we cross - identify the periodic variable stars from best with the exodat catalog using a matching radius of 1@xmath20 deg . in this way
we obtain 81 matches out of the 86 periodic variable stars .
more identifications could have been obtained by increasing the matching radius , but that would have also increased the number of mismatches .
4.7 shows a color - magnitude diagram of the exodat catalog , with the discovered periodic variable stars marked .
the stars in the @xmath21 @xmath22 versus @xmath22 diagram follow two branches : to the left the main - sequence stars and to the right the giant stars . for the faint stars
there is a mismatch between stars on the main sequence and stars on the giant branch , since we do not know anything about the distance to the stars and therefore can not provide their absolute luminosity .
there is also a mismatch for nearby late main - sequence stars , which are bright with red colors and fall within or to the right of the giant branch .
the diamonds in the figure represent eclipsing binaries , and the crosses represent periodic pulsating stars .
a vague tendency for eclipsing binaries to fall mainly on the main sequence can be seen ( 15 out of 43 ; 35% ) , compared to the number of periodic pulsating stars that fall on the main sequence ( 7 out of 49 ; 14% ) .
this is in agreement with what we expect , since most of the periodic pulsating stars we have found are cepheids , miras , or rr lyrae stars , and none of these stars fall on the main - sequence branch in the @xmath21 @xmath22 versus @xmath22 diagram .
due to their short periodicity , the identified eclipsing binaries can not be comprised of giant stars .
hence , from their location in fig . 4.7 , they range from solar - type stars ( at the typical main - sequence field star distance of _ corot fields ) to later but more nearby main - sequence stars . _
a way to evaluate the capability of ground - based surveys to detect variable stars is to compare the number of eclipsing binaries detected with a certain system with the expected number of eclipsing binaries in general .
evaluating the capability of ground - based surveys this way has the advantage that it includes not only the photometric precision of the data in the evaluation but also the phase coverage and the detection algorithms . fig .
4.8 shows the number of eclipsing binaries in the best data compared with the _ hipparcos data .
a description of how the eclipsing binaries are detected in the _ hipparcos data can be found in . _ _ the _ hipparcos stars have been separated into two different absolute magnitude bins ( @xmath23 = 2.2 to 0.9 ) and ( @xmath23 = 0.3 to 4.0 ) .
this separation corresponds to a separation into two different parallax bins . as we do not know the parallaxes for the stars observed with best
, we do not know the absolute magnitudes .
we have tried to compare the best data with the _ hipparcos data from both bins with more or less the same result , although the close _ hipparcos sample seems to fit the best data slightly better than the distant sample . _ _ _ fig .
4.8 shows that best detected more eclipsing binaries at low periods , which is caused by the higher sampling rate of best compared to _ hipparcos . the fact that best detected more eclipsing binaries than hipparcos at low periods indicates that _ hipparcos may not be perfect at low periods . at higher periods ( log @xmath24 = 0.0 to 0.5 )
best only detected 20% 30% of the number of eclipsing binaries detected by _ hipparcos .
have calculated the relative number of eclipsing binaries as a function of period in three different amplitude bins ( @xmath25 0.1 , 0.4 and 0.7 mag ) . as almost all the eclipsing binaries
we have detected have amplitudes between 0.1 and 0.4 mag , we only discuss the results from the first bin here , i.e. , the bin with the largest eclipse probabilities .
this statistical analysis shows that the best data set presented here has a completeness of 20%
30% compared with the _ hipparcos data set for periods longer than 1 day and is complete compared with the _ hipparcos data set for short - period binaries . _ _ _ _ _ as there were six periodic variable stars in the gcvs that we did not detect with our periodic variable star detection algorithm , there is reason to believe that our algorithm is not perfect .
this means that by improving the algorithm more periodic variable stars would be detected and the completeness of the data set would increase . in particular , it is seen that we detect a lower number of eclipsing binaries with periods around 1 day than what is expected .
this might mean that we have been too conservative with excluding periodic variable stars that have a period close to 1 day .
best has observed the first long - run field ( lrc1 ) of the _ corot mission for 35 nights .
we have found 92 periodic variable stars in the data : 43 eclipsing binaries and 49 periodic pulsating stars .
_ data reduction has been performed using image subtraction .
astrometric and photometric calibration were made by matching stars to the usno - a2.0 catalog .
we used the index @xmath17 defined by @xcite to identify suspected variable stars and searched these stars for periodicity with the method introduced by @xcite . in this way
we have identified 92 periodic variable stars , compared to the already known 11 variable stars in the gcvs with periods less than 120 days .
the periodic variable stars have been identified in the exodat catalog , and the color - magnitude diagram of the stars in the field shows that the eclipsing binaries have a slight tendency to lie on the main sequence compared to the periodic pulsating stars .
this reflects that most of the periodic pulsating stars detected in the data are cepheids , miras , or rr lyrae stars .
we also present a new tool to evaluate the completeness of a survey data set by comparing the number of eclipsing binaries detected in the data set with statistical estimates of the relative number of eclipsing binaries .
the analysis shows that this best data set has a completeness of 20% 30% compared with the _ hipparcos data set for periods greater than 1 day and is complete relative to _
hipparcos for shorter periods . _ _
we would like to thank t. arentoft for helpful discussions and careful reading of the manuscript and the referee , b. s. gaudi , for many useful suggestions and comments .
we thank the observatoire de haute - provence for great support of the best survey .
k. gratefully acknowledges financial support by the danish national science research council through the danish asteroseismology centre ( dasc ) and the instrument center for danish astrophysics ( ida ) . this research has made use of the simbad database , operated at cds , strasbourg , france .
alard , c. 2000 , a&as , 144 , 363 baglin , a. , auvergne , m. , barge , p. , buey , j .-
t . , catala , c. , michel , e. , weiss , w. , & corot team 2002 , esa sp-485 : stellar structure and habitable planet finding , 17 bertin , e. , & arnouts , s. 1996 , a&as , 117 , 393 bord , p. , rouan , d. , & lger , a. 2003 , , 405 , 1137 brown , t. m. 2003 , apj , 593 , l125 deleuil , m. , moutou , c. , deeg , h. , meunier , c. , surace , c. , guterman , p. , almenara , j. m. , alonso , r. , barge , p. , bouchy , f. , erikson , a. , leger , a. , loeillet , b. , ollivier , m. , pont , f. , rauer , h. , rouan , d. , & queloz , d. 2006 , esa sp-1306 : the corot mission , 341 hartman , j. d. , bakos , g. , stanek , k. z. , & noyes , r. w. 2004 , aj , 128 , 1761 hartman , j. d. , stanek , k. z. , gaudi , b. s. , holman , m. j. , & mcleod , b. a. 2005 , aj , 130 , 2241 kane , s. r. , collier cameron , a. , horne , k. , james , d. , lister , t. a. , pollacco , d. l. , street , r. a. , & tsapras , y. 2004 , , 353 , 689 kane , s. r. , lister , t. a. , collier cameron , a. , horne , k. , james , d. , pollacco , d. l. , street , r. a. , & tsapras , y. 2005 , , 362 , 117 kane , s. r. , collier cameron , a. , horne , k. , james , d. , lister , t. a. , pollacco , d. l. , street , r. a. , & tsapras , y. 2005 , , 364 , 1091 karoff , c. , rauer , h. , erikson , a. , & voss , h. 2006 , astronomical society of the pacific conference series , 349 , 261 kholopov , p. n. , et al .
1998 , in combined general catalogue of variable stars , 4.1 ed ( ii/214a ) kjeldsen , h. , & frandsen , s. 1992 , pasp , 104 , 413 mochejska , b. j. , stanek , k. z. , sasselov , d. d. , & szentgyorgyi , a. h. 2002 , aj , 123 , 3460 monet , d. , et al .
1998 , the pmm usno - a2.0 catalog .
( 1998 ) rauer , h. , eislffel , j. , erikson , a. , guenther , e. , hatzes , a. p. , michaelis , h. , & voss , h.2004 , pasp , 116 , 38 schwarzenberg - czerny , a. 1996 , apj , 460 , l107 stetson , p. b. 1996 , pasp , 108 , 851 sderhjelm , s. , & dischler , j. 2005 , a&a , 442 , 1003 valdes , f. g. , campusano , l. e. , velasquez , j. d. , & stetson , p. b. 1995 , pasp , 107 , 1119 tamuz , o. , mazeh , t. , & zucker , s. 2005 , , 356 , 1466 vs. @xmath13 magnitude .
the gray dots represents all the observed stars while the crosses represents the periodic variable stars that we have identified .
the 9112 stars above the line with a index @xmath17 larger then 2.0 were selected as suspected variable stars . ] | the _ corot space mission will monitor several target fields for up to 150 days to perform asteroseismology and to search for extrasolar planets by photometric transits .
variable stars in the target fields are important objects for additional scientific studies but can also disturb the search for planetary transits . a variability characterization of the target fields prior to _
corot observations is therefore important for two reasons : to find interesting variable stars to monitor further and to make an analysis of the impact of the variable stars on detecting extrasolar planet transits with _ corot . _
_ _ the berlin exoplanet search telescope ( best ) is a small wide - angle telescope dedicated to high - precision photometry .
it has observed a 9 deg@xmath0 field of view centered at ( @xmath1,@xmath2)=(19@xmath300@xmath400@xmath50,@xmath600@xmath701@xmath855@xmath92 ) ( j2000.0 ) over 98 nights to search for variable stars in the surroundings of the first long - run target field ( lrc1 ) of the _ corot space mission . in this data set we identified 92 periodic variable stars , 86 of which are
new discoveries and 6 of which are known from the general catalogue of variable stars ( gcvs ) . for five of the gcvs stars , variability could not be confirmed .
forty - three of the 92 detected periodic variable stars are identified as eclipsing binaries .
we have evaluated the completeness of our survey for eclipsing binaries by comparing it to the expected fraction of eclipsing binaries based on _ hipparcos observations . from this evaluation
we show that the best data set presented here has a completeness of 20% 30% for periods
longer than 1 day and is complete relative to _ hipparcos for short - period binaries . _ _ _ |
the one component plasma ( ocp ) is one of the basic models in the field of charged systems .
the ocp model is formulated as a system of point particles , interacting via the coulomb potential , which move in a uniform neutralizing background .
it has found important physical applications in a variety of fields ranging from terrestrial physics , through important technological applications to cosmology @xcite . as a reference model
it is used in many areas of soft condensed matter , such as colloidal and polyelectrolyte solutions , e.g. @xcite , etc .
all thermodynamic properties of the ocp depend only on the dimensionless plasma parameter @xmath4 , where @xmath5 is the bjerrum length ( @xmath6 is the charge of the particles , @xmath7 is the boltzmann s constant , @xmath8 is the temperature ) and @xmath9 is the ion - sphere radius with @xmath10 being the concentration of particles ( @xmath11 is the number of particles , @xmath12 is the volume of the system ) . using a field - theoretical approach , a fairly accurate and simple expression for the equation of state of the ocp has been obtained within the gaussian approximation for the effective hamiltonian @xcite ; contrary to previous calculations , e.g. @xcite , this gives a correct behavior for the thermodynamic functions in the full range of @xmath0 ( see @xcite and references therein ) and does not have fitting parameters as e.g. in @xcite . in @xcite field - theoretical calculations for the equation of state , going beyond the gaussian approximation , have been performed , showing that corrections to the gaussian theory are rather small .
a closely related model the hard - core one component plasma ( hcocp ) incorporating a hard core repulsion between ions gives a more satisfactory description of the short - range electrostatic correlations .
the importance of this model follows also from the fact that it belongs to the class of the so - called primitive models used to describe molten salts @xcite , electrolytes @xcite , liquid metals @xcite and charged colloidal solutions @xcite .
the mean spherical approximation ( msa ) @xcite was applied to account for both the hard - core and electrostatic interactions in this system @xcite .
although it has an analytical solution , it exhibits sometimes unphysical negative contact values for the pair correlation function , which has to be remedied by a rescaling procedure@xcite .
a mixed percus - yevick / hypernetted chain integral equation also has been used @xcite and it was observed that it fits better the simulation results @xcite than the msa and cluster expansion@xcite .
the former however does not have an analytical formulation .
an analytical equation of state for the hcocp has been proposed in @xcite , as a simple generalization of the hole - corrected debye - hckel theory @xcite , which in addition to the correlation hole around charged particles takes into account the hard - core repulsion .
this was used afterwards to develop a generalized van der waals theory @xcite . the basic physical idea exploited in this theory
is that due to the strong electrostatic repulsion a `` hole '' appears around a charged particle from which all other particles are expelled .
outside the hole the electrostatic interactions are not very strong and may be described on the debye - hckel level ; the size of the hole is found self - consistently .
although being physically appealing , this theory does not give a satisfactory description for large values of the plasma parameter @xmath0 for the ocp @xcite , and is also not accurate for the hcocp for large packing fractions @xmath13 ( for @xmath14 the deviations from mc data for the internal energy reach @xmath15 @xcite ) .
recently _ an exact _ low - density expansion for the free energy of the hcocp has been obtained @xcite .
this however can not be applied in the case of strong electrostatic interaction , i.e. for large @xmath0 . in the present study
we develop a theory which allows to derive a fairly accurate and simple equation of state for the hcocp in the whole range of @xmath0 from the debye - hckel limit @xmath16 up to the limit of strong coupling @xmath1 .
it reproduces within 1 - 3% accuracy the available monte carlo ( mc ) data for @xmath17 .
larger deviations occur in the region where the mc data are not very accurate ( see the discussion below ) . as in the case of the ocp
@xcite we use the hubbard - schofield transformation which yields a field - theoretical hamiltonian for the hcocp with coefficients expressed in terms of equilibrium correlation functions of a reference hard - core fluid .
although we perform the explicit calculations within the gaussian approximation for the effective hamiltonian , one can go beyond the gaussian approximation using the standard perturbation technique , provided the structural properties of the reference hard - core fluid are known . since the obtained equation of state is in good agreement with available mc data we use it to address the problem of additivity of the hard - core and electrostatic contributions to the excess free energy .
this seems to be important since it is usually assumed that these contributions are additive ( e.g. @xcite ) and the range of validity of such approximation has not been studied yet . the rest of the paper is organized as follows : in sec.2 we briefly sketch the hubbard - schofield approach which yields the field - theoretical hamiltonian for the hcocp . in sec.3
we derive the equation of state within the gaussian approximation for the effective hamiltonian , and compare the theoretical findings with available monte carlo data . in that section
we also analyse the accuracy of the widely used approximation in which the excess free energy is written as a sum of the hard - core and electrostatic part . in the conclusion
we summarize our findings .
we start from the hcocp hamiltonian with omitted ideal part which may be written as follows ( @xmath18 ) : @xmath19 where the first term in the right - hand side of ( [ 1 ] ) refers to the coulomb interactions , written in terms of collective density variables , @xmath20 where @xmath21 denotes coordinate of j - th particle , @xmath22 is the fourier - transformed coulomb potential and @xmath23 describes the hard - sphere interaction .
summation in ( [ 1 ] ) is to be performed over the wave vectors @xmath24 with @xmath25 @xmath26 , where @xmath27 are integers , @xmath28 , and the prime over the sum denotes that the term with @xmath29 is excluded @xcite .
the configurational integral may be written in terms of the configurational integral of the reference ( hard - sphere liquid ) system @xmath30 @xcite as : @xmath31 where @xmath32 denotes the averaging over the reference system . in accordance with the hubbard - schofield scheme
@xcite we use the identity : @xmath33 and arrive after some algebra at : @xmath34 where @xmath35 , and where the integration is to be performed under restriction , @xmath36 ( @xmath37 is the complex conjugate of @xmath38 ) @xcite .
applying the cumulant theorem @xcite to the factor @xmath39 one obtains : @xmath40 , with @xmath41 @xmath42 @xmath43 here @xmath44 denotes _ cumulant _ average @xcite for the reference hard sphere fluid system .
note that ( [ 5 ] ) gives the field - theoretical expression for the partition function with , @xmath38 being the fourier components of the scalar field @xmath45 @xcite .
the coefficients of the effective hamiltonian ( [ 5 ] ) are expressed in terms of the correlation functions of the reference hard - core fluid . using definitions of correlation functions of fluids @xcite and definitions of the cumulant averages @xcite
, one can _ directly _ evaluate @xmath46 ( and thus the coefficients @xmath47 ) .
it is straightforward to show that @xmath48 may be expressed in terms of the fourier transforms of the correlation functions @xmath49 of the reference system , defined as @xcite @xmath50 @xmath51 where @xmath52 are @xmath53-particle correlation functions @xcite . in particular the second - order cumulant @xmath54 may be written as : @xmath55\,,\ ] ] and the fourier transform of the function @xmath56 is related to the direct correlation function @xmath57 @xmath58\,,\ ] ] for which an explicit analytical expression is known @xcite .
now we concentrate on the gaussian part of the effective hamiltonian , i.e. we skip all the terms with a power of the field larger than two .
the accuracy of this approximation has been critically examined in ref.@xcite for the ocp without hard - core interactions by perturbatively calculating higher - order terms .
it was found that higher order terms contribute very little to the free energy and that the gaussian approximation is in fact excellent over the whole range of coupling parameters .
one can therefore assume , and this is indeed borne out by our comparison with monte - carlo data below , that the gaussian approximation should also be quite good for the present case . using ( [ h2_k ] ) and ( [ cum_2 ] ) we write for this case : @xmath59 performing ( gaussian ) integration over @xmath38 we arrive after some algebra at @xcite : @xmath60 taking the logarithm of the configuration integral one obtains the free energy , and since the analytical expression for @xmath61 is available @xcite , no additional approximation is , in principle , required .
this leads , however , to an expression which is to be evaluated numerically .
we note that owing to the long - range nature of the coulombic interactions , the main contribution to the free energy comes from the long - wave modes of the density fluctuations , which correspond to small @xmath62 .
therefore only the small-@xmath62 behavior of the direct correlation function is important .
this suggests to approximate @xmath61 by a truncated expansion @xmath63 which correctly behaves at small @xmath62 .
as we show in what follows , only wave - vectors with @xmath64 contribute to the configuration integral , thus we apply the approximation ( [ c2_k ] ) for the interval @xmath65 . by numerical evaluation of the free energy using the full expression for @xmath57
, we convinced ourselves that deviations of the quadratic form from the actual @xmath57 at larger @xmath62 do not noticeably affect the results@xcite .
we therefore do not require @xmath62 and @xmath66 to be small since the particular behavior of @xmath57 for large @xmath62 is not important .
on the other hand , the quadratic approximation allows to obtain an analytical equation of state for the hcocp that reproduces fairly well the available monte carlo data . with ( [ c2_k ] )
one can write for the configurational integral : @xmath67 the most accurate estimate for @xmath68 may be found , using @xmath69 from ( [ h2_k ] ) and the relation for the isothermal compressibility @xmath70 @xcite , where @xmath71 is the pressure of the reference system @xcite : @xmath72 the value of @xmath73 follows from the fairly accurate carnahan - starling free energy of hard - sphere fluid @xcite : @xmath74 where @xmath75 is the packing fraction and @xmath76 is the diameter of the spheres ; this yields @xmath73 as a second derivative of @xmath77 with respect to density : @xmath78 using the wertheim - thiele solution for the direct correlation function @xmath79 @xcite and definition of @xmath80 from eq.([c2_k ] ) one obtains : @xmath81 with @xmath82 from ( [ c_2 ] ) and @xmath83 from ( [ h2_k ] ) and ( [ z_0 ] ) we arrive at the following expression for the configurational integral @xmath84 where @xmath85 .
now we show that within the gaussian approximation to the effective hamiltonian , one obtains rather accurate equation of state for hcocp provided that an appropriate value of the `` ultraviolet cutoff '' in the @xmath86-space is employed . from ( [ q_end ] )
we find for the excess free energy of the hcocp : @xmath87\ ] ] now we argue that the summation in ( [ fex ] ) should be carried out over a _ finite _ number of the wave - vectors @xmath86 . in this
we follow the debye theory of the specific heat of solids ( e.g. @xcite ) .
namely , we assume that the total number of degrees of freedom in the system , @xmath88 , should be equal to the total number of _ physically different _ modes with the wave - vectors @xmath86 within the spherical shell of radius @xmath66 in the @xmath86-space .
the number of modes is twice the number of the wave - vectors , since for each @xmath86 one has a sine and cosine mode ( the amplitude of the @xmath86-th mode is a complex number ) @xcite .
thus we obtain : @xmath89 where the factor @xmath90 appears when the integration in @xmath86-space is used instead of summation . from ( [ k0_def ] )
follows that @xmath91 a similar debye - like scheme to find the cutoff @xmath66 was first proposed for plasma in @xcite , where a somewhat different value of the cutoff wave - vector was reported . using @xmath66
as obtained above we write : @xmath92\\ \phantom{*}\\ \displaystyle\qquad=\frac{\beta f_{hc}}{n}+\frac{9}{4}\int\nolimits^1_0x^2dx \left[\ln\left(\theta+\frac{b\gamma z_0}{x^2}\right)-\frac{b\gamma}{x^2}\right]\ , , \end{array}\ ] ] where @xmath93 , so that @xmath94 ( see ( [ nukdef ] ) for the definition of @xmath95 ) , and where we define the constant @xmath96 .
the last integral is easily evaluated and yields for the excess free energy @xmath97-\frac{3}{2}\left(\frac{b\gamma z_0}{\theta}\right)^{\frac{3}{2 } } \arctan\left(\sqrt{\frac{\theta}{b\gamma z_0}}\phantom{|}\right)\ ] ] with @xmath98 and for the excess internal energy , @xmath99 of the hcocp @xmath100\,.\ ] ] for @xmath101 , eqs.([fex_fin],[u_fin ] ) recover the corresponding result for the one component plasma @xcite .
as it follows from ( [ fex_fin ] ) and ( [ u_fin ] ) , for @xmath102 and @xmath103 the debye - huckel behavior is obtained . in the opposite limit @xmath1 and for any packing fraction @xmath13 eqs.([fex_fin ] ) and
( [ u_fin ] ) demonstrate a linear behavior on @xmath0 .
the leading term for this case is @xmath104 , where the constant @xmath105 reads : @xmath106 this is fairly close to the constant @xmath107 of the fits for the ocp ( see e.g.@xcite ) . for arbitrary values of @xmath13 and @xmath0
we compare our analytical expression for the excess internal energy ( [ u_fin ] ) with the available mc data for the hcocp ( fig.1 ) .
fig.2 shows the relative error of the analytical expression ( [ u_fin ] ) .
as it follows from fig.1 and fig.2 the equation of state is fairly accurate in full range of plasma parameters for which mc data are available . for @xmath108 the relative error does not exceed @xmath109 for all values of packing fraction and plasma parameter . for @xmath110
one observes the linear behavior and deviation from the numerical data for this range of @xmath0 less than @xmath111 .
the maximal deviation of the analytical expression from the mc data occurs for @xmath112 . in this range of the plasma parameter
the deviation is about @xmath113 with the maximal one of @xmath114 at the smallest value @xmath115 .
it should be noted however , that such deviation occurs for @xmath0 , where the method of mc loses its accuracy .
moreover , we expect that for very small packing fraction , @xmath116 , where the maximal deviation is observed , the contribution to the internal energy due to the hard - core interactions may not exceed @xmath111 .
therefore for this @xmath13 the difference between @xmath3 of the hcocp and @xmath3 of the ocp is less then @xmath111 .
it has been already shown that the analytical equation of state for the ocp ( which follows from ( [ u_fin ] ) for @xmath101 ) has the accuracy of @xmath117 @xcite .
thus we expect that the accuracy of our equation of state may not be worse than @xmath118 for these values of @xmath13 ; the observed deviation of @xmath114 seems to be the manifestation of the low accuracy of mc for this range of parameters . with
the equation of state obtained we can analyse the accuracy of the widely used approximation for the excess free energy , which represents this as a sum of the hard - core and electrostatic component , @xmath119 ( see e.g. @xcite ) .
the validity of this approximation may be checked by the direct comparison of the excess free energy of the hcocp @xmath120 and the above sum .
for @xmath120 we use our equation of state ( [ fex_fin ] ) and equation of state for ocp from @xcite ( this expression for @xmath121 may be obtained from eq.([fex_fin ] ) for @xmath101 ) .
for @xmath122 we use the carnahan - starling eq.([fhc ] ) . in fig.3
we compare the complete equation of state ( solid lines ) with the approximate , based on the assumption of additivity of the hard - core and electrostatic parts ( dashed lines ) .
as it follows from fig.3 for small @xmath0 the values of the free energy from the complete and approximate equations of state may differ significantly ; they may even have different signs .
the relative difference becomes smaller with increasing @xmath0 ( see inset to fig.3 ) and decreasing packing fraction .
for @xmath123 the approximation reaches the reasonable accuracy : the error does not exceed @xmath113 for all @xmath13 . for small packing fractions , @xmath124 one has the same accuracy already for @xmath110 .
a `` first - principle '' equation of state for the classical hard - core one component plasma is obtained that has a correct debye - huckel behavior at the limit of small plasma parameter @xmath0 and small packing fractions ( @xmath125 , @xmath126 ) .
it demonstrates a linear dependence on @xmath0 for any packing fraction if @xmath127 .
the obtained coefficient @xmath128 at the linear leading term in this case is close to the corresponding coefficient @xmath129 found for the one component plasma in the monte carlo simulations .
the simple analytical expression for the excess internal energy reproduces the available mc data with an accuracy of @xmath109 for the most range of @xmath0 and @xmath13 .
the maximal deviation of @xmath114 is observed for small @xmath0 and @xmath13 , where the mc method is not very accurate , and we argue that such deviation does not reflect the accuracy of our equation . to derive the equation of state we apply the hubbard - schofield transformation to obtain the field theoretical hamiltonian for the hcopc and use the gaussian approximation .
the gaussian approximation assumes that one can neglect all terms in the effective hamiltonian which contain the power of field higher than two .
as it was shown by the field theoretical calculations for the case of the one component plasma ( without the hard - core ) @xcite , the corrections to the gaussian theory are rather small , and we expect that the same is true for the system of interest .
physically , this implies that the higher - order coefficients in the field theoretical hamiltonian are smaller than that at the second order of the field . in this case
the contribution to the field integral from the field - amplitude domain where the higher - order terms dominate is small : the main contribution comes from the domain where the gaussian term prevails .
generally , the gaussian approximation fails near the critical point , where the gaussian term of the effective hamiltonian becomes small and even vanishes @xcite . for the hcocp we do not expect any criticality and thus the gaussian approximation
is expected to be valid . to obtain the analytical expression for the equation of state we also approximate the fourier transform of the direct correlation function of the reference hard - core system @xmath57 by its small-@xmath62 expansion .
although such approximation deviates from the actual dependence of @xmath57 at larger @xmath62 , this does not eventually affect the equation of state : owing to the long - range nature of the coulombic interactions , only the small-@xmath62 behavior of @xmath57 is important ; this is correctly reproduced by the approximation .
we also analyse the validity of the widely used approximation , where the free energy of the hcocp is represented as a sum of the hard - core and electrostatic component .
we show that such approximation is rather accurate for small packing fraction and large plasma parameter . in the opposite case of large @xmath13 and small @xmath0 the excess free energy
may not be adequately represented by a sum of a hard - core and electrostatic part . to obtain eq.([q_start ] ) we take into account that @xmath130 where @xmath131 and @xmath132 are respectively the real ( cosine ) and imaginary ( sine ) parts of @xmath133 . similar relation for @xmath134 is used , to obtain eq.([q_fin ] ) .
the direct correlation function for the hard sphere fluid reads @xcite : @xmath135 , when @xmath136 and @xmath137 when @xmath138 , where @xmath76 is the diameter of the hard spheres . for @xmath57 , which may be easily found from @xmath79 we do not need the complete expression .
the maximal deviation between the free energy evaluated with the full expression for @xmath139 and with the quadratic expansion is approximately @xmath140 and occurs at a packing fraction of @xmath141 and a plasma parameter @xmath142 . for all other values of the parameter , the deviations are smaller .
the value of @xmath68 may be obtained directly as the zero - order moment of the wertheim - thiele solution for @xmath79 @xcite of the hard - sphere fluid .
this is however based on the approximate percus - yevik closure @xcite and occurs to be less accurate than that related to @xmath143 , with @xmath143 taken from the carnahan - starling compressibility .
this happens because the carnahan - starling semiempirical equation of state @xcite is more accurate than those found from the wertheim - thiele solution @xcite .
note that the restriction @xmath144 does not reduce the total number of _ physically different _ modes , since the modes with the wave - vectors @xmath86 and @xmath145 correspond to that spreading in the opposite directions . on the plasma parameter @xmath4 ( @xmath5 , @xmath9 ) for different values of the packing fraction
@xmath146 .the curves from top to bottom correspond to @xmath116 , @xmath147 , @xmath148 , @xmath149 and @xmath150 , respectively .
points give the monte - carlo data @xcite : circles correspond to @xmath116 , squares to @xmath147 , triangles to @xmath148 , diamonds to @xmath149 and down triangles to @xmath150 , respectively . in the inset the same dependence
is shown for larger range of @xmath0 . ]
( [ fex_fin ] ) , for the free energy of the hcocp ( solid lines ) with the approximate one , given as the sum of the free energy of the point particle ocp and that of the hard - core fluid , @xmath152 ( dashed lines ) .
the curves from top to bottom correspond respectively to @xmath153 , @xmath154 and @xmath155 . in the inset the same dependence
is shown for larger range of @xmath0 . ] | an accurate and systematic equation of state for the hard - core one - component plasma ( hcocp ) is obtained .
the result is based on the hubbard - schofield transformation which yields the field - theoretical hamiltonian , with coefficients expressed in terms of equilibrium correlation functions of the reference hard - core fluid .
explicit calculations were performed using the gaussian approximation for the effective hamiltonian and known thermodynamic and structural properties of the reference hard - core fluid . for small values of the plasma parameter @xmath0 and packing fraction the debye - hckel result
is recovered , while for @xmath1 , the excess free energy @xmath2 and internal @xmath3 energy depend linearly on @xmath0 .
the obtained expression for @xmath3 is in a good agreement with the available monte carlo data for the hcocp .
we also analyse the validity of the widely used approximation , which represents the free energy as a sum of the hard - core and electrostatic part .
preprint accepted to journal of physics d : applied physics . |
at very small values of @xmath1 it is expected that the number density of partons within the proton becomes so large that they begin to recombine with each other .
this phenomenon of parton recombination is also referred to as absorptive corrections , non - linear effects , screening , shadowing , or unitarity corrections , all leading to saturation .
the first perturbative qcd ( pqcd ) calculations describing the fusion of two pomeron ladders into one were made by gribov - levin - ryskin ( glr ) @xcite and by mueller - qiu ( mq ) @xcite .
the glrmq equations add an extra non - linear term , quadratic in the gluon density , to the usual dglap equations for the gluon and sea - quark evolution .
the evolution of the gluon distribution is then given by @xmath6 ^ 2,\ ] ] where @xmath7 fm is of the order of the proton radius .
the glrmq equations account for all ` fan ' diagrams , that is , all possible @xmath8 ladder recombinations , in the double leading logarithmic approximation ( dlla ) which resums all powers of the parameter @xmath9 .
there has been much recent theoretical activity in deriving ( and studying ) more precise non - linear evolution equations , such as the balitsky - kovchegov ( bk ) and jalilian - marian iancu mclerran weigert leonidov kovner ( jimwlk ) equations ( see @xcite for a review ) .
note that the bk and jimwlk equations are both based on bfkl evolution .
however , for the most relevant studies in the hera and lhc domain ( @xmath10 ) , the predominant theoretical framework is collinear factorisation with dglap - evolved parton distribution functions ( pdfs ) . at very small values of @xmath1
it might be expected that the dglap approximation would break down , since large @xmath11 ( bfkl ) terms would appear in the perturbation series in addition to the @xmath12 terms resummed by dglap evolution .
however , it turns out that the resummed nll bfkl calculations of the gluon splitting function @xmath13 @xcite and the gluon transverse momentum distribution @xcite are rather close to the dglap calculations .
moreover , the convolution @xmath14 coincides with the nnlo dglap result and is close to the nlo dglap result for @xmath10 @xcite .
hence , in the analysis of current data , it is reasonable to ignore bfkl effects .
if recombination effects are significant , it is therefore important that they be incorporated into the global dglap parton analyses which determine the pdfs from deep - inelastic scattering ( dis ) and related hard - scattering data .
such a programme , based on glrmq evolution ( which accounts for gluon - induced screening only ) , was implemented some years ago @xcite , before the advent of hera .
the input gluon and sea - quark distributions were _ assumed _ to have a small-@xmath1 behaviour of the form @xmath15 at an input scale of @xmath16 gev@xmath17 .
the inclusion of shadowing effects , both in the form of the input pdfs and in the glrmq evolution , was found to significantly _ decrease _ the size of the small-@xmath1 gluon distribution in comparison with the result with no absorptive corrections .
a crucial observation is that , at that time ( 1990 ) , @xmath0 data were only available for @xmath18 , and so these results were largely dependent on the theoretical assumptions made for the starting distributions .
however , with hera , we now have @xmath0 data down to @xmath19 or less , and so the pdfs at small @xmath1 can be determined directly from the hera data .
in fact , the advent of hera data has led to a puzzling behaviour of the small-@xmath1 gluon and sea - quark pdfs at low scales @xmath20 .
if we write @xmath21 and @xmath22 , then the expectation of regge theory is that @xmath23 for low scales @xmath24 gev , where @xmath25 @xcite is the power of @xmath26 obtained from fitting soft hadron data . at higher @xmath27 gev
, qcd evolution should take over , increasing the powers @xmath28 and @xmath29 .
however , the current mrst2004 nlo @xcite and cteq6.1 m @xcite pdf sets exhibit a very different behaviour at low scales from that theoretically expected ; see fig .
[ fig : mrstcteq ] .
in fact , the mrst group has found that a negative input gluon distribution at @xmath30 gev is required in all their nlo dglap fits since mrst2001 @xcite .
the cteq group , who take a slightly higher input scale of @xmath31 gev , also find a negative gluon distribution when evolving backwards to 1 gev .
gev@xmath17 found in the mrst2004 nlo @xcite and cteq6.1 m @xcite global analyses .
the valence - like behaviour of the gluon is evident.,scaledwidth=50.0% ] since data at small @xmath32 now exist , the introduction of the absorptive corrections is expected to _ increase _ the size of the input gluon distribution at small @xmath1 to maintain a satisfactory fit to the data . to understand this , note that the negative non - linear term in the glrmq equation slows down the evolution .
therefore , it is necessary to start with a _ larger _
small-@xmath1 gluon distribution at low scales @xmath33 to achieve the _ same _ pdfs at larger scales required to describe the data .
if the non - linear term is neglected , the input small-@xmath1 gluon distribution is forced to be artificially small in order to _ mimic _ the neglected screening corrections .
we have anticipated that the introduction of absorptive corrections will _ _ enhance _ _
gev@xmath17 , then evolving upwards with the glrmq equations based on lo dglap evolution , improves the agreement with @xmath0 data at small @xmath32 and low @xmath20 compared to the standard cteq sets , and leads to an enhanced small-@xmath1 gluon distribution for @xmath34 gev@xmath17 .
note , however , that there is a large nlo correction to the splitting function @xmath35 which changes completely the relationship between the quark and gluon distributions , and so weakens the conclusion of ref . @xcite . ] the small-@xmath1 gluon at low scales , and hence could possibly avoid what appears to be anomalous behaviour at small @xmath1 .
thus , here , we perform such a study using an abridged version of the mrst2001 nlo analysis @xcite , improving on our previous analysis @xcite . first , we derive a more precise form of the glrmq equations .
\(a ) + due to the @xmath8 pomeron contribution . ( b ) application of the agk cutting rules . for simplicity , the upper parton ladder , shown in the right - hand diagram of ( a ) ,
is hidden inside the upper blob in each diagram of ( b).,title="fig : " ] + \(b ) + due to the @xmath8 pomeron contribution .
( b ) application of the agk cutting rules . for simplicity , the upper parton ladder , shown in the right - hand diagram of ( a ) ,
is hidden inside the upper blob in each diagram of ( b).,title="fig : " ] the inclusive proton structure function , @xmath36 , as measured by experiment , can be approximately written as a sum of the single pomeron exchange ( dglap ) contribution and absorptive corrections due to a @xmath37 pomeron merging ; see fig . [
fig : agk](a ) .
that is , @xmath38 in computing @xmath39 we need to sum over all possible cuts .
the abramovsky - gribov - kancheli ( agk ) cutting rules @xcite were originally formulated in reggeon field theory but have been shown to also hold in pqcd @xcite .
application of the agk rules gives the result that relative contributions of @xmath40 , @xmath41 , and @xmath42 are obtained according to whether neither pomeron , one pomeron , or both pomerons are cut ; see fig .
[ fig : agk](b ) . therefore ,
the sum over cuts is equal to _ minus _ the diffractive cut and so the absorptive corrections can be computed from a calculation of the @xmath43-integrated diffractive structure function @xmath44 , where @xmath45 and @xmath46 is the fraction of the proton s momentum transferred through the rapidity gap .
the pqcd description of @xmath47 is described in @xcite .
working in the fixed flavour number scheme ( ffns ) , it can be written as @xmath48 apart from the secondary reggeon contribution . the separation between the soft pomeron and qcd pomeron
is provided by a scale @xmath49 gev . for simplicity , we take @xmath50 to be the same as the scale @xmath51 at which the input pdfs are taken in the analysis of @xmath0 data , so @xmath52 gev , the value used in the mrst2001 nlo analysis @xcite . the contribution to the absorptive corrections arising from the soft pomeron contribution of is already included in the input pdfs , therefore @xmath53,\ ] ] where @xmath54 is the fraction of diffractive events in which the proton dissociates . in practice , we take @xmath55 and take an upper limit of 0.1 instead of 1 for @xmath46 in . is justified by a zeus comparison @xcite of proton - tagged diffractive dis data with data which allowed proton dissociation up to masses of @xmath56 gev , where @xmath57 was obtained . ] first consider the contribution to from the @xmath58 term .
it corresponds to a @xmath37 pomeron merging with a cut between the two pomeron ladders and can be written as @xmath59 where @xmath60 are the _ same _ coefficient functions as in inclusive dis .
the diffractive pdfs , @xmath61 or @xmath62 , where @xmath63 , satisfy an _
inhomogeneous _ evolution equation @xcite : @xmath64 here , @xmath65 is the perturbative pomeron flux factor ,
@xmath66 ^ 2.\ ] ] the diffractive slope parameter @xmath67 comes from the @xmath43-integration , while the factor @xmath68 accounts for the skewedness of the proton gluon distribution @xcite .
there are similar contributions from ( light ) sea quarks , where @xmath69 in is replaced by @xmath70 , together with an interference term .
a sum over all three contributions is implied in and in the second term of .
the pomeron pdfs in , @xmath71 , are evolved using nlo dglap from a starting scale @xmath72 up to @xmath20 , taking the input distributions to be lo pomeron - to - parton splitting functions , @xmath73 @xcite . from , @xmath74 where @xmath75 or @xmath76 , and @xmath77 differentiating with respect to @xmath20 gives the evolution equations for the ( inclusive )
gluon and sea - quark pdfs : @xmath78 thus is a more precise version of the glrmq equations , which goes beyond the dlla and accounts for sea - quark recombination as well as gluon recombination .
consider the recombination of gluons into gluons , for example , in the dlla where @xmath79 , then @xmath80 @xcite .
taking @xmath81 and @xmath82 , then becomes @xmath83 ^ 2.\ ] ] comparing to this is simply the glrmq equation with @xmath84 . for numerical results
we take @xmath85 ( @xmath86 ) gev@xmath87 for light ( charm ) quarks , which would correspond to @xmath88 ( 1.1 ) fm .
the procedure for incorporating absorptive corrections into a ( nlo ) global parton analysis ( in the ffns ) is as follows : 1 .
parameterise the @xmath1 dependence of the input pdfs at a scale @xmath89 gev .
2 . evolve the pdfs @xmath90 and @xmath76 using the non - linear evolution equation .
( the non - singlet distributions are evolved using the usual linear dglap equations . )
3 . compute @xmath91,\ ] ] and compare to data . here , the two terms inside the square brackets are beyond collinear factorisation , that is , they can not be written as a convolution of coefficient functions with the pdfs .
the first term inside the square brackets corresponds to the process @xmath92 .
the second term corresponds to the process @xmath93 , for light quarks with a longitudinally polarised photon .
these contributions are calculated as described in ref .
@xcite . as usual ,
these three steps should be repeated with the parameters of the input pdfs adjusted until an optimal fit is obtained .
this procedure is our recommended way of accounting for absorptive corrections in a global parton analysis .
however , in practice , available nlo dglap evolution codes , such as the qcdnum @xcite program , are often regarded as a ` black box ' , and it is not trivial to modify the usual linear dglap evolution to the non - linear evolution of .
therefore , we adopt an alternative iterative procedure which avoids the explicit implementation of non - linear evolution , but which is equivalent to the above procedure .
we model our analysis of hera @xmath0 data @xcite on the mrst2001 nlo analysis @xcite , which was the first in which a negative gluon distribution was required at the input scale of @xmath94 gev .
( the more recent mrst sets have not changed substantially at small @xmath1 . )
we apply cuts @xmath95 , @xmath96 gev@xmath17 , and @xmath97 gev@xmath17 , leaving @xmath98 data points .
the input gluon and sea - quark distributions are taken to be @xmath99 where the powers of the @xmath100 factors are taken from @xcite , together with the valence - quark distributions , @xmath101 and @xmath102 , and @xmath103 .
the @xmath104 parameter is fixed by the momentum sum rule , while the other nine parameters are allowed to go free .
since we do not fit to dis data with @xmath105 , we constrain the input gluon and sea - quark distributions , and their derivatives with respect to @xmath1 , to agree with the mrst2001 nlo parton set @xcite at @xmath106 .
this is done by including the value of these mrst pdfs at @xmath106 , and their derivatives , as data points in the fit , with an error of 10% on both the value of the mrst pdfs and their derivatives .
therefore , the pdfs we obtain are not precisely constrained at large @xmath1 , but this paper is primarily concerned with the small-@xmath1 behaviour of the pdfs . the procedure we adopt is as follows :
1 . start by performing a standard nlo dglap fit to @xmath0 data with no absorptive corrections .
2 . tabulate @xmath39 , given by , and @xmath107 , given by , using pdfs @xmath108 and @xmath109 obtained from the previous fit .
3 . perform a standard nlo dglap fit to ` corrected ' data , @xmath110 , to obtain pdfs @xmath111 . then correct these pdfs to obtain @xmath112 .
these latter pdfs @xmath113 then satisfy the non - linear evolution equations .
go to ( ii ) .
each successive iteration of steps ( ii ) and ( iii ) introduces another level of @xmath37 pomeron mergings , so that eventually all the ` fan ' diagrams are included , achieving the same effect as the procedure described at the end of section [ sec : non - linear ] . note
that the correction to the pdfs , @xmath112 , in each step ( iii ) , was omitted in our previous analysis @xcite .
consequently , the effect of the absorptive corrections on the pdfs at large scales was overestimated .
also in @xcite , the known lo @xmath114 were multiplied by free parameters ( ` k - factors ' ) , determined from separate fits to diffractive dis data , in an attempt to account for higher - order pqcd corrections to the lo pomeron - to - parton splitting functions . however , since these k - factors took unreasonable values , with some going to zero , here we have chosen to fix them to 1 .
therefore , the updated analysis , presented here , does not require a simultaneous fit to the diffractive dis data . in fig .
[ fig : gluon](a ) we show the gluon distribution at scales @xmath115 , @xmath86 , @xmath116 , and @xmath117 gev@xmath17 obtained from fits before and after absorptive corrections have been included .
both fits are almost equally good with @xmath118 values of @xmath119 and @xmath120 for the fits without and with absorptive corrections respectively . at low @xmath20
the absorptive corrections give an increased gluon distribution at small @xmath1 , apart from at @xmath121 where there are only a few data points and where additional absorptive effects ( pomeron loops ) may become important .
the non - linear term of slows down the evolution , so that by @xmath117 gev@xmath17 the two gluon distributions are roughly equal ; see fig . [
fig : gluon](a ) .
( a)(b ) + data , before and after absorptive corrections have been included .
( b ) the effect of successive iterations on the gluon distribution obtained from fits to @xmath0 , taking a positive definite input gluon at 1 gev .
each iteration introduces another level of @xmath37 pomeron mergings.,title="fig:",scaledwidth=49.0% ] data , before and after absorptive corrections have been included .
( b ) the effect of successive iterations on the gluon distribution obtained from fits to @xmath0 , taking a positive definite input gluon at 1 gev . each iteration introduces another level of @xmath37 pomeron mergings.,title="fig:",scaledwidth=49.0% ] we repeated the fits without the negative term in the input gluon distribution , that is , without the second term in .
when absorptive corrections were included , almost the same quality of fit was obtained ( @xmath122 ) , while without absorptive corrections the fit was slightly worse ( @xmath123 ) .
we conclude that absorptive corrections lessen the need for a negative gluon distribution at @xmath115 gev@xmath17 .
the gluon distributions obtained from six successive iterations of steps ( ii ) and ( iii ) above are shown in fig .
[ fig : gluon](b ) .
the convergence is fairly rapid , with only the first three iterations having a significant effect , that is , the ` fan ' diagrams which include @xmath124 pomeron mergings .
although we have seen that the inclusion of absorptive corrections has reduced the need for a _ negative _ gluon , it has not solved the problem of the _ valence - like _ gluon .
that is , the gluon distribution at low scales still decreases with decreasing @xmath1 , whereas from regge theory it is expected to behave as @xmath125 with @xmath25 .
we have studied several possibilities of obtaining a satisfactory fit with this behaviour @xcite .
the only modification which appears consistent with the data ( and with the desired @xmath126 equality ) is the inclusion of power - like corrections , specifically , a global shift in all scales by about 1 gev@xmath17 .
( note that a similar shift in the scale is required in the dipole saturation model @xcite . )
however , we do not have a solid theoretical justification for this shift .
therefore , a more detailed , and more theoretically - motivated , investigation of the effect of power corrections in dis is called for .
we thank robert thorne for useful discussions . adm
thanks the leverhulme trust for an emeritus fellowship .
this work was supported by the uk particle physics and astronomy research council , by the federal program of the russian ministry of industry , science and technology ( grant ss-1124.2003.2 ) , by the russian fund for fundamental research ( grant 04 - 02 - 16073 ) , and by a royal society joint project grant with the former soviet union . l. v. gribov , e. m. levin and m. g. ryskin , phys .
* 100 * ( 1983 ) 1 .
a. h. mueller and j. w. qiu , nucl .
b * 268 * ( 1986 ) 427 .
j. jalilian - marian and y. v. kovchegov , prog . part .
( in press ) [ arxiv : hep - ph/0505052 ] . r. s. thorne , phys . lett .
b * 474 * ( 2000 ) 372 [ arxiv : hep - ph/9912284 ] ; g. altarelli , r. d. ball and s. forte , nucl .
b * 674 * ( 2003 ) 459 [ arxiv : hep - ph/0306156 ] ; m. ciafaloni , d. colferai , g. p. salam and a. m. stasto , phys .
b * 587 * ( 2004 ) 87 [ arxiv : hep - ph/0311325 ] .
v. a. khoze , a. d. martin , m. g. ryskin and w. j. stirling , phys .
d * 70 * ( 2004 ) 074013 [ arxiv : hep - ph/0406135 ] .
g. p. salam , arxiv : hep - ph/0501097 . j. kwiecinski , a. d. martin , w. j. stirling and r. g. roberts , phys .
d * 42 * ( 1990 ) 3645 .
a. donnachie and p. v. landshoff , phys .
b * 296 * ( 1992 ) 227 [ arxiv : hep - ph/9209205 ] .
a. d. martin , r. g. roberts , w. j. stirling and r. s. thorne , phys . lett .
b * 604 * ( 2004 ) 61 [ arxiv : hep - ph/0410230 ] .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. nadolsky and w. k. tung , jhep * 0207 * ( 2002 ) 012 [ arxiv : hep - ph/0201195 ] .
a. d. martin , r. g. roberts , w. j. stirling and r. s. thorne , eur .
j. c * 23 * ( 2002 ) 73 [ arxiv : hep - ph/0110215 ] .
k. j. eskola , h. honkanen , v. j. kolhinen , j. w. qiu and c. a. salgado , nucl .
b * 660 * ( 2003 ) 211 [ arxiv : hep - ph/0211239 ] .
a. d. martin , m. g. ryskin and g. watt , phys .
d * 70 * ( 2004 ) 091502 [ arxiv : hep - ph/0406225 ] .
v. a. abramovsky , v. n. gribov and o. v. kancheli , yad .
* 18 * ( 1973 )
595 [ sov . j. nucl .
phys . * 18 * ( 1974 ) 308 ] .
j. bartels and m. g. ryskin , z. phys .
c * 76 * ( 1997 ) 241 [ arxiv : hep - ph/9612226 ] .
a. d. martin , m. g. ryskin and g. watt , eur .
j. c * 37 * ( 2004 ) 285 [ arxiv : hep - ph/0406224 ] .
a. d. martin , m. g. ryskin and g. watt , eur .
j. c ( in press ) [ arxiv : hep - ph/0504132 ] .
s. chekanov _ et al . _
[ zeus collaboration ] , eur .
j. c * 25 * ( 2002 ) 169 [ arxiv : hep - ex/0203039 ] .
a. g. shuvaev , k. j. golec - biernat , a. d. martin and m. g. ryskin , phys .
d * 60 * ( 1999 ) 014015 [ arxiv : hep - ph/9902410 ] .
m. botje , qcdnum version 16.12 , zeus note 97 - 066 ( unpublished ) , available from ` http://www.nikhef.nl/\simh24/qcdcode/ ` .
j. breitweg _ et al . _
[ zeus collaboration ] , eur .
j. c * 7 * ( 1999 ) 609 [ arxiv : hep - ex/9809005 ] ; s. chekanov _ et al .
_ [ zeus collaboration ] , eur .
j. c * 21 * ( 2001 ) 443 [ arxiv : hep - ex/0105090 ] ; c. adloff _ et al . _
[ h1 collaboration ] , eur .
j. c * 21 * ( 2001 ) 33 [ arxiv : hep - ex/0012053 ] ; c. adloff _ et al . _
[ h1 collaboration ] , eur .
j. c * 19 * ( 2001 ) 269 [ arxiv : hep - ex/0012052 ] ; c. adloff _ et al . _
[ h1 collaboration ] , eur .
j. c * 13 * ( 2000 ) 609 [ arxiv : hep - ex/9908059 ] .
j. bartels , k. golec - biernat and h. kowalski , phys .
d * 66 * ( 2002 ) 014001 [ arxiv : hep - ph/0203258 ] ; h. kowalski and d. teaney , phys . rev .
d * 68 * ( 2003 ) 114005 [ arxiv : hep - ph/0304189 ] . | we study the effect of absorptive corrections due to parton recombination on the parton distributions of the proton . a more precise version of the glrmq equations , which account for non - linear corrections to dglap evolution , is derived .
an analysis of hera @xmath0 data shows that the small-@xmath1 gluon distribution is enhanced at low scales when the absorptive effects are included , such that a negative gluon distribution at 1 gev is no longer required .
desy 05 - 136 + ippp/05/41 + dcpt/05/82 + 8th august 2005 + * effect of absorptive corrections + on inclusive parton distributions * g. watt@xmath2 , a.d . martin@xmath3 and m.g .
ryskin@xmath4 + @xmath2 deutsches elektronen - synchrotron desy , 22607 hamburg , germany + @xmath3 institute for particle physics phenomenology , university of durham , dh1 3le , uk + @xmath5 petersburg nuclear physics institute , gatchina , st .
petersburg , 188300 , russia |
the existence of peculiar properties that differentiate pure non diagonal disorder models from disorder models including diagonal disorder has a long history .
probably , the first contribution along this line corresponds to dyson s work on a one dimensional phonon model published in 1953@xcite .
the existence of a divergent density of states at the band center of such disordered systems implying a divergent localization length was pointed in later works@xcite . since
then many works on this subject have been published ( a representative list can be found in ref . ( ) ) .
the existence of some kind of delocalization transition at the band center of such models certainly disturbs the widely accepted statement saying that the specific form of disorder does not matter in single parameter scaling theory@xcite .
a bipartite lattice is a lattice that can be divided into two sublattices such that the hamiltonian changes sign under a transformation that changes the sign of the wavefunction on one sublattice . when pure non
diagonal disorder is considered on a bipartite lattice , the electron - hole symmetry of the spectrum is not destroyed by disorder .
this property has important consequences as , for example , the existence of an eigenstate at @xmath1 for any disorder realization of a system constituted by an odd number of sites ( the spectrum shows @xmath2 pairs plus a state at @xmath1 ) .
recently , several works pointing to the exotic behavior of transport properties of quantum wires showing chiral symmetry have been published@xcite . for example , scaling of the conductance strongly depends on the parity of the number of channels along the wire .
also , related activity in field theory has produced several models demonstrating a delocalization transition in the vicinity of the zero energy state@xcite . in this paper , the scaling properties of the simplest disorder model preserving chiral symmetry have been carefully analyzed at the band center of quantum wires .
numerical simulation has been used to get the whole distribution of conductances . while the mean value of the conductance decays algebraically for wires of odd width , an alternative measure of the central value of distributions -the median- shows an exponential decay .
actually , the larger part of the conductance measurements has an exponentially small value for large wires .
therefore , although the conductance distribution is certainly peculiar at the band center , i would still use the term exponential localization when referring to the scaling behavior of the conductance of long wires .
results change when width and length of the wire coincide . in this case
, the numerical simulation presented in this paper uses a dot geometry point of view keeping the size of the contacts finite while dot area scaling proceeds .
the conductance distribution converges to a well defined limit that is compatible with the predicted critical behavior of the state at the band center .
it is shown that the analytic form used to fit conductance distributions comes from a new underlying wavefunction statistics describing the distribution of weight ratios of the @xmath1 wavefunction .
the format of the paper is as follows .
section ii defines a quite simple chain model that allows some analytical results and an unbound numerical simulation .
section iii gives a more general disorder model on finite rectangular clusters of the square lattice .
the way in which the conductance is calculated is presented in section iv .
numerical results are given in section v , first for wide disordered wires and , second , for square clusters .
the last section of the paper compiles the main conclusions reached by this numerical study of conductance scaling .
let me begin with a detailed description of the scaling properties of the conductance of a chain showing pure non diagonal disorder . in this case , sign changes of the hopping parameters do not matter and changes in their absolute value should be considered .
one case that allows some analytic results consists of a chain with hopping parameters that randomly take values of @xmath3 and @xmath4 with equal probability .
starting from the selfenergy of an ideal chain at its band center @xmath5 , successive selfenergies can be obtained through the disordered part of the chain by means of the usual iterative sequence ( see , for example , eqs .
( 7)-(10 ) in ref .
( ) ) : @xmath6 where @xmath7 takes values 1 or 2 with equal probability .
since selfenergy remains purely imaginary , a more convenient form is possible : @xmath8 the conductance @xmath9 of a sample is obtained from the selfenergy at the end of the disordered chain part @xmath10 : @xmath11 ^ 2 } = { { 4 ( -\sigma_n / i ) } \over [ 1+(-\sigma_n / i)]^2 } \qquad .
\label{chain3}\ ] ] eq .
( [ chain3 ] ) shows that the conductance varies between 0 and 1 as it corresponds to a one channel system .
actually , the conductance takes the value 1 only if @xmath12 .
repeated use of eq.([chain2 ] ) shows that the form in which the random hopping elements appear in the selfenergy expression is : @xmath13 therefore , a perfect transmission through the chain is obtained when the number of normal ( hopping equal to 1 ) and strong ( hopping equal to @xmath4 ) bonds of the numerator coincides with the corresponding numbers of the denominator . solving this simple combinatorial problem ,
one gets a probability @xmath14 ^ 2 \label{probability}\ ] ] for the peak of the probability distribution of the conductance at @xmath15@xcite .
notice that the distribution is a sum of delta functions since the hopping takes just two different values ; for example , the peak below the one at @xmath15 appears at @xmath16 and corresponds to @xmath17 or @xmath18 ( one extra strong bond either in the numerator or the denominator ) .
an asymptotic expansion for large enough chain lengths can be found for the sum in ( [ probability ] ) : @xmath19 the scaling behavior of this peak is enough to explain an inverse square root law for the _ mean value _ of @xmath9 given that other peaks of the distribution decay in a similar or faster way with the size of the chain .
this is an interesting behavior since it perfectly coincides with scaling predictions for wide wires of odd number of transversal modes@xcite .
numerical simulations can be done as precise as necessary for this simple model .
this fact allows a detailed comparison with theoretical predictions .
[ toy.ps ] shows the scaling of the conductance as described by its mean value .
root mean square deviation is about some tenths of the conductance unit . in any case , the average conductance at large sizes ( @xmath20 ) shows a power law scaling with the distance that is compatible with the inverse square root law given by eq.([law ] ) . according with this result , a disordered chain with pure non
diagonal disorder shows non
standard scaling at the band center .
nevertheless , alternative measures of the central value of the distribution restore to some way the usual exponential decay of one dimensional conductances , _ even _ in the presence of chiral symmetry .
for example , the geometric mean ] ] of the same conductance distributions shows a much more pronounced decrease with length ( see fig .
[ toy2.ps ] ) although the corresponding standard deviation is of the same order as the mean making doubtful its statistical relevance .
but there are other alternatives which do give a good description of the overall scaling of the distribution . both the median is the value @xmath21 for which larger and smaller values of @xmath22 are equally probable : @xmath23 ] or any definition based on the value of the integral of the distribution between 0 and an arbitrary upper limit @xmath24 flow to exponentially small values .
the physical meaning is clear in this case : half or more of the measures are exponentially small at large chain lengths .
actually , the precise scaling law for these central value alternatives is : @xmath25 where @xmath26 gives a measure of the exponential localization length . fig .
[ toy2.ps ] shows that fits according to this law are excellent over the whole length range .
the ultimate reason for such statistically disappointing results is simply the unusual size scaling of the distribution .
while the major part of the distribution below @xmath15 decays in an algebraic form , the weight of the distribution accumulates in an exponentially small region near @xmath0 ( fig .
[ mudry_hist.ps ] in section v illustrates graphically this behavior ) . in this way , the upper part of the distribution dominates the scaling behavior of mean , root mean square deviation , etc . , while accumulation at @xmath0 gives the behavior of central value definitions based on the integral of the distribution . in my opinion , these last definitions are better suited for characterizing the whole distribution than standard averages .
ultimately , one should look at the precise experimental protocol followed to get a value of the conductance before making predictions about the result of the measurements .
the lattice hamiltonian describing random hopping on a @xmath27 cluster of the square lattice is : @xmath28 where @xmath29 creates an electron on site @xmath30 , @xmath30 and @xmath31 are nearest neighbor sites , and @xmath32 is the hopping energy from site @xmath30 to @xmath31 .
it takes values 1 and -1 with equal probability .
let me refer to this model as the random hopping sign ( rhs ) model .
obviously , the square lattice can be divided into two sublattices such that atoms belonging to one of them hop only to sites belonging to the other sublattice when described by hamiltonian ( [ rhs ] ) .
therefore , this hamiltonian changes sign under a transformation that changes the sign of the electron operators on one sublattice .
consequently , the spectrum is symmetric relative to the band center at @xmath1 for any disorder realization , i.e. , for any values of the random variables of the model @xmath33 .
the model given by eq.([rhs ] ) is probably the simplest two dimensional model showing chiral symmetry .
many changes can be done to this model preserving chirality .
for example , absolute values of the hopping can fluctuate in addition to their signs or complex values of the hopping parameters can be considered if random magnetic fluxes are simulated .
the results described below are not sensible to any of these modifications of the disorder model .
numerical values are somewhat changed but trends remain exactly the same .
randomly generated samples of @xmath27 clusters are connected to ideal leads of width @xmath34 .
typically , @xmath35 in a wire geometry .
the conductance of the whole system is obtained using kubo formula@xcite within exact one - electron linear response theory .
computational details are given elsewhere@xcite .
let me just mention that the inversion of the hamiltonian matrix needed to get the green function of the wire can not proceed slab by slab as it is usually done within an optimized code .
numerical divergences take place owing to the existence of a true eigenstate at exactly @xmath1 for any piece of the system showing an odd number of sites ( the matrix @xmath36 is locally singular ) .
nevertheless , numerical calculation proceeds straightforward when pivoting over the whole hamiltonian matrix that includes the ideal leads is allowed .
the first aim has been the recovery of some important results obtained for wide wires by mudry , brouwer , and furusaki@xcite . in particular , the exotic dependence of the scaling law on the parity of the wire width is obtained for the present model ( a simplification of their random flux model ) .
open boundary conditions have been used to get conductances of stripes of fixed width and number of open channels ( the wire width is equal to the number of channels at the band center ) .
sample lengths have been varied from 99 to 1980 in steps of 99 . as many as @xmath37 samples are necessary to get good values of means and other central values of conductance distributions .
[ mudry1.ps ] shows the scaling law for two typical odd widths ( 9 and 19 channels ) in a log - log plot .
error bars are comparable to the symbols representing the conductance averages
. a power law fit to the numerical data is compatible with a mean conductance scaling proportional to the inverse square root of the sample length @xmath38 : @xmath39 this is precisely the scaling law obtained in ref .
( ) for quantum wires of an odd section .
notice that scaling proceeds smoothly without distinguishing odd and even wire lengths .
as noted by these authors , the variance of the conductance distribution is as large as its mean , making the mean a poor characteristic of the whole distribution . actually , the predicted relationship @xmath40 is accurately reproduced by the numerical data ( see filled symbols of fig . [ mudry1.ps ] and the dashed lines that are simply @xmath41 of the previous fits ) . on the other hand ,
when an even number of channels is studied , exponential scaling of the average conductance together with an exponentially small typical deviation is obtained in complete agreement with previous work@xcite .
a deeper understanding of scaling of the conductance mean value can be obtained from the analysis of the whole conductance distribution . here ,
conductances for @xmath42 samples of width @xmath43 and lengths @xmath44 , @xmath45 , and @xmath37 have been compiled and the corresponding histograms plotted in fig .
[ mudry_hist.ps ] .
it is clear that the probability for large values of the conductance diminishes as the length of the wire increases .
actually , since the plot is semilogarithmic the roughly equal separation between filled and empty circles for one side and between empty circles and diamonds for the other side , implies a power law decrease of the probability .
the weight that the probability distribution loses for large conductance values goes close to @xmath0 .
i have checked that the divergence of the probability at @xmath15 is proportional to @xmath46 while the divergence at the origin looks also algebraic but with an exponent starting close to @xmath47 for small disorder and decreasing towards @xmath48 as disorder increases .
although the conductance corresponding to ideal ( non disordered ) wires equals their widths ( 9 and 19 , in this case ) , fig .
[ mudry_hist.ps ] shows that conductance just reaches the value of 1 for disordered systems .
it seems that just one channel is effective .
therefore , it is tempting to compare these conductance distributions with the one corresponding to one random channel within the orthogonal universality class@xcite arguing that before localization the effect of disorder is just randomizing transport coefficients .
but the comparison is very bad since apart for small differences all theories show probabilities that continuously decrease from @xmath0 to @xmath15 .
in particular , the simplest result that applies to a random channel described by scattering matrices of the circular orthogonal ensemble is : @xmath49 that is quite different from the one obtained by numerical simulation .
there is a direct mathematical reason explaining this failure .
typically , several states contribute to the green function calculated at an arbitrary energy within the spectrum of a disordered system .
but this is not the case when the green function of a chiral system is calculated at @xmath1 which is an eigenenergy of the isolated system with an odd number of sites . in this situation , both the green function of the isolated finite system and the one corresponding to the extended system including the leads ( and related to the previous one by a dyson equation )
are dominated by the pole at @xmath1 , that is , are completely determined by the @xmath1 eigenfunction . in the next subsection , i will exploit this feature to analyze the conductance distribution as a consequence of a precise wavefunction statistics .
meanwhile , let us return to the theory of ref .
( ) to analyze the numerical results .
mudry , brouwer , and furusaki give the following expression for the conductance distribution of wires with an odd number of channels in the localized regime ( see eq .
( 4.10 ) of the second paper of ref .
( ) ) : @xmath50 ^ 2 \ } } \over { g \sqrt{1-g } } } \qquad , \label{p_law}\ ] ] where @xmath51 with @xmath30 the mean free path , @xmath38 the wire length , and @xmath34 the wire width ( which coincides with the number of channels at the bandcenter )
. a second scaling parameter @xmath52 that characterizes the disorder on a microscopic scale does not appear because it vanishes for the random flux model ( rfm ) and the model studied here ( see section iii ) is just a special case of rfm model@xcite .
when @xmath53 is small enough ( @xmath54 ) the distribution is very well described by a much simpler expression ( error smaller than one percent for @xmath55 ) : @xmath56 which allows the evaluation of the mean and the variance of the distribution : @xmath57 @xmath58 using the explicit form of @xmath53 ( eq.([cte ] ) ) an expression is obtained that can be used to fit the remaining parameter @xmath30 once an enough number of widths and/or lengths have been studied : @xmath59 fig .
( [ fit.ps ] ) shows a fit to the mean and variance values obtained for sets of @xmath37 randomly generated samples of length @xmath60 and width @xmath34 from three to 37 .
while eq.([teoria ] ) works reasonably good for large widths it fails near @xmath61 .
actually , an alternative fit by a linear law works sensibly better : @xmath62 consequently , the distribution given by eq .
( [ p_law ] ) will be used with @xmath63 to describe the results obtained by numeric simulation .
probability @xmath64 is shown in fig .
[ mudry_hist.ps ] as continuous lines .
although the three numerical distributions are nicely reproduced , accuracy is better for wider wires as expected from the theory ( @xmath65 is assumed in the theory leading to eq .
( [ p_law ] ) ) .
the divergence at @xmath15 is of the inverse square root form while the apparent @xmath66 non integrable divergence at the origin is regularized by the complex numerator which vanishes at @xmath0 . in conclusion ,
eq.([p_law ] ) is a very good description of the numerical data once the constant @xmath53 is properly estimated .
let us try an alternative way of characterizing the central value of distributions of the form shown in fig .
[ mudry_hist.ps ] .
previous experience with disorder models of this kind proves that the median is more robust than the mean for some distributions showing large @xmath9 variances but smaller @xmath67 variances , i.e. , when geometric mean does it better that the usual arithmetic mean@xcite .
unfortunately , the distribution of @xmath67 is also very broad in the present case . scaling results for the toy model of section ii suggest the use of the median . fig .
[ mudry2.ps ] gives the median scaling obtained for the previously collected conductance distributions .
in can be seen that median values are exponentially smaller than mean values at large wire lengths .
the fact that the median scales towards 0 can be inferred from the scaling of the histograms given in fig .
[ mudry_hist.ps ] although , the accumulation near exponentially small values of @xmath9 is not visible in the figure .
even the scaling law of medians is the same that describes the chain model ( see , eq .
( [ law2 ] ) ) .
alternatively , the exponential decay of the median as a function of the square root of the wire length can be inferred from eq .
( [ p_law ] ) once the distribution is written as a logarithmic normal distribution using the variable change described in the next subsection .
the practical meaning of the result is quite clear , more than half of the conductance observations are exponentially decreasing as the wire length increases . actually , the fraction of the samples showing exponentially small conductances increases with wire length because probability below @xmath15 decreases monotonically .
just the difference with a conventional one dimensional scaling ( as , for example , the one obtained for wires of even width ) comes from the power law decrease of the upper part of the probability distribution that should be compared with the exponential decrease characteristic of standard scaling .
while a standard study of the scaling of the conductance in 2d would imply the calculation of conductances of increasing @xmath68 samples connected to ideal leads of width equal to the square side @xmath38 , in this work i have used a dot setup specially designed for the study of just one conducting state@xcite . certainly , when the number of incoming channels is fixed by a point contact geometry , the only factor that affects the value of the conductance is the size ( area ) of the dot .
in this way , the presumably increase of the conductance due to wider contacts does not obscure the underlying scaling law strictly due to the increased size of the system .
two different limits are well known for large values of @xmath38 .
first , ballistic transport through the sample can occur as it happens in chaotic cavities .
this limit is described in a first approximation by scattering matrices of the circular orthogonal ensemble ( coe)@xcite .
roughly speaking , a conductance about @xmath69 per channel can be expected .
second , anderson exponential localization would imply an exponentially small value of conductances for large enough dot sizes .
this limit applies to diagonal disorder , for example . at the band center of a system with chiral symmetry ,
numerical simulation shows a behavior similar to the ballistic one ( see fig .
[ g0_dist.ps ] ) .
mean conductance converges to a well defined finite limit while the whole conductance distribution is _ perfectly _ described by eq .
( [ p_law ] ) . since now the system is not quasi onedimensional as wide wires are , one is forced to conclude that there should be deep general reasons for the validity of eq .
( [ p_law ] ) in this context .
let me briefly describe the numerical procedure .
the transmission between any couple of points within the dot is calculated and the corresponding conductance distribution obtained . to this end , two clean infinite chains ( the leads )
are attached through two arbitrarily chosen lattice sites within the disordered square sample ( the dot ) . for this geometry ,
the transmission from site @xmath70 to site @xmath71 is given by the following expression : @xmath72 where @xmath73 is the wavefunction at the band center . from this equation , @xmath74 , and
the conductance is between zero and one quantum unit for this numerical simulation . while the large number of transmission evaluations ( @xmath75 ) allows for a very precise calculation of the whole conductance distribution , the dependence of the conductance on the separation between point contacts is not given by the procedure .
the eigenstate at the band center is obtained by direct inversion of the schrdinger equation ( @xmath76 ) where @xmath77 is given by ( [ rhs ] ) .
once the @xmath1 wavefunction statistics is known , eq .
( [ transmission ] ) can be used to get a conductance distribution .
for example , if the porter - thomas form@xcite were valid : @xmath78 where @xmath79 being @xmath80 the number of sites , the conductance distribution would be given by : @xmath81 which can be integrated to give the final result : @xmath82 although probability @xmath83 reproduces some features of the conductance distributions shown in figs .
[ mudry_hist.ps ] and [ g0_dist.ps ] for the smaller cluster sizes ( for example , the square root divergence at @xmath15 ) , it is clearly non comparable to the accurate result given by eq.([p_law ] ) . what comes as a surprise is the fact that a logarithmic normal distribution of the wavefunction ratios @xmath84 exactly gives the conductance distribution proposed by mudry
et al._@xcite . that is , assuming @xmath85 being @xmath86 , the conductance distribution is given by @xmath87 this integral is easily solved giving @xmath64 of eq.([p_law ] ) with @xmath88 or equivalently , @xmath89 this result gives some clue over the complicate @xmath9 dependence that happens to appear in its distribution function .
numerical simulation ( see fig . [ lognormal.ps ] ) shows that eq.([lognormal_law ] ) accurately describes the wavefunction squared ratios of large disordered two - dimensional systems and , consequently , conductance distributions of the form given by eq.([p_law ] ) are valid in this case .
let us now discuss the scaling properties of the mean conductance at the band center .
conductance has been averaged over a significant number of randomly generated samples of increasing linear sizes ( @xmath38 ) keeping almost constant the total number of _
measurements_. fig .
[ g0.ps ] shows the results obtained by this numerical procedure .
the scaling of the mean conductance can be fitted to a model of the form : @xmath90 with a very good precision .
the asymptotic value corresponding to the infinite limit is 0.245 .
finite values of the conductance of two dimensional systems showing chiral symmetry have been predicted in a number of papers@xcite .
although , the present numerical simulation nicely supports these theories some caution must be used for two main reasons .
first , a somewhat practical reason .
results heavily depend on the fact that the green function of this problem is just given by only one particular state .
while the energy of this state is well defined theoretically , it could be difficult to make an experiment at just a particular energy .
previous authors on the subject have clearly shown that chirality is lost as soon as @xmath1 is left@xcite .
second reason is a bit more technical .
scaling properties have been obtained for clusters of an odd number of sites and , therefore , a state at @xmath1 .
present computational facilities do not allow to prove that scaling of clusters of an even number of states proceeds in the same way .
( note that in this case the green function should be recalculated for any new position of the point contacts since storing of the whole green function matrix of the isolated cluster is not possible ) .
nevertheless , i have checked that odd / even differences are minimal for small systems .
the _ universal _ conductance distribution at the band center of a two dimensional chiral system shown by the thick line in fig.([g0_dist.ps ] ) is probably the most original result in the paper .
it corresponds to the @xmath91 limit of eq.([2d_law ] ) which is obtained for @xmath92 in eq.([p_law ] ) but does not differ very much from the distributions corresponding to the simulated sizes ( @xmath93 ) plotted as three thin indistinguishable lines .
all three are accurately described by eq.([p_law ] ) with @xmath94 .
although obtained for point contacts , previous experience with wide wires proves that minor differences should be expected for wider but finite contacts@xcite .
let me insist in the idea that perfect agreement with the theoretical conductance distribution is also explained by a particular distribution of wavefunction weight ratios ( see eq.([lognormal_law ] ) ) that is confirmed by numerical simulation .
[ lognormal.ps ] give the distributions obtained for a small cluster size ( @xmath95 ) -higher maximum value- and for a larger size ( @xmath96 ) together with its corresponding gaussian ( dashed line ) .
the thick line gives the _ universal _ prediction for this statistical measure of the wavefunction .
consequently , i think that unusual phenomena as they have been described in the literature are quite compatible with present results for the conductance distribution@xcite .
the scaling properties of the conductance distribution of wide wires showing chiral symmetry has been analyzed by numerical solution of a simple model hamiltonian .
predictions of a power law decrease of the mean conductance have been confirmed . besides , the strange scaling properties of the distribution give rise to unusual statistical properties that have been described in the main text .
for example , while the slow algebraic decay of the mean conductance is due to few but very large fluctuations , it has been shown that almost all conductance measurements are exponentially small for large wires .
actually , an exponential law is obtained for the median of the distribution and for any generalization of the median based on the integral of the distribution .
nevertheless , exponential decay is not proportional to the sample length but to its square root . that is a significant deviation from standard localization theory .
surprisingly , statistical conductance properties of two dimensional systems connected to point contacts are described at the band center by the same distribution that describes wide wires .
it has been shown that this statistics appears due to the existence of an underlying gaussian distribution of the logarithm of the ratio of the wavefunction weights at @xmath1 .
the main difference with quasi one dimensional systems is that now both the mean and the median conductance seem to reach a finite limit .
the _ universal _ statistical distributions of conductance and wavefunction ratios have been obtained .
although they share some properties with the corresponding statistics of the circular orthogonal ensemble , it has been shown that they differ .
this numerical result calls for some theory able to explain the origin of such a simple statistics of the wavefunction at the band center of a chiral system .
this work has been partially supported by spanish comisin interministerial de ciencia y tecnologa ( grant pb96 - 0085 ) and direccin general de enseanza superior e investigacin y ciencia ( grant 1fd97 - 1358 ) .
a. altland and b. d. simons , nucl .
b * 562 * , 445 ( 1999 ) ; t. fukui , nucl . phys .
b * 562 * , 477 ( 1999 ) ; t. fukui , nucl .
b * 575 * , 673 ( 2000 ) ; s. guruswamy , a. leclair and a. w. w. ludwig , nucl .
b * 583 * , 475 ( 2000 ) ; m. fabrizio and c. castellani , nucl . phys .
b * 583 * , 542 ( 2000 ) ; a. m. garcia - garcia and j. j. m. verbaarschot , nucl .
b * 586 * , 668 ( 2000 ) ; b. klein and j. j. m. verbaarschot , nucl . phys .
b * 588 * , 483 ( 2000 ) ; t. guhr , a. mller - groeling and h. a. weidenmller , phys . rep . * 299 * , 189 ( 1998 ) .
if the complex probability distribution of selfenergies implied by eq.([chain4 ] ) were substituted by its geometric mean , a value of 1 would be obtained .
but this does not imply that perfect transmission ( @xmath15 ) or diverging localization length in the language of other authors is the average value of the relevant transport magnitude .
actually , both the distribution of selfenergies or lyapunov exponents are broad and can not be substituted by any central value of their distributions .
( ) further illustrates this point .
r. kubo , j. phys .
japan @xmath100 , 570 ( 1957 ) ; r. kubo , s.j .
miyake , and n. hashitsume , in _ solid state physics _ , edited by h. ehrenreich and d. turnbull ( academic press , new york , 1965 ) , vol .
17 , pp . 269364 .
h. u. baranger and p. a. mello , phys .
lett . * 73 * , 142 ( 1994 ) and phys .
b * 54 * , r14297 ( 1996 ) , v. n. prigodin , k. b. efetov and s. iida , phys
lett . * 71 * , 1230 ( 1993 ) , r. a. jalabert , j - l .
pichard and c. w. j. beenakker , europhys . lett .
* 27 * , 255 ( 1994 ) , e. mccann and i. v. lerner , j. phys . : condens .
matter * 8 * , 6719 ( 1996 ) .
a conventional scaling analysis like the one followed in ref .
( ) shows an almost constant conductance average close to @xmath102 quantum units of conductance for clusters sizes from @xmath103 to @xmath104 and both periodic and open boundary conditions .
r. gade , nucl .
b * 398 * , 499 ( 1993 ) ; r. gade and f. wegner , _ ibid .
_ * 360 * , 213 ( 1991 ) ; see also , s. hikami , m. shirai , f. wegner , nucl .
b * 408 * , 413 ( 1993 ) ; d. k. k. lee , phys .
b * 50 * , 7743 ( 1994 ) . a. w. w. ludwig , m. p. a. fisher , r. shankar , and g. grinstein , phys .
b * 50 * , 7526 ( 1994 ) ; c. mudry , c. chamon , x .-
wen , nucl .
b * 466 * , 383 ( 1996 ) ; i. i. kogan , c. mudry , and a. m. tsvelik , phys .
lett . * 77 * , 707 ( 1996 ) ; c. de c. chamon , c. mudry , and x - g .
wen , phys .
. lett . * 77 * , 4194 ( 1996 ) ; y. hatsugai , x .- g .
wen , m. kohmoto , phys .
b * 56 * , 1061 ( 1997 ) . | kubo formula is used to get the scaling behavior of the static conductance distribution of wide wires showing pure non diagonal disorder . following recent works that point to unusual phenomena in some circumstances , scaling at the band center of wires of odd widths has been numerically investigated . while the mean conductance shows a decrease that is only proportional to the inverse square root of the wire length , the median of the distribution exponentially decreases as a function of the square root of the length .
actually , the whole distribution decays as the inverse square root of the length except close to @xmath0 where the distribution accumulates the weight lost at larger conductances .
it accurately follows the theoretical prediction once the free parameter is correctly fitted .
moreover , when the number of channels equals the wire length but contacts are kept finite , the conductance distribution is still described by the previous model .
it is shown that the common origin of this behavior is a simple gaussian statistics followed by the logarithm of the @xmath1 wavefunction weight ratio of a system showing chiral symmetry . a finite value of the two dimensional mean conductance is obtained in the infinite size limit .
both conductance and the wavefunction statistics distributions are given in this limit .
this results are consistent with the _ critical _ character of the @xmath1 wavefunction predicted in the literature . |
pulsed electron spin resonance ( esr ) techniques have been developed and applied in many areas , including physics , chemistry and biochemistry @xcite .
another important application is quantum information processing ( qip ) @xcite .
systems in which an electron is coupled to nuclei through the hyperfine interaction are advantageous over conventional nuclear magnetic resonance ( nmr ) qip experiments .
the electron spin polarization can be transferred to the nuclear spins @xcite , and fast gate operations can be realized using only microwaves , if the hyperfine interaction has a strong anisotropic component @xcite .
while gate operations are done through the electron via microwave control , relatively long coherence times of the nuclei can be exploited for storing information @xcite .
thus the electron - nuclear hybrid spin system is a promising for developing advanced qip architectures @xcite .
such systems permit testing the ideas of quantum control and quantum error correction @xcite in a setting unavailable to classical simulations , opening a path to the development of large scale quantum devices based on spins .
the ability to generate accurate control pulses with arbitrary amplitudes and phases is required to realize a universal set of quantum gates with high fidelity . moreover , in other spectroscopic techniques such as double electron - electron resonance ( deer ) @xcite and solution - state 2d electron - electron double resonance ( 2d - eldor ) @xcite , arbitrary waveform generation is beneficial for creating desired excitation profiles @xcite . integrating an arbitrary waveform generator ( awg ) into pulsed esr spectrometers
has been reported in several previous works @xcite . in order to achieve precise coherent control ,
efforts have also been made to overcome pulse distortions due to limited bandwidth of the resonator and amplifier non - linearity @xcite .
however , quantitative characterization of unitary gate fidelities and detailed studies of the sources of infidelity in pulsed esr systems have been rarely addressed @xcite .
randomized benchmarking ( rb ) @xcite is a well - developed , scalable approach for estimating the average error probabilities of quantum gates apart from the state preparation and measurement ( spam ) errors . when comparing to the cp / cpmg method for quantifying single - qubit rotation amplitude and phase control errors applied in refs .
@xcite , rb has the advantage of being able to assess a broader class of operations than only @xmath3 pulses . in the work of ref .
@xcite , a method was proposed to characterize the control accuracy by measuring the excitation profiles of different pulses experienced by the spins . however , the accuracy of the measured excitation profiles is limited by the esr signal linewidth , and this method can not provide a well - defined unitary fidelity that describes the gate performance .
moreover , there are other important sources of error beyond the imperfection of the pulse excitation profile . on the other hand
, rb provides a quantitative characterization of gates in terms of the average fidelity of the unitary operation , which is a relevant quantity in the context of fault - tolerant qip .
furthermore , it is unclear how to quantify multi - qubit operations in both the cp / cpmg method in refs . @xcite and the transfer function method in ref . @xcite . therefore , rb is a more general characterization protocol that produces a quantity that can be directly compared to fault tolerance thresholds and to fidelities measured in other qip implementations .
rb has been applied in a wide range of qip implementations , including trapped ions @xcite , liquid state nmr @xcite , superconducting qubits @xcite , atoms in optical lattices @xcite , @xmath4p donor in silicon @xcite , and nitrogen - vacancy ( nv ) centres in diamond @xcite . however , as yet rb results in a conventional pulsed esr system have not been reported . in this work ,
we implement a single qubit rb protocol using pulsed esr at x - band on a sample of gamma - irradiated fused quartz .
the stable , spin-1/2 defect is an unpaired electron at an oxygen vacancy . in order to have the flexibility and precision of control necessary for qip with arbitrarily shaped optimal control pulses ,
the spectrometer was custom built and includes an awg and i - q modulator for pulse generation , and a specially designed loop - gap resonator ( lgr ) for efficient , broadband control that accomodates small samples .
one challenge with pulsed esr that is not usually encountered in nmr is that the bandwidth of the microwave resonator that interacts with the sample is comparable to ( or narrower ) than the desired control bandwidth . moreover , impedance matching of all elements in the pulse generation , amplification and transmission train is more challenging at microwave frequencies . here , we identify and partially correct sources of gate error due to pulse distortions ; these distortions arise both due to the finite resonator bandwidth and to hardware imperfections
. additionally , we narrow the distribution of local fields that lead to inhomogeneous dephasing ( @xmath5 ) by applying a selection sequence , and this demonstrates the dominant role of the @xmath5 process as a source of incoherent gate error .
the initial rb result of @xmath6 error probability per gate is significantly reduced to about @xmath7 after identifying and mitigating several sources of error , with the lowest error rate being dominated by @xmath5 dephasing .
our work provides a first benchmarking of experimental unitary gate fidelities in a conventional pulsed esr system .
this result can be compared , in an unbiased way , with the gate fidelities obtained in other qip implementations .
the use of rb in concert with simulations and the selection sequence sheds light on the contributions of different error sources - @xmath8 , @xmath9 and @xmath5 processes , @xmath10 field inhomogeneity , as well as unitary errors .
this is a powerful diagnostic tool that should be broadly useful in esr applications that increasingly rely on sophisticated pulse sequences that demand high precision ; such sequences will benefit from pulses with high unitary fidelities , i.e. pulses that work as desired , independent of the input state .
schematic of the home - built x - band esr spectrometer .
a signal generated from the microwave source is mixed at the i - q modulator with @xmath11 and @xmath12 phase - shifted components of a shaped pulse from the arbitrary waveform generator .
the i - q modulator outputs the shaped pulse at upconverted frequency , and the pulse phase is accurately controlled by a method explained in the text . a twt amplifier combined with a pre - amplifier and an attenuator provides an output power up to 500 w. the amplified pulse is transmitted to the loop - gap resonator containing the sample . in the receiver
, the esr signal is mixed with the reference frequency and downconverted to an if signal , which is digitized by a fast oscilloscope.,scaledwidth=45.0% ] a schematic of the custom x - band pulsed esr spectrometer is depicted in fig .
[ fig : esrcircuit ] .
the microwave source ( rohde and schwarz smf100a ) provides a continuous wave ( cw ) output at @xmath13 ghz , and features low phase noise and output power up to @xmath14 dbm . in order to generate arbitrary shapes such as gradient ascent pulse engineering ( grape ) pulses @xcite
, we use a single - sideband ( ssb ) upconversion technique @xcite with an i - q modulator ( marki iq-0714lxp ) as the ssb mixer .
a @xmath15 gs / s arbitrary waveform generator ( awg , tektronix awg5014b with a memory expansion to 32 mpts ) provides both @xmath11 and @xmath12 phase - shifted pulse inputs from two output channels , with intermediate frequency ( if ) @xmath16 ranging from 150 - 400 mhz . applying if signals of the same amplitude but phase - shifted by @xmath17 at the i and q ports suppresses the phase error of the output pulse due to the non - linear power responses of the i and q mixers . the phase of the output microwave pulse is then controlled by the if phase generated by the awg , accurate up to 1 part in 16384 .
the i - q modulator outputs only the upconverted signal at @xmath18 , and the lower sideband is suppressed .
the signal is amplified by a travelling wave tube ( twt , applied systems engineering model 117 ) amplifier before being transmitted to the resonant sample cavity . although the maximum output power of the twt is @xmath191 kw , we typically only require tens of watts in these experiments . prior to the twt , a variable attenuator and a low - gain solid - state amplifier ( miteq afs3 - 08001200 - 10 - 10p-4 )
are used in order to adjust the input power level .
the output of the twt is directed by a circulator and travels to the loop - gap resonator ( see sec .
[ sec : lgr ] ) that contains the sample , where the pulse produces an oscillating magnetic field @xmath10 .
the circulator has a peak power limit of 500 w , hence the twt input is adjusted so that its output will not exceed this value .
the esr signal is directed by the circulator to a low - noise preamplifier ( miteq amf-5f-08001200 - 09 - 10p ) and then to the receiver ( miteq arm0812lc2c ) , where it is mixed with the reference frequency @xmath20 and downconverted to @xmath21 .
the receiver includes a second stage of low - noise amplification .
a diode switch ( advanced technical materials s1517d ) is used to protect the receiver from damage during pulsing , and is controlled by a marker channel of the awg .
a diode limiter protects the switch , which is also easily damaged at high microwave power . after the switch
, a high - pass filter removes switching transients .
finally , the esr signal is digitized using a fast oscilloscope ( lecroy wavepro 715zi ) and is recorded on a computer for further processing . the design of our loop - gap resonator ( lgr ) was adapted from ref .
@xcite , with a resonance frequency close to 10 ghz .
it is a lumped - element resonator with inductance and capacitance determined by its geometry .
the advantages of the lgr compared to conventional cavity resonators is a larger filling factor , large @xmath22 per square root power , good @xmath10 uniformity over the sample volume and larger control bandwidth ( with shorter pulse ringdown times ) due to a lower quality factor @xmath23 @xcite .
figure [ fig : lgr ] shows a schematic of the two - loop , one - gap resonator employed here .
the loops and the gap are cut by wire electric discharge machining from a 9.6 mm @xmath24 5.4 mm @xmath24 2.5 mm oxygen - free copper block .
the loop radii are 1.2 mm and 0.6 mm , and the gap is 0.1 mm wide and 3 mm long . the geometry of the lgr was chosen for a 10 ghz resonance frequency and @xmath25250 .
the microwave electric field is the strongest in the gap while the magnetic field is strongest in the smaller ( sample ) loop .
( a ) schematic of the loop - gap resonator .
it has two loops and one gap , and is made of copper .
the geometry is designed to give a resonance frequency close to 10 ghz , and a quality factor @xmath26 .
the magnetic and electric field directions are shown in the side view .
the sample is placed inside the smaller loop .
( b ) high frequency structural simulator ( hfss ) simulations of field distributions at the resonance frequency , with input power 1 w. the electric field is the strongest in the gap while the magnetic field is the strongest in the loop . , scaledwidth=40.0% ] the resonator is placed in a rectangular copper enclosure of dimensions 9.4 mm @xmath24 9.6 mm @xmath24 5.4 mm .
the lowest resonance frequency of the box is well above 12 ghz , so it does not interfere with the lgr resonance . the inner conductor of a coaxial cable is formed into a one - turn loop and brought close to the larger of the two lgr loops , where it inductively couples the microwave signal to the lgr .
the outer conductor of the coaxial cable is copper , and the inner conductor is silver - plated copper .
impedance matching is done by fine adjustment of the position of the lgr relative to the coupling inductor .
the @xmath27 control field seen by the sample does not exactly match the desired pulse shape due to non - linearities in the pulse generation and amplification , and due to the finite bandwidth ( @xmath28 ) of the resonator . in order to identify and correct these errors ,
we insert a second , smaller coaxial cable terminated with an inductive loop as a pickup coil , as shown in fig .
[ fig : probe ] .
this pickup coil measures the microwave field in the vicinity of the sample .
the signal from the pickup coil is demodulated in the same way as an esr signal , by mixing with the reference frequency .
the pulse measured by the pickup coil is compared with the desired pulse , and the input waveform adjusted to minimize the difference between the two .
this reduces pulse error substantially ; however , we note that this procedure assumes the transfer function of the pickup coil is flat over the control bandwidth , and that the pickup coil only couples to the resonant mode and does not itself introduce new modes .
we take these assumptions to be approximately true in practice ( violations are probably small enough to be ignored in this work , but can be explored in future work ) .
schematic of the custom - built esr probe .
the two - loop , one - gap resonator is placed in a rectangular copper box , and the box is placed in a cylindrical copper enclosure which makes contact with cold helium vapour when operated cryogenically .
the cylindrical enclosure is sealed using indium wire and screws , and is evacuated .
the microwave transmission line is coupled to the lgr by an inductive loop .
tuning and matching of the lgr resonance are accomplished by fine adjustment of the lgr position relative to the coupling inductor .
a second coaxial cable is terminated with a pickup coil to detect the microwave field in the vicinity of the sample loop during a pulse.,scaledwidth=40.0% ] the rectangular enclosure containing the resonator is held inside a cylindrical probe - head made of copper , with 28 mm inner diameter and 85 mm inner depth .
a schematic of the probe is illustrated in fig .
[ fig : probe ] .
it is designed to fit into the oxford instruments cf935o continuous - flow cryostat , and can be cooled to slightly below 4 k by pumping on the helium space .
experiments at cryogenic temperatures have been carried out with this setup , and electron spin polarizations following the expected curie law were observed down to about 5 k. however , the experiments reported in this paper were carried out at room temperature , since our goal here is to characterize the precision of spin manipulation by the microwave control field . in future
work we plan to perform rb experiments at low temperature and on multi - qubit samples .
the strategy of rb is to average over a set of random sequences of computational gates , with each sequence followed by a reverse ( recovery ) operation that makes the entire evolution the identity operator , in the absence of gate error .
the computational gate set should be chosen such that it forms a depolarizing channel upon averaging , which is true as long as the errors are mostly gate - independent @xcite .
the depolarizing parameter @xmath29 of the averaged channel @xmath30 is given by : @xmath31 and is related to the average gate fidelity . here
@xmath32 is the dimension of @xmath33 .
averaging over the full unitary group provides a benchmark for control fidelity over the complete set of quantum gates .
however , since the unitary group is a continuous set with the number of parameters increasing exponentially in the number of qubits @xmath34 , generating an arbitrary unitary operator is exponentially hard with increasing @xmath34 . therefore , it is more practical to use a discrete set of gates .
it is desirable that the operations can be efficiently ( classically ) simulated and can be generated from a small set of elementary one and ( in the general case ) two - qubit gates .
the clifford group , denoted as @xmath35 , is a set that is attractive for rb for several reasons .
first , a universal gate set can be generated using @xmath35 with the addition of a so - called magic state together with measurement in the computational basis , which means that there exists a universal quantum computation model in which all necessary gates can be drawn from the clifford group @xcite .
furthermore , benchmarking the clifford gates is useful since most encoding schemes for fault tolerant qip are based on stabilizer codes , in which the error correction is performed using clifford gates .
we followed the single - qubit rb protocol implemented in a liquid - state nmr experiment presented in ref .
the single qubit clifford randomization is equivalent to a randomization using the 48 operations parametrized as @xcite @xmath36 in this protocol , the @xmath17 rotations ( which belong to the symplectic group and are denoted as @xmath37 in eq .
[ eq : cliffop ] ) are computational operations while the pauli operations ( which belong to the pauli group and are denoted as @xmath38 in eq .
[ eq : cliffop ] ) serve to toggle the pauli frame and thus depolarize the noise .
@xmath39 realizes an operation in the clifford group . to quantify the average error per gate , randomly chosen sequences of @xmath40 clifford operations in eq .
( [ eq : cliffop ] ) are applied to a fixed , known initial state .
a natural choice for the initial state is the thermal state , which can be simply written as @xmath41 since the identity part of the density matrix is irrelevant in both unitary evolution and in measurement .
the initial state is tracked through a sequence of clifford operations ( eq . [ eq : cliffop ] ) , and the sequence is truncated at different lengths @xmath42 to measure the fidelity decay curve . at each truncation @xmath43 , a recovery gate is chosen at random to return the state to either @xmath44 .
then , a read - out pulse sequence that contains a @xmath17 rotation pulse , a 700 ns delay and a @xmath3 rotation pulse , is used to generate a spin echo .
the amplitude of the spin echo in the time domain , equivalent to the integrated intensity of the corresponding absorption peak in the frequency domain , is the quantity we measure . in the experiment ,
we pre - calculate the final state prior to the spin echo measurement , and choose the phase of the @xmath17 read - out pulse so that the positive eigenvalue is always measured .
the quantum circuit implementing a particular series of clifford operations is shown in fig .
[ fig : bmcircuit ] . a particular realization of a clifford gate sequence in the rb protocol . @xmath45 and @xmath46 indicate pauli and computational operations , respectively , and @xmath47 is the recovery gate that brings the state to either @xmath48 or @xmath49 .
the sequence is truncated at some length @xmath43 , and the final state is measured by the amplitude of the electron spin echo after a read - out pulse sequence , as described in the text .
the electron spin echo measurement part is denoted by ese in the figure.,scaledwidth=45.0% ] the single qubit rb protocol can be summarized as follows @xcite : 1 .
choose a maximum number of clifford operations @xmath40 , and a set of random integers @xmath50 that is the length of a truncated pulse sequence ( i.e. a subsequence ) .
the number of elements in @xmath43 is denoted @xmath51 .
2 . generate @xmath52 random sets of @xmath40 computational gates , and truncate each sequence at length @xmath53 .
3 . for each subsequence of length @xmath54 ( total @xmath55 subsequences ) , do the following : 1 . generate @xmath56 random sets of pauli gates of length @xmath57 .
2 . interleave computational gates with the @xmath58 pauli gates .
calculate recovery operation and insert it after @xmath59 pauli gate .
4 . insert final pauli randomization gate before the spin echo detection .
calculate the phase of the spin echo read - out pulse that yields the positive eigenvalue of @xmath60 .
4 . evolve the known initial state under each of the @xmath61 number of random pulse subsequences .
measure the electron spin echo amplitude at the end of each subsequence and compare with the same measurement on a reference state .
6 . for each @xmath54
where @xmath62 , calculate the average remaining spin magnetization along @xmath60 by averaging over @xmath63 subsequences .
plot the average remaining signal as a function of @xmath43 , and fit the curve to @xmath64 , where the average gate error is @xmath65 .
the state preparation and measurement ( spam ) error is absorbed in the constant @xmath66 .
as indicated in steps 6 and 7 , the average gate error is calculated from the decay rate of @xmath67 as a function of the number of gates @xmath43 .
the single qubit system employed for the benchmarking experiment is gamma - irradiated fused quartz @xcite , a paramagnetic sample in powder form where the primary defect is an unpaired electron at an oxygen vacancy .
we obtained @xmath68 @xmath69s , @xmath70 @xmath69s , and @xmath71 ns from esr experiments at room - temperature .
@xmath72 is determined from the linewidth of the thermal - state esr spectrum in the frequency domain .
the @xmath72 line - broadening here is mainly due to the anisotropy of the @xmath73-value , which produces a powder pattern of width close to 5 mhz , and also to inhomogeneity of the applied static magnetic field .
different spins within the sample volume also experience different nutation frequencies due to the inhomogeneity of the applied microwave field ( @xmath10 ) in the resonator loop . to obtain the @xmath10 distribution profile
, we measure the time - domain rabi oscillations .
simulations show that @xmath5 does not contribute significantly to the rabi decay envelope for this sample , and the decay is dominated by @xmath10 inhomogeneity . the insensitivity of the rabi oscillations to @xmath72 noise agrees with the intuition that the continuous pulse partially refocuses the dynamics due to static field inhomogeneity .
therefore , the fourier transform of the time - domain rabi oscillations gives a good estimate of the @xmath10 distribution over the sample .
this distribution is confirmed by comparing experimental and simulated rabi decays , which allows us to refine the estimate for the @xmath10 distribution profile .
the esr line - broadening due to @xmath72 and the @xmath10 distribution profile are both shown in fig .
[ fig : dist ] . using the decoherence parameters and the @xmath10 inhomogeneity data , we simulated a rb experiment with @xmath74 , @xmath75 giving a total of 98 sequences , and @xmath761 , 2 , 7 , 9 , 10 , 12 , 14 , 18 , 20 , 21 , 25 , 28 , 32 , 57 , 60 , 66 , 74 , 97 , 110 , 128@xmath77 chosen at random .
both computational and pauli gates are realized by 35 ns long gaussian microwave pulses , yielding clifford operations that are 70 ns in total length .
figure [ fig : bmsim ] summarizes the simulation results under four different conditions : ( 1 ) no @xmath10 inhomogeneity and no @xmath72 effect , ( 2 ) with @xmath10 inhomogeneity and no @xmath72 effect , ( 3 ) with @xmath72 effect and no @xmath10 inhomogeneity , and ( 4 ) with both @xmath10 inhomogeneity and @xmath72 effects . @xmath8 and @xmath9 processes are included in all four simulations by using a lindblad model , while the local field inhomogeneity effects are included in ( 2)@xmath19(4 ) using the method of weighted averaging over multiple simulations . in the absence of any local field inhomogeneities ,
the simulated error per gate is @xmath78 .
since the controls fields are ideal in this case , the imperfection is solely due to @xmath8 and @xmath9 processes .
when the local field inhomogeneities are included , the error per gate increases to @xmath79 for @xmath10 , @xmath80 for @xmath5 and @xmath81 for both @xmath10 and @xmath5 .
thus , the simulations clearly show that the @xmath5 effect contributes more significantly to the gate error rate than @xmath10 inhomogeneity for this sample .
this is not surprising , since the @xmath10 field distribution is narrower than the static field distribution ( see figures [ fig : t2star ] and [ fig : rfdist ] ) .
simulated rb results on the single qubit ( fused quartz ) spin system , plotted in semi - log scale .
the lower x - axis is the number of gates and the upper x - axis gives the actual sequence time .
the expectation value @xmath82 is measured after application of @xmath43 randomized gates and fitted using @xmath64 .
the expectation value @xmath82 for each value of @xmath43 is the result averaged over @xmath83 random sequences .
the simulations use experimentally measured values of @xmath8 , @xmath9 , and @xmath72 at room - temperature , which are 160 @xmath69s , 5 @xmath69s , and 60 ns , respectively .
all curves include the @xmath8 and @xmath9 processes of the electron , through a lindblad formalism .
the @xmath72 and @xmath10 distributions are included as follows : green ( @xmath84 ) indicates no distributions ; black ( @xmath85 ) indicates @xmath10 only ; red ( @xmath86 ) indicates @xmath72 only ; blue ( @xmath87 ) indicates both @xmath10 and @xmath72 .
the size of the error bars represents standard error of the mean averaged over the 98 sequences.,scaledwidth=47.5% ] with no optimization of the microwave pulses beyond simple power calibrations , rb was implemented experimentally in the quartz single qubit system .
the error per gate was @xmath88 , indicating a relatively poor level of quantum control . in the following section ,
we discuss identification and mitigation of pulse distortions and other factors that allow the gate error rate to be significantly reduced .
the pickup coil introduced in sec .
[ sec : probe ] directly measures the microwave field in the vicinity of the sample .
the largest pulse imperfection revealed by the pickup coil was a phase transient error .
here we use a @xmath89 ns gaussian pulse as an example .
the pulse is applied along the rotating frame y - axis , so that the amplitude of the imaginary component ( x - axis component ) of the pulse is zero . however , as shown in fig .
[ fig : ptc1a ] , the measured output pulse has a noticeable phase transient with nonzero imaginary component .
this transient effect leads to undesired spin dynamics as shown in fig .
[ fig : ptc1b ] .
figure [ fig : ptc1b ] shows the rotation of the spin magnetization vector as the power of the gaussian pulse is varied .
the y - axis pulse is applied to the spin state @xmath60 , and it is followed by a 700 ns delay and a @xmath3 pulse to form a spin echo signal .
the solid and dashed curves represent real ( x ) and imaginary ( y ) components of the measured spin echo amplitude . in the absence of phase transients ,
the imaginary part should be zero for all powers as the spin magnetization vector remains within the x - z plane .
however , a nonzero y - component in fig .
[ fig : ptc1b ] indicates that the spin is rotated out of the x - z plane by the phase transient . in order to suppress this pulse error
, we designed an input pulse with an imaginary part of equal amplitude but opposite sign to cancel the transient .
the phase transient corrected ( ptc ) pulse is designed as follows .
first , we take the imaginary part of the pulse measured by the pickup coil , and digitize it so that the time resolution of the measured pulse matches the awg sampling rate .
we then multiply the imaginary part of the digitized pulse by @xmath90 to negate the phase error .
however , the digitization step can introduce a scaling error in the pulse amplitude , and thus the multiplication by @xmath90 may not exactly cancel the unwanted imaginary component .
we find that introducing a scaling factor of @xmath91 to the ptc pulse gives the best result in minimizing the phase transient effect .
figure [ fig : ptc2a ] demonstrates that using the ptc pulse , the phase transient is significantly reduced . as shown in fig .
[ fig : ptc2b ] , the deviation of the spin evolution trajectory from the x - z plane is strongly suppressed when the same experiment is conducted with the ptc pulse .
it should be noted that the real component of the pulse is asymmetric in time , with a slowly decaying tail ; this is due to the finite bandwidth of the resonator .
however , since the integrated pulse area can be calibrated with high accuracy to yield the desired spin rotation , we only need to correct the distortion in the imaginary part of the pulse .
it should also be mentioned that due to the limited resonator bandwidth there is a trade - off between the pulse length and the ability to correct the phase transient .
we find that gaussian pulses of 35 ns length ( hence clifford gates of 70 ns length ) offer the best compromise in the particular system reported here .
another significant contribution to the gate error was the phase and amplitude droop of the twt microwave amplifier .
we found this error to be strongly dependent on the amplifier unblanking time delay .
figure [ fig : bd300n ] shows a series of phase transient corrected @xmath89 ns gaussian pulses , as measured by the pickup coil , when the twt unblanking delay is set to @xmath92 ns .
it is clear from the figure that both amplitude and phase are unstable until about 2 @xmath69s after the blanking is turned off ( at t=0 in fig .
[ fig : bd300n ] ) . figure [ fig : bd2u ] shows that when we use a 2 @xmath69s unblanking time , the amplitude and phase stabilities are much better . after the phase transient effect is corrected and
the twt unblanking delay is set to 2 @xmath69s , the rb experiment was carried out again , and the error per gate was found to be @xmath93 ( @xmath84 symbol in fig .
[ fig : bmresults ] ) .
this error rate is still about an order of magnitude larger than the rate due solely to @xmath8 and @xmath9 processes , as predicted by simulations .
a large part of the remaining error in the experimental value is due to ensemble inhomogeneity effects , which we describe in the next section .
the simulation results shown in fig .
[ fig : bmsim ] indicate that the average gate fidelity is limited by local field inhomogeneities : the static distribution of larmor frequencies ( @xmath5 ) and the inhomogeneity of the microwave field @xmath10 across the sample volume .
the @xmath10 distribution is relatively narrow , and we observed from simulations that the @xmath5 effect contributes more strongly to the error rate . to test this experimentally , we designed a pulse sequence to effectively narrow the static field distribution .
one approach to overcome field inhomogeneity is to use composite pulses @xcite .
however , useful composite pulses are typically much longer than simple pulses . given the timescales of @xmath5 and @xmath9 in our system , composite pulses do not improve gate fidelities , but worsen them .
for example , at least three simple pulses are required for a spin inversion ( @xmath3-rotation ) composite pulse that is robust to the @xmath72 field inhomogeneity @xcite , meaning that the total pulse duration of either a pauli or a computational gate is @xmath94 ns=@xmath95 ns . a simulation using the @xmath95 ns composite pulses shows that even in the absence of the field inhomogeneities , where the error solely comes from @xmath8 and @xmath9 processes , the error per gate in the rb experiment would be @xmath96 .
thus , composite pulses were not used in our experiments .
the approach taken here is to select a subset of the spin ensemble within a narrower static field distribution , similar to the idea of rf selection that has been used successfully in nmr qip @xcite .
the selection sequence is constructed as follows .
first , a @xmath97 ns grape @xcite pulse ( with 400 time points ) rotates spins that experience the on - resonance field by @xmath98 around the x - axis , with a simulated unitary fidelity ( hilbert - schmidt ( hs ) norm ) of @xmath99 . for the off - resonance spin packets , the unitary fidelity quickly decreases with offset .
thus , mostly on - resonance spins are left pointing along the z - axis after the grape pulse .
next , we wait for a duration @xmath100 in order to dephase the transverse component of the off - resonance spin packets pointing along other directions in the bloch sphere . by repeating these two steps , but keeping the total sequence time short
compared to @xmath8 , an initial state is prepared that corresponds to a narrower esr linewidth .
ideally , the dephased off - resonance spin packets will be in a fully unpolarized state and not contribute to subsequent spin signals .
we find that repeating the sequence four times and using a delay equal to @xmath9 provides the best narrowing of the esr linewidth while retaining a signal level well above the detection noise floor . in this post - selection state , @xmath72
is extended to more than twice its original length , as shown in fig .
[ fig : distsel ] .
note that the selection sequence does not necessarily improve the @xmath10 field homogeneity . in order to reduce @xmath10 field inhomogeneity
, spin packets from localized region of the sample need to be selected .
we do not see any significant improvement in the @xmath10 distribution after this sequence .
nevertheless , we show in the next section that this selection sequence does indeed improve the average gate fidelity , suggesting that the @xmath5 process is a dominant source of gate error .
summary of experimental results plotted in semi - log scale .
the lower x - axis is the number of gates and the upper x - axis gives the actual sequence time .
the expectation value @xmath82 is measured after @xmath43 randomized gates and fitted using @xmath64 .
the expectation value @xmath82 for each value of @xmath43 is the result averaged over @xmath83 random sequences . the initial experiment , before making the improvements described in the text ,
is denoted as ` none ' in the legend .
the average error rate is above @xmath101 ( @xmath102 ) .
the triangles ( @xmath87 ) indicate the result after setting the twt unblanking delay to @xmath103 @xmath69s ( denoted ub ) , where the error probability is reduced to @xmath104 . adding phase transient correction ( denoted ptc ) , the error rate
is further reduced to @xmath105 ( @xmath84 ) . finally , adding the selection sequence ( denoted sel ) to narrow the static field inhomogeneity , the gate error decreases to @xmath106 ( @xmath85 ) .
the open circles are the results of the rb simulation .
the simulation includes the selection sequence , the @xmath8 , @xmath9 , @xmath72 values measured in experiment , and the @xmath10 inhomogeneity distribution ( but otherwise assumes perfect pulses ) .
the simulation should be compared to the final experimental data with @xmath106 error rate .
error bars represent the standard error of the mean averaged over the 98 sequences .
the horizontal dashed line ( black ) corresponds to the experimental noise floor.,scaledwidth=47.5% ] the rb experimental results under several different conditions are plotted in fig .
[ fig : bmresults ] .
the semi - log plot shows the expectation value of @xmath60 measured after applying @xmath43 clifford gates .
the depolarizing constant as well as the error per gate can be calculated from the decay rate of the expectation value as a function of the number of clifford operations .
the initial experimental rb data , without any improvements , is plotted ( @xmath102 symbol ) and corresponds to an error probability of 6.7@xmath107 .
changing the twt unblanking delay ( @xmath87 symbol ) results in an error probability close to 4@xmath107 .
both of these give decay curves that deviate substantially from a single - exponential fit after about 30 gates .
hence , we only use the first few points to fit to a single exponential and extract the error rate . further control improvement is achieved by employing phase transient corrected pulses ( @xmath84 symbol ) . combining the longer unblanking delay with
phase transient corrected pulses , the error per gate is reduced to @xmath93 .
however , the non - exponential behaviour of the decay curve is still observed .
note that this non - exponential behaviour was seen in the simulated rb results when @xmath72 is included ( see the lower two curves in fig .
[ fig : bmsim ] ) .
this behaviour can be explained by the @xmath72 field distribution over the spin ensemble , which is a form of incoherent error and leads to a sum of exponential decays ( see appendix ) .
note that this type of incoherent error is not restricted to spatial ensemble systems , but also applies to single quantum systems when a parameter ( e.g. local field ) fluctuates between the many experiments that must be averaged to measure an observable @xcite . after incorporating the selection sequence to effectively increase @xmath72
as described previously , the non - exponential behaviour is reduced greatly and the decay curve fits well to a single exponential curve up to the largest number of gates ( 128 ) performed . with the selection sequence , together with the longer unblanking delay and ptc pulses , the error per gate is reduced to @xmath108 ( @xmath85 symbol in fig .
[ fig : bmresults ] ) .
this is nearly an order of magnitude improvement compared to the initial result of @xmath109 .
simulation of the rb protocol including the selection sequence , and using experimentally determined values of @xmath8 , @xmath9 , @xmath5 and the @xmath10 distribution , is also shown in fig .
[ fig : bmresults ] ( @xmath86 symbol ) .
the simulated error per gate is @xmath110 , which is in good agreement with the experimental value .
this agreement suggests that the error due pulse distortions has been reduced to the same order of magnitude as the error from intrinsic noise .
moreover , the improvement achieved by incorporating the selection sequence implies @xmath5 effect is a dominant error source .
therefore , although the standard rb protocol is not designed for explicitly distinguishing error sources , we are able to infer which mechanisms are dominant by comparison with simulation and by exploiting the selection sequence .
we constructed an x - band pulsed esr spectrometer based on arbitrary waveform generation , a special loop - gap resonator and a pickup coil for monitoring the microwave field close to the sample .
the rb protocol for the single qubit clifford gates was used to characterize the precision of control over electron spin dynamics .
the initial results indicated a relatively poor @xmath111 average gate fidelity .
active correction of phase transients together with stabilization of the microwave amplifier improved the gate fidelities significantly . by additionally selecting spin packets with a narrowed distribution of static fields , an average fidelity of @xmath2 was achieved .
comparison between experiment and simulations suggest that the remaining errors due to pulse imperfections are no larger than the intrinsic noise , i.e. noise due to inhomogeneous dephasing . in a sample with longer @xmath5 , @xmath9 and @xmath8 values ,
the extrinsic errors due to microwave imperfections could again dominate .
we expect that a suitable solid - state microwave amplifier should outperform the twt amplifier in both noise figure and amplitude / phase stability , and it would be interesting to use rb for a side - by - side comparison of these amplifier technologies in the context of quantum control .
rb protocols that can separate unitary control errors from incoherent errors due to @xmath5 , @xmath9 and @xmath8 effects have recently been developed @xcite , and can further distinguish sources of quantum gate error in future work .
the rb of single qubit control performed here is a milestone towards more complex multi - qubit esr qip experiments . in future work
, we are interested in characterizing microwave control during cryogenic operation , as high electron spin polarization is desirable .
rb can also be used to characterize control in multi - qubit systems , e.g. an electron spin coupled to one or more nuclear spins by the hyperfine interaction @xcite .
high fidelity control in these systems can be achieved with optimal control pulses , so a clear next step is to perform rb on both single qubit and multi - qubit clifford gates based on grape @xcite pulses .
if sufficiently high fidelities are achieved , then such hybrid electron - nuclear spin systems could test quantum algorithms such as heat bath algorithmic cooling @xcite and quantum error correction @xcite , in particular , multiple - round quantum error correction which is of fundamental importance for building a large scale quantum computer .
this research was supported by nserc , the canada foundation for innovation , cifar , the province of ontario , industry canada and the gerald schwartz and heather reisman foundation .
we thank david cory and troy borneman for providing the sample and for stimulating discussions ; colm ryan , yingjie zhang and jeremy chamilliard for their contributions to the spectrometer ; roberto romero and hiruy haile for assistance with machining .
incoherent error of a quantum process can be caused by classical noise , for example , a distribution over external experimental parameters @xcite such as @xmath10 inhomogeneity and @xmath5 line - broadening . in our system , after optimizing the twt unblanking delay and using the phase transient corrected pulses , @xmath72 effect remains as the dominant source of the incoherent error .
@xmath10 inhomogeneity is also present , but does not critically reduce the control fidelity at current level .
hence only the @xmath5 noise is considered in our incoherent error discussions . in this section , we adapt the analysis presented in @xcite that was used to describe @xmath10 inhomogeneity effect on the fidelity decay to explain how @xmath72 gives rise to the non - exponential decay observed in our experimental ( fig . [
fig : bmresults ] ) and simulated ( fig .
[ fig : bmsim ] ) results . due to @xmath5 local field inhomogeneity across the sample ,
unitary errors with different strengths arise on the spins experiencing off - resonance fields .
intuitively , since the spins at different larmor frequencies experience different average unitary errors , the signal decay curve should contain multiple depolarization rates , which explains the non - exponential behavior . for more concrete analysis
, we consider a single step which consists of a computational gate ( @xmath46 ) followed by a pauli gate ( @xmath45 ) in a randomized benchmarking sequence .
the superoperator describing the process can be expressed as : @xmath112 here @xmath113 is off - resonance frequency , @xmath114 is the superoperator describing the cumulative error of @xmath115 for the fraction of the system with off - resonance frequency @xmath113 , and @xmath116 is the distribution of @xmath113 .
@xmath116 can be obtained from a frequency domain thermal state spectrum ( fig .
[ fig : t2star ] for example ) . for a given @xmath113 ,
the cumulative error strength of this single step can be calculated by @xmath117 where @xmath118 is the faulty implementation of @xmath46 and @xmath45 due to the local field inhomogeneity , and @xmath119 is the gate fidelity ( hilbert - schmidt ( hs ) norm ) of the faulty implementation .
as the off - resonance fields cause undesired rotations along z - axis , which do nt commute with rotations along x - axis or y - axis , the error strength @xmath120 is different for different combinations of @xmath46 and @xmath45 .
it can be easily verifed that for a given @xmath113 there are 9 different error strengths depending on different cases of @xmath46 and @xmath45 ( e.g. whether @xmath46 and @xmath45 are rotations along z - axis , they are along parallel or anti - parallel axes , or the rotation axis of @xmath46 is clockwise or counterclockwise from that of @xmath45 ) .
the 9 types are labelled from 1 to 9 in tab .
[ tab : errortype ] .
therefore , for a fixed @xmath113 , the error is gate - dependent .
however , it is proven in @xcite that the cumulative effect of the gate - dependent errors in the randomized benchmarking can still be described as a depolarizing channel as long as the gate dependence is weak . therefore , for the spin packet with the off - resonance frequency @xmath113 , upon averaging over random gate sequences , @xmath114 in eq .
[ eq : singleps ] forms a depolarizing channel @xmath121 with the depolarizing factor @xmath122 .
@xmath121 and @xmath122 can be expressed as @xcite : @xmath123 here @xmath124 is the depolarized channel associated with the error type @xmath125 with depolarizing parameter @xmath126 @xcite .
@xmath127 is the probability for the error type @xmath125 to occur , and their values are 1/9 , 1/9 , 1/9 , 1/9 , 1/9 , 1/9 , 2/9 , 1/18 , 1/18 .
finally , by averaging over the distribution of @xmath113 , the expression for the channel constructed from @xmath34 random gates can be obtained as : @xmath128 where @xmath129 and @xmath33 is the deviation density matrix or the traceless part of the full density matrix .
therefore , the fidelity decay ( in our case , the decay of the expectation value of @xmath60 ) is the sum of multiple exponential decays weighted with the distribution function @xmath116 . using experimentally measured values for @xmath116 , we compare in fig .
[ fig : compare ] the decay curve numerically calculated from eq .
[ eq : ngate ] with the curve obtained by simulating randomized benchmarking sequences .
the evaluation of eq .
[ eq : ngate ] is carried out numerically since @xmath118 is generated by a time - dependent hamiltonian with two non - commuting terms , @xmath60 term representing the off - resonance effect and @xmath130 terms for the external control .
for the @xmath46 and @xmath45 gates , 35 ns gaussian pulses with 1 ns time steps are used .
two decay curves agree very well , indicating the @xmath5 incoherent error is responsible for the non - exponential behavior of the fidelity decay .
furthermore , in the experimental data , the decay curve fits very well to a single - exponential function within the error bar when @xmath72 is extended using the selection sequence ( @xmath85 in fig .
[ fig : bmresults ] ) .
therefore , we conclude that the non - exponential fidelity decay can be explained by the @xmath72 error .
chiang , a.j .
costa - filho , b. baird , j.h .
freed , 2d - eldor study of heterogeneity and domain structure changes in plasma membrane vesicles upon cross - linking of receptors , j. phys .
b 115 ( 2011 ) 1046210469 .
j. j. l. morton , a. m. tyryshkin , a. ardavan , k. porfyrakis , s. a. lyon , g. a. d. briggs , measuring errors in single - qubit rotations by pulsed electron paramagnetic resonance , phys . rev . a 71 ( 2005 ) 012332 .
j. j. l. morton , a. m. tyryshkin , a. ardavan , k. porfyrakis , s. a. lyon , g. a. d. briggs , high fidelity single qubit operations using pulsed electron paramagnetic resonance , phys .
95 ( 2005 ) 200501 .
h. wu , e. m. gauger , r. e. george , m. mttnen , h. riemann , n. v. abrosimov , p. becker , h .- j .
pohl , k. m. itoh , m. l. w. thewalt , j. j. l. morton , geometric phase gates with adiabatic control in electron spin resonance , physical review a 87(3 ) ( 2013 ) 032326 .
d. k. park , g. feng , r. rahimi , s. labruyre , t. shibata , s. nakazawa , k. sato , t. takui , r. laflamme , j. baugh , hyperfine spin qubits in irradiated malonic acid : heat - bath algorithmic cooling , quantum information processing 14 ( 7 ) ( 2015 ) 24352461 .
t. maly , g. t. debelouchina , v. s. bajaj , k .-
joo , m. l. mak - jurkauskas , j. r. sirigiri , p. c. van der wel , j. herzfeld , r. j. temkin , et al . , dynamic nuclear polarization at high magnetic fields , the journal of chemical physics 128 ( 5 ) ( 2008 ) 052211 .
a. b. barnes , g. de paepe , p. c. a. van der wel , k .-
joo , v. s. bajaj , m. l. mak - jurkauskas , j. r. sirigiri , j. herzfeld , r. j. temkin , et al . , high - field dynamic nuclear polarization for solid and solution biological nmr , applied magnetic resonance 34 ( 3 - 4 ) ( 2008 ) 237263 .
v. filidou , s. simmons , s. d. karlen , f. giustino , h. l. anderson , j. j. l. morton , ultrafast entangling gates between nuclear spins using photoexcited triplet states , nature physics 8 ( 8) ( 2012 ) 596600 .
l. kevan , l.d .
kisper , electron spin double resonance spectroscopy , interscience , new york , ny , 1976 .
y. chiang , a. costa - filho , j. freed , dynamic molecular structure and phase diagram of dppc - cholesterol binary mixtures : a 2d - eldor study , j. phys .
b 111 ( 2007 ) 1126011270 . a. doll , s. pribitzer , r. tschaggelar , g. jeschke , adiabatic and fast passage ultrawideband inversion in pulsed epr , j. magn .
reson . 230
( 2013 ) 2739 .
spindler , s.j .
glaser , t.e .
skinner , t.f .
prisner , broadband inversion peldor spectroscopy with partially adiabatic shaped pulses , angew .
52 ( 2013 ) 34253429 .
spindler , y. zhang , b. endeward , n. gershernzon , t.e .
skinner , s.j .
glaser , t.f .
prisner , shaped optimal control pulses for increased excitation bandwidth in epr , j. magn .
reson . 218
( 2012 ) 4958 .
j. m. chow , j. m. gambetta , l. tornberg , j. koch , l. s. bishop , a. a. houck , b. r. johnson , l. frunzio , s. m. girvin , r. j. schoelkopf , randomized benchmarking and process tomography for gate errors in a solid - state qubit , phys .
102 ( 2009 ) 090502 .
r. barends , j. kelly , a. megrant , a. veitia , d. sank , e. jeffrey , t. c. white , j. mutus , a. g. fowler , b. campbell , y. chen , z. chen , b. chiaro , a. dunsworth , c. neill , p. omalley , p. roushan , a. vainsencher , j. wenner , a. n. korotkov , a. n. cleland , j. m. martinis , superconducting quantum circuits at the surface code threshold for fault tolerance , nature 508 ( 7497 ) ( 2014 ) 500503 .
m. veldhorst , j. c. c. hwang , c. h. yang , a. w. leenstra , b. de ronde , j. p. dehollain , j. t. muhonen , f. e. hudson , k. m. itoh , a. morello , a. s. dzurak , an addressable quantum dot qubit with fault - tolerant control - fidelity , nature nano . 9
( 2014 ) 981 - 985 .
a. laucht , j. t. muhonen , f. a. mohiyaddin , r. kalra , j. p. dehollain , s. freer , f. e. hudson , m. veldhorst , r. rahman , g. klimeck , k. m. itoh , d. n. jamieson , j. c. mccallum , a. s. dzurak , a. morello , electrically controlling single - spin qubits in a continuous microwave field , sci .
adv . 1 ( 2015 ) e1500022 .
j. t. muhonen , a. laucht , s. simmons , j. p. dehollain , r. kalra , f. e. hudson , s. freer , k. m. itoh , d. n. jamieson , j. c. mccallum , a. s. dzurak , a. morello , quantifying the quantum gate fidelity of single - atom spin qubits in silicon by randomized benchmarking , journal of physics : condensed matter 27 ( 15 ) ( 2015 ) 154205 .
f. dolde , v. bergholm , y. wang , i. jakobi , b. naydenov , s. pezzagna , j. meijer , f. jelezko , p. neumann , t. schulte - herbrggen , j. biamonte , j. wrachtrup , high - fidelity spin entanglement using optimal control , nature communications 5 ( 2014 ) 3371 .
x. rong , j geng , f. shi , y. liu , k. xu , w. ma , f. kong , z. jiang , y. wu , j. du , experimental fault - tolerant universal quantum gates with solid - state spins under ambient conditions , nature communications 6 ( 2015 ) 8748 .
n. khaneja , t. reiss , c. kehlet , t. schulte - herbrggen , s. j. glaser , optimal control of coupled spin dynamics : design of nmr pulse sequences by gradient ascent algorithms , journal of magnetic resonance 172 ( 2 ) ( 2005 ) 296305 .
g. rinard , g. eaton , loop - gap resonators , in : s. eaton , g. eaton , l. berliner ( eds . ) , biomedical epr , part b : methodology , instrumentation , and dynamics , vol .
24/b of biological magnetic resonance , springer us , 2005 , pp .
1952 . c. c. lpez , b. lvi , d. g. cory , error characterization in quantum information processing : a protocol for analyzing spatial correlations and its experimental implementation , phys .
a 79 ( 2009 ) 042328 .
m. a. fogarty , m. veldhorst , r. harper , c. h. yang , s. d. bartlett , s. t. flammia , a. s. dzurak , nonexponential fidelity decay in randomized benchmarking with low - frequency noise , physical review a 92 ( 2015 ) 022326 .
m. a. pravia , n. boulant , j. emerson , a. farid , e. m. fortunato , t. f. havel , r. martinez , d. g. cory , robust control of quantum information , the journal of chemical physics 119 ( 19 ) ( 2003 ) 999310001 . | spin systems controlled and probed by magnetic resonance have been valuable for testing the ideas of quantum control and quantum error correction .
this paper introduces an x - band pulsed electron spin resonance spectrometer designed for high - fidelity coherent control of electron spins , including a loop - gap resonator for sub - millimeter sized samples with a control bandwidth @xmath0 mhz .
universal control is achieved by a single - sideband upconversion technique with an i - q modulator and a 1.2 gs / s arbitrary waveform generator .
a single qubit randomized benchmarking protocol quantifies the average errors of clifford gates implemented by simple gaussian pulses , using a sample of gamma - irradiated quartz .
improvements in unitary gate fidelity are achieved through phase transient correction and hardware optimization . a preparation pulse sequence that selects spin packets in a narrowed distribution of static fields confirms that inhomogeneous dephasing ( @xmath1 ) is the dominant source of gate error .
the best average fidelity over the clifford gates obtained here is @xmath2 , which serves as a benchmark to compare with other technologies .
electron spin resonance , quantum information processing , randomized benchmarking |
analytic objects associated with convergent sequences of combinatorial objects have recently been attracting a significant amount of attention .
this line of research was initiated by the theory of limits of dense graphs @xcite , followed by limits of sparse graphs @xcite , permutations @xcite , partial orders @xcite , etc .
further details can also be found in a recent monograph on graph limits of lovsz @xcite .
analytic methods applied to such limit objects led to results in many areas of mathematics and computer science , most importantly , extremal combinatorics @xcite and property testing @xcite . in particular , lovsz and szegedy in
@xcite characterize testable properties and parameters of graph limits ( graphons ) and they identify a class of testable graph properties , so - called flexible properties . in this paper , we focus on a question when the limit analytic object is uniquely determined by finitely many densities of substructures . this phenomenon is known as finite forcibility . in fact , questions of this kind are closely related to quasirandomness and
they were studied well before the theory of limits of combinatorial objects emerged . for example
, the results on pseudorandom graphs from work of chung , graham and wilson @xcite , rdl @xcite and thomason @xcite imply that the homomorphic densities of @xmath0 and @xmath1 guarantee that densities of all subgraphs behave as in the random graph @xmath2 . in the language of graphons ( the limit structure for graphs )
, this result asserts that the graphon identically equal to @xmath3 is finitely forcible by densities of @xmath4-vertex subgraphs .
a similar result on permutations , which was originally raised as a question by graham , was proven by the last author and pikhurko @xcite who exploited the analytic view of permutation limits .
let us now give motivation for the concepts .
the result on finite forcibility of the graphon identically equal to @xmath3 was generalized by lovsz and ss @xcite who proved that any stepwise graphon is finitely forcible .
these results were further extended by lovsz and szegedy @xcite who also gave several conditions when a graphon is not finitely forcible . in this paper , we are interested in finite forcibility of graph limits ( graphons ) and permutation limits ( permutons ) .
we start with proving an analogue of the result of lovsz and ss @xcite for permutons , which is stated as corollary [ cor - force ] .
we then focus on finite forcibility of permutons with infinite recursive structure , and on the interplay between finite forcibility of permutons and graphons , partly motivated by question 11 from @xcite .
a graph can be associated with a permutation in the following way : the vertices of the graph correspond to the elements of the permutation and two of them are joined by an edge if they form an inversion . along the same lines
, a graphon can be associated with a permuton .
for example , the graphon associated with the limit of random permutations is one of the candidates for a finite forcible `` 2-dimensional '' graphon ( see section [ sec : concl ] for more details ) .
unfortunately , we show that there exist finitely forcible permutons such that the associated graphons are not finitely forcible . in section [ sect - infinite ] , we find two families of finitely forcible permutons , which have infinite recursive structure .
then , we show that the associated graphons are not finitely forcible in section [ sect - graphon ] , where we prove this result for all graphons with the recursive structure analogous to that of permutons from section [ sect - infinite ] .
still , we believe that some permutons leading to `` 2-dimensional '' graphons can be finitely forcible and we mention two particular cases at the end of the paper . also let us remark that the methods we use in section [ sect - infinite ] were subsequently extended by jan volec and two of the authors @xcite to resolve conjecture 9 from @xcite on compactness of finitely forcible graphons .
in this section , we introduce concepts related to graphs and permutations and their limits used in the paper .
we start with the slightly simpler notion of permutation limits .
the theory of permutation limits was built by hoppen , kohayakawa , moreira , rth and sampaio in @xcite . here , we follow the analytic view of the limit as used in @xcite , which also appeared in an earlier work of presutti and stromquist @xcite . a _ permutation of order _
@xmath5 is a bijective mapping from @xmath6 $ ] to @xmath6 $ ] , where @xmath6 $ ] denotes the set of integers from @xmath7 to @xmath5 .
the order of a permutation @xmath8 is also denoted by @xmath9 .
the set of all permutations of order @xmath5 is denoted by @xmath10 . in what follows ,
we identify a sequence of @xmath5 different integers @xmath11 between @xmath7 and @xmath5 with a permutation @xmath8 by setting @xmath12 .
for example , the identity permutation of order @xmath4 is denoted by @xmath13 . if @xmath8 is a permutation of order @xmath5 , a _
subpermutation _ induced by @xmath14 in @xmath8 is a permutation @xmath15 of order @xmath16 such that @xmath17 if and only if @xmath18 .
for example , the subpermutation of @xmath19 induced by @xmath20 is @xmath21 .
subpermutations are more commonly referred to as _ patterns _ but we decided to use the term subpermutation in the paper to be consistent with the analogous concept for graphs as well as with previous work on permutation limits . a _ density _ @xmath22 of a permutation @xmath15 of order @xmath16 in a permutation @xmath8 of order @xmath5 is the number of @xmath16-tuples inducing @xmath15 in @xmath8 divided by @xmath23 .
conveniently , we set @xmath24 if @xmath25 .
an infinite sequence @xmath26 of permutations with @xmath27 is _ convergent _ if @xmath28 converges for every permutation @xmath15 .
we will see that one can associate with every convergent sequence of permutations the following analytic object : a _ permuton _ is a probability measure @xmath29 on the @xmath15-algebra @xmath30 of borel sets of the unit square @xmath31 ^ 2 $ ] such that @xmath29 has _ uniform marginals _ , i.e. , @xmath32\times [ 0,1]{\right})=\mu{\left}([0,1]\times[\alpha,\beta]{\right})=\beta-\alpha$ ] for every @xmath33 . in
what follows , we use @xmath34 for the uniform measure on @xmath30 .
more generally , if @xmath35 ^ 2 $ ] is a non - trivial convex polygon , i.e. , a convex polygon different from a point ( however , which can be a segment ) , we define @xmath36 to be the unique probability measure on @xmath30 with support @xmath37 and mass uniformly distributed inside @xmath37 . in particular , @xmath38 ^ 2}=\lambda$ ] .
we now describe the relation between permutons and convergent sequences of permutations .
let @xmath29 be a permuton .
for an integer @xmath5 , one can sample @xmath5 points @xmath39 in @xmath31 ^ 2 $ ] randomly based on @xmath29 . because @xmath29 has uniform marginals , the @xmath40-coordinates of all these points are mutually different with probability one .
the same holds for their @xmath41-coordinates .
assume that this is indeed the case .
one can then define a permutation @xmath8 of order @xmath5 based on the @xmath5 points @xmath39 as follows : let @xmath42 $ ] be such that @xmath43 and define @xmath8 to be the unique bijective mapping from @xmath6 $ ] to @xmath6 $ ] satisfying that @xmath44 if and only if @xmath45 .
we will say that a permutation @xmath8 of order @xmath5 obtained in the just described way is a _
@xmath29-random permutation _ of order @xmath5 .
uniformly random permutation _ is a @xmath34-random permutation , i.e. , each permutation of order @xmath5 is chosen with probability @xmath46 at random .
if @xmath29 is a permuton and @xmath15 is a permutation of order @xmath5 , then @xmath47 is the probability that a @xmath29-random permutation of order @xmath5 is @xmath15 .
we now recall the core results from @xcite . for every convergent sequence @xmath26 of permutations
, there exists a unique permuton @xmath29 such that @xmath48 this permuton is the _ limit _ of the sequence @xmath26 . on the other hand ,
if @xmath29 is a permuton and @xmath49 is a @xmath29-random permutation of order @xmath50 , then with probability one the sequence @xmath26 is convergent and @xmath29 is its limit . we now give four examples of the just defined notions ( the corresponding permutons are depicted in figure [ fig - permuton ] ) .
let us consider a sequence @xmath51 such that @xmath52 is the identity permutation of order @xmath50 , i.e. , @xmath53 for @xmath54 $ ] .
this sequence is convergent and its limit is the measure @xmath55 where @xmath56{\right}\}$ ] .
similarly , the limit of a sequence @xmath57 , where @xmath58 is the permutation of order @xmath50 defined as @xmath59 for @xmath54 $ ] , is @xmath60 where @xmath61{\right}\}$ ] .
a little bit more complicated example is the following : the sequence @xmath62 , where @xmath63 is the permutation of order @xmath64 defined as @xmath65$,}\\ 2(k - i ) & \mbox{otherwise } \end{array } \right.\;\ ] ] is convergent and the limit of the sequence is the measure @xmath66 , where @xmath67{\right}\}$ ] and @xmath68{\right}\}$ ] . next ,
consider a sequence @xmath69 such that @xmath70 is a random permutation .
this sequence is convergent with probability one and its limit is the measure @xmath34 . a permuton @xmath29 is _ finitely forcible _ if there exists a finite set @xmath71 of permutations such that every permuton @xmath72 satisfying @xmath73 for every @xmath74 is equal to @xmath29 .
for example , the following result from @xcite asserts that the random permuton is finitely forcible with @xmath75 .
[ thm - quasirandom ] let @xmath29 be a permuton .
it holds that @xmath76 for every @xmath77 if and only if @xmath78 . if @xmath29 is a permuton , then @xmath79 is the function from @xmath31 ^ 2 $ ] to @xmath31 $ ] defined as @xmath80\times [ 0,y]{\right})$ ] .
for example , if @xmath78 , then @xmath81 .
observe that @xmath79 is always a continuous function satisfying @xmath82 for every @xmath83 $ ] . furthermore , notice that @xmath84 implies @xmath85 , that is , the function @xmath86 determines the permuton @xmath29 .
the next theorem was implicitly proven in @xcite .
we include its proof for completeness .
[ thm - express ] let @xmath87 be a polynomial and @xmath16 a non - negative integer . there exist a finite set @xmath71 of permutations and coefficients @xmath88 , @xmath74 , such that @xmath89 ^ 2 } p(x , y)f_{\mu}^k(x , y)\dif\lambda=\sum_{\sigma\in s}\gamma_\sigma d(\sigma,\mu)\label{eq - express}\ ] ] for every permuton @xmath29 . by additivity
, it suffices to consider the case @xmath90 for non - negative integers @xmath91 and @xmath92 .
fix a permuton @xmath29 .
since @xmath29 has uniform marginals , the product @xmath93 for @xmath94 ^ 2 $ ] is equal to the probability that out of @xmath95 points are chosen randomly independently based on @xmath29 , the first @xmath91 points belong to @xmath96\times [ 0,1]$ ] , the next @xmath92 points belong to @xmath31\times [ 0,y]$ ] , and the last @xmath16 points belong to @xmath96\times [ 0,y]$ ] .
so , the integral in ( [ eq - express ] ) is equal to the probability that the above holds for a uniform choice of a point @xmath97 in @xmath31 ^ 2 $ ] .
since @xmath29 is a measure with uniform marginals , a point @xmath97 uniformly distributed in @xmath31 ^ 2 $ ] can be obtained by sampling two points randomly independently based on @xmath29 and setting @xmath40 to be the first coordinate of the first of these two points and @xmath41 to be the second coordinate of the second point .
thus , we can consider the following random event .
let us choose @xmath98 points independently at random based on @xmath29 and denote by @xmath40 the first coordinate of the last but one point , and by @xmath41 is the second coordinate of the last point .
then the integral on the left hand side of ( [ eq - express ] ) is equal to the probability that the first @xmath91 points belong to @xmath96\times [ 0,1]$ ] , the next @xmath92 points belong to @xmath31\times [ 0,y]$ ] , and the following @xmath16 points belong to @xmath96\times [ 0,y]$ ] .
we conclude that the equation ( [ eq - express ] ) holds with @xmath99 and @xmath88 equal to the probability that the following holds for a random permutation @xmath8 of order @xmath98 : @xmath100 for @xmath101 and for @xmath102 , and @xmath103 for @xmath104 . instead of sampling two additional points to get a random point with respect to the uniform measure @xmath34 , we can also sample just a single point , which is a random point with respect to @xmath29 .
this gives the following .
[ thm - express - mu ] let @xmath87 be a polynomial and @xmath16 a non - negative integer .
there exist a finite set @xmath71 of permutations and coefficients @xmath88 , @xmath74 , such that @xmath89 ^ 2 } p(x , y)f_{\mu}^k(x , y)\dif\mu=\sum_{\sigma\in s}\gamma_\sigma d(\sigma,\mu)\ ] ] for every permuton @xmath29 .
let now @xmath105 be the set of permutations of order @xmath16 with one distinguished element ; we call such permutations _ rooted_. to denote rooted permutations , we add a bar above the distinguished element : e.g. , if the second element of the permutation @xmath106 is distinguished , we write @xmath107 . note that @xmath108 .
if @xmath109 , then @xmath110 is the probability that the point @xmath97 and @xmath111 points randomly independently chosen based on @xmath29 induce the permutation @xmath15 with the distinguished element corresponding to the point @xmath97 .
observe that @xmath112 , @xmath113 and @xmath114 .
a reader familiar with the concept of flag algebras developed by razborov @xcite might recognize the notion of @xmath7-labelled flags in the just introduced notation .
similarly to theorem [ thm - express - mu ] , the following is true .
since the proof is completely analogous to that of theorem [ thm - express ] , we decided to state the theorem without giving its proof .
[ thm - express - mu - general ] let @xmath115 be a multiset of rooted permutations .
there exist a finite set @xmath71 of permutations and coefficients @xmath88 , @xmath74 , such that @xmath89 ^ 2 } \prod_{\sigma\in\sigma}f_{\mu}^\sigma(x , y)\dif\mu=\sum_{\sigma\in s}\gamma_\sigma d(\sigma,\mu)\ ] ] for every permuton @xmath29 .
the other limit structure we consider is limits of graphs .
graph _ is a pair @xmath116 where @xmath117 .
the elements of @xmath118 are called _ vertices _ and the elements of @xmath119 are called _ edges_. the _ order _ of a graph @xmath120 is the number of its vertices and it is denoted by @xmath121 . if @xmath120 and @xmath122 are graphs , then @xmath123 is the disjoint union of @xmath120 and @xmath122 and @xmath124 is the graph obtained from @xmath123 by adding all edges between @xmath120 and @xmath122 . finally , if @xmath120 is a graph and @xmath125 is a subset of its vertices , then @xmath126 is the graph obtained from @xmath120 by removing the vertices of @xmath125 and all edges containing at least one vertex from @xmath125 .
the _ density _
@xmath127 of @xmath128 in @xmath120 is the probability that @xmath129 randomly chosen vertices of @xmath120 induce a subgraph isomorphic to @xmath128 . if @xmath130 , we set @xmath131 .
we now survey basic results related to the theory of dense graph limits as developed in @xcite .
a sequence of graphs @xmath132 is _ convergent _ if the limit @xmath133 exists for every @xmath128 .
the associated limit object is called _ graphon _ : it is a symmetric @xmath134-measurable function from @xmath31 ^ 2 $ ] to @xmath31 $ ] .
here , symmetric stands for the property that @xmath135 for every @xmath136 $ ] . if @xmath137 is a graphon , then a _ @xmath137-random graph _ of order @xmath16 is obtained by sampling @xmath16 random points @xmath138 $ ] uniformly and independently and joining the @xmath50-th and the @xmath139-th vertex by an edge with probability @xmath140 . as in the case of permutations , we write @xmath141 for the probability that a @xmath137-random graph of order @xmath129 is isomorphic to @xmath128 . for every convergent sequence @xmath132 of graphs
, there exists a graphon @xmath137 such that @xmath142 for every graph @xmath128 .
we call such a graphon @xmath137 a _ limit _ of @xmath132 . on the other hand ,
if @xmath137 is a graphon , then with probability one the sequence @xmath132 where @xmath143 is a @xmath137-random graph of order @xmath50 is convergent and its limit is @xmath137 .
unlike in the case of permutations , the limit of a convergent sequence of graphs is not unique . for example , if @xmath137 is a limit of @xmath132 and @xmath144\to[0,1]$ ] is a measure preserving transformation , then the graphon @xmath145 is also a limit of @xmath132 .
let us introduce the following definition of equivalence of graphons : two graphons @xmath137 and @xmath146 are _ weakly isomorphic _ if @xmath147 for every graph @xmath128 . finally , a graphon @xmath137 is _ finitely forcible _ if there exist graphs @xmath148 such that any graphon @xmath146 satisfying @xmath149 for @xmath150 $ ] is weakly isomorphic to @xmath137 .
the densities of graphs in a graphon @xmath137 can be expressed as integrals using @xmath137 .
if @xmath137 is a graphon and @xmath128 is a graph of order @xmath16 with vertices @xmath151 and edge set @xmath119 , then @xmath152^k}\prod_{v_iv_j\in e } w(x_i , x_j)\prod_{v_iv_j\not\in e } ( 1-w(x_i , x_j ) ) \dif x_1\dots\dif x_k\label{eq - densw}\ ] ] where @xmath153 is the automorphism group of @xmath128 .
a permutation @xmath8 of order @xmath16 can be associated with a graph @xmath154 of order @xmath16 as follows .
the vertices of @xmath154 are the integers between @xmath7 and @xmath16 and @xmath155 is an edge of @xmath120 if and only if either @xmath156 and @xmath157 , or @xmath158 and @xmath159 .
if @xmath26 is a convergent sequence of permutations , then the sequence of graphs @xmath160 is also convergent .
moreover , if two convergent sequences of permutations have the same limit , then the graphons associated with the two corresponding ( convergent ) sequences of graphs are weakly isomorphic . in this way
, we may associate each permuton @xmath29 with a graphon @xmath161 , which is unique up to a weak isomorphism ( see figure [ fig - graphon ] for examples ) .
in this section , we give a sufficient condition on a permuton to be finitely forcible . a function @xmath162 ^ 2\to\rr$ ]
is called piecewise polynomial if there exist finitely many polynomials @xmath163 such that @xmath164 for every @xmath94 ^ 2 $ ] .
[ thm - force ] every permuton @xmath29 such that @xmath86 is piecewise polynomial is finitely forcible .
let @xmath29 be a permuton such that @xmath86 is piecewise polynomial , that is , there exist polynomials @xmath163 such that @xmath165 for every @xmath94 ^ 2 $ ] .
let @xmath166 be the set of all continuous functions @xmath167 on @xmath31 ^ 2 $ ] such that @xmath168 for every @xmath94 ^ 2 $ ] .
the set @xmath166 is finite .
indeed , let @xmath169 and let @xmath170 be the set of all points @xmath171 such that @xmath172 .
by harnack s curve theorem , the set @xmath170 has finitely many connected components .
bzout s theorem implies that the number of branching points in each of these components is finite and these points have finite degrees .
consequently , @xmath173 has finitely many components . if @xmath174 are all the connected components of @xmath31 ^ 2\setminus q$ ] , then each function @xmath175 coincides with one of the @xmath16 polynomials @xmath163 on every @xmath176 .
so , @xmath177 . by stone - weierstrass theorem , there exist a polynomial @xmath87 and @xmath178 such that @xmath179 ^ 2}\left ( f_\mu(x , y)-p(x , y ) \right)^2\dif\lambda & < & \varepsilon \;\mbox { , and } \label{eq - force1 } \\ \int_{[0,1]^2}\left ( f(x , y)-p(x , y ) \right)^2\dif\lambda & > & \varepsilon \mbox { for every $ f\in\ff$ , $ f\not = f_\mu$. } \label{eq - force2}\end{aligned}\ ] ] let @xmath180 be the value of the left hand side of ( [ eq - force1 ] ) .
we claim that the unique permuton @xmath72 satisfying @xmath179 ^ 2}\prod_{i=1}^k\left ( f_\mu'(x , y)-p_i(x , y ) \right)^2\dif\lambda & = & 0 \;\mbox { , and } \label{eq - force3 } \\ \int_{[0,1]^2}\left ( f_\mu'(x , y)-p(x , y ) \right)^2\dif\lambda & = & \varepsilon_0 \ ; \label{eq - force4}\end{aligned}\ ] ] is @xmath29 .
assume that @xmath72 is a permuton satisfying both ( [ eq - force3 ] ) and ( [ eq - force4 ] ) .
the equation ( [ eq - force3 ] ) implies that @xmath181 .
next , ( [ eq - force2 ] ) , ( [ eq - force4 ] ) , and yield that @xmath182 for every @xmath175 , @xmath183 . we conclude that @xmath184 and thus @xmath185 . by theorem [ thm - express ] ,
the left hand sides of ( [ eq - force3 ] ) and ( [ eq - force4 ] ) can be expressed as finite linear combinations of densities @xmath47 .
let @xmath71 be the set of all permutations appearing in these linear combinations .
any permuton @xmath72 with @xmath186 for every @xmath74 satisfies both ( [ eq - force3 ] ) and ( [ eq - force4 ] ) and thus it must be equal to @xmath29 .
this shows that @xmath29 is finitely forcible .
we immediately obtain the following corollary .
[ cor - force ] if @xmath29 is a permuton such that there exist non - negative reals @xmath187 and non - trivial convex polygons @xmath188 ^ 2 $ ] satisfying @xmath189 , then @xmath29 is finitely forcible
. let @xmath190 , @xmath191 $ ] , be the function from @xmath31 ^ 2 $ ] to @xmath31 $ ] defined as @xmath192\times [ 0,y]{\right})$ ] . clearly , each function @xmath190 is piecewise polynomial . since @xmath193
, the finite forcibility of @xmath29 follows from theorem [ thm - force ] .
a particular case of permutons that are finitely forcible by corollary [ cor - force ] is the following .
if @xmath16 is an integer , @xmath194 $ ] are reals such that @xmath195 and @xmath196 is a square matrix of order @xmath16 with entries being non - negative reals summing to @xmath197 in the @xmath50-th row and in the @xmath50-th column , we can define a permuton @xmath198 to be the sum @xmath199 where @xmath200\times [ s_{j-1},s_j]$ ] , @xmath201 $ ] and @xmath202 ( in particular , @xmath203 and @xmath204 ) .
for instance , if @xmath205 and @xmath206 we get the permuton depicted in figure [ fig - force - step ] .
in this section , we show that two particular types of permutons with infinite structure are finitely forcible .
we show that the associated graphons are not finitely forcible in the next section . for @xmath207 ,
define @xmath208 to be the permuton @xmath209 where @xmath210{\right}\}$ ] .
examples of the just defined permutons can be found in figure [ fig - mono ] .
we next show that all permutons @xmath208 are finitely forcible .
[ thm - mono ] for every @xmath207 , the permuton @xmath208 is finitely forcible .
we claim that any permuton @xmath29 satisfying @xmath211 ^ 2}\left(1-x - y+f_\mu(x , y)-\frac{\alpha}{1-\alpha}{\left}(x+y-2f_\mu(x , y){\right})\right)^2\dif\mu = 0 & \label{eq - mono2}\end{aligned}\ ] ] is equal to @xmath212 .
this would prove the finite forcibility of @xmath212 by theorem [ thm - express - mu ] .
note that the permuton @xmath208 satisfies ( [ eq - mono1 ] ) , ( [ eq - mono1b ] ) , and ( [ eq - mono2 ] ) . assume that a permuton @xmath29 satisfies ( [ eq - mono1 ] ) , ( [ eq - mono1b ] ) , and ( [ eq - mono2 ] ) .
let @xmath213 be the support of @xmath29 and consider the binary relation @xmath214 defined on the support of @xmath29 such that @xmath215 if * @xmath216 and @xmath217 , or * @xmath218 and @xmath219 , or * @xmath220 and @xmath221 .
the relation @xmath214 is an equivalence relation .
indeed , the reflexivity and symmetry is clear . to prove transitivity ,
consider three points @xmath97 , @xmath222 and @xmath223 such that @xmath215 and @xmath224 but it does not hold that @xmath225 . by the definition of @xmath214 , either @xmath218 and @xmath226 , or @xmath220 and @xmath227 .
if @xmath218 and @xmath226 , then we obtain that @xmath228 unless @xmath229 ( recall that @xmath214 is defined on the support of @xmath29 ) .
we can now assume that @xmath229 and @xmath230 . since @xmath29 has uniform marginals , the support of @xmath29 intersects at least one of the open rectangles @xmath231 , @xmath232 and @xmath233 .
however , this yields that @xmath228 in the first two cases and @xmath234 in the last case .
the case @xmath220 and @xmath227 is handled in an analogous way .
let @xmath235 be the set of equivalence classes of @xmath214 .
if @xmath236 , let @xmath237 and @xmath238 be the projections of @xmath37 on the @xmath40 and @xmath41 axes .
it is not hard to show that @xmath237 is a closed interval for each @xmath236 and these intervals are internally disjoint for different choices of @xmath236 .
the same holds for the projections on the @xmath41 axis .
since @xmath29 has uniform marginals , the intervals @xmath237 and @xmath238 must have the same length for every @xmath236 .
moreover , the definition of @xmath214 implies that if @xmath237 precedes @xmath239 , then @xmath238 also precedes @xmath240 for any @xmath241 .
we conclude that there exists a set @xmath242 of internally disjoint closed intervals such that @xmath243\in\ii } { \left}[z , z'{\right}]=[0,1]\mbox { and}\ ] ] the support of @xmath29 is equal to ( because the density of subpermutations @xmath244 and @xmath245 is zero ) @xmath243\in\ii } { \left}\{(x , z'-x+z),x\in { \left}[z , z'{\right}]{\right}\}\;\mbox{.}\ ] ] note that some intervals contained in @xmath242 may be formed by single points .
let @xmath246 be the subset of @xmath242 containing the intervals of positive length .
let @xmath247\in\ii_0 $ ] and let @xmath248{\right}\}$ ] .
the equality ( [ eq - mono2 ] ) implies that the ( continuous ) function integrated in ( [ eq - mono2 ] ) is zero for every @xmath249 .
substituting @xmath250 and @xmath251 into this function implies @xmath252 let @xmath253 be the set formed by the left end points of intervals in @xmath246 .
define @xmath254 to be the minimum elements of @xmath253 , and in general @xmath197 to be the minimum element of @xmath255 .
the existence of these elements follows from ( [ eq - zz ] ) and the fact that the intervals in @xmath246 are internally disjoint .
if @xmath253 is finite , we set @xmath256 for @xmath257 .
we derive from the definition of @xmath253 and from ( [ eq - zz ] ) that @xmath258 , ~ i\in \mathbb{n}^+{\right}\}\setminus{\left}\{[1,1]{\right}\}\;\mbox{.}\end{aligned}\ ] ] consequently , we obtain @xmath259 for @xmath260 , we define @xmath261 $ ] as follows : @xmath262 the equation ( [ eq - d21 ] ) can now be rewritten as @xmath263 hence , the equality ( [ eq - mono1b ] ) can hold only if @xmath264 for every @xmath139 which implies that @xmath265 . consequently , the permutons @xmath29 and @xmath208 are identical .
we remark that any permuton @xmath29 obeying the constraints ( [ eq - mono1 ] ) and ( [ eq - mono2 ] ) must also be equal to @xmath208 .
however , we decided to include the additional constraint ( [ eq - mono1b ] ) to make the presented arguments more straightforward .
we now show finite forcibility of another class of permutons .
its structure is similar to that of @xmath208 .
the proof proceeds along similar lines as the proof of theorem [ thm - mono ] but we have to overcome several new technical difficulties . for @xmath207 , define @xmath266 to be the permuton @xmath267\times[1-\alpha^{i-1},1-\alpha^i]}\;\mbox{.}$ ] see figure [ fig - rand ] for examples .
our goal is to show that all permutons @xmath266 are finitely forcible .
we start by proving an auxiliary lemma .
[ lm - multiint2 ] there exist a finite set @xmath71 of permutations and reals @xmath88 , @xmath74 , such that the following is equivalent for every permuton @xmath29 : * @xmath268 , * @xmath29 restricted to @xmath269\times[y_2,y_1]$ ] is a ( possibly zero ) multiple of @xmath270\times[y_2,y_1]}$ ] for any two points @xmath271 and @xmath272 of the support of @xmath29 with @xmath273 and @xmath274 .
the proof technique is similar to that used in @xcite .
let @xmath275 , i.e. , @xmath276 .
let @xmath271 and @xmath272 be two points of the support of @xmath29 with @xmath273 and @xmath274 . by cauchy - schwartz inequality , the measure @xmath29 restricted @xmath269\times[y_2,y_1]$ ] is a multiple of @xmath277\times[y_2,y_1]}$ ] if and only if it holds that @xmath278\times [ y_2,y])^2\;\dif\tilde{\lambda}_{[x_1,x_2]\times[y_2,y_1]}\right)\times \nonumber \\ & \left(\int\limits_{(x , y)}(x - x_1)^2(y - y_2)^2\;\dif\tilde{\lambda}_{[x_1,x_2]\times[y_2,y_1]}\right)- \nonumber \\ & \left(\int\limits_{(x , y)}(x - x_1)(y - y_2)\mu([x_1,x]\times [ y_2,y])\;\dif\tilde{\lambda}_{[x_1,x_2]\times[y_2,y_1]}\right)^2 = 0 \label{eq - multi - rand1}\end{aligned}\ ] ] since the left hand side of ( [ eq - multi - rand1 ] ) can not be negative , we obtain that the second statement in the lemma is equivalent to @xmath279\times [ y_2,y]{\right})^2- & & \nonumber \\ & ( x - x_1)(y - y_2)\cdot\mu{\left}([x_1,x]\times [ y_2,y]{\right})\cdot(x'-x_1)(y'-y_2)\cdot\mu([x_1,y_2]\times [ x',y ' ] ) & & \nonumber \\ & \qquad\qquad\qquad\qquad\qquad\qquad \;\dif\tilde{\lambda}_{[x_1,x_2]\times[y_2,y_1]}\;\dif\tilde{\lambda}_{[x_1,x_2]\times[y_2,y_1]}\;\dif\mu\;\dif\mu ~=~ 0 \label{eq - multi - rand2}\end{aligned}\ ] ] in the rest of the proof , we show that the left hand side of ( [ eq - multi - rand2 ] ) can be expressed as a linear combination of finitely many subpermutation densities .
since this argument follows the lines of the proofs of theorems [ thm - express][thm - express - mu - general ] , we only briefly explain the main steps .
the left hand side of ( [ eq - multi - rand2 ] ) is equal to the expected value of the integrated function in ( [ eq - multi - rand2 ] ) for two points @xmath271 and @xmath272 randomly chosen in @xmath31 ^ 2 $ ] based on @xmath29 and two points @xmath97 and @xmath222 randomly chosen in @xmath31 ^ 2 $ ] based on @xmath34 when treating the value of the integrated function to be zero if @xmath280 , @xmath281 , @xmath282 $ ] , @xmath283 $ ] , @xmath284 $ ] , or @xmath285 $ ] .
such points @xmath271 , @xmath272 , @xmath97 , and @xmath222 can be obtained by sampling six random points from @xmath31 ^ 2 $ ] based on @xmath29 since @xmath29 has uniform marginals ( see the proof of theorem [ thm - express ] for more details ) .
when the four points @xmath271 , @xmath272 , @xmath97 , and @xmath222 are sampled , any of the quantities @xmath286 , @xmath287 , @xmath40 , @xmath41 , @xmath288 , @xmath289 , @xmath290\times [ x , y]{\right})$ ] , and @xmath291\times [ x',y'])$ ] appearing in the product is equal to the probability that a point randomly chosen in @xmath31 ^ 2 $ ] based on @xmath29 has a certain property in a permutation determined by the sampled points . since we need to sample six additional points to be able to determine each of the products appearing in ( [ eq - multi - rand1 ] ) , the left hand side of ( [ eq - multi - rand1 ] ) is equal to a linear combination of densities of @xmath292-element permutations with appropriate coefficients .
we conclude that the lemma holds with @xmath293 .
analogously , one can prove the following lemma . since the proof follows the lines of the proof of lemma [ lm - multiint2 ] , we omit further details .
[ lm - multiint3 ] there exist a finite set @xmath71 of permutations and reals @xmath88 , @xmath74 such that the following is equivalent for every permuton @xmath29 : * @xmath294 , * if @xmath271 , @xmath272 , and @xmath295 are three points of the support of @xmath29 with @xmath296 and @xmath297 , then @xmath29 restricted @xmath298\times[y_2,y_3]$ ] is a ( possibly zero ) multiple of @xmath299\times[y_2,y_3]}$ ] .
we are now ready to show that each permuton @xmath266 , @xmath207 , is finitely forcible .
[ thm - rand ] for every @xmath207 , the permuton @xmath266 is finitely forcible .
let @xmath300 be the union of the two sets of permutations from lemmas [ lm - multiint2 ] and [ lm - multiint3 ] .
next , consider the following eight functions : @xmath301 to save space in what follows , we often omit parameters when no confusion can arise , e.g. , we write @xmath302 for the value @xmath303 if @xmath40 and @xmath41 are clear from the context . we claim that any permuton satisfying the following three conditions is equal to @xmath304 : @xmath305 @xmath306 ^ 2}\big((1-\alpha){\left}(f_\mu^\nearrow f_\mu^\searrow - f_\mu^\searrow f_\mu^\nearrow{\right})f_\mu^\nwarrow \nonumber\\ & \qquad - \alpha\big(f_\mu^\nwarrow f_\mu^\nwarrow f_\mu^\searrow+ f_\mu^\searrow f_\mu^\nwarrow f_\mu^\searrow+ f_\mu^\nwarrow f_\mu^\swarrow
f_\mu^\searrow+ f_\mu^\searrow f_\mu^\nwarrow
f_\mu^\nearrow\big ) \big)^2\;\dif\mu = 0 \ ; , \label{eq - rand1b}\end{aligned}\ ] ] and @xmath307 this would prove the finite forcibility of @xmath304 by theorem [ thm - express - mu - general ] .
suppose that a permuton @xmath29 satisfies ( [ eq - rand1 ] ) , ( [ eq - rand1b ] ) , and ( [ eq - rand2 ] ) .
let @xmath213 be the support of @xmath29 and consider the binary relation @xmath214 defined on the support of @xmath29 such that @xmath215 if * @xmath216 and @xmath217 , * @xmath218 and @xmath219 , or * @xmath220 and @xmath221 .
unlike in the proof of theorem [ thm - mono ] , the relation @xmath214 need not be an equivalence relation .
instead , we consider the transitive closure @xmath308 of @xmath214 and let @xmath235 be the set of the equivalence classes of @xmath308 . we define @xmath309 , where @xmath97 and @xmath222 are two points of the support of @xmath29 such that @xmath215 , as follows @xmath310\times [ y',y]{\right})}{(x'-x)(y - y ' ) } & \mbox { if $ x < x'$ and $ y > y'$ , } \\ \frac{\mu{\left}([x',x]\times [ y , y']{\right})}{(x - x')(y'-y ) } & \mbox { if $ x > x'$ and $ y < y'$ , and } \\ 0 & \mbox{otherwise . }
\end{array } \right.\ ] ] since @xmath29 satisfies ( [ eq - rand1 ] ) , lemma [ lm - multiint2 ] implies that any three points @xmath97 , @xmath222 and @xmath223 of the support of @xmath29 such that @xmath215 and @xmath224 satisfy @xmath311 . in particular , the quantity @xmath309 is the same for all pairs of points @xmath97 and @xmath222 with @xmath215 lying in the same equivalence class of @xmath308
so , we may define @xmath312 to be this common value for each equivalence class @xmath236 or for a closure of such class . as in the proof of theorem [ thm - mono ]
, we define @xmath237 and @xmath238 to be the projections of an equivalence class @xmath236 on the @xmath40 and @xmath41 axes .
the definition of @xmath214 yields that @xmath237 and @xmath238 are closed intervals for all @xmath236 and these intervals are internally disjoint for different choices of @xmath236 .
since @xmath29 has uniform marginals , the intervals @xmath237 and @xmath238 must have the same length for every @xmath236 .
as in the proof of theorem [ thm - mono ] , we conclude that there exists a set @xmath242 of internally disjoint closed intervals such that @xmath313\in\ii } [ z , z']=[0,1]\ ; \mbox{,}\ ] ] the support of @xmath29 is a subset of @xmath313\in\ii } [ z , z']\times [ z , z']\;\mbox{,}\ ] ] and the interior of each of these squares intersects at most one class @xmath314 . since some intervals contained in @xmath242 may be formed by single points , we define @xmath246 to be the subset of @xmath242 containing the intervals of positive length .
let @xmath247\in\ii_0 $ ] and let @xmath37 be the closure of the corresponding equivalence class from @xmath235 .
let @xmath315 , @xmath316 $ ] , be the minimum @xmath41 such that @xmath97 belongs to @xmath37 ; similarly , @xmath317 denotes the maximum such @xmath41 .
assume first that @xmath318 .
since @xmath29 has uniform marginals , it must holds that @xmath319 for every @xmath320 . from and lemma [ lm - multiint2 ]
we see that the functions @xmath167 and @xmath321 are non - decreasing , and similarly and lemma [ lm - multiint3 ] imply that @xmath167 and @xmath321 are non - increasing . we conclude that @xmath322\times [ z , z']{\right})$ ] and @xmath323 .
assume now that @xmath324 .
lemma [ lm - multiint3 ] and ( [ eq - rand1 ] ) imply that if @xmath325 belongs to the support of @xmath29 , then @xmath326\times [ z , y])=0 $ ] ( otherwise , @xmath318 ) .
but then @xmath97 can not be in relation @xmath214 with another point of the support of @xmath29 .
so , we conclude that the case @xmath324 can not appear .
the just presented arguments show the support of the measure @xmath29 is equal to @xmath313\in\ii } [ z , z']\times [ z , z']\ ] ] and the measure is uniformly distributed inside each square @xmath247\times [ z , z']$ ] , @xmath247\in\ii_0 $ ] .
let @xmath247 $ ] be one of the intervals from @xmath246 . recall that we have argued that @xmath327\times[0,z]\cup [ z , z']\times [ z , z']\cup [ z',1]\times [ z',1]\big)=1\ ] ] and the measure @xmath29 is uniform inside the square @xmath247\times [ z , z']$ ] ( see figure [ fig - rand - zz ] ) . by ( [ eq - rand1b ] ) , the following holds for almost every point of the support of @xmath29 : @xmath328 in particular , this holds for all points in @xmath247\times [ z , z']$ ] since the functions appearing in ( [ eq - rand1c ] ) are continuous .
let @xmath97 be a point from @xmath329 .
let @xmath330 , @xmath331 , @xmath332 , and @xmath333 ( see figure [ fig - rand - zz ] ) .
since all the quantities appearing in ( [ eq - rand1c ] ) are positive , we may rewrite ( [ eq - rand1c ] ) as @xmath334 observe that @xmath335\times [ y,1])$ ] , @xmath336\times [ y , z'])=\frac{x_1y_2}{z'-z}$ ] , and @xmath337\times [ z , y])=\frac{x_2y_1}{z'-z}$ ] .
further observe that @xmath338 plugging these observations in ( [ eq - rand1d ] ) , we obtain that @xmath339\times [ y,1])-\frac{x_2y_2}{z'-z}\right)=\alpha\frac{x_1y_2+x_2y_1+x_1y_1+x_2y_2}{z'-z}\;\mbox{. } \label{eq - rand1f}\ ] ] since @xmath340 and @xmath341\times[y , z'])$ ] , we obtain from ( [ eq - rand1f ] ) that @xmath342\times [ z',1])=\alpha\frac{(z'-z)^2}{z'-z}={\alpha}(z'-z)\;\mbox{. } \label{eq - rand1g}\ ] ] finally , we substitute @xmath343 for @xmath344\times[z',1])$ ] in ( [ eq - rand1 g ] ) and get the following : @xmath345 so , we conclude that the right end point of every interval in @xmath246 is uniquely determined by its left end point .
let @xmath253 be the set formed by the left end points of intervals in @xmath246 .
as in the proof of theorem [ thm - mono ] , for a positive integer @xmath50 , let @xmath197 be the @xmath50th smallest element of @xmath253 .
notice that the existence of minimum elements follows from ( [ eq - rand - zz ] ) . if @xmath253 is finite , we set @xmath256 for @xmath257 .
we derive from the definition of @xmath253 and from ( [ eq - rand - zz ] ) that @xmath258 , ~ i\in \mathbb{n}^+{\right}\}\setminus{\left}\{[1,1]{\right}\}\;\mbox{.}\end{aligned}\ ] ] consequently , we obtain @xmath346 analogously to the proof of theorem [ thm - mono ] , for @xmath260 , we define @xmath261 $ ] as follows : @xmath262 the equation ( [ eq - rand - d21 ] ) can be rewritten as @xmath347 hence , the equality ( [ eq - rand2 ] ) can hold only if @xmath264 for every @xmath139 , i.e. , @xmath265 .
this implies that the permutons @xmath29 and @xmath266 are identical .
in this section , we show that two types of graphons similar to the two types of finitely forcible permutons from section [ sect - infinite ] are not finitely forcible . we start with graphons @xmath348 associated with permutons @xmath208 , @xmath349 . in this subsection we focus on graphons with @xmath350 where @xmath351 is the path on @xmath352 vertices .
we start with the following lemma , which seems to be of be of independent interest .
informally , the lemma asserts that any finitely forcible graphon with zero density of @xmath351 can be forced by finitely many densities of complete graphs .
[ lm - clique ] if @xmath353 is a finitely forcible graphon and @xmath354 , then there exists an integer @xmath355 such that any graphon @xmath137 with @xmath350 and @xmath356 for @xmath357 is weakly isomorphic to @xmath353 . to prove the statement of the lemma , it is enough to show the following claim : the density of any @xmath5-vertex graph @xmath120 in a graphon @xmath137 with @xmath350
can be expressed as a combination of densities of @xmath358 , @xmath359 in @xmath137 .
we proceed by induction on @xmath360 where @xmath5 and @xmath16 are the numbers of vertices and components of @xmath120 respectively .
if @xmath361 , there exists only a single one - vertex graph @xmath362 and the claim holds .
assume now that @xmath363 . if @xmath120 is not a union of complete graphs , then @xmath364 since @xmath350 .
so , we assume that @xmath120 is a union of @xmath16 complete graphs @xmath365 , i.e. , @xmath366 . if @xmath367 , then @xmath368 and the claim clearly holds .
so , we assume @xmath369 . for @xmath370
, we denote @xmath371\setminus\{1,i\}}g_j\ ; .\ ] ] observe that the following holds : @xmath372 where @xmath373 is the probability that a set @xmath118 of randomly chosen @xmath374 vertices of the graph @xmath375 induces a complete graph and the graph @xmath376 is isomorphic to @xmath377 , and @xmath378 , @xmath379 , is the probability that a set @xmath118 of randomly chosen @xmath374 vertices of @xmath380 induces a complete graph and the graph @xmath381 is isomorphic to @xmath377 . to see ( [ eq - clique - mul ] ) , observe that the product @xmath382 is equal to the product of the probability that a @xmath137-random graph of order @xmath374 is isomorphic to @xmath383 and the probability that a @xmath137-random graph of order @xmath384 is isomorphic to @xmath377 .
this is equal to the probability that randomly chosen @xmath374 vertices of a @xmath137-random graph of order @xmath385 induce a subgraph isomorphic to @xmath383 and the remaining vertices induce a subgraph isomorphic to @xmath377 .
this probability is equal to the right hand side of ( [ eq - clique - mul ] ) . by induction , @xmath386 and @xmath387 , @xmath370
, can be expressed as combinations of densities of complete graphs of order at most @xmath5 in @xmath137 . rearranging the terms of ( [ eq - clique - mul ] )
, we obtain that @xmath388 is equal to a combination of densities of complete graphs of order at most @xmath5 in @xmath137 . for a sequence of non - negative reals @xmath389 such
that @xmath390 , define a graphon @xmath391 such that @xmath392 if and only if there exists @xmath393 such that @xmath394 we consider a particular case of this graphon @xmath395 for @xmath207 : @xmath396 for @xmath397 .
observe that @xmath398 .
the main result of this subsection asserts that unlike the associated permuton @xmath208 , the graphon @xmath395 is not finitely forcible .
although it immediately follows as a corollary of the more general theorem [ thm - quasi ] , we give its proof here to increase the readability of the paper .
[ thm - clique ] for every @xmath207 , the graphon @xmath398 is not finitely forcible .
observe that @xmath399 . by lemma [ lm - clique ] , it is enough to show that @xmath395 is not finitely forcible with @xmath400 for any @xmath401 , i.e. , by setting the densities of @xmath351 and the complete graphs of orders @xmath402 .
suppose for the sake of contradiction that for some @xmath401 the graphon @xmath395 is uniquely determined by the densities of @xmath351 and @xmath403 .
let @xmath397 .
further , define functions @xmath404 for @xmath405 as follows : @xmath406 observe that if it holds that @xmath407 , which is equivalent to @xmath408 , then it also holds that @xmath409$,}\label{eq - clique2}\ ] ] where @xmath410 is the sequence with @xmath411 for @xmath412 and @xmath413 for @xmath414 .
hence , to obtain the desired contradiction , it suffices to prove that there exist functions @xmath415 , @xmath416 $ ] , on some open neighborhood of @xmath417 such that @xmath418$.}\label{eq - clique3}\ ] ] indeed , if such functions @xmath415 , @xmath416 $ ] , exist , then ( [ eq - clique2 ] ) yields that the densities of @xmath403 in the graphon @xmath419 with @xmath420 for @xmath421 , @xmath422 and @xmath413 for @xmath414 equal their densities in the graphon @xmath391 .
this implies that @xmath391 is not forced by the densities of @xmath351 and @xmath403 .
we now establish the existence of functions @xmath423 satisfying ( [ eq - clique3 ] ) on some open neighborhood of @xmath417 .
observe that @xmath424 we consider the jacobian matrix of the functions @xmath425 with respect to @xmath426 .
the determinant of the jacobian matrix is equal to @xmath427 substituting @xmath428 for @xmath429 , we obtain that the jacobian matrix has non - zero determinant .
in particular , the jacobian is non - zero .
the implicit function theorem now implies the existence of the functions @xmath423 satisfying ( [ eq - clique3 ] ) .
this concludes the proof .
the graphons considered in the previous section were associated with the permutons @xmath208 .
we now consider graphons related to the permutons @xmath266 .
in fact , we introduce a more general concept .
let @xmath137 be a graphon .
for a sequence of non - negative reals @xmath389 such that @xmath390 , define a graphon @xmath430 as follows . informally speaking , we take the graphon @xmath431 and plant a copy of @xmath137 on each of its `` components '' . formally , for @xmath432 , let @xmath433 and @xmath434 be the integers such that @xmath435 if @xmath436 , then @xmath437 . if @xmath438 , then @xmath439 the set of pairs @xmath97 with one of the coordinates being equal to zero has mesure zero , so we can for example set the values of @xmath440 for such pairs to be equal to zero .
similarly as before , we also define @xmath441 for @xmath207 as @xmath442 where @xmath397 . if @xmath443 is the graphon identically equal to @xmath444 $ ] , then @xmath445 is @xmath391 ( we put a comma to separate the two indices , the first referring to the density of edges inside connected components , the second determining the sizes of those ) .
more generally , we use @xmath446 for the graphon @xmath447 .
examples can be found in figure [ fig - quasi ] .
our main theorem asserts that a graphon @xmath441 is not finitely forcible unless it is @xmath448 .
[ thm - quasi ] for every @xmath207 and every graphon @xmath137 , if the graphon @xmath441 is finitely forcible , then @xmath137 is weakly isomorphic to @xmath448 , i.e. , the graphon @xmath441 is identically equal to zero up to a set of measure zero .
it is enough to show that for every @xmath5 , there exists @xmath410 different from @xmath449 , @xmath397 , such that @xmath450 the proof of theorem [ thm - clique ] yields that for every @xmath5 , there exists such @xmath410 different from @xmath449 satisfying @xmath451 we claim that this @xmath410 also satisfies ( [ eq - quasi1 ] ) . also note that ( [ eq - quasi2 ] ) is non - zero only for graphs @xmath120 that are disjoint union of cliques .
let @xmath120 be a graph with @xmath5 vertices and let @xmath365 be the connected components of @xmath120 .
furthermore , let @xmath452 be the partition of @xmath453 $ ] according to the isomorphism classes of the graphs @xmath454 , i.e. , for every @xmath455 $ ] with @xmath456 and every @xmath457 and @xmath458 , the graphs @xmath459 and @xmath460 are isomorphic , and the graphs @xmath459 and @xmath461 are not isomorphic . observe that @xmath462\to\nn }
c(f ) { \left } ( \prod_{i=1}^\infty d\left(\bigcup_{j\in f^{-1}(i ) } g_j , w\right ) b_i^{{\left}|\bigcup\limits_{j\in f^{-1}(i ) } g_j{\right}|}{\right})\ ; \label{eq - quasi4}\end{aligned}\ ] ] with the normalizing factor @xmath463 } \frac{\prod\limits_{i=1}^\infty { \left}|f ^{-1}(i)\cap i_m{\right}|!}{{\left}|i_m{\right}|!},\ ] ] where we set @xmath464 and the density @xmath465 of the empty graph in the graphon @xmath137 to @xmath7 .
we consider partitions of the set of connected components of @xmath120 .
if @xmath466 is such a partition , we slightly abuse the notation and identify @xmath467 with the subgraph of @xmath120 induced by the components of @xmath467 .
in particular , @xmath468 denotes the number of vertices in this subgraph .. furthermore , we always view a partition @xmath170 as a multiset , and also allow some of the @xmath467 s to be empty . let @xmath469 be the set of all such partitions .
the identity can now be rewritten as follows : @xmath470}d(q_i , w ) d{\left}(\bigcup\limits_{i\in[k ] } k_{{\left}|q_i{\right}|},w_{\vec b}^c{\right } ) , \label{eq - quasi5}\ ] ] where @xmath471 is the empty graph .
since @xmath410 satisfies ( [ eq - quasi2 ] ) , we obtain that it satisfies , and therefore also ( [ eq - quasi1 ] ) .
we immediately obtain the following two corollaries .
[ cor - quasi ] for every @xmath207 and every @xmath472 $ ] , the graphon @xmath446 is not finitely forcible .
[ cor - grand ] for every @xmath207 , the graphon @xmath473 , which is associated with the permuton @xmath266 , is not finitely forcible .
we have shown that graphons associated with finitely forcible permutons need not be finitely forcible . in @xcite ,
question 11 asks whether there exists a `` 2-dimensional '' finitely forcible graphon and such graphons naturally arise from permutons .
jan volec with the first and the last authors @xcite constructed a finitely forcible graphon where the minkowski dimension of the associated topological space of typical points is two but the space is not connected . our discussions with the authors of @xcite led to the intuition that the graphon @xmath474 is a good candidate for a finitely forcible graphon with the associated space being connected and having dimension two .
is the graphon @xmath475 associated with the permuton @xmath34 finitely forcible ?
more generally , we suspect that all graphons associated with permutons @xmath476 constructed at the end of section [ sect - finite ] might be finitely forcible .
let @xmath196 be a square matrix of order @xmath16 with entries being non - negative reals such the sum of the entries in the @xmath50-th row is equal to that in the @xmath50-column and the sum of all the entries of @xmath196 is one .
is the graphon @xmath477 associated with the permuton @xmath476
finitely forcible ?
this work was done when the second author was visiting the university of warwick . the hospitality provided to him by the university
is strongly appreciated .
the authors would also like to thank jan volec for his comments on the topics related to those covered in the paper . c. borgs , j.t .
chayes , l. lovsz , v.t .
ss , and k. vesztergombi : convergent sequences of dense graphs i : subgraph frequencies , metric properties and testing , _ advances in mathematics _ * 219 * ( 2008 ) , 18011851 . c. borgs , j. chayes , l. lovsz , v.t .
ss , b. szegedy , and k. vesztergombi : graph limits and parameter testing , in : _ proceedings of the 38rd annual acm symposium on the theory of computing ( stoc ) , acm , new york , 2006 _ , 261270 .
presutti and w.r .
stromquist : packing rates of measures and a conjecture for the packing density of 2413 , in : _ permutation patterns , london mathematical society lecture note series _ * 376 * ( 2010 ) , 340 .
a. thomason : random graphs , strongly regular graphs and pseudorandom graphs , in : c. whitehead ( ed . ) , surveys in combinatorics 1987 , london mathematical society lecture note series , 123 , cambridge univ . press , cambridge ( 1987 ) , 173-195 . | we investigate when limits of graphs ( graphons ) and permutations ( permutons ) are uniquely determined by finitely many densities of their substructures , i.e. , when they are finitely forcible .
every permuton can be associated with a graphon through the notion of permutation graphs .
we find permutons that are finitely forcible but the associated graphons are not .
we also show that all permutons that can be expressed as a finite combination of monotone permutons and quasirandom permutons are finitely forcible , which is the permuton counterpart of the result of lovsz and ss for graphons . |
wave propagation in granular crystals has been studied quite intensively in the past ten years .
granular crystals are thought to be closely - packed chains of elastically interacting particles , which obey the fermi - pasta - ulam ( fpu ) lattice equations with hertzian interaction forces .
experimental work with granular crystals and their numerous applications @xcite stimulated theoretical and mathematical research on the granular chains of particles .
existence of solitary waves in granular chains was considered with a number of analytical and numerical techniques . in his two - page note ,
mackay @xcite showed how to adopt the technique of friesecke and wattis @xcite to the proof of existence of solitary waves .
english and pego @xcite used these results to prove the double - exponential decay of spatial tails of solitary waves .
numerical convergence to the solitary wave solutions was studied by ahnert and pikovsky @xcite .
stefanov and kevrekidis @xcite reviewed the variational technique of @xcite and proved that the solitary waves are bell - shaped ( single - humped ) .
recently , the interest to granular crystals has shifted towards periodic travelling waves as well as travelling waves in heterogeneous ( dimer ) chains , as more relevant for physical experiments @xcite .
periodic wave solutions of the differential advance - delay equation were considered by james in the context of newton s cradle @xcite and homogeneous granular crystals @xcite .
in particular , numerical approximations in @xcite suggested that periodic waves with wavelength larger than a critical value are spectrally unstable .
convergence to solitary waves in the limit of infinite wavelengths was also illustrated numerically and asymptotically in @xcite .
more recent work @xcite showed non - existence of time - periodic breathers in homogeneous granular crystals and existence of these breathers in newton s cradle , where a discrete @xmath0-schrdinger equation provides a robust approximation .
periodic waves in a chain of finitely many beads closed in a periodic loop were approximated by starosvetsky _
_ in monomers @xcite and dimers @xcite by using numerical techniques based on poincar maps .
interesting enough , solitary waves were found in the limit of zero mass ratio between lighter and heavy beads in @xcite .
it is explained in @xcite that these solitary waves are in resonance with linear waves and hence they do not persist with respect to the mass ratio parameter .
numerical results of @xcite indicate the existence of a countable set of the mass ratio parameter values , for which solitary waves should exist , but no rigorous studies of this problem have been developed so far .
recent work @xcite contains numerical results on existence of periodic travelling waves in granular dimer chains . in our present work ,
we rely on the anti - continuum limit of the fpu lattice , which was recently explored in the context of existence and stability of discrete multi - site breathers by yoshimura @xcite . by using a variant of the implicit function theorem
, we prove that every limiting periodic wave is uniquely continued with respect to the mass ratio parameter . by the perturbation theory arguments ( which are similar to the recent work in @xcite )
, we also show that the periodic waves with the wavelength larger than a certain critical value are spectrally stable .
our results are different from the asymptotic calculations in @xcite , where a different limiting solution is considered in the anti - continuum limit .
the family of periodic nonlinear waves bifurcating from the anti - continuum limit are shown numerically to extend all way to the limit of equal masses for the dimer beads .
the periodic travelling waves of the homogeneous ( monomer ) chains considered in @xcite are different from the periodic waves extended here from the anti - continuum limit . in other words , the periodic waves in dimers
do not satisfy the reductions to the periodic waves in monomers even if the mass ratio is one .
similar travelling waves consisting of binary oscillations in monomer chains were considered a while ago with the center manifold reduction methods @xcite .
the paper is organized as follows .
section 2 introduces the model and sets up the scene for the search of periodic travelling waves .
continuation from the anti - continuum limit is developed in section 3 .
section 4 gives perturbative results that characterize floquet multipliers in the spectral stability problem associated with the periodic waves near the anti - continuum limit .
numerical results are collected together in section 5 .
section 6 concludes the paper .
we consider an infinite granular chain of spherical beads of alternating masses ( a so - called _ dimer _ ) , which obey newton s equations of motion , @xmath1 where @xmath2 and @xmath3 are masses of light and heavy beads with coordinates @xmath4 and @xmath5 , respectively , whereas @xmath6 is the interaction potential .
the potential @xmath6 represents the hertzian contact forces for perfect spheres and is given by @xmath7 where @xmath8 and @xmath9 is the heaviside step function with @xmath10 for @xmath11 and @xmath12 for @xmath13 .
the mass ratio is modeled by the parameter @xmath14 . using the substitution , @xmath15 we rewrite the system of newton s equations ( [ fpu - before ] ) in the equivalent form : @xmath16 the value @xmath17 correspond to the anti - continuum limit , when the heavy particles do not move . at the limit of equal mass ratio @xmath18
, we note the reduction , @xmath19 for which the system of two granular chains ( [ eq : add ] ) reduces to the scalar granular chain ( a so - called _ monomer _ ) , @xmath20 the system of dimer equations ( [ eq : add ] ) has two symmetries .
one symmetry is the translational invariance of solutions with respect to @xmath21 , that is , if @xmath22 is a solution of ( [ eq : add ] ) , then @xmath23 is also a solution of ( [ eq : add ] ) for any @xmath24 .
the other symmetry is a uniform shift of coordinates @xmath25 in the direction of @xmath26 , that is , if @xmath22 is a solution of ( [ eq : add ] ) , then @xmath27 is also a solution of ( [ eq : add ] ) for any @xmath28 .
the system of dimer equations ( [ eq : add ] ) has the symplectic structure @xmath29 where the hamiltonian function is @xmath30 written in canonical variables @xmath31 .
we shall consider @xmath32-periodic solutions of the dimer system ( [ eq : add ] ) , @xmath33 travelling waves correspond to the special solution to the dimer system ( [ eq : add ] ) , which satisfies the following reduction , @xmath34 where @xmath35 $ ] is a free parameter .
we note that the constraints ( [ periodicity ] ) and ( [ traveling ] ) imply that there exists @xmath32-periodic functions @xmath36 and @xmath37 such that @xmath38 in this context , @xmath39 is inverse proportional to the wavelength of the periodic wave over the chain @xmath40 . the functions @xmath36 and @xmath37 satisfy the following system of differential advance - delay equations : @xmath41 a more general traveling periodic wave can be sought in the form @xmath42 where @xmath43 is an arbitrary parameter .
however , the parameter @xmath44 can be normalized to one thanks to invariance of the system of dimer equations ( [ eq : add ] ) with respect to a scaling transformation . for particular values @xmath45 , where @xmath2 and @xmath46 are positive integers such that @xmath47 , periodic travelling waves satisfy a system of @xmath48 second - order differential equations that follows from the system of lattice differential equations ( [ eq : add ] ) subject to the periodic conditions : @xmath49 this reduction is useful for analysis of stability of periodic travelling waves and for numerical approximations .
let @xmath50 be a solution of the nonlinear oscillator equation , @xmath51 because @xmath52 , bootstrapping arguments show that if there exists a classical @xmath32-periodic solution of the differential equation ( [ oscillator ] ) , then @xmath53 .
the nonlinear oscillator equation ( [ oscillator ] ) has the first integral , @xmath54 the phase portrait of the nonlinear oscillator ( [ oscillator ] ) on the @xmath55-plane consists of a family of closed orbits around the only equilibrium point @xmath56 .
each orbit corresponds to the @xmath57-periodic solution for @xmath50 , where @xmath57 is determined uniquely by energy @xmath58 .
it is well - known @xcite that , for @xmath59 , the period @xmath57 is a monotonically decreasing function of @xmath58 such that @xmath60 as @xmath61 and @xmath62 as @xmath63 .
therefore , there exists a unique @xmath64 such that @xmath65 for this @xmath66 .
we also know that the nonlinear oscillator ( [ oscillator ] ) is non - degenerate in the sense that @xmath67 ( to be more precise , @xmath68 ) .
in what follows , we only consider @xmath32-periodic functions @xmath50 which are defined by ( [ energy ] ) for @xmath66 . for uniqueness arguments , we shall consider initial conditions @xmath69 and @xmath70 , which determine uniquely one of the two odd @xmath32-periodic functions @xmath50 .
the limiting @xmath32-periodic travelling wave solution at @xmath17 should satisfy the constraints ( [ traveling ] ) , which we do by choosing for any fixed @xmath35 $ ] , @xmath71 to prove the persistence of this limiting solution in powers of @xmath72 within the granular dimer chain ( [ eq : add ] ) , we shall work in the sobolev spaces of odd @xmath32-periodic functions for @xmath73 , @xmath74 and in the sobolev spaces of @xmath32-periodic functions with zero mean for @xmath75 , @xmath76 the constraints in ( [ space-1 ] ) and ( [ space-2 ] ) reflects the presence of two symmetries ( [ symmetry1 ] ) and ( [ symmetry2 ] ) .
the two symmetries generate a two - dimensional kernel of the linearized operators . under the constraints in ( [ space-1 ] ) and ( [ space-2 ] ) ,
the kernel of the linearized operators is trivial , zero - dimensional .
it will be clear from analysis that the vector space @xmath77 defined by ( [ space-2 ] ) is not precise enough to prove the persistence of travelling wave solutions satisfying the constraints ( [ traveling ] ) . instead of this space , for any fixed @xmath35 $ ]
, we introduce the vector space @xmath78 by @xmath79 we note that @xmath80 , because if the constraint @xmath81 is satisfied , then the @xmath32-periodic function @xmath82 has zero mean . before developing persistence analysis , we shall point out three remarkable explicit periodic travelling solutions of the granular dimer chain ( [ eq : add ] ) for @xmath83 , @xmath84 and @xmath85 . for @xmath84
, we have an exact solution @xmath86 this solution preserves the constraint @xmath87 in equations ( [ eq : add ] ) thanks to the symmetry @xmath88 on the @xmath32-periodic solution of the nonlinear oscillator equation ( [ oscillator ] ) . for either @xmath83 or @xmath85
, we obtain another exact solution , @xmath89 by construction , these solutions ( [ exact-1 ] ) and ( [ exact-2 ] ) persist for any @xmath90 .
we shall investigate if the continuations are unique near @xmath17 for these special values of @xmath39 and if there is a unique continuation of the general limiting solution ( [ limit - solution ] ) in @xmath72 for any other fixed value of @xmath35 $ ] .
furthermore , we note that the exact solution ( [ exact-2 ] ) for @xmath85 at @xmath18 satisfies the constraint ( [ scalar - constraint ] ) with @xmath91 .
this reduction indicates that the function ( [ exact-2 ] ) for @xmath18 satisfies the granular monomer chain ( [ eq : mono ] ) and coincides with the solution considered by james @xcite . on the other hand , the exact solutions ( [ exact-1 ] ) for @xmath84 and ( [ exact-2 ] ) for @xmath83 do not produce any solutions of the monomer chain at @xmath18
this indicates that there exists generally two distinct solutions at @xmath18 , one is continued from @xmath17 and the other one is constructed from the solution of the monomer chain ( [ eq : mono ] ) in @xcite .
we consider the system of differential advance - delay equations ( [ eq : advance - delay ] ) .
the limiting solution ( [ limit - solution ] ) becomes now @xmath92 where @xmath50 is a unique odd @xmath32-periodic solution of the nonlinear oscillator equation ( [ oscillator ] ) with @xmath70 .
we now formulate the main result of this section .
fix @xmath35 $ ] .
there is a unique @xmath93 continuation of @xmath32-periodic traveling wave ( [ limit - solution - advanced - delay ] ) in @xmath72 , that is , there is a @xmath94 such that for every @xmath95 , there are @xmath96 and a unique @xmath32-periodic solution @xmath97 of the system of differential advance - delay equations ( [ eq : advance - delay ] ) such that @xmath98 [ theorem-1 ] by theorem [ theorem-1 ] , the limiting solution ( [ limit - solution - advanced - delay ] ) for @xmath99 is uniquely continued in @xmath72 . these continuations coincide with the exact solutions ( [ exact-1 ] ) and ( [ exact-2 ] ) .
let us first consider formal expansions in powers of @xmath72 to understand the persistence analysis from @xmath17 .
expanding the solution of the system of differential advance - delay equations ( [ eq : advance - delay ] ) , we write @xmath100 and obtain the linear inhomogeneous equations @xmath101 and @xmath102 because @xmath6 is @xmath103 but not @xmath104 , we have to truncate the formal expansion ( [ expansion ] ) at @xmath105 to indicate that there are obstacles to continue the power series beyond terms of the @xmath106 order .
let us consider two differential operators @xmath107 as a consequence of two symmetries , these operators are not invertible because they admit one - dimensional kernels , @xmath108 note that the kernel of @xmath109 is one - dimensional under the constraint @xmath67 ( see lemma 3 in @xcite for a review of this standard result ) . to find uniquely solutions of the inhomogeneous equations ( [ inhomogeneous-1 ] ) and ( [ inhomogeneous-2 ] ) in function spaces @xmath110 and @xmath111 respectively , see ( [ space-1 ] ) and ( [ space-2 ] ) for definition of function spaces , the source terms must satisfy the fredholm conditions @xmath112 the first fredholm condition is satisfied , @xmath113 d \tau = \int_0^{2\pi } v'(\varphi(\tau + 2q ) ) d \tau - \int_0^{2\pi } v'(-\varphi(\tau ) ) d \tau = 0,\ ] ] because the mean value of a periodic function is independent on the limits of integration and the function @xmath50 is odd in @xmath21 . since @xmath114 , there is a unique solution @xmath115 of the linear inhomogeneous equation ( [ inhomogeneous-1 ] ) .
the second fredholm condition is satisfied , @xmath116 d \tau = 0,\ ] ] if the function @xmath117 is odd in @xmath21 .
if this is the case , then @xmath118 and there is a unique solution @xmath119 of the linear inhomogeneous equation ( [ inhomogeneous-2 ] ) . to show that @xmath117 is odd in @xmath21 , we will prove that @xmath120 satisfies the reduction @xmath121 it follows from the linear inhomogeneous equation ( [ inhomogeneous-1 ] ) that @xmath122 where the last equality appears because @xmath50 is odd in @xmath21 . integrating this equation twice and using the fact that @xmath123 , we obtain reduction ( [ reduction - w ] ) .
note that the reduction ( [ reduction - w ] ) implies that @xmath124 , where @xmath125 is given by ( [ space-3 ] ) . to prove theorem [ theorem-1 ]
, we shall consider the vector fields of the system of differential advance - delay equations ( [ eq : advance - delay ] ) , @xmath126 we are looking for a strong solution @xmath127 of the system ( [ eq : advance - delay ] ) satisfying the reduction , @xmath128 that is , @xmath129 and @xmath130 . if @xmath131 , then @xmath132 is odd in @xmath21 .
furthermore , since @xmath6 is @xmath103 , then @xmath132 is a @xmath93 map from @xmath133 to @xmath134 and its jacobian at @xmath17 is given by @xmath135 on the other hand , under the constraints ( [ reduction - u - w ] ) , we have @xmath136 , because @xmath137 moreover , under the constraints ( [ reduction - u - w ] ) , we actually have @xmath138 because @xmath139 since @xmath6 is @xmath103 , then @xmath140 is a @xmath93 map from @xmath141 to @xmath142 and its jacobian at @xmath17 is given by @xmath143 let us now define the nonlinear operator @xmath144 we have @xmath145 are @xmath93 near the point @xmath146 . to apply the implicit function theorem near this point , we need @xmath147 at @xmath148 and the invertibility of the jacobian operator @xmath149 with respect to @xmath150 near @xmath151 .
it follows from ( [ jacobian-1 ] ) and ( [ jacobian-2 ] ) that the jacobian operator of @xmath149 at @xmath151 is given by the diagonal matrix of operators @xmath109 and @xmath152 defined by ( [ operator-1 ] ) and ( [ operator-2 ] )
. the kernels of these operators in ( [ kernels ] ) are zero - dimensional in the constrained vector spaces ( [ space-1 ] ) and ( [ space-2 ] ) ( we actually use space ( [ space-3 ] ) in place of space ( [ space-2 ] ) ) . by the implicit function theorem
, there exists a @xmath93 continuation of the limiting solution ( [ limit - solution - advanced - delay ] ) with respect to @xmath72 as the @xmath32-periodic solutions @xmath97 of the system of differential advance - delay equations ( [ eq : advance - delay ] ) near @xmath17 . from the explicit expression ( [ vector - field ] ) , we can see that @xmath153 whereas @xmath154 as @xmath155 .
the proof of theorem [ theorem-1 ] is complete .
we shall consider the dimer chain equations ( [ eq : add ] ) , which admit for small @xmath156 the periodic traveling waves in the form ( [ advance - delay ] ) , where @xmath157 is defined by theorem [ theorem-1 ] .
linearizing the system of nonlinear equations ( [ eq : add ] ) at the periodic traveling waves ( [ advance - delay ] ) , we obtain the system of linearized dimer equations for small perturbations , @xmath158 where @xmath40 . a technical complication
is that @xmath159 is continuous but not continuous differentiable .
this will complicate our analysis of perturbation expansions for small @xmath156 . note that the technical complications does not occur for exact solutions ( [ exact-1 ] ) and ( [ exact-2 ] ) .
indeed , for exact solution ( [ exact-1 ] ) with @xmath84 , the linearized system ( [ eq : add - linear ] ) is rewritten explicitly as @xmath160 for exact solution ( [ exact-2 ] ) with @xmath83 or @xmath85 , the linearized system ( [ eq : add - linear ] ) is rewritten explicitly as @xmath161 in both cases , the linearized systems ( [ eq : add - linear-1 ] ) and ( [ eq : add - linear-2 ] ) are analytic in @xmath72 near @xmath17 .
the system of linearized equations ( [ eq : add - linear ] ) has the same symplectic structure ( [ symplectic ] ) , but the hamiltonian is now given by @xmath162 the hamiltonian @xmath9 is quadratic in canonical variables @xmath31 . because coefficients of the linearized dimer equations ( [ eq : add - linear ] ) are @xmath32-periodic in @xmath21 , we shall look for an infinite - dimensional analogue of the floquet theorem that states that all solutions of the linear system with @xmath32-periodic coefficients satisfies the reduction @xmath163 where @xmath164 $ ] and @xmath165 is the monodromy operator .
let @xmath45 for some positive integers @xmath2 and @xmath46 such that @xmath47 . in this case , the system of dimer equations ( [ eq : add ] )
can be closed into a chain of @xmath48 second - order differential equations subject to the periodic boundary conditions ( [ boundary - conditions ] ) .
similarly , the linearized system ( [ eq : add - linear ] ) can also be closed as a system of @xmath48 second - order equations and the monodromy operator @xmath165 becomes an infinite diagonal composition of a @xmath166-by-@xmath166 floquet matrix , each matrix has @xmath166 eigenvalues called the floquet multipliers .
we can find eigenvalues of the monodromy operator @xmath167 by looking for the set of eigenvectors in the form , @xmath168 where @xmath169 are @xmath32-periodic functions and the admissible values of @xmath170 are found from the existence of these @xmath32-periodic functions .
the admissible values of @xmath170 are called the _ characteristic exponents _ and they define the floquet multipliers @xmath171 by the standard formula @xmath172 .
eigenvectors ( [ floquet - eigenvectors ] ) are defined as @xmath32-periodic solutions of the linear eigenvalue problem , @xmath173 the krein signature , which plays an important role in the studies of spectral stability of periodic solutions ( see section4 in @xcite ) , is defined as the sign of the @xmath174-form associated with the symplectic structure ( [ symplectic ] ) : @xmath175,\ ] ] where @xmath31 is an eigenvector ( [ floquet - eigenvectors ] ) associated with an eigenvalue @xmath176 .
note that by the symmetry of the linear eigenvalue problem ( [ eq : add - eigenvalue ] ) , it follows that if @xmath170 is an eigenvalue , then @xmath177 is also an eigenvalue , whereas the @xmath174-form @xmath178 is constant with respect to @xmath179 .
if @xmath17 , the monodromy operator @xmath165 in ( [ floquet - reduction ] ) is block - diagonal and consists of an infinite set of @xmath174-by-@xmath174 jordan blocks , because the dimer system ( [ eq : add ] ) is decoupled into a countable set of uncoupled second - order differential equations . as a result , the linear eigenvalue problem ( [ eq : add - eigenvalue ] ) with the limiting solution ( [ limit - solution ] ) admits an infinite set of @xmath32-periodic solutions for @xmath180 , @xmath181 where @xmath182 are arbitrary coefficients .
besides eigenvectors ( [ floquet - limit ] ) , there exists another countable set of generalized eigenvectors for each of the uncoupled second - order differential equations , which contribute to @xmath174-by-@xmath174 jordan blocks .
each block corresponds to the double floquet multiplier @xmath183 or the double characteristic exponent @xmath180 .
when @xmath184 but @xmath185 , the characteristic exponent @xmath180 of a high algebraic multiplicity splits .
we shall study the splitting of the characteristic exponents @xmath170 by the perturbation arguments .
we now formulate the main result of this section .
fix @xmath45 for some positive integers @xmath2 and @xmath46 such that @xmath47 .
let @xmath97 be defined by theorem [ theorem-1 ] for sufficiently small positive @xmath72 .
consider the linear eigenvalue problem ( [ eq : add - eigenvalue ] ) subject to @xmath48-periodic boundary conditions ( [ boundary - conditions ] ) .
there is a @xmath94 such that , for every @xmath95 , there exists @xmath186 such that for every @xmath187 or @xmath188 $ ] , no values of @xmath170 with @xmath189 exist , whereas for every @xmath190 , there exist some values of @xmath170 with @xmath191 .
[ theorem-2 ] by theorem [ theorem-2 ] , periodic traveling waves are spectrally stable for @xmath187 and @xmath188 $ ] and unstable for @xmath190 .
therefore , the linearized system ( [ eq : add - linear-1 ] ) for the exact solution ( [ exact-1 ] ) with @xmath84 subject to @xmath192-periodic boundary conditions ( @xmath193 and @xmath194 ) is unstable for small @xmath156 , where the linearized system ( [ eq : add - linear-2 ] ) for the exact solution ( [ exact-2 ] ) with @xmath85 subject to @xmath174-periodic boundary conditions ( @xmath193 and @xmath195 ) is stable for small @xmath156 .
the result of theorem [ theorem-2 ] is expected to hold for all values of @xmath39 in @xmath196 $ ] but the spectrum of the linear eigenvalue problem ( [ eq : add - eigenvalue ] ) for the characteristic exponent @xmath170 becomes continuous and connected to zero .
an infinite - dimensional analogue of the perturbation theory is required to study eigenvalues of the monodromy operator @xmath165 in this case .
the case @xmath83 is degenerate for an application of the perturbation theory .
nevertheless , we show numerically that the linearized system ( [ eq : add - linear-2 ] ) for the exact solution ( [ exact-2 ] ) with @xmath83 ( @xmath193 and @xmath197 ) is stable for small @xmath156 and all characteristic exponents are at least double for any @xmath156 .
we would normally expect splitting @xmath198 if the limiting linear eigenvalue problem at @xmath17 is diagonally decomposed into @xmath174-by-@xmath174 jordan blocks @xcite .
however , in the linearized dimer problem ( [ eq : add - eigenvalue ] ) , this splitting occurs in a higher order , that is , @xmath199 , because the coupling between the particles of equal masses shows up at the @xmath106 order of the perturbation theory .
regular perturbation computations in @xmath106 would require @xmath159 to be at least @xmath93 , which we do not have .
in the computations below , we neglect this discrepancy , which is valid at least for @xmath84 and @xmath85 . for other values of @xmath39 ,
the formal perturbation expansion is justified with the renormalization technique ( section 4.6 ) .
we expand @xmath32-periodic solutions of the linear eigenvalue problem ( [ eq : add - eigenvalue ] ) into power series of @xmath72 : @xmath200 and @xmath201 where the zeroth - order terms are given by ( [ floquet - limit ] ) . to determine corrections of the power series expansions uniquely
, we shall require that @xmath202 indeed , if @xmath203 contains a component , which is parallel to @xmath204 , then the corresponding term only changes the value of @xmath205 in the eigenvector ( [ floquet - limit ] ) , which is yet to be determined
. similarly , if a @xmath32-periodic function @xmath206 has a nonzero mean value , then the mean value of @xmath206 only changes the value of @xmath207 in the eigenvector ( [ floquet - limit ] ) , which is yet to be determined . the linear equations ( [ eq : add - eigenvalue ] ) are satisfied at the @xmath208 order . collecting terms at the @xmath209 order , we obtain @xmath210 let us define solutions of the following linear inhomogeneous equations : @xmath211 if we can find uniquely @xmath32-periodic solutions of these equations such that @xmath212 then the perturbation equations ( [ eq : add - order-1 ] ) at the @xmath209 order are satisfied with @xmath213 the linear equations ( [ eq : add - eigenvalue ] ) are now satisfied up to the @xmath209 order .
collecting terms at the @xmath106 order , we obtain @xmath214 where corrections @xmath215 and @xmath120 are defined by expansion ( [ expansion ] ) .
to solve the linear inhomogeneous equations ( [ eq : add - order-2 ] ) , the source terms have to satisfy the fredholm conditions because both operators @xmath109 and @xmath152 defined by ( [ operator-1 ] ) and ( [ operator-2 ] ) have one - dimensional kernels .
therefore , we require the first equation of system ( [ eq : add - order-2 ] ) to be orthogonal to @xmath204 and the second equation of system ( [ eq : add - order-2 ] ) to be orthogonal to @xmath216 on @xmath217 $ ] . substituting ( [ floquet - limit ] ) and ( [ floquet - order-1 ] ) to the orthogonality conditions and taking into account the symmetry between couplings of lattice sites on @xmath218 , we obtain the difference equations for @xmath219 : @xmath220 where @xmath221 , and @xmath222 are numerical coefficients to be computed from the projections . in particular , we obtain @xmath223 note that the coefficients @xmath224 and
@xmath225 need not to be computed at the diagonal terms @xmath205 and @xmath207 thanks to the fact that the difference equations ( [ projection - equations ] ) with @xmath226 must have eigenvectors with equal values of @xmath227 and @xmath228 , which correspond to the two symmetries of the linearized dimer system ( [ eq : add - linear ] ) related to the symmetries ( [ symmetry1 ] ) and ( [ symmetry2 ] ) .
this fact shows that the problem of limited smoothness of @xmath159 , which is @xmath229 but not @xmath93 near zero , is not a serious obstacle in the derivation of the reduced system ( [ projection - equations ] ) .
difference equations ( [ projection - equations ] ) give a necessary and sufficient condition to solve the linear inhomogeneous equations ( [ eq : add - order-2 ] ) at the @xmath106 order and to continue the perturbation expansions beyond this order . before justifying this formal perturbation theory
, we shall explicitly compute the coefficients @xmath222 of the difference equations ( [ projection - equations ] ) .
note that the system of difference equations ( [ projection - equations ] ) presents a quadratic eigenvalue problem with respect to the spectral parameter @xmath230 .
such quadratic eigenvalue problem appear often in the context of spectral stability of nonlinear waves @xcite .
we shall prove the following technical result .
coefficients @xmath231 , @xmath225 , @xmath232 , and @xmath233 are independent of @xmath39 and are given by @xmath234 consequently , @xmath235 , whereas @xmath236 .
on the other hand , coefficient @xmath224 depends on @xmath39 and is given by @xmath237 where @xmath238.\ ] ] [ lemma - coefficients ] to prove lemma [ lemma - coefficients ] , we first uniquely solve the linear inhomogeneous equations ( [ eq-1 ] ) , ( [ eq-2 ] ) , and ( [ eq-3 ] ) .
for equation ( [ eq-1 ] ) , we note that a general solution is @xmath239,\ ] ] where @xmath240 are arbitrary constants and @xmath241 is the derivative of the @xmath242-periodic solution @xmath243 of the nonlinear oscillator equation ( [ oscillator ] ) with the first integral ( [ energy ] ) satisfying initial conditions @xmath244 and @xmath245 at the value of energy @xmath66 , for which @xmath246 .
we note the equation @xmath247 that follows from the differentiation of equation @xmath248 with respect to @xmath58 at @xmath66 . to define @xmath249 uniquely
, we require that @xmath250 . because @xmath204 is even in @xmath21 ,
whereas @xmath251 and @xmath241 are odd , we hence have @xmath252 and @xmath253 .
hence @xmath249 is odd in @xmath21 and , in order to satisfy the @xmath32-periodicity , we shall only require @xmath254 , which uniquely specifies the value of @xmath255 by virtue of ( [ technical - equ ] ) , @xmath256 as a result , we obtain @xmath257.\ ] ] for equation ( [ eq-2 ] ) , we can use that @xmath258 for @xmath259 $ ] and @xmath260 for @xmath261 $ ] .
we can also use the symmetry @xmath262 . integrating equations for @xmath263 separately , we obtain @xmath264 , \\ c_+ \dot{\varphi } + d_+ \partial_e \varphi_{e_0 } , & \tau \in [ 0,\pi ] , \end{array } \right . \\
y_-(\tau ) & = & \left\ { \begin{array}{ll } a_- \dot{\varphi } + b_- \partial_e \varphi_{e_0 } , & \tau \in [ -\pi,0 ] , \\
1 + c_- \dot{\varphi } + d_- \partial_e \varphi_{e_0 } , & \tau \in [ 0,\pi ] .
\end{array } \right.\end{aligned}\ ] ] continuity of @xmath263 and @xmath265 across @xmath266 defines uniquely @xmath267 and @xmath268 . with this definition , @xmath269 , whereas condition @xmath270 sets up uniquely @xmath271 whereas constants @xmath272 are not specified . to define @xmath263 uniquely
, we again require that @xmath273 .
this yields the constraint on @xmath272 , @xmath274 as a result , we obtain @xmath275 , \\
\frac{\dot{\varphi}(\tau)}{\dot{\varphi}(0 ) } , & \tau \in [ 0,\pi ] , \end{array } \right.\ ] ] and @xmath276 , \\ 1 - \frac{\dot{\varphi}(\tau)}{\dot{\varphi}(0 ) } , & \tau \in [ 0,\pi ] , \end{array } \right.\ ] ] where @xmath277 are uniquely defined above . for equation ( [ eq-3 ] ) , we integrate separately on @xmath278 $ ] and @xmath196 $ ] to obtain @xmath279 , \\
c_+ , & \tau \in [ 0,\pi ] , \end{array } \right .
\quad \dot{z}_-(\tau ) = \left\ { \begin{array}{ll } c_- , & \tau \in [ -\pi,0 ] , \\
c_- + |\varphi(\tau)|^{\alpha } , & \tau \in [ 0,\pi ] , \end{array } \right.\ ] ] where @xmath280 are constants of integration and continuity of @xmath281 across @xmath266 have been used . to define @xmath282 uniquely
, we require that @xmath283 . integrating the equations above under this condition , we obtain : @xmath284 , \\
c_+ \tau - d_+ , & \tau \in [ 0,\pi ] , \end{array } \right .
\quad z_-(\tau ) = \left\ { \begin{array}{ll } c_- \tau + d_- , & \tau \in [ -\pi,0 ] , \\
c_- \tau - d_- - \dot{\varphi}(\tau ) , & \tau \in [ 0,\pi ] , \end{array } \right.\ ] ] where @xmath285 are constants of integration .
continuity of @xmath282 across @xmath266 uniquely sets coefficients @xmath286 .
periodicity of @xmath287 is satisfied .
periodicity of @xmath288 uniquely defines coefficients @xmath289 . as a result
, we obtain @xmath290 , \\
\dot{\varphi}(0 ) ( 2 \tau - \pi ) , & \tau \in [ 0,\pi ] , \end{array } \right.\ ] ] and @xmath291 , \\
-\dot{\varphi}(0 ) ( 2 \tau - \pi ) - 2 \pi \dot{\varphi}(\tau ) , & \tau \in [ 0,\pi ] .
\end{array } \right.\ ] ] we can now compute the coefficients @xmath222 of the difference equations ( [ projection - equations ] ) . for coefficients @xmath231 , we integrate by parts , use equations ( [ oscillator ] ) , ( [ energy ] ) , ( [ sol-1 ] ) , and obtain @xmath292 \biggr|_{\tau = -\pi}^{\tau = \pi } + \frac{2 \pi}{t'(e_0 ) } \int_{-\pi}^{\pi } \left ( \partial_e \varphi_{e_0 } \ddot{\varphi } - \partial_e \dot{\varphi}_{e_0 } \dot{\varphi } \right ) d \tau \\ & = & - \frac{4 \pi}{t'(e_0 ) } \int_{0}^{\pi } \partial_e \left ( \frac{1}{2 }
\dot{\varphi}^2 + \frac{1}{1 + \alpha } \varphi^{1 + \alpha } \right)_{e_0 } d \tau = - \frac{4 \pi^2}{t'(e_0)}.\end{aligned}\ ] ] because @xmath68 , we find that @xmath235 .
for @xmath224 , we use equation ( [ eq-3 ] ) , solution ( [ sol-3-minus ] ) , and obtain @xmath293 hence , the sign of @xmath224 depends on @xmath39 . using solution ( [ sol-3-minus ] ) , for @xmath294 $ ] , we obtain @xmath295 on the other hand , for @xmath296 $ ] , we obtain @xmath297 so that @xmath298 because the mean value of a periodic function does not depend on the limits of integration . for @xmath225 , we use equation ( [ eq-2 ] ) and obtain @xmath299 hence , @xmath300 . for @xmath232 , we use equations ( [ oscillator ] ) , ( [ energy ] ) , ( [ sol-2-minus ] ) , and obtain @xmath301 \\ & = & \frac{2 ( 2\pi - t'(e_0 ) ( \dot{\varphi}(0))^2)}{t'(e_0 ) \dot{\varphi}(0)}.\end{aligned}\ ] ] because @xmath70 and @xmath68 , we find that @xmath302 . for @xmath233 , we use equations ( [ oscillator ] ) , ( [ energy ] ) , ( [ sol-1 ] ) , and obtain @xmath303 \biggr|_{\tau = 0}^{\tau = \pi } \\ & = & \frac{2\pi - t'(e_0 ) ( \dot{\varphi}(0))^2}{\pi t'(e_0 ) \dot{\varphi}(0 ) } = \frac{1}{2 \pi } l_1,\end{aligned}\ ] ] hence , @xmath304 .
the proof of lemma [ lemma - coefficients ] is complete . because the coefficients @xmath222 of the difference equations ( [ projection - equations ] ) are independent of @xmath305 , we can solve these equations by the discrete fourier transform . substituting @xmath306 where @xmath307 $ ]
is the fourier spectral parameter , we obtain the system of linear homogeneous equations , @xmath308 a nonzero solution of system ( [ linear - equations ] ) exists if and only if @xmath230 is a root of the characteristic polynomial , @xmath309 since this equation is bi - quadratic , it has two pairs of roots for each @xmath307 $ ] . for @xmath310 ,
both pairs are zero , which recovers the characteristic exponent @xmath180 of algebraic multiplicity of ( at least ) 4 in the linear eigenvalue problem ( [ eq : add - eigenvalue ] ) . for a fixed @xmath311
, the two pairs of roots are generally nonzero , say @xmath312 and @xmath313 .
the following result specifies their location . there exists a @xmath314 such that @xmath315 for @xmath316 $ ] and @xmath317 for @xmath318 .
[ lemma - roots ] to classify the nonzero roots of the characteristic polynomial , we define @xmath319 the two pairs of roots are determined in table i. [ cols="<,<",options="header " , ] * table ii : * error between numerical and exact solutions for branch @xmath216 .
we can see from the top left panel of figure [ fig:4soln ] that a pitchfork bifurcation occurs at @xmath320 and results in the appearance of two symmetrically reflected branches ( labeled as branches @xmath174 and @xmath321 ) .
these branches with @xmath322 extend to @xmath18 ( bottom panels ) to recover two travelling wave solutions of the monomer chain ( [ eq : mono ] ) .
the solution of branch @xmath174 satisfies the travelling wave reduction @xmath323 and was previously approximated numerically by james @xcite .
the other solution of branch @xmath321 satisfies the travelling wave reduction @xmath324 and was previously obtained numerically by starosvetsky and vakakis @xcite . for @xmath194 ( @xmath84 )
, the solution of branch @xmath321 given by @xmath325 is obtained from the solution of branch @xmath174 given by @xmath326 , by means of the symmetry @xmath327 which holds for any @xmath156 .
( of course , both solutions @xmath174 and @xmath321 exist only for @xmath328 $ ] because of the pitchfork bifurcation at @xmath329 . )
the solution of branch @xmath216 is the invariant reduction @xmath330 , @xmath331 with respect to the symmetry ( [ four - particles - symmetry ] ) so that it satisfies @xmath332 for all @xmath333 . for six particles ( @xmath334 or @xmath335 ) , the nonlinear system ( [ eq : ode ] )
is written explicitly as @xmath336 we are looking for @xmath32-periodic functions satisfying the travelling wave reduction : @xmath337 we note that the system ( [ eq : six ] ) is invariant with respect to the following transformation : @xmath338 a @xmath32-periodic solution of this system satisfying ( [ reversibility - reduction - six ] ) must also satisfy @xmath339 , @xmath340 , and @xmath341 .
then , the constraints of the travelling wave reduction ( [ travelling - reductions - six ] ) yield the conditions @xmath342 and @xmath343 .
to approximate a solution of the initial - value problem for the nonlinear system ( [ eq : six ] ) satisfying ( [ reversibility - reduction - six ] ) , we only need six shooting parameters @xmath344 in the initial condition : @xmath345 this solution corresponds to a @xmath32-periodic travelling wave solution only if it satisfies the following six conditions : @xmath346 the six conditions determines the shooting method for the six parameters @xmath344 .
additional conditions , @xmath347 are to be checked aposteriori , after the shooting method has converged to a solution . : the solution of branch 1 is continued from @xmath348 to @xmath349 ( top right ) and the solution of branch 2 is continued from @xmath349 ( bottom left ) to @xmath350 ( bottom right ) . the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72 . , title="fig:",width=264,height=188 ] : the solution of branch 1 is continued from @xmath348 to @xmath349 ( top right ) and the solution of branch 2 is continued from @xmath349 ( bottom left ) to @xmath350 ( bottom right ) . the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72 . , title="fig:",width=264,height=188 ] + : the solution of branch 1 is continued from @xmath348 to @xmath349 ( top right ) and the solution of branch 2 is continued from @xmath349 ( bottom left ) to @xmath350 ( bottom right ) . the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72 . , title="fig:",width=264,height=188 ] : the solution of branch 1 is continued from @xmath348 to @xmath349 ( top right ) and the solution of branch 2 is continued from @xmath349 ( bottom left ) to @xmath350 ( bottom right ) . the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72 . , title="fig:",width=264,height=188 ]
figure [ fig:6soln ] ( top left ) shows two solution branches obtained by the shooting method .
again , @xmath351 is plotted versus @xmath72 .
branch 1 is continued from @xmath17 to @xmath18 ( top right ) without any pitchfork bifurcation in @xmath352 .
branch 2 is continued from @xmath18 ( bottom left ) starting with a numerical solution of the monomer chain ( [ eq : mono ] ) satisfying the reduction @xmath353 to @xmath350 ( bottom right ) , where the branch disappears from the radars of our shooting method .
we have not been able so far to detect numerically any other branch of travelling wave solutions near branch 2 for @xmath350 , hence the nature of this bifurcation will remain opened for further studies .
we use the same technique for @xmath354 and show similar results on figure [ fig:8soln ] .
branch 1 is uniquely continued from @xmath17 to @xmath18 ( top right ) , whereas branch 2 is continued from @xmath18 ( bottom left ) starting with a numerical solution of the monomer chain ( [ eq : mono ] ) satisfying the reduction @xmath355 to @xmath356 ( bottom right ) , where the branch terminates .
: the solution of branch 1 continued from @xmath348 to @xmath18 ( top right ) and the solution of branch 1 continued from @xmath349 ( bottom left ) to @xmath356 ( bottom right ) .
the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72.,title="fig:",width=264,height=188 ] : the solution of branch 1 continued from @xmath348 to @xmath18 ( top right ) and the solution of branch 1 continued from @xmath349 ( bottom left ) to @xmath356 ( bottom right ) .
the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72.,title="fig:",width=264,height=188 ] + : the solution of branch 1 continued from @xmath348 to @xmath18 ( top right ) and the solution of branch 1 continued from @xmath349 ( bottom left ) to @xmath356 ( bottom right ) .
the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72.,title="fig:",width=264,height=188 ] : the solution of branch 1 continued from @xmath348 to @xmath18 ( top right ) and the solution of branch 1 continued from @xmath349 ( bottom left ) to @xmath356 ( bottom right ) .
the top left panel shows the value of @xmath351 for solution branches 1 and 2 versus @xmath72.,title="fig:",width=264,height=188 ] to determine stability of the different branches of travelling periodic wave solutions of the granular dimer chains ( [ eq : add ] ) , we compute floquet multipliers of the monodromy matrix for the linearized system ( [ eq : add - linear ] ) . to do this
, we use the travelling wave solution obtained with the shooting method and the matlab function ` ode113 ` to compute the fundamental matrix solution of the linearized system ( [ eq : add - linear ] ) on the interval @xmath357 $ ] . by theorem [ theorem-2 ] ,
the travelling waves of branch @xmath216 for @xmath194 ( @xmath358 ) are unstable for small values of @xmath72 .
figure [ fig : n4stability ] ( top ) shows real and imaginary parts of the characteristic exponents associated with branch @xmath216 for all values of @xmath72 in @xmath359 $ ] .
only positive values of @xmath360 and @xmath361 are shown , moreover , @xmath362 $ ] because of @xmath216-periodicity of the characteristic exponents along the imaginary axis .
thanks to the periodic boundary conditions , the system of linearized equations ( [ eq : add - linear-1 ] ) for @xmath194 is closed at @xmath192 second - order linearized equations , which have @xmath363 characteristic exponents as follows .
the exponent @xmath180 has multiplicity @xmath192 for small positive @xmath72 , and two pairs of nonzero exponents ( one is real , the other one is purely imaginary ) bifurcate according to the roots of the characteristic equation ( [ characteristic - eq ] ) for @xmath364 .
these asymptotic approximations are shown on the top panels of figure [ fig : n4stability ] by solid curves , in excellent agreement with the numerical data .
we can see that the unstable real @xmath170 persist for all values of @xmath72 in @xmath359 $ ] .
the pitchfork bifurcation at @xmath320 in figure [ fig:4soln ] ( top left ) corresponds to the coalescence of the pair of purely imaginary characteristic exponents on figure [ fig : n4stability ] ( top right ) and appearance of a new pair of real characteristic exponents for @xmath365 on figure [ fig : n4stability ] ( top left ) .
therefore , the branch continued from @xmath17 is unstable for all @xmath366 $ ] .
bottom panels on figure [ fig : n4stability ] shows real and imaginary parts of the characteristic exponents associated with branch @xmath174 ( same for @xmath321 by symmetry ) for all values of @xmath72 in @xmath367 $ ] .
we can see that these travelling waves are spectrally stable near @xmath18 in agreement with the numerical results of james @xcite .
when @xmath72 is decreased , these travelling waves lose spectral stability near @xmath368 because of coallescence of the pair of purely imaginary characteristic exponents and appearance of a new pair of real characteristic exponents for @xmath369 .
the two solution branches disappear as a result of the pitchfork bifurcation at @xmath370 , which is again induced by the coalescence of the second pair of purely imaginary characteristic exponents . versus @xmath72 for @xmath194 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath194 for branch 1 ( top ) and
branch 2 ( bottom).,title="fig:",width=264,height=188 ] + versus @xmath72 for @xmath194 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath194 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath371 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath371 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] + versus @xmath72 for @xmath371 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath371 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] for @xmath371 ( @xmath372 ) , the system of linearized equations ( [ eq : add - linear-1 ] ) is closed at @xmath373 second - order linearized equations . besides the characteristic exponent @xmath180 of multiplicity four
, we have @xmath363 nonzero characteristic exponents @xmath170 . the characteristic equation ( [ characteristic - eq ] ) with @xmath374 and @xmath375
predicts a double pair of real @xmath170 and a double pair of purely imaginary @xmath170 .
figure [ fig : n6stability1 ] ( top ) shows @xmath360 ( left ) and @xmath361 ( right ) for solutions of branch 1 . the double pair of purely imaginary @xmath170 split along the imaginary axis for small @xmath156 . on the other hand , the double pair of real @xmath170 splits along the transverse direction and results in occurrence of a quartet of complex - valued @xmath170 for small @xmath156 .
these complex characteristic exponents approach the imaginary axis at @xmath376 ( neimark sacker bifurcation ) and then split along the imaginary axis as two pairs of purely imaginary @xmath170 for @xmath377 . at the same time , one pair of of the purely imaginary @xmath170 continued from @xmath17 approaches the line @xmath378 ( corresponding to the floquet multiplier at @xmath379 ) at @xmath380 ( period - doubling bifurcation ) and splits along the negative real axis . in summary ,
the periodic travelling wave of branch 1 for @xmath371 is stable for @xmath381 but unstable near @xmath17 and @xmath18 .
figure [ fig : n6stability1 ] ( bottom ) shows @xmath360 ( left ) and @xmath361 ( right ) for solutions of branch 2 that exists only for @xmath382 $ ] , where @xmath383 .
all four pairs of the characteristic exponents @xmath170 are purely imaginary near @xmath18 that corresponds to the numerical results for stability of travelling waves in monomer chains in @xcite .
two pairs coalesce at @xmath384 resulting in the complex characteristic exponents ( neimark sacker bifurcation ) .
one more pair crosses the line @xmath378 for @xmath385 resulting in the negative characteristic exponents ( period - doubling bifurcation ) . the last remaining pair of purely imaginary @xmath170 crosses zero near @xmath386 that indicates that termination of branch 2 is related to a local bifurcation .
however , we are not able to identify numerically any other branch of travelling wave solutions in the neighborhood of branch 2 for @xmath387 . versus @xmath72 for @xmath388 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath388 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] + versus @xmath72 for @xmath388 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath388 for branch 1 ( top ) and branch 2 ( bottom).,title="fig:",width=264,height=188 ] recall that the coefficient @xmath224 changes sign at @xmath389 , as seen in figure [ fig : coeff ] .
therefore , for @xmath390 , the characteristic equation ( [ characteristic - eq ] ) for any values of @xmath391 predicts pairs of purely imaginary @xmath170 only .
this is illustrated on the top panel of figure [ fig : n8stability ] for @xmath388 ( @xmath392 ) .
we can see that all double pairs of purely imaginary @xmath170 split along the imaginary axis for small @xmath156 and that the periodic travelling waves of branch 1 remain stable for all @xmath366 $ ] .
the figure also illustrate the validity of asymptotic approximations obtained from roots of the characteristic equation ( [ characteristic - eq ] ) .
it is interesting that figure [ fig : n8stability ] shows safe coalescence of characteristic exponents for larger values of @xmath72 . recall from remark [ remark - krein ] that the characteristic exponents have opposite krein signature for small values of @xmath72 in such a way that larger exponents on figure [ fig : n8stability ] have negative krein signature @xmath178 and smaller exponents have positive krein signature @xmath178 .
it is typical to observe instabilities after coalescence of two purely imaginary eigenvalues of the opposite krein signature @xcite but this only happens when the double eigenvalue at the coalescence point is not semi - simple .
when the double eigenvalue is semi - simple , the coalescence does not introduce any instabilities @xcite .
this is precisely what we observe on figure [ fig : n8stability ] .
after coalescence for larger values of @xmath72 , the purely imaginary characteristic exponents @xmath170 reappear as simple exponents of the opposite krein signature and the exponents with positive krein signatures are now above the ones with negative krein signatures .
figure [ fig : n8stability ] ( bottom ) shows @xmath360 ( left ) and @xmath361 ( right ) for solutions of branch 2 that exists only for @xmath382 $ ] , where @xmath393 . besides the pairs of purely imaginary characteristic exponents
@xmath170 , there exists one pair of real exponents @xmath170 near @xmath18 that corresponds to the numerical results for instability of travelling waves in monomer chains in @xcite . for
smaller values of @xmath72 , more instabilities arise for the solutions of branch 2 because of various bifurcations of pairs of purely imaginary exponents @xmath170 . versus @xmath72 for @xmath394 ( left ) and @xmath395 ( right ) .
the real part of all the exponents is zero.,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath394 ( left ) and @xmath395 ( right ) .
the real part of all the exponents is zero.,title="fig:",width=264,height=188 ] finally , figure [ fig : n10stability ] illustrate the stability of solutions of branch 1 for @xmath396 ( left ) and @xmath397 ( right ) .
not only the double pairs of purely imaginary @xmath170 split safely along the imaginary axis for small @xmath156 , various coalescence of purely imaginary exponents @xmath170 of opposite krein signature never result in occurrence of complex exponents @xmath170 .
the solutions of branch 1 remain stable for all @xmath366 $ ] .
the periodic solution with @xmath398 ( which is no longer a traveling wave but a uniform oscillation of all sites of the dimer ) is given by the exact solution ( [ exact-2 ] ) .
spectral stability of this solution is obtained from the system of linearized equations ( [ eq : add - linear-2 ] ) . using the boundary conditions @xmath399 where @xmath307 $ ] is a continuous parameter
, we obtain the system of two closed second - order equations , @xmath400 the characteristic equation ( [ characteristic - eq ] ) for @xmath83 predicts a double pair ( [ limiting - roots ] ) of purely imaginary @xmath230 for any @xmath311 .
we confirm here numerically that the double pair is preserved for all @xmath366 $ ] .
figure [ fig : q0stability ] shows the imaginary part of the characteristic exponents @xmath170 of the linearized system ( [ eq : add - linear-3 ] ) for @xmath364 ( left ) and @xmath401 ( right ) .
similar results are obtained for other values of @xmath391 .
therefore , the periodic solution with @xmath83 remains stable for all values of @xmath366 $ ] .
versus @xmath72 for @xmath364 ( left ) and @xmath401 ( right ) .
the real part of all the exponents is zero.,title="fig:",width=264,height=188 ] versus @xmath72 for @xmath364 ( left ) and @xmath401 ( right ) .
the real part of all the exponents is zero.,title="fig:",width=264,height=188 ] the pattern on figure [ fig : q0stability ] suggests a hidden symmetry in this case .
suppose @xmath402 is a characteristic exponent of the system ( [ eq : add - linear-3 ] ) for the eigenvector @xmath403 = \left [ \begin{array}{c } u_{\theta}(t ) \\
w_{\theta}(t ) \end{array } \right ] e^{\lambda_{\theta } t},\ ] ] where @xmath404 and @xmath405 are @xmath32-periodic and the subscript @xmath391 indicates that the system ( [ eq : add - linear-3 ] ) depends explicitly on @xmath391 . recall that the unperturbed solution satisfies the symmetry @xmath406 for all @xmath333 . using this symmetry and the trivial identity @xmath407 , we can verify that there is another solution of the system ( [ eq : add - linear-3 ] ) with the same @xmath391 for the characteristic exponent @xmath408 : @xmath409 = \left [ \begin{array}{c } u_{\pi - \theta}(t + \pi ) \\
e^{2i \theta } w_{\pi - \theta}(t+\pi ) \end{array } \right ] e^{\lambda_{\pi-\theta } t}.\ ] ] from the symmetry of roots ( [ limiting - roots ] ) and the corresponding characteristic exponents , we have @xmath410 .
the eigenvectors ( [ eigenvector-1-symmetry ] ) and ( [ eigenvector-2-symmetry ] ) are generally linearly independent and coexist for the same value of @xmath411 .
this argument explains the double degeneracy of characteristic exponents @xmath170 for the case @xmath83 for all values of @xmath366 $ ] .
we have studied periodic travelling waves in granular dimer chains by continuing these solutions from the anti - continuum limit , when the mass ratio between the light and heavy beads is zero .
we have shown that every limiting periodic wave is uniquely continued for small mass ratio parameters .
although the vector fields of the granular dimer chain equations are not smooth , we can still use the implicit function theorem to guarantee that the continuation is @xmath93 with respect to the mass ratio parameter .
we have also used rigorous perturbation theory to compute characteristic exponents in the linearized stability problem . from this theory ,
we have seen that the periodic waves with the wavelength larger than a certain critical value are spectrally stable for small mass ratios .
numerical computations are developed to show that the stability of these periodic waves with larger wavelengths extends all way to the limit of equal mass ratio .
on the other hand , we have also computed periodic travelling waves that are continued from solutions of the granular monomer chains at the equal mass ratio , their spectral stability , and their terminations for smaller mass ratios . among open problems
, we have not clarified the nature of bifurcation , where the solutions of branch 2 terminate at a @xmath412 for @xmath413 .
we have not been able to find another solution nearby for @xmath414 .
safe coalescence of purely imaginary characteristic exponents @xmath170 of opposite krein signatures is also amazing and we have not been able to explain the hidden symmetry that would explain why the eigenvalues at the coalescence point remain semi - simple .
these problems as well as analysis of the periodic travelling wave solutions for other values of @xmath39 will wait for further studies . | we study bifurcations of periodic travelling waves in granular dimer chains from the anti - continuum limit , when the mass ratio between the light and heavy beads is zero .
we show that every limiting periodic wave is uniquely continued with respect to the mass ratio parameter and the periodic waves with the wavelength larger than a certain critical value are spectrally stable .
numerical computations are developed to study how this solution family is continued to the limit of equal mass ratio between the beads , where periodic travelling waves of granular monomer chains exist . |
a common feature in x - ray power spectra from many x - ray binaries ( xrb ) are quasi - periodic oscillations ( qpo ) .
many `` flavours '' of qpo have been identified , with some correlations between their frequencies ( psaltis , belloni & van der klis 1999 ) .
for example , low frequency ( @xmath0 of a few hz ) qpo in black hole xrb seem to appear preferentially when the state of the source changes ( rutledge et al .
they are mostly observed in luminous states , high or very high state , when the energy spectrum is rather complex , showing both disc thermal component and high energy comptonized component ( see qpo observations review in wijnands 2001 ) .
models of qpo so far concentrated on identifying the frequencies of oscillations , with very little if any attention paid to the fact that it is the hard x rays which are modulated ( see qpo models review in psaltis 2001 ) . the results of gilfanov , revnivtsev & molkov ( 2003 ) and revnivtsev & gilfanov ( 2005 ) are perhaps the only ones where the energy spectra of qpo are discussed and interpreted .
these authors applied fourier frequency resolved spectroscopy ( revnivtsev , gilfanov & churazov 1999 , 2001 ) to show in particular that the kilo - hz qpo spectra in neutron star xrb are the same as spectra at other fourier frequencies .
this can be interpreted as common spatial origin of the x - rays showing the usual broad band noise variability and the more coherent quasi - periodic variability .
the spectral shape of these components is consistent with comptonized emission from a boundary layer .
fourier spectroscopy techniques are particularly suitable for analysis of qpo . however , the approach of revnivtsev et al .
( 1999 ) requires integrating the power density spectra over a narrow frequency interval around a chosen frequency resulting in contamination of the fourier frequency resolved spectrum of the qpo by a contribution from the broad band variability component .
thus , in this paper we construct a root - mean - square spectrum , where each point is the r.m.s .
variability integrated over the component ( e.g. a lorentzian ) describing the qpo in the broad band power spectrum .
such energy dependences of qpo amplitude in xrb has been investigated before ( e.g. belloni et al .
1997 ; cui et al .
1999 ; rao et al .
2000 ; tomsick & kaaret 2001 ; rodriguez et al .
2004 ; casella et al .
2004 ; see review in van der klis 2006 ) .
the main conclusion from these studies is that the qpo amplitude increases with energy up to at least @xmath1 kev .
this suggests immediately that the oscillations are connected with the high energy spectral component , rather than with the disk component .
[ cols="<,^,^,^,^",options="header " , ] + @xmath2 @xmath3/dof from fits of the time averaged spectrum model to the qpo spectra , allowing _ only _ the parameters of the reprocessed component and overall normalization to be free + @xmath4 best fit value of @xmath5 from fits @xmath2 + @xmath6 @xmath3/dof from fits of the thcomp model to the qpo spectra ( all parameters adjustable ) , _ without _ the reprocessed component + @xmath7 @xmath8 when cold ( @xmath9 ) reprocessed component is added , with no relativistic smearing ( i.e. , one additional parameter ) + @xmath10 best fit value of reflection amplitude ( or 95 per cent upper limit , @xmath11 , when @xmath12 in column @xmath7 ) + @xmath13 @xmath8 ( relative to fits @xmath6 ) when ionized reprocessed component is added ( 2 additional parameters ) + @xmath14 best fit value of @xmath5 from fits @xmath13 + @xmath15 best fit obtained for a pure reflected component ( @xmath16 ) ; 95 per cent lower limit shown + @xmath17 the disc blackbody component was included in the qpo model [ tab : qpo_meansp.refl ] comparing the time averaged and the qpo spectra we explicitly find that only one component , a thermal comptonization ( with its compton reflection ) is needed to describe the latter . however , the qpo spectra are in most cases significantly different than the comptonized component in the time averaged spectra .
qualitatively , we find an interesting anti - correlation : in soft spectral states the qpo spectra are harder than the mean spectra , while they are softer than the mean spectra in hard spectral states .
together with two intermediate cases this forms a continuous sequence .
quantitatively , the differences between the spectra can be described by a difference in spectral indices or seed photons temperatures , or both , with the statistical quality of the various descriptions being approximately the same .
the qpo spectra do contain the reprocessed component . in most datasets
their presence is highly statistically significant ( 99 per cent or better ) , although the reflection amplitude is rather poorly constrained .
because of that , it is not possible to determine if the amplitude is different than that in the time averaged spectrum . in paper
i we introduced a phenomenological model of qpo , where the background radial propagation model of broad band variability ( kotov et al .
2001 , ycki 2003 ) is supplemented with quasi - periodic modulation of one or more of the parameters of a comptonization spectrum .
the main parameters include the plasma heating rate , @xmath18 , the cooling rate by soft disc photons , @xmath19 ( here @xmath20 is the compactness parameters ) , and the amplitude of the reflected component ( describing the coupling between the heating and cooling ) .
our motivation for such a model is that irrespectively of the physical mechanism behind the qpo , the modulation of the emitted x - rays can only be produced by a modulation of one or more of the parameters actually determining the hard x - ray spectrum .
we can now discuss our observational results in the light of that model .
qpo spectra harder than the mean spectra can be obtained if the qpo are produced by modulation of the heating rate while the cooling rate does not respond fully to @xmath18 modulation .
this produces spectral pivoting around the low energy end of the spectrum , and , in consequence , r.m.s .
variability increasing with energy .
such a model might then be applicable to the soft spectral states discussed in this paper .
the model predicts a local maximum of the ew of the fe line at the qpo frequency ( see fig . 3 in paper i ) ,
but the quality of our data is not sufficient to test this prediction .
qpo spectra softer than the mean spectra can be produced by modulations driven by the cold disc .
one possibility here is a modulation of the cooling rate .
this results in the spectra pivoting around an energy point intermediate between the low and high energy ends of the spectrum ( see also zdziarski et al . 2002 ) . in the limited energy band ,
e.g. that of _
rxte_/pca , it may correspond to the r.m.s .
spectra decreasing with energy .
the cooling rate may be modulated with or without modulation of the seed photons temperature .
the difference between the two is that the qpo spectrum is monotonic in energy in the former case but it has a very deep minimum ( at the pivot energy ) in the latter case ( see figs .
4 and 5 in paper i ) .
additionally , qpo harmonics appear when @xmath21 is modulated , since the modulation of @xmath19 is then not sinusoidal ( even if modulation of @xmath21 is ) .
however , the harmonics are limited to the soft disc component only , unlike in some of our data , for example , xte j1550 - 564 ( i ) .
the other possibility of producing a soft qpo spectrum is to modulate the amplitude of the reflected component , @xmath5 , assuming that it also describes the feedback between illuminating hard x - rays and the re - emitted soft disc photons cooling the hot plasma . because of the coupling , this again generates modulations of @xmath19 and the characteristic spectral pivoting .
the most characteristic feature of this model is the strong fe line ( and the entire reprocessed component ) at @xmath22 , directly resulting from modulations of @xmath5 .
contrary to that , the model with @xmath19 modulation gives either generally weak fe line , or at least a minimum at @xmath22 .
the presence of the reprocessed component ( including the fe line ) in our qpo spectra makes modulation of @xmath5 an interesting possibility ( see also miller & homan 2005 ) .
we note that models from paper i involving modulation of @xmath19 predict very strong soft component in the qpo spectra , which is not observed .
it is rather hard to envision a geometry where the modulated soft flux would enter the hot plasma , but would not be directed towards an observer .
modulation of @xmath5 also produces some soft component in the qpo spectra , but it appears less prominent than the one from @xmath19 modulation models . clearly , further development of the models is necessary to fully describe the data .
another potentially useful diagnostics is the coherence function . while variations in different energy bands are perfectly coherent when @xmath5 is modulated , the coherence function shows a complex behaviour , with a number of minima , when @xmath19 and/or @xmath21 are modulated .
however , distinguishing between the models based on such predicted observables as the ew of the fe line , the coherence function or the hard x - ray time lags would require better data than currently available .
it needs to be emphasized that majority of physical models of qpo envision the qpo as driven by some kind of disc oscillations .
however , our interpretation of the energy dependencies of low-@xmath0 qpo , at least in soft spectral states , would rather point out to modulations of the heating rate in the hot plasma .
if so , then those models must find a way of transferring the disc oscillations to the hot plasma with 100 per cent efficiency , i.e. without affecting the disc emission . on the other hand ,
most of the models were formulated in the context of the high frequency qpo , where the energy dependencies might be different than those for the low-@xmath0 qpo .
the only work where a physical model of the low-@xmath0 qpo was constructed is that by giannios & spruit ( 2004 ) .
they consider oscillations of a hot ion - supported accretion flow coupled to a outer cold disc .
the coupling is realized by a number of channels : hard x - rays illuminating the cold disc , hot protons heating the cold disc , the soft disc photons cooling the hot flow .
spectral variability predicted by that model corresponds to our case of modulating the cooling rate , that is , the qpo spectra are softer than the time averaged spectra .
this would then be consistent with observed qpo behaviour in the hard state .
the overall geometry of a hot inner flow and an outer cold disc fits that state too .
considering that the usual broad band noise x - ray variability is driven by instabilities and flares in the hot plasma , it is obvious that an important piece of physical understanding is still missing .
* the qpo energy spectra are harder than the time average spectra in soft spectral states , but they are softer than the time averaged spectra in the hard state .
the qpo spectra are similar in slope to the time averaged spectra , when the latter are intermediate in slope between hard and soft ( @xmath23 ) . *
the disc component is absent in the qpo spectra .
* comparison of the observational data with simple models of spectral variability suggests that instabilities in the hot plasma drive the low-@xmath0 qpo in the soft state , while cold disc oscillations drive the qpo in the hard state .
this work was supported in part by polish committee of scientific research through grants number 1p03d01626 and 2p03d01225 . | we study energy dependencies of quasi - periodic oscillations from a number of black hole x - ray binaries .
the selected sources were observed by _ rxte _ at time periods close to state transitions and showed qpo in the 110 hz range .
we have constructed qpo root - mean - square energy spectra , which provide information about underlying physical process leading to qpo generation .
these spectra show an interesting anti - correlation with the time averaged spectra .
the qpo r.m.s .
spectra are harder than the time averaged spectra when the latter are soft , while they are softer than the time averaged spectra when the latter are hard .
we then discuss these observational results in the context of simple spectral variability models .
hard qpo spectra can be produced by quasi - periodic modulations of the heating rate of the comptonizing plasma , while soft qpo spectra result from modulations of the cooling rate by soft photons .
-1.5 cm [ firstpage ] accretion , accretion discs instabilities radiation mechanisms : thermal binaries : close x - rays : binaries |
critical behavior of absorbing phase transitions can be categorized by universality classes @xcite .
the most well - known example is the directed percolation ( dp ) class which may be ( loosely ) defined by a classification scheme known as the `` dp conjecture '' @xcite .
the conjecture states that models exhibiting a continuous transition into a _ single _ absorbing state should belong to the dp class unless symmetry or conservation is involved .
the dp conjecture is not complete in two folds .
first , some models with infinitely many absorbing states such as the pair contact process @xcite also belong to the dp class , at least in its stationary property , unless there is no clear - cut symmetry among absorbing states @xcite .
second , the conjecture itself does not resolve the controversy related to the pair contact process with diffusion ( pcpd ) @xcite .
however , this conjecture is useful enough to trigger a search for a non - dp universality class .
one can find several universality classes in ref .
@xcite which do not meet the constraint of the dp conjecture . among them , this paper interests in the system with two equivalent absorbing states ( or in general two equivalent groups of many absorbing states ) @xcite . the ising - like @xmath3 symmetry between two absorbing states which is sometimes manifest as a modulo 2 conservation of domain walls in one dimension renders the system to exhibit a non - dp critical behavior .
this universality class has several names such as the parity conserving @xcite , the directed ising ( di ) @xcite which will be used to refer to this class in this paper , the generalized voter class @xcite and so on . in higher dimensions ,
all these classes become distinct .
the ising symmetry ( or the modulo 2 conservation ) _ per se _ is not enough to force the di critical behavior .
it requires an infinite dynamic barrier between two equivalent ( group of ) absorbing states @xcite , which is analogous to the free energy barrier between two ordered states in the equilibrium ising model .
a state near one absorbing state can not evolve into the other absorbing state within a finite number of successive local changes .
in other words , a domain wall ( frustration ) in a configuration generated by pasting two absorbing states can not disappear by itself , so the system never becomes absorbing with a single frustration @xcite .
as an example without this additional feature , consider a one - dimensional system of diffusing particles with pair annihilation ( @xmath4 ) and branching of two particles by a triplet ( @xmath5 ) .
we may map this particle model onto the ferromagnetic ising spin model by interpreting a particle as a domain wall of the ising system as follows : @xmath6 where @xmath7 ( @xmath8 ) stands for a particle ( vacancy ) and @xmath9 ( @xmath10 ) represents a up ( down ) spin . clearly the number of domain walls is conserved modulo 2 and there are two absorbing states in the spin system ( all spins are up or down ) which are equivalent .
however , the states with only one ( diffusing ) particle are also absorbing in the particle model . in the spin language , these extra absorbing states can be obtained by connecting two equivalent absorbing states ( for example , @xmath11 ) , which violate the infinite dynamic barrier requirement for the di universality class mentioned above @xcite .
it is numerically shown that this system shares the critical behavior with the dp not the di @xcite , while a similar system with @xmath4 and @xmath12 belongs to the pcpd class @xcite .
it is claimed that a long - term memory plays a crucial role in differentiating these two universality classes @xcite . according to the above criterion for the di class
, one can conceive three possible routes along which the di critical behavior crosses over to the dp in one dimension . introducing a symmetry breaking field ( sb ) @xcite , allowing dynamics which breaks the modulo 2 conservation ( cb ) @xcite , or making channels connecting two equivalent absorbing states ( cc )
@xcite drives the system into the dp class .
but these three routes bring about different characteristics in the absorbing states .
the sb route maintains the absorbing states but breaks the probabilistic balance between them .
the cb route ( for example , by adding @xmath13 or @xmath14 ) invalidates the mapping between the spin and the particle model , and ends up with only one absorbing state ( vacuum ) .
the cc route is accompanied by additional ( infinitely many ) absorbing states connecting two equivalent absorbing states with both the symmetry and the mapping sustained .
to quantify the difference of these routes , we measure the crossover exponents governing the crossover behavior . in sec .
[ section : models ] , one - dimensional models are introduced belonging to the di class . in sec .
[ section : three ] , numerical results for the crossover exponents along three different routes are reported .
we discuss and summarize our results in sec .
[ section : discussion ] .
we choose two di models for studying crossover from the di class to the dp class : branching annihilating walks with two offspring ( baw2 ) @xcite and the interacting monomer model ( i m ) which is a simplified version of the interacting monomer - monomer model @xcite .
we study the macroscopic absorbing phase transitions @xcite inherent in the model .
the suitable order parameters will be defined when we explain each model in detail . for convenience , the order parameter for different models will be denoted by @xmath15 throughout the paper .
we only study one dimensional lattice systems with periodic boundary conditions . for the simulation of the baw2 , we take the dynamic branching rule of ref .
the simulation algorithm is detailed as follows : at first , we choose a particle residing at a lattice point randomly among , say , @xmath16 particles present at time @xmath17 .
it may hop to a randomly chosen nearest neighbor site with probability @xmath18 .
if the target site is already occupied , two particles are removed immediately . with probability @xmath19 ,
this selected particle branches two particles along ( color online ) semilogarithmic plot of @xmath20 vs @xmath17 near criticality for the baw2 .
the critical point is found to be @xmath21 with the figure in the parentheses as uncertainty of the last digit.,scaledwidth=45.0% ] one of two directions .
if an offspring is created on a site already occupied , both particles annihilate immediately . after the above steps , time increases by @xmath22 .
since the absorbing state is vacuum , the order parameter @xmath15 is the particle density . at criticality
, the density is expected to decay as @xmath23 with the di critical exponent @xmath24 @xcite .
figure [ fig : baw ] locates the critical point at @xmath21 which is more accurate than that obtained in ref .
@xcite .
the i m is a lattice model with two possible states at each lattice point ; either vacant or occupied by a particle .
the dynamics begins with the adsorption attempt at a randomly selected vacant site .
when both of nearest neighbor ( nn ) sites are vacant , a monomer is adsorbed with rate @xmath25 . in case
only one of the nn sites is occupied , the monomer is adsorbed on the substrate with rate @xmath26 , then reacts and leaves the substrate with the neighboring particle ( @xmath27 ) instantaneously . when both nn sites are already occupied , the adsorption attempt is rejected .
the dynamics on the substrate can be summarized as [ eq : alr ] @xmath28 where @xmath7 ( @xmath29 ) means the monomer occupied ( vacant ) site .
the absorbing states of the i m are characterized by the `` antiferromagnetically ordered '' state ( @xmath30 ) and have the @xmath3 symmetry ( poisoning even or odd sites by monomers ) .
starting with initial configurations without any @xmath31 pair , the proper order parameter @xmath15 is the density of @xmath32 pairs .
the mapping of the dynamics of the i m to that of the contact process with the modulo 2 conservation [ cp(2 ) ] @xcite is possible , though there are some subtleties which make these two models a little bit different .
if we assign a domain wall excitation , say @xmath33 , in the middle of the @xmath32 pair , the domain wall dynamics is exactly the same as the cp(2 ) rules .
a pair annihilation in the cp(2 ) corresponds to an adsorption process of a monomer as in eq . and a pair branching in the cp(2 ) to an annihilation process eq . .
note that the reaction always generates two domain walls ( two @xmath32 pairs ) , because an @xmath31 pair is not allowed in the i m .
however , there is no one - to - one correspondence between configurations of the cp(2 ) and the i m , in general .
for example , a domain wall configuration @xmath34 is not possible in the i m by definition , but possible in the cp(2 ) in principle .
nevertheless , if we start with the fully occupied domain wall state , such a configuration can not be generated by dynamics and thus the mapping becomes exact . in this case , the parameter @xmath35 in ref .
@xcite for the cp(2 ) is related to @xmath36 of the i m with suitable time rescaling .
this mapping guarantees that the i m should belong to the di class which is confirmed by fig .
[ fig : imc ] .
we found the critical point of the i m to be @xmath37 and correspondingly the more accurate critical point of the cp(2 ) as @xmath38 than that in ref .
@xcite .
( color online ) semilogarithmic plot of @xmath20 vs @xmath17 near criticality for the i m .
the critical point is found to be @xmath37.,scaledwidth=45.0% ]
to study the crossover from the di class to the dp class , we introduce dynamics with rate @xmath39 which breaks either the symmetry or the mod-2 conservation .
the crossover scaling ansatz of the order parameter at criticality is @xcite @xmath40 where @xmath41 is the deviation of a tuning parameter from the di critical point .
the meaning of @xmath41 and @xmath39 will be made clearer in the context of each crossover model .
when @xmath42 , the scaling function in eq .
should describe the di critical behavior .
hence , @xmath43 and @xmath44 are the critical exponents of the di classes ; @xmath45 and @xmath46 , respectively @xcite .
the crossover exponent is defined as @xmath47 which can be measured from observing how the phase boundary behaves near the di critical point @xmath48 where @xmath49 is the distance from the di critical point at @xmath42 to the dp critical point at nonzero @xmath39 .
first , consider the conservation - breaking ( cb ) route by introducing the mod-2 conservation breaking dynamics to the baw2 .
when a particle tries to branch offspring , the number of offspring is 2 ( 1 ) with the probability @xmath50 ( @xmath39 ) .
this crossover model with finite @xmath39 will be referred to as the bawcb .
second , the symmetry - breaking ( sb ) route is constructed by introducing the symmetry breaking dynamics to the i m in such a way that the adsorption rate eq . at even - indexed sites is different from that of odd - indexed sites . to be specific , we modify adsorption dynamics without affecting the desorption events as @xmath51 where the subscripts o and e refer to an odd- and even - indexed sites , respectively . when @xmath52 , this system prefers the absorbing state of poisoning monomers at even - indexed sites , which breaks the @xmath3 symmetry of the i m .
we will refer to the model with dynamics eqs . and as the imsb .
.[table : pb ] critical point values @xmath53 of the imsb and imcc and @xmath54 of the bawcb for various @xmath39 s .
the numbers in the parentheses indicate the uncertainty of the last digits . [
cols=">,<,<,<",options="header " , ] finally , the channel - connecting ( cc ) route can be constructed by connecting two equivalent symmetric absorbing states through infinitely many absorbing configurations @xcite . in the context of the i m model
, this route can be studied by allowing an adsorption without a pair reaction probabilistically : @xmath55 the model with dynamics of eqs . , , and will be called as the imcc . as an adsorption attempt at a vacant site between two monomers ( @xmath56 )
is still rejected , any configuration without a @xmath32 pair is absorbing and the order parameter is still the @xmath32 pair density .
in contrast to the i m , the imcc has infinitely many absorbing states characterized by the @xmath31 pair density ( auxiliary field density ) . by pasting
the two antiferromagnetically ordered absorbing states , we find either @xmath57 which is already absorbing , or @xmath58 which can evolve back to an absorbing state by adsorbing a monomer at either site of the @xmath32 pair . hence there is no infinite dynamic barrier between absorbing states and the system leaves the di class into the dp class .
now we are equipped with three different models which can show the crossover behavior from the di class to the dp class .
as explained above , the crossover exponent @xmath59 can be deduced from the dp phase boundary near the di critical point ; see eq . .
for each model , the critical points at finite @xmath39 are located numerically using the dp values of the critical exponents ( see table [ table : pb ] ) .
the critical points for the imsb and imcc are denoted by @xmath53 and for the bawcb by @xmath60 , respectively . to measure the crossover exponent
, we define @xmath61 for the imsb and the imcc , and similarly @xmath62 for the bawcb . in fig .
[ fig : pb ] , the crossover exponents are estimated for all three different routes . along the sb route , we estimate @xmath63 which is consistent with previous estimates : @xmath63 @xcite , @xmath64 @xcite , and recently @xmath65 @xcite . along the cb route ,
our estimate is @xmath66 which is close to the recent result @xmath67 @xcite . along the cc route , we find @xmath68 , which is clearly different from those for the sb and cb routes .
( color online ) log - log plot of @xmath49 vs @xmath39 for three different crossover models . for graphical clarity
, we shifted @xmath69 axis with arbitrary scale .
each line whose slope is @xmath70 shows the fitting result for the corresponding phase boundary.,scaledwidth=45.0% ]
the crossover behavior from one fixed point to another does not reflect the properties of both fixed points , in general .
rather , strictly speaking , it is related to one fixed point and its crossover operator which forces the system to crossover to the other fixed point @xcite . in many cases ,
the crossover route is unique between two fixed points , so is the crossover operator , which leads to one unique crossover exponent between two universality classes .
however , in the case studied here , we found that there are three different routes for the crossover from the di to the dp class , which yield all three different crossover exponents .
it indicates that the crossover operator associated with each route should be different from each other .
the difference between the sb and cb routes has been noticed earlier @xcite . summarizing it in the domain - wall ( particle ) representation ,
the crossover operator in the cb route is a single particle annihilation or creation operator in the action , while that in the sb route is a nonlocal string operator ( global product of particle number operators ) @xcite .
hence it is not surprising to find that the two crossover exponents are different .
( color online ) long - time behavior of the auxiliary field density @xmath71 for small @xmath39 s .
the exponent @xmath72 is the temporal decay exponent of the order parameter in the dp class .
simple extrapolation lines are given by dotted lines . , scaledwidth=45.0% ] the crossover behavior along the cc route is tricky .
first , note that the @xmath3 symmetry between the odd - indexed and even - indexed sites is not broken in this route , which implies that the cc route is different from the sb one .
second , in terms of domain walls , the cc process , eq . , clearly breaks the mod-2 conservation in the number of @xmath32 pairs ( primary field , say , @xmath33 ) , because of which one may naively guess the cb - type crossover along the cc route .
however , unlike the ordinary cb models , the cc process generates an infinite number of absorbing configurations with @xmath31 pairs ( auxiliary field , say , @xmath73 ) which give the feedback effect to the domain wall dynamics .
consider the dynamics of the imcc in terms of two fields @xmath33 and @xmath73 .
the reactions , eqs . and , can be rewritten as @xmath74 and @xmath75 , while eq .
is equivalent to @xmath76 .
the presence of @xmath31 pairs makes it possible to allow another reaction of @xmath77 ( @xmath78 ) in combination with eq . .
the last two reactions involving the auxiliary field represent the feedback which generates memory effects on the primary field .
the emergence of the auxiliary field density and the feedback mechanism may be responsible for the new crossover behavior along the cc route . to see how the auxiliary field may affect the crossover
, we measure the auxiliary field density ( natural density ) @xmath79 along the dp critical line .
we expect its long - time temporal behavior as @xmath80 starting from the @xmath7 particle vacuum @xcite , where @xmath81 is the temporal decay exponent of the order parameter in the dp class .
figure [ fig : d1 ] confirms our expectation and the asymptotic values of @xmath71 is plotted against @xmath39 in fig .
[ fig : d2 ] .
it is clear that @xmath82 at the di critical point . near @xmath42
, the power - law singularity is found as @xmath83 with @xmath84 .
we test the universality of the cc crossover scaling as well as the singularity of the auxiliary field density by investigating various different models like a cc variant of the interacting monomer - dimer model @xcite and a two - species particle model defined as below .
it turns out that these models form one crossover universality class with the same cc - type crossover exponent @xmath59 and the auxiliary field density exponent @xmath85 @xcite .
( color online ) log - log plot of @xmath71 in the steady state versus @xmath39 .
the slope of the dotted line is 0.645 .
, scaledwidth=45.0% ] to examine the exclusive role of the feedback ( memory ) effect , we introduce a two - species ( @xmath33 , @xmath73 ) particle model where the feedback process can be directly controlled . the primary particles @xmath33 can diffuse , while the secondary particles @xmath73 are immobile .
each species particles are hard core particles , but a simultaneous occupation by different species at a site is allowed .
therefore each site can be in one of four possible states such as @xmath8 , @xmath33 , @xmath73 , and @xmath86 .
the reactions dynamics is symbolically summarized as @xmath87 , @xmath75 , @xmath76 , and @xmath88 , @xmath89 , or @xmath90 . to be more specific ,
the evolution rule is given as follows : first , choose one of @xmath33 particles randomly .
with probability @xmath39 , the chosen @xmath33 spontaneously mutates or annihilates as @xmath76 or @xmath88 . with probability @xmath91
, the @xmath33 hops to one of its neighboring sites .
whenever two @xmath33 particles attempt to occupy the same site , they annihilate immediately like in the baw2 . with probability @xmath92
, the @xmath33 can branch particles in the neighborhood as @xmath93 with the dynamic branching rule @xcite . to control the feedback effect
, we introduce a parameter @xmath94 in the branching process when a @xmath73 resides along with the @xmath33 at the same site , such that @xmath95 with the relative rate @xmath96 and @xmath97 with @xmath94 . at @xmath42
, there is no spontaneous annihilation ( or mutation ) process and the model becomes exactly the same as the baw2 after some transient period needed for removing all @xmath73 particles via branching processes . at finite @xmath39 , we expect the dp scaling with a finite density of @xmath73 particles ; @xmath98 because @xmath33 particles generate @xmath73 particles via mutation processes , which can not be fully eliminated via branching processes .
so the cc - type crossover is expected in general .
however the @xmath99 point is special .
there , the presence of @xmath73 particles never affects the dynamics of @xmath33 particles , so there is no feedback mechanism to alter the primary particle density through the secondary particles .
the primary field dynamics is basically identical to the bawcb model introduced in sec .
[ section : three ] and the cb - type crossover scaling should appear at @xmath99 along with nonzero @xmath100 . at finite @xmath94 ,
the @xmath33 particle dynamics is affected by the presence of @xmath73 particles , which generates the memory effects on the @xmath33 particle density .
hence @xmath94 can be regarded as the feedback controlling parameter and @xmath39 as the crossover parameter .
( color online ) @xmath100 vs @xmath39 along the phase boundary for the two - species model with ( @xmath101 and @xmath102 ) or without ( @xmath99 ) feedback processes as well as the variant of the contact process ( see text ) in double logarithmic scales .
the slopes of the straight lines are 0.70 , 0.71 , 0.80 , and 1 from above . , scaledwidth=48.0% ] we performed numerical simulations for various values of @xmath94 .
for nonzero @xmath94 , we found that the secondary particle density behaves as @xmath103 with @xmath104 and the crossover exponent @xmath105 , both of which are consistent with the cc crossover results for the imcc model within numerical errors .
in contrast , at @xmath99 the cb crossover is found as expected along with @xmath106 different from the cc crossover case ; see fig .
[ fig : rhoy ] .
this leads to the conclusion that the feedback process ( memory ) is crucial in establishing the cc - type crossover universality class and affects the singularity characteristic of the auxiliary field density .
it is interesting to note that the numerical values of @xmath85 suggest a simple conjecture that @xmath107 for the di to the dp crossover with an emerging auxiliary field .
the mean field analysis leads to @xmath108 and @xmath109 ( see appendix a ) which is also consistent with this conjecture . the nontrivial singularity in @xmath110 is related to the nontrivial crossover from the di to the dp .
then , it is natural to ask what will happen if there is no nontrivial crossover . to answer this question , we study a two - species version of the contact process ( cp ) with the secondary particles ; @xmath111 , @xmath112 , @xmath76 , and @xmath88 , @xmath33 , or @xmath89 with the crossover parameter @xmath39 controlling the @xmath76 process and the feedback parameter @xmath94 controlling the @xmath113 .
this model belongs to the dp class for any value of @xmath39 and @xmath94 , including @xmath42 .
therefore there is no true " crossover in this model .
we found the trivial linear phase boundary ( @xmath114 ) as expected and also the trivial behavior of @xmath110 with @xmath115 regardless of the presence of the feedback processes ( see fig . [
fig : rhoy ] ) .
these results are consistent with the mean field results described in the appendix a. up to now , we have been only interested in the emergence of the auxiliary field for a finite crossover parameter @xmath39 which breaks the mod(2 ) conservation for the primary particles .
one may consider more general cases where the auxiliary field density @xmath110 is finite even at the di critical point ( @xmath42 ) .
first , we studied the crossover behavior to the dp models with a continuous variation of @xmath110 as @xmath39 increases .
we found that @xmath116 and the crossover scaling belongs to the cb class @xcite .
this implies that the vanishing @xmath110 ( not just singularity near @xmath42 ) is another important ingredient in the cc crossover .
note that @xmath100 seems to behave in the same way as the phase boundary @xmath117 with @xmath118 for the cb crossover .
second , there may be a discontinuous drop in @xmath110 in the crossover to the dp with a single absorbing state ( @xmath119 ) .
this case may be compared to the crossover from the dp with finite @xmath110 ( infinitely many absorbing states ) to the dp with @xmath119 and similarly from the di with finite @xmath110 to the di with @xmath119 @xcite . as understood in ref .
@xcite , this crossover may be characterized by a discontinuous jump in the phase boundary at the di critical point for the excitatory route or a continuous phase boundary for the inhibitory route . along the inhibitory route
, the cb crossover is observed again ( not shown here ) and we conclude that the discontinuity in the auxiliary field density does not provide any new crossover scaling @xcite . to summarize , we studied three routes of the crossover from the di to the dp class ; symmetry breaking ( sb ) , mod-2 conservation breaking ( cb ) , and channel connecting ( cc ) routes .
these three routes are characterized by three different crossover exponents .
the difference between the sb route and the cb route is clear because the symmetry breaking field can be interpreted as the spatial non - local operator affecting the domain wall dynamics , while the conservation breaking corresponds to the local operator .
the cb route and the cc route share some common features , but the feedback memory effect ( or temporal non - locality ) makes the cc crossover distinct from the cb crossover .
we also found the universal exponent @xmath120 which describes the singularity of the vanishing auxiliary field in the cc route . from the numerical study of the feedback controlling model and its mean field theory , we conjectured that @xmath107 .
scp would like to thank for the support of the korea institute for advanced study ( kias ) where this work was initiated and the support by dfg within sfb 680 _ molecular basis of evolutionary innovations_. most of computation was carried out using kias supercomputers .
this appendix provides the mean field theory for the feedback controlling model introduced in sec . [
section : discussion ] . here
@xmath121 and @xmath110 are the densities of @xmath33 and @xmath73 particles , respectively and @xmath122 is the density of sites where both @xmath33 and @xmath73 particles reside .
then one may easily write down the mean field equations as @xmath123 , \label{eq : z}\end{aligned}\ ] ] where @xmath124 .
since @xmath125 in the mean field theory , @xmath35 is the same as @xmath41 in the text . in the active phase
, the steady state density can be calculated by setting @xmath126 and so on .
the superscript @xmath127 in the following indicates the steady state density . from eq . , we get in the active phase @xmath128 hence eq . gives the steady state density of @xmath129 such that @xmath130 in the active phase , the ( stable ) steady state density @xmath129 is determined uniquely , so we define the natural density at criticality by taking the limiting process from the active side as @xmath131 . from eqs . and
, the critical point is determined by the equation @xmath132 which reads @xmath133 for any value of @xmath134 , @xmath135 which gives @xmath136 . since @xmath137 at criticality , we get @xmath138 from eq . , yielding @xmath139 . in low dimensions
, @xmath140 is expected to renormalize such that @xmath141 , which leads the conjecture @xmath142 . for comparison
, we study the two - species version of the contact process with the secondary particles introduced in sec .
[ section : discussion ] .
the mean field equations are @xmath143 where @xmath144 .
then one may easily show that the critical line and the critical steady - state density of @xmath73 particles are given as for small @xmath39 @xmath145 where @xmath146 . as expected , we find the trivial phase boundary and also the trivial value of @xmath115 , which is consistent with the simulation results shown in fig . [
fig : rhoy ] .
moreover , the renormalization can not generate a singular behavior in the denominator of eq .
, so @xmath85 is always expected to be 1 in all dimensions .
h. hinrichsen , adv .
* 49 * , 815 ( 2000 ) ; g. dor , rev . mod .
phys . * 76 * , 663 ( 2004 ) .
h. k. janssen , z. phys .
b : condens . matter * 42 * , 151 ( 1981 ) .
p. grassberger , z. phys .
b : condens .
matter * 47 * , 365 ( 1982 ) .
g. grinstein , z .- w .
lai , d.a .
browne , phys .
a * 40 * , 4820 ( 1989 ) .
i. jensen , phys .
lett . * 70 * , 1465 ( 1993 ) .
w. hwang and h. park , phys .
e * 59 * , 4683 ( 1999 ) .
m. c. marques and j. f. f. mendes , eur .
j. b * 12 * , 123 ( 1999 ) .
h. s. park and h. park , j. korean phys .
soc . * 38 * , 494 ( 2001 ) . for a review ,
see m. henkel and h. hinrichsen , j. phys .
a * 37 * , r117 ( 2004 ) .
j. kockelkoren and h. chat , phys .
lett . * 90 * , 125701 ( 2003 ) . j. d. noh and h. park , phys .
e * 69 * , 016122 ( 2004 ) . s .-
park and h. park , phys .
lett . * 94*,065701 ( 2005 ) ; phys .
e * 71 * , 016137 ( 2005 ) .
h. hinrichsen , physica a * 361 * , 457 ( 2006 ) .
park and h. park , phys .
e * 73 * , 025105(r ) ( 2006 ) .
p. grassberger , f. krause , and t. von der twer , j. phys .
a * 17 * l105 ( 1984 ) ; p. grassberger , _ ibid . _ * 22 * , l1103 ( 1989 ) .
h. takayasu and a. yu .
tretyakov , phys . rev
. lett . * 68 * , 3060 ( 1992 ) . m. h. kim and h. park , phys
lett . * 73 * , 2579 ( 1994 ) ; h. park , m. h. kim , and h. park , phys .
e * 52 * , 5665 ( 1995 ) .
n. menyhrd , j. phys .
a * 27 * , 6139 ( 1994 ) ; n. menyhrd and g. dor , _ ibid . _
* 29 * , 7739 ( 1996 ) .
i. jensen , phys .
e * 50 * , 3623 ( 1994 ) .
w. hwang , s. kwon , h. park , and h. park , phys .
e * 57 * , 6438 ( 1998 ) .
j. d. noh , h. park , and m. den nijs , phys .
e * 59 * , 194 ( 1999 ) .
o. a. hammal , h. chat , i. dornic , m. a. muoz , phys .
lett . * 94 * , 230601 ( 2005 ) . s .-
park and h. park , europhys .
j. b * xx * xxxx ( 2008 ) ( doi:10.1140/epjb / e2008 - 00022 - 4 ) . k. park , h. hinrichsen , and i .- m . kim , phys .
e * 63 * , 065103 ( 2001 ) . h. park and h. park ,
physica a * 221 * , 97 ( 1995 ) .
k.e . bassler and d.a .
browne , phys .
lett . * 77 * , 4094 ( 1996 ) ; phys .
e * 55 * , 5225 ( 1997 ) . h. hinrichsen , phys .
e * 55 * , 219 ( 1997 ) .
s. kwon , w. hwang , and h. park , phys .
e * 59 * , 4949 ( 1999 ) .
s. kwon and h. park , phys .
e * 52 * , 5955 ( 1995 ) .
j. l. cardy and u. c. tuber , phys .
lett . * 77 * , 4780 ( 1996 ) ; j. stat . phys . *
90 * , 1 ( 1998 ) .
n. inui and a. yu .
tretyakov , phys .
lett . * 80 * , 5148 ( 1998 ) .
see , e.g. , i.d .
lawrie and s. sarbach , in _ phase transitions and critical phenomena _
, edited by c. domb and j.l .
lebowitz ( academic press , london , 1984 ) , vol .
g. dor and n. menyhrd , arxiv:0804.3092 .
park and h. park , phys .
e * 76 * , 051123 ( 2007 ) .
g. dor , j. f. f. mendes , m. a. santos , and m. c. marques , phys .
e * 58 * , 7020 ( 1998 ) . s .-
park and h. park ( unpublished ) . | scaling nature of absorbing critical phenomena is well understood for the directed percolation ( dp ) and the directed ising ( di ) systems .
however , a full analysis of the crossover behavior is still lacking , which is of our interest in this study . there are three different routes from the di to the dp classes by introducing a symmetry breaking field ( sb ) , breaking a modulo 2 conservation ( cb ) , or making channels connecting two equivalent absorbing states ( cc )
. each route can be characterized by a crossover exponent , which is found numerically as @xmath0 ( sb ) , @xmath1 ( cb ) , and @xmath2 ( cc ) , respectively .
the difference between the sb and cb crossover can be understood easily in the domain wall language , while the cc crossover involves an additional critical singularity in the auxiliary field density with the memory effect to identify itself independent . |
the investigations of neutrino physics in an active medium are the subject of great interest today .
the one of the topical problem in current researches is the influence of an external magnetic field and plasma on the neutrino dispersion relation .
it this paper we analyse the neutrino dispersion properties in an active medium consisting of magnetic field and plasma .
the investigations of such type are based on calculation of neutrino self - energy operator @xmath0 .
this operator can be defined in term of the invariant amplitude of the neutrino transition @xmath1 , by the relation : m ( ) = - |(p ) ( p)(p ) , [ eq : def ] where @xmath2 is the neutrino four - momentum . using the expression ( [ eq : def ] ) , one obtains for the additional neutrino energy in magnetized plasma . ] : e = sp \ { ( ( p ) + m_)(1 - ( s)_5)(p ) } , [ eq:2 ] where @xmath3 is the neutrino energy in vacuum , @xmath4 is the neutrino mass , @xmath5 is the neutrino spin four - vector , @xmath6 are the dirac matrices in the standard presentation .
the lorentz indexes of four - vectors and tensors within parenthesis are contracted consecutively , for example , @xmath7 .
there are several parametrization for general structure of the operator @xmath8 . in the presence of homogeneous magnetized medium the operator
@xmath8 contains three independent structures . in this case
it is convenient to express @xmath8 as ( p ) = \ { a ( p ) + b ( u ) + c ( p ) } l. [ eq:3 ] here @xmath9 are the numerical coefficients , @xmath10 is the four - vector of medium velocity , @xmath11 is the dimensionless dual tensor of the magnetic field , @xmath12 is the absolute value of the magnetic field strength , @xmath13 is the left - handed projection operator . on this parametrization
the coefficients @xmath9 have a simple physical interpretation . really , performing calculation of the trace in eq.([eq:3 ] ) with @xmath8 in the form ( [ eq:2 ] )
, one can obtained for the neutrino self - energy e = b - c ( p s ) .
[ eq:4 ] where @xmath14 is the neutrino velocity , @xmath15 is the twice vector of average spin of neutrino .
it is seen that the additional energy in magnetized plasma for massless left - handed neutrino in fact depends on the parameter @xmath16 only , e = b. in the case of massive neutrino the second term in ( [ eq:4 ] ) corresponds to the additional energy caused by a neutrino magnetic moment @xmath17 , e _
_ = - ( p s ) , _ = .
so , one can speak that parameter @xmath18 determines the additional neutrino magnetic moment in the magnetized plasma .
the modification of the neutrino dispersion relation in a magnetized plasma have a long history , see for example @xcite-@xcite .
in particular , the dispersion relation of neutrino in charge symmetric plasma under physical conditions m_w^2 t^2 , eb m_e^2 , eb t^2 [ eq : cond ] is well known = .
[ eq:5 ] here @xmath19 is the neutrino momentum , @xmath20 is the angle between magnetic field direction and vector @xmath19 , @xmath21 is the plasma temperature .
the first term in ( [ eq:5 ] ) is the pure plasma contribution @xcite , while the second @xcite and third @xcite terms in brackets are caused by the common influence of the plasma and magnetic field .
as one can see the term of the second order of the field contains the infrared divergence in the massless electron limit
. however , there is a good reason to think that this large logarithmic factor @xmath22 can not arise under physical conditions ( [ eq : cond ] ) , when the electron mass is the smallest parameter of the task .
actually , under conditions ( [ eq : cond ] ) the contribution into neutrino energy is determined by electrons and positrons on exited landau levels with energy @xmath23 .
one can see that in this case the electron mass squared can be neglected in comparison to magnetic field . by this means , the availability of the logarithmic factor @xmath22 in the ( [ eq:5 ] ) has come into question , so independent investigations of the neutrino dispersion in magnetized medium it is required .
further we study the additional neutrino energy for massless neutrino in the charge symmetric plasma with the presence of the magnetic field of arbitrary strength . as a special case we consider the weak field , moderate field and strong field limits in more details .
in the case of the charge symmetric plasma the contribution into amplitude of neutrino transition @xmath1 in magnetized plasma derives from processes , shown in fig.1 .
it should be noted that in the local limit the neutrino self - energy operator @xmath8 is zero @xcite .
thus , there is a need to take into account the momentum dependence of w , z - boson propagators .
the process of neutrino transition via z boson does not sensitive to influence of external magnetic field .
thus , the contribution into operator @xmath24 from this process can be extracted from paper @xcite
. the amplitude of neutrino forward scattering on all plasma electrons and positrons can be immediately obtained from the lagrangian of @xmath25 interaction with w - boson l = ( |e _ ( 1 - _ 5 ) ) w_+ ( [ fig ] omitting the details of calculations , the general expression for the neutrino self - energy caused by the neutrino forward scattering on plasma electrons can be written as @xmath26 here @xmath23 and @xmath27 are the electron energy and z - component of electron momentum correspondingly , @xmath28 is the landau level number , @xmath20 s the angle between magnetic field direction and neutrino momentum @xmath19 , @xmath29 is the electron distribution function , @xmath30^{-1}$ ] , the sum is defined as ^_n=0 f(n ) = f(n=0 ) + ^_n=1 f(n ) .
the integral and sum in eq .
( [ eq : total ] ) can be calculated under some physical conditions : * the limit of weak magnetic field , when the magnetic field strength is the smallest physical parameter t^2 m^2_e eb .
[ eq : cond1 ] the result of calculations for the plasma contribution into neutrino self - energy in this limit is @xmath31 .
\nonumber\end{aligned}\ ] ] as one can see the expression ( [ eq : weak ] ) contains the logarithmic factor with @xmath32 , but under conditions considered ( [ eq : cond1 ] ) the electron mass is not the smallest parameter of task . so the electron mass can not be tend to zero at nonzero magnetic field b. * the moderate magnetic field , when the magnetic field being relatively weak in comparison with plasma temperature , is simultaneously strong enough on the scale of electron mass squared t^2 eb m^2_e .
[ eq : cond2 ] under conditions ( [ eq : cond2 ] ) plasma electrons and positrons occupy highest landau levels . in this case for neutrino energy one can obtained @xmath33 . \nonumber\end{aligned}\ ] ] our calculations show , that under conditions ( [ eq : cond2 ] ) the additional neutrino energy does not contain the infrared divergence in the limit @xmath34 in contrast @xcite . * the strong magnetic field limit , when from two components of active medium the field component dominates eb t^2 m^2_e .
[ eq : cond3 ] under conditions ( [ eq : cond2 ] ) the most part of plasma electrons and positrons occupy the ground landau level .
the result for the additional neutrino energy in this limit has the form : @xmath35 .
\nonumber\end{aligned}\ ] ] here the second term corresponds to the contributions of ground landau level , while the third term is caused by the first landau level .
we have studied the neutrino dispersion in the charge symmetric plasma with the presence of a constant magnetic field . the most general expression in simple analytic form for the plasma contribution into the neutrino self - energy
was obtained . in particular , we have considered the physical conditions , corresponding to the weak field and moderate field , when plasma electrons and positrons occupy the excited landau levels .
the strong magnetic field limit , when plasma electrons and positrons mainly occupy the lowest landau level , is investigated also .
it is shown that additional neutrino energy in the limit of moderate field , @xmath36 , does not coincide with previous result in paper @xcite under the same physical conditions and does not contain the infrared divergence in the limit @xmath34 in contrast to @xcite . the work was supported in part by the russian foundation for basic research under the grant no .
07 - 02 - 00285-a , and by the council on grants by the president of the russian federation for the support of young russian scientists and leading scientific schools of russian federation under the grant no . nsh-497.2008.2 . | the neutrino dispersion in the charge symmetric magnetized plasma is investigated .
we have studied the plasma contribution into the additional energy of neutrino and obtained the simple expression for it .
we consider in detail the neutrino self - energy under physical conditions of weak field , moderate field and strong field limits .
it is shown that our result for neutrino dispersion in moderate magnetic field differ substantially from the previous one in the literature . |
in the outer solar system beyond the orbit of neptune is a large belt of minor planets , the kuiper belt , whose dynamics has been understood , over the past two decades , to be significantly controlled by the gravity of the giant planets , jupiter
neptune , either through secular or resonant perturbations or by gravitational scattering . however , there are a few known kuiper belt objects [ kbos ] that are unlikely to be significantly perturbed by the known giant planets in their current orbits . in the dynamical classification scheme for the small bodies of the outer solar system ,
such objects ( semimajor axis , @xmath1 , and perihelion distance , @xmath2 , exceeding 150 au and 40 au , respectively ) would likely belong in a class known as `` detached objects '' , distinct from the classical kbos , the resonant kbos and the scattered / scattering kbos @xcite . @xcite call them `` extreme '' kbos [ ekbos , henceforth ] . upon querying the minor planet center s catalog ,
we identify six kbos having @xmath3 au and @xmath4 au ; two of these are numbered and four are unnumbered : sedna ( 90377 ) , 2010 gb174 , 2004 vn112 , 2012 vp113 , ( 148209 ) , 2013 gp136 .
their orbital inclinations are in the range 11.933.5 degrees ( to the ecliptic ) , not especially remarkable for kbos .
however all of these objects have extremely large orbital eccentricity , @xmath5 , suggesting large orbital perturbation in their past history .
this is rather at odds with their current large separations from the known giant planets .
a natural hypothesis is that these objects orbital dynamics were or are dominated by the gravitational perturbations of a planet in the heliocentric distance range of a few hundred au .
@xcite hypothesized a `` rogue '' planet , a short - lived planet larger than an earth - mass , to explain the ekbos .
@xcite proposed an extant hypothetical distant planet to explain an apparent clustering of the arguments of perihelion of the ekbos .
@xcite have also argued for this hypothesis by calling attention to the clustering of the eccentricity vectors and orbit poles of some distant kbos .
these authors suggest a planet mass of @xmath6 , and an orbit with semimajor axis @xmath7 au , and eccentricity @xmath8 . @xcite
provide an analysis of _ cassini _ tracking data , arguing that , for the favored planet parameters of @xcite , the residuals in the orbital motion of saturn are reduced with the existence of such a planet at current true longitude near @xmath9 ; however their analysis can not exclude a large range of true longitudes , @xmath10 , where the planet would be too distant to sufficiently perturb saturn . here
we report additional support for the extant distant planet hypothesis , and we narrow the range of its parameters and its current location .
we point out hitherto unnoticed peculiarities of the orbits of the ekbos : we find that the orbital period ratios of these objects are close to integer ratios .
such ratios are not dynamically significant unless the ekbos are in mean motion resonances ( mmrs ) with a massive planet .
we identify an orbital period of the hypothetical planet which is in simple integer ratios with the four most distant ekbos .
we show that the resonant dynamics provide constraints on the planet s orbital parameters and its mass and can also narrow the possible range of its current location in its orbital path .
the wide range of the perihelion to - aphelion distances of each ekbo may overlap those of the hypothetical planet [ hp , henceforth ] . in this circumstance
, the extreme eccentricities of the ekbos may be a natural consequence of gravitational scattering encounters with the planet .
the dynamical survival of scattered objects over solar system age could be simply due to the long timescales of close encounters at these distances , or it could be due to enhanced dynamical stability offered by phase - protection in mean motion resonances with hp . in their analysis , @xcite emphasized the secular effects of the planets in aligning the orbits of the ekbos , but also noted that transient resonant interactions would be important .
here we investigate the possibility of extant mmrs of the known ekbos with the unseen planet . examining the orbital periods of the six ekbos as listed in the minor planet center s catalog , and ordering them in order of decreasing orbital period ( so sedna is ekbo#1 , with the longest period orbit ) , here is what we find for the ratios @xmath11 : @xmath12 these values are close to the following rational ratios : 8/5 , 2/1 , 8/3 , 10/3 , and 6/1 , sufficiently intriguing to investigate further .
we proceeded as follows .
we computed the best - fit orbits for these six ekbos , as well as their uncertainties , using the orbit - fitting software developed by @xcite ; for these orbit - fits , we included all of the astrometric measurements for the objects listed in the minor planet center .
we find the following best - fit semimajor axes and their 1@xmath13 uncertainties ( all quoted in au ) : @xmath14 these osculating barycentric elements agree well with the heliocentric mpc orbits when both are converted to heliocentric cartesian positions and velocities .
we then consider some simple possibilities for hp s resonant orbit .
assume that its orbital period @xmath15 is longer than @xmath16 , but not by too much , else its gravitational effects would not be very strong .
so , we consider @xmath17 values of 2/1 or 3/2 or 4/3 .
we find that the case of @xmath18 ( hence @xmath19 au ) is particularly striking , for it yields period ratios close to 5/2 , 3/1 , and 4/1 for @xmath20 , @xmath21 and @xmath22 .
the other simple choices ( @xmath17 values of 2/1 or 4/3 ) yield larger integer ratios for @xmath23 . although these larger integer ratios can not be ruled out _ prima facie _
, we limit this first study to the small integer ratio mmrs ; other mmrs will be considered in a future study .
the orbit period @xmath15 of hp is assumed to be 3/2 times sedna s best - fit orbital period , @xmath24 11,411 years .
the dotted horizontal lines indicate the rational ratios 3/2 , 5/2 , 3/1 4/1 , 5/1 and 9/1 . ] in fig .
[ f : pratios ] , we plot the period ratios @xmath25 for the best - fit orbits with their uncertainties .
we observe that , within 3@xmath13 uncertainties , four of the six ekbos have periods near n/1 or n/2 ratios with hp , when we take the latter s period to be 1.5 times sedna s best - fit period .
we also observe that the error bars are not insubstantial and one can be skeptical of any perceived coincidence with simple ratios .
we remark that our procedure for computing the uncertainties weights all published observations equally and may overestimate the true errors .
secondly , one might also be skeptical of the period ratio coincidences because not all possible n/1 and n/2 period ratios in the range shown are occupied . on this point
, we can remark that ( a ) the larger n values lead to shorter period orbits in which hp s influence is diminished and neptune s influence looms larger , potentially destabilizing the resonances with hp , and ( b ) the small number of currently known ekbos will occupy only a small fraction of all possible n/1 , n/2 period ratios .
thirdly , one can ask if the proximity to the resonant ratios is different from random .
we answer this question approximately , as follows . for integer values @xmath26
, there are 31 distinct rational numbers @xmath27 in the range @xmath28 , so the mean distance between said rationals is @xmath29 .
we can compare this mean value to the distances of the observed period ratios to the resonant ratios .
taking account of the 3@xmath13 observational uncertainties , the latter are 0.007 , 0.27 , 0.27 , 0.26 , 0.21 , and 0.46 , respectively ( listed in increasing period ratio , as in fig . [ f : pratios ] ) .
the shortest period ekbo , 2013 gp136 , presents the largest distance ( from the 9/1 ratio ) , as is evident in fig .
[ f : pratios ] .
sedna s very small distance to the 3/2 ratio is explained by its very small orbital uncertainties and because its period ratio of 3/2 is fixed by hypothesis .
the remaining four values are similar to the mean distance between the rationals , indicating that they are not strongly distinct from random . however , this statistical argument depends upon how many rationals are allowed and it does not consider any physics of mmrs
. we will show later ( section 2.2 ) that the libration widths of the specific mmrs are similar to these apparent distances from exact resonant ratios ; hence , libration in mmrs is possible even with the significant apparent distances from exact resonant ratios .
we therefore proceed with investigating the implications of these mmrs for the four longest period ekbos .
1truein -2.1truein stable resonant orbits of the ekbos will avoid close approaches with hp , and will have a discrete number of geometries for the conjunctions of the ekbo with the planet .
these are described by librations of the critical resonant angle , @xmath30 as usual , in the definition of the resonant angle , @xmath31 and @xmath2 are integers , @xmath32 the mean longitudes and @xmath33 the longitudes of perihelion ; we use the prime to refer to orbital parameters of the planet , the parameters of the ekbo are unprimed .
depending upon the orbital eccentricity , @xmath34 may be stable at either 0 or at @xmath35 or at both .
we illustrate in fig .
[ f : resonantorbits ] the trace of the resonant orbits ( of zero libration amplitude ) in the frame co - rotating with the mean angular velocity of hp .
also included here are the traces of a few possible coplanar orbits of hp , with eccentricity @xmath36 and @xmath37 .
we observe that for sedna , stable libration is likely only near @xmath38 , and unlikely near @xmath39 , because the latter would allow small encounter distances .
( this is the opposite of the geometry for the low eccentricity regime in which stable ( unstable ) librations are centered at @xmath40 . )
consequently , conjunctions with hp occur near one of three longitudes , @xmath41 and @xmath35 from sedna s longitude of perihelion .
moreover , the long term stability of sedna would limit @xmath42 .
likewise , a coplanar orbit of 2010 gb174 with @xmath38 limits @xmath43 ; however , in the geometry with @xmath39 , the limit is @xmath44 . the orbit of 2004 vn112 limits @xmath45 for @xmath38 . of course , strict coplanarity of hp is possible for at most one ekbo , because the ekbos orbits are not coplanar .
[ f : hpplane ] shows the ecliptic latitude and longitude of the orbit normal vectors of each of the four potentially resonant ekbos .
we see that they are somewhat clustered , approximately within one quadrant of ecliptic longitudes ; they have relative inclinations 715 degrees to their mean plane .
the non - coplanarity of the orbits with hp s orbit plane may relax the upper limit on @xmath46 , especially if the ekbo s argument of perihelion @xmath47 ( relative to hp s orbit plane ) were near @xmath48 .
this would have the effect of lifting the ekbo s aphelion away from the hp orbit plane and thereby increasing the closest approach distances and enhancing dynamical stability , and it librates about @xmath49 with a period of about 4 megayears @xcite . ] .
such a configuration is associated with the so - called `` periodic orbits of the third kind '' envisioned by henri poincar ( as noted by @xcite ) and computed by @xcite @xcite and @xcite for the three dimensional circular restricted three body problem .
such periodic orbits are characterized by simultaneous stationarity of the resonant angle @xmath34 ( at 0 or @xmath35 ) and of the argument of perihelion @xmath47 ( at @xmath50 or @xmath35 or @xmath51 ) .
physically , we can visualize these periodic orbits as having a fixed orbit plane and a fixed eccentricity vector ( i.e. , the apsidal precession rates of both node and pericenter vanish ) . in the particular case of @xmath52 , the particle reaches aphelion when it is at its largest excursion away from the plane of the planet .
we computed the values of @xmath47 relative to the mean plane of the four longest period ekbos , finding that they are @xmath53 -1.2truein within 45 degrees of @xmath51 .
the region shaded in magenta identifies the subset for which the relative inclinations of the ekbos are also within 10 degrees of the stationary inclination values at the periodic orbit of the third kind.,title="fig : " ] -0.2truein we look for an alternative hp orbit plane for which @xmath47 is proximate to @xmath48 , within some specified tolerance , for all of the four ekbos .
we proceed as follows .
let @xmath54 be the ( unknown ) unit vector normal to the hp orbit plane .
let @xmath55 be the unit vector in the orbit plane of an ekbo and normal to its eccentricity vector .
we ask whether @xmath55 can be rotated by no more than a specified amount , @xmath56 , in the ekbo s orbit plane , such that @xmath57 changes sign ; if so , this would indicate that the ekbo has @xmath47 within @xmath56 of @xmath51 , near the periodic orbit of the 3rd kind . by demanding that all four ekbos satisfy this constraint
, we identify the range of possible @xmath54 .
the results of this search are shown in fig .
[ f : hpplane ] , in which we plot in grey the ecliptic longitude and latitude of the allowed @xmath54 s for @xmath58 .
we narrow the possible hp planes further by examining the inclinations of the ekbos to possible hp orbit planes .
consulting the results of @xcite , we find that the stationary inclinations ( relative to hp s orbit plane ) are @xmath59 , and @xmath60 , respectively , for the specific mmrs and the observed eccentricities of the four ekbos . of course , these ekbos are likely librating with non - zero amplitude about the pertinent periodic orbit , so we look for those @xmath54 s that yield inclinations within @xmath61 of these values . carrying out this search
, we find the range indicated in magenta in fig .
[ f : hpplane ] satisfies this constraint .
we conclude that an hp orbit plane with inclination near @xmath62 and longitude of ascending node near @xmath63 allows all four resonant ekbos to be in proximity of the periodic orbits of the third kind .
resonant perturbations from hp provide phase - protection if an ekbo is sufficiently close to the exact resonant orbit .
how close is close enough depends upon the planet mass .
we now quantify this condition . in the usual canonical variables , @xmath64 the orbital perturbation equations are given by @xmath65 where the perturbation hamiltonian is @xmath66 here the first equation gives the newtonian perturbation potential ( @xmath67 , @xmath68 are the position vectors of an ekbo and of hp , respectively , in jacobi coordinates ( e.g. , * ? ? ?
* ) ; i.e. , they are relative to the barycenter of the sun and the known planets ) , whereas the second equation gives an approximation for the resonant perturbation
. the dependence of the resonant perturbation on the canonical variables can be expressed as a product of an amplitude @xmath69 which is a function of only the canonical momenta , and a phase @xmath70 which is an even function of the resonant angle .
( @xmath71 also depends on the orbital parameters of hp , which will be assumed to be constant . ) without loss of generality , we can define @xmath71 and @xmath72 such that @xmath73 for @xmath74 . for the ekbos whose extreme eccentricities are near unity
, the magnitude of @xmath71 will be of order a few .
when we require numerical values ( below ) , we will adopt @xmath75 ; but this value can , in principle , be determined more accurately with numerical averaging of the perturbation potential . the mean motion of the ekbo is @xmath76 .
therefore , the resonant perturbations of the mean motion are given by @xmath77 , whence we obtain the following equation of motion for @xmath34 , @xmath78 where @xmath79 .
this equation for @xmath34 resembles the nonlinear pendulum .
the small amplitude oscillation frequency is given by @xmath80 for the resonances of interest , the corresponding libration periods are on the order of a few megayears for a few - earth - mass planet .
by analogy with the pendulum , we divine that @xmath34 ( hence the conjunction longitudes @xmath81 ) can execute librations about stable equilibrium points provided @xmath82 . thus the required condition for resonance is that the deviation of the mean motion of the ekbo from the exact resonant value , @xmath83 , should not exceed @xmath84 .
then , by kepler s third law , the resonance width in semimajor axis is given by @xmath85 computing the widths of the 3/2 , 5/2 , 3/1 and 4/1 resonances for the four ekbos of interest , we find that a @xmath86 planet yields @xmath87 values comparable to the uncertainties in the semimajor axes of these objects . the more severe condition on the planet mass obtains when we demand that the ekbos are in libration about periodic orbits of the third kind .
the frequency of small amplitude oscillations about these periodic orbits or @xmath88 mmrs , are often inaccurately referred to as `` kozai within mmr '' .
they can be considered to be the discrete sequence whose limit for @xmath89 is the kozai - lidov resonance @xcite .
[ e : w0klc ] is obtained in this limit . ]
is given by @xmath90 where @xmath91 are a resonant ekbo s mean motion , semimajor axis , eccentricity and inclination ( relative to hp s orbit plane ) , @xmath92 is the planet s semi - minor axis , and @xmath93 is a numerical coefficient whose asymptotic value is @xmath94 for @xmath95 ( @xmath93 will be larger for @xmath1 approaching @xmath96 ) . for numerical estimates ,
we adopt @xmath97 and @xmath98 .
we find that , for @xmath99 , the small amplitude oscillation periods for the four ekbos are 0.60.94 gyr , a factor of a few smaller than the age of the solar system .
we conclude that @xmath86 is approximately the minimum mass necessary to maintain periodic orbits of the third kind for the four ekbos .
each resonant ekbo constrains the longitude of hp because the critical resonant angle must librate about either 0 or @xmath100 . for illustration , consider the libration of @xmath101 about @xmath100 for sedna s 3/2 resonant orbit .
given that sedna is presently near its perihelion , @xmath102 , the planet can be in proximity to one of three longitudes : @xmath103 or @xmath100 or @xmath104 from sedna s longitude of perihelion , @xmath33 . the trace of the resonant orbit in the rotating frame ( fig .
[ f : resonantorbits ] ) shows that libration of @xmath34 would allow the planet to be located as far as @xmath105 from these longitudes .
this means that only three small zones of @xmath106 around each of three locations , @xmath33 , @xmath107 and @xmath108 are excluded for the planet s present location . for @xmath97
, we find that the excluded zone is a little larger , @xmath109 around each of the same three locations .
, and jacobi constant values as indicated .
these are phase space portraits in the vicinity of the 3/2 , 5/2 , 3/1 and 4/1 resonant orbits of eccentricity 0.85 , 0.87 , 0.85 and 0.7 , respectively .
the abscissa , @xmath110 , measures the angle between the planet and the particle s pericenter ; the ordinate is the particle s barycentric semimajor axis , in units of the circular orbit radius of the planet . ] , and jacobi constant values as indicated .
these are phase space portraits in the vicinity of the 3/2 , 5/2 , 3/1 and 4/1 resonant orbits of eccentricity 0.85 , 0.87 , 0.85 and 0.7 , respectively .
the abscissa , @xmath110 , measures the angle between the planet and the particle s pericenter ; the ordinate is the particle s barycentric semimajor axis , in units of the circular orbit radius of the planet . ] , and jacobi constant values as indicated .
these are phase space portraits in the vicinity of the 3/2 , 5/2 , 3/1 and 4/1 resonant orbits of eccentricity 0.85 , 0.87 , 0.85 and 0.7 , respectively .
the abscissa , @xmath110 , measures the angle between the planet and the particle s pericenter ; the ordinate is the particle s barycentric semimajor axis , in units of the circular orbit radius of the planet . ] , and jacobi constant values as indicated .
these are phase space portraits in the vicinity of the 3/2 , 5/2 , 3/1 and 4/1 resonant orbits of eccentricity 0.85 , 0.87 , 0.85 and 0.7 , respectively .
the abscissa , @xmath110 , measures the angle between the planet and the particle s pericenter ; the ordinate is the particle s barycentric semimajor axis , in units of the circular orbit radius of the planet . ]
however , such geometric estimates of exclusion zones do not account for the planet s gravity . to take account of this , we computed surfaces - of - section for the circular planar restricted three body problem , for several values of the planet mass in the range of @xmath111 to @xmath112 , for the neighborhood of the 3/2 , 5/2 , 3/1 and 4/1 mmrs and the extreme eccentricities of the four ekbos ( 0.85 , 0.87 , 0.85 and 0.7 , respectively ) .
[ f : sos ] shows a sample of these for planet mass @xmath113 ( approximately @xmath86 ) . from these surfaces - of - section , we determine the range of excluded longitudes of the planet by measuring the complement of the large libration zones of the 3/2 and 5/2 mmrs .
we also conservatively exclude a small range , @xmath114 , around the homoclinic points of the 3/1 and 4/1 resonant ekbos .
[ f : plotex ] shows our estimates of the excluded zones , projected on the sky , for a planet of mass @xmath113 ; we plot a representative case of the planet s low inclination orbit ( in green ) and also a representative case of its high inclination orbit ( in red ) .
( note that a range of orbital planes near each of these representative cases is allowed by our calculations . )
we find that approximately half of the ecliptic longitudes are excluded in both cases ; the excluded zone is not contiguous .
we noticed that the four known longest period kbos ( sedna ( 90377 ) , 2010 gb174 , 2004 vn112 , 2012 vp113 ) could all be in mean motion resonances with an unseen distant planet .
the set of resonances is not uniquely determined . in the present work we investigated the case of a hypothetical planet of orbit period @xmath115
17,117 years ( semimajor axis @xmath116 au ) , which would have dynamically significant n/1 and n/2 period ratios with these four ekbos .
we give a broad sketch of how resonant geometries and resonant dynamics of these high eccentricity ekbos help narrow the range of planet parameters ; many details in this analysis would benefit from deeper investigation in future work and potentially improve planet parameters further .
our calculations suggest two possibilities for the planet s orbit plane : a plane moderately close to the ecliptic and near the mean plane of the four ekbos ( @xmath117 , @xmath118 ) , or an inclined plane near @xmath119 , @xmath120 .
the former offers dynamical stability of the four ekbos by means of libration of the critical resonant angles , and the latter offers enhanced dynamical stability by means of additional libration of the argument of perihelion associated with _ periodic orbits of the third kind _ ( a class of periodic orbits of the three dimensional restricted three body problem ) . in the case of coplanar mmrs , the planet s orbital eccentricity is limited to @xmath121 by the putative 3/2 resonant orbit of sedna , and to @xmath44 by the putative 5/2 resonant orbit of 2010 gb174 . for non - coplanar orbits , including the periodic orbits of the third kind , the planet s eccentricity may be larger than these limits .
preliminary numerical analysis finds that for relative inclinations up to @xmath122 , stable resonant librations are possible but uncommon for @xmath123 .
a planet mass of @xmath124 defines resonance widths that are similar to or exceed the semimajor axis uncertainties of the observed ekbos .
resonant libration periods would be on the order of a few megayears , much shorter than the age of the solar system . in the case of the inclined planet orbit
, a planet of @xmath125 also supports librations about the periodic orbits of the third kind , with libration periods shorter than the age of the solar system .
both these estimates support the conjecture that the period ratio coincidences are of dynamical significance .
we determined exclusion zones of the current location of the planet in its orbital path ( fig .
[ f : plotex ] ) , based on the mmrs of all four ekbos ; these exclude just over half of the orbital path , assuming a @xmath86 planet .
our analysis supports the distant planet hypothesis , but should not be considered definitive proof of its existence .
the orbital period ratios have significant uncertainties , so the near - coincidences with mmrs may simply be by chance for a small number of bodies .
we encourage further observations of these ekbos to reduce the uncertainties in their orbital periods ; a factor of @xmath126 reduction in the uncertainties would provide a clear test of our specific mmr hypothesis for a @xmath124 planet .
future work should also examine other possible sets of mmrs . as a final point , we note that the long orbital timescales in this region of the outer solar system may allow formally unstable orbits to persist for very long times , possibly even to the age of the solar system , depending on the planet mass ; if so , this would weaken the argument for a resonant planet orbit .
in future work it would be useful to investigate scattering efficiency as a function of the planet mass , as well as dynamical lifetimes of non - resonant planet - crossing orbits in this region of the outer system .
nevertheless , the possibility that resonant orbital relations could be a useful aid to prediction and discovery of additional high mass planets in the distant solar system makes a stimulating case for renewed study of aspects of solar system dynamics , such as resonant dynamics in the high eccentricity regime , which have hitherto garnered insufficient attention .
rm thanks scott tremaine , daniel fabrycky , sean mills and anthony dobrovolskis for helpful comments on this work , eric christensen for an observer s perspective , and david frenkel for many clarifying discussions .
rm and kv acknowledge funding from nasa ( grant nnx14ag93 g ) .
xw acknowledges funding from national basic research program of china ( 973 program ) ( 2012cb720000 ) and china scholarship council .
this research made use of the nasa astrophysics data system bibliographic services . | the four longest period kuiper belt objects have orbital periods close to integer ratios with each other . a hypothetical planet with orbital period @xmath0 17,117 years , semimajor axis
@xmath0 665 au , would have n/1 and n/2 period ratios with these four objects .
the orbital geometries and dynamics of resonant orbits constrain the orbital plane , the orbital eccentricity and the mass of such a planet , as well as its current location in its orbital path . |
the emergence of similarities between the properties of molecular glasses , dense colloidal suspensions , foams , and granular materials , have led to the notion of jamming @xcite - the transition between solid - like and fluid - like phases in disordered systems - as a manner through which one can gain a deeper understanding of the traditional liquid - glass transition and the fascinating , complex phenomena observed in amorphous materials in general @xcite .
although there has recently been some works highlighting the differences between the onset of rigidity in jammed matter and glassiness @xcite , the emphasis in this work is on the commonalities between the two @xcite . here
, we aim to provide a heuristic physical picture that accounts for specific , long wavelength , structural features that emerge in the jammed state and relate these features to their dynamical properties .
donev et al .
@xcite found that in the _ hard sphere _
jamming transition , the structure factor , @xmath3 , exhibits a linear dependence on wavenumber @xmath4 , @xmath5 this behavior of @xmath3 suggests that the total correlation function @xmath6 , decays as @xmath7 , at large separations @xmath8 , as deduced from the asymptotic estimates of fourier transforms @xcite .
similarly it suggests a long range behavior for the direct correlation function @xmath9 .
this is indeed in contrast to standard liquid state theory for liquids whose constituents interact via a finite range potential - hence @xmath9 is short ranged - which predicts , @xmath10 .
this anomalous low-@xmath0 behavior can be interpreted as being indicative of the suppression of long wavelength density fluctuations due to hyperuniformity @xcite . here , we propose an alternative interpretation related to large length - scale collective dynamics .
this paper is arranged as follows .
we provide a brief overview of the molecular dynamics ( md ) simulations used here to generate liquid and jammed states .
we then review previous results from studies of the jamming transition pertinent to the discussion here .
this is followed by a discussion on the relevant concepts from liquid state theories that indicate that our results and those of ref . for frictionless particles
are indeed rather unusual .
we then present our results for the static structure factor , @xmath3 , at small values of the wavenumber @xmath0 , in our jammed , model glassy system .
we then put forward a conjecture that relates the asymptotic behavior of @xmath3 to an excess of vibrational modes relative to the debye model .
we end with results from ongoing work on frictional systems and conclusions .
the computer simulations performed here are for monodisperse , _ soft spheres _ of diameter @xmath11 and mass @xmath12 , interacting through a finite range , purely repulsive , one - sided , harmonic potential , @xmath13 where @xmath14 is the center - to - center separation between particles @xmath15 and @xmath16 located at @xmath17 .
the strength of the interaction is parameterized by @xmath18 , which is set to unity in this study .
most of the results presented below are for frictionless particles , with friction coefficient @xmath19 .
we also present preliminary results for frictional packings using a static friction model @xcite to compare between frictionless and frictional systems . in the frictional packings
the particle friction coefficient was varied , @xmath20 .
we simulated systems ranging in size , @xmath21 particles , in cubic simulation cells of size @xmath22 , with periodic boundary conditions , over a range of packing fractions , @xmath23 .
the jamming protocol implemented here to generate zero - temperature jammed packings is similar to other soft - sphere protocols @xcite .
initially , starting from a collection of spheres randomly placed in the simulation cell at low packing fraction @xmath24 , we compressed the system to a specified over - compressed state , @xmath25 , minimizing the energy of the system in a steepest descent manner . at this value of @xmath26 , all the particles experience overlaps with several other particles - their contact neighbors .
the packings are mechanically stable and disordered with no signs of long range order . to generate packings at @xmath27 ,
we then incrementally reduced @xmath28 in steps of , @xmath29 , minimizing the energy after each step .
this allowed us to accurately determine the location of the jamming transition where the system unjams at a packing fraction @xmath30 ( see below ) , for _ each independent realization _ , down to an accuracy of @xmath31 in @xmath28 for @xmath32 and an accuracy of @xmath33 for @xmath34 .
for the largest system , @xmath35 , we generated four independent configurations at each value of @xmath28 .
we also studied equilibrated liquids of frictionless spheres at a dimensionless temperatures @xmath36 .
this value of @xmath37 was chosen to compare and contrast our results between liquid and jammed states .
our main tool of analysis is the static structure factor @xmath3 , which we obtain in the standard way as a direct fourier transform of the particle positions , @xmath38 here , @xmath39 is the reciprocal lattice vector for the periodic simulation cell , which is restricted to @xmath40 .
although the data is noisy at the lowest wavenumbers , our largest packings ( @xmath35 ) display the anomalous low-@xmath0 behavior , as described below , well above the noise level . throughout this study , @xmath41 , and
all quantities have been appropriately non - dimensionalized .
frictionless , purely repulsive , soft sphere system have been shown to undergo a zero - temperature transition between jammed and unjammed phases as a function of density or packing fraction @xcite . in the infinite - size limit
, this transition occurs at a packing fraction coincident with the value often quoted for random close packing , @xmath42 @xcite . above the jamming threshold @xmath43 ,
the jammed state is mechanically stable to perturbations with non - zero bulk and shear moduli , whereas below the transition @xmath44 , it costs no energy to disturb the system @xcite .
the relevant parameter here is not the absolute value of the packing fraction , rather the _ distance _ to the jamming transition defined through , @xmath45 .
thus , as a jammed state is brought closer to the transition point , @xmath46 , it gradually loses its mechanical rigidity and becomes increasingly soft .
intriguingly , this loss of mechanical stability as the jamming transition is approached can be related to diverging length scales @xcite that characterize the extent of soft regions that determine the macroscopic behavior of system @xcite .
experiments @xcite have also identified growing length scales in the vicinity of the jamming transition thus promoting the idea that the jamming transition can be considered in the context of critical phenomena . in this paper , we connect the long wavelength structural features observed in @xmath3 to correlation lengths characterizing the typical length scale of collective , low - frequency vibrational modes in jammed , zero - temperature , disordered packings @xcite . in ref .
it was shown that the approach of the jamming transition in a soft sphere packing is accompanied by a dramatic increase in the number of low - frequency vibrational modes over the expected debye behavior . in traditional glasses , these excess low - frequency modes in the vibrational density of states , @xmath47
, are often referred to as the _ boson peak _ @xcite in reference to the peak observed when plotting @xmath48 . for convenience
, we also employ this language here .
a detailed study of @xmath49 for jammed sphere - packings and the appearance of the so - called boson peak can be found in @xcite . in fig .
[ fig1 ] we compare the vibrational density of states for dense , soft sphere liquids ( fig .
[ fig1](a ) ) and amorphous jammed solids , identifying the location of the boson peak , @xmath50 , for our jammed system ( fig .
[ fig1](b ) ) .
the location of the boson peak tends to zero at the jamming transition point , i.e. @xmath51 , as @xmath46 @xcite .
the two values of @xmath52 and @xmath53 correspond to actual packing fraction values of @xmath54 and @xmath55 respectively , for sample configurations used to generate this data .
, for two systems of @xmath56 soft , frictionless spheres , characterized by the distance @xmath57 , from their zero - temperature , jamming transition packing fraction .
solid line : @xmath58 . dashed line : @xmath59 .
( a ) equilibrated liquids at @xmath60 , well above their respective freezing or apparent glass transition temperatures .
data obtained from fourier transforms of velocity autocorrelation functions , and has been rescaled such that the @xmath61 intercepts coincide . a lorentzian function ( dotted ) corresponding to langevin diffusion @xcite
is also shown .
( b ) jammed packings at zero temperature .
the location of the _ boson peak _ is identified as @xmath50 , for @xmath62 . in ( b ) , the debye result ( dotted ) , corresponding to the system @xmath63 , was generated using values for the bulk and shear moduli @xcite , and is shown for comparison.,title="fig:",width=264 ] , for two systems of @xmath56 soft , frictionless spheres , characterized by the distance @xmath57 , from their zero - temperature , jamming transition packing fraction .
solid line : @xmath58 . dashed line : @xmath59 .
( a ) equilibrated liquids at @xmath60 , well above their respective freezing or apparent glass transition temperatures .
data obtained from fourier transforms of velocity autocorrelation functions , and has been rescaled such that the @xmath61 intercepts coincide . a lorentzian function ( dotted ) corresponding to langevin diffusion @xcite
is also shown .
( b ) jammed packings at zero temperature .
the location of the _ boson peak _ is identified as @xmath50 , for @xmath62 . in ( b ) , the debye result ( dotted ) , corresponding to the system @xmath63 , was generated using values for the bulk and shear moduli @xcite , and is shown for comparison.,title="fig:",width=264 ] associated with this excess in the vibrational density of states are two diverging length scales : the longitudinal correlation length , @xmath64 , characterizing the scale of collective excitations of longitudinal modes contributing to the boson peak , while @xmath65 , characterizes transverse excitations .
these correlation lengths scale with the boson peak position as @xcite , @xmath66 thus , when the debye contribution , @xmath67 , is included the corresponding dispersion relations become , @xmath68 the preceeding relations are only approximate , valid for low frequencies .
they both contribute to the boson peak : the longitudinal term by modifying the slope in the debye relation , whereas the transverse contribution contains an anharmonic term .
we discuss the relevance of these results further below . in order to avoid any misunderstanding
we define the concepts of regular and singular in the following sense @xcite .
assuming @xmath0 to be a complex variable , for any three dimensional system interacting with tempered pairwise , additive potentials , namely when @xmath69 and therefore integrable , @xmath70 .
then , the structure factor , in the long wavelength limit reads @xcite , @xmath71 where @xmath72 is the number density , @xmath73 is a constant that depends on the thermodynamic properties of the equilibrium system , and @xmath74 is an analytic function of @xmath0 , regular in the neighborhood of @xmath75 .
thus , the second term on the right hand side of eq . [ eq9 ] is the regular contribution to @xmath3 , whereas the first term - originating from the potential of interaction - is the singular term .
for instance , when @xmath76 , as in the attractive part of the lennard - jones potential , @xmath77 a result originally derived by enderby , gaskill , and march @xcite .
thus , a linear behavior of the structure factor at small values of @xmath0 corresponds to @xmath78 .
physically , this would correspond to a charge - dipole interaction , which chan _ et al .
@xcite have shown can only be present if _ long - ranged _ dipole - dipole interactions are also included .
thus , in the present work , the linear behavior of @xmath3 , at small @xmath0 , does not come from the potential of interaction , as the finite range harmonic potential used here will only give the regular contribution to @xmath3 .
however , this does suggest that the anomalous low-@xmath0 behavior originates in long - ranged correlations associated with the system .
we initially generated a number of over - compressed , @xmath43 , zero - temperature packings at various distances , @xmath79 , from the jamming transition point .
we also point out that the main distinction between this work and that of ref . , is that we study soft - spheres and approach the jamming transition from above , @xmath80 , whereas
, hard spheres necessarily approach the transition from below . in fig .
[ fig2 ] we show @xmath3 for two soft - sphere systems at @xmath36 , well above their respective apparent glass transition temperatures @xmath81 .
( the corresponding density of states are shown in fig .
[ fig1](a ) . ) for @xmath82 , @xmath83 , while for smaller @xmath84 , @xmath81 lies below this .
qualitatively the curves are similar .
we find the usual primary peak corresponding to nearest neighbors , and , as expected from standard liquid state theory @xcite , at lower @xmath0 , @xmath10 , with a non - zero intercept at @xmath75 , which corresponds to the finite compressibility sum rule , fig .
[ fig2 ] inset .
thus , for equilibrated liquids the structure factor and the long - wavelength limit are quite insensitive to the location of the zero - temperature jamming transition , at these values of the number density , @xmath85 for @xmath86 respectively .
soft - sphere systems in the liquid state , @xmath60 , for @xmath87 and @xmath88 ( see legend in panel ( b ) ) .
this temperature is well above the respective glass transition temperatures for the two systems , @xmath89 .
( a ) static structure factor @xmath3 , as a function of wavenumber @xmath0 .
the inset is a zoom of the region near the origin , where the arrows indicate the non - zero intercepts as @xmath90 .
( b ) radial distribution function , @xmath91 shows a prominent nearest neighbor peak at @xmath92 and correlations that die off rapidly.,title="fig:",width=264 ] soft - sphere systems in the liquid state , @xmath60 , for @xmath87 and @xmath88 ( see legend in panel ( b ) ) .
this temperature is well above the respective glass transition temperatures for the two systems , @xmath89 .
( a ) static structure factor @xmath3 , as a function of wavenumber @xmath0 .
the inset is a zoom of the region near the origin , where the arrows indicate the non - zero intercepts as @xmath90 .
( b ) radial distribution function , @xmath91 shows a prominent nearest neighbor peak at @xmath92 and correlations that die off rapidly.,title="fig:",width=264 ] turning our attention to the jammed phases at zero temperature @xcite , in fig .
[ fig3 ] we show the radial distribution function @xmath91 and the low-@xmath0 region of @xmath3 , for @xmath35 soft - spheres systems at three values of @xmath93 , @xmath94 , and @xmath95 .
note that in fig .
[ fig3](a ) we use log - log scales to clearly demonstrate the linear region in @xmath3 . far from the jamming transition , @xmath93 , @xmath3 plateaus near @xmath96 and tends to constant as @xmath90 . for the systems closer to the jamming transition , @xmath97 and @xmath98 ,
there is a _ qualitative change _ in the low-@xmath0 behavior of @xmath3 .
we find a linear region , @xmath1 , extending over almost an order of magnitude at low-@xmath0 .
we also point out that on closer inspection the linear region extends to lower @xmath0 for the system closest to the jamming transition point @xmath99 . at the smallest @xmath0 attainable , @xmath3 ,
flattens out again because the system is not exactly at @xmath100 .
the radial distribution function shown in fig .
[ fig3](b ) exhibits typical features characterisitc of a glassy phase ; namely a split second peak an additional shoulder on the third peak .
purely - repulsive , frictionless , soft - spheres at three different values of @xmath93 ( dotted line ) , @xmath94 ( dashed ) , and @xmath101 ( solid ) .
( a ) static structure factor @xmath3 .
far from the jamming transition , @xmath102 plateaus nears @xmath96 .
closer to jamming , @xmath103 and @xmath104 exhibits approximately linear dependence on @xmath0 over almost an order of magnitude in @xmath0 , extending down to low-@xmath0 .
a linear curve on this log - log plot is shown for comparison .
( b ) radial distribution function .
the jamming transition is characterized by a diverging nearest neighbor peak at @xmath105 , a clear splitting of the second peak @xcite , and oscillations that persist out to larger @xmath8 than for the liquid state.,title="fig:",width=264 ] purely - repulsive , frictionless , soft - spheres at three different values of @xmath93 ( dotted line ) , @xmath94 ( dashed ) , and @xmath101 ( solid ) .
( a ) static structure factor @xmath3 .
far from the jamming transition , @xmath102 plateaus nears @xmath96 .
closer to jamming , @xmath103 and @xmath104 exhibits approximately linear dependence on @xmath0 over almost an order of magnitude in @xmath0 , extending down to low-@xmath0 .
a linear curve on this log - log plot is shown for comparison .
( b ) radial distribution function .
the jamming transition is characterized by a diverging nearest neighbor peak at @xmath105 , a clear splitting of the second peak @xcite , and oscillations that persist out to larger @xmath8 than for the liquid state.,title="fig:",width=264 ] for completeness , we examine the influence of system size on the results presented here . in fig .
[ fig4 ] , we show @xmath3 at @xmath106 for three different systems sizes , @xmath107 , corresponding to @xmath108 .
the main panel of fig .
[ fig4 ] shows that the gross properties of @xmath3 do not depend on system size .
oscillations in @xmath3 persist out to the largest @xmath0 and is a consequence of the diverging nearest neighbor peak in @xmath91 @xcite .
the inset to fig .
[ fig4 ] indicates that the linear portion of @xmath3 at small @xmath0 becomes resolvable for @xmath109 . at @xmath110 for different systems sizes , @xmath111 ( line ) , @xmath112 ( @xmath113 ) , and @xmath114 ( @xmath115 ) .
oscillations persist out to large @xmath0 , reflecting the dominant nearest neighbor peak in @xmath91 shown in fig .
the inset indicates that the linear behavior at low-@xmath0 becomes better resolved with increasing system size.,width=264 ] thus the first aim of this work shows that , similarly to the findings in ref . for hard spheres just below the jamming transition point , for our soft sphere packings above @xmath116
, @xmath3 exhibits a linear behavior at small values of momentum transfer .
hence , this unusual behavior not only pertains to hard spheres , but also to soft spheres , both above and below the jamming transition point , and possibly , to finite range repulsive potentials in general .
we will attempt below to produce what we believe is a reasonable explanation for the origin of this behavior which , as stated at the beginning , does appeal to a conjecture .
since , in the long wavelength limit , the linear behavior of @xmath3 can not arise from the singular contribution of the potential used in this work @xcite , it can only be due to the collective excitations present in the jammed or glassy state @xcite .
the collective excitations we have in mind are the same that are responsible for the boson peak .
the range of @xmath0 , over which @xmath117 , characterizes the length scale over which these excitations may be considered collective .
as one approaches the jamming transition , this length scale diverges , a result consistent with work on the density of states of jammed packings @xcite . recent work by chumakov _
@xcite does indeed show that the excitations leading to the boson peak are predominantly collective ; in agreement with inelastic neutron scattering experiments @xcite .
although there have been a number of theories put forward @xcite , there is , at present , no agreed explanation as to the origins of the boson peak .
for the purposes of this work this is not necessary , except insofar as there is agreement that these excitations are predominantly collective , and that the vibrational modes are likely to be kinetically driven .
the arguments that follow are somewhat oversimplified , but we believe they are along the correct lines .
the following comments are in order : ( i ) in a system that undergoes the liquid - glass transition , the long wavelength limit of the static structure factor , @xmath118 , is not the thermodynamic compressibility @xmath119 , but a different value , say @xmath120 , which can actually be extracted from experiments @xcite .
( ii ) independently of the model or approximation used , the density fluctuations in the glass phase are kinetically , not thermodynamically , driven @xcite .
( iii ) in spite of their different origins , the kinetic temperature , say @xmath121 , defined for glasses , matches the edwards temperature , @xmath122 , used in granular matter .
they happen to coincide within mean field theories when the glass is in contact with an almost zero temperature heat bath , and the athermal grains are jammed @xcite .
the connection between the dispersion relations and the structure factor , follows from the second moment of the dynamical structure factor , @xmath123 @xcite , @xmath124 where @xmath125 ( for a system in thermodynamic equilibrium @xmath126 , where @xmath127 is boltzmann s constant ) . the static structure factor is the zero moment of @xmath123 , @xmath128 we now put forward the following conjecture : assume that there is only one very well defined collective mode with dispersion relation @xmath129 , the boson peak .
this gives us a relation between the asymptotic long wavelength behavior of @xmath3 and the boson peak , @xmath130 whence , @xmath131 this collective mode is associated with vibrational excitations with a wavelength of the order of the correlation length of the jammed state .
we expand the dispersion relation at these small values of momentum transfer @xcite @xmath132 in eq .
[ eq15 ] , @xmath133 is the speed of sound , and the second term on the rhs denotes departures from the usual debye behavior , such that both @xmath133 and @xmath134 are independent of @xmath0 .
replacing eq .
[ eq15 ] into [ eq14 ] , in the limit of @xmath90 , we find , @xmath135 thus , on comparing eqs .
[ eq15 ] and [ eq16 ] to eqs .
[ eq6 ] and [ eq7 ] , it transpires that it is the transverse modes that contribute to the linear behavior of @xmath3 . on the other hand , the longitudinal components , eq .
[ eq6 ] , only contribute a constant term to @xmath3 , modifying the slope of the debye relation .
we illustrate the differences in the dispersion behavior at two extreme values of @xmath136 and @xmath137 in fig .
these results were obtained from diagonalization of the dynamical matrix for systems with @xmath138 and then locating the peaks in the transverse components of the fourier transforms of the eigenmodes - transverse mode structure factors @xcite - and then averaging over a small range of frequencies .
the data shown in fig . [ fig5 ] distinguishes between the regular , linear dispersion relation that dominates the dispersion behavior far from the jamming transition at @xmath82 , while the quadratic contribution to the dispersion behavior is significant closer to the jamming transition , @xmath136 . and @xmath139 .
the solid lines correspond to quadratic fits to the data as in eq .
data obtained from the low - frequency portion of the transverse structure factors of the vibrational modes for @xmath140.,width=264 ] from the analysis we obtain the transverse speed of sound , @xmath141 , as a fitting parameter . at these two compressions : @xmath142 and @xmath143 .
the actual values calculated from the bulk and shear moduli data @xcite give : @xmath144 and @xmath145 .
therefore , we find reasonable agreement within this approximation .
if we use only a linear dispersion relation we find that the values of @xmath141 computed here and the actual values @xcite vary by as much as an order of magnitude .
this can be realized by the observation that a linear fit to data for @xmath146 does not pass through the origin as required in the hydrodynamic limit .
the linear feature of @xmath3 in the small-@xmath0 regime appears here as a consequence of assuming an `` excess '' relative to the debye model for the dispersion relation , and only one collective mode .
although the `` excess '' model used here appears as an approximation appropriate for the low-@xmath0 regime , we have reconciled this with existing results on the emergence of characteristic length scales associated with these collective excitations . more generally , however , it does suggest that the boson peak and the linear behavior of @xmath3 , as @xmath90 , are two sides of the same coin . therefore , this anomalous suppression of long - wavelength density fluctuations emerges as a consequence of large length - scale correlated dynamics in the low - frequency modes of the jammed solid . furthermore , although the boson peak has been traditionally associated with the glassy phase , recent low frequency raman spectroscopic studies of glassy , supercooled , and molten silica reveal that the boson peak persists into the liquid phase @xcite .
this also appears to be the case with other network systems ( see references in ref . ) which , in the molten state , show a distinctive prepeak at small @xmath0 in @xmath3 .
this prepeak is indicative of intermediate range order in those melts , representing a characteristic length in those melts that measures the correlations between the centers of `` clusters '' present in the liquid state .
this intermediate - range clustering may in fact promote longer wavelength correlated dynamics in the liquid state .
it may therefore be interesting to investigate experimentally , and by means of md simulations , whether the low @xmath0-behavior of @xmath3 in those systems is also linear , both in their glassy and liquid phases .
studies on jammed packings of _ frictional _ particles are less well - developed .
the picture that is emerging is that frictional packings undergo a similar zero - temperature jamming transition at packing fractions which now become friction dependent @xcite . in the limit of high friction coefficient
the frictional jamming transition coincides with the value associated with random loose packing , @xmath147 @xcite .
thus the relevant parameter that measures the distance to the jamming transition now becomes a friction - dependent quantity , @xmath148 @xcite .
, focusing on the low-@xmath0 region .
all packings at the same @xmath149 , contain @xmath35 purely - repulsive , monodisperse , soft - spheres , with different friction coefficients @xmath150.,width=264 ] we test these concepts through @xmath3 for packings over a range of friction coefficients at the _ same _ @xmath149 , which translates to different values of @xmath148 . our preliminary results in fig .
[ fig6 ] are data for @xmath3 for jammed packings with @xmath151 . at this fixed value of @xmath149 , @xmath148 increases with increasing friction coefficient .
therefore , we expect that the anomalous linear behavior in the low-@xmath0 region of @xmath3 should become less prominent with increasing @xmath152 .
indeed this trend is observed in fig .
[ fig6 ] , thus providing yet further evidence that the jamming transition point shifts to lower @xmath28 with increasing friction . to recover the linear behavior at low-@xmath0 in frictional packings
we would therefore need to study mechanically stable packings at lower @xmath28 .
this work in ongoing .
in conclusion , we have provided a physical , albeit naive , explanation for the observation of the apparent suppression of long wavelength density fluctuations in jammed model glassy materials .
we submit that it is the presence of low - frequency , collective excitations , mainly of transverse character , contributing to the excess of low - frequency modes , that are responsible for the linear behavior at low-@xmath0 in @xmath3 .
the relevant length scale here is the transversal correlation length that contributes to the diverging boson peak in the jammed phase .
this low-@xmath0 feature is most pronounced in the model system studied here for soft particles at packing fractions above the zero - temperature jamming transition matching the behavior on the other side of the jamming transition for hard spheres .
this connection between the long - wavelength behavior on either side of the transition reinforces the view that the jamming transition is critical in nature .
although we also point out that the jamming transition is very different from a thermodynamic liquid - gas critical point in that we do not observe a divergence in @xmath153 , but rather it becomes zero .
moreover , we expect that as the transition is approached , @xmath154 the dependence of @xmath155 on a generalized temperature - like quantity , @xmath156 , in jammed packings is consistent with the concept of the _ angoricity _ @xcite .
this plays the role of temperature in packings of elastic particles and goes to zero when the particles fall out of contact which occurs at the jamming transition .
future studies may also provide a way to understand the connection between different temperature definitions relevant to the study of thermal and jammed systems .
our preliminary data for frictional packings also provide further evidence that the location of the jamming transition occurs at lower packing fractions with increasing friction coefficient .
the underlying nature of the low - frequency modes in frictional materials has yet to be investigated in three dimensional systems and forms part of ongoing work .
when friction is present the modes will contain not only the translational character as seen in frictionless systems , but also rotational character due to the additional degrees of freedom .
this anomalous low-@xmath0 behavior is also likely to be present in other , finite range , model repulsive systems .
these features may be detected , using the appropriate spectroscopy , in ( hard - sphere ) colloidal glasses , and granular packings , which can be prepared close to their respective jamming transitions .
we are grateful to gary barker for insightful discussions on some aspects of this work .
les is especially grateful to jane and gary mcintyre for support during the course of this work and also gratefully acknowledges the support of the national science foundation cbet-0828359 . | the structural properties of static , jammed packings of monodisperse spheres in the vicinity of the jamming transition are investigated using large - scale computer simulations . at small wavenumber @xmath0
, we argue that the anomalous behavior in the static structure factor , @xmath1 , is consequential of an excess of low - frequency , collective excitations seen in the vibrational spectrum .
this anomalous feature becomes more pronounced closest to the jamming transition , such that @xmath2 at the transition point .
we introduce an appropriate dispersion relation that accounts for these phenomena that leads us to relate these structural features to characteristic length scales associated with the low - frequency vibrational modes of these systems . when the particles are frictional , this anomalous behavior is suppressed providing yet more evidence that jamming transitions of frictional spheres lie at lower packing fractions that that for frictionless spheres .
these results suggest that the mechanical properties of jammed and glassy media may therefore be inferred from measurements of both the static and dynamical structure factors . |
* schematic distribution in the complex plane of the poles , @xmath41 , given by eq .
( [ 10 ] ) , for @xmath35 ( a ) and for @xmath48 ( b ) .
the poles are located in the solid points , distributed along radial axes , as explained in the text . | we investigate the scope of a previous result concerning the behaviour of fermions hitting a general wall caused by a first - order phase transition .
the wall profile function was considered to be analytic in the real axis .
the previous result is valid for analytic functions in the whole complex plane except in certain isolated singularities located out of the real axis .
a non - analytic profile function in the real axis is studied in order to show the validity of the result for any profile which can be put as a certain limit of a function which verifies the latter .
a new understanding of the high energy behaviour of the quantum reflection caused by a sharp profile , as the step , arises from that study .
pacs . : * 11.10q - field theory , relativistic wave equations . *
11.80f - relativistic scattering theory , approximations . * 11.30q - symmetry and conservation laws , spontaneous symmetry breaking .
# 1([#1 ] ) fanse-96/10 + hep - ph/9610290 * do the profile function singularities explain the high energy reflection of fermions in a phase transition ? *
jos rodrguez - quintero departamento de fsica atmica , molecular y nuclear , universidad de sevilla , spain + the scattering of high energy fermions by a wall separating two phases of different symmetry properties has received much attention recently .
the main physical motivation of these works is the idea@xcite that the baryon asymmetry of the universe might have been produced if the cosmological electroweak phase transition has been of first order .
the transition is described in terms of bubbles of a spontaneously broken symmetry vacuum expanding in a preexisting symmetric one . in this scenario
the point to elucidate is whether there exists a cp - asymmetry that produces a different reflection and transmission probability for quarks and antiquarks in order to explain , via the standard model baryon number anomaly@xcite , the correct baryon asymmetry of the universe . in order to simplify the treatment ,
an useful assumption is to break down the process into two steps , one describing the production of cp asymmetry when the quarks / antiquarks are reflected on the wall , the second describing the transport and the eventual transformation of the cp asymmetry into a baryon asymmetry .
the first of two steps justifies the effort concentrated in the study of the scattering of fermions in the presence of first order phase transition@xcite,@xcite,@xcite .
the structure of the wall depends on the higgs field effective potential that takes into account the effects of the surrounding plasma trough the temperature of a certain thermal bath .
the wall profile obtained by solving the equation of motion with this effective potential is rather complex and depends on too many coupling constants@xcite , thus the study of the general wall profile problem@xcite@xcite is justified not only from a purely formal interest .
we have shown in a previous work@xcite the connection between the complex plane poles of the wall profile function and the behaviour of fermions hitting the wall in the high energy limit .
nevertheless , by performing an extreme simplification which allows to compute the feynman fermion propagator in an exact way @xcite , the wall profile can be approximated by a step function , i.e. a sudden jump from one phase to the other , that is called the thin wall approximation@xcite , @xcite . in this case , the reflection coefficient in the high energy limit is ] , r(e)=m_02e , [ 1 ] being the mass @xmath0 the high of the wall .
notice the power law dependence on @xmath1 of the reflection coefficient in eq .
( [ 1 ] ) . as we will see , the results given in ref .
@xcite can not be directly applied to profile functions which are not analytic in the real axis , as the step profile function .
nevertheless , in this note we show the high energy behaviour of a new kind of profile functions , which can be considered in a certain limit as analytic extensions in the complex plane of the step function , by following @xcite .
furthermore , we will investigate the appropriate limit in order to understand the result ( [ 1 ] ) through the analytic properties of the step function extensions .
next , we outline the derivation of the high energy asymptotic expression for the reflection coefficient given in ref .
@xcite , where the detailed calculation can be found . by formulating the problem in the rest frame of a wall normal to the @xmath2-axis , characterized by a general non - cp - violating wall profile , and working in the chiral basis@xcite
, we can factor the dirac equation into @xmath3 identical blocks .
thus , the problem is reduced to solve the following equation , ( i _ z + q(z ) ) _ i / ii = 0 , [ 2 ] with \(z ) = ( cc e & -m(z ) + m(z ) & -e ) = _ i _ ii e^-iet ; [ 3 ] where @xmath4 is the time - independent dirac equation solution in the chiral basis .
we express the solution of ( [ 3 ] ) as follows _
i / ii(z ) = ( z , z_0 ) ( c _
1(z_0 ) + _ 2(z_0 ) ) , [ 4 ] with ( z , z_0)=_p_e^i^z_z_0d ( ) ; [ 5 ] where @xmath5 indicates a path ordered product and @xmath6 is the position variable along the z - axis .
we consider @xmath7 , where @xmath8 is the profile wall function .
the asymptotic conditions @xmath9 and @xmath10 are required , @xmath11 decreasing exponentially when @xmath12 .
thus , after some tedious calculation we obtain from equations ( [ 3 ] ) and ( [ 5 ] ) ( z , z_0 ) = ( 1 + ( m_0 e)_2 _ z_0^z d p ( ) f ( ) e^-2i_3_^z dp ( ) .
+ + .
o ) e^i_3 _ z_0^z d p
( ) ; [ 6 ] where @xmath13 ^ 2\right)^{1/2}$ ] . following nelson _
et al_@xcite , we can obtain the reflection coefficient from the result ( [ 6 ] ) .
thus , by assuming in general that f()=f ( ) , [ 7 ] where the parameter @xmath14 gives the wall thickness , and after much calculation@xcite , we finally obtain r(e ) = 2 m_0 _
j=1^n e^-2ey_j e^2iex_j _ n=1^_j b_-n^j ( 2ie ) ^n-1 ( n - 1 ) ! ; [ 8 ] where @xmath15 and @xmath16 , for @xmath17 , are all the poles of @xmath18 with positive imaginary part and the n - power coefficient of the laurent expansion for the function in each pole , respectively .
the order of the pole @xmath19 is @xmath20 .
the result ( [ 8 ] ) is valid in general for profile functions which are analytic in the real axis , i.e. analytic functions in all the complex plane except in certain isolated singular points with non - zero imaginary part .
it is positively checked in ref .
@xcite by using the particular kink - type wall profile@xcite@xcite , @xmath21 .
we consider now the more general kind of functions f()= 11 + ( -^2n+1 ) .
[ 9 ] it is obvious that the functional behaviour of the kink - type wall is a particular case of the latter , given by taking @xmath22 .
moreover , any of these functions gives the step function if the limit @xmath23 is taken . in order to apply the result ( [ 8 ] ) to the functions ( [ 9 ] )
, we must know the singularities of these functions and their distribution in the complex plane .
taking into account eq .
( [ 7 ] ) , we obtain for the poles of @xmath18 z_jk = ^1m(1 + 2j)^1 m e^im(12 + 2k ) ; [ 10 ] with @xmath24 and @xmath25 . where @xmath26 is the power in the exponential argument in ( [ 9 ] ) .
the order of all the poles is @xmath27 and the residous , @xmath28 , can be written as b_-1^jk = 1m^1-mm e^i1-mm(12 + 2k ) .
[ 11 ] it is easy to see that these poles are distributed in the complex plane along @xmath29 radial axes , on circumferences with radius given by @xmath30^{1/m}$ ] , where @xmath31 . in fact , the distribution of the poles is symmetric with regard to the real axis .
the schematic location of the poles in the complex plane for @xmath32 is shown in fig .
1 .
it can be found that the two poles with lower positive imaginary part are z_0,0=^1me^i2 m z_0,m-12=^1me^i(1 - 12 m ) , [ 12 ] and by applying the result ( [ 8 ] ) in the range of energies @xmath33 , we obtain r(e)=2(2-_m,1)m^1mm_0 e^-2 ^ 1m(2m)e ; [ 14 ] where all the decreasing exponential terms except the two with lowest arguments has been neglected .
the factor @xmath34 arises because the two poles considered in eq .
( [ 12 ] ) are the same for @xmath35 .
the exponential law of eq .
( [ 14 ] ) is the expected behaviour for the reflection coefficient provided that the result ( [ 8 ] ) can be applied , if the energy is large enough .
therefore , the reflection coefficient for the step wall profile which presents a power law asymptotic behaviour , as we mentioned above , does not seem to be coherent with the former .
nevertheless , we must take into account that the result ( [ 8 ] ) relates the high energy behaviour of the fermions hitting a wall to the analytic properties of the profile function in the complex plane .
thus , we must appropriately extend the step profile in the complex plane to investigate if the obtained poles allow to explain the behaviour given by ( [ 1 ] ) .
it can be easily seen that the limit of the functions ( [ 9 ] ) , for all @xmath36 , when the parameter @xmath14 goes to zero is the step function .
therefore , in that limit , any of those can be considered as an analytic extension in the complex plane for the step function . by considering that the function @xmath37 is the same than @xmath8 , but scaled by the parameter @xmath14
, we have the same pattern for the distribution of poles of @xmath8 than for that of @xmath37 , but with the distance between two consecutive poles depending on @xmath14 .
in other words , the poles will be as close to each other along the same radial axes as @xmath14 small .
the result ( [ 8 ] ) can not be strictly applied for @xmath38 , but we know that if the limit when @xmath23 of the high energy reflection coefficient exists , it must be the same obtained by using ( [ 8 ] ) , which is valid for any non - zero @xmath14 .
the crucial point in this approach is that when @xmath14 is getting smaller and smaller we can not consider a large enough energy , @xmath1 , to neglect the contribution other than the closest poles to the real axis . in this way
, the contributions due to each pole have to be summed before to take the limit and there is no reason to expect an exponential final result .
indeed , we have r(e ) = 2 m_0 _ k=0^m-1_j=0^n
e^-2ey_j e^2iex_j b_-1^jk , [ 15 ] where @xmath28 is defined in eq .
( [ 11 ] ) and @xmath39 , @xmath40 are the real and pure imaginary components of @xmath41 , given by eq .
( [ 10 ] ) . by taking @xmath23 in eq .
( [ 14 ] ) , the sum for @xmath42 from @xmath43 to @xmath44 can be considered as an integral and after some appropriate transformations , we obtain r(e)=m_02e _
k=0^m-1 e^i_k _ 0^dt e^-t_ke^it_k ; [ 16 ] where & & _ k = ( 1-m)(12 + 2k ) , + & & _ k = ( 12 + 2k ) .
[ 17 ] as we only sum for the poles with positive imaginary part , the signs @xmath45 in eq .
( [ 17 ] ) are for @xmath46 and @xmath47 , respectively . by solving eq .
( [ 16 ] ) , we obtain r(e)=m_02e _
k=0^m-1 e^i(_k+2-_k ) = m_02e \ { m+12 - m-12 } + = m_02e ; [ 18 ] which agrees with eq .
( [ 1 ] ) . in consequence ,
the power law behaviour which is characteristic for the step function , as well as for any non - continuous function , and therefore non - analytic , in the real axis , can be well understood through the complex analytic properties of these functions .
the high energy reflection behaviour is caused by the sum of the contributions due to each singularity of the profile function .
for a large enough energy the contributions of one or several of these singularities can be considered as leading , provided that the singular points are isolated .
we only must take into account that the real non - continuous point is produced when the poles of the function are distributed in an infinitely compact way along lines in the complex plane which contains this point to understand the power law for the step .
thus , the smoother decreasing on the energy of the reflection coefficient for the sharp profiles can be explained because no singularity contribution can be isolated in order to lead the total result . * aknowledgements * i thank o. pne and m. lozano for essential comments and very helpful discussions .
i am specially grateful to the latter for his careful reading of the manuscript .
11 v.a .
kuzmin , v.a .
rubakov and m.e .
shaposnikov , phys.lett.*b155 * ( 1985 ) 36 .
g.t hooft , phys.rev.lett.*37 * ( 1976 ) 8 ; phys.rev.*d14 * ( 1976 ) 3432 ; + n.s .
manton , phys.rev.*d28 * ( 1983 ) 2019 ; + f.r .
klinkhamer and n.s .
manton , phys.rev.*d30 * ( 1984 ) 2212 ; + s. dimopoulos and l. susskind , phys.rev.*d18 * ( 1978 ) 4500 .
a. ayala , j.j - marian and l. mc lerran , phys .
rev .
* d49 * ( 1994 ) 5559 .
k. funakubo , a. kakuto , s. otsuki , k. takenaga and f. toyoda , phys . rev .
* d50 * ( 1994 ) 1105 .
j. rodrguez - quintero , m. lozano and o. pne , preprint fanse-96/09lpthe orsay-96/24hep - ph/9604273 ( 1996 ) , to be published in phys.lett.*b*. a. e. nelson , d. b. kaplan , a.g .
cohen , nucl . phys .
* b373 * ( 1992 ) 453 .
m. b. gavela , m. lozano , j. orloff , o. pne nucl . phys .
* b430 * ( 1994 ) 345 .
g.r.farrar and m.e.shaposhnikov , phys .
rev .
lett .
* 70 * ( 1993 ) 2833 .
g.r.farrar and m.e.shaposhnikov , cern - th.6732/93 .
j. rodrguez - quintero , m. lozano and o. pne , in preparation . |
oscillatory behavior exhibited by nonlinear chemical systems often takes the form of periodic mixed - mode oscillations consisting of @xmath0 large amplitude oscillations followed by @xmath1 small amplitude oscillations .
the distinctive difference between large and small oscillations allows one to classify the mixed - mode oscillatory states by integer pairs @xmath2 and assign them the symbol @xmath3 .
ordered progressions of such periodic patterns separated by chaotic oscillations are observed as control parameters are varied .
the progressions of different mixed - mode forms are known to be farey sequences described by farey arithmetic @xcite . since their discovery in the belousov
zhabotinsky reaction @xcite , mixed - mode oscillations have been found in a broad range of chemical and biological @xcite systems .
their ubiquity suggests the possibility of the existence of a common dynamical systems theory origin .
a number of authors @xcite identified experimentally observed mixed - mode oscillations with periodic motions arising as a result of phase - locking on a 2-torus .
this hypothesis was based mainly on the fact that the periodic states on an invariant torus form farey sequences as a control parameter runs through the quasiperiodic route from period-1 oscillations to chaos .
numerical evidence for the existence of quasiperiodicity and torus oscillations was obtained from studies of model systems which exhibit mixed - mode oscillations @xcite . however , there are results which show that mixed - mode oscillations can not be attributed to the phenomenon of quasiperiodicity in all systems . in a study of electrodissolution of copper albahadily _
_ @xcite demonstrated that the mixed - mode states emerge in parameter regions different from those where a torus is stable .
the formation of mixed - mode oscillations was observed at a point well beyond the range of parameters between the secondary hopf bifurcation and the torus break - up bifurcation , where the mixed - mode states should lie according to the hypothesis of quasiperiodicity . in a series of electrochemical studies koper _
_ @xcite found that the mixed - mode oscillations in farey sequences observed in their study are separated by chaotic states which resemble random mixtures of the adjacent periodic patterns rather than by quasiperiodic oscillations densely covering the surface of a 2-torus . in a numerical study of a model system koper @xcite found a parameter domain where both quasiperiodic oscillations lying on a 2-torus and a large - amplitude mixed - mode oscillatory state which does not belong to the torus are stable simultaneously . in order to interpret the results of the experimental study in ref .
@xcite ringland _
@xcite constructed a family of two - extremum z - maps capable of producing farey sequences in the limiting case when the central segment of the z - map is vertical .
they have shown that one may obtain farey sequences which are not related to phase locking on a torus . in this paper
we present results of investigations of a model proposed earlier for the qualitative description of the mixed - mode oscillations observed in the belousov
zhabotinsky reaction @xcite . from the analysis of the transformation of the slow manifold of the model as system parameters
are changed we demonstrate the gradual emergence of a horseshoe - type strange set in which all the mixed - mode states are embedded as periodic orbits . on the basis of this analysis we construct a three - dimensional phase space picture describing the emergence and bifurcations of mixed - mode states .
this scenario does not involve a torus and appears to be an alternative to the well - established quasiperiodic scenario . in sec .
[ model ] we introduce the model and present a brief discussion of its attractors and bifurcation structure .
section [ shoe ] is devoted to the detailed analysis of the slow manifold and a description of its important properties .
we introduce a technique that allows us to construct the slow manifold and demonstrate how it turns into a horseshoe by stretching and folding .
the repulsive flow that defines a particular shape of mixed - mode oscillations is introduced and discussed in some detail . in sec .
[ mmo ] we show that when the slow manifold is folded into a horseshoe interacting with the repulsive flow , the mixed - mode oscillations arise naturally at the point where the large - amplitude , period-1 oscillation loses its stability through a saddle - node bifurcation .
the relation of the phase flow on the horseshoe to the z - map introduced by ringland _ et al . _
@xcite is discussed . in sec .
[ conc ] we conclude that the existence of the repulsive flow allows one to consider this horseshoe as a representative of a distinct subclass of strange sets whose periodic windows are organised according to farey sequences .
we consider a three - variable extension @xcite of the boissonade de kepper model @xcite , @xmath4 system ( [ sys ] ) was originally proposed in @xcite in order to obtain a qualitative description of transient mixed - mode oscillations in the belousov zhabotinsky reaction .
depending on the values of @xmath5 and @xmath6 , ( [ sys ] ) possesses either one or three steady states .
the change in the number of stationary points occurs on the curve @xmath7 where two steady states coalesce or emerge in a saddle - node bifurcation . at @xmath8 , where two branches of the curve ( [ det ] ) meet in a cusp point
, the pitch - fork bifurcation takes place .
note that the parameters @xmath9 and @xmath10 only rescale the corresponding components of the vector field and do not influence the positions of the nullclines and steady states in phase space . in the following discussion the geometric parameters @xmath5 and @xmath6 are fixed @xmath11 while @xmath9 and @xmath10 are varied . with this choice of @xmath5 and @xmath6 system ( [ sys ] )
has only one steady state at @xmath12 , which is a stable focus in a domain of the parameter plane @xmath13 lying to the left of the hopf bifurcation line ( see fig .
[ fig01 ] ) , and a saddle - focus otherwise .
as the real characteristic exponent of the steady state is negative everywhere on @xmath13 , no second order bifurcation is possible at the chosen values of @xmath5 and @xmath6 .
figure [ fig01 ] shows the partition of the @xmath13 plane into domains of different types of dynamical behavior according to the results of numerical integration . in domain 1
a stable focus is the only attractor of the system ( [ sys ] ) .
large amplitude period-1 @xmath14 oscillations with a strong relaxational character at large @xmath9 are found in domain 2 .
finally , mixed - mode oscillations are observed in the wedge - shaped domain 3 .
transitions from period-1 to mixed - modes take place on an almost flat curve ( shown by the dashed line in fig .
[ fig01 ] ) where the trajectory of system ( [ sys ] ) performs a small - amplitude excursion only once on the entire time interval @xmath15 .
this transition can be represented symbolically as @xmath16 .
continuing to traverse domain 3 by decreasing @xmath9 at constant @xmath10 one observes pruned farey sequences of periodic mixed - mode oscillations @xmath17 interspersed by narrow , yet detectable , spans of chaos .
this periodic - chaotic progression ends suddenly in a crash onto the focus at the point where the line separating zones 1 and 3 is crossed . at large @xmath10 ( @xmath18 ) , where the crash bifurcation curve appears to be tangental to that for the @xmath19 transition
, both bifurcations occur within an extremely narrow parameter span and one observes hard termination / emergence of large - amplitude periodic oscillations .
another limiting case is observed at low @xmath10 ( @xmath20 ) where the crash curve merges with the hopf bifurcation line . here
the crash occurs from pattern @xmath21 which forms an almost homoclinic connection to the focus undergoing the hopf bifurcation . as @xmath10 varies the appearance of the phase portrait of the mixed - mode states changes considerably . at large values of @xmath10
both small- and large - amplitude excursions lie almost on the same plane .
this type of mixed - mode oscillation , referred to type-1 according to classification given in @xcite , is shown in fig .
[ fig02](a , b ) . with a decrease in @xmath10
the reinjection part of trajectory falls closer to the focus and the angle between the planes of small- and large - amplitude oscillations grows . at @xmath22
the phase portrait of the oscillation acquires all of the features characteristic of type-2 mixed - mode states . here
the large loops are mainly orthogonal to the small ones and the reinjection occurs along the one - dimensional stable manifold of the focus ( see fig .
[ fig02](c , d ) ) .
the parameter region in which complex periodic and chaotic oscillations are found can be partitioned into subdomains corresponding to particular periodic mixed - mode states @xmath23 and their chaotic mixtures .
we shall study the mechanisms responsible for the bifurcations of one mixed - mode state into another by analysing transformations of the phase flow on the system s slow manifold .
the concept of a slow manifold relies on the assumption that the relaxation of an arbitrary initial condition to the attractor has two characteristic time scales : the trajectory quickly reaches certain subset of the phase space and then slowly approaches the attractor within this subset . in the case of a strongly dissipative system with three dynamical variables , the slow set is usually represented by a 2d manifold which is often folded in the 3d phase space . to avoid complications arising from nontrivial embedding of the slow manifold , we choose for its construction the region of large @xmath10 where the slow manifold appears to have a relatively simple organization .
figure [ fig03](a ) presents the projection of the phase flow on the @xmath24 plane for @xmath25 and @xmath26 . for these parameter values
deep within domain @xmath27 the system ( [ sys ] ) is bistable . apart from the stable focus , two stationary periodic solutions exist : a stable , large - amplitude , period-1 limit cycle ( solid line ) and a saddle - type , small - amplitude cycle ( dashed line ) .
the planar , concentric organization of these cycles around the focus suggests that the slow manifold containing them is flat .
trajectories that leave the saddle cycle at @xmath28 and approach stable cycle or focus at @xmath29 lie in the slow manifold .
these trajectories are easy to calculate and one can use them for the construction of a poincar section of the slow manifold .
we choose @xmath30 shown in fig [ fig03](a , b ) as a surface of section .
intersections of the transient trajectories with @xmath31 yield a set of points which , after interpolation by a smooth function @xmath32 @xcite , becomes a continuous zero - order approximation @xmath33 to the poincar section of the slow manifold .
figure [ fig03](b ) shows a 3d phase portrait of the same trajectories as in fig .
[ fig03](a ) , as well as @xmath31 and @xmath34 .
the intersections of the flow initiated on @xmath34 with @xmath31 can be used both to improve the accuracy of the approximation and to construct the first return map ( see fig . [ fig03](c ) ) which quantitatively describes the phase flow on the slow manifold .
a complete description of the phase flow on the slow manifold can be obtained from the comparison of the phase portrait and the first return map .
consider the example presented in fig .
[ fig03](a , c ) .
the first return map @xmath35 intersects the bisectrix in three points which correspond to the stable limit cycle @xmath36 , the saddle cycle @xmath37 , and the stable focus @xmath38 .
the nearly rectangular shape of the map implies that the slow manifold comprises areas of strong expansion ( the neighborhood of @xmath39 , where the map is vertical ) and contraction ( horizontal part of the map ) . in the course of its evolution a trajectory initiated in the vicinity of the saddle cycle executes a number of small - amplitude loops around the saddle cycle and then in one turn reaches the neighborhood of the stable limit cycle . whether the trajectory approaches the limit cycle like trajectory 3 performing small transient loop or like trajectory 4 which makes a large loop depends on its position relative to a certain boundary trajectory 0 which can be identified with the point @xmath40 on the first return map where @xmath41 has a vertical tangent .
analysis of @xmath41 in a small neighborhood of @xmath40 shows that the trajectory 0 is superunstable ; _ i.e. _ , the neighboring trajectories deviate from it faster than geometrically .
figure [ fig04 ] shows @xmath42 versus @xmath43 , where @xmath44 and @xmath45 , for both left @xmath46 ( lower set of points ) and right @xmath47 ( upper set ) branches of the map shown in fig .
[ fig03](c ) .
one can see that for @xmath48 sufficiently small , @xmath49 is characterized by the power law @xmath50 where @xmath51 .
this implies an extreme divergence of trajectories which is also reflected in the phase portrait . the initial conditions @xmath52 on @xmath31 for trajectories 1 and 2 ( solid lines in fig .
[ fig03](a , b ) ) were such that @xmath53 where @xmath54 .
[ fig03 ] shows , the separation between these two trajectories on their return to the surface of section grows by @xmath55 orders of magnitude .
in fact , trajectories 3 and 4 constitute another pair with @xmath56 . the existence of the superunstable trajectory results in the observed dichotomy of trajectories belonging to the slow manifold . a 2d manifold intersecting the slow manifold along trajectory 0 plays similar separating role in the 3d phase space .
although the global structure of this repulsive flow in the phase space is very complicated , the flow can be easily determined locally by the property that it separates trajectories making large and small loops while approaching the attractor .
figure [ fig05](a ) shows an expanded view of the flat part of the first return map @xmath41 in fig .
[ fig03](c ) .
the slope of the map is everywhere positive and , thus , the map is invertible . as @xmath9 decreases at constant @xmath25 , a smooth transformation of the slow manifold takes place , resulting in the loss of invertability of the map . a fragment of map constructed for @xmath57 ( see fig .
[ fig05](b ) ) demonstrates that @xmath41 is not monotonic and a smooth maximum arises near the stable limit cycle . at @xmath58
the map is tangent to the bisectrix and a new pair of stable and unstable limit cycles emerges on the slow manifold ( cf .
[ fig05](c ) ) .
as the map s maximum grows it sharpens and the newborn limit cycle looses its stability through a cascade of period - doubling bifurcations . at @xmath59 ( see fig . [ fig05](d ) ) a chaotic attractor coexists with the period-1 limit cycle and focus .
although the entire cascade to chaos occupies very narrow span of @xmath9 , two first period - doubling bifurcations were resolved .
the tristability of the system breaks down at a certain value of @xmath9 and every trajectory initiated near the maximum of the map is attracted to the period-1 limit cycle after a chaotic transient of varying length .
figure [ fig05](e ) shows a part of the first return map for @xmath60 where oscillations suddenly crash on the focus .
two important characteristic features of the flow can be inferred from the map .
first , the maximum of the map acquires a cuspoid shape signalling of the formation of a highly compressed fold on the slow manifold .
second , from the tangency of the map and bisectrix , one can see that the stable period-1 limit cycle and the large - amplitude saddle cycle are near the point of coalescence .
the annihilation of this pair in a saddle - node bifurcation gives rise to an intermittent chaotic attractor which has the appearance of a thick limit cycle .
a decrease in the control parameter @xmath9 leads to the crash of chaotic oscillations onto the focus . summarising the information presented in fig .
[ fig05](a - e ) , one can conclude that as @xmath9 changes through the interval @xmath61 $ ] , the slow manifold transforms from a planar surface into a folded fractal structure with infinitely many leaves ( smale s horseshoe ) . making use of several auxiliary poincar sections it is possible to construct a 3d picture of the slow manifold .
figure [ fig06 ] schematically represents the manifold and three of its poincar sections at parameters close to the crash bifurcation .
part of the manifold situated between sections @xmath62 and @xmath63 is removed to facilitate visualization of its geometry . in section @xmath63
one sees a triply - folded curve which corresponds to the first structural level of the manifold .
note the difference between the two folds of this curve : while the outer fold has a smooth parabolic shape , the inner fold is a sharp cusp . increasing the resolution considerably ( by @xmath64 5 orders of magnitude )
one would be able to find that the curve is , in fact , a densely compressed fold constituted by seven leaves of the second structural level , and so on . to describe one iteration of the manifold stretching and folding we do not need to consider these higher structural levels . following the evolution of points in section @xmath63 under the flow ( clockwise in fig . [ fig06 ] ) one encounters a pleat where the manifold as a whole begins to fold into an @xmath1-shaped form ( section @xmath65 ) with two approximately equal parabolic folds .
prominent stretching of the flow in the plane of the manifold , accompanied by strong contraction in the orthogonal direction , transforms the manifold into the thin roll composed of seven leaves seen in section @xmath62 . as a result of the extreme inhomogeneity of compression the inner fold acquires sharp cuspoid shape while the outer one remains smooth and parabolic . as the flow returns to section @xmath63 more contraction along the vertical direction occurs and the second order structure shown in section @xmath62 is not discernible .
in sec . [ shoe ] we described two important characteristic features of the phase flow of system ( [ sys ] ) .
firstly , at the parameter values in the neighborhood of mixed - mode oscillations , the slow manifold is folded into a horseshoe - type set with a fractal , multi - leaved structure .
secondly , the repulsive flow divides the transient trajectories into two distinct classes .
intersection of the repulsive flow and the slow manifold yields a superunstable trajectory with peculiar dynamical properties . in the present section we show how mixed - mode states which correspond to periodic orbits embedded in the horseshoe emerge and bifurcate into each other . consider a path in the @xmath13 parameter plane which starts in domain 2 or @xmath27 and ends in the region of mixed - mode oscillations 3 ( see fig .
[ fig01 ] ) . as the domain of large - amplitude period-1 oscillations
is traversed towards the curve corresponding to the @xmath66 transition , the modifications of the slow manifold described in sec .
[ shoe2 ] take place . before the first mixed - mode is formed an important transformation of the manifold occurs at the point where a cusp - shaped fold developes .
inspection of fig .
[ fig06 ] shows that the manifold can not support periodic oscillations with small - amplitude loops since its folded part lies on the `` large - amplitude '' side of the repulsive flow ( only the superunstable trajectory is shown in fig .
[ fig06 ] ) . in order for orbits with both small- and large - amplitude excursions
to exist , the repulsive flow has to intersect the folded part of the horseshoe .
figure [ fig07 ] shows segments of the manifolds cut out by two orthogonal planes for three consequtive values of @xmath9 and fixed @xmath10 .
the case @xmath67 ( fig .
[ fig07](a ) ) corresponds to the generic transversal position of the manifolds in which the growing fold of the horseshoe lies on one side of the repulsive flow .
this case is described by a first return map of the type shown in fig .
[ fig05](d ) . at @xmath68 , in addition to the transversal intersection , a tangency between the two manifolds occurs .
the tangency manifests itself by the formation of a sharp cusp on the first return map ( see fig . [ fig05](e ) ) . for @xmath69
the fold of the horseshoe penetrates the surface corresponding to the repulsive flow and drastically increases in size due to the strong divergence from the repulsive flow .
one also sees in fig .
[ fig07](c ) that as a result of this transformation two new superunstable orbits ( thick solid lines ) are born on the slow manifold . taking into the consideration fractal nature of the horseshoe one could imagine an infinite stack of superunstable orbits born in pairs when nested folds of the horseshoe penetrate the surface of the repulsive flow one by one .
the birth of new superunstable orbits is reflected in the first return map as well .
figure [ fig08 ] shows the first return map constructed with @xmath31 as a surface of section as described in sec .
[ shoe1 ] for parameter values @xmath70 at which the transition @xmath71 occurs .
as @xmath72 is passed , the cusp on the map breaks up and the maximum acquires normal parabolic shape . at the same time the height of the maximum increases tremendously and two segments with infinite slope appear on its sides .
the analysis of the map shows that the transition @xmath73 occurs simultaneously with annihilation of the large - amplitude stable and saddle - type limit cycles . at the point of coalescence
they produce structurally unstable limit cycle which is seen on the first return map as a point of tangency with bisectrix .
this cycle combines properties of both stable and unstable limit cycles .
consider a trajectory which departs from such cycle at @xmath74 .
in course of evolution it inevitably executes one small - amplitude loop around the focus . upon the completion of this loop
the trajectory is reinjected into the stable part of the neighborhood of the cycle which is eventually approached at @xmath75 .
thus , first mixed - mode state @xmath76 is characterized by the orbit homoclinic to the structurally unstable cycle and , therefore , exists in a parameter domain of zero measure . for @xmath77 a finite number of large - amplitude loops is executed before a trajectory visits the small loop and successions of @xmath78 mixed - mode states with rapidly decreasing @xmath79 are oserved .
it is possible to establish how a particular mixed - mode orbit is incorporated in the slow manifold .
figure [ fig09 ] shows the @xmath24 projection of a @xmath80 orbit and its embedding in the horseshoe . as in fig .
[ fig06 ] , the poincar sections of the manifold are shown at different levels of resolution of the fine structure . while section @xmath63 presents the first two structural levels , in section @xmath65 only first level is shown in full .
two lines of phase points parallel to the surface of the repulsive flow and separated from it by @xmath81 are selected in section @xmath63 ( long dashes ) .
the evolution of these points shown in section @xmath65 demonstrates the extreme divergence of trajectories caused by the presence of the repulsive flow .
orbit @xmath82 consists of a dense , flat band of large - amplitude loops which consecutively shrink in size and a small loop where the trajectory segment is reinjected into the large - amplitude band .
projecting points in which the orbit intersects planes @xmath63 and @xmath65 ( shown by thick dots ) on the corresponding poincar sections of the horseshoe one can obtain insight into the mechanism of this reinjection . at a certain time the trajectory appears to the left of the repulsive flow surface ( point 1 in @xmath63 ) .
this causes it to execute a small loop ( point 2 on the cusp - shaped tongue in @xmath65 ) .
next return of the trajectory into section @xmath63 occurs in point 3 , separated far from 1 .
following the subsequent numbers one sees how the orbit returns to the large - amplitude band , gradually ascending leaves of the slow manifold .
the same mechanism holds for all @xmath83 states , although escape on the small loop occurs from different leaves .
the analysis of the reinjection scheme shows that the horseshoe ( solid lines in fig .
[ fig09 ] ) can not support mixed - mode oscillations with more than one small - amplitude loop .
indeed , regardless of how far to the left one shifts the position of point 1 , first return of a trajectory initiated at 1 into the surface of section ( point 3 ) lies to the right of the repulsive flow .
this situation changes as the cusp - shaped tongue , to which point 3 belongs , penetrates the surface of the repulsive flow through the formation of a tangency as described in sec .
( corresponding changes on both sections are shown by medium - dashed lines . )
as soon as a similar tongue of a higher structural level passes through the surface of the repulsive flow , an opportunity for the existence of the third small - amplitude loop arises , and so on .
thus , changing position of the horseshoe relative to the repulsive flow one can obtain mixed - mode states @xmath23 with any desired @xmath0 and @xmath1 . at small values of @xmath10 , where the bulk of the mixed - mode domain lies , the surface - of - section @xmath31 introduced in sec .
[ shoe1 ] becomes unsuitable for the construction of the first return map due to the strongly bent shape of the corresponding poincar section ( see section @xmath65 in fig .
[ fig09 ] ) .
instead , one can use sections @xmath84 const similar to section @xmath63 in fig .
[ fig09 ] .
figure [ fig10 ] presents two examples of the first return map @xmath85 constructed in the approximation based on the upper leaf of the horseshoe for parameter values corresponding to @xmath86 ( a ) and @xmath87 ( b ) mixed - mode states . as one can see , maps of this type share a number of common features .
the map consists of two branches with positive slope and an extremely steep , negatively - inclined segment which joins the branches . at the point corresponding to the superunstable orbit the map has an infinite slope and
can be locally described by an exponent as discussed in sec .
[ shoe1 ] .
any mixed - mode state @xmath23 may be represented on this map by a periodic trajectory with @xmath0 iterates on the right branch and @xmath1 iterates on the left branch .
this property of the map makes it particularly suitable for the illustration of mixed - mode bifurcations . at parameter values corresponding to the transition @xmath88
the right branch is tangent to the bisectrix . as
an intermittent channel opens , the trajectory funnels through it and then jumps onto the left branch from which it is reinjected onto the branch of large - amplitude loops and the cycle closes .
as @xmath9 decreases both branches fall relative to the bisectrix , reflecting the gradual growth of the number of small - amplitude loops accompanied by a reduction in the number of large - amplitude loops . in ref .
@xcite ringland _ et .
al . _ extensively studied the properties of the two - extremum map @xmath89 where @xmath90 at small values of the parameter @xmath91 the map has a smooth shape similar to that of the cubic map . as @xmath91 increases , the slope of its middle , negatively - inclined segment grows steeper while the @xmath92 distance between two map s extrema vanishes . as @xmath91 tends to infinity
, the z - map acquires zigzag shape with a vertical middle segment .
_ have shown that in the limiting case @xmath93 the attractors of the z - map form farey sequences as other map parameters are varied .
they related this property to the existence of a vertical segment since a decrease in @xmath91 leads to a gradual transformation of farey sequences into u - sequences .
although the dynamic origin of such maps was not discussed , a possible relation of the z - map to the existence of mixed - mode oscillations was inferred .
one can easily see that the z - map in its @xmath93 limit is qualitatively similar to the map @xmath94 constructed in the present section . in our map the middle segment is always vertical due to the presence of the superunstable orbit . applying results of ringland _ et .
to system ( 1 ) , we may assume that the repulsive flow defines not only the shape of mixed - mode orbits with their nonoverlapping bands of small- and large - amplitude loops , but also their property to form farey sequences .
we have analyzed in detail a model exhibiting mixed - mode oscillations . by constructing the model s slow manifold
we have shown that mixed - modes correspond to periodic orbits embedded in a horseshoe - type strange set .
this explains why chaotic oscillations are observed in transitions between adjacent periodic states .
the organization of the slow manifold into a horseshoe also accounts for the signatures of chaos in transient trajectories when the attractor is periodic . in our model the mixed - mode periodic orbits
do not lie on a 2-torus . following transformations of the slow manifold from its simple planar organization into a horseshoe
, we have also shown that a torus does not exist as an intermediate state between period-1 and mixed - mode oscillations .
thus , our scenario for the emergence of mixed - mode oscillations is an alternative to quasiperiodicity .
the main distinctive feature of mixed - mode states , the partition of their orbits into nonoverlapping bands of small- and large - amplitude loops , finds its explanation in the existence of the repulsive flow .
intersection of this flow with the slow manifold yields a complex system of superunstable orbits .
the position of the slow manifold relative to the repulsive flow determines whether the formation of small or large loops of the attractor is favored . as parameters change along a path running through the mixed - mode domain , the position of the repulsive flow changes and one observes sequences of mixed - mode states @xmath95 with monotonously increasing ( decreasing ) @xmath96 ratio . as the results of ringland _ et.al .
_ @xcite suggest , the fact that these sequences are described by farey arithmetic is also a consequence of the presence of the superunstable orbits .
thus , the existence of the repulsive flow places this horseshoe into a separate subclass of strange sets whose periodic orbits are given by @xmath23 mixed - mode states , and whose windows of periodicity are organized into farey sequences .
there exist indications that our scenario may account for the mixed - mode oscillations observed in a number of model and experimental studies .
this appears to be so for all those cases in which the mixed - mode periodic states were found to be separated by chaotic rather then quasiperiodic oscillations .
as ringland _ et.al .
_ @xcite pointed out , zigzag - shaped maps with a vertical middle segment indicative of a superunstable trajectory were found in many chemical and electrochemical systems .
this work was supported in part by a grant from the natural sciences and engineering research council of canada .
v. petrov , s. k. scott , and k. showalter , j. chem . phys . * 97 * , 6191 ( 1992 ) ; f. tracqui , acta biotheor .
* 42 * , 147 ( 1994 ) ; for review on mixed - mode oscillations in chemistry see i. r. epstein and k. showalter , j. phys .
chem . * 100 * , 13132 ( 1996 ) .
m. t. m. koper and p. gaspard , j. phys .
chem . * 95 * , 4945 ( 1991 ) ; m. t. m. koper and p. gaspard , j. chem .
96 * , 7797 ( 1992 ) ; m. t. m. koper , p. gaspard , and j. h. sluyters , j. chem . phys . * 97 * , 8250 ( 1992 ) .
polynomial piece - wise interpolation of the third to fifth order was used in our study . for the particular case considered in this section accurate approximation for the entire poincar section
is given by the polynomial @xmath97 . | a detailed study of the slow manifold of a model exhibiting mixed - mode oscillations is presented . a scenario for the emergence of mixed - mode states which does not involve phase locking on a 2-torus is constructed .
we show that mixed - modes correspond to the periodic orbits embedded in the horseshoe - type strange set and demonstrate how transformations of this set determine the transitions of mixed - mode states into each other . 2 |
outflows and jets are ubiquitous in nature . they are observed both in local universe , mostly in black hole ( bh ) x - ray binaries ( bhxrbs ) which are believed to harbour stellar mass bhs called microquasars [ 1 ] , as well as in powerful extragalactic radio sources [ 2 ] where well - collimated outflows or jets emerge continuously from the nuclear region of the host active galaxies ( agns ) or quasars harbouring supermassive bhs . the accreting hot plasma around bhs powered by extreme gravity of the central object results in the formation of outflow / jet which extracts mass , angular momentum and energy from the inner regions of the accretion flow . the outflows in microquasars are observed only in low - hard state of bhxrbs [ 3,4 ] which are radiatively inefficient .
radiatively inefficient accretion flows ( riafs ) are hot gas pressure dominated systems which are geometrically thick @xmath6 and optically thin , where @xmath7 is the scale - height of the accretion region .
the dynamics of the flow is thus strongly sub - keplerian and advection dominated [ 5,6 ] .
riafs occur when the mass accretion rate @xmath3 is very low ( presumably with @xmath8 ) , where @xmath9 is the eddington accretion rate or the accretion rate corresponding to the eddington luminosity .
outflows / jets are not likely observed in high - soft state of bhxrbs ( [ 4 ] and references therein ) , which are believed to be powered by geometrically thin and optically thick radiation pressure dominated keplerian accretion disk [ 7 ] . theoretically speaking ,
it has been argued [ 5,8,9,10 ] that a geometrically thick advective accretion flow has a strong tendency to drive bipolar outflows due to high thermal energy content of the hot gas .
they may be additionally propitious to propel outflows / jets because its vertical thick structure enhances the large - scale poloidal component of the magnetic field , which plays a critical role in launching strong and collimated outflows [ 11 ] .
apart from low - hard state of bhxrbs which power jets , at the other end of the spectrum , strong outflows and jets are observed in low excitation radio galaxies ( lergs ) harbouring supermassive bhs .
lergs , a more generalization of low luminous agns ( llagns ) seem to be accreting gaseous plasma directly from the hot x - ray emitting phase of interstellar medium ( ism ) or from the hot x - ray halos surrounding the galaxy or from the hot phase of the intergalactic medium ( igm ) quasi - spherically in a radiatively inefficient mode with near bondi rate [ 12 ] .
lergs thus resemble low - hard state of brxrbs having geometrically thick and optically thin gas pressure dominated strongly advective quasi - spherical accretion flow , accreting hot gas at a high sub - eddington accretion rate ( presumably with @xmath8 ) .
this strongly advective radiatively inefficient accretion paradigm ( riaf ) or hot mode accretion having considerable geometrical thickness , is more prone to emanate outflows / jets and is very conducive to propel matter vertically outwards out of the accreting region .
however , with the increase in @xmath3 as @xmath10 , the flow tends to be more centrifugally dominated and become moderately advective , with the central bh accreting relatively cold gas as compared to the hot mode accretion .
incidentally the moderately advective accretion flow does not occur in a radiatively inefficient mode , but where considerable amount of both gas and radiation pressure seem to be present in the system , rendering the flow to have a moderate optical depth .
geometrically , the inner advective region would then be relatively thinner than that correspond to riafs
. the moderately advective accretion paradigm may also be susceptible to eject outflows .
the difference between this paradigm and with the riaf , however , may lie in the acceleration mechanisms to eject bipolar outflows and jets , which we would eventually investigate in this study . nonetheless , with the increase in @xmath3
as the flow tends to become more rotationally / centrifugally dominated with lesser geometrical thickness , the efficacy of the disk to eject outflows diminishes . beyond @xmath11 , the flow would eventually tend towards keplerian nature ( [ 13 ] ; paper ( in preparation ) ) .
geometrically thin keplerian accretion disk plausibly fails to account for the launching and acceleration of outflows and jets ( [ 8,9,10 ] ; also see [ 14 ] for further discussion ) .
extensive work has been pursued on the origin of outflow / jet , since the seminal work of blandford & payne [ 15 ] in studying accretion powered hydromagnetic outflows , which we like to focus upon in this study .
physical understanding of the accretion powered hydromagnetic outflows have either been performed in stationary self-similar approximation [ 15,16,17 ] in quasi - analytical regime to demonstrate the importance of poloidal component of the magnetic field to launch outflowing matter from the keplerian accretion disk , or through magnetohydrodynamic ( mhd ) simulations in both nonrelativistic as well as in relativistic regimes ( for details , see the introduction in [ 8,10 ] and references therein , also see [ 18 ] and references therein ) .
in most of these studies the authors remain focused mainly on the launching of outflows / jets from the geometrically thin keplerian disk .
the formation of the accretion powered outflow and jet is directly related to the efficacy of extraction of angular momentum and energy from the magnetized accretion flow .
however , the exact mechanism by which the radial accretion is diverted into strong outflows and plausible jets still remains theoretically elusive .
notwithstanding , jet launching is completely a magnetohydrodynamic process . accreting material diffuses across magnetic field lines threading the accretion region ,
is then lifted upwards by mhd forces which then couples to the field and becomes accelerated magnetocentrifugally . however , if the accreting system is strongly gas pressure dominated , it may happen that the gas pressure gradient would play a more contributory role to lift the plasma vertically outwards along with the help of magnetic forces .
in addition , turbulent reconnection of magnetic field lines may lead to flux annihilation [ 19 ] . magnetic energy dissipates through turbulent magnetic reconnection , may also power the outflow / jet emission ( [ 20 ] and references therein ) .
most of the studies of accretion disk and outflow / jet have evolved separately , assuming these two to be apparently dissimilar objects . in the light of both deeper theoretical understanding and observational inferences [ 21,22,3 ]
it is evident that the dynamics of outflow and the underlying accretion are strongly correlated ( for details see [ 8,9,10 ] and references therein ) .
outflows and jets observed in agns and xrbs can only originate in an accretion powered system , where the accreting plasma around gravitating objects like bhs acts as a source , whereas outflow and then jet acts as one of the possible sinks [ 10 ] .
the implicit coupling between accretion and outflow is then essentially governed by conservation laws ; conservation of matter , momentum and energy .
the outflowing matter carries away mass , angular momentum and energy extracted from the accreting plasma [ 16 ] .
we do not intend to investigate the physics of jet formation and its launching mechanism which is altogether a different field of research , but would like to focus entirely on the inter - correlating dynamics of the accretion and outflow within the coupled accretion - outflow region , and the conditions / criteria for jet launching .
any proper understanding of the dynamics and the conditions of jet launching should necessarily require the robust understanding of the dynamics of the magnetized advective accretion region coupled to outflow , governed explicitly by the conservation laws .
the relevant dynamical solutions at the accretion - outflow coupled surface ( the surface from where outflow decouples from radially inward accretion flow ) would then act as boundary conditions at the base of the jet .
accretion disk - outflow / jet coupling has been studied on number of occasions from purely observational angle , in accretion powered systems [ 22,23 ] . from theoretical perspective
, few self - similar studies have been attempted in the context to accretion - outflow coupling , both in non magnetized and in magnetized regime ( see [ 18 ] and references therein ; [ 8,9,10,24,25,26 ] ) .
notwithstanding , theoretically , it is still difficult to construct reasonably satisfactory and definitive model of magnetized accretion - outflow / jet coupled region , owing to the complicated geometry and inconclusive understanding of the inflow - outflow coupled region . on the other hand , few simulations on disk - outflow / jet coupling
have also been cultivated [ 27,28,29 ] . in these simulations
how the matter gets deflected from the equatorial plane has been studied largely in the keplerian regime .
casse and keppens [ 30 ] performed an advective , resistive mhd simulation of accretion - ejection structure with the inclusion of the energy equation , however neglecting the viscosity and radiative loss from the system .
nonetheless , it is still difficult to simultaneously simulate the accretion and the outflow regions because the time scales of the accretion and outflow are in general very different . in the present work ,
we endeavour to develop a robust viscous , resistive and advective mhd accretion - induced outflow model in the 2.5 dimensional regime , in the mean field mhd approximation , in the context to accretion powered hydromagnetic outflows / jets , focusing entirely on the inter - correlating dynamics of the accretion and outflow within the coupled accretion - outflow region , without aspiring to explore the mechanism of launching and ejection of outflows and jets .
a complete 2.5-dimensional viscous , resistive , advective global mhd numerical solution of such a system is left for further work , and which is beyond the scope at present ; we confine our treatment to quasi - analytical / quasi - numerical power law self - similarity ( e.g. , [ 5 ] ) in a quasi - stationary configuration , by upholding the conservation equations .
all physical quantities are scaled as powers in @xmath12 and @xmath13 according to their dimensions , in the limit of higher order polynomial expansion .
we will perform a critical analysis of accretion - outflow region and provide a complete quasi - analytical family of solutions .
although the quasi - analytical self - similar solutions are approximate , they , however , can provide a strong physical intuitive picture of accretion dynamics coupled with the outflow , as well as physical criteria / conditions to eject outflows / jets . in the next section
, we present the formulation of our model .
describes the quasi - analytical / quasi - numerical procedure to solve the model equations of the accretion - induced outflow . in , we evaluate the coefficients of our self - similar solutions and analyze them . in , we investigate the nature and behaviour of the family of solutions for accretion - induced outflow within the bounded accretion coupled outflow region .
finally , we end up in with a summary and discussion .
we formulate the accretion - outflow coupled model by considering a 2.5-dimensional viscous , resistive , advective accretion flow geometry in the mean field mhd regime as strong outflows / jets are more likely to eject from a geometrically thick , advective region of the accretion flow . the vertical flow is explicitly included in the system .
we adopt the cylindrical coordinate system ( @xmath14 ) to describe a quasi - stationary , mean axisymmetric accretion flow . as we have incorporated outflow in our system , within the accretion - outflow coupled region in advective paradigm , all the dynamical flow variables ; namely , radial velocity ( @xmath15 ) , azimuthal velocity ( @xmath16 ) , specific angular momentum ( @xmath17 ) , vertical velocity or outflow velocity ( @xmath18 ) , isothermal sound speed ( @xmath19 ) , mass density ( @xmath20 ) , thermal pressure ( @xmath21 ) and magnetic field components ( @xmath22 ) vary in both @xmath12 and @xmath13 .
the dynamical equations are vertically integrated over an arbitrary scale height @xmath7 from @xmath23 to @xmath24 . here
@xmath7 is not a hydrostatic disk - scale height but a photospheric surface within which accretion and outflow are coupled . above @xmath7 , the outflow decouples from the accretion flow ,
gets further accelerated in the hot nonthermal magnetized corona and finally forms a relativistic jet .
we focus on this accretion - outflow coupled region within which the flow is mostly bounded .
we only include the @xmath25 component of turbulent stress , which is responsible for radial transport of angular momentum outwards ( angular momentum get transported due to the diffusion of turbulent eddies ) .
vertical transport of angular momentum occurs mainly through large - scale magnetic stresses , where the outflowing matter magnetically extracts or removes angular momentum .
the accreting mass is assumed to be much less as compared to that of the central object , and hence the flow is not self - gravitating . as the accretion flow is inherently turbulent ,
we express all the dynamical variables in mean and fluctuating part .
microscopic viscosity and resistivity are neglected compared to the large - scale diffusion of turbulent eddies .
assuming that the fluctuation in the density is very less ; in the limit of boussinesq approximation , the generic ansatz follows as @xmath26 where @xmath27 represents the mean flow which is either a time average or an ensemble average .
@xmath28 is then the fluctuation corresponding to that variable . the turbulence is defined in terms of mean reynolds and maxwell stress described through correlations given by @xmath29 \ , , \label{2}\end{aligned}\ ] ] where , @xmath30 is the net turbulent stress .
statistical averaging of mhd equations generates large number of turbulent correlation terms . in the present study
, we restrict to first order turbulent correlation and neglect second order and higher order correlation terms .
the reynolds and maxwell stresses are commonly parameterized in terms of kinetic and magnetic turbulent viscosities @xmath31 and @xmath32 respectively as @xmath33 @xmath34 where @xmath35 is the strain tensor .
the turbulent viscosities are parameterized through an @xmath5 prescription as @xmath36 where @xmath37 is the net turbulent viscosity and @xmath38 is the net turbulent viscosity parameter . with these parameterizations ,
the coupled accretion - outflow dynamical equations in quasi - stationary state are as follows : \(a ) mass transfer : @xmath39 we define the net mass flow rate which is a constant , through an integro - differential equation as @xmath40 r d\varphi dr dz = - { \dot m } \ , , \label{7}\end{aligned}\ ] ] where the first term is the signature of the radial accretion flow and the second term attributes to outflow . if we discard @xmath41 ( neglecting outflow ) , eqn .
( [ 7 ] ) reduces to a height integrated continuity equation of the accretion flow , and where @xmath42 would then be the usual mass accretion rate .
\(b ) momentum transfer : the momentum balance equation in the mean field mhd is given by @xmath43 using eqn .
( [ 6 ] ) and integrating eqn .
( 8) vertically , the radial momentum balance equation is given by @xmath44 \ , dz \
, = \ , 0 \ , .
\label{9}\end{aligned}\ ] ] in deriving this we have used divergence criteria of magnetic field .
@xmath45 is the radial component of the gravitational force . in a similar fashion
, we write the azimuthal momentum balance equation as @xmath46 the last term in the right hand side of eqn .
( [ 10 ] ) is the magnetic torque that acts at the accretion - outflow surface , and which helps in transporting the angular momentum vertically outwards .
this term is responsible for the mass loss in the wind .
next we derive the vertical momentum balance equation which is obtained from the @xmath13 component of eqn .
( [ 8 ] ) .
@xmath47 \ , dz \ , = \ , 0 \ , , \label{11}\end{aligned}\ ] ] where @xmath48 is the vertical component of the gravitational force . equation ( [ 11 ] ) is integrated from @xmath49 to @xmath50 due to the reflection symmetry of all the dynamical variables across the accretion - outflow coupled surface .
equation ( [ 11 ] ) contains the information of the outflow dynamics within the accretion - outflow region .
the matter starts to accelerate vertically outwards from just above the equatorial plane of the accretion region . if there is no outflow , then @xmath51 , and if we neglect the magnetic pressure and magnetic stresses , eqn .
( [ 11 ] ) reduces to the well known hydrostatic equilibrium condition in the disk , and the usual hydrostatic disk scale - height can be obtained .
\(c ) divergence condition : @xmath52 \,dz \ , = \ , 0 \ , .
\label{12}\end{aligned}\ ] ] the divergence condition determines the symmetry property of magnetic field components .
whether the radial and the vertical component of the magnetic field will follow odd and even symmetry or vice - versa in the @xmath13 direction , can be ascertained from the above equation .
\(d ) magnetic induction : the turbulent magnetic induction equation is derived from the mean field mhd theory ( e.g. , [ 31 ] ) . following the usual procedure and neglecting the dynamo effect , the steady state induction equation in the tensorial form is given by @xmath53 where @xmath54 is the kinetic part of the turbulent viscosity .
the above equation has been written in the most general form considering an anisotropic turbulence .
splitting eqn .
( [ 13 ] ) in radial and azimuthal direction we obtain after vertical integration @xmath55 \ , dz \ , = \ , 0 \ ,
. \label{14}\end{aligned}\ ] ] and @xmath56 \ , dz \ , = \ , 0 \ , .
\label{15}\end{aligned}\ ] ] we do not show here the vertical component of the induction equation as it is similar to that of the radial equation , and contains the same information regarding the magnetic dynamics of the flow .
note that the turbulent diffusion in the induction equation arises only from the kinetic part of the turbulent stress tensor through the reynolds stress .
this is attributed to the mean field approximation , where we split the quantities in mean and turbulent part .
\(e ) energy conservation : the poynting flux @xmath57 is given by @xmath58 using mean field mhd and the induction equation , and by discarding the microscopic resistivity , we write @xmath59 where @xmath60 is the current density .
in contrast to the other correlation terms , the last term in the right hand side of eqn .
( 17 ) is not a first order correlation term , rather a second order or higher order correlation term [ 32 ] ; we omit this term in our present study . neglecting the kinetic dynamo effect , using eqn .
( 17 ) , the steady state energy conservation equation of a turbulent magnetic fluid in a tensorial form using mean field mhd is then given by @xmath61 \ , + \frac{\bar v_j}{c } \epsilon_{jlm } { \bar b_l } { \bar j_m } \ , + \frac{\bar v_j}{c } \epsilon_{jlm } \overline{b^{\prime}_l j^{\prime}_m } - \ , \frac{1}{4 \pi } \epsilon_{jlm } \nu^{\mathcal r}_{lk } \frac{\partial { \bar b_m}}{\partial x_k } \epsilon_{jlm } \frac{\partial { \bar b_m}}{\partial x_l } \ , \
, + \frac{\partial f^{r}_j}{\partial x_j } \ ,
= \ , 0 \ , , \label{18}\end{aligned}\ ] ] where @xmath62 , the ratio of gas pressure to the total pressure in the accretion - outflow coupled region is assumed to be constant .
@xmath63 is the radiation pressure in the gas - radiation mixture .
the radiation field has been assumed to be locally isotropic .
the ratio of specific heat of the fully ionized gas and radiation are taken as @xmath64 and @xmath65 respectively .
the ` effective ratio of specific heat ' @xmath66 is related to @xmath67 through @xmath68 . in deriving eqn .
( [ 18 ] ) , we have neglected the turbulent thermal conductivity .
the last term in the above equation represents the transport of radiative flux . multiplying eqn .
( [ 8 ] ) by @xmath69 and using the continuity equation , the vertically integrated energy budget for accretion - induced outflow is thus obtained below @xmath70 dz \
, + \int^{+h}_{-h } \big [ \frac{3}{2 } ( 2 - \beta){\bar \rho } \ , { \bar v_z } \frac{\partial { \bar c}^2_s}{\partial z } \ , - { \bar v_z } { \bar c}^2_s \frac{\partial { \bar \rho}}{\partial z } \big ] \ , dz \ , \nonumber \\ =
\ , \int^{+h}_{-h } \big [ { \bar \rho } \ , \nu^{\mathcal r}_{r\varphi } \ , \bar{s}^{2}_{r\varphi } \ , + \frac{1}{4 \pi } \epsilon_{jlm } \nu^{\mathcal r}_{lk } \frac{\partial { \bar b_m}}{\partial x_k } \epsilon_{jlm } \frac{\partial { \bar b_m}}{\partial x_l } \big ] \ , dz - 2 f^{r+ } \ , .
\label{19}\end{aligned}\ ] ] left hand side of the above equation is the signature of advection of energy flux in radial and vertical direction due to accretion and outflow respectively .
first and the second term of the right hand side expresses the turbulent viscous heating due to @xmath71 component of the stress tensor and the turbulent ohmic dissipation or joule s heat loss respectively in the accretion flow .
turbulent ohmic dissipation symbolizes a resistive flow due to which constant annihilation of the magnetic flux occurs .
the last term of the equation is the flux of radiation that is escaping from the accretion - outflow surface .
similar to that of the induction eqn .
( [ 13 ] ) , the turbulent diffusion in the energy equation also arises only through the reynolds stress tensor the net heat flux generated in the accretion flow is defined by @xmath72 defining @xmath73 as a constant cooling factor scaled through a relation @xmath74 the final form of the heat flux equation is obtained by coupling eqns .
( [ 19 ] ) , ( [ 20 ] ) and ( [ 21 ] ) .
a constant @xmath73 implies that the radiative heat loss from the accretion surface has a linear proportionality with the net heat flux generated in the system .
although @xmath73 should vary radially , however , within a small inner accretion region where the accretion and outflow are coupled , this presumption is an acceptable approximation .
the parameter @xmath73 reflects the extent to which the accretion flow is advective .
@xmath75 represents an accretion flow with efficient cooling or is radiatively efficient . at the other extremity , @xmath76
epitomize a strongly advective or advection dominated system which are radiatively inefficient .
equations ( [ 7 ] ) , ( [ 9 ] ) , ( [ 10 ] ) , ( [ 11 ] ) , ( [ 12 ] ) , ( [ 14 ] ) , ( [ 15 ] ) and ( [ 19 ] ) exhibit the dynamical behavior of accretion - induced outflow in a resistive mhd paradigm in mean field approximation . in the next section ,
we explore a technique to solve these eight coupled partial differential equations using eqns .
( [ 5 ] ) , ( [ 20 ] ) and ( [ 21 ] ) , rigorously .
the equations describing the turbulent magneto - fluid in the last section are extremely complicated and it is beyond our scope to solve them numerically . the only procedure left in our hand is to explore some kind of approximate analytical method or quasi - analytical method to solve them .
we follow a power law self - similar approach ( see also ny94 ) with a polynomial expansion to solve the equations to obtain the class of solutions . for the present purpose
we invoke a generalized n@xmath77 degree polynomial expansion of all the dynamical variables , where the flow variables are functions of both radial and vertical coordinate .
we restrict ourselves in newtonian paradigm as power law self - similar solutions are valid only in the limit of newtonian approximation .
the generalized newtonian potential at any ( @xmath78 ) in the accretion - outflow region is written in the form of power series @xmath79 where @xmath80 is the mass of the central object .
the gravitational force in the radial and vertical direction is then written in the form of a polynomial expansion as shown below .
@xmath81 @xmath82 @xmath45 and @xmath48 have even and odd symmetry in @xmath13 direction , respectively .
we seek a polynomial expansion in the similar form for all dynamical variables , where the flow of matter in the accretion region is being considered to have reflection symmetry about the equatorial plane .
all the hydrodynamical variables will have even symmetry except those directly related to @xmath48 .
the outflow velocity @xmath18 will then have a odd symmetry in @xmath13 direction .
consequently , the radial component of the magnetic field @xmath83 and the vertical component of the magnetic field @xmath84 will have even and odd symmetry configuration , respectively , which is required from the divergence condition of the magnetic field . odd symmetry configuration of magnetic field has been used previously on other occasions in context to outflows / jets from the accretion flows ( e.g. , [ 33,26 ] ) .
the flow velocities , angular momentum and the density are then written in the following polynomial form @xmath85 similarly , the components of the magnetic field can be expanded as @xmath86 where , @xmath87 are the dimensionless coefficients which will be evaluated from the mhd conservation equations .
we determine the exponents @xmath88 by self comparison of various terms in the model equations . substituting the solutions from eqns .
( [ 25 ] ) and ( [ 26 ] ) in the mhd conservation equations and comparing the exponents of @xmath12 and @xmath13 , we obtain @xmath89 . using the above value of exponents and using the polynomials in eqns .
( [ 25 ] ) and ( [ 26 ] ) , eqns .
( [ 9 ] ) , ( [ 10 ] ) , ( [ 11 ] ) , ( [ 12 ] ) , ( [ 14 ] ) , ( [ 15 ] ) and ( [ 19 ] ) can be expanded in the power of aspect ratio @xmath90 . using a simple technique given by the form @xmath91
the above polynomial equations then can be written in a generic form as @xmath92 where , @xmath93 are zeroth order , first order , second order and higher order coefficients which are nonlinear functions of @xmath94 corresponding to @xmath95 , respectively . equation ( [ 28 ] ) is a linear combination in powers of @xmath96 which are linearly independent .
if we neglect all the terms of the order @xmath97 , and only keep the zeroth order term , the generic eqn .
( [ 28 ] ) after neglecting the magnetic field contribution will reduce to algebraic equations in ny94 . to exemplify , in the appendix , we have shown the polynomial expansion of the integro - differential mass transfer eqn .
( [ 7 ] ) and the radial momentum balance eqn . ( [ 9 ] ) explicitly . even if the accretion flow has considerable thickness , @xmath50 in general
would always be less than @xmath12 .
further , magnetic stresses will compress or squeeze the accretion region .
considering the expression in eqn .
( [ 28 ] ) up to the term @xmath98 would then be a reasonable approximation .
restricting the expansion up to @xmath98 corresponding to @xmath99 and neglecting the terms with orders @xmath100 in the generic expression ( [ 28 ] ) , we equate @xmath101 and @xmath102 to zero , respectively . extending this to mhd conservation equations and assuming isotropic turbulence , after rigorous algebra
, we will then have fifteen independent non - linear algebraic equations with sixteen unknown coefficients consisting of zeroth and first order only , which are shown below .
+ + equation ( [ 7 ] ) renders @xmath103 \frac{t^3}{3 } \ , = \ , - \frac{\dot m}{4\pi } \ , .
\label{29}\end{aligned}\ ] ] equation ( [ 12 ] ) renders @xmath104 @xmath105 equation ( [ 9 ] ) renders @xmath106 @xmath107 equation ( [ 10 ] ) renders @xmath108 @xmath109 equation ( [ 11 ] ) renders @xmath110 @xmath111 equation ( [ 14 ] ) renders @xmath112 @xmath113 equation ( [ 15 ] ) renders @xmath114 @xmath115 \nonumber \\
+ \alpha^{\mathcal r}_{r\varphi } \ , t \left[6 \,c_{s2 } \,b_{\varphi 2 } + \frac{15}{16 } ( c_{s2 } \,b_{\varphi 0 } + 9 \ , c_{s0 } \,b_{\varphi 2 } ) \right]\ , = \ , 0 \ , .
\label{41}\end{aligned}\ ] ] equation ( [ 19 ] ) using eqns .
( [ 20 ] ) and ( [ 21 ] ) renders @xmath116 @xmath117 \ , .
\label{43}\end{aligned}\ ] ] in the above equations @xmath118 .
using eqns .
( [ 29 ] ) , ( [ 30 ] ) , ( [ 31 ] ) , ( [ 33 ] ) , ( [ 34 ] ) , ( [ 35 ] ) , ( [ 36 ] ) , ( [ 38 ] ) , ( [ 39 ] ) , ( [ 40 ] ) , ( [ 42 ] ) , and after very complicated and tedious algebra , we systematically determine the value of all the first order coefficients in terms of zeroth order coefficients of hydrodynamic variables . substituting them in eqns .
( [ 32 ] ) , ( [ 37 ] ) , ( [ 41 ] ) and ( [ 43 ] ) , we successfully reduce fifteen equations to four nonlinear algebraic equations comprising of zeroth order coefficients @xmath119 and @xmath120 only .
we then solve the above four nonlinear equations through an iterative newton raphson method for appropriate initial guess . in these way
we can obtain the values of zeroth and first order coefficients of the corresponding dynamical variables .
the value of the coefficients are scaled by putting @xmath121 , and @xmath3 in units of @xmath9 . in the next section ,
we evaluate them for appropriate choice of parameters @xmath122 and @xmath73 .
in we have analyzed the necessity to have a sub - keplerian advective accretion regime to eject strong outflows and jets from the accretion region in the vicinity of the bhs .
the corresponding accretion regime can be possibly envisaged if the mass accretion rate or net mass flow rate is considerably sub - eddington ( @xmath123 ) .
the ratio of gas to the total pressure @xmath67 and the radiative cooling factor @xmath73 directly depends on @xmath3 .
looking meticulously into the equations from ( @xmath124 ) , we notice that the continuity equation is written in the integral form through eqn .
( [ 29 ] ) , unlike the other hydrodynamical equations .
this is being deliberately done to preserve the information of @xmath3 in the flow , as @xmath3 is the most fundamental parameter which determines the nature of the bh accretion paradigm . however , this constraints the number of dynamical equations . for our case , we have sixteen unknown coefficients but fifteen number of equations . to resolve this , we proceed in the following way .
@xmath3 carries the information of the density of the flow which means that if @xmath3 is known , in principle , density too is known .
equation ( [ 29 ] ) reveals that the density in the accretion - outflow region is a function of two unknown coefficients @xmath125 and @xmath126 .
@xmath125 is the signature of inflow whereas @xmath126 is that of the outflow .
if the outflow is discarded , @xmath126 looses its significance and @xmath3 then becomes the usual mass accretion rate from where @xmath125 can be easily calculated .
if @xmath127 represents the outer radial boundary of the accretion - outflow coupled region , then at @xmath128 the net mass flow @xmath3 is equivalent to the mass accretion rate of the flow . at @xmath128 , density of the accretion flow
is just a function of @xmath125 , which is then given by ( also see [ 5 ] ) @xmath129^{3/2}}{12 \pi \sqrt{2 } } \frac{\dot m}{\alpha_{r\varphi } } \ , .
\label{44}\end{aligned}\ ] ] @xmath130 arises due to mean field approximation .
note that , the value of @xmath125 computed in eqn .
( [ 44 ] ) is not exactly the same as that of @xmath125 in eqn .
( [ 29 ] ) , where it is coupled to @xmath126 .
however , we presume that the density in the accretion flow does not change abruptly due to emanation of the outflow and jet , as only a small fraction of matter is ejected out through the outflow . with this presumption , we calculate @xmath125 in eqn .
( [ 29 ] ) using eqn .
( [ 44 ] ) , and supply its value to the rest of other nonlinear equations for further computation . in the usual accretion flow ,
when there is no net vertical flux , the vertical height of the accretion geometry is calculated from the hydrostatic pressure balance , assuming the pressure and density at the outer accretion surface to be zero .
however , this physical condition cease to exist when the outflow is incorporated in the system , and the scale - height of the coupled accretion - induced outflow region becomes difficult to ascertain . in these circumstances accretion - outflow
coupled surface can be treated as a photospheric height , delineating between the accretion - outflow surface and the transition region leading to the outflow decoupling from the inflow .
we treat @xmath118 as a parameter to get physical plausible solutions .
pessah , chan and psaltis [ 34 ] had shown that during the exponential growth of magnetorotational instability at saturation , the ratio of maxwell stress to reynolds stress becomes @xmath131 where @xmath132 .
@xmath133 is related to the angular velocity through the relation @xmath134 .
@xmath135 signifies an keplerian flow whereas @xmath136 implies a sub - keplerian or an advective accretion flow .
thus for a strongly advective flow @xmath137 .
we use the relation given in eqn .
( [ 45 ] ) to determine the value of @xmath130 used in our equations , which is then related to @xmath138 as @xmath139 the qualitative feature of the dynamical coefficients for an accretion - induced outflow which are physically plausible should satisfy the following properties of the flow variables ; @xmath140 negative , @xmath141 positive , @xmath142 negative .
all the other variables should have a positive value .
the positivity of @xmath83 and @xmath143 is related to the open magnetic field lines threading the accretion flow , across which the accreting matter diffuses , and then gets accelerated outwards along the poloidal field lines by extracting the angular momentum from the flow . the term @xmath144 [ last term in eqn .
( [ 10 ] ) ] , which is a magnetic torque on the accretion flow is attributed to the transport of angular momentum from the @xmath145 surfaces of the accretion - outflow region outwards .
this term should be positive in order to launch a jet , which removes angular momentum from the flow and decouples from the accretion region .
this premise makes a obvious choice for @xmath146 to have a negative value .
it should be reminded that the negative value of many quantities do not necessarily mean that their magnitude is negative , but it represents the direction of their flow .
next we compute the value of the coefficients of all dynamical variables for three relevant choices of @xmath3 with appropriate values of @xmath67 and @xmath73 , conducive to form outflows and jets .
the particular choice of @xmath3 corresponds to riaf , which is linked observationally to low - hard state of bhxrbs and lergs / llagns .
this type of flow is significantly gas pressure dominated , and strongly advective .
the flow is considerably geometrically thick , optically thin and radiatively inefficient .
we choose appropriate values of @xmath67 and @xmath73 corresponding to this @xmath3 to get physical plausible solutions . as
this flow is highly sub - keplerian and strongly advective , an appropriate choice of value of @xmath148 has been taken . in tables
1.1 , 1.2 and 1.3 , we present the computed values of the dynamical coefficients for @xmath149 , @xmath150 with suitable choices of @xmath5 and @xmath151 . for @xmath152
, @xmath66 is @xmath153 .
it is interestingly found that for @xmath154 and in the range of @xmath155 , we never found any physically valid solutions .
however , for much stronger advective paradigm ( @xmath156 ) , one would still obtain physically correct solutions for @xmath155 .
we will analyze the family of analytical solutions later , however , we note that the values of the coefficients are in conformity with desired physically valid solutions ( mentioned earlier ) .
we notice that they are obtained only at high @xmath5 ( @xmath157 ) and at a reduced vertical scale - height ( @xmath158 ) of the accretion - outflow coupled region .
although @xmath151 is small , it corresponds to a geometrically thick accretion flow ( @xmath159 , @xmath151 measures the degree of flow thickness ) .
we have elucidated the necessity of a geometrically thick accretion flow to eject outflow and jet in previous paragraphs . however , corresponding to @xmath160 , we never found any physically acceptable solution for @xmath161 . for higher @xmath162 , however , we obtain a solution at a maximum value of @xmath163 .
this infers that with the increase of @xmath5 , the plausible physical solutions can be obtained with a thicker accretion geometry .
nevertheless , very high @xmath5 ( @xmath164 ) accretion flow might not be realistic in nature , and hence we restrict our study to a maximum plausible value of @xmath162 .
table 1.1 @xmath165 , @xmath166 , @xmath167 , @xmath168 , @xmath169 , @xmath170 . [ cols="^,^,^,^,^,^,^,^",options="header " , ] the earlier values of @xmath3 in the previous subsections correspond to riaf .
nonetheless , moderately advective accretion paradigm may also be susceptible to eject outflows and jets .
these accretion flows which are less advective as compared to riafs , will have lesser gas pressure content and higher cooling efficiency , and are less geometrically thick and are centrifugally more dominating .
they have a moderate optical depth .
this accretion paradigm can be presumably envisaged with @xmath10 ; for our analysis , here , we choose @xmath171 . we choose appropriate values of @xmath67 and @xmath73 corresponding to this @xmath3 to get physical plausible solutions . as
this flow is sub - keplerian and advective , @xmath133 should be greater that @xmath172 , but considerably less than in flows illustrated in previous subsections .
we choose the value of @xmath173 . for @xmath174 and for @xmath175
, we never found any physically valid solutions with @xmath176 .
however , in the stated range one may still obtain valid solutions , either for @xmath177 or for @xmath178 , respectively .
we evaluate the value of the dynamical coefficients for appropriate choice of @xmath179 with @xmath180 corresponding to @xmath181 .
the values of @xmath5 are same as before .
we found that for @xmath182 or @xmath183 , the flow becomes near keplerian ( @xmath184 ) and we get the physical solutions only at @xmath185 .
such a flow is not favourable for ejection of outflow as reasoned earlier .
hence we restrict our study to a maximum value of @xmath186 corresponding to @xmath179 .
corresponding @xmath66 for @xmath179 is @xmath172 . in tables 3.1 , 3.2 , 3.3 , we present them for @xmath186 for appropriate values of @xmath151 . for an easy comparison
we furnish the values of @xmath187 corresponding to @xmath188 for @xmath189 , with appropriate @xmath5 and @xmath151 in table 4 , whose importance we will notice as we proceed .
we found that we do not obtain any solution for @xmath190 with @xmath168 , and @xmath191 for @xmath192 .
the reason of physical solutions of accretion - induced outflow at a reduced scale - height has been stated in ( a ) . however , obtaining physically valid solutions with @xmath189 at a slightly reduced scale - height as compared to that obtained with @xmath193 is consistent with the fact , that as @xmath194 , the accretion flow tends to become more rotationally dominated with diminishing degree of advection .
if we compare the value of the dynamical coefficients for @xmath195 and @xmath196 , we found a fundamental difference in the dynamical nature of @xmath197 .
for @xmath193 , the value of the coefficient @xmath187 is always positive . on the contrary ,
the value of @xmath187 is negative for @xmath189 .
negative value of @xmath187 implies that @xmath197 decreases in @xmath13 in the coupled accretion outflow region for @xmath189 . to verify this anomaly
, we evaluated @xmath187 for lower @xmath67 and higher @xmath73 , and vice - versa .
it is revealing that for @xmath198 and @xmath199 , we always obtain a negative @xmath187 with @xmath162 .
@xmath200 corresponding to @xmath201 , @xmath202 .
with @xmath160 the negative values of @xmath187 are obtained for @xmath203 and @xmath204 .
the corresponding value of @xmath205 for @xmath206 and @xmath207 . to reassure ourselves we computed @xmath187 for @xmath208 , and we arrive at a similar result .
the above consistent findings convey that for strong gas pressure and advection dominated flows ( riafs ) , @xmath197 do not decrease in @xmath13 within the accretion - outflow coupled region . on the contrary , for flows with lesser content of gas and higher cooling efficiency , which are less advective and centrifugally more dominating
, @xmath197 decreases in @xmath13 within the coupled accretion - induced outflow region .
we comment on this apparent dichotomy in .
we also found that with a moderate decrease in @xmath5 from @xmath209 to @xmath210 , @xmath187 turns negative at a higher @xmath73 and at a lower value of @xmath67 , corresponding to @xmath211 .
in this section , we analyze the family of solutions for advective flows in the accretion - outflow coupled region with @xmath212 . although the flow variables vary in both @xmath12 and @xmath13 , we do not display three dimensional figures as they are very obscure and difficult to interpret . as outflow and jet effuse out from inner region of the accretion flow , we restrict our analysis up to @xmath213 schwarzschild radius ( @xmath214 ) within which we presume that the accretion and outflow are coupled , where , @xmath215 .
also it has been stated by kumar et al .
( 2013 ) that vlbi observations of m87 ( junor et al .
1999 ) have shown that the jet originates from the vicinity ( @xmath216 ) of a bh / compact object .
however , we do not expect any outflow out of the accreting plasma in the extreme vicinity of a bh , as the accretion flow remains highly bounded in the extreme vicinity of a gravitationally starved bh .
so we restrict our study up to @xmath217 in the inward radial direction , a quite reasonable choice for the inner radius of the accretion - outflow coupled region .
both @xmath12 and @xmath13 coordinates in the figures are expressed in units of @xmath214 .
we express @xmath3 in our entire analysis in units of @xmath9 .
the dynamical solutions are shown in the following figures with @xmath218 .
note that all the dynamical variables in our study are mean quantities , which are either time averaged or ensemble averaged .
all the flow variables in the figures represent mean quantities .
figures 1 and 2 describe the variation of vertically averaged flow variables as functions of radial coordinate @xmath12 in different accretion paradigms .
variation of @xmath219 and @xmath143 are shown only along coupled accretion - outflow surface @xmath7 as they are odd functions in @xmath13 .
figure 1 shows that with the decrease in @xmath3 , the poloidal components of the velocity ( @xmath220 ) and the sound speed consistently increases .
conversely , the magnitude of @xmath197 increases with the flow becoming less advective and more centrifugally dominated .
it is being interestingly found from the tables in , that the value of the large - scale poloidal magnetic field enhances with the increase in the geometrical thickness of the accretion flow .
as one moves from strongly advective regime to moderately advective regime , there is a sharp fall in the value of the poloidal component of the magnetic field .
this is owing to the fact that the geometrical thickness of the accretion region corresponding to moderately advective accretion flow is much less as compared to the case in strongly advective accretion paradigm , as shown in figures 2a , b .
this indicates the dominating influence of the vertical thickness of the accretion flow structure on the poloidal component of the magnetic field .
in contrast , the toroidal component of the magnetic field @xmath146 always increases with the flow becoming less advective and more centrifugally / rotationally dominated ( fig .
figure 3 shows the variation of poloidal component of velocity and magnetic field with @xmath5 for different accretion paradigms along the radial distance @xmath12 . @xmath221 and
@xmath83 are vertically averaged quantity , whereas @xmath219 and @xmath143 are along coupled accretion - outflow surface @xmath7 .
we choose the value of corresponding @xmath151 for different @xmath5 to be maximum as illustrated in .
this is because the system has a greater tendency to relax itself to the maximum possible height available to render a physical plausible solution of the coupled accretion - outflow , as the geometrically thicker accretion flow is more conducive to propel plasma vertically outwards of the accretion region .
we find that with a small increase in @xmath5 from @xmath210 to @xmath209 , the value of the poloidal component of velocity and the magnetic field increases for both accretion paradigms .
we do not show the variation of other flow variables with @xmath5 as their dependence on @xmath5 is insignificant for a particular @xmath3 , which can be verified from tables in . in figures 4 and 5 we present the variation of the few dynamical variables in @xmath13 ( which are very relevant to outflow ) at any arbitrary location in @xmath12 , for @xmath222 .
figure 4a shows the dependence of @xmath219 in @xmath13 for different @xmath3 for @xmath168 similar to that in fig .
it is found that initially @xmath219 increases rapidly in @xmath13 , however there is a sudden deceleration of @xmath219 as the flow approaches the coupled accretion - induced outflow surface .
this is due to the fact that the inward vertical component of the gravitational force @xmath223 dominates near the coupled accretion - outflow surface .
also at low @xmath3 corresponding to moderately advective accretion flow , the increase of @xmath219 in the vertical direction is much steeper .
the truncation of the curves at a particular @xmath13 represent the corresponding vertical thickness @xmath151 .
figure 4b shows the variation of @xmath219 in @xmath13 with @xmath5 for two different accretion paradigms similar to that in fig .
the nature of the curves are similar to that in fig .
nevertheless , with the increase in @xmath5 , the value of @xmath219 gets enhanced , and there is a steeper increase of @xmath219 in @xmath13 .
the nature of the variation of toroidal component of the magnetic field in @xmath13 for different @xmath3 is shown in fig .
it is seen that with the increase in @xmath3 , @xmath146 increases at a much faster rate in @xmath13 .
variation of @xmath146 with @xmath5 is insignificant ( see tables in ) and hence is not graphically displayed . in figures 4d and 4e , we show the variation of density in @xmath13 corresponding to strongly advective and moderately advective accretion regimes , respectively .
both the figures indicate that @xmath224 decreases with the increase in @xmath5 for all @xmath3 . also , there is a steeper fall of density in @xmath13 with the increase in @xmath3 .
figure 5 depicts the variation of toroidal velocity @xmath197 in @xmath13 . with the increase in @xmath3 ( @xmath225 ) as the flow becomes less advective and more centrifugally dominated , @xmath197 decreases in @xmath13 for all relevant values of @xmath5 and @xmath73 .
this feature of @xmath197 has already been remarked in . in figures 5c and 5d ,
we show them for @xmath189 .
we also find that with increase in @xmath5 the profile of @xmath197 in @xmath13 attains a steeper nature . in figures 6b and
6a we depict the profile of differential magnetic torque ( @xmath2 ) along @xmath13 , as well as differential magnetic torque acting on the coupled accretion - outflow surface @xmath226 along @xmath12 , respectively , corresponding to different accretion paradigms .
this term represents the magnetic extraction of angular momentum by the outflowing matter , what is called the magnetic braking .
the curves show that with the increase in the value of @xmath3 as well as with the increase in @xmath5 ( for a specific value of @xmath3 ) , the value of the differential magnetic torque increases . as @xmath3 increases , with the accretion flow becoming less advective and more centrifugally dominated , the extraction of angular momentum by the outflowing plasma is greatly enhanced .
consequently the gas gets centrifugally accelerated and would leave the accretion region by removing the angular momentum from the accreting matter .
this mechanism predominantly determines the outward flow of the matter for accretion flow with moderate advection .
if the system has a large residual toroidal velocity ( centrifugally dominated ) , it is possible that the angular momentum loss in the vertical direction due to the magnetic torque will be so high , owing to which the degree of angular momentum loss proportionately increases as the matter flow vertically outwards . as a consequence
, there will be an eventual decrease of the toroidal velocity @xmath197 in the vertical direction within the accretion - outflow coupled region , as seen in fig .
5c , d , corresponding to accretion flow with @xmath227 . on the other hand , if the accretion flow is predominantly gas pressure dominated as in a strongly advective regime ( with @xmath193 ) , the gas pressure gradient would play a more contributory role to lift the plasma vertically outwards with the help of magnetic forces , and the effective contribution of the magnetocentrifugal acceleration to control the dynamics of the outflowing matter gets curtailed as compared to that in a more centrifugally dominated accreting system .
figures 6d and 6c show the variation of the poloidal component of the magnetic field @xmath1 [ @xmath228 along @xmath13 and at coupled accretion - outflow surface @xmath229 along @xmath12 , respectively .
the nature of the curves indicate that as high @xmath5 renders the accretion - induced outflow to a greater geometrical thickness ( commented earlier ) , the large - scale poloidal field gets strongly augmented with the increase in @xmath5 for a specific @xmath3 due to the dominating influence of the vertical thickness on @xmath1 . thus with the increase in the turbulent viscosity parameter @xmath5 , the value of the differential magnetic torque responsible for the centrifugal acceleration of the outflowing plasma , as well as the large - scale poloidal magnetic field @xmath1 get strongly augmented ; consequently enhancing the transport of vertical flux outwards . in the moderately advective accretion paradigm with more centrifugal domination ,
the effective contribution to launch and eject the matter vertically outwards from the accretion region arises mainly from the magnetocentrifugal acceleration .
a small increase in the turbulent viscosity parameter @xmath5 from @xmath210 to @xmath209 intensifies the process of extraction of the angular momentum due to the magnetic torque .
the eventual result is the enhanced transport of the outward vertical flux with the increase in the effective angular momentum transport in @xmath13 direction .
this renders @xmath230 to become negative at a lower @xmath73 and higher @xmath67 as compared to that for @xmath168 , as stated in last two lines of 4.3 . in fig .
7a we show the variation of the ratio of @xmath219 and poloidal alfven velocity @xmath231 [ @xmath232 in @xmath13 for different @xmath3 and @xmath5 . with the increase in @xmath3
as the system becomes less advective and more centrifugally / rotationally dominated , there is an increase in the ratio of @xmath233 .
in contrast , with the increase in @xmath5 for a specific @xmath3 there is a sharp fall in the value of the above ratio .
figure 7b shows the variation of the ratio of @xmath219 and net alfven velocity @xmath234 [ @xmath235 in @xmath13 similar to that in fig .
figure 7c depicts the profile of plasma @xmath4 [ @xmath236 in @xmath13 for different @xmath3 and @xmath5 .
we find that @xmath4 always increases steadily in the vertical direction , and its value gets strongly augmented with a small increase in @xmath5 , however , it always remains mostly below equipartition for all relevant @xmath3 and @xmath5 .
moreover it is found that with the increase in @xmath3 , in general , @xmath4 decreases . in figures 8 and 9
we make a comparison of our magnetized accretion - outflow solutions with that of adaf type solutions in [ 5 ] . in figures 8 and 9
, we compare the radial profiles for radial velocity , orbital velocity , sound speed and density , respectively , obtained for our magnetized accretion - outflow solutions with that of adaf type solutions , corresponding to different @xmath3 and different @xmath5 .
we found that the magnitude of radial velocity along @xmath12 that we obtained in our accretion - induced outflow is less as compared to that obtained in self - similar adaf . on the other hand ,
the magnitude of orbital velocity along @xmath12 obtained in our case is higher as compared to that obtained in adaf .
also there is a marginal decrease in the magnitude of sound speed or equivalently the temperature of the gas along @xmath12 obtained in our accretion - induced outflow as compared to that obtained in adaf .
however , the magnitude of density of the gas along @xmath12 in our accretion - induced outflow is found to be almost similar with that obtained in case of adaf .
observationally it is found that low / hard state of bhxrbs which are supposed to be powered by geometrically thick strongly advective sub - eddington ( presumably with @xmath193 ) and consequently quasi - spherical and radiatively inefficient accretion flows ( riafs ) , emanate strong outflows and relativistic jets . outflows and jets are not observed in high / soft of bhxrbs which are powered by geometrically thin and optically thick standard keplerian accretion disk .
the physics of origin and launching of outflows / jets in galactic bh systems ( also called microquasars ) is supposed to be similar with that corresponding to smbhs in agns , as agns may be seen to be scaled up galactic bhs [ 37 ] . geometrically thick advective accretion flow
having substantial amount of gas pressure with strong advection , is more conducive to effuse and accelerate plasma in the vertical direction out of the inner accretion region .
although we do not aspire to explore the exact mechanism of launching and ejection of jets , however , it is generally conceived that the origin , launching and ejection of outflow and jet from the accretion flow is a magnetohydrodynamic process . in the present work we mainly focussed on accretion powered hydromagnetic outflows .
although the distinctive cause of the origin and launching of accretion powered hydromagnetic outflows is still inconclusive , however , it is certain that the dynamics of the outflowing matter should be intrinsically coupled to the accretion dynamics through the fundamental laws of conservation of ( matter , momentum and energy ) within the coupled accretion - induced outflow region , and should not be treated as dissimilar objects .
conservation laws are the most valuable foundation in physics , and play a significant role in understanding astrophysical outflows and jets .
this is because the physical dynamics of the coupled inflow and outflow are essentially governed by the laws of conservation .
the nature of the dynamical solutions in the accretion - outflow coupled region should then reflect upon the physical conditions / criteria to eject outflows .
for the theoretical analysis of the accretion - outflow coupling one needs to be very thoughtful about the proper modelling of the system which essentially needs to solve a complete set of magnetohydrodynamic conservation equations in 2.5-dimension viscous , resistive , advective paradigm . in
we have endeavoured to describe a robust form of accretion - outflow coupled magnetohydrodynamic set of equations in viscous , resistive , advective paradigm , upholding all the conservation laws in 2.5-dimensional mean field mhd regime without any ad hoc proposition , where the dynamical flow variables vary in ( @xmath78 ) .
the mean field approximation gives rise to the emergence of various turbulent correlation terms , where we restrict our study to first order turbulent correlation .
note that turbulent magnetic diffusivity and turbulent viscous term in induction and energy conservation equation arises only from the kinetic part of the turbulent stress tensor through the reynolds stress .
in this work , we have assumed isotropic turbulence and also neglected the contribution of other turbulent stress tensors apart from @xmath25 , which may be dynamically important .
the contribution @xmath25 component would be dynamically more dominant as is responsible for the radial transport of angular momentum outwards .
vertical transport of angular momentum occurs mainly through large - scale magnetic stresses .
nonetheless , in future we would like to examine the possibility of their inclusion , as well as investigate the nature of the flow with anisotropic turbulence .
the inflow and outflow are governed by eight coupled integro - partial differential mhd equations in the cylindrical geometry .
limited observational inputs put constraint on the boundary conditions as well as the scaling relation between accretion and the outflow . owing to the fact that it is beyond the scope to have complete global numerical solutions of the said coupled partial differential mhd equations , motivated us to invoke necessary and proper quasi - analytical method to solve them . ever since the work of [ 5 ] , use of power law self - similarity in studying the accretion flow dynamics , especially the advection dominated accretion flows ( adafs ) to explain the nature of llagns has become widely popular .
realistic strongly advective accretion flow preserves self - similarity reasonably well , within an appreciable region of the flow [ 5,38 ] , and has been widely used to explain observational features in llagns ( see [ 39 ] for a review ) .
previously , self - similar methods had been indeed used to study outflow from the accretion disk on many occasions ( see the references in in introduction ) . it is being found from many studies that self - similarity holds approximately well in context to outflows [ 40,18,28,29,41 ] .
keeping the essence of power law self - similarity , we sought a generalized n@xmath77 degree polynomial expansion of all the dynamical variables in two dimensions , and solve the complete set of coupled integro - partial differential mhd conservation equations within the accretion - outflow coupled region self - consistently , in 2.5 dimensional viscous , resistive advective paradigm , where we have restricted up to order @xmath237 ^ 2 $ ] . in sections 4 and 5 , we have analyzed the nature and the behaviour of our mhd solutions in advective accretion paradigm with @xmath212 .
although we have not intended to explore the physical mechanism of outflow / jet launching in the present study , the quasi - stationary dynamical solutions of the accretion - induced outflow carry the information about the physical conditions / criteria to propel matter vertically outwards out of the accretion - outflow region .
we have mainly focused within the accretion - outflow coupled region where the flow is essentially bounded .
we obtain solutions at a reduced vertical thickness irrespective of the nature of accretion paradigm which we have focused on .
magnetic field tends to compress or squeeze the accretion region by counterbalancing the thermal pressure gradient .
we found that the large - scale poloidal component of the magnetic field is enhanced with the increase in the geometrical thickness of the accretion flow , consistently . with the increase in @xmath3 as @xmath238 , with the accretion flow becoming less advective and more centrifugally dominated with lesser geometrical thickness i.e. , for the flow with moderate advection
, there is a sharp fall in the value of poloidal component of the magnetic field , however with a strong enhancement in the value of the toroidal component of the magnetic field and consequently the differential magnetic torque ( @xmath2 ) .
this term ( differential magnetic torque ) is responsible for the magnetic extraction of the angular momentum to magnetocentrifugally accelerate the outflowing plasma out of the radial accretion flow , and this predominantly determines the outward flow of the matter in a moderately advective accretion paradigm which is more centrifugally dominated .
however , with the decrease in @xmath3 as the flow becomes strongly advective ( @xmath193 ) and geometrically more thick with strong gas pressure and inefficient cooling , despite in the decrease in the value of @xmath2 , a consistent increase in the @xmath18 occurs , indicating that the gas pressure gradient might play a more contributory role to lift the plasma vertically outwards with the help of magnetic forces .
the plasma in the accretion flow can be lifted outwards and ejected , only if some physical process can overcome the effect of the inward vertical force due to the central gravity .
the dynamical behaviour of the solutions indicate that in the advective paradigm both magnetocentrifugal acceleration and thermal pressure gradient along with the magnetic forces , will help in lifting and accelerating the plasma vertically outwards , and the gas material will diffuse outwards across magnetic field lines . however , the effective contribution of either magnetocentrifugal acceleration or thermal pressure gradient to lift the plasma vertically outwards depends on the degree to which the flow is advective .
in fact , with the increase in mass accretion rate as the flow tends to become less advective and more centrifugally dominated with lesser geometrical thickness , in general , the efficacy of the disk to eject outflows diminishes . in paper ( in preparation )
, we have quantitatively demonstrated this aspect with the increase in @xmath3 from @xmath239 to @xmath240 ; the accretion flow with @xmath227 , least conducive to eject outflows .
we obtain dynamical solutions in accretion - outflow coupled region only at high turbulent diffusive parameter @xmath0 .
the accretion - induced outflow solutions have a profound dependence on turbulent diffusive parameter @xmath5 .
it is being interestingly found from our solutions that the enhancement in @xmath5 renders the accretion - induced outflow region to attain a greater geometrical thickness .
consequently , the poloidal component of magnetic field @xmath1 , as well as the differential magnetic torque ( @xmath2 ) get strongly augmented , enhancing the transport of vertical flux outwards . also the plasma beta @xmath4 increases steadily in the vertical direction , and its value gets strongly augmented with a small increase in @xmath5 , however , it always remains mostly below equipartition within the accretion - outflow coupled region .
although we expect the accretion flow to have a large @xmath5 owing to advective nature of the flow , however , the values of @xmath5 that we have obtained in our solution may have been slightly overestimated .
nonetheless , it is indeed being found from previous works [ 5,6,38,42 ] that strongly advective accretion flows , in general , are mostly plausible for large values of @xmath5 ( @xmath241 ) .
gu and lu [ 43 ] also shown that the transition from an outer geometrically thin keplerian disk to an advection - dominated accretion flow is possible for @xmath242 .
mckinney and narayan [ 29 ] , in their grmhd simulation of disk - outflow model , found to have large turbulent viscosity parameter in the accretion disk in the vicinity of the bh .
further , king , pringle and livio [ 44 ] suggested a typical range of @xmath243 from observational evidence .
one of the important approximation we have used in our study is to treat the scale - height of accretion - outflow coupled region as a parameter . in reality
the physical conditions to launch outflow would consistently determine the vertical height of the inflow - outflow surface , from where the outflow decouples from the accretion region .
moreover , in our study we have neglected the effect of spin of the bh in the accretion dynamics and its subsequent impact on the outflow , as the self - similar technique can only be used in the newtonian approximation .
this restricts us of using this method to investigate the physical behaviour of the system in the extreme vicinity of the bh , where general relativistic effects are indispensable .
although power law self - similarity is an analytical approximation , and the quantitative feature of the solutions may have been either overestimated or undervalued , the dynamical solutions show consistent and predictable behaviour , and do exhibit many physical insights on the nature of the accretion - induced outflow , as well reflect upon the relevant physical conditions to propel and eject plasma out of the accretion flow .
power law self - similarity , thus , seems to be reasonably good approximation , within the accretion - outflow coupled region .
nonetheless , more extensive study is required to understand the definitive criteria or condition in launching accretion powered outflows and jets .
bh spin is a very important aspect that needs to be incorporated in the conservation equations atleast through the use of pseudo - newtonian potentials ( e.g. , [ 45 ] ) , while understanding accretion powered outflow dynamics or the correlated dynamics of accretion and outflow .
it is found from the work of [ 10 ] that the spin of the bh significantly influences accretion powered outflows / jets ; for a rapidly rotating bh , the outflow power increases by @xmath244 two orders in magnitude .
moreover , it is also essential that to have complete and a more realistic understanding of the dynamics of the accretion - outflow coupled region , a global numerical solution of such a system in advective paradigm should be performed .
also , explicit inclusion of cooling / radiative processes is required for the completeness of energy conservation .
the relevant dynamical solutions at the accretion - outflow coupled surface would then necessarily act as boundary conditions at the base of the jet .
a more definitive understanding of the criteria to launch accretion powered outflows and jets , thus , requires a complete 2.5-dimensional viscous , resistive , advective global mhd numerical solution with the inclusion of bh spin , which is left for future work . in a subsequent work [ paper ( in preparation ) ]
we will investigate in detail , the energetics of the magnetized accretion - induced outflows and study the spectral behaviour of accretion powered sources .
the integro - differential continuity equation ( [ 7 ] ) , after substitution of the flow variables in the n@xmath77 polynomial order is given by @xmath245 \
, \big(\frac{h}{r}\big)^{2n+1}= - \frac{\dot m}{4 \pi } \eqno(a1)\ ] ] similarly the polynomial expansion of all the other height - integrated mhd equations are done , however we do not furnish all of them here as the structure of the equations are huge . as an example we show it for the radial momentum balance equation ( [ 9 ] ) .
+ @xmath246 @xmath247 } \
, v_{r 2l } \ , + \ , g m \sum_{n=0}^{\infty } \sum_{m=0}^{n
} { -3/2 \choose m } \ , \rho_{2 ( n - m ) } \,\ ] ] @xmath248 @xmath249 \ , b_{z [ 2(n - m)+1 ] } \ , b_{z [ 2(m-1)+1 ] } \ , - \ , \sum_{n=1}^{\infty } \sum_{m=1}^{n } \ , 2 m \ , b_{z [ 2(n - m)+1 ] } \ , b_{r2 m } \biggr ] \ , \biggr ] \ , \frac{1}{2n+1 } \ , \bigl(\frac{h}{r}\bigr)^{2n } = 0 .
\eqno(a2)\ ] ] i. f. mirabel , l. f. rodriguez , nature , * 371 * , 46 ( 1994 ) ; i. f. mirabel , l. f. rodriguez , nature , * 392 * , 673 ( 1998 ) ; s. s. eikenberry , k. matthews , e. h. morgan , r. a. remillard , r. w. nelson , apj , * 494 * , 61 ( 1998 ) s. w. allen , r. j. h. dunn , a . c. fabian , g. b.taylor , c. s. reynolds , mnras , * 372 * , 21 ( 2006 ) ; m. j. hardcastle , d. a. evans , j. h. croston , mnras , * 370 * , 1893 ( 2006 ) ; m. j. hardcastle , d. a. evans , j. h. croston , mnras , * 376 * , 1849 ( 2007 ) ; b. balmaverde , r. d. baldi , a. capetti , a&a * 486 * , 119 ( 2008 ) ; s. ghosh , p. banik , ijmpd , * 24 * , 1550084 ( 2015 ) m. wardle , a. knigl , apj , * 410 * , 218 ( 1993 ) ; j. contopoulos j , apj , * 460 * , 185 ( 1996 ) ; j. ferreira j , g. pelletier , a&a , * 295 * , 807 ( 1995 ) ; j. ferreira , a&a , * 319 * , 340 ( 1997 ) ; f. casse f , j. ferreira , a&a , * 353 * , 1115 ( 2000 ) ; r. narayan , j. c. mckinney , a. j. farmer , mnras , * 375 * , 548 ( 2007 ) e. m. de gouveia dal pino , a. lazarian , a&a , * 441 * , 845 ( 2005 ) ; e. m. de gouveia dal pino , p. p. piovezan , l. h. s. kadowaki , a&a , * 518 * , 5 ( 2010 ) ; d. giannios , d. a. uzdensky , m. c. begelman , mnras , * 402 * , 1649 ( 2010 ) f. yuan f , w. cui , r. narayan , apj , * 620 * , 905 ( 2005 ) ; e. g. krding , s. migliari , r. fender , t. belloni , c. knigge , i. mchardy , mnras , * 380 * , 301 ( 2007 ) ; j. neilsen , j. c. lee , nature , * 458 * , 481 ( 2009 ) ; p. soleri et al . , mnras , * 406 * , 1471 ( 2010 ) ; j. c. a miller - jones et al . , iaus , * 275 * , 224 ( 2011 ) ; j. c. a miller - jones et al . , mnras , * 421 * , 468 ( 2012 ) ; j. j. miller et al .
apj , * 776 * , 104 ( 2013 ) ; a. l king , j. m. miller , m. t. reynolds , k. gltekin , e. gallo , d. maitra , apj , * 774 * , 25 ( 2013 ) s. ghosh , b. mukhopadhyay , apj , * 367 * , 667 ( 2007 ) ; s. ghosh , t. sarkar , a. bhadra , phys .
d , * 92 * , 083010 ( 2015 ) ; t. sarkar , s. ghosh , a. bhadra , phys .
d , * 90 * , 063008 ( 2014 ) ; s. ghosh , t. sarkar , a. bhadra , apj , * 828 * , 6 ( 2016 ) | the correlated and coupled dynamics of accretion and outflow around black holes ( bhs ) are essentially governed by the fundamental laws of conservation as outflow extracts matter , momentum and energy from the accretion region .
here we analyzed a robust form of 2.5-dimensional viscous , resistive , advective magnetized accretion - outflow coupling in bh systems , in the mean field magnetohydrodynamical ( mhd ) regime .
we solve the complete set of coupled mhd conservation equations self - consistently , through invoking a generalized polynomial expansion in two dimensions .
we perform a critical analysis of accretion - outflow region and provide a complete quasi - analytical family of solutions for advective flows .
we obtain the physical plausible outflow solutions at high turbulent viscosity parameter @xmath0 , and at a reduced scale - height , as magnetic stresses compress or squeeze the flow region .
we found that the value of the large - scale poloidal magnetic field @xmath1 is enhanced with increasing geometrical thickness of the accretion flow . on the other hand differential magnetic torque ( @xmath2 ) increases with the increase in @xmath3 .
@xmath1 , @xmath2 as well as the plasma beta @xmath4 get strongly augmented with the increase in the value of @xmath5 , enhancing the transport of vertical flux outwards .
our solutions indicate that magnetocentrifugal acceleration plausibly plays a dominant role in effusing out plasma from the radial accretion flow in moderately advective paradigm which are more centrifugally dominated , however in strongly advective paradigm it is likely that the thermal pressure gradient would play a more contributory role in the vertical transport of the plasma . |
the study of the consumption of cultural goods in general , and that of films in particular , has been traditionally restricted to total demand empirical studies .
most of these studies have followed the original guidelines set by the earliest authors , such as baumol & bowen @xcite or moore @xcite . in these studies variations in quality of the good have been consigned to a residual status , focusing in the effect of prices and income as the main explanatory factors .
other studies , such as the one performed by blanco & baos - pino @xcite , have considered the availability of alternative leisure activities ; while on the other side , models such as the one presented by thorsby @xcite have considered the impact of quality on final attendance , incorporating it either as an expected value at the individual level or as a macro variable obtained ex - post from critic s review indexes @xcite or from online reviews @xcite .
although we acknowledge their efforts , we must say that in the latter works the authors were not concerned with the structure of the consumption cycle as an indicator of film quality .
their work is close to ours , only in the sense that reviews act as a form of social interaction , which turns out to be a key ingredient of our model .
however useful , demand studies as the ones surveyed above do not account for the dynamics of the consumption life cycle of the cultural good and , more importantly , they do not discuss the ways in which the structure of this cycle is affected by the transmission of information from members who have had access to the cultural good to potential consumers . beside economic analysis
, the sociologist lipovetsky @xcite has focused his work on the macro structure of the life cycle , mainly in terms of its duration .
basic principles which underlie this process , however , are not identified in his work . in this paper
we propose a model based on first principles , which agrees with all the observed behaviors exhibited by the empirical data of the movie industry .
the model is concerned with the primary life cycle of the good , but it differs from the others in the sense that the whole structure of the life cycle is recovered , not only the characteristic decay times or the total consumption are considered like in previous studies .
although the consumption life cycle of some cultural goods can be potentially unbounded ( we still buy copies from don quixote ) , the aim of this work is to understand the dynamics of the primary life cycle , which finishes when per period consumption decreases below a certain threshold relative to its premiere level .
a less ambiguous case is that of performing arts like theater , where producers are forced to cancel presentations when box - office revenues goes below fixed costs . the alternative cost associated with the availability of new options , effect which also applies to the film industry or best - sellers book industry , strengthens the limited duration life cycle that characterizes aggregated consumption in these environments .
we present a model that reproduces the life cycle dynamics and is determined by three basic factors : * \(i ) the size of the targeted group * \(ii ) the prior conception about the quality of the good in question ; and * \(iii ) the effect of social interactions , in the form of information about the quality of the good , between agents who have effectively experienced the good and potential consumers
. it will be the latter effect the one which determines the particular shape of the consumption life cycle . in relation with the literature studying the diffusion of technological innovations
, our model relates to the works of mansfield @xcite , and bass @xcite , with the difference that the dynamical process that underlies the shape of the cycle are identified and made accountable for it . whereas bass uses a parameterized curve that he fits into the empirical behavior , without incorporating any dynamical justifications for the curves he chooses .
word of mouth can be justified as a valid mechanism of social influence by following works like the one presented by moore @xcite .
he argues that given the scale economy that characterizes certain technologies , e.g. , database software , it is impractical ( or too expensive ) to do small - scale experiments and , thus , word - of - mouth becomes an important part of the diffusion process of these investment goods .
going to the movies , however , poses a somewhat different problem for potential consumers .
first , `` early adopters '' do not face the risks normally associated with the adoption of a novel technology and thus an important fraction of them will be willing to consume the good before peer s opinions become available .
this will be performed based solely on the pre - opening expectations generated by the media and the producers themselves .
this leads the life - cycle consumption process to diverge from the standard s - shaped behavior associated with the adoption of a novel technology .
on the other hand , new movies which become available for the general public tend to have an initial attendance which is relatively high and usually decrease all along the consumption cycle , the sole exception being small productions or independent films which show a consumption life - cycle that resembles that of novel technology adoption .
our model captures both forms of behavior . other family of models , which takes into account social interaction effects in the dynamics of consumption , is that of `` fashion cycles '' @xcite . in our case , however , the demand of a good does not decrease due to the consumption of it by other agents , the movie does not become worn out , like a fashion design , in fact what the model attempts to capture is the fact that agents who have seen the movie have an impact on the expected value of the good of agents who have not seen it . a different approach for the diffusion of innovations is to consider site percolation on regular lattices @xcite .
these models are based on the assumption that agents occupy the vertex of a regular lattice in @xmath5 dimensions and are represented by a random number .
innovations diffuse by percolating the lattice according to a certain quality value which when above the percolation threshold generates a giant consumption cluster .
although we acknowledge this efforts we believe that spatially realistic models should consider a network substrate instead of a regular lattice .
social networks exhibit topological properties which are completely different to the substrate defined by a regular lattice , such as the scaling properties of its degree distribution @xcite , the community structure @xcite and the short average path length @xcite to name a few . in this paper
we present a model that does not take this substrate into account , it could be consider a mean field solution or jelly model , regardless of this we are able to reproduce all the types of aggregated behavior .
we now present a model of cultural consumption based on the following assumptions . *
\(i ) each agent goes to see a movie at the theater only once .
( we neglect the probability of going twice assuming that the probability of going to the theater to see the same movie @xmath6 times is a rapidly decaying function of @xmath6 ) . *
\(ii ) the probability that an agent goes to the movies is affected by the interactions with agents who have already seen it .
the different quantities associated with the model will be expressed using the following notation .
@xmath7 will represent the number of agents that have not watched the movie at discrete time @xmath8 , while @xmath9 will represent the probability that an agent who has not watched the movie decides to attend it at discrete time @xmath8 .
the observable quantity that can be measured is the number of agents who have seen the movie at discrete time @xmath8 .
we will call this quantity @xmath10 . in our model , it will be given by the product @xmath11 , which is nothing more than the expected number of agents attending the good . before the movie is available for its potential consumers , agents have a prior conception about its quality , which comes from the information of pre - observable features such as its budget , the cast , and advertisement .
the availability of this information will depend on the marketing strategies of the producers and distributors .
irrespectively of the way in which these strategies are conceptualized , i.e. , whether publicity is taken as an information provider or as a persuasive device , its effects on our model are equivalent , affecting the prior conception about the movie in question and , therefore , the agents likelihood of actually purchasing it at the beginning of the process .
we will denote this likelihood by @xmath12 , and we will call the initial target population @xmath13 ; which represents the total number of potential attendants of the good before it becomes available . it can be argued that the length of the consumption cycle that the paper takes into consideration is unrealistically short , since the energy of film producers is directed to finding , successfully , new ways of lengthening it .
sedgwick @xcite claims that more than 70 per cent of film revenue is derived from non - theatrical resources .
however , given the dramatic effect that the primary cycle of consumption has in the subsequent ones , we believe that this primary cycle is worth of investigation on its own .
the simplest case is the one of an atomized society where agents decisions are independent from each other , both in terms of information the opinion of agents who have watched the movie do not influence the decision of potential consumers and in terms of consumption there are no network externalities associated with timely coordinated consumption in this case , we will also assume that restrictions on the supply side do not apply .
the whole targeted population could simultaneously go to the theater and the horizon of its exhibition has no time limit . in this case
the probability of attending the theater does not depend on time , and it is equal to its initial value @xmath12 .
the system dynamics can be described by considering that the expected attendance at a given time is given by the product between @xmath12 and @xmath7 ; and that at the beginning of the process no agents have seen the movie .
after the first time step some agents will have attended the theater , and we will therefore have to subtract them from the ones who have not .
thus the temporal variation of will be given by @xmath14 if we approximate the discrete variables by continuous ones and notice that in the first time step attendance is given by , we can conclude that the general solution of equation ( 1 ) has the traditional form @xmath15 @xmath16 in the next section we will see how social interactions alter this behavior .
once the effect of social interactions are considered , @xmath9 becomes a dynamical variable and its change in time is then due to two main contributions .
the first one is the one associated with the transmission of information about the observed value of the movie , and it represents the change in the probability of attending the theater induced by the information about it transmitted between agents . after the first period of attendance , residual potential consumers have access to the opinion of the first period audience .
this effect is cumulative as the residual consumers have potential access to the opinions of audiences from the previous time steps order where opinions are transmitted to residual consumers also via agents who have not yet participated in the consumption of the cultural good .
qualitative properties of an @xmath6 order process in which the effect s strength diminishes over time , would be equivalent to the @xmath17 order process investigated here .
wu et al.@xcite have shown that information flow remains bounded as long as the likelihood of transmitting information remains sufficiently small ] .
we assume that the change induced in @xmath9 by this effect will be proportional to the probability that agents who have not seen the movie meet with agents who have .
the proportionality constant associated with this term will be the one representing the strength of it .
the second contribution we considered is the one associated with coordinated consumption . in the case of performing arts , agents usually attend in groups ; therefore agents who have not seen the movie are less likely to go if they are not able to find other agents to keep them company .
the change induced by this effect will always reduce attendance likelihood , and it is also proportional to the probability that agents who have attended the good meet with ones who have not .
therefore , when we consider the effect of social interactions , we can write the change in the probability of attending the good as @xmath18 where @xmath19 represents the proportionality constant associated with the effects of information flow while @xmath20 is the one associated with the social coordination effects . from now on we will focus on the case in which @xmath19 can take positive and negative values representing the fact that information transmitted between agents can stimulate or inhibit the future attendance of other agents . whereas @xmath20 will always contribute negatively .
this is because coordinated consumption can only reduce the likelihood of going to the theater for un - coordinated agents .
both terms , it is worth noticing , act directly on the population , altering the likelihood that agents will attend or purchase the good .
although agents are not explicit rational optimizers who make their decision over the basis of a bayesian update of their beliefs about the movie s quality , one could understand this model as the outcome of rational searching behavior in an incomplete information environment .
the continuous version of the system can be solved analytically in the case in which @xmath19,@xmath20 do not depend on time and the effects are not cumulative . in this case ( 1 ) and ( 4 ) reduce to @xmath21 @xmath22 which in the continuous case can be represented by @xmath23 @xmath24 where @xmath25 to find a solution we notice that equation ( 7 ) can be substituted into equation ( 8) giving us a third differential equation relating @xmath26 and @xmath27 which can be solved and used to introduce the initial conditions of the system .
this relation can be expressed as @xmath28 replacing [ 9 ] back on equation ( 7 ) and integrating it , we find that the solution of the system is given by @xmath29 @xmath30 where the constant @xmath31 is defined as @xmath32 ( a ) fraction of agents attending the theater a a function of time according to the model described by eqns .
( [ 7 ] ) and ( [ 8 ] )
. the segmented line represents the atomized behavior described by eqn .
( [ 3 ] ) while the lines on top of it and below of it represent the behavior of positive and negative @xmath33 values respectively .
( b ) shows the accumulated attendance , or in other words , the integral in time of ( a).,scaledwidth=100.0% ] figure [ fig1 ] shows the attendance as a function of time as well as the accumulated attendance normalized by the total target population ( @xmath34 and @xmath35 ) .
the segmented line represents the atomized behavior occurring when @xmath9 remains constant throughout the process , it also divides the behavior of the system in two , the solutions that lie above it are examples of cases in which @xmath33 is a positive quantity , these are examples of systems which are dominated by a strongly favorable flow of information stimulating the consumption of the good . on the contrary ,
the lines that lie below the dashed one are the ones for which @xmath33 is a negative quantity and are dominated by coordination effects .
summarizing , we can see that two classes of behavior are predicted .
the first class is characterized by a monotonic decay with an exponential tail .
whereas , when the strength of social interactions @xmath33 is large enough , a second class of behavior emerges . in this case a period of increasing attendance exists until a maximum is reached . after this , the first class of behavior develops .
the latter case has an accumulated behavior that resembles the standard ogive s - shape , which characterizes technological diffusion .
the model represented by equations ( 5 ) and ( 6 ) was validated by comparing it with the u.s .
box - office data available on the internet movie database ( imdb ) web site .
we considered the 44 movies with the highest budgets of 2003 as our sample set and performed a @xmath36 estimation procedure on all movies in order to find the parameter set which most accurately fitted the empirical results .
the model has 3 free parameters , the two initial conditions @xmath13 and @xmath37 , and the constant representing the strength of social interactions @xmath33 .
the parameters used in the @xmath36 minimization process were @xmath33 and @xmath38 while @xmath12 was determined by matching the first data point on the data set with the first data point in the model-@xmath39- .
this reduced the number of free parameters to just two .
the data set considered did not contain weekly attendance as a function of time but the weekly gross collected by the movie .
we consider that the amount of money collected in the box office by a movie is a linear function of the number of people that attended it .
this allow us to do our empirical analysis based on the weekly amount of money collected by a particular movie , which is equivalent to making the analysis based on the number of agents attending it .
three examples of the fitting procedure are shown for ( a ) lord of the rings i ( b ) blade ii ( c ) kissing jessica stein.,scaledwidth=100.0% ] figure ( [ fig2 ] ) shows three plots comparing the model with the empirical data . both , the lord of the rings : fellowship of the ring , and blade ii represent two examples of the first class of behavior identified in the previous section .
it is worth noticing that the slope followed by the data changes in time .
the first points have a negative slope , which is considerably more pronounced than the ones in the tail .
a simple exponential fit will be accurate in only one section of the curve , but would fail to fit both behaviors that were expected from the model and that appear to be present in nature in order to interpret the relative values of @xmath33 , and therefore compare different motion pictures , it is important to consider the behavioral class in which the particular films are located .
blockbuster movies have a high initial attendance , which causes population finite size and social coordination effects to be very strong .
this tell us that movies in this category will usually have a negative value for @xmath33 , and the magnitude of it will represent the observed value of the good .
small negative values are associated with movies in which the decay is not accelerated due to the observed quality , but exhibit some acceleration due to social coordination effects . on the other hand ,
large negative values map into accelerated decays that we believe are due to social coordination effects plus poor values in the observed quality of the good . in the second class of behavior ,
the premier level of attendance is low , thus coordination effects do not act strongly on the system , because of the initially slow depleting of potential customers . in this case ,
small values of @xmath33 represent a movie in which agents attended in a random fashion and the decay is not accelerated or damped due to social effects . on the other hand when relatively large values of @xmath33 are observed the attendance increases during the first time steps .
this increase in consumption is due to the fact that the movie was well evaluated and that the pool of possible attendants was not depleted during the first couple of weeks . summarizing , in order to interpret the parameter associated with the observed quality of the good @xmath33 , it is important to consider two things .
first we consider the class of behavior to which the movie actually pertains , this can be done by observing if the premiere attendance represents a large fraction of the total one .
then , once the movie has been associated with a certain class of behavior , the actual value of sigma should be interpreted relative to the distribution of values associated with movies in that particular class from the empirical point of view and after observing the whole data set it becomes apparent that the values of sigma that we found tend to lie in a well defined distribution . figure ( [ fig4 ] )
shows the distribution of values for @xmath33 obtained using this data set and the procedure explained above .
given the selection criterion used , movies with the highest budgets , it is natural to see that this curve shows @xmath33 values belonging to the blockbuster class of behavior .
a negative bias is observed when we look at the distribution of @xmath33 presented in figure ( [ fig4 ] ) .
this negative bias tells us that in most cases , the coordination effect , which could also be associated with the underlying effect of cultural competition , dominates the dynamical properties of the system . for the 44 studied movies .
[ fig4],scaledwidth=50.0% ] on the basis of the conjecture that coordinated consumption has the same effect on all motion pictures inside the same behavioral class , we can interpret that the value extracted for @xmath33 represents the actual effect that the flow of the information associated with the observed value has on the film , and therefore on its consumption life cycle .
this would indicate that coordinated consumption only introduces a shift in the value of @xmath33 and that the deviation from this well - defined mean represents the actual value of the movie as given by the targeted audience .
for instance , the differences in the estimated parameter for the lord of the rings and blade ii , would indicate that the former film was much better received than the latter one . *
what is interesting about this result is that we do not claim an actual value based on the opinion of a film critic or an audience survey , we infer it from the structure of attendance behavior we observe*. this reversed engineering , definition of the observed value of subjective _ film quality _ presented here can be used to characterize consumers response to specific genres and thus help target film publicity and distribution in a more effective way .
a dynamical model representing the life cycle of motion pictures was introduced .
the assumptions of the model were that * \i ) agents do not go to the theater to see a particular movie more than once , and that * \ii ) the probability that agents go to the theater changes in a way which is proportional to the number of agents who have attended the theater times the ones who have not .
the first of these assumptions is the one that gives rise to the exponential decay that characterize the tail of this process , while the second one is the one that allows the system to adjust its decay according to the social interactions present in cultural consumption , and consequently allows the model to accurately fit the different classes of observed behavior , namely , * \i ) a monotonic decay with an exponential tail , and * \ii ) an exponential adoption followed by an exponential tail which is traduced into an ogive s - shaped behavior when the accumulated theater attendance is observed .
in addition , under certain assumptions , our model can be used to infer a quantitative estimator of the subjective reception of a particular film .
an estimate which is independent of critics review and is inferred only from the shape of the consumption life cycle .
further research in this area can be directed in a variety of ways .
a natural extension of the model analyzed here would be to consider the more general setting in which the agents face a number of cultural options and a limited budget for a given period of time .
it is also an area interest to investigate the link between the primary life of consumption and subsequent phases of consumption .
another route for more empirically oriented research could be associated with the investigation of longitudinal processes in order to explore how the structure of the consumption life cycle has evolved along the last decades , and transversal processes to see the differences and correlations between the reception of particular films in different geographical regions .
a more complete analysis of the movie industry database should be carried out in order to understand the nature of these structures .
we thank the comments of samuel bowles , peter richerson , marcos singer and the participants of the workshop on evolution and the social sciences at the ceu , budapest .
we also acknowledge financial support from fundacion andes grant c-13960 . | we model the consumption life cycle of theater attendance for single movies by taking into account the size of the targeted group and the effect of social interactions .
we provide an analytical solution of such model , which we contrast with empirical data from the film industry obtaining good agreement with the diverse types of behaviors empirically found .
the model grants a quantitative measure of the valorization of this cultural good based on the relative values of the coupling between agents who have watched the movie and those who have not .
this represents a measurement of the observed quality of the good that is extracted solely from its dynamics , independently of critics reviews . _
@xmath0 department of physics and center for complex network research , university of notre dame , notre dame , in 46556 .
+ @xmath1 kellogg institute , 130 hesburgh center , notre dame , in , 46556 .
+ @xmath2 department of physics , university of michigan , ann arbor 48104 .
+ @xmath3 department of sociology , pontificia universidad catlica de chile , vicua mackenna 4860 , macul , santiago chile .
+ @xmath4santa fe institute , 1399 hyde park road , santa fe , nm 87501 .
usa _ |
since the first discovery of the two clearest examples of moving groups in the solar neighborhood ( the hyades and the ursa ) by proctor ( 1869 ) , the studies of moving group have been paid attention for a long time .
there are two main techniques to detect moving groups .
one is the convergent point ( brown 1950 , jones 1971 ) technique for proper motion data ; the other is searching kinematic structure in velocity space .
most works adopted the second technique .
for example , g@xmath0 et al . (
1990 ) used the sem algorithm to decompose the sample into the sum of tridimensional gaussians in the ( u , v , w ) velocity , and concluded that the observed distribution of the residual velocity can be explained as the sum of four independent distributions which were related to several open clusters in the solar neighborhood .
chen et al . ( 1997 ) developed an algorithm using a non - parametric kernel estimator to describe the the stellar distribution in a 4-dimensional space ( velocities and age ) and four moving groups near the sun ( pleiades , sirius , hyades , ic2391 ) have been identified without assuming any prior knowledge of moving groups , neither the velocity distribution nor other physical properties .
after the release of the hipparcos catalogue ( esa 1997 ) , the research of moving groups in the solar neighborhood has made great progress since accurate parallaxes and proper motions for a large number of stars are available .
( skuljan et al . 1999 ) use the wavelet transform technique to analyze the distribution derived by an adaptive kernel method and also find several moving groups .
recent works by famaey et al .
( 2005 , 2007 ) have identified five moving groups based on 6000 giants and investigated the mass distribution of the hyades stream based on the geneva - copenhagen survey of 14000 dwarfs from nordstrm et al .
klement et al .
( 2008 ) identified at least four ` phase space overdensities ' of stars on very similar orbits in the solar neighborhood using the @xmath1 rave public data release .
different authors use different techniques and different samples , but they all report the presence of moving groups in the solar neighborhood . with the newly - developed wavelet transform technique by skuljan et al .
( 1999 ) , we attempt to identify the stellar moving groups in the solar neighborhood by combining both the dwarf samples from nordstrm et al .
( 2004 ) and the giant samples from famaey et al .
the main goals are to investigate whether the locations in velocity space of those groups derived from different samples are consistent as well as finding new streams based on large samples of stars and new method .
since the sample is in the solar neighborhood , the position of the stars does not provide any discriminant information for the detection of moving groups , so the detection in our paper mainly depends on the two components of the stellar velocity - the u component pointing towards the galactic , v component towards the direction of the galactic rotation .
first , the probability density function f(u , v ) is estimate by a kernel function , and next the wavelet transform will be carried out , so we can finally recognize the moving groups on the basis of the analysis of wavelet coefficients .
we use the kernel function to decide the probability density at any given point .
the type of kernel function used in our method is a radial basis function ( equation 1 ) .
h is a smoothing parameter , ( @xmath2 ) is the covariance matrix .
the probability density of the ( u , v ) panel is then derived by equation 2 .
n is the number of stars in our sample . @xmath3 @xmath4 after deriving the density of any given ( u , v ) point , in order to achieve the ( u , v ) center and the dispersion of the possible moving groups , we use two - dimension wavelet transform technique to analyze them ( skuljan et al .
wavelet analysis ( daubechies 1988 ; chui 1992 ; ruskai et al .
1992 ) is a fantastic tool and is becoming more and more popular in astronomical researches .
the wavelet transform provides an easily interpretable visual representation of @xmath5 .
moreover the continuous wavelet transform can be used in singularity detection . in this paper , we use it to find some structures working at different scales . to process a wavelet transform of a function f(x , y ) , we must define an analyzing wavelet @xmath6 called mother wavelet in advance .
@xmath7 is scale variable .
the wavelet coefficient of the point @xmath8 then can be derived by equation 3 .
@xmath9 the actual choice of the analyzing wavelet @xmath10 depends on the particular application .
a mother wavelet named mexican hat ( skuljan et al .
1999 ) is the second derivative of a gaussian and generally gets good results when applied to find singularities .
so when we intend to search for certain groups from a given data distribution , a two - dimensional mexican hat is selected as the mother wavelet ( equation 4 ) .
@xmath11 the main characteristic of the function @xmath10 is that the total volume is equal to zero , which is what enables us to detect any over - densities in our data distribution .
the wavelet coefficients will be all zero if the analyzed distribution is uniform ; but if there is any significant ` bump ' in the distribution , the wavelet transform will give a positive value at that point .
two data sets are selected as our samples .
one is the 14000 dwarf ( nordstrm et al .
2004 , hereafter nord04 ) ; the other is 5311 k and 719 m giant ( famaey et al .
the information of dwarf samples came from the geneva - copenhagen survey , in which the ages , metallicities and kinematic properties are provided , while that of the giant sample was obtained from coravel / hipparcos / tycho-2 .
first the probability density was derived by the method ( sec 2.1 ) . for equation 1
, the smoothing parameter h should be carefully selected because small values of h emphasize noisy structure while large h value will smooth out all the details of the distribution .
our selection of the value h adopts the method of asiain et al ( 1999 ) .
the value of h is set to be 0.145 for the dwarf sample and 0.117 for the giant sample .
the density distribution in the ( u , v ) plane has some irregularities and can not be described by a unique velocity ellipsoid , and some ` bumps ' can be clearly recognized .
famaey et al.(2008)derived the threshold of noise wavelet coefficients to have a value of @xmath12 .
thus , the coefficients smaller than @xmath12 in the present work will be rejected .
fig.1 shows the contour map of the positive wavelet coefficients obtained in the ( u , v ) space from @xmath13 = 4km@xmath14 .
after obtaining the probability density of the samples and the wavelet coefficient contour map , the results show several features . however , are these features real or are they caused by some noise ?
there are several factors that will affect the observed number of stars in each bin : such as the measurement error , statistical fluctuations related to the finite sample , etc .
we can expect that the number of the observed stars in each bin will subject poisson distribution with an average of @xmath15 which can be derived by the probability density function multiplied with factor ns , where n is the total number of stars and s is area covered by a square bin . to investigate how probable are the features showed in fig.1
, we generate a large number ( n=2000 ) of poisson random copies .
for each copy , we get a wavelet positive coefficient map ( after rejecting those smaller than @xmath12 ) .
fig.2 is a example of one copy .
the probability of those features is shown in table 1 .
typically , only those features with 90 percent or better probability are considered to be real . from fig.1 ,
it is clear that stars are clumped at different locations in the ( u , v ) space .
table 1 shows the central ( u , v ) of 22 possible moving groups detected by either the dwarf sample or the giant sample by using the wavelet technique . we analyzed the dwarf sample and the giant star sample separately because the star numbers in the solar neighborhood of dwarf and giant stars are different due to different lifetimes with more dwarf stars than giant stars . in view of this , merging the two types of stars into one sample will reduce the statistically grouped structure in the ( u , v ) space . in some cases
, the giant sample can give better structure than the dwarf sample as we will show later ( for group 7 ) .
it seems that a more reasonable way is the comparison of these grouped structures detected from both the dwarf and giant sample and check if they can give consistent results . in this sense , we have found that the central ( u , v ) values of these groups between the two samples are quite consistent from 12 groups . among these groups , groups 1 to 5
are quite strong and they have been widely reported to be moving groups in the literature .
as we can compare our results with previous works by eggen ( 1991 , 1992a , b , c ) based on the fk5 catalogue and dehnen ( 1998 ) , we found good agreement for the pleiades ( group 1 ) , hyades ( group 2 ) and sirius ( group 4 ) .
specifically , our values are quite consistent with those of dehnen ( 1998 ) , who gave mean motions of ( -12 , -22 ) , ( -40 , -20.0 ) , ( 9 , 3 ) , km@xmath14 for the pleiades , hyades and sirius groups , respectively , based on the sample of 14369 stars observed by the hipparcos satellite .
eggen ( 1991 , 1992a , b , c ) showed mean motions of ( -11.6 , -20.7 ) , ( -40.0 , -17.0 ) , ( 14.9 , 1.3 ) km@xmath14 for the pleiades , hyades and sirius are somewhat updated by using hipparcos parallaxes .
our values are also similar to those values in recent works by famaey et al .
( 2007 ) and antoja et al ( 2008 ) .
mean motions of group 3 is also in agreement with the hercules stream centered at ( -35 , -45 ) found by fux ( 2001 ) .
it is noticed that the mean motions of group 3 are slightly different between the dwarf and giant groups .
considering that the giant group shows better structure than that of the dwarf group in fig . 1 , we suggested that the mean motions of giant groups have more reasonable values for group 3 .
our group 5 with mean motions of ( -11 , -7 ) in the present work is very similar to the castor group with ( -10 , -10 ) , which is classified as a young group by famaey et al .
( 2007 ) with mean motion of ( -10 , -12 ) . besides the five known moving groups , two moving groups ( group 6 and group 7 ) with mean motions of ( 38 , -20 ) and ( -57 , -45 ) are quite significant from both dwarf and giant samples and the probability from our simulation is larger than @xmath16 . for our group 6 ,
antoja et al . (
2008 ) also suggest the structure centered at ( 35 , -20 ) to be a new group . however , since the structure is weak in their work , they considered it as part of the elongation of the sirius or coma berenices structure . however , group 6 is very prominent and has a well - defined shape in the present work both for the dwarf sample and giant sample .
moreover , their mean motions are far away from those of sirius and the coma berenices structures without significant connections in the contour plot of fig .
1 . note that the giant sample of group 6 has some extension to the south - west direction , which disappears in the simulation contour of fig.2 after adding noise while the center part of group 6 persists .
it seems that the center part of group 6 is real while the extension to the south - west direction may be noise . in this sense
, we suggest that group 6 is real in the giant sample , and it is very significant in dwarf sample .
the other five groups ( 8 , 9 , 10 , 12 , 14 ) also have some features in both the dwarf sample and the giant sample and the simulation shows that four of them are significant with probability above 90% of being real , except for the group 10 with the probabilities of 86% and 82% . the rest of the groups are shown only in the dwarf sample ( 15 and 16 ) or only in the giant sample ( 11 , 13 , 17 , 18 , 19 , 20 , 21 , 22 ) ; but four groups ( 11 , 13 , 17 , 21 ) are not very significant with the probability lower than 90% by simulation .
interestingly , some of these features have been reported by previous works .
for example , our group 14 is consistent with the ic 2391 stream , which shows a mean motion of ( -20.8 , -15.9 ) by eggen ( 1992b ) and of ( u=-20 , v=-12 ) by chen et al.(1997 ) .
moreover , our groups 18 , 20 and 22 may be the same groups as no . 7 , 14 and 13 in table 3 of antoja et al ( 2008 ) , which summarizes the possible grouping structures from dehnen(1998 ) and eggen s serial papers .
a catalog of 22 possible moving group candidates are listed in table 1 , although some of them are not so significant ( 10 , 11 , 13 , 17 , 21 ) for a statistical requirement of the probability above 90% .
certainly , these groups with low probabilities from our statistical analysis need further study .
the metallicity is only available in the dwarf sample , based on which we present the peak and scatter of [ fe / h ] for all the candidates in table 1 .
although all the works to detect the moving groups do not take into account the w velocity , this component may bring some new information .
therefore , we have generated the ( v , w ) and ( u , w ) contour plots with the same method .
as expected , the distribution of w velocity is limited to a narrow range from -20 km / s to 5 km / s for both the dwarf and giant samples .
moreover , the dispersion in the w velocity for these groups is about 10 - 15 km / s , which is significantly larger than those of the u and v components of 4 - 6 km / s in defining these groups .
thus , the ( v , w ) and ( u , w ) contour plots have no advantages in identifying possible moving groups .
several groups of the 22 candidates actually belong to the same groups in the ( v , w ) and ( u , w ) contour plots .
note that the groups of 9 and 12 have the largest w dispersion of 20 km / s because they have very weak features in our fig.1 .
in general , these moving group candidates based on the ( u , v ) contour are considered to be more reliable by taking into account the large scatter in the w component distribution .
our group 3 and group 7 appear to coincide with the hercules group identified as green lumps in fig.1 of famaey et al 2008 and they think the green lumps are both the hercules group . however , the argument is not significant and we suggest that they may be distinct group .
fig.3 shows the statistical result of the @xmath17 distribution with the cut of @xmath18 kpc .
it seems that the peak of group 3 in the @xmath17 distribution is at 0.1 - 0.2 kpc while there is a large contribution from stars with @xmath17 of 0.2 - 0.3 kpc for group7 .
moreover , from famaey et al ( 2004 ) , it is clear that the @xmath19 of the hercules group is significantly higher than those of the hyades or sirius , which could indicate an overlapping of more than one distinct group . in the present work ,
the @xmath19 is nearly the same for group 3 and group7 as well as other groups .
finally , there is also a hint of [ fe / h ] deviation with the average [ fe / h ] of -0.16 dex in group 3 versus -0.22 dex for group 7 .
further study on this topic is necessary before firm conclusions can be drawn .
using the dwarfs and giants in the solar neighborhood , we illustrate a detailed analysis of the uv distribution .
this analysis reveals 22 possible grouping structures identified by the kernel estimator and wavelet technique .
the location in velocity space of 12 possible moving groups from both dwarf and giant samples are consistent .
11 groups , including the five well - known groups , pleiades , hyades , hercules , sirus and castor streams , are consistent with previous works .
eight groups , not reported by previous works , are presented and most of them are thought to be significant in term of statistics . in summary ,
a catalog of 22 moving group candidates with the centers ( u , v , w ) , their dispersions and mean metallicity are given .
this study is supported by the national natural science foundation of china under grants no .
10521001 , 10673015 , 10433010 , the national basic research program of china ( 973 program ) no .
2007cb815103/815403 and by the ministry of science and technology of china under grant no . 2006aa01a120 .
antoja t. , figueras f. , fernandez d. , torra j. , 2008 , , 179 , 326 asiain r. , figueras f. , torra j. , chen b. , 1999 , a&a , 341 , 427 brown , a. , 1950 , , 112 , 225 chen b. , asiain r. , figueras f. , torra j. 1997 , a&a , 318 , 29 chui c. h. , 1992 , wavelet analysis and its application ( new york : academic ) daubechies i. , 1988 , pure appl . math .
41 , 909 dehnen , w. , 1998 , , 115 , 2384 eggen o.j . , 1971 , , 83 , 251 eggen o.j . , 1991 , , 102
, 2028 eggen o.j .
, 1992a , , 103 , 1302 eggen o.j . , 1992b , , 104 , 1482 eggen o.j . , 1992c , , 104 , 1493 eggen o.j . , 1983 , , 88 , 642 eggen o.j . , 1996 , , 112 , 1595 [ 1997)esa97 , esa , the hipparcos catalogue , 1997 , esa sp-1200 famaey b. , jorissen a. , luri x. , mayor m. , udry s. , dejonghe h. and turon c. , arxiv:0411540 famaey b. , jorissen a. , luri x. , mayor m. , udry s. , dejonghe h. and turon c. , 2005 , a&a , 430 , 165 famaey b. , pont f. , luri x. , udry s. , mayor m. , jorissen a. , 2007 , a&a , 461 , 957 famaey b. , siebert a. , jorissen a. , 2008 , a&a , 483 , 453 fux r. , 2001 , a&a , 373 , 511 girardi l. , bressan a. , bertelli g. and chiosi c. , a&as , 141 , gomez a.e . , delhaye j.,grenier s. , jaschek c. , arenou f. , jaschek m. , 1990 , a&a , 236,95 grebel g. k. , roberts w. j. , 1995 , a&as , jones , d. h. p. , 1971
, , 152 , 231 klement r. , fuchs b. , and rix h. w. , 2008 , , 685 , 261 nordstrm b. , mayor m. , andersen j. , holmberg j. , pont f. , j@xmath20rgensen b. r. , olsen e. h. , udry s. and mowlavi n. , 2004 , a&a , 418 , 989 paulson d. b. , sneden c. , cochran w. d. , 2003 , , 125 , 3185 proctor , r. a. , 1869 , proceedings of the royal society of london , 18 , 169 ruskai m. b. , beylkin g. , coifman , r. daubechies i. , mallat s. , meyer y. , raphael l. , 1992 , wavelets and their applications ( boston : jones bartlett ) skuljan j. , hearnshaw j. b. , cottrell p. l. , 1999 , , 308,731 taylor b.j . , 1999 ,
aas , 159 , 100 taylor b.j . , &
honer m.d . , 2005 , , 159 , 100 cccccccccccccc 1&-38,-18,-10&6,6,10&100%&-38,-17,-11&6,6,12&100%&-0.09&0.17&0.1 - 0.2&hyades + 2&-12,-23,-10&6,6,10&100%&-15,-23,-10&6,6,12&100%&-0.17&0.17&0.1 - 0.2&pleiades + 3&-32,-48,-15&5,5,12&100%&-35,-51,-11&5,5,15&100%&-0.16&0.20&0.1 - 0.2&hercules + 4&10,3,-11&6,6,10&100%&10,3,-14&6,6,13&100%&-0.21&0.15&0.0 - 0.1&sirius - uma + 5&-11,-7,-12&5,5,10&100%&-13,-6,-10&5,5,15&97%&-0.20&0.17&0.0 - 0.1&coma(or castor ) + 6&38,-20,-15&4,4,12&100%&37,-25,-12&4,4,14&98%&-0.24&0.21&0.1 - 0.2 & a08(new feature ) + 7&-57,-45,-16&5,5,13&100%&-55,-50,-16&5,5,14&98%&-0.22&0.29&0.2 - 0.3&f08(f.3 green ) + 8&57,-5,-10&3,3,12&93%&60,-5,-10&5,5,18&87% & -0.28&0.18&0.2 - 0.3 & + 9&-56,7,1&3,3,21&97%&-50,4,3&3,3,20&90%&-0.28&0.17&0.1 - 0.2 & + 10&30,-5,-11&3,3,6&87%&28,0,-11&6,6,5&86%&-0.25&0.13&0.2 - 0.3 & + 11&&&&65,-20,-7&5,8,16&63%&&&0.2 - 0.3 & + 12&-53,-19,-12&3,3,21&92%&-50,-20,-18&3,3,10&93%&-0.20&0.24&0.2 - 0.3 & + 13&&&&20,-50,-5&3,3,22&82%&&&0.1 - 0.2 & + 14&-16,-15,-12&4,3,10&90%&-18,-15,-16&4,3,5&91%&-0.13&0.10&0.0 - 0.1&ic 2391 + 15&9,-15,-9&5,6,12&97%&&&&-0.09&0.22&0.0 - 0.1 & + 16&42,8,-29&3,3,15&92%&&&&-0.20&0.22&0.2 - 0.3 & + 17&&&&55,-50,10&4,4,2&86%&&&0.2 - 0.3 & + 18&&&&20,-20,-15&5,5,4&90%&&&0.1 - 0.2&a08(t3.no7 ) + 19&&&&8,-27,-8&5,5,13,2&93% & & & 0.2 - 0.3 & + 20&&&&-75,-46,-13&5,5,11&91% & & & 0.1 - 0.2&a08(t3.no14 ) + 21&&&&-17,-49,-10&5,5,13&82% & & & 0.2 - 0.3 & + 22&&&&-69,-2,-10&5,5,22&90% & & & 0.1 - 0.2&a08(t3.no13 ) + ccccccccc 1&-40,-20&-40.4,-16.0&-42,-18&&-37,-17&-35,-18&-25,-15@xmath21&tab.3 no2 + 2&-12,-22&-11.6,-20.7&-13,-19&&-15,-25&-16,-23&-25,-15@xmath21&tab.3 no1 + 3&&-30.0,-50.0&-35,-45&-42,-51&-30,-50&-35,-51&-20,-50&tab.3 no16 + 4&9,3&14.9,1.4&9,3&6.5,3.9&10,-5&5,1.5&6,4&tab.3 no3 + 5&-10,-5&&-3,-4&&-10,-10&&&tab.3 no4 + 6&&&&&&&&35,-20 + 7&&&&&&-55,-51 & & + 8 & & & & & & & & + 9 & & & & & & & & + 10 & & & & & & & & + 11 & & & & & & & + 12 & & & & & & & & + 13&15,-60&5.8,-59.6&&&&&&tab.3 no6 + 14&&-20.8,-15.9&&&&&&tab.3 no12 + 15 & & & & & & & & + 16 & & & & & & & & + 17 & & & & & & & & + 18&20,-20&&&&&&&tab.3 no7 + 19 & & & & & & & & + 20&-70,-50&&&&&&&tab.3 no14 + 21 & & & & & & & & + 22&-70,-10&&&&&&&tab.3 no13 + | based on kernel estimator and wavelet technique , we have identified 22 moving group candidates in the solar neighborhood from a sample which includes around 14000 dwarfs and 6000 giants .
six of them were previously known as the hercules stream , the sirus - uma stream , the hyades stream , the caster group , the pleiades stream and the ic 2391 ; five of them have also been reported by other authors . eleven moving group candidates ,
not previously reported in the literature , show prominent structures in dwarf or giant samples .
a catalog of moving group candidates in the solar neighborhood is presented in this work . |
a record is an entry in a discrete time series that is larger ( _ upper record _ ) or smaller ( _ lower record _ ) than all previous entries .
thus , records are extreme values that are defined not relative to a fixed threshold , but relative to all preceding events that have occurred since the beginning of the process .
statistical data in areas like meteorology @xcite , hydrology @xcite and athletics @xcite are naturally represented in terms of records .
records play an important role in the public perception of issues like anthropogenic climate change and natural disasters such as floods and earthquakes , and they are an integral part of popular culture . indeed , the _ guinness book of records _ , first published in 1955 , is the world s most sold copyrighted book .
the mathematical theory of records was initiated more than 50 years ago @xcite , and it is now a mature subfield of probability theory and statistics ; see @xcite for reviews and @xcite for an elementary introduction .
most of this work has been devoted to the case when the time series under consideration consists of independent , identically distributed ( i.i.d . )
random variables ( rv s ) . for the following discussion
, it will be useful to distinguish between the _ record times _ at which the current record is broken and replaced by a new one , and the associated _ record values_. one of the key results of record theory is that the statistical properties of record times for real - valued i.i.d .
rv s are completely independent of the underlying distribution .
to illustrate the origin of this universality , we recall the basic observation that the probability @xmath5 for a record to occur in the @xmath0th time step ( the _ record rate _ ) is given by @xmath6 for i.i.d .
rv s , because each of the @xmath0 first entries @xmath7 , including the last , is equally likely to be the largest or smallest . the mean number of records up to time @xmath0 , @xmath8 , is therefore given by the harmonic series @xmath9 with @xmath10 .
further considerations along the same lines lead to a remarkably complete characterization of record times , which will be briefly reviewed below in section [ sec : iid ] .
the universality of record times can be exploited in statistical tests of the i.i.d .
property of a given sequence of variables , without the need for any hypothesis about the underlying distribution @xcite .
by contrast , distributions of record values fall into three distinct universality classes , which are largely analogous to the well - known asymptotic laws of extreme value statistics for distributions with exponential - like tails ( _ gumbel _ ) , bounded support ( _ weibull _ ) and power law tails ( _ frchet _ ) , respectively @xcite .
the decay of the record rate ( [ rate ] ) with increasing @xmath0 implies that the record breaking events form a non - stationary time series with unusual statistical properties , which will be further discussed below in section [ sec : iid ] . _ record dynamics _
has therefore been proposed as a paradigm for the non - stationary temporal behaviour of diverse complex systems ranging from the low - temperature relaxation of spin glasses to the co - evolution of biological populations @xcite .
in fact , records appear naturally in the theory of biological adaptation , because any evolutionary innovation that successfully spreads in a population must be a record , in the sense that it accomplishes some task encountered by the organism in a way that is superior to all previously existing solutions .
consequently the statistics of records and extremes has been invoked to understand the distribution of fitness increments in adaptive processes @xcite as well as the timing of adaptive events @xcite . in the biological context
the universality of record time statistics is particularly attractive , because genotypical fitness is a somewhat elusive notion that is hard to quantify in terms of explicit probability distributions .
surprisingly few result on record statistics are known that go beyond the standard setting of i.i.d .
rv s , and thus consider correlated and/or non - identically distributed rv s . in the present article we focus exclusively on the latter issue , while maintaining the independence among the entries in the sequence .
a simple example of this type was introduced by yang in an attempt to explain the frequency of occurrence of olympic records , which is much higher than would be expected on the basis of the i.i.d .
theory @xcite . in his model
a specified number of i.i.d .
rv s become available simultaneously in each time step , corresponding , in the athletic context , to a variable ( growing ) population from which the contenders are drawn .
much of the standard theory can be extended to this case @xcite ( see section [ sec : growing ] for a brief review ) .
in particular , one finds that the record rate becomes asymptotically constant for exponentially growing populations .
an application of yang s model to evolutionary searches in the space of genotypic sequences can be found in @xcite . a second line of research has addressed the case of sequences with a linear trend , in which the @xmath0th entry is of the form @xmath11 with i.i.d .
rv s @xmath12 and @xmath13 @xcite . also in this case the record rate becomes asymptotically constant , see section [ sec : growing ] for details .
the effect of trends on the occurrence rate of records is a key issue in the ongoing debate about the observable consequences of global warming @xcite . in this context
it has been pointed out that climate _ variability _ is presumably a more important factor in determining the frequency of extreme events than averages @xcite .
it is therefore of considerable interest to investigate the record statistics of sequences of uncorrelated rv s in which the _ shape _ of the underlying probability distribution changes systematically with time .
to initiate such an investigation is the goal of the present paper . throughout
we assume that the probability density @xmath14 of the @xmath0th entry @xmath15 is of the form @xmath16 where @xmath17 is a fixed normalized distribution and the @xmath18 usually have a power - law time dependence @xmath19 so that @xmath20 ( @xmath21 ) corresponds to a broadening ( sharpening ) distribution .
after a brief review of a few important classic results of the theory of records in section [ sec : classic ] , our new results for non - indentically distributed random variables will be presented in section [ sec : record ] .
we focus on the asymptotic behaviour of the record rate @xmath5 and the mean number of records @xmath8 .
preliminary numerical results for the variance of the number of records are reported in section [ sec : numerics ] , but a more complete characterization of record times and record values is left for future work . finally , some concluding remarks are offered in section [ sec : discussion ] .
given the distributions @xmath22 of the entries @xmath23 in a sequence of independent rv s , the probability @xmath5 that the @xmath0th entry is an upper record is equal to the probability that @xmath24 for all @xmath25 .
hence @xmath26 where @xmath27 is the cumulative distribution of @xmath23 .
similarly the probability that @xmath15 is a lower record reads @xmath28.\ ] ] equations ( [ upper ] ) and ( [ lower ] ) form the basis for most of what follows . for i.i.d .
rv s the integral ( [ upper ] ) can be performed by noting that @xmath29 and @xmath30 , which yields the universal result ( [ rate ] ) .
to arrive at a characterization of the record time process beyond the mean number of records @xmath8 we introduce the _ record indicator variables _ @xmath31 , which take the value @xmath32 iff @xmath15 is a record , and @xmath33 else .
it turns out that the @xmath31 are independent @xcite , and hence they form a bernoulli process with success probability @xmath5 . to see why this is so , consider the two - point correlation function @xmath34 and assume that @xmath35 .
then the key idea is that the right hand side of @xmath36\ ] ] can be split into independent events according to @xmath37 \times { \mathrm{prob}}[x_j = \max(x_{i+1}, ... ,x_j ) ] \times \nonumber \\ \times { \mathrm{prob}}[\max(x_1, ... ,x_i ) < \max(x_{i+1}, ... ,x_j)].\end{aligned}\ ] ] following the symmetry argument used to derive ( [ rate ] ) , the first two factors are @xmath38 and @xmath39 , respectively , and the third factor can be written as @xmath40 = \frac{j - i}{j}.\ ] ] we conclude that @xmath41 higher order correlations can be shown to factorize in the same way .
the number @xmath42 of records up to time @xmath0 can then be expressed in terms of the indicator variables as @xmath43 and the variance of @xmath42 is @xmath44 for @xmath45 .
the _ index of dispersion _
@xmath46 of the record time process @xcite , defined as the ratio of the variance to the mean @xmath47 thus tends to unity , and the distribution of the @xmath42 becomes poissonian with mean @xmath1 for large @xmath0 .
the record times form a _ log - poisson _
process @xcite .
a second useful observation concerns the ratios between consecutive record times .
let @xmath48 denote the time of the @xmath49th record , with @xmath50 by convention . repeating the symmetry argument used to derive ( [ rate ] )
, we expect that given @xmath48 , the preceding @xmath51th record occurs with equal probability anywhere in the interval @xmath52 $ ] .
this is not quite correct , because the previous @xmath53 records also have to be accomodated , but since @xmath54 this is a small correction which can be neglected for large @xmath49 .
it is therefore plausible ( and can be proved @xcite ) that the ratio @xmath55 tends to a uniformly distributed rv @xmath56 $ ] for large @xmath49 .
moreover the @xmath57 become independent in this limit @xcite .
this allows us to highlight a peculiar property of the sequence of record breaking events : the expected value of @xmath58 , given @xmath48 , is @xmath59 but the reverse conditioning yields an _ infinite _ expectation , because @xmath60 in this sense , the occurrence of records can be predicted only with hindsight , but not forward in time . in the model for growing populations introduced by yang @xcite and elaborated by nevzorov @xcite , a number @xmath61 of of i.i.d .
rv s becomes available simultaneously at time @xmath0 .
the symmetry argument in section [ sec : intro ] is easily extended to this case : because of the i.i.d .
property , the probability that there is a record among the @xmath61 newly generated rv s is equal to the ratio of @xmath61 to the total number of rv s that have appeared up to time @xmath0 , and hence @xmath62 the independence of the record indicator variables @xmath12 introduced above in section [ sec : iid ] continues to hold @xcite , so again the sequence of record breaking events is a bernouilli process with success probability @xmath5 . to give a simple example for the consequences of ( [ rate_yang ] ) , suppose the @xmath61 grow exponentially as as @xmath63 with @xmath64 .
this could model a sequence of athletic competitions in an exponentially growing population , where each athlete is assumed to be able to participate only in one event @xcite .
then the evaluation of ( [ rate_yang ] ) yields @xmath65 and the distribution of inter - record times @xmath66 is geometric . in his analysis of olympic records yang estimated a growth factor of @xmath67 for the four - year period between two games , and concluded that this growth rate was insufficient to explain the observed high frequency of records . motivated by this outcome , ballerini and resnick @xcite considered a model of _ improving _ populations , where the sequence of rv s displays a linear drift according to ( [ trend ] ) .
they showed that the record rate tends to an asymptotic limit @xmath68 given by @xmath69 where @xmath70 is the probability density of the i.i.d .
rv s @xmath71 in ( [ trend ] ) and @xmath72 = \lim_{n \to \infty } \prod_{k=1}^{n-1 } q(y + ck),\ ] ] with @xmath73 .
the function @xmath68 has the obvious limits @xmath74 and @xmath75 , but the explicit evaluation is generally difficult .
a simple expression is obtained when @xmath76 is of gumbel form , @xmath77 $ ] , which yields @xmath78 . for further details on the model ( [ trend ] ) and applications to athletic data
we refer to @xcite .
results for specific distributions and an application to global warming can be found in @xcite .
in this section we want to evaluate the record rates ( [ upper ] ) and ( [ lower ] ) for distributions of the general form ( [ shape ] ) .
introducing the cumulative distribution corresponding to @xmath17 , @xmath79 the record rates of interest can be written as @xmath80 and @xmath81,\ ] ] which makes clear the obvious fact that the overall scale of the @xmath18 s is without importance . in some special cases
the record rates can be evaluated exactly for arbitrary choices of the @xmath18 s .
for example , for the exponential distribution @xmath82 we have @xmath83 , and the evaluation of the lower record rate ( [ lower1 ] ) yields @xmath84 inserting the power law behaviour ( [ power ] ) we see that the denominator converges to the riemann zeta function @xmath85 for @xmath86 , so that @xmath87 for large @xmath0 , and the expected number of lower records @xmath88 remains finite for @xmath45 . for @xmath89 we have instead that @xmath90 for large @xmath0 , and hence @xmath91 asymptotically .
as would be intuitively expected , the occurrence of lower records is enhanced for sharpening distributions ( @xmath21 ) and suppressed for broadening distributions ( @xmath20 ) .
finally , in the borderline case @xmath92 we find @xmath93 which is our first example of a nontrivial asymptotic law that differs qualitatively from the i.i.d . result ( [ harmonic ] ) .
a simple explicit expression for the upper record rate @xmath5 can be obtained for the uniform distribution characterized by @xmath94 when the @xmath95 are increasing , in the sense that @xmath96 for all @xmath25 , i.e. for the case of a sharpening uniform distribution .
then the arguments of @xmath97 on the right hand side of ( [ upper1 ] ) are all less than unity , and direct integration yields @xmath98 inserting the power law form ( [ power ] ) with @xmath21 one finds that the record rate decays exponentially as @xmath99 , and hence the asymptotic number of records is finite for all @xmath21 . in this section
we focus on broadening distributions , @xmath20 , and evaluate the upper record rate ( [ upper1 ] ) asymptotically for representatives of all three universality classes of extreme value statistics .
the starting point is to replace the product on the right hand side of ( [ upper1 ] ) by the exponential of a sum of logarithms , and to replace the latter by an integral .
it then follows that the asymptotic behaviour of the record rate is given by @xmath100 the second representation will prove to be useful in the final evaluation of @xmath5 . note that @xmath101 can always be expressed in terms of @xmath97 because @xmath102 .
the function @xmath103 is given by @xmath104 where in the second step it has been used that the integral in ( [ asym ] ) is dominated for large @xmath0 by the region where @xmath105 and @xmath106 .
it is therefore clear that the asymptotic behaviour of the record rate depends only on the tail of @xmath97 , and hence universality in the sense of standard extreme value statistics should apply .
the evaluation of ( [ galpha ] ) is straightforward for the frchet class of distributions with power law tails .
we set @xmath107 and obtain @xmath108 inserting this into ( [ asym ] ) yields @xmath109 for large @xmath0 , and hence @xmath110 this result remains valid for negative @xmath111 as long as @xmath112 .
when @xmath113 the evaluation of @xmath103 shows that @xmath114 and thus the asymptotic number of records remains finite .
the gumbel class comprises unbounded distributions whose tail decays faster than a power law @xcite .
a typical representative is the exponentical distribution ( [ exponential ] ) with @xmath115 .
evaluation of ( [ galpha ] ) yields @xmath116 where @xmath117 denotes the incomplete gamma function . for large @xmath101
we have @xcite @xmath118 so that @xmath119 which yields @xmath120 = \int_0 ^ 1 dv \ ; \exp[-n v/(\alpha \ln(1/v))].\ ] ] to further evaluate the integral we substitute @xmath121 and obtain @xmath122 \to \nonumber \\
\to \frac{\ln n}{n } \int_0^\infty dw \ ; e^{-w/\alpha } = \frac{\alpha \ln n}{n}\end{aligned}\ ] ] for @xmath45 .
correspondingly the mean number of records grows as @xmath123 a second important representative of the gumbel class is the gaussian ( normal ) distribution , for which @xmath124 proceeding as before , we find @xmath125 which becomes identical to ( [ galpha_exp ] ) upon replacing @xmath111 by @xmath126 .
we conclude that @xmath127 and @xmath128 for the gaussian case .
although this does not constitute a strict proof , it strongly indicates that the behaviour @xmath129 is _ universal _ within this class of probability distributions . as a representative of the weibull class of distributions with finite support we first consider the uniform distribution ( [ uniform ] ) .
the integral on the right hand side of ( [ galpha ] ) can then be evaluated without approximating @xmath130 by @xmath131 , and one obtains @xmath132 this is a negative monotonically increasing function which vanishes quadratically in @xmath133 near @xmath134 , @xmath135 the evaluation of the record rate ( [ asym ] ) then yields @xmath136 \approx \sqrt { \frac{\alpha \pi}{2 n}}\ ] ] for large @xmath0 , and the number of records grows asymptotically as @xmath137 . the specific power is clearly related to the quadratic behaviour of @xmath103 near @xmath134 , which in turn reflects the behaviour of @xmath138 near the upper boundary @xmath139 .
more generally we may consider bounded distributions of the form @xmath140 with @xmath141 and the uniform case corresponding to @xmath142 . to extract the leading order behaviour of @xmath103 for @xmath143 we write @xmath144 for @xmath143 .
hence the record rate decays as @xmath145 and the mean number of records grows as @xmath146 the asymptotic laws ( [ r_frechet ] , [ mean_exp ] , [ weibull_mean ] ) were first discovered in simulations , and they have subsequently been numerically verified for a variety of parameter values . as an example , we show in figure [ figure1 ] numerical data for the mean number of records obtained for distributions in the gumbel class .
there are significant corrections to the asymptotic behaviour for the gaussian distribution as well as for the exponentical distribution with @xmath147 .
this is not surprising in view of the approximations used in the derivation of ( [ mean_exp ] ) ; for example , the last step in ( [ pn_exp2 ] ) requires that @xmath148 which is true only for enormously large values of @xmath0 . ,
@xmath92 and @xmath149 .
the dashed line shows data obtained for the gaussian distribution and @xmath92 .
the thin dotted line is the harmonic series ( [ harmonic ] ) which applies universally for @xmath150 .
the short bold dotted lines show the predicted slope @xmath151 for the exponential case and @xmath111 in the gaussian case .
all data were obtained from @xmath152 realizations of time series of length @xmath153.,scaledwidth=70.0% ] simulations have also been used to investigate the occurrence of correlations in the record time process for @xmath20 .
we have seen in section [ sec : iid ] that the poisson statistics of @xmath42 is a consequence of the fact that the record indicator variables @xmath31 are independent in the i.i.d . case . in particular , ( [ rn_variance ] ) shows that the variance of @xmath42 is asympotically equal to the mean whenever the @xmath31 are uncorrelated and the record rate @xmath5 tends to zero for @xmath45 in such a way that @xmath8 diverges .
as this is true for @xmath20 in all cases that we have considered , the index of dispersion ( [ rho ] ) can be used as a probe for correlations .
the data displayed in figure [ figure2 ] clearly show that the asymptotic value of @xmath46 is less than unity and independent of @xmath111 for the uniform distribution .
similar results have been obtained for the exponential distribution , whereas we find that @xmath154 for the power law case .
we conclude that , at least in certain cases , the record time process becomes more regular than the log - poisson process when the underlying distribution broadens with time .
, 1/2 , 1 and 2 . while the data for @xmath150 approach the asymptotic poisson limit of unity according to ( [ rn_variance ] ) , the data for @xmath20 converge to a universal sub - poissonian value .
the data were obtained from @xmath153 realizations of time series of length @xmath152.,scaledwidth=70.0% ]
the main results of this paper are the asymptotic laws ( [ r_frechet ] , [ mean_exp ] , [ weibull_mean ] ) for the mean number of records in sequences of random variables drawn from broadening distributions . in all three cases
the exponent @xmath111 governing the time dependence of the width of the distribution enters only in the prefactors and does not affect the functional form of the result . comparing the three cases , we see that the effect of the broadening on @xmath8 is stronger the faster the underlying distribution @xmath17 decays for large arguments : for fat - tailed power law distributions the number of records remains logarithmic , for exponential - like distributions it changes from @xmath1 to @xmath2 , while for distributions with bounded support the logarithm speeds up to a power law in time . apart from the presentation of new results , a secondary purpose of this paper has been to advertise record dynamics as a paradigm of non - stationary point processes with interesting mathematical properties and wide - spread applications ranging from fundamental issues in the dynamics of complex systems to the consequences of climatic change . in the present work we have combined the intrinsic non - stationarity of record dynamics with an explicit non - stationarity of the underlying sequence of random variables .
this turns out to be a relevant modification which may alter the basic logarithmic time - dependence of the mean number of records , and it can induce correlations among the record times , as detected in deviations of the index of dispersion ( [ rho ] ) from unity . it is worth noting that evidence for such correlations can also be found in recent applications of record dynamics in simulations of complex systems @xcite .
an analytic understanding of the origin of correlations in the models presented here is clearly an important goal for the near future .
i am grateful to kavita jain for her contributions in the early stages of this project , and to sid redner for useful correspondence and discussions .
this work was supported by dfg within sfb - tr 12 _ symmetries and universality in mesoscopic systems_.
10 hoyt d v 1981 _ climatic change _ * 3 * 243 bassett jr .
gw 1992 _ climatic change _ * 21 * 303 benestad re 2003 , _ climate research _ * 25 * 3 redner s and petersen mr 2006 _ phys .
e _ * 74 * 061114 matalas nc 1997 _ climatic change _ * 37 * 89 vogel rm , zafirakou - koulouris a and matalas nc 2001 _ water res .
research _ * 37 * 1723 gembris d , taylor jg and suter d ( 2002 ) _ nature _ * 417 * 506 chandler kn 1952 _ j. roy . stat
b _ * 14 * 220 glick n 1978 _ amer . math . monthly _ * 85 * 2 nevzorov vb 1987 _ theory probab . appl . _
* 32 * 201 arnold bc , balakrishnan n and nagaraja hn 1998 _ records _
( new york : wiley ) nevzorov vb 2001 _ records : mathematical theory _
( providence : american mathematical society ) schmittmann b and zia rkp 1999 _ am . j. phys . _
* 67 * 1269 galambos j 1987 _ the asymptotic theory of extreme order statistics _ ( malabar : r.e .
krieger ) sornette d 2000 _ critical phenomena in natural sciences _
( berlin : springer ) sibani p and littlewood p 1993 _ phys .
lett . _ * 71 * 1482 sibani p and dall j 2003 _ europhys .
lett . _ * 64 * 8 anderson pe , jensen hj , oliveira lp and sibani p 2004 _ complexity _ * 10 * 49 gillespie jh 1991 _ the causes of molecular evolution _ ( new york : oxford university press ) orr ha 2005 _ nature rev
* 6 * 119 kauffman sa and levin s 1987 _ j. theor .
biol . _ * 128 * 11 sibani p , brandt m and alstrm p 1991 _ int . j. mod
b _ * 12 * 361 krug j and karl c 2003 _ physica a _ * 318 * 137 krug j and jain k 2005 _ physica a _ * 358 * 1 jain k and krug j 2005 _ j. stat .
mech . : theory and experiment _
p04008 sire c , majumdar sn and dean ds 2006 _ j. stat .
mech . : theory and experiment _
l07001 yang mck 1975 _ j. appl .
* 12 * , 148 ballerini r and resnick s 1985 _ j. appl .
* 22 * 487 ballerini r and resnick s 1987 _ adv .
_ * 19 * 801 borovkov k 1999 _ j. appl .
prob . _ * 36 * 669 katz rw and brown bg 1992 _ climatic change _ * 21 * 289 cox dr and isham v 1980 _ point processes _ ( london : chapman and hall ) tata mn 1969 _ z. warsch .
* 12 * 9 shorrock rw 1972 _ j. appl .
prob . _ * 9 * 316 gradshteyn is and ryzhik i m 2000 _ table of integrals , series and products _
( san diego : academic press ) | in the context of this paper , a record is an entry in a sequence of random variables ( rv s ) that is larger or smaller than all previous entries . after a brief review of the classic theory of records , which is largely restricted to sequences of independent and identically distributed ( i.i.d . )
rv s , new results for sequences of independent rv s with distributions that broaden or sharpen with time are presented . in particular , we show that when the width of the distribution grows as a power law in time @xmath0 , the mean number of records is asymptotically of order @xmath1 for distributions with a power law tail ( the _ frchet class _ of extremal value statistics ) , of order @xmath2 for distributions of exponential type ( _ gumbel class _ ) , and of order @xmath3 for distributions of bounded support ( _ weibull class _ ) , where the exponent @xmath4 describes the behaviour of the distribution at the upper ( or lower ) boundary .
simulations are presented which indicate that , in contrast to the i.i.d .
case , the sequence of record breaking events is correlated in such a way that the variance of the number of records is asymptotically smaller than the mean . |
the x - ray source ax j1420.1@xmath06049 is one of two extended , hard spectrum sources whose positions are consistent with the @xmath1ray source gev j1417@xmath06100/3eg j1420@xmath06038 @xcite .
radio images of the region from the molonglo synthesis telescope ( most ) galactic plane survey @xcite and the australia telescope compact array ( atca ) ( roberts et al .
1999 , hereafter r99 ) reveal a complex of compact and extended radio sources .
this complex , known as the kookaburra , consists of a shell with two wings " stretching to the northeast and the southwest .
there is also a bright hii region and several compact extra - galactic sources in the field .
ax j1420.1@xmath06049 is coincident with a slight enhancement ( k3 of r99 ) in the northeast wing .
low resolution @xmath2 60 @xmath3 m imaging of this region shows evidence of thermal emission from the shell but shows a lack of emission from the wings .
however , the low resolution and artifacts from the hii region prevent strong constraints to be derived from this image . the paucity of infrared emission along with the spectral index of the wing containing ax j1420.1@xmath06049 ,
although poorly constrained , hint at a non - thermal origin for the radio emission . the 68.2 ms pulsar psr j1420@xmath06048 was recently discovered during the parkes multibeam survey at a position consistent with ax j1420.1@xmath06049 @xcite .
timing measurements have determined a characteristic age of @xmath4 yr , and a very high inferred spin - down energy @xmath5 ergs / s . this energy output is higher than all the known @xmath6-ray pulsars except for the crab , and is similar to the recently discovered pulsar and pulsar wind nebula ( pwn ) psr j2229 + 6114 , the likely counterpart of the @xmath1ray source 3eg j2227 + 6122 @xcite .
psr j1420@xmath06048 has been monitored since 1998 october , and was observed to glitch in late 2000 . in this letter , we report on x - ray spectral and timing analyses of a 1999 asca observation of ax j1420.1@xmath06049 , sensitive radio pulse profile and polarization measurements made with the parkes radio telescope , high - resolution 20 cm radio imaging with the atca , and high resolution infrared imaging at 60 @xmath3 m and 8.3 @xmath3 m .
on feb . 13 - 15 , 1999 , asca re - observed gev j1417@xmath06100 centered on the two x - ray nebulae . approximately 63 ks ( @xmath7ks with high - bit rate telemetry ) of exposure was obtained with the gas imaging spectrometers ( gis2,gis3 ) in standard ph mode with no timing bits .
image and spectral analyses of the field have previously been reported in @xcite , and we use similar reduction and analysis procedures here . since the source fell near the edge of the sis detectors , we restrict our analysis to the gis data .
ax j1420.1@xmath06049 is seen by asca to be significantly extended .
however , there is a central concentration in ax j1420.1@xmath06049 whose centroided position is @xmath8 from the radio position of the pulsar ( see below ) .
this is larger than the nominal 90% position uncertainty @xcite , but is consistent with the 97% contour ; we assume the peak of the x - ray emission is at the position of the pulsar in the following analsysis . in general , timing solutions for young , glitching
, pulsars such as psr j1420@xmath06048 can not be reliably extrapolated back to previous observation epochs .
therefore , timing studies of the earlier asca dataset ( 1997 epoch ) and the @xmath9 data are problematic .
fortunately , contemperaneous radio data is available for the 1999 asca observations @xcite .
these were observed in default observing mode which has a nominal time resolution of @xmath10 .
however , the reason for such a large nominal uncertainty in arrival time is a timing ambiguity which is only important for sources with extremely high count rates ( @xmath11 cps ) . for low count rate sources such as psr j1420@xmath06048 ,
the time assignment precision of the gis is equal to the telemetry output period ( @xmath12 ms for high bit - rate data , @xmath13 ms for medium bit - rate , hirayama et al .
1996 ) . using the high bit - rate gis data set , we extracted 1 - 10 kev photons from a 6 pixel radius region ( @xmath14 , roughly corresponding to the half - power radius of the asca psf ) centered on the position of the x - ray peak .
the small extraction region was chosen to minimize the photon contribution from the surrounding nebula .
the resulting 314 photon arrival times were then corrected to the solar system barycentre . using the radio period and period derivative to determine arrival time phases ,
an h - test analysis was performed to determine the significance of any periodicity @xcite resulting in a test statistic value of 12.95 .
the formal chance probablity for a value this large is 0.0056 , which we verified by examining the h - test values of @xmath15 independent nearby periods on the same dataset .
the telemetry times used by asca are delayed from the true arrival time by an offset of 8.5 ms @xmath16 , where @xmath17 has an approximately uniform probability of being any value between @xmath18 s and @xmath19 s ( @xmath20 s for the medium bit - rate data ) @xcite . to extract the best folded light curve , we take into account this delay as well as the arbitrary placement of bin boundaries by assigning a fractional value to adjacent bins according to the probability of a photon belonging in each .
the result is shown in figure [ pulse ] .
the absolute phasing of this method was verified with an offset observation of the crab pulsar taken in the same mode and folded using the jodrell bank ephemeris . for psr j1420 - 6048 , the x - ray pulse ranges in phase from @xmath21 referenced to the radio pulse maximum . a similar analysis with the medium bit - rate data resulted in a profile consistent in amplitude and phase with a highly smoothed version of the high bit - rate data .
the formal h - test chance probability of the medium bit - rate data is only 0.14 .
the above phase window was used to make on - pulse and off - pulse 1 - 10 kev images from the high bit - rate data .
the off - pulse image was scaled and subtracted from the on - pulse image to create a pulsed - only image , which was then exposure corrected and smoothed with a 1.7 pixel gaussian ( roughly the width of the gis psf peak ) . a particle background image made from the night earth calibration files
was subtracted from the off - pulse image , which was also exposure corrected and smoothed .
the resulting pulsed - only and off - pulse images are presented in figure [ onoff ] .
a radial profile comparison of these images to point - source calibration images for nearby positions on the detector , ( one such is added as a reference to the images in figure [ onoff ] ) show the pulsed - only profile to be consistent with that of the calibration source while the off - pulse profile is virtually flat out to @xmath22 and then slowly falls off .
this provides additional support to the pulse detection . the overall x - ray flux for the entire data set of high and medium bit - rate data is adequately fit by a power - law with @xmath23 with significant absorption ( @xmath24 ) and a 2 - 10 kev flux of @xmath25 .
weak evidence for an fe k line ( ew@xmath26 ev ) suggests that some thermal emission may also be present in the region . in order to estimate the magnitude of the apparent pulsed flux , a @xmath27 extraction radius was used to create on - pulse and off - pulse spectra from the high bit - rate data using a slightly larger phase range of 0.575 - 0.775 to maximize inclusion of the pulsed emission
this extraction region is roughly twice the point spread function full width half max , but is smaller than the nominal @xmath28 in order to limit noise due to photons from the surrounding nebula .
using the off - pulse spectrum as a background , and fixing the @xmath29 at the value for ax j1420.1@xmath06049 as a whole , the resulting spectrum is adequately fit by either a power law or blackbody , with a 1 - 10 kev flux of roughly @xmath30 ergs @xmath31 .
psr j1420@xmath06048 was observed on may 20 , 2001 with the parkes 64 m radio telescope as part of a program of high time - resolution observations of vela - like pulsars @xcite . the central feed of the 13 beam , 21 cm multi - beam system was used .
the signal was split in two and sent to two different back - ends .
one was a dual polarization , 512 channel filterbank receiver with a central observing frequency of 1517.5 mhz and a 256 mhz bandwidth , where the two polarizations were summed , fast sampled , one - bit digitized and written to storage tape for off - line de - dispersion and pulse folding .
this was used for the high - resolution profile .
the second back - end was a 128 mhz bandwidth correlator with somewhat lower frequency resolution , but which retained the full polarization information and was used for the polarization analysis . a 4 hour integration on psr j1420@xmath06048 folded at the contemporary radio ephemeris produced the high resolution pulse profile in figure 1 .
the pulsar has a broad profile ( @xmath32 = 16 ms at 10% of the peak flux ) with two well separated components .
we have set radio phase 0 ( the maximum of the second peak ) to the phase of the radio pulse at the ( pre - glitch ) ephemeris used to fold the x - ray data .
figure [ pks ] shows the polarization profiles of the radio peaks at @xmath33ms effective time resolution .
the pulse shows a high fractional linear polarization ( @xmath34 ) , with a smooth position angle swing .
high linear polarization appears to correlate with high spindown luminosity @xmath35 @xcite and psr j1420@xmath06048 supports this trend .
a simple @xcite rotating vector model fit to the position angle produces an adequate fit to the sweep for small impact parameters ( @xmath36 ) and small magnetic inclinations @xmath37 , although the latter angle is relatively poorly determined .
we have also extracted the rotation measure @xmath38 , which together with the previously measured dispersion measure gives a mean line - of - sight magnetic field strength of @xmath39 , pointed away from the observer .
the pulsar also shows very strong ( right handed ) circular polarization , particularly in the second peak where the circular polarization reaches @xmath40 .
this value is among the highest recorded ( han et al .
1998 ) , certainly the highest among pulsars classified as ` conal doubles ' . the right - handed circular polarization and increasing position angle swing
follow the correlation between handedness and sense of position angle change found by han et al .
( 1998 ) for cone - dominated pulsars . with no reversal of sign ,
the phase averaged circular polarization fraction is quite large at 23% .
r99 presented results of radio continuum observations at 20 cm and 13 cm of the kookaburra complex with the atca in three antenna configurations ( 1.5a , 0.375 , and 0.75a configurations on 11 jan .
, 28 mar . , and 22 apr .
, respectively . )
three additional 20 cm continuum observations have since been performed ( 6d , 0.375 , and 0.75c configurations on 20 aug .
1999 , 13 jan .
2001 , and 03 mar .
all these data were reduced and analysed with the miriad software package @xcite . a high resolution ( @xmath41 beam , 1375 mhz mean frequency ) image was produced by combining all of the 20 cm data .
figure [ hires ] shows the upper wing of the kookaburra which we name g313.6 + 0.3 ( k2 of r99 ) .
the pulsar is clearly seen as a point source near the southeast corner at r.a.(j2000)=14h 20 m 8.23(6)s , dec.(j2000)=@xmath42 @xmath43 @xmath44 , consistent with the timing position of psr j1420@xmath06048 @xcite .
stokes v parameter imaging finds a circular polarization fraction consistent with the time averaged value from the parkes data . using the dual - frequency 1998 data and only the longest ( @xmath45 km ) baselines to resolve out the continuum flux , we obtain pulse average fluxes of @xmath46 mjy ( 20 cm ) and @xmath46mjy ( 13 cm ) for a net spectral index @xmath47 .
@xcite have used a low ratio of infrared to radio flux as an indicator of non - thermal emission in order to identify supernova remnant candidates .
we have obtained _ iras _ hi - res @xmath48 m and _ msx _
@xmath49 m images from the nasa / ipac infrared science archive , and compared them to the radio maps .
the shell is clearly seen in infrared , but there is no excess emission associated with g313.6 + 0.3 ( figure [ hires ] contours ) in either image .
@xcite noticed a plausible association between the @xmath1ray source and a complex of young star forming regions and snr at a distance @xmath50 kpc . however , the pulsar dispersion measure ( dm=360 pc @xmath51 , damico et al . 2001 ) indicates a much larger distance @xmath52 kpc @xcite .
the measured x - ray absorption of @xmath53 is consistent with the total galactic @xmath29 towards that direction estimated from the ftools @xmath54 tool derived from the hi map by @xcite .
however , the prevalence of hii regions and dense clouds in this direction may mean the dm and @xmath29 are dominated by local sources in the carina arm , implying a closer distance .
we can compare the x - ray and radio fluxes of our sources with other psr / pwn .
saito ( 1997 ) fits an empirical relation for pulsed asca fluxes of @xmath55 erg / s , where @xmath56 is the spin - down power in units of is @xmath57erg / s .
this matches our observed pulsed flux at @xmath58 kpc .
the radio flux can also be compared to known pwn .
r99 suggested that the k3 site , a @xmath59 mjy enhancement above the general flux in g313.6 + 0.3 , was the radio counterpart to ax j1420.1 - 6049 . assuming a typical radio pwn spectral index @xmath60 , @xcite
note the fraction @xmath61 of the spin - down energy which goes into radio emission is related to the 20 cm flux density by @xmath62 mjy .
this suggests an efficiency @xmath63 at a distance of 8 kpc which is rather low compared to most other detected radio pwn @xcite , and would be even lower if the true distance is closer .
however , the entire wing has a flux of @xmath64 jy , the nature of which is uncertain .
it is possible that a substantial fraction of the flux is due to the pulsar wind .
a pwn flux of @xmath65 mjy would make for a fairly typical value of @xmath66 at 8 kpc , and reasonable values at closer distances .
the implied size of the nebula at 8 kpc , @xmath67 pc , would then be unusually large .
the morphology of g313.6 + 0.3 is curious , being a rectangular plateau of radio emission with a question mark shaped enhancement starting at the pulsar and reaching up to nearly the top , reminiscent of a trailing bow - shock .
the lack of ir emission , along with the apparent excess of polarized flux seen by r99 in portions of the wing and their estimate of @xmath68 for the spectral index , suggest the entire wing may be a supernova remnant .
however , the large scatter in the t - t plot of r99 makes their spectral index value problematic .
the expanded data set confirms the presence of linear polarizaton near the top of the enhancement , but seems to suggest that g313.6 + 0.3 has a mixture of thermal and non - thermal spectral components .
given the rarity of such high @xmath35 pulsars , it is natural to assume a connection with gev j1417@xmath06100/3eg j1420@xmath06038 .
the average @xmath69 mev flux for this source is @xmath70 .
models of @xmath6-ray pulse emission ( e.g. romani 1996 ) allow one to predict the typical @xmath6-ray flux for pulsar spin parameters and the magnetic inclination angle @xmath71 . for psr j1420@xmath06048 ( @xmath72yr , surface magnetic field @xmath73 g ) , an assumed 1 steradian beam , and an efficiency of @xmath74 ( somewhat less than vela ) the observed @xmath69mev flux is expected for @xmath75 if the distance were as close as 2 kpc . at a distance of 8 kpc , the inferred efficiency would need to be significantly higher if the @xmath1ray emission is 100% pulsed .
if @xmath71 is truely less than @xmath76 , the expected @xmath6-ray efficiency is quite small and in an outer gap picture one would not expect the @xmath6-ray beam to cross the line of sight for small @xmath77 . in this case
the @xmath6-ray emission could have an unpulsed , pulsar wind origin .
it is important to note that the @xmath6-ray flux appears variable .
a significant fraction of the emission may therefore be associated with a pulsar wind shock , where instabilities might produce luminosity fluctuations on @xmath78month time scales .
similar variability is seen in other unidentified _ egret _ sources containing pwn @xcite . a second pulsar in the nearby rabbit pwn candidate , unseen in the radio ,
might also contribute to the @xmath1ray flux .
we thank bryan gaensler , vicky kaspi , and the referee eric gotthelf for useful comments .
we also thank froney crawford for the 20 aug .
1999 atca data .
the australia telescope is funded by the commonwealth of australia for operation as a national facility managed by the csiro .
mser is a quebec merit fellow .
this work was supported in part by nasa grants nag5 - 3263 and a cottrell scholars grant from the research corporation .
case , g. & bhattacharya , d. 1999 , , 521 , 246 crawford , f. , manchester , r. n. , & kaspi , v. m. 2001 , astro - ph/0106544 , in press .
damico , n. et al .
2001 , , 552 , l45 de jager , o. c. 1994 , , 436 , 239 dickey , j. m. & lockman , f. j. 1990 , 28 , 215 frail , d. a. & scharringhausen , b. r. 1997 , , 480 , 364 gaensler , b. m. , stappers , b. w. , frail , d. a. , moffett , d. a. , johnston , s. , & chatterjee , s. 2000 , , 318 , 58 gotthelf , e. v. , ueda , y. , fujimoto , r. , kii , t. , & yamaoka , k. 2000 , , 543 , 417 green , a. j. , cram , l. e. , large , m. i. , & ye , t. , 1999 , , 122 , 207 halpern , j. p. , camilo , f. , gotthelf , e. v. , helfand , d. j. , kramer , m. , lyne , a.g . , leighly , k. m. , & eracleous , m. 2001 , , 552 , l125 han , j. l. , manchester , r. n. , xu , r. x. , & qiao , g. j. 1998 , , 300 , 373 hirayama , m. , nagase , f. , gunji , s. , sekimoto , y. , & saito , y. 1996 , _ asca news _ ,
johnston s. & romani r.w 2001 , apj , in preparation radhakrishnan , v. & cooke , d. j. 1969 , , 3 , 225 roberts , m. s. e. & romani , r. w. 1998 , , 496 , 827 roberts , m. s. e. , romani , r. w. , johnston , s. , & green , a. j. 1999 , , 515 , 712 ( r99 ) roberts , m. s. e. , gaensler , b. m. , & romani , r. w. 2001 , in preparation .
roberts , m. s. e. , romani , r. w. , & kawai , n. 2001 , , 133 , 451 romani , r.w .
1996 , , 470 , 469 saito , y. 1997 , ph.d .
thesis univ . of tokyo sault , r.j .
& killeen , n. e. b. 1998 , the miriad user s guide ( sydney : australia telescope national facility ) taylor , j. h. , & cordes , j. m. 1993 , , 88 , 529 whiteoak , j. b. z. , cram , l. e. , & large , m. i. 1994 , , 269 , 294 whiteoak , j.b.z . &
green , a.j .
1996 , , 118 , 329 yadigaroglu , i .- a . &
romani , r.w .
1997 , , 476 , 347 | we present x - ray , radio , and infrared observations of the 68 ms pulsar psr j1420@xmath06048 and its surrounding nebula , a possible counterpart of the @xmath1ray source gev j1417@xmath06100/3eg j1420@xmath06038 .
pulsed x - ray emission at the radio period is marginally detected by asca from a source embedded in the hard spectrum x - ray nebula ax j1420.1@xmath06049 . at radio wavelengths ,
the pulsar is found to be strongly linearly and circularly polarized , and the polarization sweep is measured .
a comparison of high resolution atca radio imaging of the kookaburra s upper wing ( g313.6 + 0.3 ) , which contains the pulsar and the x - ray nebula , with infrared images suggests the radio emission is partly non - thermal . |
electronic devices based on organic semiconductors , such as polymeric field effect transistors ( fets ) and light emitting diodes ( leds ) , have attracted much interest as possible inexpensive and flexible alternatives to inorganic devices@xcite . while there has been considerable improvement in device properties ,
the detailed mechanism of electronic transport in organic thin film devices remains a subject of active investigation .
charge motion in undoped organic field effect devices is often characterized in the linear ( small source - drain bias ) regime by a field - effect mobility . in disordered polymer semiconductors
this mobility , @xmath0 , is strongly temperature dependent near room temperature , consistent with thermally assisted hopping between localized states dispersed throughout the polymer film@xcite .
these localized states are likely to be polaronic@xcite .
several different models for hopping transport in these materials have been used to interpret experimental data from @xmath2200 k to room temperature , including simple thermal activation@xcite , 2d variable range hopping ( vrh)@xcite , and percolative vrh with an exponential density of states@xcite . while some models work better with some specific samples , generally distinguishing between them is difficult .
the situation is complicated by the fact that the effective mobilities inferred in this manner depend on @xmath1 , gated charge density@xcite , source - drain bias , and contact effects .
parasitic contact resistances in particular can be important .
experiments in several organic semiconductor / electrode combinations@xcite have shown that contact resistances can be a significant fraction of the total source - drain resistance in the linear regime in short devices .
data in this shallow channel limit must be corrected accordingly for this parasitic series resistance , @xmath3 , to find the true mobility within the semiconductor . as with @xmath0
, one must bear in mind that @xmath3 typically depends on temperature , gate voltage , and local electric field@xcite .
when examining transport properties of organic fets , it is therefore important to determine whether the devices are dominated by the bulk ( channel ) or the contacts .
we report _ nonlinear _ transport measurements in field - effect devices made from high quality , solution cast , regio - regular poly(3-hexylthiophene ) ( p3ht ) , with channel lengths from 3 @xmath0 m to 200 nm and aspect ratios of 10 .
higher temperature properties at low source - drain fields are consistent with those observed by other investigators . from 200 k to 10 k , we observe gate - modulated nonlinear @xmath4 characteristics .
to understand the effects of contacts in our series of devices , we examined both the devices described above , and an additional series of fixed - width fets to obtain the channel and contact resistances as a function of temperature .
these data demonstrate that ( a ) the fixed - apect - ratio devices are bulk ( not contact ) limited at high temperatures ; and ( b ) the contribution of contacts relative to the channel actually _ decreases _ as the temperature is lowered , so that bulk - limited devices tend to remain so as @xmath1 is decreased . since the nonlinearities in the @xmath4 curves become more pronounced at low temperatures and in wider devices ( for which contact resistance are proportionately less important ) , it is unlikely that these nonlinearities are due to contact effects in these geometries .
we find that the nonlinear data are consistent with a model of poole - frenkel ( pf ) type conduction in the space - charge limited ( scl ) regime . within this framework ,
the strong gate and temperature dependence of this conduction are consistent with a density of localized valence states that varies approximately exponentially in energy .
still within this model , at temperatures below 30 k there appears to be a crossover from thermally assisted hopping to a nonthermal mechanism .
these nonlinear data over a broad temperature range constrain any other models of transport in such devices . studying fets in this nonlinear regime
allows comparisons between models not readily performed with linear transport data .
devices are made in a bottom - contact configuration ( see fig . [
fig : mobt ] inset ) on a degenerately doped @xmath5 silicon wafer to be used as a gate .
the gate dielectric is 200 nm of thermal sio@xmath6 .
source and drain electrodes are patterned using standard electron beam lithography .
the electrodes are deposited by electron beam evaporation of 4 nm ti and 25 nm of au followed by liftoff .
the fixed - aspect - ratio devices have channel lengths ranging from 3 @xmath0 m down to 200 nm , with the channel width scaled to maintain @xmath7 larger fet devices ( @xmath8 1 mm , @xmath9 50 @xmath0 m ) are also prepared for comparison .
we also examine a second set of devices with fixed width @xmath10 m and channel lengths varying from 5 @xmath0 m up to 40 @xmath0 m , specifically for probing contact resistance issues .
the organic semiconductor is 98% regio - regular p3ht@xcite , a well studied material@xcite .
p3ht is known from x - ray scattering to form nanocrystalline domains with sizes on the order of 20 nm@xcite , and more ordered films correlate with higher measured mobilities . rr - p3ht is dissolved in chloroform at a 0.02% weight concentration , and is solution cast@xcite onto ozone - cleaned , choloroform - swabbed substrates .
the fixed - aspect - ratio series of devices are from one casting , while the fixed - width series are from a second casting .
the resulting film thicknesses over the channel region are tens of nm as determined by atomic force microscopy ( afm ) , though casting produces somewhat nonuniform films .
all devices are stored in vacuum dessicators until use .
the measurements are performed in vacuum ( @xmath11 torr ) in a variable - temperature probe station using a semiconductor parameter analyzer ( hp4145b ) .
at room temperature , the devices operate as standard @xmath12-type fets in accumulation mode@xcite . with the source electrode as ground , in the linear regime we extract@xcite an effective mobility from the transconductance .
that is , from data of source - drain current , @xmath13 versus the gate voltage , @xmath14 , at a fixed low drain voltage , @xmath15 , we compute @xmath16 , where @xmath17 is the transconductance , @xmath18 is capacitance per unit area of the gate insulator , and @xmath19 is the aspect ratio .
the relevance of parasitic contact resistances will be addressed below . as is reported elsewhere@xcite ,
the mobility is gate voltage dependent , increasing with increasing @xmath14
. it also increases with increasing source - drain voltage .
effective mobilities are typically between @xmath20 and @xmath21 @xmath22/vs , and apparent threshold voltages ( @xmath23 ) , though not necessarily meaningful@xcite , are low ( @xmath24 v ) . for @xmath25 50
@xmath0 m fets operated in the saturation regime , the on / off ratio is typically @xmath26 , comparing between gate voltages of -95 v and 0 v. as temperature is reduced below 300 k , the off - current drops to undetectable levels by 150 k , as the unintentional carriers ( due to slight doping from air exposure ) freeze out . over the moderate temperature range of 300 k to @xmath2 200 k
, the mobility as inferred above at fixed small @xmath15 is found to depend steeply on temperature .
a small representative set of this data is shown in fig .
[ fig : mobt ] , where @xmath0 is plotted vs. inverse temperature for @xmath27 and -50 v , for the @xmath28 nm , @xmath29 m device at constant source - drain electric field of @xmath30 v / m .
these data are approximately equally consistent with the three models mentioned above : simple arrhenius behavior ( the dashed line ) with an activation energy @xmath2 100 mev ; vrh for a 2-d system ( dotted line ) , of the form @xmath31 , where @xmath32 and @xmath33 are fit parameters ; and finally the more sophisticated percolative vrh theory ( solid lines ) developed by vissenberg _
et al_@xcite .
the vissenberg model s underlying assumptions include an exponential density of ( localized ) states ( dols ) , @xmath34 , with transport of carriers dictated by percolative hopping .
since the gate voltage controls the fermi level in the channel , and hence the occupation of the localized states , one finds that transport in the channel is strongly affected by @xmath14 .
relevant fit parameters @xcite are : @xmath33 , describing the energy dependence of the dols ; a prefactor @xmath35 with units of conductivity ; and @xmath36 , an effective overlap parameter for hopping .
values used in the fits shown are @xmath37 s , @xmath38 418 k , @xmath39 m@xmath40 , consistent with those seen by other investigators in p3ht@xcite .
as temperatures are lowered from 200 k down to 10 k , over a broad range of source - drain and gate voltages , @xmath13 evolves from approximately linear to a strongly nonlinear ( superquadratic ) dependence on @xmath15 .
an example of this evolution is shown in fig .
[ fig : nonliniv ] for the @xmath41 m , @xmath42 m device , comparing data at 300 k and 70 k. we note that , at the lowest temperatures , smaller devices transport current _ more easily _ ( larger currents at smaller gate voltages for a fixed @xmath43 ) than larger devices , as we will discuss later .
analysis below shows that the nonlinear @xmath4 characteristics are decribed well by a model incorporating space - charge limited currents , modified by a poole - frenkel - like exponential dependence of effective mobility on square root of the local electric field ( sclpf ) .
this conduction mechanism has been seen repeatedly in _
two_-terminal devices@xcite .
room temperature experiments@xcite on p3ht fets with 70 nm channel lengths also show indications of scl currents .
we find that within this model , the temperature and gate voltage dependence of the data support a strongly energy dependent dols such as that in the vissenberg picture .
other models may be possible , but they are constrained by the dependences presented below .
charge transport in a device is space charge limited if the injected carriers significantly alter the local electric field from the average field imposed by the electrode potentials , and correspondingly limit the current .
if , instead , the bottleneck in charge transport is injection at the contacts , a device is said to be contact limited , and is expected to exhibit ohmic behavior at low source - drain fields . in principle , modeling our devices requires the full solution of the steady state charge and electric field profile in a three - terminal accumulation mode transistor , including field- and temperature dependent effective mobility , and field- and temperature dependent contact properties .
this general problem is very complex@xcite ; here we consider a simpler model and compare with the transport data . in a system that is
not contact limited , when the effective mobility varies as a function of electric field @xmath44 , the space - charge limited current in a _ two - terminal _ device ( a 1d model ) is determined by the numerical solution of @xmath45 where @xmath46 is the local hole density , @xmath47 , and the appropriate boundary condition on @xmath48 . for approximately ohmic contacts , @xmath49 .
we note that the case of an effective mobility that varies as @xmath50 and ohmic contacts has been solved@xcite , and that the exact numerical solution is very well approximated by : @xmath51 .
\label{eq : sclpf}\ ] ] here @xmath52 is the device cross - sectional area , and @xmath53 is the interelectrode distance . in our geometry
@xmath53 is the channel length , @xmath52 is an effective cross - sectional area for the device ( proportional to channel width ) , and @xmath54 is the relative dielectic constant of the semiconductor ( chosen to be 3 in our analysis ) .
the appearance of @xmath55 in this equation does not imply that the electric field is constant over the device length .
rather , eq . [ eq : sclpf ] suggests a means of plotting @xmath4 data to quickly ascertain consistency with the detailed numerical solution to the two - terminal sclpf problem .
note that @xmath32 can depend on temperature ; in a picture of hopping it should be proportional to the effective dols at the injecting contact .
the dependence of mobility on @xmath56 has long been seen in semiconducting and conducting polymers@xcite , and is associated with the charge carriers and disorder in these materials@xcite . the numerical solution of ref .
@xcite ( approximated by eq .
[ eq : sclpf ] ) should be valid as long as the functional form of the field - dependent mobility remains @xmath56 .
this equation is derived@xcite assuming that the charge distribution is determined by source - drain electrostatics only .
the electric field from the gate certainly plays a nontrivial role in our devices , clearly affecting charge injection , and allowing the formation of a channel at higher temperatures ( since any charge present in the channel below dopant freeze - out has to be injected from the source and drain ) . for simplicity
, however , in this model we will assume that the gate dependence will manifest itself through @xmath32 , and that for fixed gate voltage we may treat the source - drain conductance like a two - terminal device .
figure [ fig : sclpf]a shows a representative log - log plot of @xmath13 vs. @xmath15 for one sample with @xmath57 nm and @xmath58 v , for several temperatures .
the solid lines are the numerical solution to eq .
[ eq : coupled ] , with parameters @xmath59 and @xmath60 chosen at each temperature to give the best fit .
the numerical solution is virtually indistinguishable from the analytic form of eq .
[ eq : sclpf ] .
data on this and other samples for different gate voltages are qualitatively similar , with very good agreement between the numerical solution of the sclpf model and the data . fig .
[ fig : sclpf]b shows the same data and fits replotted as suggested by eq .
( [ eq : sclpf ] ) . the quality of this agreement between the sclpf model and the data in several devices over a broad range of @xmath1 , @xmath14 , and @xmath43 is striking . clearly for a given ( large ) @xmath14 and ( low )
@xmath1 , @xmath61 .
alternate explanations of the nonlinear conduction are strongly constrained by this dependence .
the temperature and gate voltage range over which this form of source - drain nonlinearity occurs varies systematically with sample size , as described below .
one must consider whether the nonlinear @xmath62 characteristics result from nonlinear contact resistances as the temperature is decreased . for several reasons ,
discussed below , we do not believe this to be the case .
first we consider directly inferring the contact and channel resistances ( @xmath3 and @xmath63 , respectively ) in the linear regime , and examining the temperature variation of their relative contributions . for a series of devices with fixed width
, these resistances are calculated as follows@xcite : at a given @xmath1 , the total resistance , @xmath64 , is calculated for each device at a small @xmath15 and is plotted as a function of @xmath53 for each gate voltage .
the channel resistance per unit length , @xmath65 , at a given @xmath1 and @xmath14 is the slope ( @xmath66 ) of such a graph , and the intercept ( @xmath67 extrapolated to @xmath68 ) gives the parasitic series resistance , @xmath3 . for a fixed - aspect - ratio series of devices
, one may follow an analogous procedure . the total source - drain resistance @xmath69 , where @xmath70 is the resistance per square of the channel .
at a given @xmath1 and @xmath14 , @xmath71 is plotted versus @xmath53 for the series of devices .
the slope of such a graph gives @xmath70 , and the intercept gives the total parasitic contact resistivity , @xmath72 .
this analysis is shown in the fixed - aspect - ratio devices in fig .
[ fig : olddatarvsl ] at room temperature for several gate voltages . for our geometry of @xmath73
, the inset shows the inferred @xmath74 as a function of @xmath14 , for a device with @xmath75 m .
_ our series of fixed - aspect - ratio devices is clearly channel - limited at room temperature . _
this analysis may be repeated at different temperatures to examine the evolution of @xmath3 and @xmath63 .
we find that the nonlinear conduction at large average source - drain fields shown in the previous section makes this difficult to measure over a broad temperature range in the fixed - aspect - ratio device series .
however , the fixed - width devices with longer channel lengths are well - suited to this approach down to 100 k. figure [ fig : rsrchvst ] shows @xmath74 as a function of temperature for the @xmath76 m device from the @xmath10 m series .
near room temperature , @xmath77 for this device .
as @xmath1 decreases , while both @xmath3 and @xmath78 increase significantly , @xmath3 falls below @xmath63 near 100 k. the results of this experiment and others to be published elsewhere@xcite demonstrate that the contact contributions become _ less _ important at low temperatures .
this strongly suggests that the nonlinearities in the @xmath62 curves of short channel length devices observed at lower temperatures are unlikely to be contact effects .
furthermore , the trends in transport with sample dimensions also support this conclusion .
we find at low temperatures that the smallest devices actually transport charge considerably _
better _ than larger devices . for example , at 50 k for fixed @xmath43 , the fixed - aspect - ratio @xmath28 nm device exhibits measurable conduction for gate voltages as small as -25 v , while the @xmath79 m device requires @xmath80 v. this trend is the _ opposite _ of what one would expect for contact - limited conduction@xcite . since @xmath19 is held constant in this set of devices , shorter channel devices have significantly smaller contact areas as well , further emphasizing this point .
the data are , however , consistent with the suggestion of space charge effects seen in 70 nm channel length p3ht fets@xcite . finally , the voltage and temperature dependence of the data in fig .
[ fig : sclpf ] is not consistent with the forms for either classical schottky contacts or fowler - nordheim emission .
detailed theory@xcite and experiments@xcite on injection into disordered polymer semiconductors show that injection efficiency can actually _ improve _ as temperature is decreased , consistent with our contact resistance data described above . coupled with the size dependence ,
this supports the idea that low temperature transport in our devices is _ bulk _ limited rather than contact limited .
continuing within the sclpf model and our analysis of the fixed - aspect - ratio devices , we note that , for identical effective mobilities and @xmath60 parameters , eq . [ eq : sclpf ] implies that two samples with the same aspect ratio , thickness , and average source - drain field should give the same currents , independent of channel length , even deep in the nonlinear regime .
device - to - device variability in the effective mobility and @xmath60 , presumably due to differences in p3ht thickness and microstructure , make this challenging to check directly in our devices . for reasonable values of @xmath60 and @xmath43 ,
a 10% variation in @xmath60 would lead to more than a factor of 2 variation in predicted current at low temperatures because of the exponential dependence in eq .
[ eq : sclpf ] .
however , if one fixes @xmath43 and @xmath14 , and normalizes measured currents by room temperature mobilities , one does indeed find scaling .
for example , the @xmath81 nm and @xmath75 m currents in the nonlinear regime normalized this way agree well all the way down to 10 k. within this model , the parameter @xmath60 should depend only on the hopping mechanism ( _ e.g. _ thermal activation ) and the nature of the localized states .
we therefore expect @xmath60 to be independent of gate voltage for a given sample , and this is indeed seen in the inset to fig .
[ fig : slopet ] . at high temperatures (
@xmath82 k ) , the data for all gate voltages and all samples look roughly linear in @xmath83 , consistent with thermally activated hopping .
the magnitude of the slope of @xmath60 vs. @xmath83 is approximately 0.12 k(m / v)@xmath84 .
this is consistent in magnitude with coefficients found in other semiconducting polymers such as poly(phenylene vinylene)@xcite .
however , @xmath60 vs. @xmath83 deviates significantly from a straight line at lower temperatures for all samples .
this is consistent with a crossover from thermally activated hopping to a much less steep temperature dependence .
a natural candidate is field enhanced tunneling between the localized states .
we now consider the gate and temperature dependence of the parameter @xmath59 found by the numerical analysis above . the effective cross - section for current flow , @xmath52 , is assumed to be temperature and gate voltage independent for each sample . in the linear regime at moderate temperatures it is known that the mobility inferred from the transconductance is gate voltage dependent , as seen in fig .
[ fig : mobt ] .
this dependence on @xmath14 continues in the apparent sclpf regime , as shown in fig .
[ fig : intvgt]a on an @xmath85 nm device .
the amount of variation of @xmath32 with @xmath14 , roughly a 5% increase of @xmath32 per volt of @xmath14 for this sample , shows no strong trend with temperature .
the magnitude of this variation of mobility with @xmath14 is consistent with that seen at higher temperatures in fig .
[ fig : mobt ] for this sample .
this exponential dependence of @xmath32 on @xmath14 is seen throughout the apparent sclpf regime . as in the linear regime of an accumulation fet ,
gate voltage modulation of the conduction is _ unipolar _ : higher currents result only when @xmath14 is made more negative .
the temperature dependence of @xmath32 is also very strong , as shown in fig .
[ fig : intvgt]b for the same 300 nm channel sample at five gate voltages .
the variation of @xmath86 shown is much closer to an exponential in @xmath1 to some slightly sublinear power rather than an arrhenius or vrh form .
this strong temperature dependence of @xmath32 holds for all samples in the apparent sclpf regime .
the strongly energy dependent dols employed in , for example , the vissenberg model offers a natural explanation for these steep dependences of @xmath32 on @xmath14 and @xmath1 . in the absence of any gate effect ,
the _ effective _ density of localized states available for hopping transport at some temperature @xmath1 is given by @xmath87 where energy is measured from the band edge .
an exponential dols of the vissenberg model would then lead to an exponential dependence of @xmath32 on @xmath1 .
gate voltage dependence in this case comes from electrostatic modulation of the fermi level in this rapidly varying dols .
it is difficult to understand otherwise how an exponential dependence of the prefactor on @xmath1 or @xmath14 could arise .
deriving a quantitative relationship between @xmath14 and the local fermi level would require solving the full electrostatic problem of scplf conduction in the presence of the transverse gate field .
one can consider whether the proposed sclpf conduction is a _ bulk _ process or one dominated by conduction in the thin channel layer active in standard fet operation .
for the sclpf mobility parameter @xmath32 to coincide with the zero field mobility found at high temperatures in the linear regime , an effective cross - sectional area @xmath52 considerably larger than @xmath88 a few nanometers is required .
this is also true for the data of ref .
@xcite , in which a considerably different device geometry was used , if analyzed using the sclpf model .
however , the model of eqs .
( [ eq : coupled]),([eq : pflike ] ) does not account for the presence of a gate electrode or complicated source and drain geometries , and full computational modeling in this regime may be required to quantitatively account for this .
one test for bulk vs. channel conduction would be to search for a correlation between p3ht thickness and currents in the apparent scl regime .
this investigation is ongoing .
for a series of field effect devices with channel lengths ranging from 3 @xmath0 m to 200 nm , we find gate modulated nonlinear conduction at low temperatures and high average source - drain electric fields .
analysis of channel and contact resistances as a function of temperature , and the dependence of conduction on sample size at low temperatures support the conclusion that this nonlinearity is unlikely to be a contact effect .
we find that the data are well described by a model of gate modulated space - charge limited currents with poole - frenkel - like behavior of mobility . within this model ,
the @xmath14 and @xmath1 dependence of the mobility prefactor is consistent with a very strongly energy dependent density of localized states .
finally , the temperature dependence of the poole - frenkel - like term within this model suggests a crossover from thermal hopping to quantum tunneling at low temperatures .
further studies of the field effect electrostatics problem , the metal - semiconductor contacts , and the low temperature nonthermal hopping process should lead to increased understanding of the conduction processes at work in these materials .
the authors gratefully acknowledge the support of the robert a. welch foundation . | we measure hole transport in poly(3-hexylthiophene ) field effect transistors with channel lengths from 3 @xmath0 m down to 200 nm , from room temperature down to 10 k. near room temperature effective mobilities inferred from linear regime transconductance are strongly dependent on temperature , gate voltage , and source - drain voltage . as @xmath1 is reduced below 200 k and at high source - drain bias , we find transport becomes highly nonlinear and is very strongly modulated by the gate .
we consider whether this nonlinear transport is contact limited or a bulk process by examining the length dependence of linear conduction to extract contact and channel contributions to the source - drain resistance .
the results indicate that these devices are bulk - limited at room temperature , and remain so as the temperature is lowered .
the nonlinear conduction is consistent with a model of poole - frenkel - like hopping mechanism in the space - charge limited current regime .
further analysis within this model reveals consistency with a strongly energy dependent density of ( localized ) valence band states , and a crossover from thermally activated to nonthermal hopping below 30 k. |
compressive sensing ( cs ) signal acquisition paradigm asserts that one can successfully recover certain signals sampled far below their nyquist frequencies given they are sparse in some dictionary @xcite .
the fourier dictionary for frequency sparse signals is an example of this .
encouraged by this assertion , the usual sample and then compress setup can be combined into a single efficient step .
signals acquired in this fashion do , however , have to be reconstructed which , in the noiseless case , entails a non - convex optimisation problem of the form : @xmath2 where @xmath3 is the reconstructed signal , @xmath4 is a known measurement matrix , and @xmath5 is the measured signal with @xmath6 . in the cs context
, @xmath7 is the number of samples sensed while @xmath8 is the number of samples in the original signal .
we take @xmath9 to denote the @xmath0 pseudo norm from @xcite , i.e. the number of non - zero entries in @xmath10 .
solving the combinatorial problem in ( [ eq : compressive_sensing - general_problem ] ) by an exhaustive search is generally infeasible .
one feasible approach in reconstructing the signal is to relax the problem in ( [ eq : compressive_sensing - general_problem ] ) by substituting the @xmath1 norm for the @xmath0 making the problem a linear program ( lp ) @xcite .
another feasible approach is taken by the family of so - called iterative greedy algorithms . in these , the problem in ( [ eq : compressive_sensing - general_problem ] ) is reversed by minimising the residual of the energy of @xmath11 subject to some sparsity enforcing constraint .
abstractly , the greedy algorithms can be separated into two classes @xcite : 1 ) simple one stage algorithms which use a single greedy step in each iteration . examples are matching pursuit ( mp ) @xcite and iterative hard thresholding ( iht ) @xcite .
2 ) composite two stage algorithms which combine a greedy step with a refinement step in each iteration .
examples are orthogonal matching pursuit ( omp ) @xcite and cosamp @xcite .
the main advantage of the greedy algorithms over the @xmath1 approach is that they are computationally less complex @xcite and require less computation time than state - of - the - art lp solvers @xcite .
in addition to the computation time , a measure of the reconstruction quality must be considered .
recently , the measure of phase transition @xcite has become a standard way to specify reconstruction capabilities , see e.g. @xcite , @xcite , @xcite , @xcite .
phase transitions evaluate the probability of successful reconstruction versus the indeterminacy of the constraints @xmath12 and the true sparsity of @xmath10 . in general
, the main advantage of the @xmath1 approach over the greedy algorithms is that it is superior in terms of phase transition @xcite . in search of a fast algorithm with a phase transition similar to that of the @xmath1 approach
, it has been proposed to solve ( [ eq : compressive_sensing - general_problem ] ) by approximating the @xmath0 norm with a continuous function @xcite .
the resulting smoothed @xmath0 norm ( sl0 ) algorithm has a better phase transition than the greedy algorithms while requiring considerably less computation time than the state - of - the - art lp solvers . in this paper , we show that a few key parameters must be carefully selected and knowledge of the indeterminacy exploited to fully unleash the potential of sl0 .
we provide a set of empirically determined recommended parameters for a modified sl0 algorithm that may dramatically improve its phase transition . through extensive simulations , the claim of superiority of the recommended parameters is supported .
finally , we discuss implementation strategies that speed up the algorithm by exploiting knowledge of the indeterminacy .
the paper is organised as follows . in section [ sec :
sl0 ] , we restate the sl0 algorithm and present the proposed algorithm .
implementations of sl0 that yield reduced computation time are discussed in section [ sec : fast_implementations ] .
section [ sec : simulation_framework ] describes the setup used for simulations while section [ sec : results ] provides the simulation results .
a discussion of the results is given in section [ sec : discussion ] .
finally , conclusions are stated in section [ sec : conclusions ] .
sl0 attempts to solve the problem in ( [ eq : compressive_sensing - general_problem ] ) by approximating the @xmath0 norm with a continuous function . consider the continuous gaussian function @xmath14 with the parameter @xmath15 : @xmath16 the parameter @xmath15 may be used to control the accuracy with which @xmath14 approximates the kronecker delta . in mathematical terms
, we have @xcite : @xmath17 define the continuous multivariate function @xmath18 as : @xmath19 since the number of entries in @xmath10 is @xmath8 and the function @xmath18 is an indicator of the number of zero - entries in @xmath10 , the @xmath0 norm of the reconstructed vector @xmath10 is approximated by : @xmath20 substituting this approximation into ( [ eq : compressive_sensing - general_problem ] ) yields the problem : @xmath21 the approach is then to solve the problem in ( [ eq : compressive_sensing - sl0_problem ] ) for a decreasing sequence of @xmath15 s .
the underlying thought is to select a @xmath15 which ensures that the initial solution is in the subset of @xmath22 over which the approximation is convex and gradually increase the accuracy of the approximation . by careful selection of the sequence of @xmath15 s , ( hopefully ) non - convexity and thereby local minima are avoided . in the sl0 algorithm stated below ,
we let @xmath23 denote the moore - penrose pseudo - inverse of the matrix @xmath24 and let @xmath25 denote the hadamard product ( entry wise multiplication ) of the vectors @xmath26 and @xmath27 . furthermore
, we let @xmath28^t$ ] and @xmath29 . * sl0 algorithm : * * initialise : * @xmath30 , @xmath31 , @xmath32 , @xmath33 @xmath34 @xmath35 @xmath36 [ alg : sl0-sl0_basic_algorithm - d ] @xmath37 [ alg : sl0-sl0_basic_algorithm - x1 ] @xmath38 [ alg : sl0-sl0_basic_algorithm - x2 ] @xmath39 for each @xmath15 , the problem in ( [ eq : compressive_sensing - sl0_problem ] ) is solved by repeatedly taking an unconstrained gradient descent step in line [ alg : sl0-sl0_basic_algorithm - x1 ] ( @xmath40 is a normalised gradient of @xmath18 ) and projecting @xmath10 back onto the feasible set in line [ alg : sl0-sl0_basic_algorithm - x2 ] .
substituting line [ alg : sl0-sl0_basic_algorithm - x1 ] into line [ alg : sl0-sl0_basic_algorithm - x2 ] results in the expression : @xmath41 where @xmath42 is the identity matrix .
the matrix @xmath43 is a projector onto the null space of @xmath24 .
consequently , as pointed out in @xcite , the unconstrained gradient descent step followed by projection back onto the feasible set is equivalent to a direct gradient descent step on the feasible set .
the reason is that the projector @xmath43 restricts the gradient descent step to be on the feasible set .
the parameter @xmath44 is the constant ( typically chosen in the interval @xmath45 ) in the geometric sequence of @xmath15 s while @xmath46 relates to the final value used for @xmath15 and hence this controls the quality of the reconstruction .
the optimal choice of @xmath44 and @xmath46 is problem dependent while an initial @xmath47 as well as @xmath32 and @xmath33 are problem independent recommendations given in @xcite .
a comprehensive proof of the convergence of the sl0 algorithm exists for a specific set of parameters provided that an asymmetric resticted isometry property ( arip ) constraint is satisfied @xcite .
the authors of the proof conclude that though theoretically satisfactory , the arip constraint leads to an unnecessarily pessimistic choice of parameter values . motivated by this conclusion ,
we have carried out an extensive empirical analysis with the objective of finding the optimal parameter values .
this analysis has shown that the standard step - size @xmath32 and iteration number @xmath33 provided in @xcite result in a suboptimal phase transition . instead , we suggest a modification to the sl0 algorithm which may greatly improve the phase transition .
our initial observation is that the sl0 algorithm presented here is outperformed by the greedy iht algorithm described in @xcite in terms of phase transition for indeterminacy @xmath48 . by introducing an update of the @xmath49 parameter
, the phase transition of the sl0 algorithm can be improved to lie between that of the greedy iht algorithm and that of the @xmath1 approach .
specifically , we obtained this improvement by multiplying @xmath49 with a factor @xmath50 after each update of @xmath15 . by carefully selecting sequences of @xmath51 s and @xmath15 s
, the phase transition of the sl0 algorithm can be even further improved to consistently lie on or above that of the @xmath1 approach . in brief , we chose the step - size in the order of @xmath52 for the first few @xmath15 s and in the order of @xmath53 for the last @xmath15 s .
also , we chose the initial @xmath15 based on knowledge of the indeterminacy @xmath54 to improve the phase transition across all @xmath54 . based on our experimental results ,
we propose the following strategy .
we choose a step - size of @xmath55 for the first three @xmath15 s and a step - size of @xmath56 for the last @xmath15 s .
we use a sequence of @xmath57{^{\mathrm{t}}}$ ] where the number of entries equals the number of @xmath15 s used .
furthermore , an inversely proportional relation between @xmath54 and the initial value of @xmath15 yielded the most promising phase transition .
specifically , we choose an initial @xmath58 and combine this choice with a @xmath59 . finally , a gradually increasing @xmath49 for decreasing @xmath15 still provides an improvement for the updated parameter choices . here
, we choose a geometric sequence starting with @xmath60 and increasing by a factor of @xmath61 for each update of @xmath15 . with an increased value of @xmath44 and gradually increasing values of @xmath49 , the computation time is bound to increase .
this effect can , however , be counteracted by introducing a stopping criterion in the inner loop of the sl0 algorithm .
therefore , we choose to terminate the inner loop when the relative change @xmath62 falls below @xmath63 where @xmath64 .
generally , this measure has proved to be a good indicator of convergence and significantly reduced the average number of iterations taken in the inner loop .
we now propose the smoothed @xmath0 norm algorithm with modified step - size ( sl0 mss ) which incorporates all of the above findings .
* sl0 mss algorithm : * * initialise : * @xmath59 , @xmath31 , @xmath65 , + @xmath61 , @xmath66 , @xmath67 @xmath34 [ alg : sl0_mod - ls_init ] @xmath68{^{\mathrm{t}}}$ ] @xmath69 @xmath70 @xmath71 [ alg : sl0_mod - inner_loop ] @xmath72 @xmath36 @xmath73 [ alg : sl0_mod - x_update ]
@xmath74 @xmath75 @xmath76 @xmath77
through experiments , we have found the most time consuming parts of the sl0 mss algorithm to be the computation of @xmath23 and the matrix - vector products in the gradient descent step .
we now provide a strategy for exploiting the known indeterminacy @xmath78 to reduce the number of required computations .
consider the full qr decomposition of @xmath79 @xcite : @xmath80 \begin{bmatrix } { \mathbf{r } } \\
{ \mathbf{0 } } \end{bmatrix } \end{aligned}\ ] ] where @xmath81 and @xmath82 form bases for the row space and null space of @xmath24 , respectively and @xmath83 is upper triangular . note how the dimensions of @xmath84 , @xmath85 , and @xmath86 change with @xmath54 . from the qr decomposition , the pseudo - inverse @xmath23 can be found : @xmath87 three equivalent expressions for the projector that projects onto the null space of @xmath24 are then : @xmath88 we now propose the following scheme for an implementation of the sl0 mss algorithm . when @xmath89 , use the projector @xmath23 and split the update in line [ alg : sl0_mod - x_update ] in an infeasible gradient descent step followed by a projection onto the feasible set : @xmath90 when @xmath91 , use the projector @xmath92 and split the update in line [ alg : sl0_mod - x_update ] in two steps : @xmath93 to appreciate the split procedure , note that fewer arithmetic operations are needed in computing the matrix - vector product @xmath94 this way when : @xmath95 where @xmath96 denotes the number of floating point operations in the computation of the expression @xmath97 .
the split procedure does not require the explicit forming of @xmath23 since the initial least squares ( least norm ) solution @xmath98 in line [ alg : sl0_mod - ls_init ] may be efficiently computed by solving the following system of equations by substitution ( since @xmath86 is triangular ) : @xmath99 followed by computing : @xmath100 avoiding the explicit forming of @xmath23 eliminates the need for computing the inverse of @xmath101 which becomes progressively more computationally expensive as @xmath102 .
on the other hand , only the reduced qr decomposition @xmath103 is needed when @xmath23 is explicitly formed .
the reduced qr decomposition becomes progressively less computationally expensive than the full qr decomposition as @xmath104 .
these two observations motivate the change of method at @xmath105 .
we evaluate the reconstruction capabilities of an algorithm by use of the phase transition measure @xcite which provides the following framework .
let @xmath106 denote the number of non - zero entries in the true signal @xmath26 .
define the measure of indeterminacy @xmath78 and the generalised measure of sparsity ( density ) @xmath107 . given a success criterion , the success rate is then evaluated on the phase space @xmath108 ^ 2 $ ] . in general , reconstruction is easier for larger @xmath54 and smaller @xmath109 and becomes increasingly difficult when decreasing @xmath54 and increasing @xmath109 .
somewhere in - between , the phase transition curve separates the phase space into a phase where reconstruction is likely and a phase where it is unlikely .
this phase transition curve is continuous in @xmath54 for fixed @xmath8 . obviously , it is desirable to have a phase transition curve which is as close to @xmath110 as possible .
different suites of problems , i.e. different combinations of ensembles of @xmath24 and @xmath26 generally result in different phase transitions .
choosing @xmath24 from the uniform spherical ensemble ( use ) and the non - zero entries in @xmath26 from the rademacher distributed generally yields the most difficult problem suite in terms of obtaining good phase transitions @xcite . in the simulations , we consider this problem suite along with a the problem suite where @xmath24 is chosen from the use and the non - zero entries in @xmath26 are chosen from the zero mean , unit variance gaussian ensemble . in @xcite , it is shown that the probability of reconstruction versus @xmath109 for fixed @xmath8 and @xmath54 can be modelled accurately by logistic regression for a specific set of algorithms .
logistic regression is used more generally in @xcite to determine the location of the phase transition curve .
we adopt the logistic regression approach to estimate the location of the phase transition curve and fix @xmath111 as proposed in @xcite .
we then attempt reconstruction on a uniform grid @xmath112 in the phase space specified by : @xmath113 for each point in the grid , we do 10 monte carlo simulations where each simulation features a new draw of @xmath24 and @xmath26 . from @xcite we adopt that an attempted reconstruction is considered successful when : @xmath114 where @xmath10 and @xmath26 are the reconstructed and true signal , respectively . if the criterion is not met , the attempted reconstruction is considered unsuccessful , i.e. , the attempted reconstruction can not be considered indeterminate . to evaluate the required computation time for the different algorithms
, we measure the absolute time spent on reconstruction when it succeeds .
the problem suite formed by choosing @xmath24 from the use combined with rademacher distributed non - zero entries in @xmath26 is used in this test . a uniform grid @xmath112 in the phase space is formed by : @xmath115 an algorithm is tested on all points in the grid that are at least 0.025 below its empirically determined phase transition ( measured on the @xmath109-axis ) . the time spent on reconstruction for the successful part of 10 monte carlo simulations in each point
is then averaged .
considered problem sizes are : @xmath116 the simulations have been conducted on an intel core i7 970 6-core based pc with ddr3 ram .
the os used is 64-bit ubuntu 12.04 lts linux and the enthought python distribution ( epd ) 7.2 - 2 ( 64-bit ) .
all simulations are carried out in double precision . to validate the results obtained from our simulation framework
, we have simulated the iterative hard thresholding algorithm presented in @xcite .
the phase transition obtained in our simulation framework has then been compared with the phase transition obtained in the simulation framework of @xcite . due to the non - deterministic nature of the monte carlo simulations used in both simulation frameworks , the two phase transitions will inevitably differ slightly .
however , we have observed that they are almost identical and therefore concluded , that our simulation framework works as intended .
four algorithms have been simulated : 1 ) sl0 std which is the sl0 algorithm presented in section [ sec : sl0 ] .
2 ) sl0 min which is the same algorithm except it is modified such that @xmath49 is multiplied by @xmath117 each time @xmath15 is decreased .
3 ) sl0 mss which is the algorithm presented in section [ sec : improving_phase_transition ] .
4 ) iht which is the iterative hard thresholding algorithm described in @xcite . in the case of the sl0 mss algorithm ,
two implementations have been simulated : the sl0 mssslowromancap1@ implementation based on ( [ eq : sl0_mod_mssi_1 ] ) and ( [ eq : sl0_mod_mssi_2 ] ) and the sl0 mssslowromancap2@ implementation based on ( [ eq : qr_methods - split1 ] ) and ( [ eq : qr_methods - split2 ] ) .
the experimental results are presented in figure [ fig : results - rademacher - phase ] , [ fig : results - gaussian - phase ] , [ fig : results - computation_times ] , and [ fig : results - scaling ] .
figure [ fig : results - rademacher - phase ] shows the phase transitions for rademacher distributed non - zero entries in @xmath26 while figure [ fig : results - gaussian - phase ] shows the phase transitions for zero mean , unit variance gaussian non - zero entries in @xmath26 . in both figures , the theoretical @xmath1 curve from @xcite is included for reference .
figure [ fig : results - computation_times ] shows the measured average computation times versus indeterminacy @xmath78 and figure [ fig : results - scaling ] shows the measured average computation times versus problem size @xmath8 . note the abrupt ending of the sl0 std curve in figure [ fig : results - computation_times ] which is due to failure of reconstruction in the tested grid for @xmath118 .
also , note that the measured computation times of the sl0 min implementation have been divided by 20 . in summary ,
sl0 mss shows the best phase transition among the tested algorithms for both rademacher and gaussian non - zero entries in @xmath26 . in large portions of the phase space
it even surpasses the theoretical @xmath1 curve .
the only exception is in the gaussian case for @xmath119 where iht shows better phase transition . regarding computation time , iht is faster than the sl0 approaches among which sl0 min is consistently more than 20 times slower than sl0 mss
. furthermore , iht scales slightly better with problem size @xmath8 than the sl0 approaches .
for rademacher non - zero entries in @xmath26 , figure [ fig : results - rademacher - phase ] reveals that sl0 std , sl0 min , and iht are by far outperformed by sl0 mss in terms of phase transistion .
even the theoretical @xmath1 curve is surpassed by sl0 mss at around @xmath120 .
for @xmath121 , sl0 mss shows the same phase transition as the theoretical @xmath1 curve . the curve for sl0 min is a clear example of the improvement in phase transition obtainable using more iterations in the inner loop of the sl0 algorithm . however , sl0 mss further improves on this , especially for @xmath122 and @xmath120 . the results in figure [ fig : results - computation_times ]
settle that sl0 does indeed require more computation time than iht .
iht is around two to four times faster ( depending on @xmath54 ) than the fastest sl0 implementation for a problem size of @xmath123 .
the important thing to note though , is that sl0 provides a trade - off between phase transition and computation time .
the price paid in computation time for using a lot of iterations to get better phase transition is clear from the sl0 min curve .
this is , however , not the case for sl0 mss , which requires less than or about the same computation time as sl0 std depending on @xmath54 .
thus , a much better phase transition is obtained using largely the same computation time in going from sl0 std to sl0 mss .
to obtain such a result , it is necessary to switch from sl0 mssslowromancap1@ to sl0 mssslowromancap2@ at around @xmath105 .
an assessment of the scaling of average computation time with problem size @xmath8 reveals that all three sl0 algorithms seem to scale in an equivalent way .
iht scales better than sl0 and hence requires relatively less computation time as the problem size @xmath8 increases .
the scaling depicted in figure [ fig : results - scaling ] is for @xmath105 which provides a rough average computation time across all values of @xmath54 as can be seen from figure [ fig : results - computation_times ] .
the parameters for sl0 mss stated in section [ sec : improving_phase_transition ] are ( locally ) optimal in terms of phase transition for rademacher non - zero entries in @xmath26 .
gaussian non - zero entries in @xmath26 are known to be in favour of greedy algorithms @xcite , which is also the case for iht in our simulations , especially for @xmath119 where the iht curve surpasses the theoretical @xmath1 curve . for
@xmath124 , sl0 min and sl0 mss demonstrate the best phase transition among the shown algorithms .
min and sl0 mss phase transitions are about the same , though .
comparing our results for sl0 mss in figure [ fig : results - gaussian - phase ] with the ones given for sl0 in figure 6 in @xcite shows about the same phase transition .
the slightly better phase transition for @xmath119 in @xcite may be due to the sl0 mss parameters not necessarily being optimal for gaussian non - zero entries in @xmath26 .
although the above simulations are quite encouraging , they are based on an empirically tuned algorithm . thus , to reach a final verdict of the success of sl0 mss , the validity of the simulation results must be exhaustively studied for a broader set of problem suites .
alternatively , more sound mathematical proofs must be presented .
we have proposed a new compressive sensing reconstruction algorithm named sl0 mss based on the smoothed @xmath0 norm .
it turns out that sl0 phase transitions heavily depend on parameter selection .
sl0 mss attempts to improve on phase transition by exploiting the known indeterminacy @xmath125 combined with carefully selected parameters .
a trade - off between phase transition and computation time is provided by sl0 .
improved phase transition has been measured for sl0 mss compared to standard sl0 while maintaining the same computation time .
10 [ 1]#1 url@samestyle [ 2]#2 [ 2 ] l@#1=l@#1#2 e. j. cands and m. b. wakin , `` an introduction to compressive sampling , '' _ ieee signal processing magazine _ , vol .
25 , no . 2 , pp .
2130 , mar .
d. l. donoho , `` compressed sensing , '' _ ieee transactions on signal processing _ ,
52 , no . 4 , pp .
12891306 , apr .
, `` for most large underdetermined systems of linear equations the minimal 1-norm solution is also the sparsest solution , '' department of statistics , stanford university , tech . rep .
2004 - 9 , sep . 2004 .
[ online ] .
available : http://statistics.stanford.edu/~ckirby/techreports/gen/2004/2004-09.pdf a. maleki and d. l. donoho , `` optimally tuned iterative reconstruction algorithms for compressed sensing , '' _ ieee journal of selected topics in signal processing _ , vol . 4 , no . 2 , pp .
330341 , apr .
s. g. mallat and z. zhang , `` matching pursuits with time - frequency dictionaries , '' _ ieee transactions on signal processing _ , vol .
41 , no . 12 , pp . 33973415 , mar .
t. blumensath and m. e. davies , `` iterative hard thresholding for compressed sensing , '' _ applied and computational harmonic analysis _
27 , no . 3 , pp . 265274 , 2009 . j. a. tropp and a. c. gilbert , `` signal recovery from random measurements via orthogonal matching pursuit , '' _ ieee transactions on information theory _ , vol .
53 , no . 12 , pp .
46554666 , dec . 2007 .
d. needell and j. tropp , `` cosamp : iterative signal recovery from incomplete and inaccurate samples , '' _ applied and computational harmonic analysis _ ,
26 , no . 3 , pp . 301321 , 2009 . w. dai and o. milenkovic , `` subspace pursuit for compressive sensing signal reconstruction , '' _ ieee transactions on information theory _
55 , no . 5 , pp .
22302249 , may 2009 .
t. blumensath and m. e. davies , `` normalized iterative hard thresholding : guaranteed stability and performance , '' _ ieee journal of selected topics in signal processing _ , vol . 4 , no . 2 , pp .
298309 , apr .
d. l. donoho and j. tanner , `` precise undersampling theorems , '' _ proceedings of the ieee _ , vol .
98 , no . 6 , pp . 913924 , jun .
d. l. donoho , a. maleki , and a. montanari , `` message passing algorithms for compressed sensing : ii .
analysis and validation , '' in _ ieee information theory workshop ( itw ) _ , cairo , egypt , jan . 68 , 2010 , pp . 15 .
p. jain , a. tewari , and i. s. dhillon , `` orthogonal matching pursuit with replacement , '' in _ twenty - fifth annual conference on neural information processing systems _ , granada , spain , dec .
1215 , 2011 , pp . 16721680
. b. l. sturm , m. g. christensen , and r. gribonval , `` cyclic pure greedy algorithms for recovering compressively sampled sparse signals , '' in _
45th ieee asilomar conference on signals , systems , and computers _ , pacific grove ( ca ) , usa , nov . 69 , 2011 , pp . 11431147 .
g. h. mohimani , m. babaie - zadeh , and c. jutten , `` a fast approach for overcomplete sparse decomposition based on smoothed l0 norm , '' _ ieee transactions on signal processing _ ,
57 , no . 1 ,
pp . 289301 , jan .
z. cui , h. zhang , and w. lu , `` an improved smoothed l0-norm algorithm based on multiparameter approximation function , '' in _
12th ieee international conference on communication technology ( icct ) _ , nanjing , china , nov . 1114 , 2010 , pp .
. h. mohimani , m. babaie - zadeh , i. gorodnitsky , and c. jutten , `` sparse recovery using smoothed l0 ( sl0 ) : convergence analysis , '' _ arxiv _ , 2010 , submitted ( on 24 january 2010 ) to ieee transactions on information theory .
[ online ] .
available : http://arxiv.org/abs/1001.5073 s. boyd and l. vandenberghe , _
convex optimization_.1em plus 0.5em minus 0.4emcambridge university press , 2004 , 9th printing .
d. donoho and j. tanner , `` observed universality of phase transitions in high - dimensional geometry , with implications for modern data analysis and signal processing , '' _ phil .
a _ , vol .
1906 , pp . 42734293 ,
j. tanner .
phase transitions of the regular polytopes and cone : tabulated values .
the university of edinburgh . accessed : 22 - 05 - 2012 .
[ online ] .
available : http://ecos.maths.ed.ac.uk/polytopes.shtml | signal reconstruction in compressive sensing involves finding a sparse solution that satisfies a set of linear constraints .
several approaches to this problem have been considered in existing reconstruction algorithms .
they each provide a trade - off between reconstruction capabilities and required computation time . in an attempt to push the limits for this trade - off
, we consider a smoothed @xmath0 norm ( sl0 ) algorithm in a noiseless setup .
we argue that using a set of carefully chosen parameters in our proposed adaptive sl0 algorithm may result in significantly better reconstruction capabilities in terms of phase transition while retaining the same required computation time as existing sl0 algorithms . a large set of simulations further support this claim .
simulations even reveal that the theoretical @xmath1 curve may be surpassed in major parts of the phase space . |
the paper is devoted to the asymptotic study of the flow in the hele - shaw cell with presence of moving obstacle in the flow .
the hele - shaw problem ( @xcite ) deals with the description of the free boundary encircling the domain occupied by incompressible fluid in the so called hele - shaw cell ( see , e.g. @xcite , @xcite ) , i.e. in a narrow space between two parallel plates .
different driving mechanisms can be considered for the fluid flow , e.g. presence of a source / sink in the fluid domain .
various physical assumptions lead to different formulations of the respective boundary value problems .
a comprehensive discussion on this topic can be found in the recent book by gustafsson and vasilev @xcite .
there exist two basic mathematical models for the flow in the hele - shaw cell .
the _ complex - analytic model _ is formulated as a nonlinear mixed boundary value problem with respect to a family of conformal mappings of the canonical domain onto the domain occupied by the fluid .
this approach goes back to the work by polubarinova - kochina @xcite and galin @xcite .
the proof of the existence ( locally in time ) and uniqueness of analytic solutions to this model was done by kufarev @xmath1 vinogradov @xcite ( rediscovered later by richardson @xcite ) on the basis of the method of successive approximations .
simplified proof of existence and uniqueness of an analytic solution was given by reissig @xmath1 von wolfersdorf @xcite ( see also @xcite , @xcite ) . in this work
the model was interpreted as a special case of an abstract cauchy - kovalevsky problem , which was solved by a variant of the cauchy - kovalevsky theorem ( @xcite , @xcite , @xcite ) .
see also @xcite and references therein for the survey of recent results on the complex - analytic hele - shaw problem in doubly connected domains .
let us note that there are some similarities between movement of the rigid body and movement of bubbles in the flow ( for the discussion of the latter process we refer , e.g. , to the article @xcite and references therein ) .
in our study we use the _ real - variable model _ proposed by gustafsson @xcite the flow is described by a family of parametrizations of the boundary of fluid domain ( see also @xcite ) .
this model was generalized to multi - dimensional case by begehr @xmath1 gilbert @xcite . among variants of the proof of existence and uniqueness for this model we have to point out the papers by reissig @xcite and by escher @xmath1 simonett @xcite . in the most general form ,
the proof of the existence , uniqueness of the classical solution and the regularity of the fluid boundary was given by antontsev , gonalves and meirmanov @xcite , @xcite .
_ variational _ formulation of the hele - shaw model was proposed by gustafsson @xcite , who proved the weak solvability of the problem ( see also @xcite , @xcite ) .
the classical ( real - variable ) hele - shaw model can be reinterpreted as a mixed boundary value problem for laplace equation with respect to unknown parametrization of the boundary and corresponding green s function of this problem in the reference domain .
when assuming the presence of a moving obstacle in the flow , we have to add an additional equation describing this movement .
the aim of our work is to perform an asymptotic analysis for such a variant of the model and to construct an efficient and robust numerical routine to tackle the problem .
application of asymptotic methods to approximation of green s function goes back to the classical paper by j. hadamard @xcite , where the method of regular perturbation was performed .
recently , v. mazya and a. movchan obtained a number of asymptotic formulas for green s function related to different boundary value problems for a number of differential operators in the case of singular perturbations of the domains ( see @xcite , @xcite , @xcite and references therein ) .
those results were used in @xcite to model the hele - shaw flow with a fixed circular obstacle . to construct the computational scheme , we choose green s formula with the neumann condition on the external boundary of doubly connected domain , and with dirichlet condition on its internal boundary .
then a preliminary transformation @xmath2 was made , which led to creation of a system of differential equations for the original problem approximation .
the existence of the reformulated hele - shaw problem follows immediately from the results of the paper escher - simonett ( 1997b ) .
the approximate system defined in this way was reduced to the system of first order odes , and tackled by a proposed numerical scheme .
this approach proved its ability to solve the analyzed hele - shaw problem , providing sufficiently good accuracy of computations .
however , the scheme itself exhibited some disadvantages : \1 ) in the case of sink the life - time of the approximate process was very short ; \2 ) in the case of source we get numerical result only on a bounded interval of time .
the first difficulty could have been expected from the theory of the hele - shaw problem , but the second one is a consequence of the numerical scheme instability . in the case of moving obstacle
this led to even worse results .
in this paper we consider the hele - shaw flow with a rigid inclusion moving in the direction of the flow without rotation .
the friction between the limiting planes and the obstacle is accounted for . to avoid problems
appeared in the case of the fixed obstacle , we use here the same mazya - movchan approach , but with different uniform asymptotic formula for green s function ( without making any preliminary transformations , and thus interchanging the role of boundary conditions ) .
the paper is organized as follows .
2 describes the problem s geometry and presents the ( real - variable ) hele - shaw model in a domain with a moving obstacle .
the model is further reduced to the form containing an unknown parametrization of the boundary of the fluid domain , an unknown green s function of the corresponding mixed boundary value problem for the laplace operator and unknown trajectory of the inclusion .
therefore we have to consider a system of equations consisting of the equation for the free boundary ( the standard hele - shaw equation ) and the equation of motion for the obstacle . in sec .
3 we present uniform asymptotic formula by mazya - movchan and describe its components . in sec .
4 we determine the values of the components of the mazya - movchan formula corresponding to the considered model .
the final form of the approximate system of differential equations is presented in subsec .
4.4 . finally , in sec .
5 the aforementioned system is implemented in a numerical scheme which illustrates the obtained results . the numerical scheme to obtain the solution employs reduction of the system of governing equations to the system of odes of the first order , where the velocity of the inclusion is introduced as an additional dependent variable .
in order to solve the dynamic system we utilize the standard ode solver of matlab package : ode45 .
respective conformal mappings of the boundary curve are performed by means of the schwartz - christoffel toolbox .
the derivatives of the mapping along the free boundary are computed by our own subroutines , based on the spline approximation .
we show that the used asymptotic expansion for the green function is effective and the computations based on that approach are stable and robust .
we consider the slow flow in the hele - shaw cell ( i.e. in the narrow gap between to parallel plates of distance @xmath3 ) .
the flow is caused by a source / sink ( situated at the origin @xmath40 of intensity @xmath5 .
the fluid of the viscosity @xmath6 occupies the bounded doubly connected domain @xmath7 at the time instant @xmath8 that takes the form @xmath9 , where @xmath10 is a bounded simply - connected domain , and the compact set @xmath11 is a small obstacle embedded within the fluid . the obstacle is moving in the direction of flow rotation free and with friction coefficient @xmath12 . to avoid technical difficulties
, we accept a circular shape of the obstacle of the radius @xmath13 and of center @xmath14 at each instant of time @xmath15 .
suppose that our initial geometry satisfies the following conditions and @xmath16 do not depend on @xmath17 . ]
@xmath18 @xmath19 following @xcite , the initial free boundary @xmath20 is to satisfy the smoothness assumptions @xmath21 with certain fixed @xmath22 . omitting the standard description of the ( real - variable ) hele - shaw model ( see , e.g. @xcite , @xcite , cf .
@xcite ) we arrive at the following problem with respect to unknown parametrization @xmath23 of the free boundary @xmath24 ( i.e. the boundary value of the conformal mapping of the unit disc @xmath25 onto the fluid domain @xmath7 ) , green s function @xmath26 of the domain @xmath7 and the center @xmath14 of the obstacle . , @xmath27 , @xmath28 depend on time @xmath15 from a right - sided neighborhood @xmath29 of @xmath30 .
in fact , for our problem we need to determine the value of @xmath26 only at the point @xmath31 , but we keep the extra variable @xmath32 for computational reasons . ] * problem ( @xmath33 ) .
* find a triple @xmath34 ; @xmath35 , satisfying the following conditions \(i ) @xmath36 for all @xmath37 ; \(ii ) @xmath38 is a @xmath39-diffeomorphism for each fixed @xmath40 ;
\(iii ) @xmath41 is a given @xmath42-diffeomorphism of the unit circle @xmath43 , which describes the boundary @xmath44 of initial domain @xmath45 ; \(iv ) @xmath26 is green s function of the operator @xmath46 in the doubly connected domain @xmath47 with the homogeneous neumann condition on @xmath48 and the homogeneous dirichlet condition on @xmath49 ; \(v ) @xmath50 for all @xmath51 ; \(vi ) @xmath52 , \(vii ) @xmath53 , @xmath54 .
the aim of our study is to get an approximate solution to the problem @xmath33 .
in order to replace the system of equation @xmath55 of the problem ( @xmath33 ) by the approximate system we use one the results by mazya and movchan . for further convenience , we introduce here a small parameter @xmath13 equal to the radius of the inclusion , and denote by @xmath56 the rescaled obstacle @xmath57 .
note that for each @xmath40 we have @xmath58 . for the reader s convenience ,
we reformulate in our notation the theorem by v. mazya and a. movchan providing uniform asymptotic approximation of green s function with the neumann data on the boundary of the obstacle @xmath59 and the dirichlet data on the boundary of the domain @xmath60 .
[ mamov ] ( * ? ? ?
2.1 ) let @xmath61 be a bounded doubly connected domain in @xmath62 with a smooth boundary , @xmath63 , where @xmath60 is a simply connected domain and @xmath64 is a compact set ( obstacle ) with diameter smaller than the distance of @xmath65 to @xmath66 .
green s function @xmath67 of the laplace operator @xmath68 with the neumann zero - data on @xmath65 and the dirichlet zero - data on @xmath66 has the following uniform asymptotic representation @xmath69 @xmath70 where @xmath71 in what follows we use this formula in the fluid domain @xmath7 for all @xmath72 $ ] , for which the solution to the hele - shaw problem exists . in the case of the flow without obstacle or with a fixed obstacle we refer for the existence to @xcite . in our case
there is no rigorous proof of the existence , but it can be obtained similarly to that for the flow of bubbles ( see , e.g. @xcite and references therein ) . in our case
we accept in this theorem the following notation for each instant of time @xmath40 .
@xmath73 is green s function of the laplace operator @xmath68 for the simply connected domain @xmath74 with zero dirichlet data on @xmath75 : @xmath76 with @xmath77 being the regular part of green s function , i.e. harmonic function solving the following boundary value problem @xmath78 @xmath79 @xmath80 is the neumann function for the exterior of the re - scaled obstacle @xmath56 : @xmath81 where @xmath82 is the regular part of this function satisfying @xmath83 @xmath84 @xmath85 the vector - function @xmath86 is the solution of the following boundary value problems in the exterior of the re - scaled obstacle @xmath56 : @xmath87 @xmath88 @xmath89 here @xmath90 are components of the inward unit vector normal to the boundary of disc @xmath91 .
in this subsection we analyze the components of the representation ( [ mazmov_uniform ] ) .
let us first consider the neumann function @xmath80 having in this case an explicit representation ( see , e.g. , @xcite ) : @xmath94.\ ] ] satisfying the conditions ( [ neumann_regular ] ) , ( [ neumann_regular1])([neumann_regular3 ] ) and symmetric @xmath95 its regular part @xmath82 is also symmetric and calculated explicitly yields @xmath96.\ ] ] green s function @xmath97 for the interior simply connected domain @xmath10 can be represented in the form @xmath98 where @xmath99 is the conformal mapping of @xmath10 onto the unit disc @xmath25 , satisfying the following normalizing conditions @xmath100 , and @xmath101 . in our case ,
@xmath102 stands for the source / sink point ( we again note that from computational point of view it is better to keep extra - variable @xmath103 up to the final formula ) . from the numerical point of view
it is customary to start with an arbitrary conformal mapping @xmath104 and determine the normalized one : @xmath105 the vector - function @xmath106 can be found by using integral representation of the solution to the exterior neumann problem for the unit disc ( see , e.g. , @xcite ) .
first we note that the inward unit normal vector on the boundary of the unit disc @xmath91 is @xmath107 where @xmath108 is the angular coordinate of polar system on the unit circle @xmath109 .
then the solutions to the problems ( [ dipole1])([dipole3 ] ) ( @xmath110 ) are represented in the form @xmath111 where @xmath112 .
it is easy to see that the above functions satisfy all conditions ( [ dipole1])([dipole3 ] ) .
we calculate ( [ dipole_sol1 ] ) using formula ( * ? ? ?
* ( 4.397.6 ) ) : @xmath113 here we calculate derivatives of green s function which are used in equation ( v ) of the problem ( @xmath33 ) .
we start with green s function @xmath114 . by
applying representation ( [ green - int - sol ] ) we have ( @xmath110 ) @xmath115 substituting @xmath116 , @xmath117 and taking into account the properties of the function @xmath118 we finally obtain @xmath119 @xmath120 @xmath121 the neumann function @xmath122 depends on the `` scaled '' variables @xmath123 hence @xmath124 using explicit representation of the neumann function ( [ neumann - sol ] ) we get the following value of the derivatives @xmath125 @xmath126 @xmath127.\ ] ] now substitute @xmath128 , @xmath129 , @xmath110 ( in order to simplify representation we omit internal variables @xmath130 and @xmath15 in the right hand - side of this relation ) and calculate derivatives of two terms of mazya - movchan asymptotic formula ( see ( 3.1 ) ) at @xmath131 with @xmath117 @xmath132 these derivatives ( denoted @xmath133 ) are are equal @xmath134 @xmath135 + \varepsilon^2 w_j}{[(w_1 - z_{0,1})^2 + ( w_2 - z_{0,2})^2 - \varepsilon^2][z_{0,1}^2 + z_{0,2}^2 - \varepsilon^2 ] + \varepsilon^2 ( w_1 ^ 2 + w_2 ^ 2 ) } - \right.\ ] ] @xmath136 thus for @xmath137 we have @xmath138 . for regular part @xmath139 of green s function @xmath140
we have the representation ( [ green_simply ] ) , i.e. @xmath141 therefore @xmath142 @xmath143 @xmath144 analogously , @xmath145 @xmath146 in this case the right hand - side of the last relation does depend on @xmath147 .
now we have to calculate the derivatives with respect to @xmath147 of the following expression @xmath148 @xmath149 where only first multiplier in each summand depends on @xmath147 .
derivatives of @xmath150 in @xmath151 is connected with that in @xmath152 @xmath153 by the direct calculation we have @xmath154 @xmath155 and @xmath156 @xmath157 where the derivatives @xmath158 are presented in ( [ gr_reg_der_x1 ] ) .
at last we have @xmath159 @xmath160 here , only second multiplier in each summand depends on @xmath147 , and @xmath161 hence @xmath162 therefore , since @xmath163 and using ( [ dipole_repr ] ) we have @xmath164 in this subsection we calculate integrals from the right hand - side of the equation ( vi ) in the representation of * problem ( @xmath33 ) * : @xmath165 were @xmath166 , @xmath117 .
in our calculations we use components of formula ( [ mazmov_uniform ] ) and their representations obtained in subsec . [ components ] .
first we calculate the integral @xmath167 employing representation ( [ green - int - sol ] ) : @xmath168 note that for each fixed @xmath40 the function @xmath169 is an analytic function with respect to variable @xmath170 in the disc @xmath171 . hence using taylor expansion of @xmath169 at @xmath172 we have
@xmath173 similar calculations can be performed for the integral @xmath174 in the integral @xmath175 \cos \theta d \theta \left|_{{\bf y } = o}\right.,\ ] ] we use representation ( [ neumann - sol ] ) @xmath176 \cos \theta d \theta = \ ] ] @xmath177 @xmath178 since @xmath179 and @xmath180 , then we have @xmath181 the function @xmath182 is an analytic function with respect to variable @xmath170 in the disc @xmath171 .
hence @xmath183 analogously for the integral @xmath184 we have @xmath185 \sin \theta d \theta \left|_{{\bf y } = o}\right.=\ ] ] @xmath186 next we calculate @xmath187 taking into account exact values ( [ dipole_repr ] ) of the functions @xmath188 , @xmath189 and parametrization of the boundary of @xmath190 ( @xmath191 ) we obtain @xmath192 finally , using ( [ gr_reg_der_x1 ] ) @xmath193 similar calculations lead @xmath194 @xmath195 the last integral from ( [ green - int1 ] ) is calculated by using ( [ dipole_repr ] ) @xmath196 @xmath197 \cos \theta d \theta.\ ] ] since @xmath198 is an analytic function with respect to variable @xmath170 in the disc @xmath171 then @xmath199 similar result we have for the integral @xmath200 @xmath201 note that @xmath202 @xmath203 @xmath204 therefore , combining all above calculations we have @xmath205 = \frac{q \varepsilon^2}{4 m } i + o(\varepsilon^4),\ ] ] @xmath206 = \frac{q \varepsilon^2}{4 m } \tilde{i } + o(\varepsilon^4),\ ] ] @xmath207 @xmath208 it follows from the potential theory ( see , e.g. ( * ? ? ?
* ch . 8) , cf .
* lemma 2.3 ) ) , that for any compact subset @xmath209 , @xmath210 thus , the * problem ( @xmath33 ) * can be asymptotically approximated by the following system @xmath211 @xmath212 @xmath213 @xmath214 with initial conditions @xmath53 , @xmath54 . here
@xmath215 is an unknown parametrization of the external boundary @xmath75 , @xmath216 is an unknown position of the center of the moving obstacle and @xmath133 , @xmath217 , @xmath29 , @xmath218 are defined in ( [ k_j ] ) , ( [ dipole_derj1_1 ] ) , ( [ dipole_derj1_2 ] ) , ( [ dipole_derj2 ] ) , ( [ int_all ] ) , ( [ int_all_tilde ] ) .
in this section we provide only a short illustration of efficiency of the proposed methods for applications . the numerical scheme to obtain the solution employs reduction of the system of governing equations - to the system of odes of the first order , where the velocity of the inclusion is introduced as an additional dependent variable .
in order to solve the dynamic system of the first order derived in this way we utilize the standard ode solver of matlab package : ode45 . it is based on an explicit runge - kutta formula .
respective conformal mappings of the boundary curve are performed by means of the schwartz - christoffel toolbox @xcite , @xcite .
the derivatives of the mapping along the free boundary are computed by our own subroutines , based on the spline approximation . to investigate the accuracy of the proposed numerical scheme we use the classical benchmark by polubarinova - kochina @xcite , which describes the fluid domain induced by a source or a sink without inclusion ( @xmath219 ) .
evolution of the free boundary in the considered case is illustrated in fig.[domain_shape ] .
we analyze three different densities of the spatial meshing , described by the number of the nodes , @xmath220 , distributed at uniform angular distances : @xmath221 , @xmath222 , @xmath223 .
moreover , both the fluid source and fluid sink variants are considered . in the first case the free boundary evolves from the internal to external shape ( see fig.[domain_shape ] ) . in the second one it moves in reverse direction .
the results of computations illustrated by the relative error of the radius vector defining the free boundary , @xmath224 , are shown in fig.[rho_error ] .
it shows that the solution accuracy is of one order of magnitude better for the expansion ( @xmath227 ) than that for the contraction ( @xmath228 ) of the domain .
thus , the proposed algorithm is capable of tackling both cases with a satisfactory solution accuracy . in the next step
we investigate to what degree the presence of an immobile inclusion inside the domain affects the fluid flow .
now , we restrict ourselves only to the case of fluid source and consider a circular inclusion of the radius @xmath229 inside the domain encircled by the internal curve 2 from the previous benchmark .
we retain the same source intensity and time interval assuming zero initial conditions ( @xmath230 ) in the absence of any forces in the right - hand sides of - .
two various locations of the inclusion are considered : @xmath231 and @xmath232 .
the graphical illustration of the problem is shown in fig .
[ imm_incl_shapes ] , where the final shapes for the free boundary for both variants are compared with the case of undisturbed flow depicted with markers .
relative deviations of the radius vector , @xmath224 , from the one obtained for the undisturbed flow are shown in fig .
[ ro_rel_dev1 ] .
as can be expected , the maximal distortion of the boundary curve takes place approximately along the direction source - inclusion .
moreover , the shorter the distance between the source and inclusion is , the more pronounced deviation from the reference value is obtained .
since in both cases the source supplies the same volume of fluid in the considered time , one can check the accuracy of computations in terms of the fluid balance .
the respective areas are : @xmath234 the relative deviations of @xmath235 from the benchmark value were : @xmath236 and @xmath237 for the first and second location of the obstacle , respectively .
we believe that the second value is greater due to the integration error itself , as the relative deviation of @xmath233 has a much sharper maximum in this case ( compare fig .
[ ro_rel_dev1 ] ) . however , both obtained results suggest very good accuracy of the solutions as well as very good quality of the green s function approximation even for relatively large magnitude of the small parameter @xmath13 .
note , that accuracy of the uniform asymptotic formula increase ( decrease ) with time in case of source ( sink ) .
next , we consider the inclusion with two degrees of freedom ( translations ) analysing its movement in two cases . in the first of them , the friction term in equations -
is neglected .
the second variant accounts for the friction phenomenon .
we assume in the computations that @xmath238 and thus the multipliers in both terms representing forces are the same and equal to @xmath239 . in both cases
we assume that initial position of the inclusion is @xmath240 and its initial velocity is zero . the evolution of the free boundary and the obstacle movement for the frictionless variant are shown in fig .
[ incl_mov_shapes ] .
starting from zero initial velocity , the obstacle moves rectilinearly along the line : the source - center of the inclusion .
we do not present a respective picture for the second variant of the problem , as the free boundary shape is hardly distinguishable from the former .
relative deviations from the benchmark values of @xmath241 for both cases are depicted in fig .
[ ro_rel_dev_mov ] .
the balance equation was satisfied this time to the level of @xmath242 .
the influence of friction on the inclusion movement is shown in fig .
[ s_t]-fig .
[ v_t_a_t ] .
as the obstacle moves in both cases along the straight line , it is sufficient to present the evolution in time of : the covered distance ( fig .
[ s_t ] ) -@xmath243 , the absolute value of the velocity ( fig .
[ v_t_a_t]a))-@xmath244 , and the absolute value of the acceleration ( fig .
[ v_t_a_t]b ) ) - @xmath245 .
as anticipated , the influence of friction becomes more pronounced along with the velocity increase , however in the considered time interval it is still far away from making the inclusion movement uniform . obviously by increasing the value of friction coefficient one
can obtain the steady state much faster . in the last part of our analysis we consider the case
when the initial velocity of the inclusion has a non - zero value , and its vector is not collinear with the line : center of inclusion - source .
we investigate the evolution of obstacle track , velocity and acceleration caused by the fluid flow .
it is assumed that the initial position of the obstacle is @xmath248 , while its initial velocity yields @xmath249 .
we consider two variants of the problem depicted in fig .
[ mov_incl_track ] . in the first one ,
the fluid flow is driven by the source and initial shape of the free boundary is described by the internal curve .
the second variant assumes the domain contraction caused by the fluid sink . here ,
the initial domain is defined by the external curve .
this time we shall rather concentrate on the inclusion movement , than on the evolution of the free boundary .
the relative deviations of the radius vector from respective reference ( benchmark ) values are shown in fig .
[ ro_dev_track ] .
naturally , the variant with the sink gives more pronounced deformation of the final shape , as the distance between the inclusion and the boundary is much smaller than in the opposite case .
the fluid balance equation was satisfied to the level of @xmath242 for the variant of domain expansion , and @xmath250 for domain contraction .
the traces of inclusion for both considered cases are shown by markers in fig .
[ mov_incl_track ] .
it should be emphasized that the imposed initial conditions do not imply kinematic equivalence between both variants of the problem .
it is a consequence of different initial accelerations resulting from equations - . as can be seen in fig .
[ a_t ] , fig .
[ a_t_abs ] , although the magnitudes of initial accelerations are very close to each other , their vectors directions are almost opposite . the curvatures of the tracks ( bend directions ) and the signs of respective components of acceleration are determined by the source / sink activity . in the case of domain expansion the fluid flow direction magnifies the velocity of obstacle .
thus the distance covered is greater than that for fluid sink . obviously , for other configurations of the initial velocity vector one can expect different trends .
concluding this section , we have shown that the method utilized the uniform asymptotic expansion for the green function delivered in @xcite is effective and the computations based on that approach are stable and robust .
kufarev , yu . p. vinogradov , on a filtration problem .
_ prikl . mat .
* 12 * , 1948 , 181198 ( in russian ) ( english translation : university of delaware , applied mathematics institute , technical report 182a , 1984 ) .
v. mazya , a. movchan , uniform asymptotics of green s kernels for mixed and neumann problems in domains with small holes and inclusions .
isakov , victor ( ed . ) , _ sobolev spaces in mathematics .
iii : applications in mathematical physics .
_ new york , ny : springer ; novosibirsk : tamara rozhkovskaya publisher .
international mathematical series * 10 * , 2009 , 277316 .
v. mazya , a. movchan , m. nieves , uniform asymptotic formulae for green s tensors in elastic singularly perturbed domains with multiple inclusions .
_ rendiconti .
accademia nazionale delle scienze detta dei xl , memorie di matematica e applicazioni , @xmath256 _ , vol .
* xxx * , 2006 , 103158 .
m. reissig , about a nonstationary mixed problem for holomorphic functions arising by the study of a potential flow past a circular cylinder with permeable surface , _ math .
_ , * 164 * , 1993 , 283297 .
m. reissig , the existence and uniqueness of analytic solutions for moving boundary value problem for hele - shaw flows in the plane , _ nonlinear anal .
theory , methods @xmath1 appl .
_ , * 23 * , no . 5 , 1994 , 565576 .
m. reissig and s.v .
rogosin , with an appendix of f. huebner , analytical and numerical treatment of a complex model for hele - shaw moving boundary value problems with kinetic undercooling regularization , _ euro j. appl .
_ , * 10 * , 1999 , 561579 . | asymptotic analysis of the hele - shaw flow with a small moving obstacle is performed .
the method of solution utilises the uniform asymptotic formulas for green s and neumann functions recently obtained by v. mazya and a. movchan .
theoretical results of the paper are illustrated by the numerical simulations .
_ keywords _ : hele - shaw flow , moving obstacle , green s function , neumann function , asymptotic analysis _ ams 2010 mathematics subject classification _ : 76d27 , 35j25 , 35r35 moving stone in the hele - shaw flow @xmath0 |
aggregated tax income , italy , benford s law
in the present information age the collection , processing and dissemination of data from different financial activities , of utmost importance for human welfare , is quite easy and convenient .
however , for policy makers , the extraction of meaningful information , critical for plausible strategic decisions , from the sea of available data is a formidable challenge .
the far reaching , adverse , consequences of using flawed data have been exemplified by the 2007 financial crisis in the initiation of which data of questionable quality being used in corporates and governments played a significant role @xcite . one statistical tool which can serve as a first check on the quality of ( large ) numerical data and thereby , to a great extent , simplifies the deciphering of anomalies present in ( large ) data sets is the so - called benford s law @xcite .
the law describes the counter - intuitive uneven distribution of numbers in large data sets .
usually , it appears that the occurrence of the first digits of numbers has nothing to do with their abundance within the data .
in fact , in a large data set , the appearance of each digit from 1 to 9 as first digit is equally likely with a proportion of about 11% .
however , according to benford s law , the appearance of digits is such that the distribution of the first digits tends to be logarithmic with numbers having smaller first digits appearing more frequently than those having larger first digits .
thus , the distribution is heavily skewed in favor of smaller digits with digits 1 , 2 and 3 taking about 60% of the total occurrences as first digits and the remaining six digits i.e. 4 to 9 left with only 40% of the occurrences .
benford s law was first reported by newcomb @xcite following his observation that the pages of logarithmic table books get progressively cleaner as one moves from initial to latter pages , the first page being the dirtiest .
about four decades latter , benford rediscovered the phenomenon through a similar observation and established it on a more solid footing by testing its accuracy on a large volume of data he collected from diverse fields , e.g. physical constants , atomic and molecular masses , street addresses , length of rivers etc . and concluded that the occurrence of first significant digits follow a logarithmic distribution @xcite
@xmath0 where p(d ) is the probability of a number having the first non - zero digit d and @xmath1 is logarithmic to base 10 .
+ the theoretical proportions for each of the digits from 1 to 9 to be first significant digit are as shown in table 1 .
the development of research on benford s law into a full fledged field is as fascinating as the discovery of the law itself .
firstly , though the mathematical form of the law is very simple a complete mathematical understanding is yet to be achieved @xcite .
furthermore , the law represents the only distribution of leading digits invariant under scale and base transformations@xcite .
secondly , numerous data sets from diverse fields conform to the law @xcite . for an exhaustive list
refer to @xcite .
yet , there is _ a priori _ no set of criteria to predict what type of data should conform to the law .
necessary and sufficient conditions are much in debate .
nevertheless , ( i ) the presence of sufficient number of entries in the data , for digits to manifest themselves , ( ii ) spanning of several orders of magnitude , and ( iii ) the absence of any human restrictions , i.e. there are no built in minima or maxima , on the appearance of numbers in data are some properties that data under investigation must posses for conformity to the law@xcite .
after the seminal work of benford , the first digit phenomenon again came to prominence following the efforts of nigrini who reported its frequent emergence in financial data @xcite .
furthermore , nigrini provided the first practical application of benford s law in the detection of tax evasion by hypothesizing that to save on their tax liabilities individuals might understate income items and overstate deduction items in their tax return files leading to overall distortion of digital frequencies .
a benford s law test successfully captured the manipulation of digital frequencies in the submitted data @xcite .
furthermore , falsification of financial documents , manipulated trade invoices and tax returns submitted by companies have been clearly unraveled @xcite .
today the law is routinely used by forensic analysts to detect error , incompleteness and wilfull manipulation of the financial data .
the basic premise of the test is that first digits in real data , in general , have a tendency to approach the benford distribution whereas people intending to play with the numbers , when unaware of the law , try to place the digits uniformly .
thus any departure from the law raises suspicion about the quality of the data @xcite and/or in the process involved in its generation @xcite .
benford s law has been successfully utilized to expose the intention of cheating in both corporates and the governments . in order to attract investment
, firms must posses a strong financial basis .
moreover , to project a healthy , though sometimes superficial , picture they report enhanced profits and reduced losses which are not always the actual values .
such a manipulation of financial statements , known as cosmetic earning management , is achieved through the rounding of numbers .
this was detected first for firms in new zealand where the frequency of zeros and nines as second digit of reported earnings was respectively more and less than could be expected from benford s law @xcite .
subsequently this unethical phenomenon has been reported to be the practice of the day for the firms worldwide @xcite .
a recent example of corporate data manipulation , on a truly global scale , is the 2011 libor scandal in which a cartel of banks distorted the interest pricing process of the inter - bank loans @xcite .
countries , like firms , also falsify economic data when it is strategically advantageous @xcite .
thus , questions have been raised about the data submitted by greece to the eurostat to meet the strict deficit criteria set by the european union ( eu ) . among all the member states of eu
, the data submitted by greece has been found to have the greatest deviation from benford s law @xcite .
similarly , the macroeconomic data of china is a subject of much debate , as it has been alleged to be overstating its gdp numbers to mislead investors @xcite .
furthermore , a benford s law based assessment of the macroeconomic data submitted by countries to the world bank hints at deliberate falsification of the data from the developing countries @xcite .
the manipulative behavior percolates down to the local governments as recent studies using benford s law have uncovered deficiencies in the data of municipalities and states of several countries .
the cases in point are the valejo city , orange county and jefferson county in u.s . , whose local authorities have filed for bankruptcy .
the digit distributions of the financial statements of all the three municipalities have been shown to have significant departures from that expected on the basis of benford s law , thereby raising questions about the credibility of the statements on revenues and spending @xcite .
further indications of data tampering by local governments have been uncovered through a benford analysis of the official financial reports of the fifty states of u.s .
another study from brazil analyzing the digit distribution of 134,281 contracts issued by 20 management units in two states found significant deviations from benford s law and concluded that there is a tendency to avoid conducting the bidding process and the rounding in determining the value of contracts @xcite . in italy ,
the municipalities represent the lowest level of the government responsible for providing services to the local residents .
some of these activities are property issues , such as building permits , street lump , garbage collection , public transport , etc . , and social activities , such as child and elderly care services , etc . @xcite .
tax collection is a fundamental source of revenues for local governments enabling the efficient delivery of services @xcite . on the other hand , tax evasion
is known to be widespread across italy with studies estimating between quarter and half of the country s gdp to be hidden from authorities in the form of underground economy @xcite .
the evasion of taxes is detrimental to the financial health of municipalities . to contain any financial distress ,
the municipalities resort to scaling down of expenditure like cutting on the number of employees , reducing the salaries of those that continue to be employed and even complete stopping of some of the services @xcite .
thus , any financial distress of municipalities has severe repercussions on the lives of the taxpayers and municipal employees .
it is important to have better oversight of the quality of financial statements and accountability over the use of funds .
the concerns on data quality and the poor auditing procedures being used have resurfaced more vigorously following the bankruptcy of number of local government bodies across several industrialized countries during recent financial crisis . in the present study
, we analyze the yearly aggregated tax data of all the italian municipalities ( cities ) for a period of five years from 2007 to 2011 to see if there are any deviations from benford s law .
beforehand , given the scale of tax evasion in italy , one expects that the digit distribution of tax data would be in complete disagreement of the predictions of benford s law .
however , we find that the tax data of all the italian cities shows an impressive compliance to the law . furthermore , we also analyze the yearly ( city cumulated ) data from three italian regions of calabria , campania and sicily .
the municipalities of these regions have , from common knowledge or assumption , a low level of governance , poor delivery of services , and substantial presence of mafia and organized crime @xcite .
thus given the poor tax administration in the municipalities of these regions , we again anticipated to find some hints of tax evasion through deviations from benford s law .
surprisingly , the data from these three regions also satisfy the law , except the years 2007 and 2008 data for campania which show large deviations from benford s law .
the italian state is organized in four levels of government ( i ) a central government ( ii ) 20 regional governments ( iii ) 110 provincial governments and ( iv ) more than 8000 municipalities , at the time of this writing .
the municipalities represent the lowest level of the government in the administrative structure of italy .
each municipality belongs to one and only one province , and each province is contained in one and only one region .
the total number of municipalities has slightly varied over the years .
this is due to occasional administrative reorganization , through the acts of the italian parliament , leading to the creation of new municipalities and sometimes also the merger of two or more municipalities into one @xcite .
thus we have a total of 8101 , 8094 , 8094 , 8092 and 8092 municipalities respectively for year 2007 , 2008 , 2009 , 2010 and 2011 , respectively . during this time interval ,
7 municipalities have changed from a province to another one , - in so doing also changing from a region ( marche ) to another ( emilia - romagna ) , in 2008 .
we have analyzed the yearly aggregated tax income ( ati ) data of all municipalities for the period of five years from 2007 to 2011 .
the data has been obtained by ( and from ) the research center of the italian ministry of finance and economy ( mfe ) .
we have disaggregated contributions at the municipal level to the italian gdp
. the standard methodology of a benford analysis is to first count the appearances of each digit from 1 to 9 as the first digit of numbers in the data .
then the corresponding theoretical frequency of each digit as first digit is determined from benford s law .
this is followed by estimating the goodness of fit of the theoretical and observed digit distributions both graphically and also using suitable fitness test .
these steps are explained through the analysis of the yearly tax data for all the italian cities in table 2 .
the @xmath2 , the number of times each digit from 1 to 9 appears as the first significant digit in the corresponding data are shown for each yearly data set in columns 2 , 4 , 5 , 7 , 8 of table 2 .
also shown are @xmath3 , the corresponding counts ( in brackets ) for each digit as predicted by benford s law : @xmath4 where @xmath5 for each year is the total number of records i.e. the number of municipalities .
for example , the total number of municipalities for year 2007 is @xmath6 , as shown in column 2 of table 2 .
the root mean square error ( @xmath7 ) is calculated from the binomial distribution @xcite @xmath8 the observed count for digit 1 as first significant digit is 2433 for 2007 , whereas the expected count from benford s law is 2438.64 with an error of about 41.29 .
the expected count from benford s law and the corresponding error depends only on the number of records in the sample of data . for the yearly ati data
the number of records for years 2008 and 2009 is 8094 and is 8092 for years 2010 and 2011 .
thus , the expected count from benford s law and the corresponding errors are shown only once in column 4 and 6 respectively for such cases . from a visual inspection of table 2
, it is found that the observed and expected digit distributions are in reasonable agreement within the margins of the calculated error .
this is further illustrated in fig .
1 , where the observed proportion of the first digits are compared with those expected from benford s law .
contrary to the _ a priori _ expectations , for the ati data of all these years , the agreement between the observed and theoretically predicted distributions is quite remarkable . in view of the statistical quantification of the closeness between the
observed and predicted digits distributions , the pearson s @xmath9 , the most widely used test in benford s law literature , @xmath10 is thereafter used .
of course , zero is not the significant digit when it occupies the extreme left of a number .
thus , in case of the first digit analysis , we have @xmath11 `` data points '' , whence @xmath12 degrees of freedom .
the critical value of @xmath9 , under @xmath13 confidence level , for acceptance or rejection of null hypothesis - the observed and theoretically predicted digit distribution are same , is @xmath14=15.507 .
if the value of the calculated @xmath9 is less than the critical value , then we accept the null hypothesis and conclude that the data fits benford s law . for the 2007 ati data ( column 2 of table 2 ) ,
the @xmath9 is 15.36 ( the last row and column 2 of table 2 ) , a value smaller than @xmath15 ; whence the null hypothesis is accepted indicating in turn that the tax data for year 2007 follows benford s law .
the calculated @xmath9 for the year 2008 is 27.52 and for 2009 is 18.96 which are far greater than the critical value of 15.507 .
therefore , the null hypothesis must be rejected .
furthermore , the @xmath9 for year 2010 is 12.29 and for 2011 is 13.72 and both values are less than the critical value , such that the null hypothesis must be accepted . from a visual examination of fig .
1 , showing the comparison of the observed proportion of the first digits with those expected from benford s law , the conclusion is of an excellent compliance for all the years .
thus for the 2008 and 2009 ati data , conclusions on compliance to benford s law drawn from fig.1 and the @xmath9 test appear to be contradictory .
however , it is known that the results of the @xmath9 test are sensitive to the number of records in the data under analysis .
the rejection of the null hypothesis becomes difficult for samples of small sizes , called type ii error , whereas for large samples the test suffers from `` excessive power '' , wherein even a small deviation from benford s law turns out to be significant , called type i error .
this leads to a wrongful rejection of the null hypothesis .
the large data sets require increasingly better fits to pass the @xmath9 threshold for conformity @xcite , although by inspection they give better fits than small data sets , and often fail a @xmath9 test that the small dataset passes .
thus , larger values of @xmath9 for the tax data of years 2008 and 2009 , in our study , are likely due to this excessive power , despite the fact that they visually appear to show clear and extraordinary conformity to benford s law .
in fact , it is due to this excessive power that the calculated @xmath9 for 2007 , 2010 and 2011 yearly data sets are only marginally smaller than the critical value of 15.507 , though the close visual conformity , evident from fig . 1 , compels one to expect much smaller values .
thus , rather than being an indication of departure of the ati data from benford s law , the high @xmath9 values are a manifestation of the limitations of the pearson s @xmath9 test itself .
a second analysis is in order .
it is widely accepted that the so - called underground or black economy is larger in the southern regions of italy than elsewhere .
studies have shown that personal income tax and value added tax evasion is highest in calabria , sicily and campania @xcite . in anticipation to find some support for this conjecture through the deviations from benford s law , we specifically analyzed the data from these three regions .
calabria consists of 409 municipalities grouped in 5 provinces .
we show the analysis for the tax data of all years for calabria in table 3 . again
since the number of records is same for all the yearly data sets the benford expected frequency of each digit and the corresponding error are same and are only shown once in column 6 of the table 3 .
the pearson s @xmath9 are all much smaller than the critical value of 15.507 .
therefore , we conclude that the ati data follows benford s law , a fact which is also clearly attested by fig .
2 , where the observed and expected digit distributions are compared .
the analysis for sicily region is shown in table 4 . again
both the calculated @xmath9 and the corresponding graphical representation of fig .
3 show an excellent compliance to the law .
the results for the campania region are shown in table 5 . here , the respective @xmath9 for 2007 and 2008 are large than the critical value for acceptance of null hypothesis .
thus , for these two years the ati data of campania region clearly deviates from benford s law .
the @xmath9 for years 2009 to 2011 are less than the critical value of 15.507 , though only marginally .
it may be noted here that unlike the case of 2008 and 2009 yearly data for all the municipalities , for which the results of the @xmath9 test are contradictory to the inferences , evident from the corresponding figure fig .
1 , for campania , the calculated @xmath9 and the observations from the corresponding figure fig . 4
are complimentary .
the departure of campania data from benford s law can be clearly seen from fig .
4 where the frequency of digit 3 is much less and that of digit 6 is much greater than expected from benford s law .
though tax evasion in italy is a phenomenon of gigantic proportions @xcite its estimation is easier said than done @xcite . nevertheless , the literature on estimation of tax evasion in italy is plentiful .
the levels of tax evasion have been shown to vary across different regions of italy @xcite and also across different sections of the italian society along its economic landscape @xcite .
furthermore , political and social factors responsible for the thriving of tax evasion have been pointed out .
for example , the lack of governance @xcite and the levying of higher taxes are known to discourage taxpayers @xcite .
the congestion of tax evaders is another factor encouraging evasion as tax authorities are overburdened thereby reducing the risk of detection @xcite .
the modeling of tax evasion is difficult as evasion in itself consists of partial or complete concealment of a significant proportion of economic activities from the authorities .
thus researchers rely on sample surveys believing that people will disclose their true incomes when promised anonymity @xcite .
the data from surveys are then compared with the official data maintained by italian mfe .
the latter is in fact the primary source of data invariably employed in modeling the different aspects of tax evasion .
however , there is some general concern and much skepticism on the quality of tax data being maintained by mfe .
the existence of several inconsistencies related to format , syntax and semantic have been pointed out .
more serious , from the point of view of the subject in this present study , are the missing , obsolete , or incorrect data values and undiscussed outliers @xcite .
the present study assesses the quality of the mfe tax data relating to municipalities through the application of benford s law in order to see whether there are in the ati set , numerical anomalies which might hint at deliberate attempts of manipulation .
any manipulation , if found , would in turn likely signal the intention of tax evasion .
when people resort to under - reporting their taxable incomes , the occurrence pattern of digits in the tax return files will be altered in a major way .
this has been shown to be true by nigrini , in the first ever practical application of benford s law , after analyzing tax return files of individuals on u.s .
internal revenue services .
it was revealed that tax payers with low incomes resorted to blatant manipulation of line items at filing time to a greater extent than high income tax payers @xcite .
furthermore , prior studies suggest that benford s law successfully captures the manipulation of tax data to the point that the deviations from the law are acceptable as evidence in courts of several usa states .
our present study shows that overall tax data of italian municipalities are in complete agreement with benford s law , thereby negating the presence of evasion . at first , these results are somewhat surprising , since large scale presence of tax evasion in italy is well documented .
however , on a more close scrutiny the reason for the compliance of data in our study becomes apparent . in our analysis ,
each municipality reports only one record of tax data for a particular year ( in contrast to nigrini analysis of the data records of individual tax payers ) .
since municipalities are composed of hundreds and thousands of inhabitants the tax value is an aggregate value which , of course , is a result of combinations of tax receipts of a large number of citizens .
thus , any individual deviation from benford s law might be lost due to the multiple mathematical operations .
in fact , the numbers obtained after multiplying a lot of numbers together have been found to agree with benford s law .
this has been conjectured to be one of the possible mathematical explanations for the validity of the law @xcite .
it is relevant to mention here that another type of evasion comprises of complete concealment of incomes from the authorities .
such incomes usually arise from underground illegal activities like thefts , drugs , kickbacks , skimming and contract rigging .
since no record / trace of transactions exists at all the detection of tax evasion through such `` underground '' economic activities can not be achieved with benford s law .
previous studies have shown that personal income tax and value added tax evasion is highest in calabria , sicily and campania i.e. the southern regions of italy .
the relative backwardness of these regions may be one reason for the high tax evasion since inefficiency of the municipalities in delivering services discourages the payment of taxes as tax payers do not see any proper return for their paid taxes .
the weak governments of poorer regions are generally less efficient in tax administration @xcite further affecting the realization of tax revenues due to the strong presence of black economy , arising out of illegal and underground activities of mafia , ( always ) hidden from the authorities .
furthermore , extortion by mafia compels legal businesses to evade taxes .
fearful of their extensive reach and the deadly consequences for not obeying them businesses pay taxes to mafia rather than to government @xcite .
furthermore , the mafia infiltration of local governments of municipalities across these regions are pervasive and in order to restore the law enforcement against such infiltration the central government from time to time has resorted to the dissolution of the municipal administration . over the years
there have been a total 217 dissolution of local municipalities across italy .
however , most of the dissolutions have been invoked in the municipalities of south - italian regions with only 4 dissolutions being reported outside these regions @xcite . in the light of the above discussion
, there are sufficient reasons to suspect that the tax data of these regions would show large departures from benford s law . on the contrary
, we found that income tax data from these regions shows strong submission to benford s law except for the 2007 and 2008 yearly data of campania .
this is somewhat surprising since the calabrian ndrangheta is the most powerful amongst all the italian mafias @xcite .
we have analyzed the yearly aggregated tax income data of all the italian municipalities to search whether there are anomalies which might hint at deliberate attempts of manipulation for tax evasion , a phenomenon widespread across the country , according to common knowledge .
however , the overall data showed excellent compliance with benford s law , thereby negating the presence of manipulations at this aggregate level .
it might be that the aggregation process hides some individual breakdowns .
however , the aggregation is not a multiplicative process .
we have also analyzed the municipality tax data of three regions ( calabria , campania and sicily ) known for the strong presence of mafia .
again the data showed compliance to benford s law except for the 2007 and 2008 data for campania which showed significant departures from the law .
our findings suggests to reconsider the campania data .
no need to say that the other ( 17 ) regions might be similarly investigated .
moreover , one possibility for further probing the data and maybe concluding on some reason for the breakdown of benford s law , in a few cases , would be to investigate the province level .
the reader will easily understand that this demands 110 or so tests / year .
this activity goes beyond the scope of the present report and is left open for further research .
up to now , necessary and sufficient conditions for the application , and the more so , explanation of benford s law are not known , in spite of intense applications .
the present report indicates that more theoretical and numerical investigations are still of interest .
this paper is part of ausloos s scientific activities in cost action cost action is1104 , `` the eu in the new complex geography of economic systems : models , tools and policy evaluation '' and in cost action td1210 `` analyzing the dynamics of information and knowledge landscapes '' @xmath16 .
s. newcomb , note on the frequency of use of different digits in natural numbers , am . j. math .
4 ( 1881 ) 39 - 40 .
f. benford , the law of anomalous numbers , proc . am .
78 ( 1938 ) 551 - 572 .
a. berger , t. p. hill , benford s law strikes back : no simple explanation in sight for mathematical gem , the mathematical intelligencer , 33 ( 1 ) ( 2011 ) 85 - 91 .
e. canessa , theory of analogous force on number sets , physica a 328 ( 1 ) ( 2003 ) 4452 .
n. gauvrit , j - p .
delahaye , pourquoi la loi de benford nest pas myst ' erieuse , math . & sci .
182 ( 2008 ) 7 - 15 .
n. gauvrit , j - p .
delahaye , loi de benford g ' en ' erale , math . & sci .
sci . 186 ( 2009 ) 5 - 15 .
t. p. hill , base - invariance implies benford s law , proc . am .
123 ( 3 ) ( 1995 ) 887 - 895 .
t. p. hill , a statistical derivation of the significant - digit law , stat .
10 ( 4 ) ( 1995 ) 354 - 363 .
r. s. pinkham , on the distribution of first significant digits , the annals of mathematical statistics , 32(4 ) ( 1961 ) 1223 - 1230 .
l. pietronero , e. tosatti , v. tosatti , a. vespignani , explaining the uneven distribution of numbers in nature : the laws of benford and zipf , physica a 293 ( 2001 ) 297 - 304
. t. a. mir , the law of the leading digits and the world religions , physica a 391 ( 2012 ) 792 - 798
. t. a. mir , the benford law behavior the religious activity data , physica a 408 ( 2014 ) 1 - 9
. t. a. mir , the leading digit distribution of worldwide illicit financial flows , @xmath17 j. c. pain , benford s law and complex atomic spectra , phys .
e. 77 ( 2008 ) 012102 .
l. shao , b. q. ma , first digit distribution of hadron full width , mod .
a 24 ( 2009 ) 3275 - 3282 .
m. sambridge , h. tkalci , a. jackson , benford s law in the natural sciences , geo .
a 37 ( 2010 ) l22301 .
g. judge , l. schechter , detecting problems in survey data using benford s law , journal of human resources .
44 ( 2009 ) 1 - 24
. w. r. mebane , jr . , the wrong man is president !
overvotes in the 2000 presidential election in florida , presp . on polit . 2 ( 2004 ) 525 - 535
. b. f. roukema , benford s law anomalies in the 2009 iranian presidential election , @xmath18 .
w. k. t. cho , b. j. gaines , breaking the ( benford ) law : statistical fraud detection in campaign finance , the american statistician 61 ( 2007 ) 218 - 223 .
p. clippe , m. ausloos , benford s law and theil transform of financial data , physica a 15 ( 2012 ) 6556 - 6567 e. den heijer and a. e. eiben .
using aesthetic measures to evolve art . in : ieee congress on evolutionary computation ( cec 2010 ) , barcelona , spain , 18 - 23 july 2010 .
ieee press .
f. sandron , do populations conform to the law of anomalous numbers ?
population - e 57 ( 2002 ) 755 - 761 .
t. alexopoulos , s. leontsinis , benford s law and the universe , in press journal of astronomy and astrophysics ( 2014 ) .
m. ausloos , c. herteliu , b. ileanu , breakdown of benford s law for birth data . submitted .
t. p. hill , a. berger , benford online bibliography , @xmath19 c. durtschi , w. hillison , c. pacini , the effective use of benford s law to assist in detecting fraud in accounting data , journal of forensic accounting 1 ( 2004 ) 17 - 34 . m. j. nigrini , taxpayer compliance application of benford s law , the j. am .
18 ( 1 ) ( 1996 ) 72 - 92
. m. j. nigrini , l. j. mittermaier , the use of benford s law as an aid in analytical procedures , auditing : j. pract .
theory 16 ( 2 ) ( 1997 ) 52 - 67 .
m. j. nigrini , benford s law : applications for forensic accounting , auditing and fraud detection , new jersey , usa : wiley publications ( 2012 ) .
t. p. hill , the difficulty of faking data , chance 12 ( 3 ) ( 1999 ) 27 - 31 .
h. varian , benford s law , american statistician 23 65 - 66 ( 1972 ) . c. carslaw , anomalies in income numbers : evidence of goal oriented behavior , accounting rev .
63(2 ) ( 1988 ) 321 - 327 .
j. kinnunen , m. koskela , who is miss world in cosmetic earnings management ?
an analysis of small upward rounding of net income numbers among 18 countries , j. intern .
res . 2 ( 2003 ) 39 - 68 .
abrantes - metz , g. judge , s. villas - boas , tracking the libor rate , applied economics letters , 10(10 ) ( 2011 ) 893 - 899 .
t. michalski , g. stoltz , do countries falsify economic data strategically ?
some evidence that they might , the review of economics and statistics , 95 ( 2013 ) 591 - 616 .
b. rauch , m. gttsche , g. brhler , s. engel , fact and fiction in eu - governmental economic data , german economic review 12 ( 2011 ) 243 - 255 . c. a. holz , the quality of china s gdp statistics , stanford university , scid working paper 487 , 2 december 2013 .
j. nye , c. moul , the political economy of numbers : on the application of benford s law to international macroeconomic statistics .
be journal of macroeconomics , 7(1 ) , article 17 ( 2007 ) .
a. h. haynes , detecting fraud in bankrupt municipalities using benford s law , scripps senior theses . , paper 42 .
( 2012 ) available at @xmath20 g. g. johnson , j. weggenmann , exploratory research applying benford s law to selected balances in the financial statements of state govenments , academy of accounting and financial studies journal , 17(1 ) ( 2013 ) 31 - 44 .
j. costa , j. santos , s. travassos , an analysis of federal entities compliance with public spending : applying the newcomb - benford law to the 1st and 2nd digits of spending in two brazilian states , r. cont .
- usp , sao paulo , 23(60 ) , 187 - 198 , set./out./nov./dez . 2012 .
d. bartolini , r. santolini , political yardstick competition among italian municipalities on spending decisions , the annals of regional science , 49(1 ) ( 2012 ) 213 - 235 .
e. padovani , e. scorsone , measuring financial health of local governments a comparative framework , year book of swiss administrative sciences ( 2011 ) .
f. schneider , the increase of the size of the shadow economy of 18 oecd countries : some preliminary explanations , ifo working paper , 306 ( 2006 ) .
g. jones , italy approves decree to stave off bankruptcy for rome council , feb .
28 ( 2014 ) .
available at @xmath21 .
g. brosio , a. cassone , r. ricciuti , tax evasion across italy : rational noncompliance or inadequate civic concern , public choice , 112(3 ) ( 2002 ) 259 - 273 .
f. calderoni , where is the mafia in italy ? measuring the presence of the mafia across italian provinces , global crime , 12(1 ) ( 2011 ) 41 - 69 .
@xmath22 f. schneider , the value added of underground activities : size and measurement of the shadow economies of 110 countries all over the world , world bank working paper washington , d.c .
( 2000 ) . p. di caro , g. nicotra , knowing the unknown across regions : spatial tax evasion across italy , @xmath23 m. r. marino , r. zizza , the personal income tax evasion in italy : an estimate by taxpayer s type . in : michael pickhardt and aloys
prinz ( eds . ) , tax evasion and the shadow economy , cheltenham : edward elgar , 2011 . c. v. fiorio , f. damuri , workers tax evasion in italy , giornale degli economisti e annali di economia , 64 ( 2/3 ) ( 2005 ) 247 - 270 .
r. galbiati , giulio zanella , the tax evasion social multiplier : evidence from italy , journal of public economics , 96(5 ) ( 2012 ) 485 - 494 .
p. missier , g. lalk , v. verykios , f. grillo , t. lorusso , p. angeletti , improving data quality in practice : a case study in the italian public administration , distributed and parallel databases , 13(2 ) ( 2003 ) 135 - 160 .
j. boyle , an application of fourier series to the most significant digit problem , am .
month . 101
( 1994 ) 879 - 886 .
p. d. scott , m. fasli , benford s law : an empirical investigation and a novel explanation , csm technical report 349 , csm technical report , department of computer science , university essex .
available at @xmath24 , 2001 .
m. alexeev , e. janeba , s. osborne , taxation and evasion in the presence of extortion by organized crime , j. comparative economics , 32 ( 2004 ) 375 - 387 .
b. geys , g. daniele , organized crime , institutions and political quality : empirical evidence from italian municipalities .
workshop paper , dept . of political science , stanford university , available at | the yearly aggregated tax income data of all , more than 8000 , italian municipalities are analyzed for a period of five years , from 2007 to 2011 , to search for conformity or not with benford s law , a counter - intuitive phenomenon observed in large tabulated data where the occurrence of numbers having smaller initial digits is more favored than those with larger digits .
this is done in anticipation that large deviations from benford s law will be found in view of tax evasion supposedly being widespread across italy .
contrary to expectations , we show that the overall tax income data for all these years is in excellent agreement with benford s law .
furthermore , we also analyze the data of calabria , campania and sicily , the three italian regions known for strong presence of mafia , to see if there are any marked deviations from benford s law .
again , we find that all yearly data sets for calabria and sicily agree with benford s law whereas only the 2007 and 2008 yearly data show departures from the law for campania .
these results are again surprising in view of underground and illegal nature of economic activities of mafia which significantly contribute to tax evasion .
some hypothesis for the found conformity is presented . |
this paper presents a relativistic theory of gravitation that is developed by considering the concrete physical mechanisms responsible for time dilation and spatial contraction .
the theory is developed for the central gravitational field of a non spinning mass .
the mathematical framework is formulated from thought experiments in which the inverse square acceleration of newtonian gravitation and the principles of relativity are combined .
the euler lagrange equations based on the resulting lagrangian for extremal proper time are shown to be equivalent to the geodesic equations of general relativity with the schwarzschild metric . however , unlike general relativity this theory is nonsingular at the schwarzschild radius and therefore uniquely provides a description of particle dynamics there .
the schwarzschild metric of general relativity is shown to be an first order approximation to the lagrangian developed here .
relativistic principles , terms , and concepts used herein are given below with comments and references where appropriate .
since numerous references exist in both classical and modern publications for these only a sampling is provided where indicated .
they are identified here as essential logical building blocks of the sections that follow .
* the principle of equivalence - the original version of this principle given by einstein @xcite equated gravitational and inertial masses . since that time a number of logically related statements of equivalence have appeared @xcite @xcite @xcite @xcite . for purposes of this paper we will take the meaning of this principle to include the equivalence of gravitational and inertial masses , independent of their nature and we also utilize the `` strong '' version of equivalence @xcite stating that a laboratory that is freely falling in a gravitational field is equivalent to one in an inertial , or lorentz frame . a mental tool used to conduct thought experiments will be an elevator , an einstein elevator acted on only by gravitational forces and in free fall .
this principle therefore states that results of experiments conducted entirely within the elevator be identical to those conducted in an unaccelerated frame of reference - ignoring tidal effects caused by spatial gradients of the gravitational field .
the lorentz transformations @xcite will be applicable within this elevator . * the mass - energy equivalence of special relativity - will be used to associate a mass of an object to its total energy .
objects may be photons . *
the weak field approximation - a parameter that appears in the following theory as well as in general relativity is chosen to quantitatively establish the gravitational field
strength. at this point it is easiest to extract this parameter from the schwarzschild metric , which in polar coordinates is given by @xcite @xcite : + @xmath1 + here , @xmath2 is the space - time interval , @xmath3 is the radial coordinate of a test mass in a spherical coordinate system centered on a mass @xmath4 , @xmath5 its differential , @xmath6 and @xmath7 are the angular coordinate differentials , @xmath8 is the schwarzschild radius calculated for a mass m as @xmath9 , @xmath10 is the gravitational constant , and @xmath11 is the speed of light . within general relativity
the variation of the integral of @xmath12 over a path length is minimized , or in other words particles follow geodesic paths in a warped space - time .
the effects of space - time warping , and therefore observable differences in predictions in general relativity and newtonian gravitational theories are quantified by the ratio @xmath13 in the scwarzschild metric above .
the weak field limit is defined here as the regime where @xmath14 . for use later in this paper , the gravitational radius will be defined as one half the scwarzschild radius , and the weak field definition applies as well to this radius .
original and subsequent experimental tests of general relativity using the sun as the central gravitational body ( light deflection , orbital perturbations , red shifting ) are necessarily in the weak field regime as defined above .
consider gravitational deflection of light passing the sun s limb ; since the physical radius of the sun is on the order of 700,000 km , and the calculated schwarzschild radius approximately 3 km , the ratio of @xmath13 is on the order of @xmath15 , or about 4 parts per million .
this ratio s value is even smaller for planetary distances and confirmations based on those ( e.g. , perihelion advance of mercury ) . *
newtonian gravitation - this law describes gravitational effects as a radial force ( no azimuthal component ) that varies inversely with the square of the distance from a mass @xmath4 to a test mass @xmath16 which is assumed small relative to @xmath4 .
while the radial - only nature of this law is retained here , it will be viewed as an inverse - square _ acceleration law _
, applying to any mass regardless of its type .
* paths of extremal proper time - in this paper , this principle refers to requiring particles in a gravitational field to follow paths resulting in variationally stationary , or extremal proper time .
this principle is analogous to seeking geodesic paths in space time @xcite .
however we wish to utilize this principle a - priori and independently of any other theory . + non relativistic physics provides examples of such a principle including fermat s least time description of photon propagation as well as the least action formulation by maupertuis @xcite .
+ that paths are those of extremal time may also be seen from maupertuis s least action principle in another , semi - classical way .
seeking paths that minimize the action given by : + @xmath17 + where @xmath18 refers to the form of variation of the integral where the endpoints may also vary , @xmath19 the particle s momentum , and @xmath20 a generalized coordinate .
if we associate a quantum mechanical wavelength @xmath21 with this particle according to @xmath22 , where @xmath23 is plank s constant , substitution into the equation above results in ( ignoring the multiplicative constant @xmath23 since it does not affect the final result of the variation ) : + @xmath24 + the reciprocal of @xmath21 may be viewed as the number of oscillations , or cycles of time the particle experiences , it s proper time , per unit of @xmath20 .
the equation above then yields paths of stationary proper time . + all of these results may be made relativistic by the principle of equivalence . by this principle
an observer in an accelerating frame will identify the same particle path as a ( proper ) elapsed time extremal as would an observer in a non - accelerating frame .
the development of the theory presented here emphasizes visualization of physical processes through the use of thought experiments .
transformations are derived relating time and spatial coordinates in a non accelerating reference frame @xmath25 to those in a freely falling frame of reference @xmath26 . a third frame
is then introduced , @xmath27 which is in uniform motion with respect to @xmath26 .
the lorentz transformation is utilized to transform time and spatial coordinates from @xmath26 to @xmath27 , and then ultimately to @xmath25 .
these result are then used to produce an expression for an increment of time in the accelerating and moving frame @xmath28 which becomes the lagrangian for arriving at paths of stationary proper time generated by : @xmath29 it is then shown that this lagrangian reduces to eq.([schwarzschild_metric ] ) , the schwarzschild metric in the weak field limit and therefore provides an equivalent description of gravitation as general relativity in this regime .
it is further shown that this lagrangian provides non - singular solutions at the scnwarzschild radius as well as over all radii ( except at the coordinate origin ) .
in special relativity one may generate time and spatial transformations between a reference and uniformly moving frame by considering a michaelson interferometer with one arm oriented parallel to the velocity vector and the other arm oriented 90 degrees from that .
in such an approach the paths light follows along these two arms are calculated from the perspective of the stationary frame of reference , and then by utilizing the principles of relativity and the postulate of the constancy of the speed of light one obtains both the temporal and spatial transformations . at first
it would appear this approach would be suitable for developing transformations between a non accelerating and an accelerating frame of reference .
however an accurate description of the path light in a gravitational field , particularly a path involving both radial and transverse motion is indeed a major goal of this paper .
a different approach will therefore be taken based on photons in free fall as presented below .
we will make use of an einstein elevator which is in free fall in a gravitational field , i.e. the only force acting on this elevator and its contents is gravitation .
an observer inside this elevator uses an atomic clock to measure time in his frame of reference which we will designate as @xmath26 . while an atomic clock is clearly only one possible way to measure time , it is based on fundamental processes and so will be considered a manifestation of time itself . as observed from @xmath25 ,
@xmath26 is accelerating in the radial direction only toward the center of mass of @xmath4 as are photons within @xmath26 .
a test mass @xmath16 will undergo a change in energy @xmath30 due to an incremental change in radius @xmath5 according to : @xmath31 where @xmath10 is the gravitational constant , @xmath3 is the radial coordinate of the test mass in a spherical coordinate system centered on the center of mass of @xmath4 .
it is assumed that @xmath16@xmath32@xmath4 .
we may associate a mass with a photon emitted by the atomic clock according to einstein s mass - energy equivalence @xmath33 where @xmath34 is the photon energy , and @xmath11 is the speed of light . writing the photon energy as the sum of its energy at @xmath35 and its potential energy at a finite @xmath3 , we have for its mass : @xmath36 where @xmath37 is the photon frequency at @xmath35 , and @xmath23 is plank s constant .
substitution of this result into eq .
( [ potential ] ) gives : @xmath38 @xmath39 defining the gravitational radius @xmath40 and using the condition that @xmath41 at @xmath35 , we have for @xmath42 : @xmath43 since the energy of the photon is the sum of it s energy at @xmath35 and @xmath42 , and since it is also given by @xmath44 , we have : @xmath45 therefore : @xmath46 this expression indicates the emission frequency @xmath47 at a radial coordinate of @xmath3 is lower than when at @xmath35 , and has a value of zero at @xmath48 .
this lowering of frequency is otherwise known as gravitational red shifting .
an observer in @xmath25 may determine the frequency of these photons as the number of wave crests @xmath49 passing a fixed location per unit of time .
( [ red_shift ] ) above above he notices that an observer in @xmath26 will observe a reduced number of crests in that same unit of his time owing to its reduced frequency .
however the principle of equivalence requires observers in @xmath26 measure the same frequency as in @xmath25 .
therefore the rate of time in @xmath26 must have changed according to : @xmath50 @xmath51 this is the transformation for a time differential between @xmath26 and @xmath25 .
one of the teachings of special relativity is that there is a time dilation effect due to uniform motion on synchronization of spatially separated clocks .
this effect is sometimes known as the relativity of simultaneity .
we now evaluate whether this mechanism results in a significant time dilation effect in a free - falling frame .
assume that spatially separated clocks in @xmath26 are synchronized by the transmission of light pulses from a single clock designated as the master clock .
upon reception of these timing pulses each of the other clocks within @xmath26 adjusts their clocks to the previously agreed to time of transmission of these pulses after accounting for their respective propagation delays . as seen from @xmath25 ,
however the reception time of a reference pulse is not only determined by the propagation time , but also by an additional effect due to acceleration .
this is illustrated in figure [ spatialtime ] .
the ( a ) portion of this figure illustrates how a timing pulse is transmitted at time @xmath52 by the master clock located at a radial coordinate @xmath3 , to a remote clock located at a radial coordinate of @xmath53 .
the ( b ) portion of this figure shows that at the time the pulse is received , @xmath54 the remote clock s radial position has decreased by @xmath55 due to acceleration , @xmath56 .
this causes the timing pulse to be received earlier than by considering the propagation time alone . since in time
@xmath57 light has propagated a distance of @xmath58 , we have : @xmath59 @xmath60 the solution for @xmath54 is given by : @xmath61 \label{time received}\ ] ] we may define the factor describing the reduction in arrival time compared with the propagation time alone as @xmath62 expanding the radical term in the eq.([time received ] ) we have : @xmath63\ ] ] for a positive @xmath64 as shown in the figure above we may choose the @xmath65 sign in the expanded radical , and retaining to second order : @xmath66 substituting @xmath67 , and @xmath40 we have : @xmath68 the ( c ) portion of the above figure shows a similar effect for the clock located at @xmath69 . the same analysis
as the above shows this clock is synchronized to a time that is greater than that in the absence of acceleration by the factor @xmath70 that is given by : @xmath71 and so we see that the spatial ( radial ) dependence in time due to synchronization effects is second order in @xmath64 , and so can be ignored compared with the first order time dilation effect represented by eq .
( [ ttransform ] ) .
spatial dependence of synchronization in the angular ( orthogonal ) coordinates are zero since acceleration is zero there .
as before we will utilize a spherical coordinate system centered on the center of mass of @xmath4 .
we once again consider a closed elevator located at a radial coordinate @xmath3 whose acceleration ( and that of its contents ) toward the the origin is given by @xmath67 .
the effect of this acceleration on a measurement of distance is presented below .
consider a thought experiment in which an observer located within an elevator measures the speed of light with a light source and reflecting mirror oriented in the radial direction as shown in figure [ transforms ] .
he measures the distance between the source and mirror @xmath72 using a metering rod provided by an observer in @xmath25 who measured its length prior to acceleration as @xmath73 .
the measurement is performed by directing a light pulse from a source toward a mirror as shown in figure [ transforms]a .
as observed from @xmath25 the mirror translates a distance s1 during the travel time of the light pulse that is given approximately by @xmath74 .
the pulse is reflected and returns to the source s location by which time it has translated a total distance of s2 ( figure [ transforms ] b and c ) .
we have for the total upward and downward distances traveled by the light pulse given respectively by : @xmath75 @xmath76 @xmath77+\frac{a}{2 } \left(\frac{2l}{c}\right)^2\ ] ] @xmath78 if the time required for light to travel @xmath79 is @xmath80 we have in frames @xmath25 and @xmath26 : @xmath81}{\delta t}=\frac{2l^\prime}{{\delta t}^\prime}\ ] ] substituting the relationship between time increments from eq ( [ ttransform ] ) , @xmath67 , and @xmath40 and then solving the above equation for @xmath82 , the length of a metering rod in @xmath26 is : @xmath83 we see again that the effect of acceleration on the measurement of a standard rigid rod of length @xmath72 is second order .
therefore this second order factor can be ignored .
this is especially true since a standard of length should be `` small '' so as to afford precision in the measurement of lengths .
the expression above therefore reduces to : @xmath84 since radial distance , as well as an incremement of radial distance is the number of standard metering rods given by @xmath72 , we have that : @xmath85 this expression therefore is the desired transformation for radial differentials between frames @xmath25 and@xmath26 .
the same methodology applied to dimensions orthogonal to @xmath3 , i.e. @xmath86 and @xmath87 , where acceleration is assumed to be zero , and the principle of relativity yields : @xmath88 the previous sections have addressed time and spatial transformations due to gravitational acceleration only .
the effects of uniform motion will now be added .
we do so by introducing the frame of reference @xmath27 which is also in free fall and in uniform motion with respect to @xmath26 with velocity @xmath89 .
since frame @xmath26 is stationary with respect to @xmath25 , @xmath89 = @xmath90 . while the effects of uniform motion ( special relativity ) could be introduced from first principles , it is more efficient to utilize the lorentz transformations which must be valid in @xmath26 according to the equivalence principle .
taking the special relativistic time transformation between @xmath27 and @xmath26 as : @xmath91}^\frac{1}{2}dt^\prime\ ] ] in polar coordinates this may be written as : @xmath92}\right]}^\frac{1}{2}dt^\prime\ ] ] @xmath93}^\frac{1}{2}\ ] ] for brevity we substitute for the angular dimension ( orthogonal to radius ) @xmath94 .
the equation above becomes : @xmath95}^\frac{1}{2}\ ] ] the transformation to @xmath25 is made by substituting for @xmath96 , @xmath97 , @xmath98 , and @xmath99 from eqs ( [ ttransform]),([rtransform ] ) , and ( [ angular_transforms ] ) to obtain : @xmath100 } ^\frac{1}{2}\ ] ] substituting @xmath101 where @xmath8 is the schwarzschild radius , we finally arrive at an expression for the proper time increment in a frame which is both accelerating and in uniform motion : @xmath102 } ^\frac{1}{2}\ ] ]
from the preceding sections , paths resulting in stationary proper time are given by : @xmath103 differential equations of motion may therefore be obtained by using eq.([dt_double_prime ] ) as a lagrangian with substitution into the euler lagrange equation , or : @xmath104 } ^\frac{1}{2}\ ] ] with : @xmath105 where @xmath20 is a generalized coordinate , and dot refers to time differentiation .
once again it is important to recall the unit analysis results presented in the radial transformation section regarding the true units of the denominator in this expression .
examples of the resulting differential equations are not provided here due to space limitations and mathematical complexity , however limiting cases are compared with general relativity in the following section .
as shown previously the lagrangian of eq .
( [ lagrangian ] ) provides equations of motion resulting from this theory .
we will show that the square of this lagrangian provides the same equations of motion .
we will then show that this squared form of the lagrangian reduces to the schwarzschild radius in the weak field limit .
recall that the frames of reference @xmath26 and @xmath27 are both in free fall with respect to @xmath25 , and that @xmath27 is in uniform motion with respect to @xmath26 with velocity @xmath89 .
consider a particle which is at rest in @xmath27 .
as seen from @xmath26 this particle will be in uniform motion with speed @xmath89 .
an observer in @xmath26 who wishes to describe the motion of this particle would write a lagrangian as the particle s constant kinetic energy only resulting in the variational equation : @xmath106 where @xmath16 is the particle s mass .
solutions to this equation are not changed by a multiplicative constant . without loss of generality
we may therefore replace the constant @xmath107 with @xmath108 to obtain : @xmath109 and since the elapsed proper time for this particle over its path is given by : @xmath110 which is a function of @xmath111 and @xmath112 only .
we choose the variational formulation such that the end points are fixed and hence we may write : @xmath113 taking the difference of eqs .
( [ proper prime ] ) and ( [ ke only ] ) we obtain : @xmath114 @xmath115 substituting : @xmath116 eq . ( [ variation prime ] ) becomes : @xmath117 note that the integrand in eq .
( [ variation prime ] ) is exactly the square of that in eq .
( [ variation ] ) which is the variation of elapsed proper time .
we will now show that these two equations are equivalent for the generation of equations of motion .
differential equations of motion may be obtained from eq .
( [ lagrangian prime squared ] ) from the euler lagrange equation to obtain , where the dot notation refers to differentiation with respect to @xmath118 : @xmath119 @xmath120 since the particle is moving with constant speed in @xmath26 , @xmath72 is constant and may be factored to leave : @xmath121 this is precisely the euler lagrange equation that would follow from eq .
( [ variation ] ) demonstrating the equivalence of eqs .
( [ variation ] ) and ( [ variation prime ] ) .
we now transform the previously defined lagrangian : @xmath122 to frame @xmath25 using the transformations from eqs ( [ ttransform]),([rtransform ] ) , and ( [ angular_transforms ] ) and substitute into eq .
( [ lagrangian prime squared ] ) to obtain : @xmath123}=0\ ] ] taking the limit of eq.([full ] ) for @xmath124 , using the first order approximation for the exponential terms and simplifying we have : @xmath125 = 0\ ] ] this is equivalent to minimizing the variation of eq .
( [ schwarzschild_metric ] ) , the schwarzschild metric which leads to the equations of motion for geodesics in general relativity .
hence the non - metric theory presented here reduces to general relativistic gravitation for the weak field limit .
the exact form is the lagrangian of eq .
( [ lagrangian ] ) .
transformations are developed from a non - accelerating reference frame and a frame in free fall in a gravitational field which is also moving with a relative velocity @xmath90 with respect to the reference frame .
these transformations are everywhere finite and free of singularity except at the radial coordinate origin . these lead to a lagrangian that reduces to the schwarzschild metric in the weak field limit , but that is nonsingular at the schwarzschild radius .
the theory shows that particles may indeed escape from any initial radial location ( to infinity ) .
a particle s energy is reduced by gravitational red shifting by a factor determined by its initial radial location . from eq .
( [ red_shift ] ) this factor is @xmath126 at the gravitational radius , and @xmath127 at the schwarzschild radius
. therefore black holes are not actually black .
the instantaneous effects of gravitation in the theory presented here and general relativity are related by the factors @xmath128 and @xmath129 respectively .
a numerical comparison of these factors shows that the two theories become distinguishable in the approximate regime of @xmath130 .
the current state of the art as indicated by recent direct observations of object motion near the milky way s central black hole @xcite and observations of multiple black hole objects with the event horizon telescope @xcite indicate that the required regime is being approached although it has not yet been achieved .
additional methods including cumulative effects ( e.g. orbital rotation ) , or gravitational red shift , or doppler effects may also be required .
the author is greatly indebted to dr .
john anderson without whose critique important corrections and improvements would not have been made .
i am also thankful for numerous helpful and formative discussions with dr
. john s. scott .
* gillessen , stefan , f. eisenhauer , s. trippe , t. alexander , r. genzel , f. martins , and t. ott . * `` monitoring stellar orbits around the massive black hole in the galactic center . '' _ the astrophysical journal _ 692 , no . 2 ( 2009 ) : 1075 . | a gravitational theory is formulated by considering the physical processes underlying relativistic dilation of time and contraction of space .
it is shown that the point mass solution of general relativity s field equation - the schwarzschild metric - is a weak - field approximation to the more general lagrangian resulting here . unlike general relativity
the resulting theory does not exhibit a singularity at the schwarzschild radius and furthermore shows that photons escape from that radius with a red shift of @xmath0
. consequently black holes are not black .
experimental confirmations of general relativity that have been conducted in the weak field domain may also be considered as confirmations of the theory presented here since the two theories provide nearly identical predictions there .
testable differentiation of these two theories will require observations in the near vicinity of the scnwarzschild radius of a black hole .
issues related to experimental confirmation are discussed . |
although very few magnetic systems can be solved exactly in three dimensions , there is a reasonable understanding of the physics of simple systems without competing interactions .
numerical estimates for physical quantities exist to at least the accuracy that experiments currently attain .
many systems do exhibit competing interactions however , that is interactions that do not all favour the same ordered state . sufficiently strong competition can lead to new physics which is manifested by the appearance of non - collinear ordering , novel critical exponents , rich phase diagrams , or an absence of long range order at low temperatures .
there are several ways in which frustration can arise . in this article attention
is limited to geometric frustration arising from triangular arrangements of magnetic moments with each pair coupled antiferromagnetically .
figure 1a shows such a situation with three atoms forming an equilateral triangle .
atoms 1 and 2 form a state of lowest energy when their moments are aligned antiparallel , but then atom 3 can not align itself simultaneously antiparallel to the moment on both atoms 1 and 2 , so it is frustrated .
this frustration is most severe for an ising system where spins can align only in one direction . for classical vector spins confined to the plane of the paper ( xy interactions )
there are two degenerate solutions for the lowest energy of the system for a given spin vector at atom 1 .
these are shown in figures 1b and 1c .
both have the three spin vectors at @xmath0 to each other , and the degeneracy corresponds to two different chiral states .
the overall degeneracy of the ground state involves the product of this state and orientation of spin 1 in the plane ; this corresponds to the group @xmath1 . for three dimensional vector spins and isotropic heisenberg interactions
the degeneracy is greater still since the spin on atom 1 can be in any direction and then the spins on atoms 2 and 3 can be in any plane which contains the direction of spin 1.this degeneracy corresponds to the group so(3 ) .
if the hamiltonian has easy axis anisotropy , the plane of the three spins will contain the z axis and the angle between the three spin vectors will no longer be 120@xmath2 .
a general feature of these frustrated systems is that the ground state has extra degeneracy over and above that found in the analogous non - frustrated systems .
this is what gives rises to the possibility of new physics .
there are several ways that triangles of antiferromagnetic interactions can be built into a crystal lattice . in this article
we limit consideration to the effects of frustration in the stacked triangular lattice .
this is a lattice containing two - dimensional triangular sheets coupled by antiferromagnetic interactions @xmath3 , with the sheets stacked perpendicular to the plane with coupling @xmath4 between neighbouring atoms in different planes .
almost all the work on this lattice involves one of two crystal structures , one with composition abx@xmath5 where a is an alkali metal , b is a transition metal , and x is a halogen atom , and the other with composition bx@xmath6 .
the crystal structure of these materials is shown in figure 2 . .
in abx@xmath5 compounds there are chains of magnetic b atoms along the z direction coupled by superexchange interactions through three equivalent anions x. there is no direct superexchange coupling between atoms on neighbouring chains .
the magnitude of the interplanar interactions @xmath4 is typically two to three orders of magnitude greater than intraplanar interactions @xmath3 , so that at high temperatures the magnetic properties become quasi one dimensional . in bx@xmath6 compounds
the strong superexchange coupling is intraplanar and the ratio of the magnitude of the interactions is reversed . at high temperatures the magnetic properties are quasi two dimensional . in a third known type of triangular antiferromagnets , abo@xmath6 compounds , the stacking of two - dimensional triangular sheets along the z direction is different from abx@xmath5 and bx@xmath6 compounds , the stacking sequence is rhombohedral of abcabcabc ... type .
the crystal structure dictates strong two dimensional character of magnetic system in abo@xmath6 compounds : the exchange path of the interplane interaction @xmath3 , b
o a o b , is much longer than that of the intraplane interaction @xmath4 , b b and b o b . at low temperatures
almost all mentioned above materials form three - dimensional ordered magnetic structures that indicate the presence of magnetic frustration in the triangular lattice . in many ,
though not all , cases the low temperature structure is built from one of the structures shown in figure 1 .
kawamura @xcite predicted that the extra degeneracy in the ground state of triangular antiferromagnets leads to new physics , which can be described in terms of universality classes based on the symmetry of the order parameter .
monte carlo work @xcite supports kawamura s prediction for heisenberg symmetry as does 4-@xmath7 renormalization group calculations @xcite .
however the non - linear @xmath8 model in 2+@xmath7 dimensions@xcite indicates non - universal behavior , likely with mean - field tricritical exponents , and 3d renormalization group calculations with a resummation technique show a first - order phase transition @xcite .
the situation is no less clear for xy symmetry .
experimentally stacked triangular antiferromagnets are found to have critical phase transitions with critical exponents that do not correspond to any of the standard universality classes for both heisenberg @xcite and xy @xcite symmetry .
in fact the critical points are found to be tetracritical in character @xcite so that they clearly can not belong in the same universality class as unfrustrated systems .
this shows that the physics of magnetic systems can be changed by frustration .
both experiment and theory indicate that the ordered states found at low temperatures in frustrated systems have reduced ordered moments .
any detailed theory must therefore include quantum effects in determining the ground state , and classical theory may be substantially in error . in fact much of the physics of the stacked triangular lattice is similar to that predicted for the two - dimensional triangular lattice @xcite and the presence of the third dimension is not an essential ingredient of most of the ideas that are used in the field .
however , for all well - characterized materials which have triangular magnetic lattices , the long - range order at low temperatures is three - dimensional in nature . for nearest-heis.texneighbour ferromagnetic or antiferromagnetic interactions , there is no extra frustration involved by going to three dimensions , and the main influence of the presence of the third dimension is that the two - dimensional ordering process is stabilized . in this article
we describe in some detail the magnetic properties of materials with stacked triangular lattices , paying special attention to cases where there are novel physical phenomena absent in the unfrustrated case .
emphasis is placed on the experimental results , though where possible the discussion is put in the appropriate theoretical context .
an additional information about triangular antiferromagnets may be found in the related reviews : @xcite .
we classify the stacked triangular lattice materials by the nature of the magnetic hamiltonian .
chapter [ heis ] describes materials with heisenberg interactions , chapter [ ising ] describes systems with ising hamiltonians , chapter [ sgs ] describes singlet ground state magnets and chapter [ ferro ] considers cases with ferromagnetic stacking of the planes .
chapter [ impurity ] describes magnetic properties of the diluted triangular antiferromagnets .
chapter [ concl ] presents conclusions from these studies .
we conclude this section by listing the materials that have been studied in the context of stacked triangular antiferromagnetism .
\a ) table [ listofmagnstr ] shows the magnetic structure of abx@xmath5 triangular compounds .
the cases where b is a chromium atom or a is a thallium atom are omitted since these compounds do not form with a triangular lattice .
\b ) four bx@xmath6 have been studied : vcl@xmath6 and vbr@xmath6 are close to heisenberg systems @xcite , though there is weak easy - axis anisotropy , the magnetic structure of vi@xmath6 is not clear @xcite .
mnbr@xmath6 has a complex magnetic structure which is not triangular in nature @xcite .
\c ) among abo@xmath6 compounds magnetic properties of only three triangular antiferromagnets have been investigated in details .
licro@xmath6 and cucro@xmath6 demostrate 120@xmath2 magnetic structure with weak easy - axis anisotropy @xcite , agcro@xmath6 has slightly modulated 120@xmath2 structure @xcite .
the stacked triangular magnetic lattice is shown in figure 3 .
its magnetic properties for heisenberg - type antiferromagnetism with single - ion anisotropy can be descibed on the basis of the following hamiltonian : @xmath9 where @xmath10 is a spin of the magnetic ion , @xmath4 is the exchange integral along the @xmath11 axis of the crystal , @xmath3 is the exchange integral in the perpendicular direction , @xmath12 is the anisotropy constant , whose sign determines the orientation of the spin plane relative to the crystal axes .
the first sum describes the exchange energy along the chain , the second sum describes the exchange energy in the basal plane , and the third and fourth sums represent the single - ion anisotropy energy and the zeeman energy of the spins in a external magnetic field @xmath13 respectively .
the case @xmath14 corresponds to a pure heisenberg system .
all real triangular magnets have non zero @xmath12 , but if @xmath15 is small compared with both @xmath16 and @xmath17 the magnetic properties will be close to those of heisenberg systems except at very low temperatures @xmath18 or very close to the critical point .
if @xmath19 the ground state in zero field has spins confined to the @xmath20 plane and in the critical region the fluctuations will only diverge for spin components within the @xmath20 plane .
if @xmath21 the anisotropy energy will be minimised for spins aligned perpendicular to the @xmath20 plane . this term will compete with the antiferromagnetic @xmath3 exchange term and lead to additional frustration .
table [ heis_parameters ] lists the parameters @xmath4 , @xmath3 ( both assumed nearest neighbour interactions only ) and @xmath12 that have been determined experimentally in units of frequency , or energy divided by @xmath22 .
the data confirm our previous observation that abx@xmath5 compounds are quasi - one - dimensional with @xmath23 and that vx@xmath6 compounds are quasi - two - dimensional with @xmath24 . in the abo@xmath6 compounds only the in - plane exchange , @xmath3 has been measured reliably .
vcl@xmath6 and vbr@xmath6 crystallize in the cdi@xmath6 structure with a space group p3m1 .
vi@xmath6 exists in two modifications , black and red .
the black modification is composed of a statistical alternating layer structure of the cdi@xmath6 and the cdcl@xmath6 type , while the red modification crystallizes in the cdi@xmath6 structure @xcite .
this feature may explain the fact that the majority of work on vi@xmath6 was done on powder , rather than on single crystal
it is difficult to prepare good quality single crystal of vi@xmath6 .
vcl@xmath6 and vbr@xmath6 are important because they are the stacked triangular materials with hamiltonians closest to the heisenberg form .
both materials show critical phase transitions from ordered to paramagnetic states in zero field @xcite . despite numerous efforts to determine magnetic structures of bx@xmath6 antiferromagnets , including nmr spectra and relaxation measurements @xcite , esr measurements @xcite and direct neutron scattering measurements @xcite
there are still some open question about spin - structures .
neutron polarization analysis has shown that below @xmath25 the spin structure of vcl@xmath6 is the 120@xmath2 structure in the @xmath26-plane @xcite , while in the case of vbr@xmath6 neutron scattering results are consistent with both the 120@xmath2 structure in the @xmath26-plane and a partially disordered structure whose spins ca nt from the @xmath11-axis by 45@xmath2 @xcite .
vi@xmath6 has only been looked by neutron scattering at in powder form .
two neutron powder measurements @xcite give different patterns .
solution of the structure will probably need single - crystal data , the simple triangular structure is not observed .
there are two phase transitions , a critical transition at 16.3 k and a first - order transition at 14.4 k @xcite .
vcl@xmath6 has weak easy - axis anisotropy as is shown by the data in table [ heis_parameters ] and also by the splitting of the critical point from the single tetracritical point of the frustrated heisenberg hamiltonian into two ordinary critical transitions at @xmath27 k and at @xmath28 k in zero field @xcite .
this splitting of about one part in 450 will give rise to crossover behaviour in the critical region @xcite .
for reduced temperatures @xmath29 of magnitude more than about 1/450 above @xmath30 or below @xmath31 the critical behaviour will be that of the frustrated heisenberg system .
nearer to @xmath30 or @xmath31 the critical behaviour will reflect the fact that the correlation length only diverges along the @xmath32 direction , not in the xy plane , and will be different .
thus it is possible in principle to measure three sets of critical exponents in vcl@xmath6 , though in practice only the heisenberg exponents for @xmath33 have been measured . in vbr@xmath6
the splitting of @xmath34 k is even smaller and has not been observed @xcite .
the measured critical exponents should correspond to those of the frustrated heisenberg system .
table [ critexp ] lists the observed critical exponents in vcl@xmath6 and vbr@xmath6 and compares them with theoretical predictions .
it is clear that the frustrated heisenberg system has critical exponents that are far from those of the unfrustrated system , confirming that frustration changes the physics .
neither the so(3 ) nor tricritical model are in complete accord with experiment ; the so(3 ) model is in disagreement with experiment for @xmath35 and @xmath36 and the tricritical exponents are in disagreement for @xmath37 and @xmath38 . in every case
both these models are significantly better than the unfrustrated heisenberg model .
mnbr@xmath6 also does not order in the triangular structure .
its structure , based on single - crystal neutron data , is complex based on arrangements with two up spins followed by two down spins @xcite .
there are two magnetic phase transitions : second order at @xmath392.32 k and first order at @xmath402.17 k. licro@xmath6 has been studied by single crystal susceptibility and neutron diffraction measurements @xcite , optical measurements @xcite , esr @xcite and neutron diffraction measurements @xcite on a powder sample
. three dimensional magnetic ordering , characterized by a double-@xmath41 120@xmath2 structure with non - equivalent wave vectors @xmath41 of @xmath42 and @xmath43 @xcite , is observed below the single - phase - transition temperature @xmath44 k. polarization analysis of neutron scattering data shows that spins triangulars are confined in a plane including the @xmath11 axis , that is the magnetic anisotropy is of the easy - axis type .
on the other hand , absence of splitting of @xmath25 and of anisotropy in the susceptibility above @xmath25 demonstrate that magnetic anisotropy @xmath12 is much smaller than exchange interactions @xmath4 and @xmath3 .
no direct measurements of @xmath12 and @xmath3 have been reported , the only an attempt to estimate the @xmath45 ratio from the phase transition temperature has been made by angelov and doumerc @xcite , which gives only a very rough estimate .
the ratio @xmath45 could also be estimated from susceptibility measurements since it is proportional to the @xmath46=5% ratio ( note the difference between licro@xmath6 and vx@xmath6 compounds : in licro@xmath6 @xmath47 ) .
the optical and esr measurements were mostly devoted to the problem of finding of the characteristic point - defects known as @xmath48 vortices , predicted theoretically by kawamura and miyashita @xcite .
it is believed the line width of the exciton magnon absorption @xcite and epr linewidth @xcite exhibit @xmath48-vortex induced broadening .
the rest of the abo@xmath6-type triangular antiferromagnets have been studied only in a powder form .
agcro@xmath6 has been reported to order magnetically below 24 k. it forms a slightly modulated 120@xmath2 structure @xcite with magnetic peaks at ( 0.327 0.327 0 ) .
the width of peaks indicates that the development of the true long - range magnetic order is suppressed .
results of neutron powder diffraction studies on cucro@xmath6 below @xmath49 k are consistent with the 120@xmath2 structure in the @xmath50-@xmath11 plane with moment @xmath51 @xcite .
no long range magnetic ordering has been found in natio@xmath6 and linio@xmath6 down to 1.4 k @xcite and in nacro@xmath6 and kcro@xmath6 down to 2 k @xcite .
an interesting issue is a magnetic moment reduction at low temperature .
there are three general reasons which effect the value of magnetic moments in abx@xmath5 and vx@xmath6 compounds : 1 ) covalency reduction , 2 ) effect of frustration and 3 ) quantum fluctuations enhanced by low - dimensionality of the magnetic system . according to high temperature susceptibility measurements @xcite , the moment of v@xmath52
can be estimated as 3.96 , 4.07 and 4.07 @xmath53 for vcl@xmath6 , vbr@xmath6 and vi@xmath6 respectively , suggesting small covalency reduction . from the low - temperature neutron diffraction data an average moment in vcl@xmath6 is @xmath54 @xcite , and in vbr@xmath6 @xmath55 or @xmath56 depending upon magnetic structure assumed @xcite .
the relatively small reduction of the magnetic moment in the two - dimensional systems in comparison with almost 50% reduction in some abx@xmath5 one - dimensional systems ( see next section ) indicates dimensionality ( quasi - two - dimensional rather than quasi - one - dimensional ) and quantum fluctuations are the major influences on the moment reduction in the stacked triangular lattice .
note , that in theory an average magnetic moment of two - dimensional magnetic system is nonzero , while in one - dimensional system @xmath57 .
in this section we describe materials with a hamiltonian ( equation [ h_h ] ) that has an exchange term and a single - ion anisotropy with @xmath21 .
this negative value of @xmath12 makes it energetically favourable for the moments to align perpendicular to the @xmath58 plane .
this breaks the isotropic symmetry of the heisenberg hamiltonian and leads to changes in the physics .
there are five materials with this hamiltonian , all of composition abx@xmath5 , with space group @xmath59 and a stacked triangular lattice of spins .
these are csnicl@xmath5 , rbnicl@xmath5 , csnibr@xmath5 , rbnibr@xmath5 and csmni@xmath5 .
all exhibit the quasi - one - dimensional behaviour typical of this crystal structure with the value of the exchange interaction , @xmath4 , along @xmath11 more than an order of magnitude greater than the intra - planar exchange interaction , @xmath60 , and with the magnitude of @xmath12 of the same order as that of @xmath60 .
experimental values of @xmath4 , @xmath60 and @xmath12 were given in table [ heis_parameters ] .
in all cases the exchange interaction is antiferromagnetic . in some cases there
are major discrepancies between experimental results ; we will look at this later in this section .
all five materials have a magnetic phase diagram as shown in figure 4 . at low temperatures and magnetic fields the magnetic structure
is observed to be the triangular structure with the @xmath11 axis in the plane of the triangle .
the anisotropy favours alignment of spins perpendicular to the basal plane resulting in structures where the angle @xmath61 in the figure is less than 60@xmath62 .
each @xmath58 plane has a net moment perpendicular to it which can be observed by neutron diffraction .
however the planes are stacked antiferromagnetically because of the large antiferromagnet exchange interaction , @xmath4 , along the chains .
this results in the whole crystal being an antiferromagnet . for a classical system @xmath63 at @xmath64
the triangular structure collapses into a collinear structure with two spins along @xmath65 and one spin along @xmath66 .
the anisotropy is not large enough for this to happen in any of the materials considered in this section .
chapter [ ising ] describes collapsed cases . at a temperature @xmath31
there is a critical phase transition to the colinear structure as shown in figure 4 .
thus for classical spins , as @xmath67 increases to 3@xmath3 , @xmath31 decreases to zero . at a higher temperature , @xmath68
, there is a second critical phase transition , this time to the paramagnetic state .
the difference @xmath69 is a measure of the relative strength of the anisotropy , @xmath12 , and the exchange , @xmath3 .
the colinear structure has energy independent of applied field , while for all the competing structures the energy decreases as the field increases .
thus it becomes less stable as the field increases . as a field
is applied along the @xmath11 axis at low temperature the angle @xmath61 changes in different directions for successive planes ; in the plane shown in figure 4 the angle @xmath61 increases on application of a field along @xmath32 . at a field @xmath70
there is a first - order phase transition to the canted structure shown in figure 4 . in this structure
the @xmath20 component of the moments form an equilateral triangle , while the @xmath32 components are aligned ferromagnetically along the field direction . for a classical system with @xmath4 large compared with @xmath3 or @xmath67 , tanaka _ et al . _ @xcite show that @xmath71 at @xmath72 .
it should be noted that the expression given by tanaka _
_ differs from that given above by a factor of 2 because they define @xmath4 to be half of our @xmath4 .
there is a multicritical point at @xmath73 , @xmath74 where all four phases meet .
alternatively , this point can be described as the intersection of three lines of critical phase transitions and one line of first - order phase transitions .
the experimentally - determined parameters of the phase diagram , @xmath30 , @xmath31 , @xmath73 , @xmath70 , @xmath74 and @xmath61 are given in table [ hea_tab ] .
there is reasonable agreement between different experimental measurements here , in contrast to the descrepancies noted between some of the measurements of the parameters of the hamiltonian ( table [ heis_parameters ] ) .
now that we have given an overview of the properties of materials with easy - axis anisotropy , we go back to examining the experimental values of the parameters in the hamiltonian , and to checking consistency with phase - diagram parameters .
one reason for inconsistency is that it is believed that the neutron scattering values of -13.0 ghz for csnicl@xmath5 and -0.5 for csmni@xmath5 are in error because of incorrect branch assignments .
a second problem is that the resonance experiments actually measure @xmath70 and then derive @xmath12 from equation [ hc ] and known values of @xmath4 .
unfortunately the equation may not hold for @xmath75 .
the difficulty arises from the need to incorporate quantum fluctuations into the theory .
a well - established technique for taking these into account involves an expansion in powers of @xmath76 @xcite .
zhitomirsky and zaliznyak @xcite show that inclusion of this term in equation [ hc ] yields a negative value for @xmath77 .
although this is non - physical , the result serves notice that equation [ hc ] is not good for values of @xmath78 as small as 1 .
furthermore the first term in the quantum correction for @xmath75 renormalizes @xmath4 by 18@xmath79 and @xmath12 by -50@xmath79 .
most experiments will measure the renormalized values , not the bare values .
a further difficulty with the nickel compounds is that because of their quasi - one - dimensional properties , there may be vestiges of the haldane effect in their low - temperature properties .
of course a full quantum treatment will take this into account , but it may not be apparent in the first term of a @xmath76 expansion .
several neutron scattering experiments claim to see haldane - gap effects in the three - dimensional ordered phase @xcite and conventional spin - wave theory ( i.e. spin - wave theory without quantum corrections ) does not fit the measured dispersion relations . without further theoretical guidance
there must be some doubts about the reported values of @xmath4 , @xmath3 and @xmath12 for the nickel compounds even in cases where the experimental data is irreproachable .
affleck@xcite has argued haldane correlations will still be effective in the three - dimensionally ordered phase .
his treatment shows much better qualitative agreement with experiment than linear spin - wave theory and provides a theoretical underpinning for the claims that the experiments constitute solid confirmation of the existence of the haldane gap . in large fields , @xmath80 and @xmath73 , the quantum fluctuations are reduced!@xcite and the interpretational difficulties are less severe . in csmni@xmath5
the quantum corrections are smaller ; in leading order by a factor of 2.5 so that they are of the same order as the experimental uncertainties .
the spin wave dispersion relations follow conventional spin wave theory @xcite .
the values of @xmath4 , @xmath3 and @xmath12 quoted in the table are the directly measured values without application of any quantum corrections .
it is interesting to consider the effects of increasing departures of the hamiltonian from the heisenberg form in easy - axis systems .
this is equivalent to increasing @xmath15 while keeping the exchange parameters @xmath4 and @xmath3 fixed .
even small values of @xmath15 break the symmetry by aligning the moments triangles in a plane that contains the @xmath81 axis and splitting the critical point into two critical points at @xmath30 and @xmath31 .
@xmath31 decreases until it becomes zero at @xmath82 .
@xmath30 and @xmath73 increase as @xmath15 increases , as can be seen for instance by comparing csnicl@xmath5 with csnibr@xmath5 where @xmath4 and @xmath3 are similar but the values of @xmath12 differ by an order of magnitude .
@xmath70 increases as @xmath83 and experiment seems to indicate that @xmath84 .
we now discuss the phase diagram and critical properties of the easy - axis magnets .
the theory is constructed assuming that the parameters of the system are such as to give a phase diagram of the type shown in figure 4 with the magnetic structures that we have described .
thus antiferromagnetic heisenberg interactions @xmath4 and @xmath3 are taken on a stacked triangular lattice with easy - axis anisotropy @xmath85 . then , building on the ideas of chiral universality classes described earlier , theory makes a number of predictions based on landau - type theories , scaling and renormalization group calculations .
there is a multicritical point connecting a line of first - order phase transitions and three lines of critical phase transitions @xcite .
all three lines of critical points come in tangentially to the first order spin - flop line @xcite .
the lines of phase transitions through @xmath30 and @xmath31 both follow the @xmath20 universality class @xcite .
4.the phase transition from the paramagnetic state to the spin flop state follows the chiral @xmath20 universality class @xcite .
\5 . at the multicritical point
the phase transition should follow the chiral heisenberg universality class @xcite .
the transition region around the multicritical point should not be large @xcite .
the first of these predictions is obeyed by all the materials studied in this section . the experimental evidence from the highest accuracy phase diagrams in csnicl@xmath5 @xcite ,
csnibr@xmath5 @xcite and csmni@xmath5 @xcite suggests that the second prediction is also good .
the experiments all clearly show that the two lines of phase transitions to paramagnetism come in parallel to the line of first - order phase transitions and give clear indications that there is curvature in the line of phase transitions through @xmath31 close to the multicritical point such as to make the predicted effect likely .
table [ hea_crit ] lists the experimental determinations of the critical exponents @xmath35 , @xmath37 , @xmath36 and @xmath38 and compares the results with predictions for various universality classes . as well as the critical exponents ,
the table also compares experiment and theory for the amplitude ratio , above and below @xmath25 , of the specific heat divergence .
inspection of the table shows that there is generally good agreement between experiment and theory for the specific heat data , but that the agreement with the neutron data at @xmath86 for @xmath35 , @xmath36 and @xmath37 is not good .
recent neutron measurements @xcite for @xmath35 at the multicritical point and for the spin - flop - to - paramagnetic phase transition agree well with theoretical predictions .
we conclude that the theoretical predictions 1,2,4 and 5 above are confirmed by experiment .
prediction 6 has not been tested .
experiment does not bear out prediction 3 that the critical phase transitions at @xmath30 and @xmath31 follow the @xmath20 universality class .
the measured indices do not fall within any standard universality class , but they do seem to be the same at the two transitions as predicted . the neutron data could be described numerically by chiral heisenberg critical exponents , suggesting that the crossover from the multicritical point takes place slowly , but the specific heat data and theoretical prediction number 6 both argue against this possibility .
the presence of an anisotropy term in the hamiltonian with @xmath19 favours the confinement of spins to the @xmath20 plane .
the ordered structure in the absence of a field is an equilateral triangle of spins with two chirally - degenerate states as was shown in figure 1 . for any atom with spin aligned along its local @xmath81 direction ,
the local @xmath87 degeneracy of the heisenberg hamiltonian is broken since a rotation of spins in the plane costs no anisotropy energy while a rotation out of the plane does cost anisotropy energy .
this splits a degeneracy in the spin wave excitations , with one acoustic branch having zero energy at magnetic reciprocal lattice points while the other exhibits a gap . near the critical point
the fluctuations will tend to diverge in the @xmath20 plane , but not in the @xmath32 direction , so that the critical exponents will correspond to the chiral @xmath20 , or @xmath1 , model .
if @xmath12 is small there should be a crossover to heisenberg exponents further away from the critical point .
this effect has not been seen to date since only one easy - plane material , csmnbr@xmath5 , has had its critical exponents measured extensively , and in this material @xmath12 is not small compared with @xmath3 .
the behaviour of easy - plane triangular antiferromagnets in a small field perpendicular to @xmath32 involves a competition between the in - plane exchange energy @xmath3 and the anisotropy energy @xmath12 .
the @xmath3 term is of lowest energy when the moments are aligned in the 120@xmath2 structure perpendicularly to the field with a canting of the moments towards the field direction , while the anisotropy term favours 120@xmath2 structure aligned in the @xmath20 plane .
what happens as the field is increased depends on the relative magnitudes of @xmath12 and @xmath3 @xcite . at low temperatures , if @xmath88 , a field in the @xmath20 plane larger than @xmath89 will flip the plane of the spin triangle so that it is perpendicular to the field .
this costs anisotropy energy , has virtually no cost in exchange energy and gains energy from a canting of the spins along the field direction .
the spin - flop phase transition will be of first order at @xmath90 if the anisotropy energy is larger , @xmath91 , the ground state above a critical field @xmath70 is one where the spins remain in the plane , but the triangular structure collapses to a colinear structure with two spins in one direction in the plane normal to the field direction and one spin in the opposite direction . this structure costs exchange energy @xmath3 , has no cost in anisotropy energy , and gains energy by a canting of the spins towards the field direction . the phase transition will be continuous with @xmath92 it is clear that of these two cases , it is the latter with @xmath91 that corresponds the more closely to the chiral @xmath20 model .
table [ hep_table ] lists the easy - plane triangular antiferromagnets .
all have the abx@xmath5 structure with strong exchange interactions @xmath4 and relatively weak in - plane interactions @xmath3 . in the table
we give the type of antiferromagnet , the space group , the nel temperature , the critical magnetic field and the aligned magnetic moment at low temperatures . in the rest of this section
we discuss separately the three cases : small @xmath12 , large @xmath12 and those where the crystal structure is distorted .
there are four materials which are known to be in this category , csvcl@xmath5 , csvbr@xmath5 , csvi@xmath5 and rbvcl@xmath5 . the crystal and magnetic structures were established by zandbergen @xcite and by hauser _
low field susceptibility measurements by niel _
@xcite on powder samples in the quasi - one - dimensional region ( @xmath93 ) established the intrachain exchange parameter @xmath4 for the cesium compounds .
@xcite have measured the spin wave dispersion relations in csvcl@xmath5 , csvbr@xmath5 and csvi@xmath5 by neutron inelastic scattering .
the measurements were confined to the ( @xmath94 ) direction of reciprocal space at low temperatures . the results fitted reasonably with the predictions of linear spin wave theory given by kadowaki _
et al._@xcite for nearest neighbour interactions and single - ion anisotropy .
the restriction of the data to the plane results in the fitting only giving values for @xmath95 and @xmath96 and a value of @xmath4 was taken from the paramagnetic magnetization measurements of niel _ et al . _ in order to determine the values of @xmath97 and @xmath12 given in table [ hep_table ] .
no zero - point - motion correction was made for the reduced moment and one - dimensionality so the real values of @xmath98 and @xmath99 are lower than those quoted . on average
the observed ordered moment on the vanadium atom , as taken from table [ hep_table ] , is 1.9 @xmath53 . for @xmath100 with @xmath101
the maximum moment is 3 @xmath53 , so the reduction is 63% and the correction to linear spin - wave theory is appreciable .
little is known experimentally of the phase diagrams or critical properties of these materials .
there has been no investigation of the nature of the phase transition in zero field .
as pointed out earlier , this is an unresolved area in the theory , with predictions of chiral xy critical exponents , tricritical exponents and a weak first - order phase transition in the literature
. it would be interesting to discover what experiment has to say .
the phase diagram for a field @xmath102 applied in the @xmath20 plane is expected to be as shown in figure [ hep_smalld_phd ] .
phase @xmath103 is the plane triangular antiferromagnetic structure , phase @xmath104 is the colinear structure , phase @xmath105 is the spin - flopped triangular structure and phase @xmath106 is paramagnetic .
molecular field calculations by plumer _
_ @xcite predict this phase diagram or similar ones differing only by the presence of a narrow region of an extra phase near the first order transition line .
the prediction was actually for the @xmath107 hamiltonian which is slightly different from our hamiltonian ( eq.[h_h ] ) , but the differences are not expected to affect the overall pattern significantly . at @xmath108 and @xmath109
there is a meeting of four phases since at negative @xmath102 there is a similar phase to phase @xmath104 with the canting in the opposite direction ; thus it is a tetracritical point .
the tetracritical nature of the transition shows that it does not belong in the same universality class as the @xmath20 model .
only one critical exponent , @xmath110 , has been determined in these materials . near the tetracritical point
the two phase boundaries are predicted to behave as @xcite @xmath111 where @xmath112 has one of two values depending on whether @xmath113 refers to the phase boundary @xmath104 to paramagnetic or @xmath103 to @xmath104
. actual determination of these exponents is quite sensitive to the value chosen for @xmath25 and to the range of data over which the fit is made @xcite .
the sum of the two exponents @xmath110 is much less sensitive to the value of @xmath25 than is the individual values of @xmath110 . in csvbr@xmath5 ,
tanaka _ et al . _
@xcite have determined the critical exponents @xmath110 in a field perpendicular to the @xmath11 direction from susceptibility data .
they find @xmath114 and @xmath115 . scaling and renormalization group theory @xcite gives @xmath116 , with @xmath117 ;
for the @xmath20 chiral model @xmath36 is expected to be only slightly greater than one , probably about 1.1 .
thus the experimental values of @xmath110 are significantly lower than predicted by theory .
the discrepancy can not lie in the sensitivity of the fits to the value of @xmath25 , since the sum of the two individual values of @xmath110 is just as far from the prediction as is the individual values .
this is the unusual and probably the most interesting case .
although there is only one example of undistorted triangular easy - plane antiferromagnet which satisfies this condition , csmnbr@xmath5 , it has attracted a lot of both theoretical @xcite and experimental attention through neutron scattering @xcite , magnetization @xcite , esr @xcite and specific heat measurements @xcite . as was pointed out in the introduction to this chapter , if @xmath118 the application of a field perpendicular to @xmath11 no longer produces the spin - flop phase shown in fig .
[ hep_smalld_phd ] .
instead there is a critical phase transition to an almost colinear structure at a critical field which is given by equation [ chub_hc ] at @xmath72 .
the phase diagram is shown in fig 6 . at low temperatures and fields the magnetic structure is the 120@xmath62 stacked triangular structure with the anisotropy confining the moments to within the @xmath58 plane .
the critical point at @xmath109 and @xmath108 is a tetracritical point just as for the case @xmath119 described in the previous section .
in the high - field phase two of the three moments on the triangular lattice become parallel with the third in the antiparallel direction .
the moment directions are in the basal plane almost perpendicular to the applied field , but with a small canting towards the field direction proportional to @xmath120 . in this phase
there is a softening of one of the exchange branches found by esr @xcite .
the nature of the phase diagram was discovered through neutron - scattering measurements of the temperature dependence of the intensity of magnetic bragg peaks in different fields as shown in fig .
[ csmnbr3_int ] @xcite .
these results were later confirmed by measurements of the field and temperature dependence of the magnetization as shown in fig .
[ csmnbr3_m ] a ) and b ) .
the magnetization process is described satisfactorily in terms of linear spin - wave theory except for two features : ( i ) the measured torques are significantly smaller than predicted and ( ii ) in high fields , h@xmath121h@xmath122 , there is a considerable anisotropy between the magnetization when the field is applied along and perpendicular to the @xmath11 axis .
it has been suggested by abarzhi _
_ @xcite that quantum fluctuations are responsible for both these effects .
this claim has bee supported by subsequent work @xcite but recently has beem disputed by santini _
et al . _
@xcite , at least for csmnbr@xmath5 and rbmnbr@xmath5 , where they claimed that the anisotropy is already present at the classical level , provided thermal fluctuations are taken into account .
santini s work is based on a hamiltonian similar to ( [ h_h ] ) with the @xmath3 term replaced by a dipole - dipole interaction .
a peculiar feature of the phase transition from a triangular to a colinear structure is that it survives even if the field direction deviates significantly from the basal plane .
the value of the critical field follows the equation @xmath123 where @xmath124 and @xmath125 is the angle between the magnetic field and the basal plane , rather than simple field projection on the spin s plane , @xmath126 @xcite .
kawamura @xcite predicted that the easy - plane materials would not follow the @xmath20 universality class in their critical properties in zero field because of the extra chiral degeneracy of the order parameter .
instead a new universality class , known as the chiral @xmath20 class , would apply . the discovery that the critical point is tetracritical confirms in a simple way that regular @xmath20 universality does not apply .
the most recent monte - carlo work , done for the case of ferromagnetic interactions along the c direction , favours a weak first - order phase transition @xcite , with effective critical exponents that are reasonably in agreement with those given by kawamura for the chiral xy model , less than three errors apart in all cases .
all the known undistorted easy - plane triangular materials are reported experimentally to show critical phase - transitions in zero field in agreement with kawamura s predictions .
of course it is never possible experimentally to rule out a first - order phase transition if it is sufficiently weak , but critical behaviour is observed to persist at least down to reduced temperatures of 2@xmath12710@xmath128 @xcite .
the measured values of the critical indices are listed in table [ critind ] and compared with the predictions of four possible models : chiral @xmath20 , chiral heisenberg , regular @xmath20 and tricritical in mean - field theory .
the exponent @xmath129 is the mean of @xmath130 and @xmath131 .
we use this parameter because gaulin @xcite has shown that while experimental determinations of @xmath130 and @xmath131 are highly sensitive to the value of @xmath25 , the mean of the two is much less sensitive , and hence is determined more reliably in experiments .
the exponent @xmath81 is the dynamic critical exponent measured from the temperature dependence of the energy of long - wavelength spin waves @xcite near @xmath25 .
the experimental results fit both the chiral @xmath20 and the mean - field tricritical models satisfactorily , while they do not fit the other two models .
it is not understood why theory and experiment agree so well for the easy - plane case , but less satisfactorily for heisenberg and easy - axis materials .
it is not uncommon for the magnetic abx@xmath5 compounds to experience a crystal distortion .
a different types of crystal structure distortions have been found at sufficiently low temperature in knicl@xmath5 , rbmnbr@xmath5 , rbvbr@xmath5 , rbvi@xmath5 , rbtii@xmath5 and rbfebr@xmath5 .
generally lattice deformations due to a structural phase transition lead to some modifications of magnetic interactions ( the crystal distortions break the symmetry and so change the exchange interaction between neighbouring in - plane magnetic ions ) and , consequently , to a partial lifting of frustration on a stacked triangular lattice
. study of such partially frustrated systems is of a fundamental interest , because they do not simply correspond to an intermediate case between unfrustrated and frustrated magnets but show novel physical phenomena absent in the two limiting cases .
phase transitions to lattices of lower symmetry with decreasing temperature are characterized usually by displacements of chains of magnetic atoms as a whole without deformation so that the intrachain distance between spins remains unchanged .
the typical structural transition to the lattice of @xmath132 space group is accompanied by the shift of one from the three adjacent chains upward along the @xmath11-axis while the two others shift downward keeping the crystal center of mass undisplaced .
this primary distortion is shown in figure [ knicl_dist ] .
the crystal unit cell in the basal plane is enlarged to become @xmath133 ( fig.[knicl_dist ] ) , preserving the hexagonal symmetry . because chains are shifted usually on a small distance from the basal plane ( @xmath134 in rbfebr@xmath5 @xcite ) , magnetic properties may be considered by placing spins on the same stacked triangular lattice and changing interactions in the hamiltonian ( [ h_h ] ) in accordance with a reduced symmetry of the crystal structure .
the distortion has little effect on the exchange @xmath4 along the @xmath81 direction but in the xy plane the interaction @xmath3 is split into two different interactions , @xmath135 and @xmath136 as shown in figure [ mag_dist]a .
the @xmath137 triangular structure corresponds to @xmath138 and small departures from this condition give triangular structures with angles not equal to 120@xmath2 .
mean - field investigations have been carried out to determine the full phase diagram as the ratio of @xmath3 to @xmath139 is changed @xcite .
the hamiltonian of this centered honeycomb model " ( in terms of zhang _
et al_. @xcite ) is obtained by the evident replacement of the second term on the r.h.s . of eq .
( [ h_h ] ) : @xmath140 without field the spin ordering occurs in two steps with additional intermediate collinear phase between @xmath30 and @xmath31 which is either ferromagnetic or partially disordered @xcite .
the splitting of @xmath25 was clearly seen in rbvbr@xmath5 @xcite .
the behaviour of the system in the applied magnetic field depends on the relative strength of exchange constants @xmath3 and @xmath139 .
note , that due to superexchange character of the interchain interaction , it depends in a complicated way from the interatomic distances and bond angles .
the critical exponent @xmath35 at t@xmath141 is observed @xcite to be 0.32(1 ) , while theory predicts that this phase transition should follow the xy model with @xmath35 = 0.35 . if @xmath142 , that is coupling between in - plane spins ( a@xmath143a@xmath6 in fig .
[ mag_dist]a ) is stronger than coupling between in - plane and out - of - plane spins ( a b ) , then the presence of the distortion does not change nature of the phase transition to a colinear phase .
critical field is given in this case by the formula @xcite : @xmath144 which is very similar to a formula ( [ chub_hc ] ) when @xmath3 and @xmath139 are close to each other . if @xmath142 , then at low magnetic field spin of the out - of - plane atom ( marked b in fig . [ mag_dist]a ) is aligned parallel to the field .
such a configuration is energetically unfavourable at higher fields , therefore an additional phase transition occurs at @xmath145 , when sublattice b starts to deviate from the field direction @xcite .
finally , the transition to a collinear phase occurs at @xmath146 which is again very similar to a @xmath70 in undistorted triangular structure .
the above described theory was developed to explore the magnetic consequences of the crystal phase transition @xmath147 .
distortions of this type were found in the low temperature phase of rbfebr@xmath5 @xcite , at the room temperatures in knicl@xmath5 @xcite , and in rbmnbr@xmath5 @xcite and , probably , in rbvbr@xmath5 @xcite .
but , as it will be evident from the discussion given below , in at least two compounds , rbmnbr@xmath5 and knicl@xmath5 , further crystal phase transitions just below room temperature play an important role . in knicl@xmath5
dielectric anomalies indicating structural phase transitions are found at 274 k , 285 k , 561 k and 762 k @xcite .
a single crystal x - ray study on the _ low - temperature _ structure of knicl@xmath5 @xcite shows clearly the existence of two crystal structure distortions , as originally found by neutron scattering @xcite .
one phase ( denoted as phase _ a _ ) is hexagonal and does not differ much from the room temperature structure , the other phase ( phase _ b _ ) is orthorhombic . in a phase
_ a _ the unit cell is rotated through @xmath148 about the @xmath11 axis from the room temperature unit cell and enlarged to @xmath149 , @xmath149 and @xmath11 ; in a phase _ b _ the low temperature unit cell has sizes @xmath150 , @xmath50 and @xmath11 . the main feature of the phase _
b _ is a sinusoidal modulation of the ion chains in the basal plane : instead of the room temperature sequence 0 - 0-up-0 - 0-up-0 - 0 , where 0 " means ion in the basal plane and up " means ion slightly shifted above the basal plane along the @xmath11-axis , at low temperature the sequence is 0-up-0-dn-0-up-0-dn-0 , where dn " means the ion is shifted below basal plane .
possible space groups are @xmath151 and @xmath152 @xcite . as a consequence of the existence of two different crystal modifications , two different magnetic structures have been observed with @xmath153 k and @xmath154 k in phases _ a _ and _ b _ respectively @xcite .
magnetization measurements @xcite and measurements of esr @xcite showed that in phase _ a _ magnetic structure is a distorted triangular with @xmath142 .
magnetic structure of phase _ b _ we discusse below , for now we just note , that in both cases magnetic structure is commensurate . in rbmnbr@xmath5
the situation with crystal distortion is very similar : the neutron measurements of heller _ et al .
_ @xcite show a crystal which seems to contain both the _ a _ and _ b _ phases , while the measurements of kato _ et al .
_ @xcite show a crystal where only the _
b _ phase is present .
moreover , recent x - ray scattering measurements @xcite shows total identity of rnmnbr@xmath5 and knicl@xmath5 phase _
b _ crystal structures .
such an identity of crystal structure makes it hard to explain difference in magnetic structure . from a symmetry point of view _
b_-phase corresponds to the row model " of zhang _
@xcite shown on fig .
[ mag_dist]b .
anticipation of this model results in appearance of incommensurate magnetic structure @xcite . in rbmnbr@xmath5
the magnetic structure is indeed incommensurate at low magnetic fields and only if the magnetic field exceeds 3 t it became commensurate @xcite .
incommensurate - commensurate phase transition is accompanied by the hysteresis phenomena in magnetization @xcite , resonance power absorption @xcite and magnetic bragg - peaks intensity @xcite .
the overall @xmath155 phase diagram , which is much complicated and includes two incommensurate phases , two commensurate phases and paramagnetic one , was successfully explained in terms of landau theory using row model " @xcite . in phase _
b _ of knicl@xmath5 the magnetic structure is commensurate even in a zero magnetic field @xcite and identical to the high - field structure of rbmnbr@xmath5 .
what causes the stabilization of the commensurate spin configuration in knicl@xmath5 remains unknown .
the triangular antiferromagnets of type abx@xmath5 with b a cobalt atom have properties that are like those of ising antiferromagnets .
the cobalt cation lies in an octahedron of x anions with a slight trigonal distortion .
the strong crystal field splits the lowest lying @xmath156f configuration so that a @xmath156t@xmath143 state has the lowest energy .
this corresponds to a kramer s doublet which is effectively an @xmath157 state with the moment lying either parallel or antiparallel to the @xmath11 axis .
there is a mixing between this state and a state of higher energy also with @xmath156t@xmath143 symmetry and the resultant exchange hamiltonian at low temperature can be described by the equation @xcite @xmath158 + j'\!\sum_{k , l}^{\text{planes}}s^z_k s^z_l - g \mu_b h\sum_i s^z_i \label{h_i}\ ] ] with @xmath159 . in every case @xmath7 is small , about 0.1 , so that the first term in the hamiltonian , which is of ising type , is the dominant term . as in all the abx@xmath5 triangular systems
the magnitude of the inter - chain exchange constant @xmath3 is small compared with the exchange constant @xmath4 along the chains .
@xmath102 is an external magnetic field applied along the @xmath81 direction .
a weak single - ion exchange - mixing term can be neglected for the purposes of this article .
the parameter @xmath7 has a weak temperature dependence , decreasing as the temperature is raised @xcite .
there are four abx@xmath5 compounds that have the magnetic hamiltonian given above , cscocl@xmath5 , rbcocl@xmath5 , cscobr@xmath5 and rbcobr@xmath5 .
there has been extensive work on the first three of these , and table [ ising_par ] lists the experimentally - determined values of the parameters in the hamiltonian for each compound .
we also list values of phase - transition temperatures in zero field , @xmath25 , and of applied fields , @xmath70 , at which there are phase transitions .
frustration effects are more acute in ising triangular antiferromagnets than in xy or heisenberg systems . in the latter cases the frustration can be partially relieved by the formation of spin triangle
as was shown in figure [ geom_frust ] b ) and c ) , while this is not possible for the ising antiferromagnet where spins are confined to directions parallel and antiparallel to @xmath81 .
in contrast to the unfrustrated case , wannier @xcite showed that the two - dimensional nearest - neighbour triangular antiferromagnet is disordered at all finite temperatures and has a critical point at @xmath72 . again
the frustration changes the underlying physics . in the stacked triangular lattice
there is no frustration for nearest - neighbour interactions along the stacking direction , whether they are ferromagnetic or antiferromagnetic .
this lessening of the overall frustration allows long - range ordering at low temperatures .
figure [ ising_spin_arr ] shows three examples of possible ordering in the basal plane where sites marked @xmath160 have @xmath161 , sites marked @xmath162 have @xmath163 , sites marked @xmath164 have @xmath165 randomly distributed with @xmath166 , and sites marked @xmath167 have @xmath165 randomly distributed with @xmath168 . in each case the unit cell is @xmath169 by @xmath169 , where @xmath50 is the lattice constant of the triangular lattice .
the states shown in a ) and c ) are not fully ordered , while that shown in b ) is fully ordered and ferrimagnetic .
there is no net ferrimagnetism over the whole crystal in this state however since the strong antiferromagnetic exchange along @xmath11 ensures that the magnetic moment in a given plane is cancelled by an equal and opposite moment on the next succeeding plane .
the three ordered arrangements shown in figure [ ising_spin_arr ] have the same energy , as do many other configurations , so the ground state has a high degree of degeneracy .
the unit cell for the three states is the same and their neutron diffraction patterns are similar . in real materials rather small effects may enable the degeneracy to be split in favour of one particular ground state . for instance small ferromagnetic next - nearest - neighbour in - plane interactions will stabilize state b ) , while such small antiferromagnetic terms will stabilize state a ) . in the known ising abx@xmath5 compounds neutron diffraction in fact
indicates the presence of long - range magnetic order .
there are sharp peaks of magnetic origin in the neutron - diffraction pattern below a temperature @xmath30 .
these peaks are located in reciprocal space at ( @xmath170 @xmath170 @xmath171 ) , where @xmath22 is an integer not divisible by three and @xmath172 is an odd integer .
the condition @xmath172 odd implies antiferromagnetic stacking along the @xmath81 direction , since the abx@xmath5 structure has @xmath11 spacing equal to twice the interplanar spacing .
this stacking reflects the strong antiferromagnetic coupling along the @xmath11 axis .
the @xmath170 factors indicate that the unit cell in the xy plane is @xmath169 by @xmath169 .
this is not an unexpected result in view of the discussion earlier regarding figure [ ising_spin_arr ] . in figures [ fyelon ] and [ fmekata ] , taken from yelon _ et al .
_ @xcite and mekata and adachi @xcite respectively the temperature dependence of some observed neutron - scattering bragg peaks is shown for cscobr@xmath5 and for cscocl@xmath5 .
these measurements show that the temperature dependence of the ordering is not simple .
some rearrangement of the ordered structure at temperatures below @xmath30is taking place , without changing the unit cell or broadening the magnetic bragg peaks .
similar results have also been reported by farkas _
@xcite in cscobr@xmath5 and by yoshizawa and hirakawa @xcite in cscocl@xmath5 .
detailed analysis has shown that the high temperature ordered structure in both materials is similar to that shown in figure [ ising_spin_arr ] a ) .
neutron critical scattering has been observed around @xmath173 at temperatures close to @xmath30 in cscobr@xmath5 @xcite . as is shown in the figures
, a feature of the low - temperature neutron scattering from cscobr@xmath5 and cscocl@xmath5 is the presence of a ( 111 ) magnetic bragg peak .
this peak has the scattering from the three magnetic sites in each basal plane in phase , while there is a phase reversal between successive planes .
thus its intensity reflects the net magnetic moment in each plane . at temperatures below @xmath174 ( 12 k in cscobr@xmath5 and 5.5 k in cscocl@xmath5 )
each plane is ordered ferrimagnetically and the ordering corresponds to that shown in figure [ ising_spin_arr ] b ) .
neutron critical scattering has been observed around ( 111 ) at temperature close to @xmath174 in cscocl@xmath5 @xcite .
the absence of magnetic intensity in the ( 001 ) bragg peak indicates that the moments are ordered in the c direction . at temperatures between @xmath174 and @xmath31 = 13.5 k in cscocl@xmath5
there is clearly some structural rearrangement taking place .
this is less evident in cscobr@xmath5 , but the data does seem to indicate some small reduction of the magnetic intensities as the temperature is lowered from @xmath31=16 k @xcite to @xmath174 . exactly what is happening in the intermediate region between @xmath31 and @xmath174 has not been established by the experimental work to date . in assigning values to @xmath31 and @xmath174 in cscobr@xmath5
we have reinterpreted the data of yelon _ et al . _ to be on a consistent basis with the assignments by other authors @xcite .
two experimental features should be mentioned here .
first in cscobr@xmath5 yelon _ et al .
_ @xcite indicate that there is also a small component of ordered moment in the xy plane , but this has not been reported in the work on rbcobr@xmath5 or cscocl@xmath5 .
if confirmed , such ordering would arise from the term involving the parameter @xmath7 in equation ( [ h_i ] ) and indicates departure from the ising form .
second , a perhaps surprising experimental result is that , although the specific heat has a clear anomaly at @xmath30 , there is no feature corresponding to the transitions at @xmath31 and @xmath174 @xcite .
theory has concentrated on the solution to the true ising system on the basis of the hamiltonian given in equation ( [ h_i ] ) with @xmath175 .
there has also been work with next - nearest - neighbour interactions in the basal plane included , so as to reduce the ground state degeneracy .
the ground state in the landau - ginzburg - wilson ( lgw ) model is case c ) of figure [ ising_spin_arr ] @xcite .
there is a transition at temperature @xmath31 to a second ordered state corresponding to case a ) of figure [ ising_spin_arr ] where one sublattice has a zero mean value of @xmath165 .
plumer , caill and hood @xcite show that , if the lgw treatment is expanded to higher orderin the spin density , the phase transition can split into two transitions at @xmath31 and @xmath174 , although the splitting , @xmath176 is small , of order 1% .
the specific heat effects are of opposite sign and tend to cancel , agreeing with the experimental findings . as discussed earlier , experiments @xcite at low temperatures indicates the presence of an ordered state that corresponds to case b ) of figure [ ising_spin_arr ] in contradiction to the predictions of the lgw model .
kurata and kawamura @xcite have recently shown that an extension of mean - field theory to include correlation effects in the xy plane can give the observed ferrimagnetic ground state . even after this difficulty is taken care of , however , the lgw treatment still has shortcomings as it does not give the correct low temperature state , which coppersmith @xcite has shown involves some disorder on every site .
the various experiments refered to above report similar values of the ordered moment at low temperature with a mean of 3.2(2 ) @xmath177 .
there is no appreciable difference between the predicted maximum ordered moment of @xmath178 @xcite and the measured moment , so the amount of the disorder is not large .
however recent nmr work of kohmoto _ et al . _
@xcite shows direct evidence of the presence of disorder at low temperatures as predicted .
another approach to the solution of the frustrated ising model is to use monte - carlo methods .
@xmath180 @xcite have shown that this gives the ferrimagnetic phase at low temperatures and two other ordered phases at higher temperatures , in good agreement with experiment .
the high - temperature ordered phase , between @xmath30 and @xmath31 , is a randomly modulated phase ( rmp ) .
although the long - range order persists on a @xmath169 by @xmath169 cell , there is a random modulation of @xmath165 on all three sites .
the order parameter is the magnetic structure factor , @xmath181 with @xmath182 .
the low - temperature structure , with t@xmath183 , also has @xmath184 as the order parameter , but with @xmath185 .
the transitions at @xmath30 and @xmath174 are critical phase transitions , but the nature of the transition at @xmath31 is not clear . between @xmath31 and @xmath174
the structure is complex with the characteristics of the rmp phase present . but with an anomalous temperature dependence of the order parameter @xmath186 .
neutron scattering is a direct technique for characterizing the phase transitions at @xmath30 and at @xmath174 because it can measure the scattering around @xmath173 and ( 001 ) directly .
it is an unanswered question whether there is a true phase transition at @xmath31 , though there clearly is a region just above @xmath174 where the magnetic order has unusual temperature dependence .
there is qualitative agreement between the neutron - scattering data and the monte carlo work both with regard to the temperature dependence of the magnetic structure factors and to the ratio of @xmath30 to @xmath31 and to @xmath174 .
further the monte carlo computations show no specific heat anomaly at @xmath31 or at @xmath174 , in agreement with experiment .
the nature of the phase transition at @xmath30 has received much attention , both theoretically and experimentally .
@xcite used renormalisation group arguments to predict that the transition is in the same universality class as the order - disorder phase transition in the three - dimensional xy model .
early monte carlo work on the phase transition was done by matsubara and inawashiro @xcite and by hienonen and petschek @xcite but the most accurate analysis comes from the work of bunker _
et al . _
@xcite and plumer and mailhot @xcite .
there is controversy concerning the accuracy of some of these results , with plumer and mailhot s work showing satisfactory agreement with the three - dimensional xy model and bunker _
_ showing significant discrepancies .
table [ i_crit ] lists the values obtained by these authors and a comparison with the xy model and with experiment .
apart from those of farkas _ et al .
_ , the experimental results for @xmath35 cover the same range as the theoretical values , with a weak bias towards the higher values .
the experimental value for @xmath38 claims higher accuracy than any of the theoretical values but again no definitive conclusions can be made .
the monte carlo simulations of plumer _ et al . _ @xcite indicate that the critical region at @xmath30 is smaller than usual due to the proximity of another ordered phase that is would be stabilized by next - nearest - neighbour interactions of order 10% of @xmath60 . the recent neutron scattering work of rogge @xcite on cscobr@xmath5 shows results incompatible with a normal critical phase transition , in that the critical fluctuations can not be described in terms of a model with a single length scale
this whole situation is unclear and more work is needed to resolve it . because of the ising nature of the hamiltonian and the quasi - one - dimensional nature of the magnetic interactions , the excitations in these compounds are predominantly of the soliton type .
solitons have been observed in both cscocl@xmath5 @xcite and cscobr@xmath5 @xcite and the results used to determine the parameters @xmath4 and @xmath7 in the hamiltonian ( equation ( [ h_i ] ) ) .
boucher _ et al .
_ @xcite show that soliton excitations are present at temperatures down to @xmath174 .
the neutron scattering data near the phase transition at @xmath30 has been interpreted in terms of a soliton condensation on to one sublattice to give a magnetic structure of the type shown in figure [ ising_spin_arr ] b ) with the solitons on the chains corresponding to the sites with zero mean moment @xcite .
this result is not in accord with the idea of a randomly modulated phase @xcite from monte carlo simulations .
since all the monte carlo work has been carried out for much - less - one - dimensional hamiltonians ( smaller values of @xmath187 ) than is found in actual abx@xmath5 compounds , and since the soliton ideas are products of the one dimensionality , the soliton measurements @xcite raise questions about whether the monte carlo simulations map on to the real materials , particularly as the excitations are fundamentally different .
we conclude this section with a discussion of the effect of a magnetic field on triangular ising antiferromagnets .
consider first a one - dimensional ising antiferromagnet with nearest neighbour interactions .
the hamiltonian is @xmath188 yang and yang @xcite give an exact solution of this hamiltonian , but here we just treat the basic ideas . at low temperature and
h=0 the system will form antiferromagnetic chains with a few solitons breaking the long - range order .
the field will have only small effects until it reaches a critical value @xmath70 where it can break antiferromagnetic bonds without cost in energy .
@xmath189 at field @xmath70 there should be a phase transition from antiferromagnetism to ferromagnetism . this transition is observed in the abx@xmath5 ising compounds , though the field @xmath70 is large ( @xmath190 t ) because the exchange @xmath4 along the chains is large . in the ferrimagnetic - plane low - temperature structure ( figure [ ising_spin_arr ]
b ) ) the small interchain exchange @xmath3 results in there being two critical fields , @xmath191 and @xmath192 . at field @xmath193 one of the chains
marked `` + '' in the figure becomes ferromagnetic without cost in interchain exchange energy .
the magnetisation per cobalt atom is @xmath194 . then at a higher field
@xmath192 the other two chains become ferromagnetic , with @xmath195 and the magnetization per cobalt atom is g@xmath196@xmath197s/2 .
figure [ ising_h_c ] shows the magnetization plotted against the applied field as observed in cscocl@xmath5 by amaya _
the two steps in the magnetization are not seen , but instead there is a rounding out between @xmath191 and @xmath192 which is not expected from the simple arguments that we have given .
as the figure shows , this rounding becomes more pronounced at higher temperatures in the ordered phase .
the critical fields , @xmath191 and @xmath192 are usually identified with the two maxima in @xmath198 , and it is these values that are shown in table [ ising_par ] .
table [ ising_par ] shows that the values of @xmath4 derived from @xmath191 agree reasonably with those found by other experimental techniques , but the values of @xmath3 derived from @xmath199 seem to be significantly higher than values from other experimental techniques .
they give @xmath45 of order 10 , which is surprisingly small .
even the larger ratios found from neutron and raman scattering are appreciably smaller than is found in heisenberg abx@xmath5 compounds ( table [ heis_parameters ] ) , indicating that the cobalt compounds are less one dimensional in magnetic properties .
this chapter is devoted to the description of the magnetic properties of four compounds from the afex@xmath5 family : csfecl@xmath5 , csfebr@xmath5 , rbfecl@xmath5 and rbfebr@xmath5 . at room temperature
they all have the same crystal structure with space group @xmath59 and , as usual for all abx@xmath5 hexagonal compounds , at low temperature they exhibit quasi - one - dimensional magnetic behaviour .
a characteristic property of these four crystals is the large value of the magnetic anisotropy in comparison with the exchange interaction . in some cases
this prevents the advent of long range magnetic ordering ( lro ) even at zero temperature .
there are other compounds ( nd@xmath200febr@xmath5 , nd@xmath200fecl@xmath5 , tlfebr@xmath5 , tlfecl@xmath5 and csfei@xmath5 ) that probably may be described as singlet - ground - state magnets @xcite , but they have been investigated less thoroughly and a comprehensive understanding of their physical properties has not yet been developed .
a free fe@xmath52 ion in the afex@xmath5 family has a @xmath201 ground state .
a cubic crystal field splits this into an upper orbital doublet and a lower orbital triplet with an energy difference of order 1000 @xmath202 .
spin - orbit coupling , @xmath203 , causes a further splitting of the triplet according to the effective total angular momentums @xmath204 , 2 and 3 .
the lowest state with @xmath204 is split still further by a trigonal component of the crystal field , @xmath205 , to produce a singlet ground state ( @xmath206 ) and an excited doublet ( @xmath207 ) as shown in fig.[sgs_split ] .
the hamiltonian representing these splittings of the triplet state may be written as @xmath208 since the energy separation between the ground state and the second excited state is of order of 100@xmath202 @xcite , at low temperature only the first excited doublet is appreciably populated and the following effective spin hamiltonian can be used to describe the magnetic properties of afex@xmath5 compounds : @xmath209 + \!\sum_{k , l}^{\text{planes}}\![j'_\perp(s_k^x s_l^x+s_k^y s_l^y ) + j'_\parallel s_k^z s_l^z ] \ ] ] @xmath210 \label{sgs_h}\ ] ] where @xmath75 is fictitious spin and @xmath12 , the value of which is positive , equals to the energy gap between the @xmath206 and @xmath207 states . however , some authors prefer to use the heisenberg hamiltonian ( [ h_h ] ) to describe magnetic properties of the linear chains antiferromagneticaly coupled fe@xmath52 ions in csfebr@xmath5 and rbfebr@xmath5 . as @xmath211 in the absence of an external magnetic field there are two regimes separated by a phase transition . \1 . for @xmath212
the system has a magnetic ground state with an easy - plane type of anisotropy .
this is the case for rbfecl@xmath5 and rbfebr@xmath5 .
\2 . for @xmath213
the system has a singlet ground state and consequently does not order magnetically even at @xmath72 .
this is the case for csfecl@xmath5 and csfebr@xmath5 .
the equality @xmath214 was derived as a condition at which the lowest excitation energy gap at the magnetic zone center becomes zero @xcite .
the application of an external magnetic field along @xmath11-axis on the sgs materials leads to a phase transition to an ordered state .
this happens at a field @xmath70 , when one of the excited doublet levels crosses the ground state singlet level , as it shown on fig.[sgs_h_split ] . in csfecl@xmath5
a commensurate 120@xmath2 ordered structure appears after intermediate transitions through two incommensurate structures , while in csfebr@xmath5 the phase transition leads directly to commensurate order .
if the external magnetic field is applied perpendicular to the @xmath11-axis , the singlet level remains below the excited levels at all fields so that no lro is expected . before going to the detailed description of each compound
we summarize some characteristics of each material in table [ sgs_tab ] .
an attempt to analyze the correlation between the structural and magnetic parameters of the afex@xmath5 family of compounds is made by visser and harrison in ref.@xcite . [ [ csfebr_3 ] ] csfebr@xmath5 + + + + + + + + + + + + + + + + + + + + + + in csfebr@xmath5 and rbfebr@xmath5 all exchange interactions are antiferromagnetic .
except for an early susceptibility measurement by takeda _
@xcite , the works on csfebr@xmath5 are devoted to the investigation of the magnetic excitations .
the excitation spectrum was studied both theoretically @xcite and experimentally by means of inelastic neutron scattering in a zero field @xcite and in an external magnetic field @xcite .
the lowest frequency excitation mode softens with decreasing temperature but stabilizes at 0.11 thz below 2.5 k down to 80 mk @xcite .
this fact indicates that csfebr@xmath5 remains a sgs system for @xmath215 in zero field . at 1.6 k in an external magnetic field of 4.1
t applied along @xmath11-axis a well defined bragg peak appears at ( 2/3 2/3 1 ) indicating a phase transition to the long range commensurate 120@xmath2 structure @xcite . but
the correlation lengths do not diverge at that field .
instead , they exhibit a flat maximum over about 0.3 t around 4.1 t and decrease again at higher field @xcite .
the nature of this phenomena is not yet understood .
[ [ rbfebr_3 ] ] rbfebr@xmath5 + + + + + + + + + + + + + + + + + + + + + + this compound can be considered as an intermediate case between the sgs antiferromagnet and the heisenberg antiferromagnet with easy - plane anisotropy ; the exchange interaction is strong enough to produce three - dimensional order at temperatures below 5.5 k @xcite . at a temperature of 108
k rbfebr@xmath5 undergoes a structural phase transition to a distorted phase with space group @xmath59 @xcite , which results in the appearance of two kinds of nearest neigbours exchange in the basal plane ( @xmath3 and @xmath139 ) .
this produces a distortion of a spin triangles with the angle between nearest spins not exactly 120@xmath2 .
the spin frustration is partially released ( see paragraph [ distort ] for details ) . the low temperature crystal phase is found to be ferroelectric @xcite .
the specific heat measurements revealed two successive magnetic phase transitions at @xmath216 k and @xmath217 k @xcite .
this may be caused by the splitting between @xmath3 and @xmath139 , but in fact the @xmath218 ratio remains experimentally unknown : the inequivalency of the fe@xmath52 sites is not sufficiently large to be distinguished by mssbauer spectroscopy @xcite ; the resolution of the inelastic neutron scattering experiments @xcite was not good enough to see the influence of the splitting on the dispersion of magnetic excitations .
the energies and intensities of the excitations can be described well using the dynamical correlated effective - field approximation neglecting the splitting between @xmath3 and @xmath139 .
[ [ csfecl_3 ] ] csfecl@xmath5 + + + + + + + + + + + + + + + + + + + + + + in csfecl@xmath5 the exchange interaction between fe@xmath52 ions is ferromagnetic along the c - axis , while the interchain exchange is weakly antiferromagnetic @xcite .
the results of inelastic neutron scattering @xcite show that at t=5k the lowest excitation energy at the magnetic zone center has a peak about 190 ghz confirming the absence of long range magnetic order . from the dispersion relations the parameters in the effective spin hamiltonian ( [ sgs_h ] ) may be obtained , but the results depend strongly upon the theoretical model used to analyze the experimental data .
an exciton model , correlated effective field analysis @xcite , self - consistent random - phase approximation @xcite and dynamical correlated effective - field approximation @xcite give substantially different values for the exchange interaction and the magnetic anisotropy .
the parameters of the spin hamiltonian can be estimated also from measurements of the mssbauer effect @xcite and of the nuclear spin - lattice relaxation time @xcite ( see table [ sgs_tab ] for details ) .
further inelastic neutron scattering investigation of the dispersion curves in the csfecl@xmath5 @xcite has shown that the minimum of the dispersion curve does not occur at the @xmath219-point but is shifted slightly away .
this effect can be explained by the inclusion of dipolar forces @xcite .
the transition to a magnetically ordered phase in an external magnetic field was detected by magnetization @xcite , specific heat @xcite , mssbauer @xcite and nuclear spin - lattice relaxation @xcite measurements . at @xmath220 t
the ordered state was observed at @xmath221 k. because measurements are necessarily made at nonzero temperature , lro appears over a wide region of magnetic field around @xmath70 .
for example at @xmath222 k magnetic susceptibility shows two anomalies at 3.8 and 4.6 t when lro appears , and the same two anomalies at 11.2 and 11.6 t when lro disappears in an increasing field .
the step structure in the @xmath223 curve observed at 3.8 and 4.6 t is caused by successive phase transitions from the nonmagnetic phase to a thermally frustrated incommensurate phase and then to a commensurate three - sublattice antiferromagnetic phase @xcite .
an incommensurate magnetic phase ( double - modulated and single - modulated ) was found between the nonmagnetic phase at low field and the commensurate phase at higher field by means of elastic neutron scattering @xcite . at t=0.7
k the magnetic phase transitions take place at @xmath224 t , @xmath225 t and @xmath226 t. possible explanation of the nature of the incommensurate phases has been given in a framework of the correlation theory @xcite .
an additional magnetic phase transition at 33 t was observed in a magnetization measurements @xcite .
the high field magnetization can not be explained within the framework of the fictitious @xmath75 spin states . since the magnitude of themagnetization at 33
t is large , the anomalous increase in @xmath227 is attributed to the upper excited @xmath228 spin state @xcite .
all five possible transitions for @xmath229 and @xmath230 between the ground state and the excited doublet ( see fig.[sgs_h_split ] ) have been observed in submillimetre wave esr - measurements @xcite , while at higher frequencies only two absorption lines were observed using far infrared fourier spectroscopy @xcite .
one of the absorption seems to come from the excitation between ground state and second excited doublet .
the mechanism of the second absorption is still unclear .
[ [ rbfecl_3 ] ] rbfecl@xmath5 + + + + + + + + + + + + + + + + + + + + + + unlike csfecl@xmath5 the isomorphous compound rbfecl@xmath5 reveals three - dimensional long range magnetic order below @xmath231k @xcite the signs of the exchange interactions are the same as in csfecl@xmath5 ferromagnetic along the c - axis and antiferromagnetic in the plane .
an inelastic neutron scattering study @xcite showed clear softening of the magnetic excitations in a small region around the zone center when @xmath232 .
the influence of the softening on the nuclear spin - relaxation time @xmath233 of @xmath234rb was observed in an nmr experiment @xcite . according to an elastic neutron scattering study @xcite at zero magnetic field ,
rbfecl@xmath5 undergoes three transitions at @xmath235 k , @xmath236 k and @xmath237 k. two different incommensurate structures have been found at @xmath238 and @xmath239 , while at @xmath240 the 120@xmath2 in - plane triangular structure is observed @xcite .
very similar values for the phase transition temperatures have been found from specific heat and susceptibility measurements @xcite .
the existence of thermal hysteresis at @xmath174 indicates that the incommensurate - commensurate transition is first order .
from this study and susceptibility measurements @xcite a phase diagram in @xmath241 coordinates may be derived as shown in fig .
[ rbfecl3_phd ] .
there are two main features a ) the 120@xmath2-structure and paramagnetic phase are always separated by one or two incommensurate phases ; b ) application of a magnetic field parallel to the @xmath11-axis increases the phase - transition temperature until a maximum is reached at 7.5 t ; subsequently there is a rapid decrease and the transition temperature goes to zero at 13 t. shiba and suzuki @xcite have proposed a theory , which explain the experimental phase diagram reasonably well from the viewpoint of a conical - point instability due to the dipole - dipole interaction .
they show that even a small dipole - dipole interaction can transform the 120@xmath2 structure to an incommensurate structure at intermediate temperatures , although the low temperature phase still should have the 120@xmath2 structure . very recently
the phase diagram of rbfecl@xmath5 for @xmath230 was reinvestigated by neutron scattering @xcite .
good agreement between the results of two experiments was found for a field less than 1.0 t. above 1.0 t the neutron scattering study gives a slightly higher value of the transition field from the commensurate phase to the incommensurate phase .
the high field magnetization of rbfecl@xmath5 exhibits an anomaly around 31 t @xcite , similar to those in csfecl@xmath5 @xcite .
the situation regarding the spin hamiltonian parameters derived from experiment for rbfecl@xmath5 is very similar to that in csfecl@xmath5 .
different authors analyzed their experimental data on the basis of different approximations and there has been controversy about the value of the exchange interactions and the anisotropy .
the results of a susceptibility measurements were analyzed using the molecular field approximation @xcite or pair approximation @xcite , where a fe@xmath52 chain was represented by an assembly of isolated pairs of nearest neighbour spins .
the m " ossbauer and susceptibility data @xcite were analyzed using the correlated effective - field approximation , developed by lines @xcite .
the results of inelastic neutron scattering @xcite were analyzed using this theory , the three sublattice spin - wave approximation and the exciton model .
the parameters so determined depend strongly on which approximation is used ; they are not universal ; one set of parameters can not describe all available experimental data .
suzuki @xcite has made an attempt to see to what extent one can understand the various observed magnetic properties on the basis of a single set of parameters .
he used thedynamical correlated effective - field approximation @xcite and has found values of @xmath242 and @xmath243 which can reproduce experimental data reasonably well .
originally he has considered for the sake of simplicity only a single chain of fe@xmath52 ions , so that the approach has applicable only in the paramagnetic phase .
later he has included an interchain coupling in the model and derived a consistent set of exchange parameters which explain the behaviour of rbfecl@xmath5 above and below @xmath25 @xcite .
except for csfecl@xmath5 and rbfecl@xmath5 as described in the previous chapter , there are just two triangular antiferromagnets with ferromagnetic interactions along the chains , cscucl@xmath5 and csnif@xmath5 .
each gives rise to its own different physics , and will be dealt with in separate subsections of this chapter .
cscucl@xmath5 has been one of the most extensively studied of the triangular antiferromagnets .
the spins in the copper chains are coupled ferromagnetically , but the planar interactions are antiferromagnetic so that frustration effects are of similar importance to the common case of antiferromagnetic chains .
what makes cscucl@xmath5 unique is that below 423 k the triangular crystal structure is distorted through the jahn - teller effect to give a crystal structure with space group @xmath244 @xcite .
the distortion from the stacked triangular lattice involves small in - plane displacements of copper atoms so as to form a helix with axis along @xmath11 with one turn of the helix every six layers . below @xmath25=10.66(1 )
k @xcite the zero - field magnetic structure shows @xmath58 planes with the @xmath137 triangular antiferromagnetic structure and with moments in the plane .
magnetic neutron bragg peaks are observed at ( @xmath245 ) with @xmath246=0.085 @xcite .
this corresponds to the triangular spin arrays that are rotated about c by @xmath247 between successive planes .
the period of rotation is 11.8c or 214 .
it is believed @xcite that this rotation arises from the dzyaloshinsky - moriya ( dm ) interaction that gives rise to a term in the hamiltonian given by @xmath248 in cscucl@xmath5 @xmath249 is a vector along @xmath11 which is non zero only when atoms @xmath112 and @xmath250 are nearest neighbours along the helical chains . in the absence of the jahn - teller distortion , symmetry requires the dm term to vanish ; this is why it has not been included in the hamiltonian for other materials discussed in this work .
the jahn - teller distortion gives a second , smaller , effect on the hamiltonian in that it causes the vector @xmath251 to deviate slightly from from the c axis with a period of six lattice spacings @xcite .
this effect is small and often neglected in the literature .
neutron scattering measurements of the spin - wave dispersion relations by mekata _
et al . _
@xcite confirm this hamiltonian and give the intra - chain exchange @xmath252 ghz , @xmath253 ghz and the in - plane exchange interaction @xmath254 ghz ( no quantum corrections were included in deriving these parameters ) .
it is apparent that cscucl@xmath5 is less one dimensional than is usual for abx@xmath5 materials , since @xmath255 is about 6 and in other materials it is one to two orders of magnitude larger .
the ferromagnetic interactions along the chain give a minimum energy when all the moments are aligned parallel , but the dm interaction is minimised when neighbouring moments are aligned perpendicularly .
the sum of these two terms in the hamiltonian gives a minimum energy classically for a helical magnet with the tangent of the turn angle between neighbouring moments equal to @xmath256 @xcite .
the observed magnetic structure corresponds to these helices along @xmath11 stacked on a @xmath0 triangular lattice .
the parameter @xmath246 is observed to be independent of temperature @xcite and field ( @xmath257 ) @xcite .
spin - wave dispersion relations have been calculated by rastelli and tassi @xcite and by stefanovskii and sukstanskii @xcite and shown to agree with antiferromagnetic resonance data .
the maximum ordered moment for @xmath157 and g=2.10 @xcite is 1.05 @xmath53 .
@xcite reported the ordered moment as @xmath258 @xmath53 , extrapolated to 0 k , but recent work gives higher values , 0.85 @xmath53 @xcite for zero field and low temperature , and @xmath259 @xmath53 at @xmath260 k and @xmath261 t @xcite .
the two latter values are higher than is typical for frustrated systems , perhaps because the ferromagnetic chain coupling leads to smaller moment reductions than does antiferromagnetic coupling . an applied field along @xmath11 gives rise to two first - order phase transitions at fields @xmath191 and @xmath192 .
@xcite gives @xmath262 t and @xmath263 t at @xmath264 k , while chiba _
et al . _
@xcite gives @xmath265 t at @xmath266 k. for fields less than @xmath191 the magnetic structure is that found in zero field together with a canting of each spin towards the direction of the applied magnetic field . at @xmath191
the triangular layers break down into a colinear structure on the triangular lattice based on that shown in figure [ ising_spin_arr ] , where two spins are aligned in one direction and the third is in the opposite direction .
there is a canting of all three spins towards the field direction @xcite .
the helical stacking of the planes remains in this structure .
neutron scattering work @xcite confirms this description of the magnetic structure above @xmath191 . on passing from the low - field to the medium - field structure the ( @xmath267 )
lines lose intensity and new lines appear at ( @xmath268 ) . above @xmath192 the magnetic structure is believed to be almost ferromagnetic , with the helical stacking destroyed .
the phase transition at h@xmath269 is not predicted for a classical system with the appropriate hamiltonian ; nikuni and shiba @xcite show however that when quantum fluctuations are taken into account the phase transition is to be expected .
the field h@xmath269 decreases slowly as the temperature increases @xcite , and this line of first - order phase transition in the h - t phase diagram meets the paramagnetic phase transition at a bicritical point .
extrapolation of the neutron scattering measurements of stuesser _ et al .
_ @xcite indicate that the bicritical point occurs at @xmath270 t and @xmath271 k , while the specific heat measurements of weber _ et al . _
@xcite give the same value for h@xmath272 and t@xmath273 = 10.59 k. the behaviour of cscucl@xmath5 in a field perpendicular to the @xmath11 axis has been treated theoretically by jacobs _
the field splits the degeneracy of the orientation of the spin triangle in the @xmath58 plane and causes the triangles to no longer have angles of @xmath137 .
this results in @xmath246 varying with the applied field @xmath102 .
an anomaly is observed @xcite in the low - temperature magnetization at 12 t which is believed to involve a transition to a commensurate state . since cscucl@xmath5 is a frustrated system with s=@xmath274 , quantum fluctuations
would be expected to be important .
_ @xcite confirm this by evaluating the first term in a @xmath275 expansion and showing that quantum ( and thermal ) fluctuations lift a nontrivial degeneracy and stabilize the commensurate state .
we finish this subsection by describing work on the critical phase transition .
the magnetism in cscucl@xmath5 shows two chiral degeneracies , one arising from the triangular structure in each plane and one from the helix along the @xmath11 axis .
@xcite argue that the structural helix will not affect the critical properties and that the phase transition should be that of the chiral xy model .
this model predicts @xmath276 and @xmath277 ( table v ) , while if the non - universality model holds tricritical exponents would be observed with @xmath278 and @xmath279 ( table iii ) . the early neutron - scattering measurements of adachi _ et al . _
@xcite gave @xmath280 , but this value has not been confirmed by recent work that gives much lower values ; mekata _ et al . _
@xcite give @xmath281 and stuesser _ et al . _
@xcite give @xmath282 .
recent specific heat measurements of weber _ et al .
_ @xcite in zero field can be described well by a critical exponent @xmath283 except very close to t@xmath284 ( @xmath285 ) where the transition seems to go over to being weakly first order .
the exponent is compatible with the chiral xy model and not with the tricritical model .
the small canting of the vector @xmath251 from the @xmath11 axis will lower the symmetry from @xmath286 to @xmath287 .
thus the critical properties should eventually exhibit a crossover effect from @xmath286 to @xmath287 behavior , though this effect has not been observed to date .
it is not clear whether the observed weakly - first - order effects are a fundamental property of the frustrated @xmath20 model or if they are a consequence of the expected crossover .
as the field increases towards h@xmath273 the amplitude of the specific heat divergence becomes smaller and the first order part of the transition becomes even weaker . at a field above h@xmath273 , @xmath288 , the specific heat measurements show a critical phase transition with @xmath289 , contrary to the expected value for regular xy behaviour where @xmath290 .
the origin of this discrepancy is not clear .
this is the only fluoride abx@xmath5 compound which crystallises with the stacked triangular lattice @xcite .
the magnetic structure corresponds to an easy - plane antiferromagnet with no distortions at low temperatures . as usual the interactions are much stronger along @xmath11 than within the @xmath58 plane , giving rise to quasi - one - dimensional properties above @xmath25 .
the interactions along the chain and the soliton properties have been much studied in the quasi - one - dimensional temperature region ( see kakurai _
_ @xcite and references therein ) . in this article
only the three - dimensional ordered properties will be discussed .
these are unique for an easy - axis material since the ordering consists of ferromagnetic @xmath26 planes stacked antiferromagnetically with moments aligned along @xmath50 @xcite , as shown in the figure [ csnif3 ] .
this ordering implies that the low - temperature ordered state breaks the hexagonal symmetry . in practice domains
are formed favouring one of the three equivalent a directions in the basal plane . in terms of the hexagonal unit cell , bragg peaks are observed in the low - temperature neutron - scattering pattern with indices ( h/2,k , l ) where h , k and l are all integers .
the magnetic order can not arise from a hamiltonian containing just nearest - neighbour interactions and easy - plane anisotropy , as has been shown in earlier chapters .
scherer and barjhoux @xcite and suzuki @xcite show that it will occur if the interactions , other than those along the chains , are predominantly dipolar in character .
the idea of this predominance is supported by the low value of the nel temperature , 2.7 k , which shows weak interchain interactions .
the value of the ordered moment , 2.26 @xmath291 @xcite , for s=1 and g=2.28 , is higher than is found in frustrated triangular structures .
a theoretical treatment which includes magnetoelastic effects has been given by caill and plumer .
a mean - field treatment of the phase diagram is given by trudeau and plumer .
the various experimentally - determined values of @xmath4 , @xmath3 , @xmath12 and @xmath25 are listed in the table [ cnf ] .
it is clear that the in - plane exchange , @xmath3 , is small compared with the exchange along the chains , @xmath4 , but the magnitude is similar to that found in other abx@xmath5 easy - plane materials ( see table [ heis_parameters ] ) .
the anisotropy , @xmath12 , is however larger than in other easy - axis materials .
the dipolar forces are long range in nature and their relative influence , which comes from sums over a large number of moments , is greater for ferromagnetic than for antiferromagnetic chain interactions .
theory @xcite predicts the correct structure for the parameters given in the table together with dipolar interactions . because the chains are ferromagnetic
, it would be expected that applied magnetic fields will have a relatively greater influence than in other abx@xmath5 materials .
this is indeed the case for fields applied in the basal plane , though the strong easy - plane anisotropy makes the effect of fields applied along c less .
the neutron scattering measurements of steiner and dachs @xcite show that small fields , @xmath102 , applied in the basal plane to the ordered material influence the relative sizes of the three domains .
a field of around 0.05 t is sufficient to produce a single - domain sample with moments approximately perpendicular to @xmath102 .
larger fields destabilize the antiferromagnetic state and there is a phase transition at critical field , @xmath70 , to a paramagnetic state with imposed alignment of the moments along h. @xmath70 is around 0.2 to 0.3 t at 2.0 k @xcite if the field is applied along @xmath11 , larger values of @xmath102 are needed to destroy the antiferromagnetism because the field is opposed by the easy - plane anisotropy .
a similar phase transition occurs in this case , but the critical field , @xmath70 is 8 to 25 times larger than when @xmath102 is applied in the basal plane @xcite .
the critical properties of csnif@xmath5 are not simple .
the hamiltonian has xy symmetry in zero field or with applied field @xmath257 along @xmath11 .
for an applied field in the plane sufficient to produce a single - domain sample the symmetry becomes ising like @xcite with @xmath81 in a direction perpendicular to both @xmath11 and @xmath102 .
there are two complicating factors however .
first the long - range nature of the dipolar interactions can lead to mean - field exponents and to crossover behaviour .
second the presence of three equivalent domains in the structure changes the critical properties and makes the phase transition first order @xcite .
neutron scattering measurements in zero field @xcite show a critical phase transition with a crossover at @xmath292 values near @xmath293 from exponents , @xmath294 and @xmath295 , at large @xmath7 , to different values , @xmath296 , @xmath297 and @xmath298 , at small @xmath7 .
although the errors on the exponents are not small , the first set is consistent with mean - field exponents and the second set with xy behaviour .
there is no experimental evidence for a first - order phase transition .
strain effects have been observed in csnif@xmath5 crystals @xcite which might suppress domain fluctuations , but this would not be expected to completely change the nature of the transition .
there is one other set of measurements of critical properties .
lussier and poirier @xcite have measured the phase boundary @xmath70 as a function of temperature and of the direction of @xmath102 . for fields perpendicular to @xmath11 , @xmath70 is large enough for the sample to be single domain and ising symmetry is to be expected .
plumer and caill @xcite show that @xmath70 varies as the order parameter , so that the measurements should result in an ising critical exponent @xmath35=0.326 .
the measured exponent is [email protected] in not - too - bad agreement with theoretical expectations .
the situation with @xmath102 along @xmath11 is puzzling however .
theory predicts an xy exponent , @xmath35=0.345 , for h@xmath122 while the experiment gives a higher value , @[email protected] .
it is apparent from previous chapters ( or at least authors hope it is apparent ) that there is a plenty of triangular magnets , physical properties of which depend on type and relative strength of exchange and anisotropic interactions .
those properties are not totally established yet .
even less understanding of properties of diluted and mixed triangular magnets is currently achieved
they found to be sophisticated as much again .
nevertheless , some interesting results and ideas have been found in the process of their investigation .
it is useful to draw the analogy here between the triangular antiferromagnets and another example of frustrated magnetic system , xy square - lattice antiferromagnet dominated by second - neighbor antiferromagnetic exchange . in the latter case dilution acts against thermal and quantum fluctuations , producing an effect known as `` ordering due to disorder '' @xcite .
present chapter summarises briefly characteristics of diluted and mixed magnets on triangular lattice .
the obvious method to `` disturb '' magnetic system is to introduce small amount of nonmagnetic impurities .
the presence of an impurity results either in mechanical distortion of the structure of the original crystal or in disruption of some part of the interaction between the magnetic atoms , which is reflected in the collective behaviour of the spin system .
random field effect due to non - magnetic impurities on spin correlation was studied by elastic neutron scattering and magnetic susceptibility measurements in ising antiferromagnet , cscocl@xmath5 , doped by mg or zn @xcite and by measurements of diffuse scattering in cscocl@xmath5 doped by mg @xcite .
the reduction of the upper magnetic transition temperature , @xmath30 , has been found in samples with impurities : @xmath300 k ,
@xmath301 k and @xmath302 k for the pure crystal , crystal with 0.58% of mg and crystal with 1.7% of mg respectively @xcite , while temperature of the lower magnetic transition , @xmath31 , could not be found down to 1.6 k @xcite .
the mg concentration was determined by a chemical analysis . on the other hand , detailed esr and magnetization measurements of heisenberg triangular antiferromagnet , rbnicl@xmath5 , doped with 1% of mg showed that the neel temperature remained unchanged from pure crystal , in which @xmath303 k @xcite .
however , the influence of the impurity at low temperature is still well pronounced : 1 ) the spin - flop region became much broader , 2 ) the gap @xmath304 of one of the resonance branches has increased from 55 ghz to 61 ghz , 3 ) the gap of another resonance branch has found to be 20 ghz , while in the pure crystal it has not been observed and the estimated value for pure crystal is 0.2 ghz .
such a dramatic changing of the resonance spectrum was successfully described by introducing a two - ion anisotropy of the form @xmath305 into the spin hamiltonian ( [ h_h ] ) .
it has been postulated that an impurity which does not occupy a site in the crystal lattice strongly distorts the electrical interactions within the crystal , altering the character of the anisotropic interactions .
in contrast to the case of doped cscocl@xmath5 , where mg concentration was different in different crystals , a @xmath36 activation analysis reviled approximately the same concentration ( about 1% ) of mg in all investigated crystals of rbnicl@xmath5 . very unusual result has been reported recently by yamazaki _
et al . _
@xcite for a csv@xmath306mg@xmath307cl@xmath5 , ( @xmath308=0.000 - 0.357 ) .
temperature - dependence of the magnetic susceptibility for a sample with @xmath308=0.026 suggests that the ordering temperature is about 35 k , while in a pure sample @xmath309 k @xcite .
such a dramatic increase of the ordering temperature obviously has to be confirmed by some other techniques .
the influence of diamagnetic dilution on the magnetic ordering process of the induced - moment antiferromagnet rbfecl@xmath5 was studied in ref .
single crystals of the solid solution rbfe@xmath306mg@xmath307cl@xmath5 ( @xmath308=0.02 , 0.03 and 0.05 ) were investigated by means of elastic neutron scattering . the @xmath308=0.02 and 0.03
samples showed transitions from paramagnetic to the ic@xmath143 phase at the same temperature , @xmath231 k ( see fig . [ rbfecl3_phd ] ) , then transitions to the ic@xmath6 and c phases at temperatures that decreased sharply with @xmath308 .
the @xmath308=0.05 sample also shows a transition to the ic@xmath143 phase at @xmath231 k , but no further transitions down to the lowest temperature of 1.38 k. an additional elastic diffuse magnetic scattering component centred at the vector @xmath42 has been found to persist to temperatures well above @xmath25 in all samples .
another method of investigation of the magnetic system consists of introducing small amount of magnetic impurities .
a substitution of magnetic ions can modify the amplitude , or even sign , of the effective single - ion anisotropy .
for example , in pure rbnicl@xmath5 the splitting between @xmath30 and @xmath31 is very small , 0.14 t @xcite , while addition of only 5% of co results in as much as 9 k between @xmath30 and @xmath31 @xcite .
magnetic phase diagrams of heisenberg triangular antiferromagnet , csni@xmath310cl@xmath5 ( @xmath311=co , fe , mg ) have been determined by heat capacity @xcite and ultrasonic velocity @xcite measurements .
as expected , when comparing with pure csnicl@xmath5 , the co - doped crystal shows enhanced ising effective single - ion anisotropy the spin - flop field @xmath312 and the splitting between @xmath30 and @xmath31 are sufficiently increased ; the fe - doped crystal behaves as a heisenberg antiferromagnet with @xmath20 type of anisotropy the phase diagram is very similar to those in csmnbr@xmath5 ; in the mg - doped crystal nonmagnetic impurity causes decrease of effective single - ion anisotropy and values of @xmath312 and @xmath313 consequently .
magnetic phase diagrams of csni@xmath310cl@xmath5 for @xmath314 is shown on fig .
[ deluted ] .
apart from addition of small amount of impurities into antiferromagnet on triangular lattice , some investigations had been done on mixed systems , where the impurity concentration is not small . in ref .
@xcite the magnetic phase diagram of rb@xmath306k@xmath307nicl@xmath5 has been studied by susceptibility and torque measurements .
pure rbnicl@xmath5 is a heisenberg antiferromagnet with easy - axis type anisotropy , while knicl@xmath5 demonstrates large easy - plane type anisotropy , see chapter [ heis ] .
the transition between two different types of anisotropy has been found at @xmath315 .
the interpretation of the observed phase diagram at large @xmath308 is quite complicated due to crystal structure distortions of pure knicl@xmath5 .
a.harrison and coauthors @xcite had investigated magnetic ordering effects in the mixed singlet - groud - state magnets afex@xmath5 , where a is a mixture of cs and rb or x is the mixture of cl and br .
as it was shown in the chapter [ sgs ] csfecl@xmath5 has `` truly '' nonmagnetic ground state , while in rbfecl@xmath5 sufficiently strong exchange interaction causes magnetic ordering at @xmath316 k. therefore , the replacement of rb in rbfecl@xmath5 by cs should decrease neel temperature or even suppress magnetic ordering .
indeed , it was found experimentaly @xcite , that as low as 5% of cs is sufficient to destroy long - range magnetic ordering . in rbfecl@xmath317br@xmath307
there is an obvious competition between ferro and antiferromagnetic sign of the intrachain interactions : in rbfecl@xmath5 the exchange along @xmath11-axis is ferromagnetic , while in rbfebr@xmath5 it is antiferromagnetic ( see chapter [ sgs ] ) .
again , low concentrations of either type of dopant , @xmath318 , destroys the magnetic long - range order @xcite .
note , that at intermediate compositions a singlet ground - state phase has been observed , rather than expected a spin - glass phase .
as it was mentioned in the introduction , there are several ways that triangles of antiferromagnetic interactions can be built into a crystal lattice .
the distinguished feature of the majority of antiferromagnets on a stacked triangular lattice , described in this article , is the appearance at sufficiently low temperature of the three - dimensional long - range magnetic ordering , despite the frustration of the exchange interaction .
the transition temperature of this sort of antiferromagnets is typically an order of magnitude lower than the curie - weiss temperature , but such a big reduction of the transition temperature usually arises from a combination of two effects , frustration and the low - dimensionality of actual stacked - triangular materials ( they have quasi 1d or 2d character ) .
the influence of the frustration on magnetic properties of other triangular antiferromagnets , which have more complicated structures , is apparently more substantial .
for example , in kagome - lattice antiferromagnet srcr@xmath319ga@xmath320o@xmath321 , with @xmath322 @xcite , the low - temperature ground state is a spin glass . in gadolinium
gallium garnet , gd@xmath5ga@xmath323o@xmath324 , where the magnetic gd ions are on two interpenetrating corner - sharing triangular sublattices , long range magnetic order has been found only in the applied magnetic field around 1 t @xcite , while in a lower field magnetization is different for a field cooling and zero field cooling @xcite , which is typical for a spin glass .
many pyrochlores with the chemical formula a@xmath6b@xmath6o@xmath325 also undergo a phase transition to a spin glass state @xcite .
because triangular antiferromagnets on a stacked triangular lattice still can be described at low temperature in terms of neel - type ordering , rather than spin - glass or a short - range order , they form a good basis for the study of the effects of frustration on magnetic systems .
frustration leads to new physics with novel phase diagrams and critical properties .
zero - point fluctuations are found in ising systems ; these become large in @xmath20 and heisenberg systems .
theory and experiment seem to be in good accord in describing the ground state and the excitations of the triangular antiferromagnets except in the case of s=1 quasi - one - dimensional nickel materials where spin - wave theory seems to be inadequate .
correction terms for ground - state fluctuations are larger than can be dealt with confidently and vestiges of the one - dimensional haldane effect are believed to be present .
theoretical predictions of the nature of the phase diagram as a function of @xmath102 and @xmath326 are in accord with experiment , but the critical properties at phase transitions are often not described satisfactorily . for materials with the heisenberg hamiltonian ,
the theoretical consensus favours kawamura s so(3 ) chiral universality class , but experiments show significant discrepancies . for the @xmath20 hamiltonian
the theoretical situation is controversial with three contending scenarios , chiral @xmath20 properties , tricritical properties and a weak first - order phase transition .
experiment , which is largely confined to one material , csmnbr@xmath5 , shows a critical phase transition with exponents that can be taken to be in agreement with either the chiral @xmath20 model or tricritical exponents ; the two theories give quite similar predictions for the exponents . there is a need for measurements of critical exponents of weak ( @xmath119 ) easy - plane materials and for strong easy - plane materials in a field . for easy - axis materials the experimental values and the theoretical predictions for the critical indices @xmath35 , @xmath37 and @xmath36 at both of the two zero - field phase transitions are irreconcilable .
the experimental values are not in accord with the scaling laws and a confirmation of the single report of values of @xmath37 and @xmath36 would be desirable . a number of cobalt quasi - one - dimensional triangular antiferromagnets have hamiltonians that are close to the ising hamiltonian .
the zero - field properties of these materials is not simple with three phase transitions .
only the upper one of these shows a specific - heat anomaly .
most theoretical works have predicted critical exponents at the upper critical temperature that follow the three - dimensional unfrustrated @xmath20 model , and most experiments agree with these predictions .
one recent monte - carlo study suggested slightly different exponents , but unfortunately the experiments are not sufficiently accurate to distinguish .
one group has recently reported experimental determinations of @xmath35 that are significantly lower than those given in other reports , and that are not in accord with any theoretical predictions .
it has long been a puzzle why one triangular antiferromagnet , rbmnbr@xmath5 shows an incommensurate magnetic structure .
recently we and our coworkers have shown this to arise from small structural distortions of the lattice along planes perpendicular to the basal plane .
these distortions , which map on to the row model of zhang , saslow and gabay , allow relatively small changes in exchange parameters , arising from the distortions , to destroy the simple triangular magnetic structure .
similar distortion effects are also found in knicl@xmath5 .
we are grateful to drs .
r. feyerherm , b.d .
gaulin , j. gardner , m.l .
plumer , d. visser and m. zhitomirsky for useful discussions about some aspects of this manuscript .
we thank the natural sciences and engineering research council of canada for financial support .
one of the authors ( mfc ) wishes to thank the chalk river laboratories for hospitality during a sabbatical visit while the this article was written . 199 h.kawamura , j. phys .
japan * 56 * , 474 ( 1987 ) .
h.kawamura , j. phys .
japan * 58 * , 584 ( 1979 ) .
h.kawamura , j. phys .
japan , * 61 * , 1299 ( 1992 ) .
t.bhattacharya , a.billoire , r.lacaze and th.jolicoeur , j. phys ( paris ) * i 4 * , 122 ( 1994 ) .
d.loison and h.t.diep , phys .
b*50*,16453 ( 1994 ) .
a. mailhot , m.l .
plumer and a. caill , phys .
b * 50 * , 6854 ( 1994 ) .
h.kawamura , phys .
b * 38 * , 4916 ( 1988 ) .
p.azaria , p.delamotte and t.jolicoeur , phys .
* 64 * , 3175 ( 1990 ) .
s.a.antonenko and a.i.sokolov , phys .
b * 49 * , 15901 ( 1994 ) .
h.kadowaki , k.ubukoshi , k.hirakawa , j.l.martinez and g.shirane , j. phys .
japan * 56 * , 4027 ( 1987 ) .
t.e.mason , b.d.gaulin and m.f.collins , phys .
rev . b*39 * , 586 ( 1989 ) b.d.gaulin , t.e.mason , m.f.collins and j.f.larese , phys .
lett . * 62 * , 1380 ( 1989 ) .
d.h.lee , j.d.joannopoulos , j.w.nagele and d.p.landau , phys .
lett . * 52 * , 433 ( 1984 ) ; phys .
b * 33 * , 450 ( 1986 ) . a.v.chubukov and d.i.golosov , j. phys.:condens .
matter * 3 * , 69 ( 1991 ) .
t.momoi , j. stat . phys . * 75 * , 707 ( 1994 ) .
b.d.gaulin , in _ magnetic systems with competing interactions _ , edited by h.t.diep ( world scientific , singapore ) , 286 ( 1994 ) .
m.l.plumer , a.caill , a.mailhot and h.t.diep , in _
magnetic systems with competing interactions _ , edited by h.t.diep ( world scientific , singapore ) , 1 ( 1994 ) .
a.p.ramirez , annu . rev .
* 24 * , 453 ( 1994 ) .
w.e.hatfield , w.e.estes , w.e.marsh , m.w.pickens , l.w .
ter haar and r.r.weller , in _ extended linear chain compounds _ , edited by j.s.miller , plenum press , new york , * 3 * , 43 ( 1983 ) .
s.r.kuindersma , c.haas , j.p.sanchez and r.al , solid state com . * 30 * , 403 ( 1979 ) .
k.hirakawa , h.kadowaki and k.ubukoshi , j. phys .
japan * 52 * , 1814 ( 1983 ) .
t.sato , h.kadowaki , h.masuda and k.iio , j. phys .
japan * 63 * , 4583 ( 1994 ) .
j.l.soubeyroux , d.fruchart , j.c.marmeggi , w.j.fitzgerald , c.delmas and g.le flem , phys . stat . sol .
* 67 * , 633 ( 1981 ) .
h.kadowaki , h.takei and k.motoya , j. phys . : condens .
matter * 7 * , 6869 ( 1995 ) .
h.kadowaki , h.kikuchi and y.ajiro , j. phys . : condens .
matter * 2 * , 4483 ( 1990 ) .
y.oohara , s.mitsuda , h.yoshizawa , n.yaguchi , h.kuriyama , t.asano and m.mekata , j. phys . soc .
japan * 63 * , 847 ( 1994 ) .
h.hori , h.mikami , m.date and k.amaya , physica b * 165&166 * , 237 ( 1990 ) .
w.j.crama , j. solid state chem . * 39 * , 168 ( 1981 ) .
h.yoshizawa , w.kozukue and k.hirakawa , j. phys .
japan * 49 * , 144 ( 1980 ) .
m.melamud , j.makovsky , h.shaked and s.shtrikman , phys .
b * 3 * , 821 ( 1971 ) .
w.b.yelon and d.e.cox , phys .
b * 6 * , 204 ( 1972 ) .
a.hauser , u.falk , p.fischer and h.gudel , j. solid state chem . * 56 * , 343 ( 1985 ) .
m.mekata and k.adachi , j. phys .
japan * 44 * , 806 ( 1978 ) .
k.adachi , n.achiwa and m.mekata , j. phys .
japan * 49 * , 545 ( 1980 ) .
m.melamud , j.makovsky and h.shaked , phys .
b * 3 * , 3873 ( 1971 ) .
v.j.minkiewicz , d.e.cox and g.shirane , solid state commun .
* 8 * , 1001 ( 1970 ) v.j.minkiewicz , d.e.cox and g.shirane , j. phys .
( paris ) c*1 * c1 - 892 ( 1971 ) .
a.harrison and d.visser , j. phys . : condens .
matter * 4 * , 6977 ( 1992 ) .
c.j.glinka , v.j.minkiewicz , d.e.cox and c.p.khattak , magnetism and magnetic materials , * 18*th ann.conf . h.t.witteveen and j.a.r .
van veen , j. phys .
solids * 35 * 337 ( 1974 ) .
w.b.yelon , d.e.cox and m.eibschutz , phys . rev .
b * 12 * , 5007 ( 1975 ) .
b.dorner , d.visser , u.steigenberger , k.kakurai and m.steiner , z.phys.b - condens .
matter * 72 * , 487 ( 1988 ) .
m.eibshtz , r.c.sherwood , f.c.l.hsu and d.e.cox in magnetism and magnetic materials ( denver .
1972 ) proc .
18th annual conf . on magnetism and magnetic materials , aip conf .
c.d.graham jr . and j.j.rhyne ( aip , new york , 1973 ) p 684 .
r.brener , e.ehrenfreund , h.shechter and j.makovsky , j. phys .
solids * 38 * , 1023 ( 1977 ) .
h.w.zandbergen , j. solid state chem . * 37 * , 308 ( 1981 ) . in ref .
@xcite cited as : h.w.zandbergen , thesis , university of leiden ( 1981 ) unpublished .
h.w.zandbergen , j. solid state chem . * 35 * , 367 ( 1980 ) .
h.tanaka , y.kaahwa , t.hasegawa , m.igarashi , s.teraoka , k.iio and k.nagata , j. phys .
japan * 58 * , 2930 ( 1989 ) .
o.a.petrenko , m.f.collins , c.v.stager and z.tun , phys .
b. * 51 * , 9015 ( 1995 ) .
m.steiner , w.kruger and d.babel , solid state commun .
* 9 * , 227 ( 1971 ) .
h.tanaka , k.iio and k.nagata , j. magn . and
magn . mat . *
104 - 107 * , 829 ( 1992 ) .
r.m.morra , w.j.l.buyers , r.l.armstrong and k.hirakawa , phys .
b * 38 * , 543 ( 1988 ) . k.katori , y.ajiro , t.asano and t.goto , j. phys
japan * 64 * , 3038 ( 1995 ) .
i.a zaliznyak , l.a.prozorova and a.v.chubukov , j. phys . : condens .
matter * 1 * , 4743 ( 1989 ) .
k.kakurai , physica b * 180 - 181 * , 153 ( 1992 ) . z.tun , w.j.l.buyers , a.harrison and j.a.rayne , phys .
b * 43 * , 13331 ( 1991 ) . e.cohen and m.d.sturge , solid state commun . *
24 * , 51 ( 1977 ) .
k.nakajima , k.kakurai , h.hiraka , h.tanaka , k.iio and y.endoh , j. phys .
japan * 61 * , 3355 ( 1992 ) .
o.a.petrenko , s.v.petrov and l.a.prozorova , sov .
jetp * 71 * , 406 ( 1990 ) .
k.iio , m.sano and k.nagata , j. magn .
japan * 11 * , suppl.59 ( 1987 ) .
s.maegawa , t.kohmoto and t.goto , phys .
b * 44 * , 12617 ( 1991 ) .
a.harrison , m.f.collins , j.abu-dayyeh and c.v.stager , phys .
b * 43 * , 679 ( 1991 ) .
t.inami , k.kakurai , h.tanaka , m.enderle and m.steiner , j.phys.soc .
japan * 63 * , 1530 ( 1994 ) .
z.tun , t.c.hsu and j-g.lussier , j. appl .
phys . * 75 * , 6063 ( 1994 ) .
m.enderle , k.kakurai , k.n.clausen , t.inami , h.tanaka and m.steiner , europhys .
lett . * 25 * , 17 ( 1993 ) .
s.i.abarzhi , m.e.zhitomirsky , o.a.petrenko , s.v.petrov and l.a.prozorova , sov .
jetp * 77 * , 521 ( 1993 ) .
m.f.collins and b.d.gaulin , j. appl .
phys . * 55 * , 1869 ( 1984 ) .
b.d.gaulin , m.f.collins and w.j.l.buyers , j. appl .
phys . * 61 * , 3409 ( 1987 ) .
u.falk , a.furrer , h.u.gudel and j.k.kjems , phys .
b * 35 * , 4888 ( 1987 ) .
t.inami , k.kakurai , h.tanaka , m.enderle and m.steiner , physica b * 213&214 * , 167 ( 1995 ) .
l.heller , m.f.collins , y.s.yang and b.collier , phys .
b * 49 * , 1104 ( 1994 ) . a.g.abanov and o.a.petrenko , phys . rev .
b * 50 * , 6271 ( 1994 ) .
m.niel , c.cros , m.pouchard and j .- p.chaminade , j. solid state chem .
* 20 * , 1 ( 1977 ) .
r.feile , j.k.kjems , a.hauser , h.u.gudel , u.falk and a.furrer , solid state com . * 50 * , 435 ( 1984 ) . h.tanaka and k.kakurai , physica b * 213 - 214 * , 188 ( 1995 ) . h. tanaka and k. kakurai , j. phys .
* 63 * , 3412 ( 1994 ) .
h.kadowaki , k.hirakawa and k.ubukoshi , j. phys .
japan * 52 * , 1799 ( 1983 ) .
s.itoh , y.endoh , k.kakurai and h.tanaka , phys
lett . * 74 * , 2375 ( 1995 ) ; physica b * 213 - 214 * , 161 ( 1995 ) .
m.niel , c.cros , g. le flem , m.pouchard and p.hagenmuller , physica * 86 - 88b * , 702 ( 1977 ) .
i.yamada , k.ubukoshi and k.hirakawa , j. phys .
japan * 53 * , 381 ( 1984 ) .
h.kadowaki , k.ubukoshi and k.hirakawa , j. phys .
japan * 54 * , 363 ( 1985 ) .
c.delmas , f.menil , g. le flem , c.fouassier and p.hagenmuller , j. phys .
solids * 39 * , 51 ( 1978 ) and c.delmas , g. le flem , c.fouassier and p.hagenmuller , j. phys .
solids * 39 * , 55 ( 1978 ) .
s.angelov , j.darriet , c.delmas and g.le flem , solid state comm . * 50 * , 345 ( 1984 ) .
p.doumerc , a.wichainchai , a.ammar , m.pouchard and p.hagenmuller , mat .
bull * 21 * , 745 ( 1986 ) .
d.juza , d.giegling and h.schfer , z. anorg .
allgem . chem . *
366 * , 121 ( 1961 ) .
k.takeda , n.uryu , k.ubukoshi and k.hirakawa , j. phys .
japan * 55 * , 727 ( 1986 ) .
j.wosnitza , r. deutschmann , h.v lohneysen and r.k.kremer , j. phys . : condens .
matter * 6 * , 8045 ( 1994 ) .
f.tabak , a.lascialfari and a.rigamonti , j. phys . : condens .
matter * 5 * , b31 ( 1993 ) .
m.nishi , y.ito , h.kadowaki and k.hirakawa . j. phys
japan * 53 * , 1214 ( 1984 ) .
m.f.collins , _ magnetic critical scattering _ , oxford university press , oxford , ( 1989 ) .
n.kojima , k.ito , i.mogi , m.takeda , g.kido , y.nakagawa , m.sakai , n.kuroda and y.nishina , j. phys .
japan * 62 * , 4137 ( 1993 ) .
m.suzuki , i.yamada , h.kadowaki and f.takei , j. phys . : condens .
matter * 5 * , 4225 ( 1993 ) y.ajiro , h.kikuchi , s.sugiyama , t.nakashima , s.shamoto , n.nakayama , m.kiyama , n.yamamoto and y.oka , j. phys .
japan * 57 * , 2268 ( 1988 ) .
j.l.soubeyroux , d.fruchart , c.delmas and g.le flem , j. magn . and
magn . mater .
* 14 * , 159 ( 1979 ) s.angelov and j.p.doumerc , solid state comm . * 77 * , 213 ( 1991 ) . h.kawamura and s.miyashita , j. phys .
japan * 53 * , 4138 ( 1984 ) .
k.hirakawa , h.kadowaki and k.ubukoshi , j. phys .
japan * 54 * , 3526 ( 1985 ) .
h.kadowaki , k. ubukoshi and k.hirakawa , unpublished .
cited by @xcite r.h.clark and w.g.moulton , phys .
b * 5 * , 788 ( 1972 ) .
d.beckmann , j.wosnitza , h von lohneysen and d.visser , phys .
lett . * 71 * , 2829 ( 1993 ) .
p.b.johnson , j.a.rayne and s.a.friedberg , j. appl . phys . * 50 * , 1853 ( 1979 ) .
h.kadowaki , k.ubukoshi and k.hirakawa , j. phys .
japan * 56 * , 751 ( 1987 ) .
d.e.cox and v.j.minkiewicz , phys .
b * 4 * , 2209 ( 1971 ) .
w.b.yelon and d.e.cox , phys .
b * 7 * , 2024 ( 1972 ) .
m.poirier , a.caill and m.l.plumer , phys .
b * 41 * , 4869 ( 1990 ) .
i.a.zaliznyak , v.i.marchenko , s.v.petrov , l.a.prozorova and a.v.chubukov , jetp lett . , * 47 * , 211 ( 1988 ) h.tanaka , s.teroaka , e.kakehashi , k.iio and k.nagata , j. phys .
japan * 57 * , 3979 ( 1988 ) .
t.kambe , h.tanaka , s.kimura , h.ohta , m.motokawa and k.nagata , j. magn .
magn . mater . *
140 - 144 * , 1967 ( 1995 ) .
y.trudeau , m.l.plumer , m.poirier and a.caill , phys .
b * 48 * , 12805 ( 1993 ) .
y.oohara , h.kadowaki and k.iio , j. phys .
japan * 60 * , 393 ( 1991 ) .
h.a.katori , t.goto and y.ajiro , j. phys .
japan * 62 * , 743 ( 1993 ) .
m.sano , k.iio and k.nagata , phys .
b * 39 * , 9753 ( 1989 ) .
h.kadowaki , t.inami , y.ajiro , k.nakajima and y.endoh , j. phys .
japan * 60 * , 1708 ( 1991 ) .
m.enderle , g.fortuna and m.steiner , j. phys . condens .
matter * 6 * , l385 ( 1994 ) .
y.ajiro , t.inami and h.kadowaki , j. phys .
japan * 59 * , 4142 ( 1990 ) .
t.oguchi , phys . rev .
* 117 * , 117 ( 1960 ) .
p - a . lindgard and a.kowalska , j. phys .
c : solid state phys .
* 9 * , 2081 ( 1976 ) .
m.e.zhitomirsky and i.a.zaliznyak , phys .
b * 53 * , 3428 ( 1996 ) .
z.tun , w.j.l.buyers , r.l.armstrong , k.hirakawa and b.briat , phys .
b * 42*,4677 ( 1990 ) .
i.affleck , j. phys . : condens
. matter , * 1 * , 3047(1989 ) , i.affleck and g.f.wellman , phys . rev . ,
b*46 * , 8934(1992 ) .
i.a.zaliznyak , solid state comm .
* 84 * , 573 ( 1992 ) .
m.l.plumer , k.hood and a.caill , phys.rev . lett .
* 60 * , 45 ( 1988 ) .
h.kawamura , a.caill and m.l.plumer , phys .
b * 41 * , 4416 ( 1990 ) .
h.kawamura , phys .
b * 47 * , 3415 ( 1993 ) .
m.enderle , r.schneider , y.matsuoka and k.kakurai , physica b , to appear ( 1997 ) .
h.weber , d.beckmann , j.wosnitza , h.v.lhneysen and d.visser , intern . j. modern phys .
b * 9 * , 1387 ( 1995 ) .
h.tanaka , h.nakano and s.matsuno , j. phys .
japan * 63 * , 3169 ( 1994 ) a.n.bazhan , i.a.zaliznyak , d.v.nikiforov , o.a.petrenko , s.v.petrov and l.a.prozorova , sov .
jetp * 76 * , 342 ( 1993 ) .
t.kato , t.ishii , y.ajiro , t.asano and s.kawano , j. phys
japan * 62 * , 3384 ( 1993 ) .
o.a.petrenko , m.f.collins , c.v.stager , b.f.collier and z.tun , j. appl . phys . * 79 * , 6614 ( 1996 ) .
m.l.plumer , a.caill and k.hood , phys .
b * 39 * , 4489 ( 1989 ) m.kerszberg and d.mukamel , phys . rev .
b * 18 * , 6283 ( 1978 ) . j.a.oyedele and m.f.collins , can .
* 56 * , 1482 ( 1978 ) .
a.v.chubukov , j. phys .
c : solid state phys . * 21 * , l441 ( 1988 ) .
t.e.mason , m.f.collins and b.d.gaulin , j. appl . phys . * 67 * , 5421 ( 1990 ) .
s.i.abarzhi , sov .
jetp * 76 * , 676 ( 1993 ) .
t.e.mason , c.v.stager , b.d.gaulin and m.f.collins , phys .
b * 42 * , 2715 ( 1990 ) .
t.e.mason , y.s.yang , m.f.collins , b.d .
gaulin , k.n .
clausen and a.harrison , j. magn . and magn .
mater * 104 - 107 * , 197 ( 1992 ) t.goto , t.inami and y.ajiro , j. phys .
japan * 59 * , 2328 ( 1990 ) .
b.ya.kotyuzhanskii and d.v.nikiforov , j. phys . : condens .
matter * 3 * , 385 ( 1991 ) .
s.i.abarzhi , a.n.bazhan , l.a.prozorova and i.a.zaliznyak , j. phys . :
matter * 4 * , 3307 ( 1992 ) .
i.a.zaliznyak , l.a.prozorova and s.v.petrov , sov .
jetp * 70 * , 203 ( 1990 ) .
p.santini , z.domanski , j.dong and p.erdos , phys .
b * 54 * , 6327 ( 1996 ) . m.l .
plumer and a. mailhot , phys .
b * 50 * , 16113 ( 1994 ) .
e.h.boubcheur , d.loison and h.t.diep , phys .
rev . b*54 * , 4165(1996 ) .
plumer and a.mailhot , phys .
( submitted 1996 ) .
t.e.mason , m.f.collins and b.d.gaulin , j. phys .
c * 20 * , l945 ( 1987 ) .
y.ajiro , t.nakashima , y.unno , h.kadowaki , m.mekata and n.achiwa , j. phys
japan * 57 * , 2648 ( 1988 ) .
t.kato , t.asano , y.ajiro , s.kawano . t.ishii and k.iio , physica b * 213&214 * , 182 ( 1995 ) .
h.kadowaki , s.m.shapiro , t.inami and y.ajiro , j. phys .
japan * 57 * , 2640 ( 1988 ) .
j.wang , d.p.belanger and b.d.gaulin , phys .
lett . * 66 * , 3195 ( 1991 ) .
r.deutschmann , h. von lohneysen , j.wosnitza , r.k.kremer and d.visser , euro .
* 17 * , 637 ( 1992 ) .
m.eibschtz , g.r.davidson , d.e.cox , aip conf . proc . * 18 * , 386 ( 1973 ) .
h.kawamura , prog .
* 101 * , 545 ( 1990 ) .
w.zhang , w.m.saslow , and m.gabay , phys .
b * 44 * , 5129 ( 1991 ) ; w.zhang , w.m.saslow , m.gabay , and m.benakli , _ ibid_. * 48 * , 10204 ( 1993 ) .
m.l.plumer , a.caill and h.kawamura , phys .
b * 44 * , 4461 ( 1991 ) m.e.zhitomirsky , o.a.petrenko and l.a.prozorova , phys . rev .
b * 52 * , 3511 ( 1995 ) .
d.visser , g.c.verschoor and d.j.w.ijdo , acta . cryst . *
b36 * , 28 ( 1980 ) .
h. von fink and h .- j .
seifert , acta crystallogr .
* b38 * , 912 ( 1982 ) . k. machida , t. mitsui , t. kato and k. iio , solid state comm . * 91 * , 17 ( 1994 ) .
o.a.petrenko , m.a.lumsden , m.lumsden and m.f.collins , j. phys . : condens .
matter * 8 * , 10899 ( 1996 ) > i.m.vitebskii , o.a.petrenko , s.v.petrov and l.a.prozorova , sov .
jetp * 76 * , 178 ( 1993 ) .
m.e.zhitomirsky , phys .
b * 54*,353 ( 1996 ) .
s.e.nagler , w.j.l.buyers , r.l.armstrong and b.briat , phys .
b * 27 * , 1784 ( 1983 ) .
g.h.wannier , phys .
b * 79 * , 357 ( 1950 ) .
u.tellenbach and h.arend , j. phys c : solid state phys .
* 10 * 1311 ( 1977 ) .
h.yoshizawa , and k.hirakawa , j. phys .
japan * 46 * 4487 ( 1979 ) .
h.yoshizawa , k.hirakawa , s.k.satiya and g.shirane , phys .
b * 23 * , 2298 ( 1981 ) h.yoshizawa , j.d.axe and g.shirane , solid state commun . * 38 * , 241 ( 1981 ) .
h.hori , h.mikami , m.date , y.ajiro and n.mori , j. magn .
magn . mat . *
104 - 107 * , 815 ( 1992 ) .
j.b.ward , v.h.mccann , q.a.pankhurst , w.l.hassett and d.c.price , j. phys .
c : solid state phys . * 20 * , 1689 ( 1987 ) .
m.mekata , s.okamoto , s.onoe , s.mitsuda and h.yoshizawa , j. magn . and
magn . mat .
* 90&91 * , 267 ( 1990 ) .
k.amaya , h.hori , i.shiozaki , m.date , m.ishizuka , t.sakakibari , t.goto , n.miura , h.kikuchi and y.ajiro , j. phys .
japan * 59 * , 1810 ( 1990 ) .
s.bocquet , j.b.ward and v.h.mccann , j. phys .
c : solid state phys . * 21 * , 367 ( 1988 ) .
d.j.lockwood , i.w.johnstone , h.j.labbe and b.briat , j. phys .
c : solid state phys . * 16 * , 6451 ( 1983 ) .
s.e.nagler , w.j.l.buyers , r.l.armstrong and b.briat , phys . rev . * 28 * , 3873 ( 1983 ) .
r.rogge , phd thesis , mcmaster university , hamilton , canada ( 1994 ) .
a.farkas , b.d.gaulin , z.tun and b.briat , j. appl . phys . * 69 * , 6167 ( 1991 ) .
j.p.boucher , l.p.regnault , j.rossat-mignod , y.henry , j.bouillot and w.g.stirling , phys .
b * 31 * , 3015 ( 1985 ) .
j.wang , d.p.belanger and b.d.gaulin , phys .
b * 49 * , 12299 ( 1994 ) .
a.n.berker , g.s.grest , c.m.soukoulis , d.blankschtein and m.ma , j. appl . phys . * 55 * , 2416 ( 1984 ) ; d.blankschtein , m.ma , a.n.berker , g.s.grest and c.m.soukoulis , phys . rev .
b * 29 * , 5250 ( 1984 ) .
m.l.plumer , a.caill and k.hood , phys.rev .
b * 40 * , 4958 ( 1989 ) .
m.melamud , h.pinto , j.makovsky and h.shaked , phys .
( b ) * 63 * 699 ( 1974 ) .
m t.kurata and h.kawamura , j. phys .
japan * 64 * , 232 ( 1995 ) .
s.n.coppersmith , phys .
b * 32 * , 1584 ( 1985 ) .
t.kohmoto , t.goto , s.maegawa , n.fujiwara , y.fukuda , n.kunitomo and m.mekata , phys .
b * 52 * , 12526 ( 1995 ) .
f.matsubara and s.ikeda , phys .
b * 28 * , 4064 ( 1983 ) .
f.matsubara and s.inawashiro , j. phys .
japan , * 53 * , 4373 ( 1984 ) .
f.matsubara and s.inawashiro , j. phys .
japan , * 56 * , 2666 ( 1987 ) . o.heinonen and r.g.petschek , phys .
b * 40 * , 9052 ( 1989 ) .
a.bunker , b.d.gaulin and c.kallin , phys .
b * 48 * , 15861 ( 1993 ) ; _ ibid _ * 52 * , 1415 ( 1995 ) .
m.l.plumer and a. mailhot , phys .
b * 52 * , 1411 ( 1995 ) .
m.l.plumer , a.mailhot , r.ducharme , a.caill , and h.t.diep , phys .
b * 47 * , 14312 ( 1993 ) .
w.j.l.buyers , m.l.hogan , r.l.armstrong and b.briat , phys .
b * 33 * , 1727 ( 1986 ) . z.tun ,
b.d.gaulin , r.b.rogge and b.briat , j. magn . and
magn . mat . *
104 - 107 * , 1045 ( 1992 ) . c.n.yang and c.p.yang , phys
. rev . * 151 * , 258 ( 1966 ) .
a.harrison , c.v.stager and d.visser , j. appl . phys . * 69 * , 5998 ( 1991 ) . d.visser and a.harrison , j. magn . and
magn . mat . *
116 * , 80 ( 1992 ) .
d.visser and a.harrison , j. de physique col . * c8 * , 1467 ( 1988 ) .
m. eibschtz , m.e.lines and r.c.sherwood , phys .
b * 11 * , 4595 ( 1975 ) .
m.steiner , k.kakurai , w.knop , b.dorner , r.pynn , u.happek , p.day and g.mcleen , solid state commun . * 38 * , 1179 ( 1981 ) .
w.knop , m.steiner and p.day , j. magn . and
magn . mater . *
31 - 34 * , 1033 ( 1983 ) . in ref.@xcite cited as : k.takeda _ et al .
_ , private communication about a specific heat measurements on csfecl@xmath5
. in ref.@xcite cited as : d.p.dickson , private communication ( 1981 ) .
p.a.montano , h.shechter , e.cohen and j.makovsky , phys .
b * 9 * , 1066 ( 1974 ) .
b.schmid , b.dorner , d.petitgrand , l.p.regnault and m.steiner , z. phys .
b * 95 * , 13 ( 1994 ) .
m.chiba , y.ajiro , k.adachi and t.morimoto , j. phys .
japan * 57 * , 3178 ( 1988 ) and j. de physique * 49 * , colloque c8 , 1445 ( 1988 ) . n.suzuki and j.makino , j. phys .
japan * 64 * , 2166 ( 1995 ) .
b.schmid , b.dorner , d.visser and m.steiner , z. phys .
b - condens .
matter * 86 * , 257 ( 1992 ) and j. magn . and
magn . mater . *
104 - 107 * , 771 ( 1992 ) .
b.dorner , d.visser , u.steigenberger , k.kakurai and m.steiner , physica b * 156 & 157 * , 263 ( 1989 ) .
d.visser , b.dorner and m.steiner , physica b * 174 * , 25 ( 1991 ) .
g.r.davidson , m.eibschutz , d.e.cox and v.j.minkiewicz , aip conf . proc . * 5 * , 436 ( 1971 ) .
n.suzuki , j. phys .
japan * 50 * , 2931 ( 1981 ) .
m.e.lines and m.eibschts , phys .
b * 11 * , 4583 ( 1975 ) .
k.adachi , k.takeda f.matsubara , m.mekata and t.haseda , j. phys .
japan * 52 * , 2202 ( 1983 ) .
a.harrison and d.visser , phys .
a. * 137 * , 79 ( 1989 ) .
y.takeda , m.shimada , f.kanamura and m.koizumi , j. phys .
japan * 37 * , 276 ( 1974 ) . n.papanicolaou and p.spathis , j. phys . : condens .
matter * 1 * , 5555 ( 1989 ) . n.papanicolaou and p.spathis , j. phys . : condens .
matter * 2 * , 6575 ( 1990 ) .
a.lindgrd and b.schmid , phys .
b * 48 * , 13636 ( 1993 ) .
m.liu and b .- l.zhou , z. phys .
b - condens .
matter * 93 * , 395 ( 1994 ) .
b.dorner , d.visser and m.steiner , z.phys.b - condens .
matter * 81 * , 75 ( 1990 ) .
t.mitsui , k.machida , t.kato and k.iio , j. phys .
japan * 63 * , 839 , ( 1994 ) .
h.shiba , solid state commun .
* 41 * , 511 ( 1982 ) . h.shiba and n.suzuki , j. phys
japan * 51 * , 3488 ( 1982 ) .
t.haseda , n.wada , m.hata and k.amaya , physica * 108b * , 841 ( 1981 ) .
t.tsuboi , m.chiba , h.hori , i.shiozaki and m.date , j. de physique * 49 * , colloque c8 , 1443 ( 1988 ) .
j.a.baines , c.e.johnson and m.f.thomas , j. phys .
c * 16 * , 3579 ( 1983 ) .
a.lindgard , j. magn . and
magn . mater . *
54 - 57 * , 1227 ( 1986 ) .
h.ohta , n.makita , k.yoshida , t.nanba and m.motokawa , international j. of infrared and millimeter waves , * 13 * , 457 ( 1992 ) .
h.ohta , d.donnelly and m.motokawa , j.magn . and magn.mat . ,
* 104 - 107 * , 777 ( 1992 ) .
t.goto , s.maegawa and t.kawai , j. phys .
japan * 55 * , 1066 ( 1986 ) .
n.wada , k.ubukoshi and k.hirakawa , j. phys .
japan * 51 * , 2833 ( 1982 ) .
n.wada , k.sumiyoshi , t.watanabe and k.amaya , j. phys .
japan * 52 * , 1893 ( 1983 ) .
b.dorner , b.schmid , k.kakurai and d.petitgrand , can
. j. phys . *
73 * , 800 ( 1995 ) .
k.amaya , m.ishizuka , t.nakagawa , s.saratani , t.sakakibara , s.takeyama , k.nakao , t.goto , n.miura , y.unno and y.ajiro , j. phys
japan * 57 * , 38 ( 1988 ) .
n.achiwa , j. phys .
japan * 27 * , 561 ( 1969 ) .
p.a.montano , e.cohen , h.shechter and j.makovsky , phys . rev .
b * 7 * , 1180 ( 1973 ) . m.e.lines , phys . rev .
b * 11 * , 1134 ( 1975 ) .
n.suzuki , j. phys .
japan * 45 * , 1791 ( 1978 ) .
n.suzuki , j. phys .
japan * 52 * , 3907 ( 1983 ) .
n.suzuki , j. phys .
japan * 52 * , 1002 ( 1983 ) .
a.w.schlueter , r.a.jacobson and r.e.rundle , inorg . chem . * 5 * , 277 ( 1966 ) .
c.j.kroese , w.j.a.maaskant and g.c.verschoor , acta cryst .
b*30 * , 1053 ( 1974 ) .
m.mekata , y.ajiro , t.sugino , a.oohara , k.ohara , s.yusada , y.oohara and h.yoshizawa , j. magn . and
magn . mat . *
140 - 144 * , 1987 ( 1995 ) .
h.tanaka , k.iio and k.nagata , j. phys .
japan , * 54 * , 4345(1985 ) .
n.stuesser , u.schotte , k.d.schotte and x.hu , physica b * 213 - 214 * , 164 ( 1995 ) . e.rastelli and a.tassi , z. phys .
b * 94 * , 139 234(1994 ) . e.p.stefanovskii and a.l.sukstanskii , jetp * 77 * , 628 ( 1993 ) .
h.ohta , s.imagawa , m.motokawa and h.tanaka , j. phys .
japan * 62 * , 3011 ( 1993 ) .
h.nojiri , y.tokunaga and m.motakawa , j. phys .
( paris ) * 49 * suppl .
c8 , 1459 ( 1988 ) .
m.chiba , y.ajiro and t.morimoto , physica b * 211 * , 196 ( 1995 ) . t.nikuni and h.shiba , j. phys .
japan * 62 * , 3268 ( 1993 ) .
239 n.v.fedoseeva , r.s.gekht , t.a.velikanova and a.d.balaev , jetp lett . * 41 * , 406 ( 1985 ) .
r.s.gekht , n.v.fedoseeva , v.a.dolina and a.d.belaev , phys
. stat . sol . *
155 * , 639 ( 1989 ) .
h.weber , t.werner , j.wosnitza , h.v.lhneysen and u.schotte , phys .
b , * 54 * , 15924 ( 1996 ) .
a.e.jacobs , t.nikuni and h , shiba , j. phys .
japan * 62 * , 4066 ( 1993 ) .
t.ohyama and a.e.jacobs , phys .
b * 52 * , 4389 ( 1995 ) .
d.babel , z. anorg .
, * 369 * , 117 ( 1969 ) .
k.kakurai , r.pynn , b. dorner and m.steiner , j. phys .
c : solid state phys . , * 17 * , l123 ( 1984 ) .
m.steiner , solid state comm . ,
* 11 * , 73 ( 1972 ) .
c.scherer and y.barjhoux , phys .
( b ) * 80 * , 313 ( 1977 ) .
n.suzuki , j. phys .
soc . japan,*52 * , 3199 ( 1983 ) .
m.l.plumer and a.caill , phys .
b * 37 * , 7712 ( 1988 ) .
m.steiner and j.k.kjems , j. phys .
c : solid state phys . , * 10 * , 2665 ( 1977 ) .
m.steiner , phase transitions * 1 * , 269 ( 1980 ) .
j.v.lebesque , j.snel and j.j.smit , solid state comm . ,
* 13 * , 371 ( 1973 ) .
c.dupas and j-p.renard , j. phys .
c : solid state phys . , * 10 * , 5057 ( 1977 ) .
h.yamazaki , e.soares , h.panepucci and y.morishige , j. phys .
japan , * 47 * , 1464 ( 1979 ) .
b.lussier and m.poirier , phys .
b * 48 * , 6199 ( 1993 ) .
m.steiner and h.dachs , solid state comm . , * 14 * , 841 ( 1974 ) . j.m.lovelock and j.villain , j. phys
( paris ) * 38 * , l77 ( 1977 ) .
d.loison and h.t.diep , j. appl
. phys . * 73 * , 5642 ( 1993 ) .
c.l.henley , phys .
, * 62 * , 2056 ( 1989 ) m.mekata , t.tatsumi , t.nakashima , k.adachi and y.ajiro , j. phys . soc .
japan * 56 * , 4544 ( 1987 ) .
m.e.zhitomirsky , o.a.petrenko , s.v.petrov , l.a.prozorova and s.s.sosin , sov .
jetp * 81 * , 185 ( 1995 ) .
h.yamazaki , k.minami and k.katsumata , j. phys . : condens .
matter * 8 * , 8407 ( 1996 ) .
a.harrison and d.visser , j. phys . : condens .
matter * 2 * , 10487 ( 1990 ) . y.oohara and k.iio , j. phys .
japan * 63 * , 4597 ( 1994 ) .
j.takeuchi , t.wada , i.hiromitsu and t.ito , solid state commun * 87 * , 899 ( 1993 ) .
y.trudeau , m.l.plumer , m.poirier and j.takeuchi , phys .
b * 52 * , 378 ( 1995 ) .
h.tanaka , t.hasegawa and k.nagata , j. phys . soc .
japan * 62 * , 4053 ( 1993 ) .
a.harrison , d.visser , p.day , w.knop and m.steiner , j. phys .
c : solid state phys . * 19 * , 6811 ( 1986 ) .
a.harrison and d.visser , j. phys . : condens .
matter * 1 * , 733 ( 1989 ) .
a.harrison , j. phys . : condens .
matter * 1 * , 6695 ( 1989 ) .
a.p.ramirez , g.p.espinosa and a.s.cooper , phys .
lett . * 64 * , 2070 ( 1990 ) .
s.hov , h.bratsberg and a.t.skjeltorp , j. magn . and
magn . mat . *
15 - 18 * , 1980 ( 1980 ) .
p.schiffer , a.p.ramirez , d.a.huse , p.l.gammel , u.yaron , d.j.bishop and a.j.valentino , phys .
lett . * 74 * , 2379 ( 1995 ) . | in this article we review the effects of magnetic frustation in the stacked triangular lattice .
frustration increases the degeneracy of the ground state , giving rise to different physics .
in particular it leads to unique phase diagrams with multicritical points and novel critical phenomena .
we describe the confrontation of theory and experiment for a number of systems with differing magnetic hamiltonians ; heisenberg , heisenberg with easy - axis anisotropy , heisenberg with easy - plane anisotropy , ising and singlet ground state .
interestingly each leads to different magnetic properties and phase diagrams .
we also describe the effects of ferromagnetic , rather than antiferromagnetic , stacking and of small distortions of the triangular lattice .
pacs 75.25.+z , 75.30.kz , 75.4-.-s . |
diblock copolymers are known to self - assemble into a variety of ordered structures , with a length scale ranging from nanometers to micrometers these phases have potential applications in nanolithographic templates,@xmath4 waveguides @xmath5 and dielectric mirrors.@xmath6 the length scale and morphology in the melt can easily be adjusted by controlling the fraction @xmath7 of a monomers in an a / b chain of @xmath8 monomers , and the temperature @xmath9.@xmath10 in experiments one often encounters samples in the lamellar phase ( made up of alternating planar a- and b - rich domains ) in which the melt is only partially ordered .
ordered microdomains , or grains ( typical size in the micrometer range ) , with grain boundaries between them are defects that cost energy . in order to anneal these defects , and to create a perfect alignment of the lamellae , several techniques have been used . in the bulk ,
mechanical shear proved to be a successful technique .
alignment by application of an electric field @xmath11 is also possible , but for macroscopic samples it requires high voltage difference between the two bounding electrodes .
nevertheless , this technique is especially suitable for thin films , because the thickness involved makes the required large fields ( typically @xmath12 ) accessible .
we consider in this paper thin films of lamellar diblock copolymers , under the influence of a perpendicular electric field .
initially , the lamellae are parallel to the confining surfaces , because of preferential short - range interactions with the surfaces . in section [ 7:wsl ]
we consider diblock copolymers in the weak segregation regime .
we show in section [ 7:rowsm ] that electric field applied perpendicular to the surfaces can cause the melt to transform from a parallel to a perpendicular orientation through a first - order phase transition .
the critical field @xmath0 for this transition is caused by a competition between the electric field and surface interactions . in section [ 7:ssl ]
we investigate the thin - film alignment for diblocks in the strong segregation regime . in this regime ,
the surface correlations are finite , and thus the range of parallel ordering induced by the surfaces is finite as well .
we give the transitions between parallel , perpendicular and mixed lamellae in terms of the system parameters , using a phenomenological model . in both weak and strong segregation regimes ,
large distortions are present in the copolymer film , and these could be observed in experiments .
the copolymer order parameter @xmath13 is the deviation of the a monomer local volume fraction from its average @xmath14 . above the order - disorder transition ( odt )
temperature , the melt is in the disordered , homogeneous state , with @xmath15 .
as the temperature is reduced below the odt temperature , the system goes through a first order phase - transition to the lamellar phase provided that @xmath16 is small enough .
close to the odt temperature and in the single - mode approximation , the block copolymer ( bcp ) order parameter is then given by @xmath17 where @xmath18 is the period of lamellar modulations , and @xmath19 is their amplitude ( to be determined later ) . consider a copolymer melt below the odt confined by two flat parallel surfaces at @xmath20 , as in fig . 1 .
the surfaces reduce chain entropy , but also chemically interact with the polymers . the difference in the a and b - block surface interactions
, @xmath21 and @xmath22 , defines the parameter @xmath23 , @xmath24 in general , @xmath25 can be different for the two surfaces , and hence @xmath26 is defined on the @xmath27 surface and @xmath28 is defined for @xmath29 .
the surface interaction , @xmath30 , can be written as an integral over the bounding surfaces ( in units of @xmath31 ) @xmath32 where @xmath33 and @xmath34 .
terms which do not depend on the copolymer order parameter are not important to subsequent calculations , and were dropped out .
the interaction in eq [ 7:fs ] is short - range , localized at the surfaces .
a positive @xmath35 induces adsorption of the b monomers ( @xmath36 ) , and @xmath37 the adsorption of the a monomers ( @xmath38 ) .
we restrict ourselves to homogeneous surfaces , for which @xmath39 are taken to be constants over each of the two surfaces .
first let us consider a bcp film without an external electric field .
we have already considered this case in a previous publication,@xmath40 and briefly review here the bcp behavior .
if the surface affinities @xmath39 are sufficiently large , the lamellae will order in a parallel arrangement .
these lamellae stretch or compress , increasing the bulk free energy , in order to decrease surface energy .
we use below an adaptation of the strong stretching approximation used by turner @xmath41 and walton _ et al .
_ @xmath42 to describe these lamellae .
the lamellar period is @xmath18 and @xmath43 is the closest integer to @xmath44 .
depending on the values of @xmath39 , an integer ( @xmath45 ) or half integer ( @xmath46 ) number of lamellae exist between the two surfaces . in the former case
the ordering is symmetric ( the same type of monomers wet both surfaces ) , while in the latter it is antisymmetric ( a monomers wet one surface , whereas b monomers wet the other surface ) .
the parallel lamellae are described by an order parameter @xmath47 given by @xmath40 @xmath48\\\ ] ] the wavenumber is @xmath49 , and the choice of @xmath50 sign in eq [ 7:phipar ] is such that the surface interactions , eq [ 7:fs ] , are minimized .
the amplitude of sinusoidal modulations @xmath19 is equal to the amplitude of density modulations in a bulk system , and is given below [ eq [ 7:phil ] ] .
we consider now the case where an electric field is turned on , in a direction that is perpendicular to the surfaces ( @xmath51-axis in fig . 1 ) .
under conditions of constant voltage difference across the electrodes situated at the two bounding surfaces , the minimum of the free energy is obtained by maximizing the capacitance . noting that since the a- and b - monomers ( blocks ) have different dielectric constants , the effect of the electric field is to align the bcp layers parallel to the field , i.e. perpendicular to the surfaces . at a certain field strength , @xmath0 , this tendency balances the preference for parallel lamellae as induced by the surfaces . further increase of @xmath52 above the critical value @xmath0 gives rise to a perpendicular lamellar ordering .
close to the odt , the copolymer ordering is weak , and the energetic cost of compressing or bending the lamellae is small . in this regime
the inter - material dividing surface ( imds ) of the lamellar phase , given by the requirement @xmath53 , can be substantially perturbed from its flat state .
this reasoning leads us to the following superposition ansatz .
for zero electric field @xmath52 , surface interactions orient the lamellae in a parallel orientation .
the order parameter is then given by @xmath54 .
the dimensionless amplitude @xmath55 is determined by the strength of the surface interactions , and can be larger than unity if @xmath39 are sufficiently large .
namely , the surface induced order can be stronger than in the bulk . upon increase of the electric field , the function @xmath56 of this parallel state diminishes , while the function @xmath57 of perpendicular lamellae ( parallel to the electric field ) increases .
this can be modeled by using the superposition ansatz for the order parameter @xmath58 in the presence of the field @xmath52 : @xmath59 the order parameter of the perpendicular lamellae is @xmath60 .
it depends only on the wavenumber @xmath61 , and is given in the single - mode approximation ( weak segregation ) by @xmath62 the wavenumber @xmath61 is yet to be determined . without the electric field , @xmath63 , the amplitude of perpendicular modulations vanishes , @xmath64 . as the limit @xmath65
is approached , the effect of the confining surfaces becomes negligible , and the bcp ordering is given by the bulk perpendicular lamellae @xmath66 ( @xmath67 ) .
hence , the weight amplitudes @xmath56 and @xmath57 satisfy the following limits as function of @xmath52 : @xmath68 in order to get explicit expression for the weight functions @xmath69 and @xmath70 we need to consider a specific model .
the free energy we use in this section is applicable to the weak segregation regime , and is given by @xmath71,@xmath72 with @xmath73 from eq [ 7:fs ] , and @xmath74 being the bulk contribution .
this part of the free energy has a polymer ( non - electrostatic ) and electrostatic contributions ( in units of @xmath31 ) , @xmath75 hereafter we restrict ourselves to symmetric ( @xmath76 ) diblock copolymers , where the polymer part of the free energy is @xmath77 @xmath78 the parameters in eq [ 7:fpol ] are given by @xmath79 denoting @xmath80 as the monomer size , the radius of gyration for gaussian chains is @xmath81 .
the polymerization index is @xmath82 , the chain density of an incompressible melt is @xmath83 , and @xmath84 is the flory parameter .
the amplitude @xmath19 in the parallel and perpendicular states is @xmath85 as is obtained by inserting eq [ 7:bulk ] in @xmath86 of eq [ 7:fp ] and minimizing with respect to @xmath19 .
the dimensionless ratio @xmath87 and @xmath88 are of order unity , and are taken to be equal exactly to one in the remaining of the paper , @xmath89 .
the correlation length @xmath90 is assumed to be larger than the film thickness , @xmath91.@xmath92 the electrostatic contribution in units of @xmath93 is @xmath94 @xmath95 here @xmath96 is the fourier transform of @xmath97 : @xmath98 , and @xmath99 is a unit vector in the @xmath100-direction .
copolymer modulations with a non - vanishing component of the wavenumber @xmath100 along the electric field , have a positive contribution to the free energy .
in other words , there is a free energy penalty for having dielectric interfaces in a direction perpendicular to the electric field . in eq [ 7:beta ] ,
@xmath101 and @xmath102 are the dielectric constants of the pure a and b - blocks , respectively , @xmath103 and @xmath104 is the material average dielectric constant for the bcp film , @xmath105 throughout the remaining of this paper we will focus only on symmetric melts ( @xmath76 ) , having an average dielectric constant @xmath106 . for small concentration variations , valid in the weak segregation regime , @xmath107 varies linearly with the local copolymer composition @xmath108 , @xmath109 it is possible to perform the spatial integration in eqs .
[ 7:fs ] , [ 7:fpol ] and [ 7:fel ] , yielding the free energy per unit volume @xmath110 for a general order parameter @xmath97 , eq [ 7:ansatz ] : @xmath111 \nonumber\\ & + & \frac{u\phi_l^4}{64}(w^4+g^4)+\frac{u\phi_l^4}{16}w^2g^2 + \frac{w\sigma}{l } \label{7:f}\end{aligned}\ ] ] the quantities @xmath112 and @xmath113 are positive and given by @xmath114 @xmath115 is related to the surface interaction , and is negative .
the @xmath50 sign is determined from the @xmath50 sign of the order parameter in eq [ 7:phipar ] .
the free energies @xmath116 and @xmath117 of the parallel and perpendicular states , respectively , are given as limiting cases @xmath118
w^2+\frac{u\phi_l^4}{64}w^4+\frac{w\sigma}{l}\\ f_\perp = f(w=0,g=1)&=&\frac14\phi_l^2(\tau+c_\perp ) + \frac{u\phi_l^4}{64}\label{7:f_perp}\end{aligned}\ ] ] as discussed in the introduction , we concentrate on the interesting case where in the absence of electric field , the bcp has a parallel ordering given by @xmath119 ( with some weight @xmath69 ) .
this is equivalent to saying that there exist @xmath69 such that @xmath120 . by inserting @xmath63 and @xmath121 in eqs .
[ 7:f_par ] and [ 7:f_perp ] we get that in this case @xmath122w^2+\frac{u\phi_l^4}{64 } w^4 + \frac{w\sigma}{l}~<~-\frac{\tau^2}{u}&&\end{aligned}\ ] ] therefore , the free energy of the parallel lamellae is expected to be lower than the free energy of the perpendicular lamellae .
this assumption is valid for strong enough surface interactions , @xmath123 . in the next section
we proceed to find the weight functions @xmath56 and @xmath57 for any @xmath124 .
the free energy ( eq [ 7:f ] ) is minimized with respect to @xmath69 and @xmath70 to yield two coupled algebraic equations : @xmath125w-\tau w^3 - 2\tau g^2w+2\sigma / l\phi_l^2=0\end{aligned}\ ] ] as we will see below , there is a critical value of the electric field , @xmath0 , which separates between the small and large field behavior .
one can estimate @xmath0 as the field in which the bulk electrostatic energy @xmath126 ( favoring perpendicular lamellae ) balances the surface interactions @xmath123 , and hence @xmath127 .
equations [ 7:gov_g ] and [ 7:gov_f ] are analyzed separately for small and large electric fields .
+ + * ( i ) small electric fields : @xmath128 . * for zero electric field ,
the lamellae are in their parallel state , namely @xmath64 .
the solution with @xmath129 and @xmath130 [ in eqs .
[ 7:gov_g ] and [ 7:gov_f ] ] corresponds to the minimum of the free energy if @xmath52 is below a certain threshold value @xmath0 , which is to be determined later . denoting @xmath131 and @xmath132 as the weight amplitudes for electric fields @xmath128 , they satisfy the following equations : @xmath133w_<(e)-\tau w_<^3(e)+2\sigma / l\phi_l^2&=&0 \label{7:wg_below}\end{aligned}\ ] ] * ( ii ) large electric fields : @xmath134 . *
there is a solution to eqs .
[ 7:gov_g ] and [ 7:gov_f ] with perpendicular lamellae , namely with a nonzero @xmath57 .
this solution gives the minimum of the free energy above a critical field , @xmath134 .
the weight amplitudes @xmath135 and @xmath136 are then given by : @xmath137w_>(e)+3\tau w_>^3(e)+2\sigma / l\phi_l^2~=~0\end{aligned}\ ] ] the above solution is valid provided that @xmath138 is small enough , as given by the inequality @xmath139 since @xmath56 is a decreasing function , from eq [ 7:c_perp ] we see that increasing the electric field @xmath52 from zero , the natural mode of the perpendicular state , @xmath121 , is the first to become critical . namely , the inequality above is obeyed first when @xmath61 is equal to the bulk mode @xmath140 , and only later by other @xmath141-modes . for this reason we assume hereafter that @xmath142 , yielding @xmath143 and @xmath144 .
thus , bcp modulations in a direction parallel to the surfaces have the bulk ( free ) periodicity @xmath145 .
equation [ 7:wg_above ] is a cubic equation for @xmath138 and has an analytical solution .
it is convenient to express the solution via a parameter @xmath146 defined as @xmath147\equiv -\frac{\sigma / l\phi_l^2}{3\tau } \left(\frac{\tau - c_\parallel(e)}{9\tau}\right)^{-3/2}<0\label{7:theta}\ ] ] and take it to be in the range @xmath148 .
recalling that @xmath149 and @xmath150 are negative , the solution @xmath138 to eq [ 7:wg_above ] is then simply given by @xmath151 curves of @xmath57 and @xmath56 are shown in fig . 2 within the assumption @xmath121 .
for zero electric field , the film has parallel lamellae , @xmath152 .
with increasing electric field @xmath52 , the amplitude @xmath132 of parallel lamellae decreases monotonically , as is given by eq [ 7:wg_below ] , while @xmath131 remains zero . at the critical field ,
@xmath153 , there is a first - order phase transition , and perpendicular lamellae appear .
the weight amplitudes @xmath56 and @xmath57 are discontinuous , the jump in their values is determined by the degree of segregation @xmath154 , inter - surface separation @xmath155 and surface parameters @xmath39 .
further increase of the electric field causes @xmath57 to increase while @xmath56 decreases . for very large electric fields , @xmath65 ,
the perpendicular state saturates to its value @xmath67 , and parallel lamellae completely disappear , @xmath156 .
the critical field @xmath0 is determined by the condition @xmath157 where the free energy @xmath158 is taken from eq [ 7:f ] .
namely , it is the field where the two values of the free energy cross .
the critical field @xmath0 as a function of surface separation @xmath155 is shown in fig . 3 a for weak segregations ( @xmath159 ) .
as the surface separation @xmath155 increases @xmath0 decreases with typical oscillations of period @xmath145 .
these oscillations are caused by the frustration occurring when the surface separation is incommensurate with the lamellar period .
figure 3 b shows a log - log plot of the same curve .
the dashed line shows a fit to a @xmath127 scaling .
this scaling can be obtained by balancing the surface energy @xmath123 with the electrostatic contribution @xmath160 .
there is good fit between this @xmath161 line and the peak positions , as expected for unstrained films.@xmath162 however , for frustrated films ( where the film thickness is incommensurate with the lamellar period ) the deviation from @xmath127 becomes increasingly important as the surface separation @xmath155 is reduced below roughly @xmath163 for the parameters used .
we define @xmath164 as the value of @xmath69 just before the transition ( @xmath165 ) , and @xmath166 the value after the transition ( @xmath167 ) , and similarly for @xmath70 .
the jump in the weight amplitudes @xmath168 and @xmath169 is @xmath170 these quantities are plotted as a function of surface separation @xmath155 in fig .
3 c , for the same parameters as in fig .
3 a. as @xmath155 increases , the critical field @xmath0 decreases , the jump in @xmath69 and @xmath70 gets larger , and the transition from parallel to perpendicular lamellae becomes more abrupt . in fig .
4 we show how the film changes its orientation and morphology as the electric field increases . in part
a we show a contour plot of the copolymer order parameter @xmath171 , for @xmath128 , but only slightly below it .
the ordering in the film is parallel to the surfaces , as @xmath172 . in part
b the field is slightly increased above its threshold value @xmath0 , and the undulations created by the appearance of the perpendicular state are prominent .
in contrast to the classical helfrich - hurault undulations @xmath173 calculated for smectic and cholesteric liquid crystals , here the lateral wavelength is finite , and is equal to the free periodicity @xmath145 . note that the correlation length @xmath174 is larger than the film thickness , @xmath175 . in addition
, the modulations of the adjacent inter - material dividing surfaces ( imds ) , given by @xmath15 , are shown in part c to be out - of - phase with each other and are pronounced only for @xmath176 .
although the perpendicular ordering may be very strong , some parallel ordering is still present ( finite @xmath55 ) . as the electric field is further increased , there is only little reminiscence of the parallel ordering , and the perpendicular lamellae are nearly perfect .
this can be seen in fig .
4 d , where we choose @xmath177 .
in this section we consider the same alignment phenomenon under an electric field as in the previous section , but the bcp melt is assumed to be in the strong segregation limit , i.e. @xmath178 . in this regime
the lamellae are not easily deformed , and the effect of the surface field is important only close to the confining walls , in contrast to the long - range ordering induced by the surfaces in the weak segregation .
we discuss first the alignment phenomenon ignoring the effect of incommensurability between the film thickness @xmath155 and the free lamellar period @xmath145 . in the following section generalization to strained films
will be presented as well . for simplicity
, we also assume that the two walls are chemically identical , and preferring the b monomers : @xmath179 , @xmath35 . for sufficiently strong @xmath52 fields we expect to nucleate a region in the middle of the film with lamellae perpendicular to the walls , as is shown schematically in fig . 5 .
hence two regions of a `` t - junction '' morphology will exist in the film , in the vicinity of the bounding surfaces . a positive energy penalty ( per unit area ) @xmath180
is associated with each of the two t - junction defects . in principle
, other types of defects might exist in the film , but this can only affect the value of @xmath180 and not the system behavior as is described below .
we denote the size of the region that is perpendicular to the surfaces by @xmath181 . if @xmath182 the film has only parallel ordering , while for @xmath183 the perpendicular ordering spans the whole film
. the free energies per unit area , @xmath116 and @xmath117 , for these two extreme cases are : @xmath184 @xmath185 is the polymer free energy per unit volume of a bcp in the lamellar phase , and in the above we have used two different depolarization factors , @xmath186 and @xmath187 , for the two orientations @xmath188 these two depolarization factors can be obtained by calculating the electrostatic energy @xmath189 of the film and using the boundary condition that the displacement field @xmath190 is continuous across the two dielectric boundaries . note that @xmath191 , meaning that @xmath192 , and the perpendicular state is favored for @xmath65 .
let us focus on the mixed state as is illustrated in fig .
5 . for intermediate values of @xmath181 , @xmath193
, there are two surface regions of parallel orientation and a central region of perpendicular orientations .
the free energy of this mixed state @xmath194 ( per unit area ) is : @xmath195 the system can be regarded as being composed of two capacitors of parallel lamellae and one capacitor of perpendicular lamellae connected in series , yielding the constant @xmath196 @xmath197^{-1}\ ] ] we note that a similar expression for the electrostatic part of the mixed state free energy was previously derived by pereira and williams .
@xmath162 the value of the surface energies @xmath21 , @xmath22 and @xmath180 must depend on @xmath181 . to see this consider , for example , @xmath198 and @xmath199 . in this case
@xmath180 is some constant .
but as @xmath200 , the value of @xmath180 must approach zero , because when @xmath183 the t - junction does not exist , and the energy associated with it is zero .
we denote by @xmath201 the cutoff length , which is the characteristic width at which @xmath180 goes to zero . from the same reason @xmath180
must also tend to zero as @xmath202 , with another cutoff length . for simplicity
we assume that this cutoff length is also equal to @xmath201 .
similarly , the surface interaction energies @xmath21 and @xmath203 tend to zero as @xmath200 , with the same cutoff length .
the qualitative forms of these parameters are shown in fig .
6 a. the smooth monotonic decay to zero at @xmath183 and @xmath182 is only suggestive .
@xmath204 has oscillations with period @xmath205 . in principle
, one should minimize @xmath194 with respect to the size of perpendicular domain @xmath181 .
this generalized ( mixed ) state would correspond to the parallel state when the minimum is at @xmath182 , and to the perpendicular state when the minimum is at @xmath183 .
however , we do not know from a molecular description how @xmath21 , @xmath22 and @xmath180 fall off to zero , and therefore use the approximation shown in fig .
6 b. @xmath206 and @xmath207 are constant for @xmath208 and zero for @xmath209 .
@xmath210 is constant in the range @xmath211 , and zero otherwise . to make the notation simpler we drop hereafter the superscript zero of @xmath21 , @xmath22 and @xmath180
note that the effect of incommensurability between the surface spacing @xmath155 and the lamellar period @xmath145 , appears as undulations in the plot of @xmath204 ( fig .
6 a ) , and these undulations are neglected here and will be addressed below separately . with the above assumptions , the size @xmath181 of a perpendicular domain that minimizes the mixed free energy @xmath194 is @xmath212 .
the free energy of the mixed configuration is thus taken from eq .
[ 7:f_mixed_strong ] with @xmath213 given by @xmath214^{-1}\ ] ] using the assumption that the b - polymer is adsorbed at the surfaces in the parallel state ( @xmath215 ) , we plot in fig .
7 @xmath116 , @xmath117 and @xmath194 as a function of electric field strength @xmath52 .
the dimensionless parameter @xmath216 measuring the difference in a and b - block surface interactions is defined as : @xmath217 based on the value of @xmath216 , we now discuss two cases : + * ( i ) strong surface fields : @xmath218*. if @xmath216 is larger than a threshold value @xmath219 , given by @xmath220 there are two distinct critical fields @xmath1 and @xmath2 given by @xmath221^{1/2 } \label{7:e1a}\\\nonumber \\ e_2&=&\left[\frac{\gamma_{_t}(\delta-2)}{l(\lambda_\perp- \lambda_m)}\right]^{1/2 } \label{7:e2a}\end{aligned}\ ] ] the smallest of these two fields , @xmath1 , obtained for @xmath222 , ( fig . 7 a ) is the field required to create the t - junction defect in the film .
this field is independent of @xmath216 , and for thick films ( @xmath223 ) it scales as @xmath224 . for @xmath52 fields in the range @xmath225 , the film has a region of size @xmath212 with perpendicular lamellae , while parallel lamellae are localized in a small region of size @xmath201 near the two walls .
the second field , @xmath3 , obtained for @xmath226 , is larger than @xmath1 and corresponds to the electric field that is required to destroy the parallel surface layer ( of width @xmath201 ) .
note that although these fields are large ( typically between @xmath227@xmath228 v/@xmath229 m ) , they can be readily achieved in thin - film experiments @xcite .
for thick films @xmath3 is independent of @xmath155 and obtains the asymptotic value @xmath230^{1/2}\ ] ] note also the appearance of @xmath231 and @xmath232 in the denominator of the two critical fields in eqs [ 7:e1a ] and [ 7:e2a ] . if the dielectric contrast between a and b domains is small , namely @xmath233 , then @xmath234 , and the critical fields are large . if , on the other hand , @xmath235 , then the critical fields required to achieve mixed and perpendicular lamellae are small . for many polymer surface combinations , @xmath180 is large and therefore @xmath216 is small . in this case
the system is classified as having weak surface fields as is discussed below .
+ * ( ii ) weak surface fields : @xmath236*. for small values of @xmath216 , namely @xmath236 , the free energies @xmath117 and @xmath116 are always smaller than @xmath194 .
namely , the mixed state is not a possible minimum of the film free energy .
there is only one transition at a critical field @xmath237 occurring when @xmath117 intersects @xmath116 ( fig .
this field is given by : @xmath238^{1/2}\nonumber\\ & = & \left[\frac{\sigma}{l(\lambda_\perp-\lambda _ \parallel)}\right]^{1/2}\sim~\delta^{1/2}~l^{-1/2}\end{aligned}\ ] ] when @xmath239 , a direct transition occurs from a state where the whole film has parallel lamellae ( @xmath240 ) to a state where the whole film has perpendicular lamellae without any surface regions ( @xmath241 ) .
the system behavior can be summarized in a phase diagram which depends on the three system parameters : the electric field @xmath52 , the surface interaction parameter @xmath216 and the film thickness @xmath155 . in fig .
8 a , we show a 2-dimensional cut through the phase diagram , varying @xmath52 and @xmath216 while keeping @xmath155 constant . for small electric fields ,
the lamellae are in the fully parallel configuration .
if @xmath216 is small , a first - order phase transition to the fully perpendicular state occurs when @xmath52 is increased above @xmath237 , eq [ 7:e1b ] .
if @xmath216 is large enough , namely @xmath218 , there are two transitions when the field is increased . in the regime @xmath242 ,
the film is in a mixed state , and layers of thickness @xmath201 with parallel lamellae still exist close to the surfaces . as @xmath52 is further increased above @xmath3 , the film has a fully perpendicular state .
note that when @xmath218 , @xmath1 is the field required to initiate the t - junction defect , and so is independent of @xmath23 , while @xmath3 is the field that destroys the surface layer , and therefore @xmath243 . as we have seen , the condition @xmath218 is required for the existence of the mixed state , and hence for the existence of two critical fields @xmath1 and @xmath3 .
alternatively , this condition can be viewed as a restriction on the film thickness @xmath155 for a given @xmath216 .
therefore , a mixed state does not exist if @xmath244}{\lambda_\parallel [ ( \delta+2)\lambda_\perp + ( \delta-2)\lambda_\parallel]}\ ] ] in the limit @xmath245 and in the case of polystyrene ( @xmath246 ) and polymethyl methacrylate ( @xmath247 ) diblock copolymer , one obtains @xmath248 .
it is smaller for smaller values of @xmath216 .
hence the mixed state exists for film thickness larger than the range of surface induced ordering @xmath201 .
we see here again why this mixed state is not expected to occur in the weak segregation limit .
an alternative cut through the phase diagram is shown in fig .
8 b , where @xmath216 is fixed while @xmath52 and @xmath155 are allowed to vary . for small surface separations , @xmath249 ,
there is a transition from parallel to perpendicular lamellae at @xmath239 [ eq [ 7:e1b ] ] . for larger separations , @xmath250
, the mixed state appears at @xmath251 [ eq [ 7:e1a ] ] .
the transition to a fully perpendicular state occurs at @xmath252 [ eq [ 7:e2a ] ] .
the main difference between this diagram and the one obtained by pereira and williams,@xmath162 is that we always find the perpendicular state to be favored for large enough electric fields .
strained films occur when the film thickness @xmath155 does not match the lamellar period @xmath145 , namely @xmath44 is not an integer or a half - integer number .
the effect of this mismatch , which was neglected in the previous section , is considered below . the strong segregation theory of turner @xmath41 and walton _ et al _ @xmath42 has been successful in describing confined lamellae , and therefore we use it in this section to include the effect of lamellae frustration , and to modify the free energy @xmath116 from section [ 7:unstrained ] .
we use the convention that @xmath43 is the closest integer to @xmath44 , yielding @xmath116 which is the minimum of @xmath253 and @xmath254 , the symmetric and antisymmetric parallel states free energies , respectively : @xmath255 + 2\sigma_{\rm bs}-\lambda_\parallel le^2 \nonumber\\ & & { \rm for}~~ n = m,~~{\rm symmetric}\label{7:48}\\ & & \nonumber\\ f^{\rm as}_\parallel & = & \sigma_{\rm ab}\left[\left(\frac{l}{d_0}\right)^3\cdot \frac{1}{n^2}+2n\right ] + \sigma_{\rm as}+\sigma_{\rm bs}-\lambda_\parallel le^2 \nonumber \\ & & { \rm for}~~ n = m\pm \frac12,~~{\rm antisymmetric}\label{7:49}\end{aligned}\ ] ] note that in the above expressions for the free energy , eqs [ 7:48 ] and [ 7:49 ] , the electrostatic part is exactly like in sec . [
7:unstrained ] . in the symmetric state
, both surfaces are wetted by b monomers , while in the antisymmetric state a monomers adsorb to one surface and b monomers adsorb to the second surface . in fig .
9 we present the phase diagrams when the mismatch between @xmath155 and @xmath145 is taken into account , plotted with same parameters as in fig .
for very thin films , @xmath256 , the perpendicular state is favored over the parallel state , even for @xmath63 . apart from this ,
9 a is similar to fig
. 8 a. however , fig .
9 b is different than fig . 8 b. the critical fields @xmath1 and @xmath237 separating the parallel state from the other two states have oscillations with period @xmath145 .
these curves are similar to the critical field curve for weakly segregated lamellae , fig . 3 , computed in sec . 3 , and also to the curve calculated by ashok _ et al .
_ using a similar model and different parameters.@xmath257 note that from the discussion of unstrained films in section [ 7:unstrained ] , @xmath1 and @xmath237 have oscillations around lines which decay as @xmath258 for large @xmath155 .
the field @xmath3 separating the mixed state with the perpendicular state depends only weakly on @xmath155 .
in this paper we study the influence of an applied electric field on the morphology of thin film diblock copolymers . in the absence of an electric field
the lamellae are taken to be ordered parallel to the confining surfaces .
strong enough @xmath52 fields in the direction perpendicular to the surfaces will eventually orient the lamellae in a perpendicular direction .
however , the response of weakly segregated lamellae is different than the response of strongly segregated lamellae . in the former case ,
the bcp order parameter is obtained as a function of electric field strength @xmath52 and other system parameters ( such as the degree of segregation @xmath154 , strength of surface interactions @xmath39 and film thickness @xmath155 ) .
the field applied in the perpendicular direction diminishes the amplitude of the parallel bcp state . above the critical field ,
@xmath134 , a first - order phase transition occurs , from the parallel into the perpendicular state . the first @xmath141-mode which becomes stable in the perpendicular state @xmath61 has a finite periodicity , equal to the bulk spacing @xmath18 .
moreover , modulations of adjacent inter - material dividing surfaces ( imds ) are out - of - phase with each other .
this is different than the strong segregation instability studied by onuki and fukuda,@xmath259 where the slowly varying phase @xmath260 $ ] is used , and the free energy is expanded in small @xmath261 . at the onset of instability , they found that adjacent imds lines are in - phase with each other .
the weak segregation imds undulations studied here appear because in this regime it is relatively easy to deform the lamellae .
for @xmath134 , parallel and perpendicular lamellae coexist in the film .
this superposed state is yet to be verified in experiments .
the critical field @xmath0 as a function of inter - surface separation @xmath155 decays with characteristic oscillations of period @xmath145 .
these oscillations are the result of lamellar frustration , occurring when the period @xmath262 is different than @xmath145 ( while @xmath61 is taken always to be equal to @xmath140 ) .
the deviation of the critical field @xmath0 from the @xmath263 scaling ( of unstrained films ) is important only for small surface separations . for large @xmath155 values ,
the @xmath161 scaling describes well the system behavior .
the weak segregation treatment we present relies on mean - field theory .
it is valid close to the critical point , but not too close where critical fluctuations become important.@xmath264 we have ignored the deviations of the inter - material dividing surface from the perfect flat shape that occur near the surfaces .
these deviations can be important for small surface separations ( @xmath265 ) , or close to the odt,@xmath266 and should be properly accounted for . in the second part of the paper we consider bcp film in the strong segregation regime .
in contrast to the weak segregation case , here the surface induced ordering has a finite range of length @xmath201 .
this added length ( taken in the weak segregation limit to be much larger than the film thickness @xmath155 ) results in the existence of two critical fields . provided that the parameter @xmath216 is large enough , @xmath218 [ eq [ 7:deltastar ] ] , for small electric fields , @xmath267 , the film has parallel lamellae . as the field
is increased above @xmath1 but below @xmath3 , a region of size @xmath268 of perpendicular lamellae nucleates in the middle of the film . in this mixed state ,
a layer of parallel lamellae of thickness equal to the cutoff length @xmath201 still persists close to the surfaces . when @xmath269 the system is in the perpendicular state . the full phase diagram in the @xmath52-@xmath216 and the @xmath52-@xmath155 planes is given in fig .
8 . as discussed above , for large @xmath218
there is a transition from parallel to mixed lamellae at @xmath251 [ eq [ 7:e1a ] ] and from mixed to perpendicular lamellae at @xmath252 [ eq [ 7:e2a ] ] . another scenario of a direct transition from a parallel to a perpendicular state at @xmath239 [ eq [ 7:e1b ] ] is realized for small @xmath216 , @xmath236 .
these diagrams do not take into account the frustration effects caused by a mismatch between the wall separation and the natural lamellar periodicity in the bulk . in this respect , they are qualitatively applicable to non - polymeric systems such as ferrosmectics in magnetic fields,@xmath270 and to bcp in the hexagonal phase , as investigated experimentally .
@xmath271 the difference between the hexagonal and lamellar phase behavior comes from the different value of the depolarization factors @xmath186 , @xmath187 , and different numerical coefficients in eqs .
[ 7:f_para_strong ] , [ 7:f_perp_strong ] and [ 7:f_mixed_strong ] .
these modifications only change the critical fields @xmath1 , @xmath3 and @xmath237 by a numerical factor , but do not alter the phase behavior .
indeed , two critical fields between which a regime of mixed state exists have been found experimentally for a bcp is its hexagonal phase by russell and co - workers.@xmath271 in their experiment , the threshold field ( called here @xmath3 ) for transition from the mixed to the fully perpendicular state has no noticeable dependence on surface separation @xmath155 , in accord with our calculations as shown in fig
. 8 b and fig .
9 b. when the surface separation @xmath155 is not an integer or half integer number of the bulk lamellar period @xmath145 , the phase diagrams are given by fig
while in the @xmath52-@xmath216 plane the behavior is only slightly changed , in the @xmath52-@xmath155 plane the border of the parallel state has prominent oscillations with period @xmath145 .
finally , we have not considered the copolymer density undulations as was done by onuki and fukuda.@xmath259 these might change the boundary lines between the parallel , perpendicular and mixed phases . the main prediction of this paper is the dependence of the transition fields on the block copolymer film thickness @xmath155 and the phase diagrams of fig .
9 . it will also be of interest to check experimentally under what conditions the mixed state exists .
in addition , the energy penalty @xmath180 of the defect created in the film as well as the difference in surface energies of the two blocks , @xmath272 , can be deduced in experiment from the measured values of @xmath1 , @xmath3 and @xmath237 using eqs . [ 7:delta ] , [ 7:e1a ] and [ 7:e2a ] .
we would like to thank t. taniguchi and t. yukimura for critical remarks .
we benefitted from fruitful discussions with b. ashok , m. cloitre , j. derouchey , m. doi , l. leibler , m. muthukumar , r. rosensweig , t. p. russell , m. schick , f. tournilhac and t. thurn - albrecht .
partial support from the u.s .- israel binational foundation ( b.s.f . ) under grant no .
98 - 00429 and the israel science foundation founded by the israel academy of sciences and humanities centers of excellence program is gratefully acknowledged .
fink , y. ; winn , j. n. ; fan , s. ; chen , c. ; michel , j. ; joannopoulos , j. d. ; thomas , e. l. _ science _ * 1998 * , _ 282 _ , 1679 .
urbas , a. ; sharp , r. ; fink , y. ; thomas , e. l. ; xenidou , m. ; fetters , l. j. _ adv . matter .
_ * 2000 * , _ 12 _ , 812 .
de gennes , p. g. ; prost , j. _ the physics of liquid crystals _ ,
oxford university : new york , 1993 .
helfrich , w. _ appl .
lett . _ * 1970 * , _ 17 _ , 531 .
helfrich , w. _ j. chem . phys . _ * 1971 * , _ 55 _ , 839 .
hurault , j. p. _ j. chem .
phys . _ * 1973 * , _ 59 _ , 2068 . * * figure 1 : * schematic illustration of the system .
the two confining surfaces are at @xmath20 , and have surface interaction parameters @xmath39 .
the electric field points in the perpendicular @xmath51 direction , and is produced by a potential difference between the two surfaces ( electrodes ) .
[ 7:1 ] * * figure 2 : * weight functions @xmath56 and @xmath57 . the horizontal dash - dot line @xmath273 is the value of @xmath69 below which a nonzero @xmath57 is possible .
surface separation is @xmath274 and the flory parameter is @xmath159 .
the film is symmetric , @xmath275 .
[ 7:2 ] * * figure 3 : * ( a ) critical field @xmath0 [ in units of @xmath276 required for the appearance of perpendicular lamellae , as a function of surface separation @xmath155 .
( b ) log - log plot of the same curve as in part a. the dashed line is a straight line corresponding to @xmath277 ( see text ) .
( c ) the jumps @xmath169 ( dashed line ) and @xmath168 ( solid ) in the amplitudes @xmath70 and @xmath69 at the critical field @xmath0 , as a function of surface separation @xmath155 .
the film is symmetric , and @xmath39 and @xmath154 are as in fig . 2 . [ 7:3 ] * * figure 4 : * contour plots of the bcp order parameter @xmath278 for symmetric film .
the surfaces are at @xmath279 , and the field is in the @xmath51 direction . in part
a the field is a little smaller than the critical field , @xmath280 , and the film has a perfect parallel ordering . in part b the field is just above the threshold , @xmath281 .
the film morphology is a superposition of parallel and perpendicular lamellae .
( c ) a plot of the imds [ given by @xmath282 of part b. in part d @xmath177 , and the lamellae are in the perpendicular state with small distortions .
the surface fields are @xmath283 , and the flory parameter is @xmath159 corresponding to correlation length @xmath284 .
the b monomers ( colored black ) are attracted to the two symmetric surfaces .
[ 7:4 ] * * figure 5 : * illustration of the lamellae in the film in the strong segregation regime . a region with perpendicular lamellae exists in the middle of the film , with size @xmath181 .
two regions of size @xmath201 each are the parallel lamellar regions .
[ 7:5 ] * * figure 6 : * ( a ) qualitative dependence of @xmath180 and @xmath22 on the thickness of perpendicular domain @xmath285 , defined in fig . 5 .
( b ) simplified curves of part a. the cutoff width is @xmath201 , the total film thickness is @xmath155 , and @xmath286 and @xmath287 are the mean values of @xmath180 and @xmath288 , respectively , for @xmath289 .
[ 7:6 ] * * figure 7 : * sketch of the dependence of @xmath116 , @xmath117 and @xmath194 on the field @xmath52 , in solid line , circles and dashed line , respectively .
( a ) @xmath218 : there are two critical fields , obtained when @xmath290 and @xmath291 .
( b ) @xmath236 : @xmath292 is always larger than @xmath293 , and there is only one critical field obtained when @xmath294 . above this field the most stable state is the perpendicular ordering .
[ 7:7 ] * * figure 8 : * ( a ) phase diagram in the @xmath52-@xmath216 plane .
if @xmath295 , there is a transition between parallel and perpendicular lamellae at @xmath239 . for @xmath218 ,
there is a transition from the parallel to the mixed state at @xmath251 , followed by a second transition from the mixed to the perpendicular state when @xmath296 .
surface separation is chosen as @xmath297 and @xmath298 .
( b ) similar diagram , but in the @xmath52-@xmath155 plane , with @xmath299 . in both parts ,
the effect of mismatch between @xmath155 and @xmath145 is neglected , and the electric fields are scaled by @xmath300 .
[ 7:8 ] * * figure 9 : * phase diagrams as in fig . 8 , but for strained films .
the critical fields @xmath1 and @xmath237 have oscillations as a function of @xmath155 , in contrast to fig .
8 b , but follow on average a similar @xmath301 scaling behavior . | we examine the alignment of thin film diblock copolymers subject to a perpendicular electric field .
two regimes are considered separately : weak segregation and strong segregation . for weakly segregated blocks and below a critical value of the field , @xmath0 ,
surface interactions stabilize stacking of lamellae in a direction parallel to the surfaces . above the critical field ,
a first - order phase transition occurs when lamellae in a direction perpendicular to the confining surfaces ( and parallel to the field ) become stable .
the film morphology is then a superposition of parallel and perpendicular lamellae .
in contrast to helfrich - hurault instability for smectic liquid crystals , the mode that gets critical first has the natural lamellar periodicity .
in addition , undulations of adjacent inter - material dividing surfaces are out - of - phase with each other . for diblock copolymers in the strong segregation regime , we find two critical fields @xmath1 and @xmath2 . as the field
is increased from zero above @xmath1 , the region in the middle of the film develops an orientation perpendicular to the walls , while the surface regions still have parallel lamellae .
when the field is increased above @xmath3 the perpendicular alignment spans the whole film . in another range of parameters ,
the transition from parallel to perpendicular orientation is direct .
= 1truecm |
active galactic nuclei ( agn ) that exhibit h@xmath0o maser emission at 22 ghz have been objects of particular interest since discovery that the maser emission can trace circumnuclear accretion disks , as in ngc 4258 @xcite .
the presence of maser emission enables direct mapping of agn dynamics within 1 pc of supermassive black holes via very long baseline interferometry ( vlbi ) ( see reviews of * ? ? ?
* ; * ? ? ?
the additional discovery that h@xmath0o maser emission can be associated with shocked gas in jets and outflows has also provided a tracer for a second component in agn @xcite .
maser emission at 22 ghz is believed to be predominantly collisionally - pumped at gas kinetic temperatures of t@xmath5 = 400 1000 k , hydrogen densities of n(h@xmath0 ) = 10@xmath6
10@xmath7 @xmath8 , and dust temperatures t@xmath9 @xmath10 100 k ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in agn accretion disks ,
gas heating is thought to arise from x - ray irradiation of the disk surfaces , whereas masers in jets and outflows may be heated via shock compressions @xcite .
since the physical conditions that give rise to emission at 22 ghz also pump other h@xmath0o maser transitions , notably those at 183 , 321 and 325 ghz ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , 22 ghz maser emission is unlikely to be alone in agn disks , jets and outflows .
we note that , in galactic regions of star formation and evolved stars , inversion of the 22 ghz transition is accompanied by inversion of other rotational transitions from the ground and vibrationally excited h@xmath0o states ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in agn
, some of the masers should occur in broadly the same regions as at 22 ghz , but others could probe uncharted regions of the central engines of agn , including regions at smaller radii .
line ratios of two or more maser transitions originating from the same gas would constrain radiative transfer models far better than is now possible .
previous searches for extragalactic millimeter / submillimeter h@xmath0o masers have been performed , but have been hampered by r.m.s .
sensitivities worse than 1 jy ( menten , private communication ) .
ngc 3079 is an almost edge - on spiral seyfert 2/liner galaxy with optical heliocentric velocity of 1125 @xmath11 6 km s@xmath12 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , at a distance of 16 mpc @xcite .
it hosts 22 ghz h@xmath0o maser emission at velocities between @xmath1940 and 1350 km s@xmath12 , which are believed to trace a disk - like structure at radii of 0.4 1.3 pc from a @xmath13 @xmath14 central engine .
the pc - scale maser disk is aligned with a dense molecular kpc - scale disk @xcite and is located at the apex of a kpc - scale superbubble @xcite .
the bubble is believed to be inflated by a wide - angle outflow @xcite that is also responsible for the radio continuum lobes along the minor axis of the galaxy ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
within a parsec of the nucleus , the relativistic jet appears aligned with the inner wall of the bubble and may be responsible for weak 22 ghz maser emission away from the disk ( * ? ? ? * and references therein ) .
all of these emission sources lie within the primary beam of the sma observations reported here .
we observed the ground - state 3@xmath15 @xmath16 2@xmath17 transition of para - h@xmath0o ( e@xmath18/@xmath19 @xmath1200 k ) at a rest frequency of 183.310 ghz ( @xmath20 : 1.6 mm ) towards ngc 3079 using the sma on 2005 march 1 .
the compact configuration included six antennas , resulting in projected baselines of 8 72 m. the sma receivers operate in double - sideband mode with two 2 ghz sidebands that we tuned to 183.310 and 193.310 ghz ( band - center ) .
the correlator provided continuous coverage of each sideband ( in 24 separate chunks of 256 channels ) and 0.4 mhz channel spacing .
the phase center of the observations was @xmath21 , @xmath22 @xcite .
the zenith opacity measured with the nrao tipping radiometer located at the caltech submillimeter observatory was @xmath23(225 ghz)=0.03 - 0.05 throughout .
we calibrated the data within the idl superset mir developed for the owens valley radio observatory @xcite and adapted for the sma ; the imaging was performed in miriad .
data reduction took into account the effect of the pressure - broadened terrestrial h@xmath0o absorption line centered at 183.310 ghz .
fortunately , the redshift of the galaxy reduces the line frequency of the maser transition by @xmath10.7 ghz , greatly lowering the corresponding atmospheric opacity . for bandpass calibration we used jupiter and saturn .
phase calibration was performed via frequent observations of the quasar 0923 + 392 about 18@xmath24 from the phase center .
we performed the amplitude calibration separately for upper and lower sideband data .
for the upper sideband , we used 0923 + 392 ; for lower sideband data , we performed amplitude calibration on each of the 24 spectral chunks separately using saturn and jupiter .
the calibration of the lower sideband both as a function of elevation and frequency was designed to reduce the effect of variation in terrestrial h@xmath0o absorption across the sideband .
we derived the flux scale using observations of callisto with an estimated accuracy of 20% .
single sideband system temperatures corrected to the top of the atmosphere were between 400 and 850 k. we observed the ground - state 6@xmath25 @xmath16 5@xmath26 transition of ortho - h@xmath0o ( e@xmath18/@xmath19 @xmath11100 k ) at a rest frequency of 439.151 ghz ( @xmath20 : 0.7 mm ) towards ngc 3079 using the 15 meter jcmt on 2004 november 18 in @xmath27=0.041 0.045 weather .
the instantaneous bandwidth was 1840 mhz , corresponding to @xmath11200 km s@xmath12 , with a channel spacing of 1250 khz or 0.85 km s@xmath12 .
the integration time on source was 2 hrs in beam - switching mode with a throw of 60@xmath28 .
the average t@xmath29 during the observations was 2650 k. we reduced the data using the starlink spectral - line reduction software , specx .
we detected millimeter ( 193 ghz ) continuum emission with a peak signal to noise ratio ( s / n ) of @xmath110 ( fig .
[ f : continuum ] ) at @xmath30 , @xmath31 ( statistical uncertainties ) .
systematic uncertainties of ( @xmath32,@xmath33 ) were estimated through measurement of an apparent position for 1150 + 497 against the phase calibrator 0923 + 392 . these place
the peak of the millimeter continuum within 1@xmath2 of the 22 ghz maser position .
we partially resolved the nucleus in the north - south direction and the elongation is approximately parallel to the position angles of both the pc - scale accretion disk traced by 22 ghz maser emission and the kpc - scale molecular disk detected using e.g. , co @xcite .
we detected 183 ghz h@xmath0o emission with a peak s / n of @xmath17 in both the real part of the amplitude spectrum and the interferometer phase spectrum ( figs .
[ f : continuum ] and [ f : line ] ) .
the peak flux density is 0.55 jy at 1017 km s@xmath12 .
the line covers 9 frequency channels and overlaps a persistent doppler component appearing in published 22 ghz maser spectra for dates 1998 may 8 , 2000 march 18 where it is especially prominent at @xmath12.5 jy ( both * ? ? ?
* ) , 2001 march 23 @xcite and 2002 april 12 @xcite .
we note possible detection of additional spectral features at 1208 ( 5 channels wide ) and 1265 km s@xmath12 ( 4 channels wide ) .
both features appear in spectra of real amplitude and phase ( near zero degrees , which indicates positions similar to the 1017 km s@xmath12 emission ) , and both lie within the velocity interval of 22 ghz emission .
note that the correspondence of features between 22 and 183 ghz observations is not precise and is also not expected to be true for all features under certain pumping scenarios .
the peak maser emission is at @xmath34 , @xmath35 , which is within 1@xmath2 of the 193 ghz continuum peak . at our resolution
, we can not determine whether 183 ghz emission originates from the disk or outflow , or from both .
we note that in addition to tracing a disk , a subset of 22 ghz maser emission is associated with the outflow within a few parsecs of the nucleus @xcite .
22 ghz emission at around 1000 and 1200 km s@xmath12 is present both in the disk and outflow .
possible angular extension of the 183 ghz emission to the southeast is marginally detected ( 2@xmath2 ) , which lies along the axis of the relativistic flow ( e.g. , * ? ? ?
we note that @xcite detect oh absorption associated with the outflow at an overlapping velocity , and that we can not rule out shocks in circumnuclear material as an origin for 183 ghz emission .
we obtained a tentative detection ( 5@xmath2 ) of the 439 ghz maser at 1157 km s@xmath12 ( fig .
[ f:439ghz ] ) . with peak s@xmath36 @xmath12.5 jy , 439
ghz emission is of comparable strength to typical values for peak 22 ghz emission towards ngc 3079 and occurs within the 22 ghz velocity interval .
the millimeter continuum of ngc 3079 is probably dominated by thermal emission from optically thin dust . as in the 1.2 mm observations of @xcite at 11@xmath28 resolution , we find dust concentrated in the nuclear region of ngc 3079 ( fig .
[ f : continuum ] ) and place an upper limit of 8 mjy ( 3@xmath2 ) on emission associated with the radio lobes @xmath11 kpc from the emission peak .
@xcite estimates the dust temperature to be @xmath130 k at kpc radii
. however , in two other agn the presence of hotter dust at 100 times smaller radii , of @xmath1300 k , has been inferred from vlt observations @xcite . adopting a dust temperature of 100 k for ngc 3079 on intermediate scales and a dust opacity index @xmath37
, we estimate the enclosed dust mass of the inner 200 pc is @xmath38 @xmath1 10@xmath39 m@xmath40 using the method described in @xcite . using co nobeyama observations by @xcite , who find a gas mass of @xmath41 @xmath110@xmath42 m@xmath40 within the central 150 pc , we estimate a gas - to - dust ratio of @xmath43 , which is comparable to the galactic value of 160 and to the ratios derived in bright local galaxies @xcite .
the 183 ghz emission we detect in the nucleus of ngc 3079 is likely from a maser process .
it arises from one , and possibly more , narrow unresolved features in the same range of velocity as that of known 22 ghz maser emission . from radiative transfer models
we know that the same high densities and temperature conditions that produce maser emission at 22 ghz also strongly invert the 183 ghz h@xmath0o transition . noting that the 22 ghz feature at 1017 km s@xmath12 varies between 0.4 2.6 jy ( hagiwara et al .
2002 ) , and assuming an ortho : para h@xmath0o ratio of 3:1 , pumping models could reproduce likely line ratios ( e.g. , * ? ? ?
narrow linewidths ( @xmath108 km s@xmath12 ) at 183 ghz , small compared with the characteristic rotational and apparent random velocities for molecular gas in this nucleus @xcite , also indicate that emitting regions should be compact .
our beam size is too large to determine whether emission is thermal or maser based on brightness temperature arguments . for the emission to be thermal , with a brightness temperature of @xmath44 2000 k such that h@xmath0o molecules are not dissociated
, the emitting region must be @xmath45 0@xmath461 ( 10 pc ) .
439 ghz emission is probably also due to maser amplification .
emission occurs over a narrow velocity range ( fig .
[ f:439ghz ] ) , and conditions for strong 439 ghz inversion are likely to exist in the galactic nucleus .
the 22 ghz spectrum peaks to the blue of the v@xmath29 of 1125 km s@xmath12 , yet the 439 ghz spectrum peaks slightly above v@xmath29 at 1157 km s@xmath12 ( fig .
[ f:439ghz ] ) .
we ascribe this difference to the different pump mechanisms operating for each maser . whereas 22 ghz emission is collisionally - pumped and is increasingly quenched by local dust temperatures @xmath45100 k , the 439 ghz maser is predominantly radiatively - pumped and is most strongly inverted in the presence of @xmath47 , @xmath48 and n@xmath49 = 5 x 10@xmath50 @xmath8 @xcite .
dust temperatures of this order and higher are reasonable in the central parsecs of the nucleus @xcite .
peak 439 ghz emission ( 1157 km s@xmath12 ) is therefore likely to originate from a hot , dusty region around the central engine , possibly at radii @xmath511 pc , separated spatially from the gas emitting most strongly at 22 ghz ( 956 km s@xmath12 ) . at 22 ghz , features at these velocities are indeed separated spatially by @xmath11 pc projected @xcite .
the observations in fig .
[ f:439ghz ] were taken on different dates , and we note that the 22 ghz feature near 1157 km s@xmath12 is not present at all epochs ( e.g. , in fig .
[ f : line ] ) . however , the feature line ratio in fig .
[ f:439ghz ] of @xmath1100 is accommodated by the vicinity of hot dust @xcite .
we thank the all the sma and jcmt staff .
we thank s. paine for opacity data , j. braatz for 22 ghz data and g. melnick for useful discussions .
b. acknowledges support from the emmy - noether program of the deutsche forschungsgemeinschaft ( dfg - be2578/1 ) . | we report the first detection of an extragalactic millimeter wavelength h@xmath0o maser at 183 ghz towards ngc 3079 using the submillimeter array ( sma ) , and a tentative submillimeter wave detection of the 439 ghz maser towards the same source using the james clerk maxwell telescope ( jcmt ) .
these h@xmath0o transitions are known to exhibit maser emission in star - forming regions and evolved stars .
ngc 3079 is a well - studied nuclear h@xmath0o maser source at 22 ghz with a time - variable peak flux density in the range 3 12 jy .
the 183 ghz h@xmath0o maser emission , with peak flux density @xmath10.5 jy ( 7@xmath2 detection ) , also originates from the nuclear region of ngc 3079 and is spatially coincident with the dust continuum peak at 193 ghz ( 53 mjy integrated ) .
peak emission at both 183 and 439 ghz occurs in the same range of velocity as that covered by the 22 ghz spectrum .
we estimate the gas to dust ratio of the nucleus of ngc 3079 to be @xmath3150 , comparable to the galactic value of 160 .
discovery of maser emission in an active galactic nucleus beyond the long - known 22 ghz transition opens the possibility of future position - resolved radiative transfer modeling of accretion disks and outflows @xmath4 pc from massive black holes . |
gamma - ray bursts ( grbs ) are one of the most energetic explosion in the universe , but there are still many issues to be understood such as gamma - ray emission mechanism of prompt emission , physical composition of jet outflow , and environment of progenitor .
grbs are also known to be originated at cosmological distance and they would be useful to explore the chemical evolution of high - z universe .
_ astro - h _ is the sixth x - ray observatory from japan , which is scheduled to be launched by the end of japanese fiscal year 2015 .
four onboard instruments of _ astro - h _ , the high - resolution x - ray micro - calorimeter ( soft x - ray spectrometer : sxs ) , x - ray ccd camera ( soft x - ray imager : sxi ) , hard x - ray imager ( hxi ) , and soft gamma - ray detector ( sgd ) realize wide - band and high - sensitivity observation from 0.3 to 600 kev energy band .
the high - resolution spectroscopy by sxs and x - ray observations with enough photon statistics by sxi could be very powerful tool to investigate spectral features and detail of x - ray absorption structure in the afterglow spectrum of grbs . and also
, high - sensitive hard x - ray observation by hxi might observe interesting features from afterglow in hard x - rays .
in addition to such afterglow observations by focal plane instruments , _ astro - h _ is also able to observe the prompt gamma - ray emission utilizing sgd .
therefore , _ astro - h _ will bring us a comprehensive observation of grbs from prompt gamma - ray emission to subsequent x - ray afterglow emission . in this paper
, we demonstrate a capability of grb observation by _ astro - h_.
our understanding of gamma - ray emission mechanism of grbs , especially for short duration grbs is still poor .
one of key observation to solve such problem is time - resolved spectroscopy as was performed for long duration grbs . however , photon statistics of short grbs is too low to perform such time - resolved analysis , and therefore , observation of short grbs with large effective area is important . _
astro - h _ has capability to observe prompt gamma - ray emission of short grbs with large effective area and good time - resolution utilizing sgd .
the main detector , compton camera of the sgd is surrounded by large 25 bgos to reduce background by anti - coincidence technique as shown in fig [ fig : sgdpic ] .
thanks to its large geometrical area and high gamma - ray stopping power of bgo crystal , the effective area of those shield detectors retain @xmath0 800 @xmath1 even at 1 mev .
therefore , the sgd shield detector acts as a powerful all - sky monitor like suzaku wam@xcite .
we have developed the sgd shield detector so that we can observe short transients such as short grbs or soft gamma repeaters with many advantages compared with suzaku wam .
table [ tab : wamspec ] shows some specifications of the sgd shield detector as an all - sky monitor comparing with suzaku wam .
the main advantage of the sgd shield detector is that it can obtain spectral information with very large effective area .
we also improved data acquisition timing of grb data of the sgd shield so that we can transfer grb data to the spacecraft soon ( @xmath010 min ) after trigger and we can set the trigger to be ready for the next grb .
this enable us to improve the efficiency of grb observation .
.performance of the sgd shield detector as all - sky monitor comparing with suzaku - wam [ cols="^,^,^",options="header " , ] [ tab : wamspec ] @xmath2 by the sgd shield .
each hatched region show the time window for the demonstration of time - resolved spectral analysis in the below figure.,width=321 ] figure [ fig : lcsim ] shows an example of simulated light curve of bright short grb with peak flux of about a few times of 10 photons s@xmath3 @xmath2 in 1 second time scale . in this simulation
, we consider poisson fluctuation in each time bin of the sgd shield ( 16 ms ) and we assumed simple band function with low - energy index @xmath4 and high - energy index @xmath5 .
the peak energy e@xmath6 has changed depending on the flux .
figure [ fig : specsim ] shows the time - resolved spectrum extracted with 0.1 s time windows , which are shown by hatched area in the figure [ fig : lcsim ] .
we can see that the simulated light curve exhibit fine time structure and extracted time - resolved spectra show clear evolution of e@xmath6 .
therefore , we can expect to have such grb data with the sgd shield .
after launch of _ astro - h _ , grb data observed by the sgd shield will be publicly available as well as suzaku wam .
the grb observation by the sgd shield can provide complementary dataset to fermi - gbm . based on simultaneously detection rate between suzaku - wam and fermi - gbm , about a half of grbs detected by fermi - gbm
are expected to be also detected by the sgd shield .
as for the x - ray afterglow of grbs , which is widely believed that the x - ray emission is coming from synchrotron emission due to accelerated electrons in the external forward shock . therefore , most of x - ray spectrum of afterglow show featureless simple power - law shape .
however , there are several reports of marginal detection of spectral features such as iron - k emission line , its recombination edge , and several lines due to light metals @xcite,@xcite,@xcite .
although , they are still controversial probably because of limited statistics and/or spectral resolution , such spectral features would be very important to investigate physical conditions of grb jet and composition of environment of grb host , and also they are useful to determine the redshift of grb by x - ray observation itself .
in addition to such spectral features , behar et al .
( 2011 ) and starling et al .
( 2013 ) have pointed out the evidence of excess absorption in soft x - ray energy band using huge sample of swift x - ray afterglow observation .
one possibility of origin for such excess absorption is contribution of absorption by intergalactic medium ( igm ) . therefore , detail spectroscopy in soft x - ray band could give important information to investigate the property of igm in high - z universe .
an x - ray observation with high spectral resolution is a key to solve above open questions in grb afterglow .
those emission line and/or absorption line spectral features can be investigated by unprecedented high energy resolution spectroscopy by _ astro - h _ sxs , and detail of continuum structure can be determined by sxi .
figure [ fig : afterglowsim ] shows a 100 ks _ astro - h _ simulation with sxs and sxi .
here we assumed grb 991216 spectrum as the baseline model fro simulation . in this model , 2@xmath710 kev flux
is set to be @xmath8 erg @xmath2 s@xmath3 .
this grb has been reported to have iron - k line and its recombination edge@xcite and thus we include those spectral features in the simulation .
we also added soft x - ray lines reported by reeves et al .
( 2003 ) for grb 011211 , and intergalactic warm absorbers ( whim ) with the temperature of 10@xmath9 k , the column density of n@xmath10 of 10@xmath11 @xmath2 , and we put those absorption material on redshift of z=0.1 . from this simulation , we can see that the iron - k related spectral features can be detected clearly by _ astro - h _ if they are really exist .
in addition , some resonance absorption lines due to whim are also detectable with about 4 sigma significance level , thanks to high energy resolution of sxs .
figure [ fig : felinesim ] shows the same simulation with figure [ fig : afterglowsim ] but with shorter exposure of 10 ks and we changed intrinsic line width from 5 ev to 30 ev .
we can clearly detect the iron - k line emission if it is intrinsically narrow with @xmath12 ev with short exposure of 10 ks .
this indicates that we can investigate the time variability of such iron - k line emission , which is useful to discuss the environment of host galaxy o grbs .
as we shown in previous section , _ astro - h _ has a capability of detection of spectral features from grb afterglow such as iron - k emission line and resonance absorption lines due to intergalactic warm absorbers .
then , we have to estimate how many number of grbs we can observe with such interesting spectral features by _
astro - h_. for this purpose , we calculated a luminosity function of grb afterglow based on 572 samples of 6-years swift - xrt data base which is publicly available in the web page @xcite .
figure [ fig : grblumin ] shows the luminosity functions of grb afterglow for several times after grb trigger . from this result , we can see that about 10 grbs / year are expected which have 10@xmath13 erg s@xmath3 @xmath2 flux level , which corresponds to that of we used in the iron - k line simulation in fig [ fig : felinesim ] , even 30 hours after the trigger .
this means that if we can slew the _ astro - h _ spacecraft within 1-day after the trigger , we could have 10 grbs / year samples for possible iron - k line search with _ astro - h_.
in this paper , we demonstrated the capability of grb observation by _ astro - h_. as for the prompt gamma - ray emission , the sgd shield detector will act as powerful grb monitor with very large effective area .
especially for the short grbs , time - resolved spectroscopy with good photon statistics can be performed by the sgd shield and such grb data can be a complimentary data set to fermi - gbm . about a half of grbs that are detected by fermi - gbm
are also expected to be observed by the the sgd shield simultaneously .
high resolution spectroscopy by _ astro - h _ sxs and sxi is expected to reveal the existence of spectral features in the grb afterglow spectrum such as emission lines from iron - k and/or other light metals , and absorption by intergalactic medium .
the expected event rate of grbs which can be used for such search of spectral features is estimated to be @xmath0 10 grbs / year , if we can slew the spacecraft within 1-day after the grb trigger .
more details about _ astro - h _ observation of grb afterglow can be found in the _ astro - h _ white paper@xcite . | _ astro - h _ , the sixth japanese x - ray observatory , which is scheduled to be launched by the end of japanese fiscal year 2015 has a capability to observe the prompt emission from gamma - ray bursts ( grbs ) utilizing bgo active shields for the soft gamma - ray detector ( sgd ) .
the effective area of the sgd shield detectors is very large and its data acquisition system is optimized for short transients such as short grbs .
thus , we expect to perform more detailed time - resolved spectral analysis with a combination of _ astro - h _ and fermi lat / gbm to investigate the gamma - ray emission mechanism of short grbs .
in addition , the environment of the grb progenitor should be a remarkable objective from the point of view of the chemical evolution of high - z universe .
if we can maneuver the spacecraft to the grbs , we can perform a high - resolution spectroscopy of the x - ray afterglow of grbs utilizing the onboard micro calorimeter and x - ray ccd camera . |
in the recent past eremin and voloshin@xcite proposed that both nucleons and individual nuclei could be considered as superposition of the constituents ( of the particles ) called partons / quarks / valons etc .
normally it is assumed that the nucleons are built of three such constituent partons and the mesons are composed of two of them .
such a constituent picture helps us to interpret the deep inelastic collision processes involving lepton - hadron , hadron - hadron and hadron - nucleus collisions .
much earlier , bialas et al studied to explore the hypothesis , within the framework of the additive quark model(aqm)@xcite , that the constituent partons(quarks ) provide the universal elements not only of nucleon - nucleon and nucleon - nucleus reactions but also nucleus- nucleus interactions at high energies ; this study ended up with surely an affirmative indication to such a possibility.but , as the focus was on different issues , the central theme of this work deviated from what it is intended to be here.we will simply look into the behaviour of some chosen observables studied here for nucleus - nucleus collisions alone from the viewpoint of such partonic constituent pictures .
the centrality - dependence of multiplicity - density offers a very fundamental observable and we present here our studies on production of pions , kaons , protons and antiprotons separately in both @xmath0(including deuteron production for this collision ) at cern - sps and @xmath1 collisions at rhic - bnl at @xmath2 and 200 gev respectively . besides
, very recently the data on deuteron production in @xmath0 collision have been obtained , for which they have also been taken here into account .
this apart , we will also compare the performance of the present approach with those obtained on the basis of some standard - version programmings like hijing , venus etc which are essentially much more complicated than the one dealt with here .
our objective here is to examine the efficacy of this approach in explaining some relevant data on production of these very important kind of secondaries in the few now available high energy nuclear collisions .
the approach was also successfully tested for production of neutral particles , like photons , by netrakanti and mohanty@xcite in the very recent past and was found to be in accord with the contention of eremin and voloshin@xcite .
in course of our calculations for the number of participant nucleons , denoted by @xmath3 , and the number of participant partons , denoted by @xmath4 we follow exactly the ways adopted by eremin and voloshin@xcite and netrakanti and mohanty@xcite , wherein a woods - saxon nuclear density profile is taken into account .
this is given by @xmath5}}\ ] ] where @xmath6 , @xmath7 , @xmath8 .
the number of participant nucleons(@xmath3 ) for a nucleus - nucleus(@xmath9 ) collision at an impact parameter @xmath10 is calculated using the expression@xcite , @xmath11^b\}\\ & + & ~ \int ~ d^2s ~ t_b(\vec{s}-\vec{b } ) ~ \{~ 1 ~ -[~ 1 ~ - ~ \frac{\sigma_{nn}^{inel } t_a(\vec{s})}{a } ~ ] ^a\ } , \end{array}\ ] ] where @xmath12 and @xmath13 are the mass numbers of the two colliding nuclei , @xmath14 is the thickness function , and @xmath15 is the inelastic nucleon - nucleon cross section .
the number of participant partons is also calculated in a similar manner by taking into account the following changes of the facts related to physical realities , viz . , ( i ) the density is changed to three times that of nucleon density with @xmath16 ; ( ii ) the cross sections are changed to @xmath17 and ( iii ) the mass numbers of the colliding nuclei are changed to three times of their values keeping the size of the nuclei same as in the cases of @xmath3@xcite . in the same way , the number of participant partons in a @xmath18 or @xmath19 collision can also be calculated by using @xmath20 and @xmath21 , and considering nucleons as hard spheres of uniform radii 0.8 @xmath22@xcite . for the present work we use
@xmath23 30 mb in the range of c.m .
energies @xmath24 4 - 17.3 gev as suggested in ref.@xcite . for higher energies
, we first fit the data on total cross - sections for @xmath18 and @xmath19 collisions@xcite at energies beyond @xmath25 gev with the expression@xcite @xmath26 + ( 1.1\pm0.1)\ln{(\frac{s_{nn}}{10 ~ gev^2 } ) } + ( 0.19\pm0.01)\ln^2{(\frac{s_{nn}}{10 ~ gev^2})}\ ] ] with @xmath27 , where @xmath28 no .
of data @xmath29 no . of parameters .
we also fit the elastic cross - section data for the same interactions@xcite with a similar type of expression which is given by , @xmath30- ( 0.47\pm0.04)\ln{(\frac{s_{nn}}{10 ~ gev^2 } ) } + ( 0.11\pm0.01)\ln^2{(\frac{s_{nn}}{10 ~ gev^2})}\ ] ] with @xmath31 . the fits along with the data are shown in fig.1 .
hence , one can obtain the energy - dependence of inelastic cross - section(@xmath15 ) for nucleon - nucleon interaction by subtracting eqn.(4 ) from eqn.(3 ) .
the values of @xmath15 , obtained in this manner , at some energies are given in table-[tab : table1 ] .
the obtained values fall within @xmath32 error of those suggested in ref.@xcite . but a point is to be noted . though some sort of justification for proposing the nature of the total cross section in the form of expression ( 3 ) given in ref.@xcite , none such explanation for expression ( 4 ) is possible .
this is simply assumed with a view to providing only a best fit to the available data on elastic cross section shown in fig.1 and extracting a workable phenomenological form of expression for the nature of inelastic cross section , @xmath15 .
in fact , these twin relations help us to arrive at the usable values of inelastic cross sections needed for our necessary calculations of the relevant observables in a systematic manner even at energies for which no measured data on cross section values are available .
the values of the number of participant nucleons and those of the participant partons for three nucleus - nucleus collisions at various energies and different centralities are presented in the tables([tab : table2]- [ tab : table4 ] ) .
a comparison of the results obtained on the basis of the present work with those of na49@xcite , wa98@xcite and phenix@xcite groups is made in fig.-2(a ) to fig.-2(d ) , in terms of @xmath33 versus @xmath34 .
it is observed that the agreements in cases of au+au collisions are modestly fair , whereas for pb+pb interaction the disagreement is quite strong and prominent .
this discrepancy could , for the present , be attributed to the much less c.m energy in lead - lead reaction which falls roughly at 17.2 gev only .
the integrated yields of the identified hadrons at central rapidity region produced in @xmath0 and @xmath1 collisions at different centralities are presented diagrammatically in fig.3-fig.5 . in subfigures labelled as ( a ) and ( b ) of each category ,
the experimental data on integrated yields are normalized by half of the number of participant nucleons(@xmath3 ) while those in subfigures labelled as ( c ) and ( d ) are normalized by half of the number of participant partons(@xmath4 ) .
as we are here to make comparison between these two cases , we should use values of both the observables obtained from a single model . and
that is why we put into use the values of @xmath35 obtained on the basis of the present work , instead of the values indicated by na49@xcite or phenix groups@xcite .
now , to make a comparison between these two cases we fit the data in fig.3-fig.5 by two phenomenological expressions which are given by , @xmath36 and @xmath37 where @xmath38 , @xmath10 , @xmath39 and @xmath40 are four constants .
the fitted values of these parameters are given in tables([tab : table5]-[tab : table7 ] ) and depicted by solid curves in the figures .
as could be seen from fig.2 that the most peripheral values of @xmath3 for @xmath0 collision at @xmath41 gev and @xmath1 collision at @xmath42 gev obtained by the present model show larger discrepancies with respect to those obtained by na49 and phenix collaborations , we keep most peripheral data in these two collisions out of the range of both the fits provided by eqn.(5 ) and eqn.(6 ) . obviously , the factors @xmath39 and @xmath40 provide the slopes of the fits .
as the magnitude of any of these factors takes value nearer to zero , the fit will exhibit better flatness of the data with respect to @xmath3 or @xmath4 implying a superior compliance of scaling , i.e. , a better scaling .
a look at table-[tab : table5 ] reveals that the data on @xmath43 , @xmath44 and @xmath45 in @xmath0 collisions show a better degree of conformity with scaling when normalized by half of @xmath4 , as in all cases @xmath46 is less than @xmath47 .
but , for the case of deuteron production in @xmath0 collision , the scenario is just the opposite . besides
, the integrated yields for charged pions produced in @xmath1 interactions at both the rhic energies[tables [ tab : table6 ] - [ tab : table7 ] ] do not exhibit scaling satisfactorily , when the data are normalized by half of @xmath4 instead of @xmath3 , as the values of @xmath47 , in these cases , are quite small with respect to @xmath46 .
however , the data on @xmath44 and @xmath45 produced in the same collisions favour the @xmath4-scaling over the @xmath3-scaling , though the cases of @xmath45 at @xmath42 gev are not as prominent as in the other cases . in fine , our net finding from the present study is the pions do not agree , while the kaons and protons do .
the fig.6(a ) deserves some special attention and comments .
the plots on integrated yields for charged hadrons produced in both nucleus - nucleus(@xmath48 ) and @xmath49 collision deflect from each other when data on both are separately normalized in terms of number of participating nucleons(@xmath3 ) .
the data on integrated yields in central nucleus - nucleus collisions normalized by the number of participant nucleon - pair have been used here from fig.3 of ref.@xcite . to our purpose , this has to be normalized in terms of the basic parton(quark)-constituents , denoted by @xmath50 , for which the ( charged)pseudo- rapidity density terms for various nucleus - nucleus collisions normalized by participant nucleon - pair , i.e. , the factors @xmath51 are to be multiplied by @xmath52 with the values of them as given in table-[tab : table8 ] . and ,
in this conversion we have utilized those particular values of @xmath35 as were used in ref.@xcite , so that we can normalize the exact values of @xmath53 by @xmath54 . in calculating @xmath4 for @xmath49 collisions ,
we use equation(2 ) and the values of @xmath4 for @xmath49 collisions at three different c.m .
energies are given in table-[tab : table9 ] as a function of relative probability .
and when partonic considerations are used in normalization , the data on both @xmath49 and @xmath48 collisions come to an agreeable state .
the @xmath50-values that are to be used for normalization of the most central @xmath49 collisions , i.e.,@xmath55 central collisions are depicted in table-[tab : table10 ] . in order to check the nature of agreement between the data on @xmath48 and @xmath49 collisions
, we try to obtain a fit by taking into account both sets of data , normalized by @xmath54 ; and the desired fit is to be described here by the following expression : @xmath56 with @xmath57 . the goodness of the fit is also shown in fig.6(b ) with the help of the expression @xmath58 as a function of @xmath59 which reveals that both sets of data are in good agreement with respect to the fit . hence ,
both @xmath48 and @xmath49 data exhibit a common @xmath59-dependence when the data on the former are normalized by the number of participant parton - pairs .
very recently , sarkisyan and shakarov@xcite adopted a somewhat similar type approach in dealing with data on pseudorapidity density per participant - pair in cases of hadron - hadron , nucleus - nucleus collisions and also in the cases of @xmath60 interactions .
the study thus reinforced the idea of similarities of the bulk observables in the widest possible ranges of particle - interactions .
let us now summarize our observations made here : ( i ) the secondaries , excluding deuterons , produced in @xmath0 collision at cern - sps behave modestly well vis - a - vis @xmath4 scaling .
the behaviour is more consistent towards the highest values of centrality , with centrality maximum at 0 - 5@xmath61(represented by solid boxes in the graph ) .
( ii ) similar statements could be made about @xmath1 interactions at both energies(fig.4 and fig.5 ) except the cases for pions .
( iii ) it is quite noticeable that there are modest degree of divergences in the @xmath3-values between the calculations done by us at different centralities and those by others , as indicated in the text ( in the second and third columns of the various tables presented in this work ) . for the present
, these discrepancies could be attributed only to our much simpler programming than what is resorted to by the big groups like na49 , wa98 , phenix etc .
the plain fact is we are now simply unable to take up such rigorous studies as might be desired or advisable due to various reasons beyond our capacity and control .
but , we recognize the urgency and importance of such studies in order to arrive at a decision about the merit of @xmath4 scaling and of the viewpoints expressed by eremin and voloshin@xcite .
( iv ) the fig.6 highlights on how the data on both @xmath49 and @xmath48 collisions negotiate better the idea of partonic - participant scaling , when the constituent - participant number for both @xmath49 and @xmath48 collisions are taken into account .
thus , the present study does essentially provide modest support to the viewpoints expressed by eremin and voloshin in their work of ref.@xcite .
the authors are very grateful to the anonymous referee for his / her very patient readings , constructive criticisms and kind valuable instructions for the improvements in the previous drafts of the manuscript .
the authors would also like to express their thankful gratitude to professor d. mi@xmath62kowiec for providing some helpful suggestions about running the fortran code developed by him in calculating the number of participants , @xmath63 .
one of the authors , bd , is thankful to the imsc for offering him the support by a post - doctoral fellowsip of the imsc , wherein a part of this work was done .
* s. eremin and s. voloshin : phys . rev . * c 67*(2003 ) 064905 .
a. bialas , w. czyz and l. lesniak : phys . rev . * d 25*(1982)2328 . p. k. netrakanti and b. mohanty : phys . rev .
* c 70*(2004 ) 027901 ; nucl - ex/0401036 . c. y. wong : introduction to high - energy heavy ion collisions , pg.161 , world scientific , 1994 .
d. mi@xmath62kowiec : http://www-linux.gsi.de/@xmath64misko/overlap/manual.ps .
particle data group : http://pdg.lbl.gov/2004/hadronic-xsections/hadron.html b. de , s. bhattacharyya and p. guptaroy : jour .
* g 27*(2001)2389 .
j. bachler et al(na49 collaboration ) : nucl .
phys . * a 661*(1999)45c
. m. m. aggarwal et al(wa98 collaboration ) : eur .
* c 23*(2002)225 .
k. adcox et al(phenix collaboration ) : phys . rev . * c 69*(2004)024904 .
k. adcox et al(phenix collaboration ) : phys .
. s. adler et al(phenix collaboration ) : phys .
* c 69*(2004)034909 .
b. b. back et al(phobos collaboration ) : phys .
lett . * 88*(2002 ) 022302 and the references therein .
e. k. g. sarkisyan and a. s. sakharov : hep - ph/0410324 v3 .
j. klay : ph .
d. thesis , university of california , davis dissertation , 2001 , pg.127 ; http://nuclear.ucdavis.edu/@xmath64jklay/cv/[email protected] .
s. v. afanasiev et al(na49 collaboration ) : phys .
* c 66*(2002)054902 .
b. b. back et al(phobos collaboration ) : phys .
85*(2000 ) 3100 .
t. anticic et al(na49 collaboration ) : phys .
rev . * c 69*(2004)024902 .
.[tab : table1]values of inelasitc nucleon - nucleon cross - section at different energies .
the second column provides the values obtained by subtracting eqn.(4 ) from eqn.(3 ) .
the last column gives the values as cited in ref.@xcite .
[ cols="^,^,^ " , ] and @xmath19 interactions as a function of c.m .
data points are taken from ref.@xcite .
the solid curve depicts the fit for total cross - sections on the basis of eqn.(3 ) while the dashed one is for elastic cross - section on the basis of eqn.(4).,width=302 ] | the present study is both a reanalysis and an extension of the approach initiated by eremin and voloshin(referenced in the text ) .
we attempt to interpret here the rapidity - spectra of the various particles produced in both @xmath0 and @xmath1 collisions at cern - sps and rhic - bnl .
the study made here is wider in scope and is more species - specific than the earlier ones for which the results obtained here have been compared with those suggested by some previous works based on hijing , venus etc , at various centralities .
the study reconfirms that the constituent parton picture of the particles provides a better and more unified description of the rapidity - density yields for the various secondaries , including some light cluster particles like deuteron even in heavy ion collisions . |
single - layer ( sl ) transition metal dichalcogenides are currently receiving considerable attention mainly due to their electronic similarity to graphene , but with the important distinction of having a seizeable band gap .
they constitute a promising materials platform for exploring spin and valley degrees of freedom @xcite and the consequences of strongly bound excitons in two - dimensions @xcite .
particular focus has been on mos@xmath0 @xcite , a material for which sls can be readily obtained by a micromechanical exfoliation method similar to that used for the initial isolation of graphene .
sl mos@xmath0 has been used in applications such as field effect transistors @xcite and optoelectronic devices @xcite . moreover ,
mos@xmath0 nanoclusters had already been grown in sls and used in catalysis even before the advent of graphene @xcite . the key difference between the electronic structure of sl mos@xmath0 and its three - dimensional ( 3d ) parent compound is the nature of the band gap .
3d mos@xmath0 in the stable 2h - structure has an indirect band gap with the valence band maximum ( vbm ) at the @xmath4 point of the brillouin zone ( bz ) and the conduction band minimum ( cbm ) between the @xmath4 and @xmath5 points .
sl mos@xmath0 , on the other hand , has both the vbm and cbm at the @xmath1 point of the two - dimensional ( 2d ) bz ( corresponding to the @xmath5@xmath6 line in the bulk bz ) @xcite .
the band character change between bulk and sl happens mainly because the original vbm and cbm are shifted to lower and higher energy , respectively .
in contrast , the states around @xmath1 ( or @xmath5@xmath6 ) are not significantly affected by the structural change and turn into the new vbm and cbm of the sl .
this band gap change can be understood qualitatively by considering the bulk band structure of the material : the bands forming the vbm and cbm of bulk mos@xmath0 show dispersion in the direction perpendicular to the mos@xmath0 sheets of the crystal and are thus subject to quantum confinement for thin crystals ( to see this , compare the dispersion of the bands in the @xmath4@xmath5 and @xmath7@xmath6 directions in e.g. ref .
the states forming the vbm and cbm of the sl , on the other hand , show hardly any dispersion in the @xmath5@xmath6 direction and are therefore unaffected by quantum confinement effects .
these differences are inherently related to the orbital character of the different states , where higher dispersion in the @xmath5@xmath6 direction indicates a larger contribution from out - of - plane orbitals . in this simple picture , placing a sl of mos@xmath0 onto a substrate can be expected to have a severe impact on the band structure since it strongly modifies the boundary condition for quantum confinement on one side of the sl . moreover , placing the sl on a metallic substrate has the potential to alter the band structure and band gap of the sl due to increased screening from the substrate @xcite . in this article
, we explore the band structure and band gap of epitaxial sl mos@xmath0 grown on au(111 ) .
such growth results in the large - area and high - quality sl materials necessary for applications in devices . starting from nanoscale clusters of mos@xmath0 on au(111 )
@xcite , we have recently developed a procedure to nearly saturate almost the entire surface of au(111 ) by sl mos@xmath0 @xcite and we have presented initial experimental results of the system s electronic structure determined by angle - resolved photoemission spectroscopy ( arpes ) in ref .
overall , the ( occupied ) electronic structure of the supported sl closely resembles the theoretical predictions for free - standing sl mos@xmath0 .
details such as the size of the spin - orbit splitting at the @xmath1 point of the bz are in near quantitative agreement with the predictions @xcite .
however , some notable differences with respect to free - standing sl mos@xmath0 were observed , suggesting a significant mos@xmath0-substrate interaction . in the experimental part of this paper
, we give a complete description of the band structure of mos@xmath0 on au(111 ) as determined by arpes and provide additional data on the band gap size by scanning tunnelling microscopy and spectroscopy ( stm and sts ) . in the theoretical part ,
we compare these results to first principles calculations .
we clarify the nature of the mos@xmath0-substrate interaction , establish a theoretical understanding of the transition metal dichalcogenide - metal interaction and explain the deviations between the arpes results and the band structure of a free - standing layer .
epitaxial sl mos@xmath0 has been grown on au(111 ) by methods described elsewhere @xcite .
the total mos@xmath0 coverage for most of the arpes data shown here and in ref.@xcite was kept below one monolayer ( ml ) at @xmath8 ml in order to avoid the growth of 2 ml islands .
the epitaxial sl mos@xmath0 samples are stable in air and could thus be removed from the dedicated growth chamber , transported to the sgm-3 end station of the synchrotron radiation source astrid2 @xcite and cleaned via mild annealing to 500 k. this procedure has been verified to yield atomically clean surfaces by stm @xcite .
arpes data were collected at 80 k with an energy and angular resolution better than 20mev and 0.2@xmath9 , respectively .
the measurements presented here were performed at two different photon energies , 49ev and 70ev .
the dispersion of the highest valence band of mos@xmath0 is most intense in the 49ev data while the 70ev data is dominated by the au bulk bands .
although not presented in this article , photon energy scans were performed to confirm the lack of @xmath10 dispersion of the sl mos@xmath0 bands to distinguish them from the underlying dispersing au bulk bands .
stm and sts measurements were carried out using a commercial low - temperature system ( createc fischer & co. gmbh @xcite ) , operating at a base temperature of 6k .
an etched w tip was used for all measurements , which was flashed _ in - situ _ and further conditioned by dipping it into the au substrate prior to the measurements .
sts was performed using lock - in detection applied to the sample with a typical frequency range of 741900hz . 1 unit cell of mos@xmath0 on a 1@xmath111 cell of au(111 ) .
the different positions considered of the bottom s atoms with respect to the underlying au lattice ( on - top , hcp , and fcc ) are shown .
b ) top view of the mismatched model consisting of a @xmath12 r13.9@xmath9 cell of mos@xmath0 on a 4@xmath114 cell of au(111 ) .
side view of the mos@xmath0/au(111 ) interface indicating the relevant distance @xmath13 between the lower s layer and the outermost surface layer of au(111 ) .
turquoise , yellow , and brown spheres indicate the position of mo , s , and au atoms , respectively.,title="fig : " ] + periodic first principles calculations based on density functional theory ( dft ) were performed for the different sl mos@xmath0 on au(111 ) models shown in fig.[fig:1 ] : the matched " models and the mismatched " model .
the simple matched models ( fig.[fig:1](a ) ) consist of a ( 1@xmath111 ) cell of mos@xmath0(0001 ) on au(111 ) , with different relative positions of the mos@xmath0 layer with respect to the substrate ( the s atoms in the lower layer are located in either the on - top , fcc , or hcp sites of the au(111 ) surface ) . in order to achieve matching of the mos@xmath0 and au(111 ) lattices , the lattice parameter of the au substrate ( modelled by 10 atomic layers ) was expanded by 9@xmath14 to fit the theoretical value for mos@xmath0 ( 3.18 ) .
such an expansion results in an artificial destabilization of the au surface and to shifts of au surface and bulk bands to higher energies .
although such artifacts indicate that this simple model is inadequate , especially for describing the substrate , it is a useful tool to directly study the mos@xmath0-au interaction and its effect on the band structure of the sl mos@xmath0 .
we also performed calculations with the more complex and representative mismatched model shown in fig.[fig:1](b ) .
the lattice mismatch between sl mos@xmath0 and the underlying au gives rise to a moir pattern @xcite .
the resulting moir structure consists of a ( 10@xmath1110 ) supercell of mos@xmath0 on a ( 9@xmath119 ) supercell of au(111 ) .
however , since the model required for such structure is computationally unfeasible , our mismatched model consists of a ( @xmath12 r13.9@xmath9 ) supercell of mos@xmath0 supported on a ( 4@xmath114 ) supercell of au(111 ) ( modelled by 4 atomic layers ) .
this model only requires a 0.15% contraction of the au lattice and was suggested by farmanbar _ _
et al . _
_ @xcite in order to avoid the artificial distortion of the work function of the au surface .
although it is slightly smaller than the experimentally observed supercell , it shares most of the structural features observed in experiments , including the different s - au contact regions - i.e. , on - top , hcp , and fcc , within the unit cell .
all the periodic electronic structure calculations presented here were carried out using the vasp code @xcite .
the valence electrons were described with plane - wave basis sets with a kinetic energy threshold of 415ev , and the interaction between the valence and frozen core - electrons was accounted for by means of the projector - augmented - wave ( paw ) method of blchl @xcite .
the pbe approximation to the exchange - correlation functional was used @xcite . for the matched models
, the reciprocal space was sampled with a ( 20@xmath1120@xmath111 ) mesh of @xmath15-points , whereas for the mismatched model , the geometry of the supercell was optimized using a ( 4@xmath114@xmath111 ) mesh , and the charge density was subsequently recalculated with a single point calculation using a denser ( 10@xmath1110@xmath111 ) mesh of @xmath15-points .
an energy threshold of @xmath16ev was used to define self - consistency of the electron density .
all atomic positions for the matched models were relaxed until the forces on all atoms were smaller than 0.01ev@xmath17 , whereas for the mismatched only the atoms corresponding to the mos@xmath0 layer were relaxed .
the band structure of the matched models can be directly compared to the measured dispersion , but for the larger mismatched supercell , the band structure along the high symmetry directions of the sl mos@xmath0 primitive cell is folded " into the smaller reciprocal lattice of the supercell . in order to unfold " the band structure with the symmetry of the primitive unit cell of sl mos@xmath0
, we have calculated the effective band structure proposed by popescu and zunger @xcite as implemented in the bandup code @xcite .
spin - orbit coupling has been included for all band structure calculations .
on au(111 ) along different high - symmetry directions of the bz , measured at a photon energy of 70ev ; inset : hexagonal bz .
( b ) data acquired at a photon energy of 49ev in the energy region of the top valence band of mos@xmath0 .
( c ) the same data as in ( b ) but with the theoretical dispersion for a free - standing layer ( solid yellow lines ) and projected band gap edges of the ( 111 ) surface of au ( solid orange lines ) @xcite superimposed on the experimental data .
the measured difference in the top of the valance - band at @xmath1 compared to @xmath2 is noted to be 0.31ev.,title="fig : " ] + figure [ fig:2 ] shows the photoemission intensity of epitaxial sl mos@xmath0 on au(111 ) along different high symmetry directions in the 2d bz , measured at two different photon energies , 70ev and 49ev ( for constant binding energy cuts through the bz for a photon energy of 49ev see fig .
1(a ) of ref . @xcite . )
the au substrate contributes with the @xmath18 valence band , the surface state in the l - gap of the bulk bz ( see supplementary material of ref .
@xcite ) , and the deeper lying @xmath19-bands visible in the lower part of fig . [ fig:2](a ) . projecting the bulk band structure of au onto the ( 111 ) surface
gives rise to several projected band gaps @xcite .
wide projected gaps are found near the fermi energy at the @xmath2 , @xmath1 and @xmath3 points .
these gaps appear as a reduction of background intensity and are outlined by the solid orange lines in fig .
[ fig:2](c ) . the upper valence band of sl mos@xmath0 is clearly identified in the measurements acquired with 49ev photons in the energy range between ( [email protected])ev ( top of the valence band at @xmath1 ) and ( [email protected])ev ( bottom of the valence band between @xmath2 and @xmath3 ) . for comparison ,
the calculated upper valence band for free - standing sl mos@xmath0 is superimposed on the photoemission data as yellow lines in fig . [ fig:2](c ) .
the overall agreement between experiment and this calculation is excellent . in particular , the dispersion close to @xmath1 and the size of the spin - splitting there is reproduced almost quantitatively .
but there are also a few points of disagreement .
the first is the relative energy of the local valence band maxima at @xmath2 and @xmath1 . in the calculation for
the free - standing layer , these states are almost degenerate whereas the state at @xmath2 is measured to be at significantly higher binding energy ( ( [email protected])ev ) than the state at @xmath1 for sl mos@xmath0 on au(111 ) . following the arguments given in the introduction , it is reasonable to assume that this ( 0.31ev ) change is due to the boundary condition given by the mos@xmath0-au(111 ) interface and we shall see below that this is indeed the case .
note that such a distortion of the uppermost valence band does not take place for exfoliated sl mos@xmath0 deposited on sio@xmath0 substrates @xcite , presumably due to the weaker bonding of the sulphur atoms to the substrate .
a second point of disagreement between the measured dispersion and the calculation for the free - standing layer is the dispersion of the bands around @xmath3 . in the calculation
, the spin - splitting of the bands reduces to zero at @xmath3 .
this is expected at @xmath3 ( and at @xmath2 ) due to the combination of time - reversal and crystal symmetry .
@xmath3 is a so - called time - reversal invariant momentum in the 2d bz @xcite , meaning that two @xmath3 points can be connected both by an inversion of the @xmath21-direction and by a reciprocal lattice vector .
the combination of these symmetries enforces a spin - degeneracy in the 2d electronic states .
this degeneracy is observed at @xmath2 but the situation is somewhat unclear at @xmath3 .
we shall return to these two points of disagreement below when we discuss the calculations for sl mos@xmath0 adsorbed on au(111 ) .
the arpes data only gives very limited information on the character and size of the sl mos@xmath0 band gap since only states below the fermi energy are accessible . as a consequence , we can only conclude that the band gap must be larger than @xmath221.3ev , i.e. corresponding to the energy of the vbm .
we have previously attempted to determine this band gap by alkali atom doping @xcite and by time- and angle - resolved photoemission spectroscopy , i.e. by pumping electrons into the conduction band and subsequent emission by a second photon @xcite .
adsorbing potassium atoms onto the surface does indeed lead to a sufficiently strong charge donation to push the cbm below the fermi level , resulting in the observation of a direct band gap at @xmath1 with a size of ( [email protected])ev .
however , it is questionable if this measured gap corresponds to the gap of the pristine system as potassium adsorption also leads to a distortion of the occupied bands in which states with out - of - plane orbital character ( e.g. the upper valence band at @xmath2 ) are shifted more than states with in - plane orbital character ( e.g. the vbm at @xmath1 ) .
indeed , it has been found that sodium intercalation in bulk mos@xmath0 does not only lead to electron donation but to a change of the band gap as such @xcite and can even trigger a transition from the semiconducting 2h - phase to the metallic 1t - phase @xcite .
the band gap determined by pump / probe arpes , which does not require doping with alkali metals , is ( [email protected])ev @xcite .
it is important to note that this electronic band gap is a fundamentally different quantity from the optical gap , which is generally measured by absorption and photoluminescence experiments @xcite . in simple terms
, the optical band gap corresponds to the energy required to create an exciton while the electronic band gap also requires the breaking of the exciton and is thus higher due to the exciton binding energy .
while the electronic band gap for free - standing sl mos@xmath0 is not known , an optical band gap of @xmath23ev has been determined for free - standing sl mos@xmath0 @xcite as well as for sl mos@xmath0 on graphite @xcite , for which weaker interactions than for au(111 ) are expected .
this optical gap measurement is in excellent agreement with a quasi - particle calculation of the optical gap of a free - standing layer @xcite , which in addition predicts an electronic gap of 2.8ev . on au(111 ) .
image parameters : v@xmath24 = @xmath251.5v , i@xmath26 = 0.2na .
( b ) atomically resolved image of a sl mos@xmath27 island revealing the atomic lattice and moir pattern .
image parameters : v@xmath24 = @xmath250.12v , i@xmath26 = 0.1na .
( c ) representative sts ( current and differential conductance ( @xmath28 ) ) near the center of a sl mos@xmath27 island .
the tip was stabilized at v@xmath24 = 1.9v , i@xmath26 = 0.4na .
a modulation amplitude of 42mv was used.,title="fig : " ] + another reliable approach for determining the electronic band gap of a material is sts .
we therefore performed sts measurements on numerous mos@xmath0 islands to simultaneously determine the electronic structure in occupied and unoccupied regimes as well as its spatial dependence .
figure [ fig:3 ] shows stm and sts data from the sample . acquiring position - dependent sts spectra on the islands ( which correspond to the brighter patches with the distinct moir pattern in fig .
[ fig:3](a ) ) , we observe small spatial variations within an energy range of 200 mev .
however , we found no statistical correlation between variations in intensity or peak location and the spatial location within the unit cell of the moir pattern . therefore , we only considered spatially averaged spectra taken near the center of mos@xmath0 islands and away from domain boundaries ( edges ) for determining the energy gap .
sts acquired in a @xmath202ev energy window ( fig .
[ fig:3](c ) ) shows a regime of very low conductance around the fermi energy , i.e. between @xmath251ev and 0.5ev , with onsets of the conductance at higher absolute energies .
the sts spectra greatly differ from those acquired on the clean au(111 ) surface areas which only show the typical surface state onset below the fermi energy .
sts spectra of the sl mos@xmath0 islands show clear conductance features resulting from the emergence of sl mos@xmath0 bands at different energies these are indicated by arrows in fig .
[ fig:3](c ) .
the sharp rise in conductance is due to the onsets of the sl mos@xmath0 valence and conduction bands outside of the mos@xmath0 band gap . in similar sts experiments involving sl mose@xmath0 on graphene ,
features in the @xmath28 versus @xmath29 plot are harder to discern but it is possible to determine the exact position of the band edges by plotting @xmath30 versus @xmath29 instead @xcite .
in the present case , the metallic character of the support and its strong hybridization with sl mos@xmath0 lead to a net ( albeit low ) conduction in the band gap and to smoother conductance variations , which hinders the possibility of assigning small increases to the band edges . we therefore approximate onsets to the peaks in the sts spectra . in the valence band regime
, we can directly compare the resonances at vb1=@xmath31ev and vb2=@xmath32ev to the top of the valence band at @xmath1 and @xmath2 in the arpes data shown in fig .
[ fig:2](c ) .
the energy difference between vb2 and vb1 is 0.36ev and in agreement with the @xmath220.31ev determined from arpes data for the same sample . in the conduction band ,
we see three onsets ( cb1 - 3 ) , specifically a step centered at @xmath33ev ( cb1 ) and two resonances at @xmath34ev ( cb2 ) and @xmath35ev ( cb3 ) . correlating cb1 and vb1 to the @xmath1 band onsets yields a band gap value of ( [email protected])ev , where the larger error stems from the uncertainty in defining the position of the cb1 step . while the gap agrees reasonably well with the pump / probe arpes results of ref .
@xcite , sts measurements on similar samples , but carried out at room temperature , have been reported to yield a smaller band gap of 1.3ev @xcite .
this difference can most likely be ascribed to the different experimental temperature ( 6k here vs room temperature in ref . )
that affects the size of the gap as such ( the spectra differ mainly in the position of vb1 ) but also has a significant impact on stability and energy resolution . the smaller band gap ( 1.74ev ) measured here for mos@xmath0 on au in comparison to the 2.15ev measured for mos@xmath0 on graphite using sts @xcite supports the interpretation of a stronger band structure renormalization induced by the au(111 ) support @xcite . layer in the matched ( on - top ) and mismatched mos@xmath0/au(111 ) models and for the free - standing monolayer .
( b ) pdos on mos@xmath0 for the on - top , hcp , and fcc positions for the matched model .
( c ) dos projected on the different atom types of the mismatched model , i.e. au , mo , and s atoms from the upper and lower layers .
the pdos of the s atoms has been scaled @xmath113 for clarity .
peaks in ( c ) have been labeled according to the assignments made in fig .
[ fig:3 ] .
the onsets of vb1 and cb1 are -1.05ev and 0.54ev , respectively .
the peaks of vb2 , cb2 , and cb3 are -1.83ev , 1.15ev , and 1.43ev , respectively.,title="fig : " ] + the experimental results on the geometry and electronic structure are now compared to first principles calculations based on dft .
we have optimized the geometry for the different models of a sl mos@xmath0 on au(111 ) in fig .
[ fig:1 ] ( three matched models for adsorption in different sites and a mismatched model ) .
a structural optimization of the matched models results in the on - top position model being the most stable one by 199mev and 215mev per unit cell , compared to the hcp and fcc models , respectively .
this higher stability is accompanied by a smaller optimized distance between the lower s layer and the outermost au layer ( distance @xmath13 in fig [ fig:1](c ) ) for the on - top structure ( 2.51 ) , as for the fcc and hcp structures ( 3.20 and 3.03 , respectively ) .
interestingly , such height differences are in close to quantitative agreement with those measured in experimental stm images of sl mos@xmath0 on au(111)@xcite .
however , the intuitive assignment made in ref .
@xcite of the different areas within the moir pattern , where topographically higher regions are assigned to on - top bonding modes , may be called into question in view of the present results , where the on - top configuration leads to the smallest @xmath13 of the matched models . for the mismatched model
the optimized @xmath13 is 3.29 , which is larger than for the matched models , probably as a result of s atoms being , on average , in less favorable positions .
[ fig:4](a ) show the density of states projected onto the mos@xmath0 layer ( pdos ) for a free - standing layer , a matched layer in the optimum position ( on - top ) and the mismatched layer , whereas fig .
[ fig:4](b ) compares the pdos for the different matched models . despite of the structural differences ,
the overall shape of the pdos for the different systems is very similar and only the relative position with respect to the fermi level differs .
the different strength of the sl mos@xmath0 interaction with the au surface leads to binding energies ( calculated as the difference in energy between the mos@xmath0-au system and the separated mos@xmath0 and au ) of @xmath250.25 , @xmath250.03 , and @xmath250.04 ev for the on - top , fcc , and hcp configurations , respectively . in a generic semiconductor ,
the position of the fermi level , or rather the chemical potential , is dictated by band structure , doping and temperature . in the present situation
, the interface to the underlying au plays an important role for the position of the mos@xmath0 bands . in all cases studied here ,
the interaction between mos@xmath0 and au(111 ) gives rise to a @xmath36-type contact with the fermi level pinned near the conduction band of mos@xmath0 , which is in agreement with similar studies @xcite .
however , the resulting energy of the vbm with respect to the fermi level ( e@xmath37 ) differs for the cases considered in this work and for the different models used in the literature .
such shifts of the band alignment can be caused by the different work functions of the underlying metal substrates @xcite .
work function differences for the same kind of metal surfaces can be induced by strain and we do in fact find marked differences between the position of the vbm calculated for matched and mismatched models : for the matched model used here and in ref .
@xcite , which involves a 9 % au(111 ) expansion , the cbm of mos@xmath0 is almost touching the fermi level . on the other hand , for our mismatched model and the ones used in refs .
@xcite , which involve a compressive strain of the au(111 ) lattice , the cbm is found at higher energies with respect to the e@xmath37 ( [email protected] ) .
since the mismatched model involves only very little strain of the au lattice ( 0.15% ) , it should give a more realistic description .
in fact , this small strain was one of the reasons why the mismatched model was originally suggested in ref . .
the pdos results for the different matched models in fig .
[ fig:4](b ) indicate that the fermi level pinning not only depends on the work function of the au(111 ) surface , but also , and even more strongly , on the interaction with the mos@xmath0 layer .
this interaction was reported to lead to the formation of interface dipoles due to charge redistribution and to the formation of gap states with mo @xmath19-orbital character @xcite . here
, we also observe such charge redistribution , which is strongly localized at the interface and extends to the s - mo - s layer but not to the vacuum region ( not shown ) . for the on - top matched model the stronger interaction with the surface shifts the mos@xmath0 states towards lower energies than for the other sites .
this is due to the more favorable s - au bonds , which involve stronger charge transfer and redistribution at the interface . for the matched models with the hcp and fcc positions , the interaction and the resulting pdos shifts are weaker but they are similar to each other , consistent with a similar s - au bonding geometry .
the s - au bonding for the mismatched model , on the other hand , is on average a mixture of top , hcp , and fcc bonding sites . as a result ,
the position of the vbm with respect to the e@xmath37 for this system is very similar to the hcp and fcc cases despite the differences in work function of the underlying au substrate models .
the small strain of the underlying au(111 ) and the different s - au bonding regions make the mismatched model more representative of the experimental situation .
its dos is thus more suitable for comparison to the onset values with respect to the fermi level determined by sts and arpes .
we have therefore labeled the bands in the atomically decomposed pdos spectra of the mismatched model according to the labels used for the sts spectra ( fig .
[ fig:4](c ) ) .
the vb1 and cb1 have contributions mainly from the mo atoms , whereas s atoms also contribute to the vb2 , cb2 , and cb3 .
in addition , vb2 overlaps with the @xmath19-band of au , which , as we show in the calculated band structures below , leads to a strong hybridization between s and au states .
the onset and peak values of the different bands agree quite well with the sts values extracted from fig .
[ fig:3](a ) ; the calculated ( experimental ) onsets of vb1 and cb1 are -1.05ev and 0.56ev ( -1.24ev and 0.50ev ) , respectively , whereas the peaks of vb2 , cb2 , and cb3 are at -1.83ev , 1.15ev , and 1.43ev ( -1.60ev , 0.99ev , and 1.63ev ) , respectively .
the over- and underestimation of the position of vb1 and cb2 , respectively , result in a band gap ( 1.61ev ) that is underestimated by 0.13ev compared with the value determined by sts ( ( [email protected])ev ) .
the position and character of the bands can be seen more clearly by analyzing the calculated band structures for free - standing sl mos@xmath0 and the sl adsorbed on au(111 ) shown in fig .
[ fig:5 ] ( first and second panels ) .
the band structure for the sl on au(111 ) corresponds to the matched model in the most stable on - top binding geometry ( fig .
[ fig:1 ] ) .
the calculated band structure for the free - standing layer is in excellent agreement with previous results using the same exchange - correlation functional @xcite , and the resulting band gap ( 1.58ev ) is also very similar .
this value significantly underestimates the electronic gap of 2.8ev determined by more sophisticated quasi - particle ( gw ) calculations @xcite .
interestingly , calculations with the local density approximation ( lda)@xcite or with another exchange - correlation functional ( pw91 ) within the generalized gradient approximation ( gga ) lead to slightly ( @xmath39ev ) larger values @xcite than the 1.58 ev calculated here .
band gap underestimation is a well - known shortcoming of the lda and gga approaches , which can be partially remedied by including a fraction of hartree - fock exchange , leading to the so - called hybrid functional formalism .
indeed , calculations with such functionals yield improved band gaps of 2.05ev @xcite .
nevertheless , studying the mos@xmath0/au(111 ) with standard dft methods provides valuable insight into the effect of the substrate on the band structure despite the gap underestimation .
+ for the on - top matched model , the adsorption of the mos@xmath0 layer on au(111 ) leads to pronounced changes in its band structure ( fig .
[ fig:5 ] , second panel ) but , surprisingly , the fundamental gap near @xmath1 ( 1.58ev ) is only 0.02ev lower than for the free - standing layer .
moreover , the bands forming the vbm and cbm of the layer remain well - defined , something that is ascribed to their greater contribution from mo orbitals ( which do not directly bond to the au substrate ) , to the in - plane nature of such states for the vbm and , most importantly , to their position in a projected gap of the au(111 ) band structure .
the binding energies of the occupied bands for the matched model are generally in good agreement with the experimental values from arpes , and the position in the projected band gap is consistent with the vbm states being very sharp in arpes .
however , the exact energy of the bands for the matched model is affected by the artificial strain induced in the au(111 ) slab , as described above , and should thus not be directly compared to the arpes values
. the unchanged size of the band gap with respect to the free - standing sl mos@xmath0 , on the other hand , is a common feature of all models , and the same was found for other mismatched models with compressed au(111 ) surfaces when studied with similar levels of theory @xcite .
in contrast to the situation around the fundamental gap , the upper valence band near @xmath2 is strongly affected by the adsorption on au(111 ) . the band is still well - defined as it enters the bulk continuum and its spin - splitting is lost . however , very close to @xmath2 it strongly merges with the bulk bands .
the band maximum at @xmath2 is very diffuse but significantly lower ( by @xmath220.4ev ) than at the @xmath1 point .
these observations are in excellent agreement with the experiment where a similar energy shift and a strong broadening , indicative of mos@xmath0-substrate interaction , is observed .
in fact , this band structure change is consistent with the simple expectation that the adsorbate - substrate interaction should mainly affect the boundary condition for the quantization of the out - of - plane states in the sl .
note that the mos@xmath0 character is retained for higher energy states near @xmath2 and the resulting dispersion of these states is very similar to the arpes results in fig . [
fig:2](a ) . the band structure can be further explored by decomposing the states of the sl mos@xmath0/au(111 ) model into the contribution from the mo atoms and from the lower ( in contact with au ) and upper ( in contact with the vacuum ) layer of s atoms .
[ fig:5 ] shows such weighted band structures , separated into out - of - plane orbitals and in - plane orbitals .
the vbm of the mos@xmath0 layer is mostly formed from in - plane mo 4@xmath19 states at the @xmath1 point ( see also refs .
@xcite ) which , together with their position in the projected band gap , explains why it remains relatively unaffected by the interaction with the au(111 ) surface .
the cbm at the @xmath1 point , on the other hand , is formed by out - of - plane mo 4@xmath19 states , which are expected to be distorted by out - of plane interactions . considering this and the low gap measured by sts ,
a gap variation upon interaction with the au substrate is expected .
nevertheless , the size of the fundamental gap at @xmath1 calculated here is unaffected by the interaction with the metallic substrate .
this points to the inefficiency of the semi - local lda or gga approaches in capturing subtle effects of out - of - plane interactions in band gaps involving correlated states .
these methods suffer from an inherent self - interaction error and only partially account for electronic correlation , neglecting long - range exchange . in contrast , gw calculations account for long - range coulomb interaction and are able to predict subtle screening effects . for sl mose@xmath0 ( which has a very similar electronic structure to mos@xmath0 )
a small direct gap reduction is calculated upon interaction with a support @xcite , and lower @xmath1-@xmath1 gaps are predicted when increasing the number of layers or decreasing the interlayer distance in mos@xmath40@xcite .
in contrast to the situation at @xmath1 , the vbm at @xmath2 contains significant contributions from the sulphur @xmath41 states ( as well as some contribution for out - of - plane mo 4@xmath42 and 4@xmath43 orbitals ) and here we observe a strong asymmetry for the top and bottom sulphur atoms . while the top - atoms still participate in the formation of the upper valence band states at @xmath2 , the contribution of the @xmath41 orbitals of the bottom sulphur atoms is completely hybridized and merged with the au(111 ) states . as one might have expected , this is an indication of a strong au - s bonding in the system .
this strong bond will presumably limit the possibilities to modify the electronic properties of epitaxial mos@xmath0 on au(111 ) compared to the case of graphene on transition metal surfaces , where a large number of atomic species can be intercalated between the graphene and the surface in order to change the graphene s electronic properties @xcite .
the strong interaction with the bulk state continuum can also explain the failure to observe the expected spin - degenerate upper valence band at @xmath3 . as can be seen in fig .
[ fig:5](second panel ) , the upper valence band mixes so strongly with the au states that it does not exist as a well - defined state at @xmath3 .
this is in contrast to exfoliated sl mos@xmath0 deposited on sio@xmath0 where the upper valence band can be observed in the entire 2d bz , albeit without a well - resolved spin - orbit splitting @xcite . 4 au(111 ) surface and ( b ) the mismatched model consisting of a @xmath12 r13.9@xmath9 cell of mos@xmath0 on a 4@xmath114 cell of au(111 ) .
the position of relevant mos@xmath0 bands ( vb1 , vb2 , and cb1 ) has been indicated according to the labels in figs . [ fig:3](a ) and [ fig:4](c).,title="fig : " ] + let us now compare these results to the band structure of the mismatched model . as mentioned in the introduction , the band structure of such a supercell is folded into its smaller bz . in order to recover the primitive cell picture of the band structure of au - supported mos@xmath0
, we have calculated the so - called effective band structure ( ebs ) for this system . to help differentiate between mos@xmath0 and au bands , fig .
[ fig:6](a ) shows the band structure for the au(111 ) slab ( without the mos@xmath0 layer ) unfolded on the primitive cell of mos@xmath0(0001).the intensity in the ebs plots for an energy interval @xmath44 and a @xmath15-vector of the primitive cell depends on the number of states of the supercell that have the same character as the primitive cell k - vector in that energy interval , i.e. that are related by the unfolding operator @xcite .
unfolding the band structure of the au(111 ) supercell on the primitive cell of mos@xmath0(0001 ) results in an ebs with several bands , which shows that despite the mismatch , there are states of the au(111 ) slab that have a similar character as the primitive cell of mos@xmath0(0001 )
. in particular , the @xmath19-band continuum of au appears diffuse below @xmath22 - 2.0ev and several bands cross the fermi level , some of which have minima at @xmath2 .
this suggests that one of these bands - probably the one with the minima closest to the fermi level - corresponds to the surface state of au(111 ) .
however , the energy of the surface state of au(111 ) is not well - reproduced and this is due to the model used . for this mismatched model , the au(111 ) surface
is represented by just four atomic layers whose structure is fixed to their truncated bulk positions . using a much thicker au(111 ) slab model and relaxing its structure would correctly reproduce the surface state , but such calculations would be too computationally demanding .
in addition , it should be noted again that the unfolding is done on the unit cell of mos@xmath0(0001 ) and this means that the unfolded band structure does not exactly correspond to that of the au(111 ) primitive cell .
the ebs for the mos@xmath0/au(111 ) mismatched model in fig .
[ fig:6](b ) shares most of the vbm features of the band structure of the on - top matched model .
the position of relevant mos@xmath0 bands has been indicated in fig [ fig:6](b ) by using the same labels ( vb1 , vb2 , and cb1 ) as in in figs .
[ fig:3](a ) and [ fig:4](c ) , to help differentiating between au and mos@xmath0 states .
the characteristic spin - split ( by 151mev ) bands at @xmath1 ( vb1 ) are clearly visible , whereas bands with lower energy ( vb2 ) merge with the au @xmath19-band , becoming more diffuse closer to @xmath2 and practically disappear .
the same states found crossing the fermi level in the ebs of the bare au(111 ) can also be recognized in the presence of the mos@xmath0 layer , but shifted towards higher energies .
this is not surprising given the n - doping of the mos@xmath0 layer , which leads to a concomitant p - doping of au .
interestingly , the interaction of vb2 with one of the au bands not only shifts the latter towards higher energies but also leads to avoided crossings between these bands near @xmath2 .
in addition , the mixing of vb2 with au states at @xmath2 destroys the local maximum of vb2 at @xmath2 .
this indicates that the interaction with the au substrate leads to vb2 having two local maxima near @xmath2 instead of just the maximum right at @xmath2 characteristic of free - standing sl mos@xmath0 .
this is consistent with the arpes spectra , where the vbm is hardly discernible at @xmath2 .
furthermore , the band gap at @xmath1 from the ebs is 1.61ev , which is very similar to those calculated for the free - standing monolayer ( 1.60ev ) or for the on - top matched mos@xmath0/au(111 ) model ( 1.58ev ) .
it is therefore clear that the choice of model does not affect the fundamental band gap at @xmath1 .
in summary , arpes and sts and first principles calculations have been used to study the electronic structure of epitaxial single - layer mos@xmath0 on au(111 ) .
we find that the interaction with a metallic support strongly distorts the vbm at @xmath2 and reduces the fundamental band gap at @xmath1 with respect to a free - standing layer .
the resulting band gap is measured by sts to be ( @xmath45)ev , which is in relatively good agreement with the ( @xmath46)ev determined by time - resolved arpes @xcite .
our dft calculations yield a smaller band gap of 1.61 ev and reveal that the mo 4@xmath19 states forming the vbm at @xmath1 of the single layer do not hybridize with the au(111 ) substrate .
this is due to the position of the mos@xmath0 vbm which does not overlap with the au bands , and the central position of the mo atoms in the mos@xmath0 layer . in turn
, the formation of au - s bonds leads to a strong hybridization of the s 3@xmath47 orbitals with the @xmath19 band of au , which causes a distortion of the mos@xmath0 vbm at @xmath2 .
we gratefully acknowledge financial support from the villum foundation , the danish council for independent research , natural sciences under the sapere aude program ( grant nos . dff-4090 - 00125 , dff-4002 - 00029 , and 0602 - 02566b ) , the lundbeck foundation , the innovation fund denmark ( cat - c ) and haldor topse a / s .
ab acknowledges support from the european research council under the european union s seventh framework programme ( fp/2007 - 2013 ) / marie curie actions / grant no .
626764 ( nano - design ) .
nh , dw , and aak acknowledge financial support from the alexander von humboldt foundation , the emmy noether program ( kh324/1 - 1 ) via the deutsche forschungsgemeinschaft , and from fom which is part of nwo , and the nwo vidi program .
the authors are grateful to dr .
paulo v. c. medeiros for the open - source distribution of the bandup code and for the useful technical support . | the electronic structure of epitaxial single - layer mos@xmath0 on au(111 ) is investigated by angle - resolved photoemission spectroscopy , scanning tunnelling spectroscopy , and first principles calculations . while the band dispersion of the supported single - layer is close to a free - standing layer in the vicinity of the valence band maximum at @xmath1 and the calculated electronic band gap on au(111 )
is similar to that calculated for the free - standing layer , significant modifications to the band structure are observed at other points of the two - dimensional brillouin zone : at @xmath2 , the valence band maximum has a significantly higher binding energy than in the free mos@xmath0 layer and the expected spin - degeneracy of the uppermost valence band at the @xmath3 point can not be observed .
these band structure changes are reproduced by the calculations and can be explained by the detailed interaction of the out - of - plane mos@xmath0 orbitals with the substrate . |
ion beams have many application in research and industry . being able to increase the delivered intensity to the application and reducing size and cost are important aspect of developing new accelerators and enabling new application for particle accelerators . among the most ambitious applications for high - intensity beams
are driving fusion reactions for the purpose of electricity generation .
there are roughly two approaches to the challenge of fusion energy production : magnetic confinement ( mfe ) and inertial confinement ( ife ) . in both approaches ,
high - intensity ion beams have been proposed as methods to heat the fusion fuel . in magnetic confinement , typified by tokamaks and other magnetic confinement schemes , a plasma of , for example , deuterium and tritium is heated to a temperature of several where the fusion cross - sections are high enough for the generation of energy from the reaction products exceeds the energy required to sustain the plasma .
the plasma and fusion reactions are sustained for a long time ( minutes to hours ) .
various methods are being developed to heat the plasma , including the deposition of energy in the plasma from high current deuteron beams .
for iter , two deuterium beam systems are being constructed with ion kinetic energy of .
each beamline occupies a volume exceeding @xmath0{m^3}$]@xcite .
the idea of using accelerators for inertial fusion came in the 1970 s shortly after rapid progress in laser technology and the recognition that lasers might drive inertial fusion targets@xcite . in laser and ion beam driven targets ,
the fuel is a spherical shell a few millimeters in diameter .
ion beam energy of the order of have to be delivered in @xmath1{s}$ ] to achieve ignition , and simulations indicate that 100@xmath2 more energy may be created from the fusion reactions@xcite .
in contrast to the magnetic fusion approach , the process here is pulsed , with a repetition rate of several hertz .
the ion beam requirements are constrained by the target design , where for heavy ions ( a>100 ) a kinetic energy of several and kiloamperes of current are required . for lighter ions and for a given total beam energy ,
the current must be increased .
there has been interest in magneto - inertial fusion ( mif ) , where aspects of magnetic and inertial fusion approaches are merged .
the initially low density plasma is confined by a magnetic field .
the plasma and embedded magnetic field are compressed by , e.g. , a metal liner , directed plasma or beams , with a confinement time longer than characteristic of ife , and shorter than mfe . for the ife and mif fusion energy applications ,
the final current density must be @xmath3 and the total energy per pulse needs to be @xmath4{mj}$ ] .
the accelerators generating the fusion driver beams are usually the most costly aspects of the fusion energy power plant system .
most ion accelerators have average acceleration gradients of over their length . comprising focusing magnets and acceleration cells ( rf or induction accelerators )
, they have transverse dimensions approximately , not including power supplies , vacuum pumps and other ancillary equipment . for non - fusion applications ,
the beam requirements tend to be less demanding but state of the art accelerator architecture in these applications have the same footprint and cost .
this work is motivated by advances in microelectromechanical systems ( mems ) techniques , in particular the potential for low - cost fabrication combined with excellent dimensional tolerances . in order to generate high power beams
, we propose to accelerate many parallel beams in a common accelerator structure and vacuum system , closely nested toward a lower - cost and reliable accelerator architecture for the applications above .
the idea of a scaled down accelerator was presented by maschke @xcite .
he proposed a meqalac structure consisting of electrostatic quadrupoles and rf - acceleration .
the current limit for an electrostatic quadrupole focusing lattice is given by @xmath5 where @xmath6 is the beam aperture , @xmath7 the maximum electric field , and @xmath8 the ion beam velocity@xcite . the electric field is bound by the breakdown field of the quadrupole structure .
measured breakdown fields stay constant or increase when scaling down the gap distance @xcite .
therefore the current density of an accelerator structure can be increased by scaling down the size of the structures and the total current can be increased by using multiple beamlets .
although magnetic quadrupoles offer better focusing than electrostatic ones at higher beam velocities , they do not scale well to smaller sizes , i.e. the current density needed to create the magnetic field increases when scaling down the quadrupole size . in practice
the smallest useful size for the quadrupole structures will be limited by alignment and fabrication errors of the structure .
mems technology pushes the boundaries of these fabrication errors into the micrometer range and therefore allows drastic improvements in transportable currents .
in fact , the theoretical current density transport limits achievable with an aperture of the order of @xmath9{\mu m}$ ] a unit cell size of @xmath10{\mu m}$ ] and an applied field of @xmath11{v}{m}$ ] ( taken as 50% of the breakdown limit from fig . 4 in [ sl02 ] ) will be @xmath12{a}{cm^2}$ ] for xe ions and @xmath13{a}{cm^2}$ ] for h ions , both at .
normal ion sources produce of the order of @xmath14{ma}{cm^2}$ ] beam currents for xe , for hydrogen up to @xmath15{a}{cm^2}$ ] have been achieved@xcite
. the ion source will therefore be the limiting factor in most applications and bunching , funneling or other methods of increasing the current density might be utilized to overcome this limitation .
the proposed accelerator structure consists of two main components : rf - units and esq doublets .
the rf - units will provide the acceleration for the ion beam and the esqs will allow effective beam transport along the accelerator structure .
each rf - wafer unit consists of two vacuum gaps between ring electrodes that are used for acceleration and a field - free drift region of a specific length that allows the rf - field to adjust its phase , so that the ions entering the second gap will also see an accelerating field . for each rf - unit
the ions enter and leave the unit at ground potential .
all rf - wafers share the same design making batch fabrication possible : each beamlet corresponds to a through hole aperture in the wafer that has a ring electrode at its entrance and exit on the surface of the wafer . in silicon
this would be a deposited metal ring and in our circuit board design the on board copper is utilized for this purpose . to create an rf - unit ,
four wafers are stacked together .
both sides of the outer wafers are grounded and all four sides of the inner wafers are connected to the rf - source ( see fig .
[ fig : rf - concept ] ) .
we use gaps as acceleration gaps .
the drift distance is calculated depending on the expected ion energy , ion mass , and rf - frequency .
precision washers are used between the wafers to define the gap distances . to focus the beam we rely on esqs .
each consists of two pairs of electrodes which are biased positive and negative respectively . to implement esq components , we form four electrodes around each beam aperture and run electrical connections to the front and back of each wafers for the positive and negative voltages ( see fig .
[ fig : esq - concept ] ) .
the esq structure is therefore completely contained in a single wafer .
a single esq wafer focuses the beam in one direction and defocuses the beam in the other .
we therefore use two esq wafers to form a focusing doublet to provide an overall focusing effect on the beam .
for initial prototyping we utilize pc board ( fr-4 ) to fabricate the rf and esq wafers . using laser micromachining ,
top and bottom metal layers are patterned and holes are drilled through the board .
alignment between top and bottom is achieved by using an integrated vision system and pre - fabricated alignment fiducials .
furthermore , by using the integrated camera of the tool , top and bottom layers can be registered for alignment .
steps of the process to fabricate rf wafers are given in fig .
[ fig : rf - fab ] . in this process , we start with fr-4 based board that has copper on both sides as seen in the cross section .
the circular holes are created using a laser tool .
then laser cutting is used to define top and bottom metal routing . a top and bottom view of the fabricated rf wafer is shown in fig . [ fig : rf - fab ] .
the main steps of the process to fabricate esq wafers are given in fig .
[ fig : esq - fab ] . in this process
, the fr-4 based board also has copper on both sides .
the holes are then created using the laser tool . as the holes in the pcbs are created using a scanned laser beam rather than a milling tool ,
arbitrary hole shapes can easily be realized . after defining of holes ,
a copper layer is evaporated onto the board in a conformal evaporator with a rotating chuck system on both sides ( @xmath16{\mu m}$ ] thick ) .
furthermore , for better sidewall coverage , copper is also electroplated from both sides .
then the laser is then used to remove the metal on certain parts of the sidewalls of the holes to isolate the electrode .
finally , the laser is used again to define top and bottom metal routing .
to test the major components of this technology , a compact test stand has been set up .
a filament driven , multi - cusp ion source@xcite is used to generate a plasma .
the source operates using argon or helium at a pressure in the plasma chamber of around ( granville - phillips convectron atm 385003 ) that is connected directly to the source body . to create the plasma we rely on continuous gas flow .
the filament is then operated at several amperes and a arc pulse is applied for a certain time ( normally @xmath17{\mu s}$ ] ) to ignite the plasma .
ions are extracted during the arc pulse by floating the source body to a high voltage ( ) and using a three electrode extraction system ( see fig . [
fig : setup ] ) .
the plasma facing electrode is not electrically connected to a fixed voltage source and therefore floats to the plasma potential during operation .
a voltage of has been measured during plasma operation .
the second electrode is used to extract the ion beam and is biased at a negative voltage relative to the source body .
a third electrode has been implemented that can be utilized to gate the extracted beam and short , uniform ion pulses of @xmath18{\mu s}$ ] bunch length with no electron pre - pulse have been achieved . however , in the work reported here , we do not utilize the gating mechanism and the third electrode is biased slightly above the second grid , so that ions are extracted continuously during the arc - pulse .
the ions then gain their remaining kinetic energy when they are accelerated from the source potential to a grounded exit aperture .
to demonstrate multiple parallel beamlets , we use a multi - aperture extraction system . in this first prototype a @xmath19 array of beam extraction apertures is implemented for each electrode , where each extraction aperture has a diameter and the electrodes are positioned with an inter - hole distance of .
array of beamlets .
a faraday - cup is used to measure the beam current .
alternatively a scintillator and a ccd camera can be used to image the beam . ] after the ions leave the beam extraction section we utilize the beam to test our accelerator components . for the rf - tests , the test structure consisted of two rf - units .
the drift gap length is adjusted to match a argon ion beam , a driving rf - frequency of and an acceleration voltage of .
since the acceleration gaps are a serious of _ pull _ and _ push _ stages , the applied rf - voltage needs to change phase by .
the drift gap between the center points of the gaps is given by @xmath20 where @xmath21 is the relative ion velocity @xmath22 , and @xmath23 is the rf wavelength .
we summarize the implemented distances in our two rf - unit ( four acceleration gaps ) setup in table [ tab : drifts ] .
.[tab : drifts]drift distances used between rf - wafers , designed for , rf signal and a ion beam .
[ cols= " < , < " , ] the rf high - voltage for our experiments is generated by amplifying the output of a signal generator ( agilent 33520a ) to an output power of max .
( mini circuit zhl-2 - 8 ) where it can be used to directly drive the gate of an rf mosfet and excite the tuned resonant circuit at the same frequency resulting in another gain of 40 .
the losses in the mosfet and the rf wafer capacitance ( @xmath24{pf}$ ] per gap ) are the dominant loads .
the frequency used for the experiments presented here is . a peak amplitude on the printed circuit board of up to has been measured .
in the absence of a pre - bunching ( or chopper ) section between the ion source will deliver a constant ion current ( compared to the rf - period ) , therefore some fraction of ions are expected to get accelerated and others decelerated .
a retarding potential analyzer was used to measure the beam energy distribution .
this has been implemented by adding a biased grid after the rf - units followed by a faraday cup to measure the beam current . by scanning the grid voltage ,
the faraday - cup at the end of the beamline selectively detects the current of ions with a kinetic velocity that is higher than the applied voltage on the grid .
the esq wafers are tested in a similar fashion . instead of the rf - units a single esq or a doublet wafer
is mounted behind the ion source .
we then use a scintillator ( rp 400 plastic scintillator ) and a fast image intensifying camera ( princeton instruments ) to look at the beam output .
since a filament driven ion source is being used , light from the filament also reaches the scintillator , by looking at the scintillator at an angle , we avoid overlapping the light from the scintillator with light output from the ion beam hitting the scintillator . as both have roughly the same amplitude , this avoids the need for background subtraction .
voltage scans on the esq electrodes then result in beam deflection that can be measured .
to characterize the ion source , we scanned the extraction voltages of the ion source .
the source was operated at using argon .
the filament was on for to create a stable and high enough filament temperature .
then the arc voltage was pulsed at for .
the source was floated at and then extraction voltages on the second and third electrode was scanned . here , the third electrode was always held higher than the extraction electrode . fig .
[ fig : source ] shows the resulting beam current measured in the source without any rf - unit or esq - units present .
as you can see , the ion current increases according to the child - langmuir law of space - charge limited extraction at the beginning and then , depending on the plasma density , changes at one point into an emission limited regime . operating the source in the emission limited regime , will generate more noise from shot - to - shot , since the output level with depend on the current gas pressure , filament conditions , etc .
therefore , an operating point at an extraction voltage of was chosen for an arc current of , which provided a very stable source performance .
next a two stage rf - unit was tested .
the beam energy profile was measured at three different conditions : 1 ) minimum rf - amplitude ( lowest setting on the rf - generator ) 2 ) rf - amplitude at amplitude and 3 ) rf - amplitude at amplitude .
the results are shown in fig .
[ fig : rf - result ] together with simulated results ( solid lines ) .
the simulation uses a very simply 1d - model to estimate the beam energy of several nanosecond long beam pulses and a perfect energy filter in front of the faraday - cup .
the simulated results are based on the rf - frequency and amplitude , the ion energy and mass , as well acceleration gap positions and contains no free parameters .
results from using several rf - gaps as well discussion of the simulation program will discussed in a future publication . even for the case with almost no rf ,
most of the feature in the signal still can be attributed to the few volts of remaining rf - amplitude .
the remaining discrepancy is due to the energy spread that is intrinsic to the ion source .
the measured beam energy can be estimated to be .
for the cases where the rf was turned on , we measured a wider energy spread .
as can be seen , some particles arrive at the correct time to achieve full acceleration in each gap , whereas other arrive at different phases of the rf and either get partial accelerated , do not feel any acceleration or get decelerated .
the measured energy distribution fits very well with our simulated results from the 1d - model .
as one can see , the maximum beam energy is roughly given by the starting beam energy and four times the maximum rf - amplitude .
the fact that the measured energy is a bit lower can be attributed to the fact that the gap distances were designed for a slightly higher rf - amplitude . for the esq tests ,
we first utilized a single esq wafer and demonstrated the typical elliptical deformation of a round beam that is the result of focusing the beam in one direction and at the same time defocusing the beam in the other direction , see fig .
[ fig : esq - result ] for results from applying different polarities to the esq electrodes . combining two esqs into a doublet
then allows the beam to be focused in both directions . to demonstrate this
, we were able to choose the voltages on an esq doublet so that the initial round beam is again focused to a round beam after passing to energies esq ( not shown ) .
in this paper we have shown that the basic components needed to implement a mems based meqalac work .
this opens the path to compact , high current accelerator that can be used , for example , to drive fusion reactions with ion beams as well as for other applications .
we envision this technology to be applicable for ions beam in the to several range , with average beam current densities up to @xmath25{ma}{cm^2}$ ] .
the technology can also be used for lower energies , however below it will be more effective to use a single high voltage gap to directly accelerate the ions . above beam energy , the advantages of having cheap components and lower voltages will be more and more important . for this technology to be competitive
we believe that an rf - amplitude of several kilovolts is needed per acceleration gap .
this way , we will be able to achieve gradients of for the accelerator structure at frequencies in the range . to accomplish these high gradients we are currently investigating the use of on board resonators with a high q that already have been shown to produce the required voltages .
first prototypes are currently being designed and we will test these devices in the coming month to be able to integrate most of the rf - stack onto the wafers . switching from pcboard to silicon will also provide a path for mass fabrication and better manufacturing precision .
we have shown that a compact rf - cell without a resonant cavity can be used to accelerate an ion beam .
for this proof of concept printed circuit board structures have been used , whereas for a final accelerator we envision the use of silicon wafers .
this will allow smaller beamlets packed to a higher density on a wafer for increased effective beam - current densities .
furthermore , focusing elements in the form of electrostatic quadrupoles will be added to the accelerator to allow for beam transport and refocusing of the beam along the beamline . | a new approach for a compact radio - frequency(rf ) accelerator structure is presented .
the idea is based on the multiple electrostatic quadrupole array linear accelerator ( meqalac ) structure that was first developed in the 1980s .
the meqalac allowed scaling of rf - structure down to dimensions of centimeters while at the same time allowing for higher beam currents through parallel beamlets .
using micro - electro - mechanical systems ( mems ) for highly scalable fabrication , we reduce the critical dimension to the sub - millimeter regime , while massively scaling up the potential number of parallel beamlets .
the technology is based on rf - acceleration components and electrostatic quadrupoles ( esqs ) implemented in a silicon wafer based design where each beamlet passes through beam apertures in the wafer .
the complete accelerator is then assembled by stacking these wafers .
this approach allows fast and cheap batch fabrication of the components and flexibility in system design for different applications . for prototyping these technologies ,
the components have been fabricated using printed circuits board ( pcboard , fr-4 ) . in this paper
, we present first proof of concept results of the major components : rf - acceleration and esq focusing . ongoing developments on implementing components in silicon and scaling of the accelerator technology to high currents and beam energies will be discussed . |
demand - side management helps utilities to regulate increasing energy demand by utilizing existing power grid infrastructure .
recent efforts of demand - side management include load - shifting methods , load curtailing methods and energy conservation strategies @xcite . distributed energy resources such as energy storage devices and renewable energy resources provide vast opportunities for demand - side management by storing extra energy generated by renewable resources that can be dispatched to support peak energy demand
. in general , effectiveness of consumer - driven demand - side management methods depends on active participation of users .
however , in the long run , users may change their participating behavior leading to unexpected outcomes such as lower peak energy reduction and economic benefits .
therefore , designing successful demand - side management approaches have often been challenging with volatile user behavior @xcite . in this paper
, we investigate impacts of realistic energy user behavior , which is not completely rational , on a decentralized energy trading system proposed to regulate electricity demand of a residential community . in the energy trading system , users with photovoltaic ( pv ) energy generation can decide to participate across time to trade energy with a community energy storage ( ces ) device .
first , we elaborate a non - cooperative stackelberg game to study the energy trading between the ces operator and participating users where the ces operator acts as the leader and the users are their followers .
then we develop another non - cooperative game between users to explore their behavior in determining optimal energy trading starting times that minimize personal daily energy costs under two different user - behavioral models : expected utility theory and prospect theory .
the contributions of this work are : * with time - varying subsets of active participating users that depend on their decisions of participating - time , the energy trading system attains a unique stackelberg equilibrium across time where the ces operator maximizes revenue while users minimize energy costs .
* benefits of the energy trading system are robust to users participating - time strategies that significantly deviate from complete rationality .
game - theoretic demand - side management methods have been widely investigated in literature @xcite .
these studies assume that users act rationally and ideally obeying the strategies predicted by game - theoretic systems .
however , social studies have proved that the rationality assumption of game theory can be violated in real world when users face uncertainty in decision making @xcite .
abundant research using prospect theory has shown how real life user behavior contravenes the conventional game theoretic rationality assumption @xcite . in @xcite ,
a prospect theoretic study for a load - shifting approach showed that deviations of users decisions to participate from conventional game - theoretic decisions result in significantly different outcomes .
in contrast to @xcite , we apply prospect theory to study users behavior of choosing to participate across time in a stackelberg game - theoretic energy trading system that does not intend to shift regular energy consumption of users . in this regard , we show that the outcomes of the energy trading system are indistinguishable under both prospect theory and expected utility theory , even though users decisions to choose to participate differ between the two models .
the stackelberg game - theoretic energy trading system between a ces device and users in @xcite assumes users participate from the beginning of day and hence the number of users remain consistent over time . here
, we extend the stackelberg energy trading system to study users decisions of selecting energy trading starting times incorporating prospect theory .
the ces - user stackelberg game in this paper differs from that in @xcite because the number of active participating users is time - variant depending upon each user s decision of choosing an energy trading starting time .
the community consists of two types of energy users : participating users @xmath0 and non - participating users @xmath1 .
the users @xmath2 have rooftop pv panels and they are the players in the energy trading optimization who trade energy with the grid and the ces device .
the users @xmath3 are conventional grid users without behind - the - meter energy generation and are not players in the energy trading optimization . depending on net pv energy after consuming
, the users @xmath2 are classified into surplus users @xmath4 and deficit users @xmath5 those are time - dependent . for the energy trading optimization , the entire control time period @xmath6 , usually a day ,
is partitioned into @xmath7 number of equal time slots with granularity of @xmath8 .
we assume that pv power generation and demand forecasts of the following day are available to the users @xmath2 to decide their day - ahead energy trading strategies .
if @xmath9 and @xmath10 are the pv energy and the regular energy demand of user @xmath11 at time @xmath12 , respectively , then they sell / buy energy amount @xmath13 to / from the ces device at time @xmath14 such that , @xmath15 where @xmath16 is the grid energy consumption of the user .
note that @xmath17 when the user buys energy from the grid and @xmath18 when the user sells energy to the grid . if the surplus energy of the user @xmath19 is @xmath20 , each user @xmath21 sells energy to the ces device and user @xmath22 buys energy from the ces device such that , @xmath23 the ces operator trades @xmath24 energy with the grid at each time @xmath14 where @xmath25 if the ces device is charged ( discharged ) . here , we use the same ces model given in @xcite that is similar to the energy storage model in @xcite . in this regard ,
per - slot energy trading amounts are given as @xmath26 and @xmath27 where @xmath28 and @xmath29 are the per - slot charging energy amounts , and @xmath30 and @xmath31 are the per - slot discharging energy amounts .
we define a charging efficiency @xmath32 , a discharging efficiency @xmath33 and a leakage rate @xmath34 for the energy storage .
denoting @xmath35 is the charge level at the beginning of day , the energy capacity limit of the ces device gives , @xmath36{\boldsymbol{\beta}}\preceq { \boldsymbol{b } } , \label{eq : id6}\ ] ] where @xmath37 with elements of maximum energy capacity of the ces device @xmath38 .
@xmath39 with elements @xmath40_l=\tau^l$ ] and the @xmath41 entry of the lower triangular matrix @xmath42 is @xmath43_{l , m}=\tau^{l - m}$ ] .
@xmath44^t$ ] , @xmath45 is the @xmath46 zero matrix and @xmath47 with elements @xmath48 , respectively .
we define to ensure the continuity of the ces device operation of the following day and to avert its over - charging or over - discharging across @xmath6 such that , @xmath49 where @xmath50 is the charge level at the end of day .
readers are referred to @xcite for detailed description of the ces model .
the pricing mechanism of the grid is similar to @xcite and in particular the unit energy price at time @xmath14 depends on the total load on the grid at time @xmath14 , @xmath51 where @xmath52 is the total grid load of the users @xmath3 and . then at time @xmath14 , the unit energy price of the grid is , @xmath53 where @xmath54 and @xmath55 .
the ces operator also adopts a unit energy price @xmath56 for their energy transactions with the users @xmath2 such that any user @xmath57 receives @xmath58 from the ces operator for their selling energy @xmath13
. then the energy cost of the user @xmath11 at time @xmath14 is , @xmath59 the ces operator obtains a revenue from the energy trading with the users @xmath2 and the grid that is given by , @xmath60 here , we assume that the energy trading between the ces operator and the grid uses the energy rate of the grid .
in the energy trading system , the ces operator maximizes their revenue in by choosing optimal @xmath56 and @xmath24 . following the strategies of the ces operator , each user @xmath11
is supposed to minimize their energy cost in at each time @xmath12 by determining optimal @xmath13 .
based on contractual agreements with the system owners , the users @xmath2 can individually choose a time @xmath61 to start energy trading with the system such that their total daily energy cost is minimized ( this process is explained in section [ sec3 ] ) .
after participating at @xmath62 , they continue to trade energy for @xmath63 . given the opportunity to choose energy trading starting times
, the number of active participating users at each time @xmath14 may not be uniform , and we denote the number of active participating users at time @xmath14 is @xmath64 where @xmath65 . using the pricing signal @xmath66 $ ] and the grid energy trading profile @xmath67 $ ] broadcasted by the ces operator , the users @xmath68 at each time @xmath69 $ ] minimize their personal energy costs in .
let us consider a single time slot @xmath14 where @xmath70[multiblock footnote omitted ] .
then for user @xmath71 , the cost function is quadratic with respect to @xmath72 , @xmath73 where @xmath74 and @xmath75 using and . here
, @xmath76 is the total grid energy load at time @xmath14 excluding the load of the user @xmath77 and @xmath78 .
clearly , is interdependent on each other s behavior and we study the energy trading coordination between the users @xmath68 using a non - cooperative game @xmath79 . here
, @xmath80 is the strategy set available to the users @xmath68 and @xmath81 is the strategy set of the user @xmath77 subject to .
@xmath82 is the set of cost functions given by @xmath83 .
each user @xmath71 determines the optimal energy trading amount from @xmath81 such that their energy cost @xmath84 is minimized . here , @xmath85 denotes the strategy profile of the opponents of the user @xmath77 that is given by @xmath86
. then the optimization problem of each user @xmath71 is to find , @xmath87 note that the game @xmath88 is similar to the non - cooperative subgame between users in @xcite .
however , the subsets of players @xmath68 are not uniform for the game @xmath88 played at each time @xmath12 in contrast to @xcite . although the number of players is time - variant , using the same rationale in @xcite we can prove that the game @xmath88 played at any particular time @xmath14 has a unique nash equilibrium for any feasible @xmath56 and @xmath24 . at the nash equilibrium of the game @xmath88 , the optimal energy trading amount of the user @xmath77 , @xmath89
can be found by setting the first derivative of with respect to @xmath72 to zero that gives , @xmath90 solving for all users @xmath68 simultaneously , we can obtain , @xmath91 where @xmath92 the ces operator also maximizes their revenue in by determining optimal @xmath93 and @xmath94 . by substituting in
, we can write the objective of the ces operator as , @xmath95={\operatornamewithlimits{argmax}}_{{\boldsymbol{a}},{\boldsymbol{l_q}}\in\mathcal{q}}{\sum_{t=1}^k ( \lambda_1 a_t^2+\lambda_2 a_t+\lambda_3 l_{q , t}^2+\lambda_4l_{q , t } ) } , \label{eq : id14}\ ] ] where @xmath96 , @xmath97 , @xmath98 , and @xmath99 .
@xmath100 is the strategy set available to the operator subject to and .
there is a unique solution for the objective function of the ces operator , since is strictly concave because of the negative definite hessian matrix with respect to all feasible @xmath101 and the strategy set @xmath100 is convex due to linear constraints and .
the ces operator first sets optimal @xmath102 $ ] to maximize and broadcasts them to the users @xmath103 .
then the users @xmath68 at each time @xmath12 follow these signals to find optimal @xmath72 by playing the game @xmath88 .
we model this hierarchical interaction between the ces operator and the users @xmath2 using a non - cooperative stackelberg game @xmath104 . in the game @xmath104 ,
_ players _ are the ces operator and the users @xmath2 where the ces operator is the leader and the users @xmath2 are the followers . as the _ strategies _ , the ces operator determines @xmath102\in \mathcal{q}$ ] to maximize and at time @xmath14 , user @xmath71 selects @xmath105 to minimize cost in .
the _ utilities _ are as defined in for the ces operator and for the user @xmath71 .
let @xmath106 $ ] be the solution of and @xmath107^t,\dotsm,[{\boldsymbol{\hat x_k}}]^t\}$ ] where @xmath108 be the solution of the game @xmath88 at time @xmath12 .
then the point @xmath109 $ ] is a stackelberg equilibrium if and only if , @xmath110 @xmath111 the game @xmath104 has a unique stackelberg equilibrium .
the game @xmath88 played at any time @xmath14 has a unique nash equilibrium for feasible @xmath56 and @xmath24 .
further , the revenue maximization of the ces operator in has a unique solution .
hence , the game @xmath104 converges to a unique stackelberg equilibrium once the ces operator obtains optimal strategy @xmath112 while the users @xmath2 attain their k - tuple of unique nash equilibrium solutions @xmath113 .
before the stackelberg game in section [ sec2 ] takes place , the users @xmath2 individually select optimal times @xmath62 to start energy trading such that their total daily energy costs are minimized .
similar to @xcite , to explicitly study such user behavior with respect to choosing to participate across time in our system , we develop a non - cooperative game @xmath114 between the users @xmath2 that has the strategic form @xmath115 and study it under expected utility theory and prospect theory . here , @xmath116 is the set of available strategies to the users @xmath2 i.e. , energy trading starting times .
@xmath117 where @xmath118 ; @xmath119 .
@xmath120 is the set of cost functions that captures the daily energy costs for each user @xmath11 .
note that the stackelberg equilibrium described in section [ sec2 ] depends on temporal distribution of the users @xmath68 which is a result of how the users @xmath2 begin energy trading across time . at the stackelberg equilibrium corresponding to an action ( i.e. , energy trading starting time ) profile @xmath121 of the users @xmath2 where @xmath122 , the daily energy cost of user @xmath19 , @xmath123 is , @xmath124 here , @xmath125 are the grid price , ces energy price , user @xmath19 s grid load and their ces energy trading amount at the stackelberg equilibrium obtained for @xmath126 , respectively .
note that @xmath127 is the action profile of the users @xmath2 except user @xmath19 . in the game @xmath114
, each user @xmath11 chooses an energy trading starting time @xmath62 for given @xmath127 such that their energy cost in is minimized .
after the users @xmath2 decide to participate as per optimal energy trading starting times determined by playing the game @xmath114 , the stackelberg energy trading in section [ sec2 ] takes place that ultimately achieves a stackelberg equilibrium . in the long run , the users @xmath2 may change their behavior with respect to choosing an energy trading starting time .
hence , we investigate a solution for the non - cooperative game @xmath114 that captures empirical frequencies of actions followed by the users @xmath2 .
the straightforward interpretation is that each user @xmath11 assigns a probability for each action in @xmath128 .
in such a paradigm , users face uncertainty to make decisions and we characterize solutions for the game @xmath114 based on mixed strategies under two different user - behavioral models : expected utility theory and prospect theory . under the notion of mixed strategies ,
each user @xmath11 determines the optimal probability distribution over the actions in @xmath128 to minimize expected daily energy cost . here
, we explore how the users @xmath2 decide probabilities of energy trading starting times according to expected utility theory assuming that all users make rational choices by objectively viewing their opponents behavior . according to the theory ,
the expected daily energy cost of the user @xmath19 can be given as , @xmath129 where @xmath130 , @xmath131 $ ] and @xmath132 is the probability of choosing @xmath62 by the user @xmath19 .
@xmath133 is the probabilities of the users @xmath2 except user @xmath19 . the intuition behind the cost in relies on the assumption that the user @xmath19 assesses their neighbours empirical frequencies of actions identical to their objective probabilities of choosing actions .
however , this generalization may not be valid in the real world as people overweight outcomes with low probabilities and underweight outcomes with high probabilities .
these observations are clearly explained under prospect theory @xcite . in practice ,
the users @xmath2 may subjectively evaluate their neighbors actions to minimize energy costs .
this characteristic is more realistic than assuming users act rationally and perceive their neighbors behavior objectively @xcite . in this regard ,
we study actual user behavior as to when they select their energy trading starting time using prospect theory . to this end ,
probability weighting functions are used to model the subjective behavior of users when they make decisions under risk and uncertainty . in this regard
, the probability weighting function @xmath134 implies the subjective evaluation of the user @xmath19 about an outcome with @xmath135 probability .
we use the prelec function @xcite to model the subjective perceptions of users on each other s behavior that is given by , @xmath136 here , @xmath137 is a parameter that decreases as the user s subjective evaluation deviates from the objective probability .
if the user s subjective and objective probabilities are equal , then @xmath138 and this corresponds to expected utility theory . assuming that the subjective probabilities of user @xmath11 about their own actions are equal to their objective probabilities , the expected daily energy cost of user @xmath19 under prospect theory is , @xmath139 after defining the expected daily costs of the users @xmath2 , we now analyze the solutions for the game @xmath114 played under expected utility theory and prospect theory . due to
computational usefulness @xcite , here we study the existence of @xmath141nash equilibria . for the game @xmath114 , a mixed strategy profile @xmath142 is an @xmath141nash equilibrium if it satisfies , @xmath143 where @xmath144 is the set of all mixed strategy profiles over @xmath145 and @xmath146 . in general , @xmath141nash equilibria always exist @xcite and for the game @xmath114 , we are interested to find @xmath140-nash equilibrium located close to a mixed strategy nash equilibrium under both expected utility theory and prospect theory .
we use the iterative algorithm proposed in @xcite that was proved to converge to an @xmath140-nash equilibrium close to a mixed strategy nash equilibrium under both expected utility theory and prospect theory . in summary , the algorithm is given by , @xmath147 where @xmath148 is the iteration number , @xmath149 is the inertia weight .
@xmath150 of which , @xmath151 where @xmath152 is the expected cost when the user @xmath19 selects the pure strategy @xmath62 in response to the mixed strategies of other players at iteration @xmath153 i.e. , @xmath154 .
note that for prospect theory , @xmath154 considers the weighted probabilities of other users mixed strategies at @xmath153 .
as the algorithm converges , @xmath140-nash equilibrium with respect to strategy profile @xmath155 is obtained under both expected utility theory and prospect theory . given the equilibrium probabilities of participating - time decisions of the users @xmath2 , we can define the expected revenue of the ces operator under both prospect theory and expected utility theory . in this regard ,
if @xmath156 and @xmath157 are the @xmath140-nash equilibriums under expected utility theory and prospect theory , respectively , then the subsequent expected daily ces revenue @xmath158 in each case can be obtained by , @xmath159 where @xmath160 is the ces revenue as per at the stackelberg equilibrium corresponds to @xmath126 , @xmath161 for expected utility theory and @xmath162 for prospect theory .
in simulations , we consider real data of average pv power and user demand of the western power network in australia on a summer day @xcite ( see fig .
[ fig : pvdemand ] ) and we assume that all users have power profiles same to these average profiles .
further , @xmath163 , @xmath164 , @xmath165 kwh , @xmath166 kwh , @xmath167 , @xmath168 and @xmath169 @xcite .
peak hours of the grid are between 16.00 and 23.00 and we select @xmath170 such that @xmath171 .
we choose @xmath172 such that the predicted grid price range is same to the reference time - of - use price range in @xcite and @xmath173 is set to a constant such that the average predicted grid price is equal to the average reference price .
the community has 10 households where 6 users are participating users @xmath2 in the system .
the allowable energy trading starting times for the users @xmath2 are 01.00 , 12.00 and 17.00 so that @xmath174 . for comparisons ,
we use a baseline without a ces device where the users @xmath2 trade energy directly with the grid that uses the same energy cost model . for the algorithm ,
we use @xmath175;~\forall n\in \mathcal{p}$ ] and @xmath176 . fig .
[ fig : expcostsav ] illustrates expected cost savings of the users @xmath2 under expected utility theory , and under prospect theory for three different @xmath177~(\text{i.e.,}~0.7,~0.4~\text{and}~0.1)$ ] tends to @xmath178 users become more subjective deviating from the objective evaluation assumption in expected utility theory . ] assuming @xmath179 .
here , cost savings are calculated compared to the baseline . when @xmath180 , and even when @xmath181 with significant non - ideal behavior , the expected cost savings remained almost 28% under both models because for all users , participation probabilities at each time in @xmath128 using prospect theory do not significantly deviate from those obtained under expected utility theory as shown in table [ table 1 ] . when @xmath182 , the participation probabilities at @xmath183 are significantly increased for the fourth and fifth users compared to those predicted using expected utility theory ( see table [ table 1 ] ) . as a result ,
the expected cost savings reduced from 28% to 21.5% for all users . .
] . ] . ]
user&@xmath62 = 1&@xmath62 = 12&@xmath62 = 17&@xmath62 = 1&@xmath62 = 12&@xmath62 = 17&@xmath62 = 1&@xmath62 = 12&@xmath62 = 17&@xmath62 = 1&@xmath62 = 12&@xmath62 = 17 + 1 & 0.9966&0.0005&0.0029 & 0.9988&0.0005&0.0007 & 0.9989&0.0005&0.0006 & 0.9979&0.0009&0.0012 + 2 & 0.9966&0.0005&0.0029 & 0.9988&0.0005&0.0007 & 0.9989&0.0005&0.0006 & 0.9979&0.0009&0.0012 + 3 & 0.9966&0.0005&0.0029 & 0.9988&0.0005&0.0007 & 0.9989&0.0005&0.0006 & 0.9979&0.0009&0.0012 + 4 & 0.0070&0.9924&0.0006 & 0.0076&0.9918&0.0006 & 0.0070&0.9924&0.0006 & 0.9979&0.0009&0.0012 + 5 & 0.0070&0.0005&0.9925 & 0.0076&0.0005&0.9919 & 0.0095&0.0005&0.9900 & 0.9979&0.0009&0.0012 + 6 & 0.9966&0.0005&0.0029 & 0.9988&0.0005&0.0007 & 0.9989&0.0005&0.0006 & 0.9979&0.0009&0.0012 + fig .
[ fig : avgexpcostsav ] , fig .
[ fig : ces rev ] and fig .
[ fig : parred ] depict the variations in different aspects of system performance across the range of possible @xmath184 values . here
, larger @xmath184 tending to 1 reflects that the users behave closer to the rationality assumption in expected utility theory , and smaller @xmath184 tending to 0 implies that their evaluations of opponents actions are more distorted from that of expected utility theory .
[ fig : avgexpcostsav ] shows that under expect utility theory , the average of expected cost savings of the users achieved by participating in the system is 28.1% . on the other hand
, even if the users weighting effects on their opponents actions are getting larger , i.e. , when @xmath184 is getting smaller , the expected cost savings will not significantly fluctuate and remain almost at 28% except for @xmath185 .
[ fig : ces rev ] shows that , when @xmath186 , expected revenue for the ces operator retains nearly unchanged compared to the expected revenue calculated under expected utility theory . in terms of demand - side management of the grid , the expected peak - to - average ratio reduction compared to the baseline will not change notably from the peak - to - average ratio reduction predicted using expected utility theory when @xmath186 .
this is because as shown in table [ table 1 ] , for @xmath186 , users prospect theoretic probabilities of participation at each time remain almost the same as those in expected utility theory . when @xmath185 , the fourth and fifth users will more likely to start energy trading from the beginning under prospect theory , which is not the case under expected utility theory
. however , this behavioral change will only reduce the expected peak - to - average ratio reduction from 17.7% to 16.55% ( see fig .
[ fig : parred ] ) .
in this paper , we have studied effects of realistic , non - ideal , behavior of users , with respect to choosing energy trading starting times , on a game - theoretic demand - side management energy trading system between a community energy storage ( ces ) device and users .
first , we have developed the non - cooperative stackelberg game to study the energy trading interaction between the users and the ces operator based on users decisions as to whether to participate across time .
next we have studied a non - cooperative game to explore how the users make decisions to participate in the above energy trading system under two user - behavioral models : prospect theory and expected utility theory .
simulation results show that the benefits of the energy trading system are robust to users strategies of participating - time that significantly deviate from complete rationality .
we postulate that the energy trading system can be scaled to any number of participating users and present similar performance trends
. a. mohsenian - rad , v. wong , j. jatskevich , r. schober , and a. leon - garcia , `` autonomous demand - side management based on game - theoretic energy consumption scheduling for the future smart grid , '' _ ieee trans .
smart grid _ , vol . 1 , no . 3 , pp .
320331 , dec . 2010 .
a. haney , t. jamasb , j. wu , l. platchkov , and m. pollitt , `` demand - side management strategies and the residential sector : lessons from the international experience , '' in _ the future of electricity demand _ , t. jamasb and m. pollitt , eds.1em plus 0.5em minus 0.4em cambridge : cambridge university press , 2011 , ch .
337378 .
i. atzeni , l. ordonez , g. scutari , d. palomar , and j. fonollosa , `` demand - side management via distributed energy generation and storage optimization , '' _ ieee trans .
smart grid _ , vol . 4 , no . 2 ,
866876 , june 2013 . h. k. nguyen , j. song , and z. han , `` demand side management to reduce peak - to - average ratio using game theory in smart grid , '' in _ proc .
computer commun . workshops ( infocom wkshps ) _ , march 2012 , pp . 9196 .
y. wang , w. saad , n. mandayam , and h. poor , `` integrating energy storage into the smart grid : a prospect theoretic approach , '' in _ proc .
conf . on acoustics , speech and signal processing ( icassp )
_ , may 2014 , pp .
77797783 .
c. p. mediwaththe , e. r. stephens , d. b. smith , and a. mahanti , `` competitive energy trading framework for demand - side management in neighborhood area networks , '' _ arxiv e - prints _ , dec .
[ online ] .
available : http://adsabs.harvard.edu/abs/2015arxiv151203440m b. jones , n. wilmot , and a. lark , `` study on the impact of photovoltaic ( pv ) generation on peak demand , '' western power , australia , tech .
, april 2012 .
[ online ] .
available : http://www.westernpower.com.au | this paper investigates effects of realistic , non - ideal , decisions of energy users as to whether to participate in an energy trading system proposed for demand - side management of a residential community .
the energy trading system adopts a non - cooperative stackelberg game between a community energy storage ( ces ) device and users with rooftop photovoltaic panels where the ces operator is the leader and the users are the followers .
participating users determine their optimal energy trading starting time to minimize their personal daily energy costs while subjectively viewing their opponents actions . following a non - cooperative game , we study the subjective behavior of users when they decide on energy trading starting time using prospect theory .
we show that depending on the decisions of participating - time , the proposed energy trading system has a unique stackelberg equilibrium at which the ces operator maximizes their revenue while users minimize their personal energy costs attaining a nash equilibrium .
simulation results confirm that the benefits of the energy trading system are robust to decisions of participating - time that significantly deviate from complete rationality . |
after the discovery that perfect hydrodynamics can be valid for describing the phenomenology of relativistic heavy ion collider ( rhic ) at brookhaven national laboratory @xcite , people are now interested in relativistic hydrodynamics for _ dissipative _ systems ; see the recent excellent review articles @xcite . recently , tsumura , kunihiro ( the present authors ) and ohnishi ( abbreviated as tko ) @xcite derived generic covariant hydrodynamic equations for a viscous fluid from the relativistic boltzmann equation in a systematic manner with no heuristic arguments on the basis of the so - called renormalization group ( rg ) method @xcite .
although the hydrodynamic equations they derived are the so - called first - order ones , the equations have remarkable aspects : the generic equation derived by tko can produce a relativistic dissipative hydrodynamic equation in any frame with an appropriate choice of a macroscopic flow vector @xmath2 ( @xmath3 ) , which defines the coarse - grained space and time ; the resulting equation in the energy frame coincides with that of landau and lifshitz @xcite , while that in the particle frame is similar to , but slightly different from , the eckart equation @xcite .
let @xmath4 and @xmath5 be the dissipative term of the symmetric energy - momentum tensor and the particle - number vector , respectively . owing to the ambiguity in the separation of the energy and the mass inherent in relativistic theories , one must choose the local rest frame ( lrf ) where the flow velocity @xmath6 with @xmath7 is defined : one of the typical frame is the energy ( landau ) frame in which @xmath8 with @xmath9 and @xmath10 , i.e. there is no dissipative energy flow .
on the other hand , another typical frame is the particle ( eckart ) frame in which @xmath11 , i.e. there is no dissipative particle flow .
both in the energy and particle frames , the dissipative terms of the energy - momentum tensor and the particle - number vector are usually assumed to satisfy the constraints , @xmath12 and @xmath13 .
these phenomenological ansatz have been employed as the matching conditions even in the subsequent `` derivations '' of the so - called second - order equations @xcite ; note that even in the grad s moment method @xcite , some ansatz are needed to @xmath4 and @xmath14 as the matching conditions , for which different proposals exist @xcite . here
we emphasize that the matching conditions touch on the fundamental but not yet fully understood problem how to define the lrf in the relativistic fluid dynamics for a viscous system .
the way how to define the lrf or equivalently to fix the matching condition is unsolved yet , and remains a nontrivial and fundamental problem in the field of nonequilibrium relativistic dynamics , although there have been no serious consideration on this difficult problem in the literature .
actually , we shall argue that these phenomenological ansatz , especially eq.([eq : pt - e ] ) , can be false and actually is not compatible with the underlying kinetic equation .
in fact , it is found that the tko equation in the particle frame derived from the relativistic boltzmann equation satisfies @xmath15 but does not satisfy eq.([eq : pt - e ] ) .
one should here note that the derived condition ( [ eq : pt - tko ] ) is identical to a matching condition postulated by marle @xcite and advocated by stewart @xcite in the derivation of the relativistic hydrodynamics from the relativistic boltzmann equation with use of the grad s moment method .
in their paper @xcite , tko proved that the eckart constraint ( [ eq : pt - e ] ) in the particle frame can not be compatible with the underlying relativistic boltzmann equation for the first - order hydrodynamic equation . in spite of the first - order one ,
the tko equation in the particle frame is free from the pathological properties @xcite in contrast to the original eckart equation with which the thermal equilibrium becomes unstable for a small perturbation @xcite .
one may naturally ask if the eckart constraint ( [ eq : pt - e ] ) should be replaced with ( [ eq : pt - tko ] ) even for the so - called second - order equation like israel - stewart one . and are any modifications needed to the constraints in the landau frame ?
a purpose of this letter @xcite is to answer these questions both by phenomenological and microscopic analyses .
we shall see that the eckart constraint should be replaced with the new one even in the second - order equation , while no modification is necessary for the constraints in the energy frame .
we shall derive the second - order dissipative relativistic hydrodynamic equations in a generic frame with a continuous parameter @xmath16 from the relativistic boltzmann equation .
we shall derive the relaxation terms for a generic frame with the new constraint , and present explicitly those in the energy and particle frames .
we shall show that the viscosities are frame - independent but the relaxation times are generically frame - dependent in accordance with the observation by betz et al .
@xcite , although the constraint to @xmath4 is quite different .
let @xmath17 and @xmath18 be the symmetric energy - momentum tensor and the particle - number vector of the system we consider , respectively .
the total number of independent variables is fourteen , and the dynamical evolution of these variables are governed by the respective balance equations ; @xmath19 with use of an arbitrary four vector @xmath6 with @xmath7 , @xmath17 and @xmath18 can be cast into the tensor - decomposed forms , @xmath20 respectively .
here , @xmath21 , @xmath22 , and @xmath23 are the internal energy , pressure , and particle - number density in the dissipative system ; @xmath24 , @xmath25 , and @xmath26 , with @xmath27 , @xmath28 , and @xmath29 being the corresponding quantities in the local equilibrium state characterized by the temperature @xmath30 and the chemical potential @xmath31 .
note that we have made it explicit by @xmath32 , @xmath33 , and @xmath34 that the dissipations may cause corrections to all these quantities , although only the correction to the pressure has been considered in the literature ; @xmath33 is identified with the bulk pressure @xmath35 .
we emphasize that there is no persuading reasoning that only the pressure acquires corrections due to the dissipative process .
the dissipative parts of the energy - momentum tensor and particle - number vector are identified as @xmath36 and @xmath37 , respectively . the energy flow relative to @xmath6
is denoted by @xmath38 , @xmath39 is the flow of particle number relative to @xmath6 , and finally @xmath40 is the shear stress tensor ; @xmath41 , @xmath42 , and @xmath43 . here
the space - like , symmetric and traceless tensor @xmath44 is introduced .
one can easily confirm that @xmath45 , @xmath46 , @xmath47 , and @xmath48 .
this implies that the total number of independent components of @xmath38 , @xmath39 , and @xmath40 is eleven .
since @xmath17 and @xmath18 have the fourteen components in total , @xmath32 , @xmath33 , and @xmath34 have only one independent component other than @xmath30 and @xmath31 .
we take @xmath49 as the independent component as a natural choice , then @xmath32 and @xmath34 can be expressed as @xmath50 and @xmath51 , where @xmath52 and @xmath53 are functions of @xmath30 and @xmath31 ; @xmath54 and @xmath55 . here
we have assumed that the dissipative order of @xmath32 and @xmath34 are the same as that of @xmath33 at most .
we remark that although @xmath52 and @xmath53 may take finite values generically , the functional forms of @xmath52 and @xmath53 can not be determined by the phenomenological theory , as those of @xmath56 , @xmath57 , and @xmath58 can not , either .
all the previous analyses assumed @xmath59 , which has not been recognized so far .
now we shall show that the just usual phenomenological derivation of the hydrodynamic equations in which the second law of thermodynamics is utilized allows the existence of @xmath32 and @xmath34 , i.e. , finite values of @xmath52 and @xmath53 , in the relativistic dissipative hydrodynamic equations .
it is found that the essential point of the proof is the same for the first- and second - order equations where the entropy current @xmath60 is at most linear and bilinear with respect to @xmath35 , @xmath38 , @xmath39 , and @xmath40 , respectively , although the resulting mathematical expressions are much more complicated in the second - order one @xcite .
thus we here take the first - order equation , for the sake of simplicity .
the second - order equations with finite @xmath52 and @xmath53 will be derived microscopically later in this article .
so the entropy current is given by @xmath61 the second law of thermodynamics reads @xmath62 .
the divergence of @xmath60 is found to take the form @xmath63 + q^\mu \ , \bigg [ \frac{1}{t } \ ,
du_\mu + \nabla_\mu \frac{1}{t } \bigg ] - \nu^\mu \ , \nabla_\mu \frac{\mu}{t } + \pi^{\mu\nu } \ , \frac{1}{t } \ , \nabla_\mu u_\nu , \end{aligned}\ ] ] where @xmath64 and @xmath65 . here ,
we have used the conservation laws , eq.s ( [ eq:2 - 001 ] ) and ( [ eq:2 - 002 ] ) , and the first law of thermodynamics , @xmath66 .
the frames define the flow velocity @xmath67 of the fluid : the flow velocity in the particle frame and the energy frame are defined by setting @xmath68 and @xmath69 , respectively @xcite . by these settings , a closed system of the relativistic dissipative hydrodynamic equations is obtained .
note that @xmath68 ( @xmath69 ) is equivalent to @xmath70 ( @xmath71 ) . in the particle frame where @xmath70 , eq.([eq:2 - 016 ] ) is reduced to @xmath72 + q^\mu \ , \bigg [ \frac{1}{t } \ , du_\mu + \nabla_\mu \frac{1}{t } \bigg ] + \pi^{\mu\nu } \ , \frac{1}{t } \ , \nabla_\mu u_\nu .
\end{aligned}\ ] ] it is found that the following constitutive equations , @xmath73,\\ \label{eq:2 - 022 } q^\mu & = & - \lambda \ , t^2 \ , \bigg [ \frac{1}{t } \ , du^\mu + \nabla^\mu \frac{1}{t } \bigg],\\ \label{eq:2 - 023 } \pi^{\mu\nu } & = & 2 \ , \eta \ , \delta^{\mu\nu\rho\sigma } \ , \nabla_\rho u_\sigma , \end{aligned}\ ] ] guarantees the second law of thermodynamics , @xmath62 , with @xmath74 , @xmath75 , and @xmath76 being the bulk viscosity , heat conductivity , and shear viscosity , respectively .
this is because the divergence @xmath77 now becomes positive semi - definite ; @xmath78 thus we realize that there is nothing wrong with the resultant relativistic dissipative hydrodynamic equations with finite @xmath52 and @xmath53 , or equivalently finite @xmath32 and @xmath34 , which is compatible with the second law of thermodynamics .
eq.s ( [ eq:2 - 021])-([eq:2 - 023 ] ) with a restricted condition @xmath59 are identical to the constitutive equations proposed by eckart that are commonly used . in the energy frame
where @xmath71 , we can obtain the constitutive equations in the same way as the particle - frame case with @xmath52 and @xmath53 being kept finite .
the resultant equations are given by eq.s ( [ eq:2 - 021 ] ) , ( [ eq:2 - 023 ] ) , and @xmath79 with @xmath80 being the enthalpy .
it is noted that these equations are reduced to the constitutive equations by landau if we can set @xmath59 . by applying the above argument to the entropy current at most bilinear with respect to @xmath35 , @xmath38 , @xmath39 , and @xmath40 ,
we can obtain the relaxation equations with @xmath52 and @xmath53 being finite , which make up the so - called second - order relativistic dissipative hydrodynamic equations together with the conservation laws in eq.s ( [ eq:2 - 001 ] ) and ( [ eq:2 - 002 ] ) @xcite .
now the dissipative part of the energy - momentum tensor satisfies @xmath81 and @xmath82 .
as emphasized before , the values of @xmath52 and @xmath53 can be determined only from a microscopic theory .
the phenomenological theory can not proceed further because no such logic to determine them is implemented in the theory . in the following section
, we shall show that the microscopic theory gives @xmath83 together with @xmath84 in the particle frame while @xmath59 in the energy frame , and hence @xmath0 but @xmath85 in the particle frame .
this fact tells us that the usual constraint employed for the particle frame must be abandoned , and all the analyses based on this constraint should be redone .
the argument so far is in the stage of thermodynamics where the argument is robust but the parameters such as @xmath52 and @xmath53 as well as the equations of state @xmath56 , @xmath57 and @xmath58 appearing in the theory remain undetermined .
the problem which we encounter is how to reduce a dynamical equation to a slower one described with fewer dynamical variables . for this purpose
, we will investigate the infrared limit of the relativistic boltzmann equation with use of a powerful reduction method , the `` rg method '' @xcite .
the rg method is a systematic reduction theory of the dynamics leading to the coarse - graining of temporal and spatial scales .
the full presentation of the reduction of the relativistic boltzmann equation to the second - order hydrodynamic equation is technical and involved .
so we here only present main results with key several equations , leaving the detailed account to another publication @xcite , although the derivation of a wide class of the first - order equations is presented in ref.@xcite .
we start with the simple relativistic boltamann equation , @xmath86_p(x ) , \end{aligned}\ ] ] where @xmath87 denotes the one - particle distribution function defined in the phase space @xmath88 with @xmath89 being the four momentum of the on - shell particle .
the right - hand side of eq.([eq:2.1.1 ] ) is the collision integral , @xmath90_p(x ) \equiv \frac{1}{2 ! } \ ,
\sum_{p_1 } \ , \frac{1}{p_1 ^ 0 } \ , \sum_{p_2 } \ , \frac{1}{p_2 ^ 0 } \ , \sum_{p_3 } \ , \frac{1}{p_3 ^ 0 } \ , \omega(p \,,\ , p_1|p_2 \,,\ , p_3 ) \ , ( f_{p_2}(x ) \ , f_{p_3}(x ) - f_p(x ) \ ,
f_{p_1}(x ) ) $ ] , where @xmath91 denotes the transition probability owing to the microscopic two - particle interaction .
we are interested in the hydrodynamical regime where the time- and space - dependence of the physical quantities are small . in another word ,
the time and space entering the hydrodynamic equation are the ones coarse - grained from those in the kinetic equation .
thus we are lead to introduce a macroscopic lorentz vector , @xmath92 which specifies the covariant coordinate system and we call the _ macroscopic - frame vector_. with use of @xmath92 , we define the macroscopic covariant coordinate system @xmath93 as @xmath94 and @xmath95 .
we note that the small quantity @xmath96 has been introduced to tag that the space derivatives are small for the system we are interested in .
@xmath96 may be identified with the ratio of the average particle distance over the mean free path , i.e. , the knudsen number . in this coordinate system ,
eq.([eq:2.1.1 ] ) can be cast into @xmath97_p(\tau \,,\ , \sigma ) - \varepsilon \ ,
\frac{1}{p \cdot \ren{a}_p(\tau \,,\ , \sigma ) } \ , p \cdot \ren{\nabla } f_p(\tau \,,\ , \sigma ) , \end{aligned}\ ] ] where @xmath98 , @xmath99 , and @xmath100 . since @xmath96 appears in front of @xmath101 , eq.([eq:2.1.6 ] )
has a form to which the perturbative expansion with respect to @xmath96 can be applied . in the perturbative expansion
, we shall take the coordinate system where @xmath102 has no @xmath103 dependence , i.e. , @xmath104 .
the zeroth - order approximate solution we construct is a stationary solution , which is identical to a local equilibrium distribution function given by the juetner function @xmath105 $ ] .
note that this solution contains five would - be integration constants , @xmath30 , @xmath31 , and @xmath6 with @xmath7 , which can be identified with the temperature , the chemical potential , and the fluid velocity , respectively .
the collision integral is expanded around the zeroth - order solution and is reduced to the linear operator @xmath106_p$ ] .
furthermore , it is found to be convenient to convert @xmath107 to @xmath108_{pq}$ ] , with the diagonal matrix @xmath109 .
we also define the inner product between arbitrary vectors @xmath110 and @xmath111 by @xmath112 with this inner product , we can define a normed linear space .
now the first - order solution is given in terms of the five zero modes of @xmath113 , @xmath114 .
the corresponding variables are just @xmath30 , @xmath31 , and @xmath6 with @xmath115 .
the zero modes span a linear space @xmath116 , which is an invariant manifold for the asymptotic dynamics of the relativistic boltamann equation in the terminology in the dynamical systems @xcite .
then the second - order solution is given by incorporating the next slow modes , which span a linear space @xmath117 .
we naturally require @xmath117 is orthogonal to @xmath118 , that is , @xmath119 .
we find that @xmath117 is expanded by the bilinear forms of momenta ; @xmath120_p$ ] , where @xmath121 , and @xmath122 is the projection to complement to @xmath116 . by definition , @xmath123 is satisfied .
note that the dimension of @xmath124 is nine , which correspond to the number of the new would - be integration constants , @xmath35 , @xmath125 with @xmath126 , and @xmath40 with @xmath47 and @xmath127 .
a generic choice of the macroscopic frame vector is @xmath128 , where @xmath16 is a parameter defining the frame .
for example , @xmath129 ( @xmath130 ) gives the energy ( particle ) frame .
the resultant generic relaxation equations of the second - order hydrodynamic equation with @xmath16 being kept are @xmath131 here , @xmath132 , @xmath133 , and @xmath134 are the thermodynamic forces ; their simple forms retaining only @xmath132 , @xmath133 , and @xmath134 are the usual constitutive equations .
the relaxation equations of @xmath35 , @xmath125 , and @xmath40 are characterized by the relaxation times @xmath135 , @xmath136 , and @xmath137 , while @xmath138 , @xmath139 , @xmath140 , and @xmath141 mean the relaxation lengths .
the correction to the thermodynamic forces @xmath132 , @xmath133 , and @xmath134 are given by @xmath142 , @xmath143 , @xmath144 , @xmath145 , @xmath146 , @xmath147 , @xmath148 , @xmath149 , and @xmath150 .
the continuity equations of the second - order equation in the energy frame is found to be given by setting @xmath129 as in the first - order case @xcite and read @xmath151 and @xmath152 , where @xmath153 the thermodynamic forces are @xmath154 , @xmath155 , and @xmath156 , which clearly show that @xmath84 and @xmath157 as was anticipated .
the energy - momentum tensor and particle - number vector in the particle frame with @xmath130 read @xmath158 and @xmath159 , @xmath160 , and @xmath156 , where @xmath161 is the ratio of the specific heats .
thus we find that @xmath83 with @xmath84 , as we announced .
although we have obtained the relaxation equations for the dissipative forces @xmath35 , @xmath162 , and @xmath163 for arbitrary @xmath16 @xcite , we shall only write down them for two typical frames , i.e. , the energy ( @xmath164 ) and the particle ( @xmath165 ) frames for the sake of the space
. \(a ) in the energy frame ( @xmath164 ) : @xmath166 - 2 \ ,
\kappa_j^{(1 ) } \ , \delta^{\mu abc } \ , \nabla_b u_c - 2 \ , \omega^{\mu a } \bigg\ } \ , j_a\nonumber\\ & & { } - \frac{1}{2 } \ , \ell_{j\pi } \ , \bigg\ { \delta^{\mu cab } \ , \big ( \kappa^{(0)}_{j\pi } \ , \nabla_c \frac{\mu}{t } - \kappa^{(1)}_{j\pi } \ , d u_c \big ) + \frac{\lambda\,\hat{h}^{-2}}{\ell_{j\pi } } \ , \delta^\mu_{\,\,\,c } \ , \partial_d \big(\frac{\ell_{j\pi}}{\lambda\,\hat{h}^{-2 } } \ , \delta^{cdef}\big ) \ , \delta_{ef}^{\,\,\,\,\,\,ab } \bigg\ } \ , \pi_{ab},\\ \label{eq : ef_pi } \pi^{\mu\nu } & = & 2 \ , \eta \ , \delta^{\mu\nu ab } \ , \nabla_a u_b - \ell_{\pi j } \ , \delta^{\mu\nu ab } \ , \nabla_a j_b - \tau_\pi \ , \delta^{\mu\nu ab } \ , d \pi_{ab}\nonumber\\ & & { } - \frac{1}{2 } \ , \ell_{\pi\pi } \ , \bigg\ { - \kappa_{\pi\pi } \ , \delta^{\mu\nu ab } \ , \nabla_a u_b \bigg\ } \ , \pi\nonumber\\ & & { } - \frac{1}{2
} \ , \ell_{\pi j } \ , \bigg\ { \delta^{\mu\nu ba } \ , \big ( \kappa^{(0)}_{\pi j } \ , \nabla_b \frac{\mu}{t } - \kappa^{(1)}_{\pi j } \ , d u_b \big ) + \frac{\eta\,t}{\ell_{\pi j } }
\ , \delta^{\mu\nu}_{\,\,\,\,\,\,bc } \ , \partial_d \big(\frac{\ell_{\pi j}}{\eta\,t } \ , \delta^{bcde}\big ) \ , \delta_e^{\,\,\,a } \bigg\ } \ , j_a\nonumber\\ & & { } - \frac{1}{2 } \ , \tau_\pi \ , \bigg\ { \delta^{\mu\nu ab } \ , \bigg [ \kappa^{(0)}_{\pi } \ , \nabla^c u_c + \frac{\eta\,t}{\tau_\pi } \ , \partial_c \big(\frac{\tau_\pi}{\eta\,t } \ , u^c\big ) \bigg ] - 4 \ , \kappa^{(1)}_\pi \ ,
\delta^{\mu\nu ce } \
, \delta_e^{\,\,\,dab } \ , \delta_{cd}^{\,\,\,\,\,\,fg } \
, \nabla_f u_g - 4 \ , \delta^{\mu\nu ce } \
, \delta_{e}^{\,\,\,dab } \ , \omega_{cd } \bigg\ } \ , \pi_{ab},\nonumber\\ \end{aligned}\ ] ] where @xmath167 is the vorticity .
\(b ) in the particle frame ( @xmath165 ) : @xmath168 - 2 \ ,
\kappa_j^{(1 ) } \ , \delta^{\mu abc } \ , \nabla_b u_c - 2 \ , \omega^{\mu a } \bigg\ } \ , j_a\nonumber\\ & & { } - \frac{1}{2 } \ , \ell_{j\pi } \ , \bigg\ { \delta^{\mu cab } \ , \big ( \kappa^{(0)}_{j\pi } \ , \nabla_c \frac{\mu}{t } - \kappa^{(1)}_{j\pi } \ , d u_c \big ) + \frac{\lambda\,t^{2}}{\ell_{j\pi } } \ , \delta^\mu_{\,\,\,c } \ , \partial_d \big(\frac{\ell_{j\pi}}{\lambda\,t^{2 } } \ , \delta^{cdef}\big ) \ , \delta_{ef}^{\,\,\,\,\,\,ab } \bigg\ } \ , \pi_{ab } , \end{aligned}\ ] ] and eq.([eq : ef_pi ] ) .
note that the effective bulk viscosity @xmath169 @xcite appears in eq.([eq : pf_pi ] ) .
here we have introduced the new coefficients , @xmath170 , @xmath171 , @xmath172 , @xmath173 , @xmath174 , @xmath175 , @xmath176 , @xmath177 , @xmath178 , @xmath179 , @xmath180 , @xmath181 , @xmath182 , @xmath183 , @xmath184 , @xmath185 , and @xmath186 , which are complicated functions of @xmath30 and @xmath31 @xcite .
we have found that the relaxation times are frame dependent except for @xmath137 while the transport coefficients such as the viscosities and the thermal conductivity are frame independent . for a demonstration of the frame - dependence of the relaxation times ,
we show in fig.[fig:1 ] the frame ( @xmath16 ) dependence of @xmath135 and @xmath136 , which tends to increase when the frame changes from the energy to particle frame . the @xmath16 dependence of @xmath135 and @xmath136 at @xmath187 and @xmath188 .
we normalized the relaxation times by the corresponding transport coefficients .
the energy and particle frames correspond to @xmath164 and @xmath189 . ]
the @xmath16 dependence of @xmath135 and @xmath136 at @xmath187 and @xmath188 .
we normalized the relaxation times by the corresponding transport coefficients .
the energy and particle frames correspond to @xmath164 and @xmath189 . ]
in summary , we have derived the second - order dissipative relativistic hydrodynamic equations in a generic frame with a continuous parameter @xmath16 ; the generic frame is reduced to the energy and particle frame with the parameter choice @xmath164 and @xmath189 , respectively .
a notable point of our result is that the dissipative part of the symmetric energy - momentum tensor @xmath4 in the _ particle _ frame satisfies the equality @xmath0 , in contrast to the usual choice @xmath1 , while @xmath4 of our derived equation in the _ energy _ frame satisfies the usual constraint @xmath190 .
we emphasize that this novel equality in the particle frame is a consequence of the derivation based on the renormalization - group method , a powerful method for the reduction of dynamical systems .
we note that the same constraints were also derived for the first - order dissipative relativistic hydrodynamic equation @xcite .
we have also shown that the phenomenological derivation based on the second law of thermodynamics allows that @xmath191 can be proportional to the bulk pressure @xmath35 and non - vanishing in the particle frame .
indeed , our microscopic derivation shows that @xmath192 . we have presented the relaxation equations in the energy and particle frames , explicitly as typical examples , although we have obtained the microscopic expressions for them in a more generic frame @xcite .
we have shown that the viscosities are frame - independent but the relaxation times are generically frame - dependent , as depicted in fig.[fig:1 ] .
the detailed derivation of the equations and discussions on the phenomenological consequences of the hydrodynamic equations thus obtained will be discussed in forthcoming papers @xcite .
we thank tetsu hirano for his interest in this work and discussions .
thanks dirk rischke for his interest in our work .
this work was partially supported by a grant - in - aid for scientific research by the ministry of education , culture , sports , science and technology ( mext ) of japan ( no.20540265 ) , by yukawa international program for quark - hadron sciences , and by the grant - in - aid for the global coe program `` the next generation of physics , spun from universality and emergence '' from mext .
k. tsumura and t. kunihiro ; talk presented at jps meeting at yamagata university , september 20 , 2008 ; t. kunihiro , y. minami and k. tsumura ; talk presented at quark matter 2009 , knoxville , march , 2009 . | we derive the second - order dissipative relativistic hydrodynamic equations in a generic frame with a continuous parameter from the relativistic boltzmann equation .
we present explicitly the relaxation terms in the energy and particle frames .
our results show that the viscosities are frame - independent but the relaxation times are generically frame - dependent .
we confirm that the dissipative part of the energy - momentum tensor in the particle frame satisfies @xmath0 obtained for the first - order equation before , in contrast to the eckart choice @xmath1 adopted as a matching condition in the literature .
we emphasize that the new constraint @xmath0 can be compatible with the phenomenological derivation of hydrodynamics based on the second law of thermodynamics . |
the replica method @xcite , originally devised as a trick to compute thermodynamical quantities of physical systems in presence of quenched disorder , has found applications in the analysis of systems of very different nature , as neural networks , combinatorial optimization problems @xcite , error correction codes @xcite etc . although many physicists believe that the method , within the replica symmetry breaking scheme of parisi @xcite , is able to potentially give the exact solution of any problem treatable as a mean field theory , the necessary mathematical foundation of the theory is still lacking , after more then 20 years from its introduction in theoretical physics .
the last times have seen a growing interest of the mathematical community in the method , leading to important but still partial results , confirming in certain cases the replica analysis , with more conventional and well established techniques @xcite .
apart the remarkable exception of the analysis of the fully connected @xmath0-spin model in ref . @xcite and the rigorous analysis of random energy models @xcite
, the analysis of the mathematicians has been , as far as we know , restricted to the high temperature regions and/or to problem of replica symmetric nature .
very welcomed have been the techniques recently introduced by guerra and toninelli @xcite which allow rigorous analysis not relying on the assumption of high temperature , and valid even in problems with replica symmetry breaking . along these lines ,
an important step towards the rigorous comprehension of the replica method , has been undertaken in @xcite , where it has been shown how in the case of the sherrington - kirkpatrick model , and its @xmath0-spin generalizations , the replica free - energies with arbitrary number of replica symmetry breaking steps constitute variational lower bounds to the true free - energy of the model .
as stated in that paper , the analysis is restricted to fully - connected models , whose replica mean field theory can be formulated in terms of a single @xmath1 matrix . however , in recent times , many of the more interesting problems analyzed with replica theory pertain to the so called `` diluted models '' where each degree of freedom interacts with a finite number of neighbors .
the introduction of a `` population dynamics algorithm '' @xcite has allowed to treat in full generality -within statistical precision- complicated sets of probabilistic functional equations appearing in the one step symmetry broken framework of diluted models .
the same algorithm has been used as a starting point of a generalized `` belief propagation '' algorithm for optimization problems @xcite .
furthermore , at the analytic level , simplifications due to graph homogeneities in some cases @xcite , and to the vanishing temperature limit in some other cases @xcite have led to supposedly exact solutions of the ground state properties of diluted models , culminated in the resolution of the random xor - sat on uniform graphs in @xcite and the random k - sat problem in @xcite within the framework of `` one - step replica symmetry breaking '' ( 1rsb ) .
the aim of this paper , is to show that the replica analysis of diluted models provides lower bounds for the exact free - energy density , and ground state energy density .
we analyze in detail the cases of the diluted @xmath0-spin model on the poissonian degree hyper - graphs also known as random xor - sat problem and the random k - sat problems .
we expect that along similar lines free - energy lower bounds can be found for many other diluted cases .
the guerra method we use sheds some light on the meaning of the replica mean field theory . the physical idea behind the method is that within mean field theory one can modify the original hamiltonian weakening the strength of the interaction couplings or removing them partially or totally , and compensate this removal by some auxiliary external fields . in disordered systems
these fields should be random fields , taken from appropriate probability distributions and possibly correlated with the original values of the quenched variables eliminated from the systems .
one is then led to consider hamiltonians interpolating between the original model and a pure paramagnet in a random field , and by means of these models achieving free - energy lower bounds .
we will see that the rs case corresponds to assuming independence between the random fields and the quenched disorder .
the parisi rsb scheme , assumes at each breaking level a peculiar kind of correlations , and gives free - energy bounds improving the rs one .
our paper is organized in this way : in section 2 we introduce some notations that will be extensively used in the following sections . in section 3
we introduce the general strategy to get the replica bounds ; we then specialize to the replica symmetric and the one step replica symmetry broken bounds , giving the results in the @xmath0-spin and the @xmath2-sat cases .
conclusions are drawn in section 4 . in the appendices some details of the calculations in both the @xmath0-spin and
the @xmath2-sat cases are shown .
our results will be issue of explicit calculations .
although at the end we will get bounds , formalizable as mathematical theorems , the style and most of the notations of the paper will be the ones of theoretical physics .
the spin models we will consider in this work are defined by a collection of @xmath3 ising @xmath4 spins @xmath5 , interacting through hamiltonians of the kind @xmath6 where the indices @xmath7 are i.i.d
. quenched random variables chosen uniformly in @xmath8 .
we will call each term @xmath9 a clause .
the subscript @xmath10 in the clauses indicates the dependence on a single or a set of quenched random variables , as it will be soon clear .
the number of clauses @xmath11 will be taken to be proportional to @xmath3 .
for convenience we will choose it to be for each sample a poissonian number with distribution @xmath12 .
the fluctuations of @xmath11 will not affect the free - energy in the thermodynamic limit , and this choice , which slightly simplify the analysis , will be equivalent to choosing a fixed value of @xmath11 equal to @xmath13 .
the clauses themselves will be random .
the @xmath0-spin model@xcite has clauses of the form @xmath14 this form reduces to @xmath15 in the case of the viana - bray spin glass @xmath16 . in both cases the @xmath17
will be taken as i.i.d .
random variable with regular symmetric distribution @xmath18 .
notice that for @xmath19 $ ] the model reduces to the random xor - sat problem @xcite of computer science .
the random k - sat clauses have the form @xcite @xmath20 where the @xmath21 are i.i.d . with symmetric probability .
( the number @xmath0 of spin appearing in a clause is usually called @xmath2 in the k - sat problem , for uniformity of notation we will deviate from this convention ) .
notice that in all cases , on average each spin participate to @xmath22 clauses , and that the set of spins and interactions defines a random diluted hyper - graph of uniform rank @xmath0 and random local degree with poissonian statistics in the thermodynamic limit . at high enough temperature ,
the existence of the free - energy in the thermodynamic limit for models of this kind has been proved in by talagrand in @xcite , together with the validity of the rs solution . a proof valid at all temperature based on the ideas presented in this paper ,
can be obtained for even @xmath0 in analogy of the analysis in @xcite for long range models .
we sketch it in appendix c in the case of the @xmath0-spin model . in establishing the free - energy bounds we will need several kind of averages : * the boltzmann - gibbs average for fixed quenched disorder : given an observable @xmath23 @xmath24 where @xmath25 and @xmath26 is the inverse temperature .
+ obviously , @xmath27 , as well as @xmath28 will be functions of the quenched variables , the size of the system and the temperature . this dependence will be made explicit only when needed . *
the disorder average : given an observable quantity @xmath29 dependent on the quenched variables appearing in the hamiltonian , we will denote as @xmath30 its average .
this will include the average with respect to the @xmath31 variables and the choice of the random indices in the clauses as well as with respect to other quenched variables to be introduced later .
* we will need in several occasion the `` replica measure '' @xmath32 and some generalizations that we will specify later .
* we will occasionally use other kinds of averages , as well as other notations , for which we will use an angular bracket notation , with a subscript indicating the variable(s ) over which the average is performed .
e.g. an average over a random variable @xmath33 with probability distribution @xmath34 will be denoted equivalently as @xmath35 .
analogously , averages over distribution families of @xmath36 will be denoted as @xmath37 .
subscripts will be omitted whenever confusion is not possible . *
another notation we will have the occasion to use in the one for the overlaps among @xmath38 spin configurations @xmath39 , out of a population of @xmath40 @xmath41 : @xmath42 and in particular @xmath43 this notation will be extended to multi - overlaps in the 1rsb case , as we will specify in section [ tarta ] . in the following
we will need to consider averages where some of the variables are excluded , e.g. the averages when a variable @xmath44 is erased .
these average will be denoted with a subscript @xmath45 e.g. if an @xmath46 average is concerned the notation will be @xmath47 .
other notations will be defined later in the text whenever needed .
our interest will be confined to bounds to the free - energy density @xmath48 and the ground state energy density @xmath49 $ ] valid in the thermodynamic limit , so that @xmath50 will be often implicitly neglected in our calculations .
the strategy to get the replica bound is a generalization of the one introduced by guerra in the case of fully connected models @xcite
. we will consider models which will interpolate between the original ones we want to analyze and pure paramagnet in random fields with suitably chosen distribution .
the underlying idea is that , given the mean field nature of the models involved , if one was able to reconstruct the real local fields acting on a given spin variable via a given hyper - edge , and to introduce auxiliary fields acting on that variable in such a way to energetically balance the deletion of the hyper - edge , then it would be possible to have an exact expression for the free - energy in terms of such auxiliary fields even when the whole edge set was emptied .
however , if the replacement is done with some approximate form of the auxiliary fields distribution function , the real free - energy will be the one calculated using the approximate fields plus an excess term at every step of the graph deletion process .
the proof of the definite sign of this excess term gives a way to determine bounds for the thermodynamic quantities .
we will prove the existence of replica lower bounds to the free - energy density of the p - spin model and the random k - sat problem . in this last case
our result proves that the recent replica solution of @xcite gives a lower bound to the ground state energy and therefore an upper bound for the satisfiability threshold .
the proofs will strictly hold in the @xmath51 limit , due to the presence of corrections of order @xmath52 in the calculated expressions for any finite size graph .
moreover , our proofs will be restricted to the @xmath0-spin model the the k - sat with even @xmath0 . in the cases of odd @xmath0
the same bound would hold if one could rely on some physically reasonable assumptions on the overlap distribution ( see below ) .
our analysis will start from the tap equations for the models @xcite , and their probabilistic solutions implied by the cavity , or equivalently the replica method at various degrees of approximation .
we will consider in particular the replica symmetric ( rs ) and one step replica symmetry broken solutions , but it should be clear from our analysis how to generalize to more steps of replica symmetry breaking . in the tap / cavity equations one singles out the contribution of the clauses and the sites to the free - energy and defines cavity fields @xmath53 and @xmath54
respectively as the local field acting on the spin @xmath55 in absence of the clause @xmath56 and the local field acting on @xmath55 due to the presence of the clause @xmath56 only .
if we define @xmath57 $ ] as the partition function of a given sample with @xmath3 spins where all but the spin @xmath55 are integrated , @xmath58 the free - energy of the corresponding systems where the spin @xmath59 and all the clauses to which it belongs are removed , we can write , @xmath60 & = & \e^{-\beta f_{n ,- i}}\prod_{\mu\in t_i } \sum_{s_{i_2^\mu}, ... ,s_{i_p^\mu } } \e^{-\beta h_{j^{(\mu)}}(s_{i^\mu},s_{i_2^\mu}, ... ,s_{i_p^\mu})+\sum_{l=2}^{p } h_{i_l^\mu}^{(\mu)}s_{i_l^\mu } } \nonumber\\ & = & \e^{-\beta f_{n ,- i } } \prod_{\mu\in t_i } b_{\mu}^{(i ) } \e^{\beta u_\mu^{(i)}s_i } \label{c1}\end{aligned}\ ] ] where @xmath61 is the set of clauses containing the spin @xmath55 , and the constant @xmath62 can be interpreted as suitable shifts in the free - energy due to the contribution of the clause @xmath56 for fixed value of the spin @xmath55 .
we notice that denoting @xmath63 as @xmath31 , and renaming the fields in ( [ c1 ] ) into @xmath64 , eq .
( [ c1 ] ) defines functions @xmath65 the equation are closed by the self - consistent condition : @xmath66 these equations are at the basis of iterative algorithms such as the `` belief propagation '' or `` sum - product '' know for a long time in statistical inference @xcite and coding theory @xcite and the more recently proposed algorithm of `` survey propagation''@xcite . conditions ( [ c1 ] ) and ( [ c2 ] ) can be diagrammatically represented as in fig.([cavity ] ) .
= 0.4 the cavity fields solutions of ( [ c1],[c2 ] ) are random variables which fluctuate for two reasons @xcite . first , they differ from sample to sample .
second , within the same sample the equations can have several solutions which can level - cross .
the cavity / replica method provides under certain assumption probabilistic solutions . in the rs approximation
, one just supposes a single solution to give the relevant contribution in a given sample .
the sample to sample fluctuation induce probability distributions @xmath67 and @xmath36 whose relations implied by ( [ c1],[c2 ] ) are : @xmath68 where @xmath69 denotes the average over the random variables appearing in a clause .
in addition to sample to sample fluctuations , the 1rsb solution assumes fluctuations of the fields from solution to solution of the equations , so that the functions @xmath67 and @xmath36 will be themselves randomly distributed according to some functional probability distributions @xmath70 and @xmath71 related by the self - consistency equations @xcite @xmath72 @xmath73 where : @xmath74 \int dh_1\ ; p_1(h_1 ) ... dh_{p-1}\ ; p_1(h_{p-1 } ) b_{\bf j}(h_1, ...
,h_{p-1})^m \delta ( u - u_{\bf j}(h_1, ... ,h_{p-1 } ) ) \\ & & p(h|{q_1}, ... ,{q_k } ) = { \cal n}_{q
, k}[{q_1}, ... ,{q_k } ] \ , \ , ( 2\cosh(\beta h))^m \int \prod_{l=1}^k d u_l \ ; \frac { { q_l } ( u_l)}{(2\cosh(\beta u_l))^m } \delta ( h - \sum_{l=1}^k u_l ) \label{innerprobs1}\end{aligned}\ ] ] where @xmath75 $ ] and @xmath76 $ ] insure normalization and @xmath77 is a rescaling term of the form ( [ ub ] ) that can be reabsorbed in the normalization in the case of the @xmath0-spin model .
its form for the @xmath2-sat case is given in the appendix .
@xmath78 is a number in the interval @xmath79 $ ] , which within the formalism selects families of solutions at different free - energy levels .
the physical free - energy is estimated maximizing over @xmath78 .
the interpretation of these equations has been discussed many times in the literature @xcite .
we will show here , that such choices in the field distributions result in lower bounds for the free - energy analogous to the ones first proved by guerra in fully connected models . in order to prove these bounds
, we will have to consider auxiliary models where the number of clauses @xmath13 will be reduced to @xmath80 ( @xmath81 ) , while this reduction will be compensated in average by some external field terms of the kind : @xmath82 where the numbers @xmath83 will be i.i.d .
poissonian variables with average @xmath84 . as the notation suggests , the fields @xmath85 will play the role of the cavity fields @xmath54 of the tap approach , and they will be i.i.d
. random variables with suitable distribution .
indeed , for each field @xmath86 we will chose in an independent way @xmath87 primary fields @xmath88 ( @xmath89 ) and clause variables @xmath90 such that the relation @xmath91 is verified .
notice that the compound hamiltonian @xmath92={\cal h}^{(\alpha t)}[{\bf s}]+{\cal h}^{(t)}_{ext}[{\bf s } ] \label{compoundh}\ ] ] will constitute a sample with the original distribution for @xmath93 , while it will consist in a system of non interacting spins for @xmath94 .
the key step of the procedure , consists in the choice of the distribution of the primary fields @xmath95
. we will also find useful to define fields @xmath96 verifying @xmath97 the field @xmath33 are related to the @xmath98 s by a relation similar to ( [ c1 ] ) , while the @xmath99 s are related to the @xmath33 s by a relation similar to ( [ c2 ] ) .
of course , the statistics of the fields @xmath99 and the @xmath98 s do coincide in the tap approach .
it is interesting to note that the bounds we will get , are optimized precisely when their statistical ensemble coincide .
as we mentioned , various replica bounds are obtained assuming for the fields @xmath95 the type of statistics implied by the different replica solution .
so , the replica symmetric bound is got just supposing the field as quenched variables completely independent of the quenched disorder and with distribution @xmath100 . for the one - step rsb bound on the other hand the distribution @xmath101
will itself be considered as random , subject to a functional probability distribution @xmath102 $ ] .
more complicated rsb estimates , not considered in this paper , can be obtained along the same lines .
the case of the fully connected models considered by guerra can be formalized in this way where the various field distributions involved are gaussian .
we consider in this case i.i.d .
fields @xmath33 and @xmath99 distributed according probabilities @xmath36 and @xmath67 verifying the following relation with the distribution @xmath103 of the primary fields . @xmath104
the distribution @xmath100 will be chosen to be symmetric under change of sign of @xmath98 , and regular enough for all the expression below to make sense .
the rs bound can now be obtained following a procedure to the one of guerra for the sk model , and considering the @xmath105 dependent free - energy ; with obvious notation : @xmath106 where @xmath107 represents the average over all the quenched variables , the one defining the clauses and the external fields .
we then consider the @xmath105 derivative of @xmath108 @xmath109 as in @xcite we will then write @xmath110 and show , by an explicit computation , that , up to @xmath50 terms that will be systematically neglected , the expression coincides with the variational rs free - energy plus a remainder . in fortunate cases
this term will have negative sign and neglecting it will immediately result in a lower bound for the free - energy .
this happens in the viana - bray model , the @xmath0-spin and the k - sat for even @xmath0 . in the cases of odd @xmath0 we were not able to prove the sign definiteness of the remainder ,
although as we will discuss we believe this to be the case on a physical basis .
the time derivative of @xmath111 take contributions from the derivative of the distribution of the number of clauses @xmath11 @xmath112 and the distribution of the number of @xmath33 fields on each site @xmath113 so that : @xmath114 where we have denoted as @xmath115 the average with respect to all the quenched variables except @xmath11 and with @xmath116 the average with respect to all the quenched variables except @xmath83 , and simply @xmath117 the partition function of the @xmath3 spin system @xmath118 . in the first term of ( [ der ] ) we can single out the @xmath11-th clause , and write @xmath119 , where by @xmath120 we denote the partition function of the system in absence of the @xmath11-th clause , and @xmath121 is the canonical average in absence of the @xmath11-th clause . in the following terms we single out the @xmath83-th field @xmath33 term , @xmath122 , where @xmath123 is the partition function in absence of the field @xmath124 and analogously for the average @xmath47 . finally ,
rearranging all terms we find @xmath125 \nonumber \\ & & - p \alpha \sum_i\sum_{k_i } \pi(k_i-1,\alpha p(1-t ) ) e''_i\log[\omega(\e^{\beta u_i^{k_i}s_i})_{-u_i^{k_i } } ] .
\label{pippo}\end{aligned}\ ] ] where we have used @xmath126 .
we notice at this point that the statistical ensemble defined by @xmath127 can be substituted with the original one @xmath107 and the average of the variables appearing in the clause we have singled out .
to be more precise , we remark that the average @xmath128 depends on the quenched variables @xmath129 appearing in the hamiltonian .
writing explicitly this dependence as @xmath130 , and denoting as @xmath131 all the quenched variables except the ones appearing in the @xmath11-th clause , our statement is that thanks to the poissonian distribution of @xmath11 and the uniform choice of the indices of each clause , @xmath132= e\left ( \frac{1}{n^p } \sum_{i_1, ... ,i_p } \langle \log [ \omega(\e^{-\beta h_j(s_{i_1}, ... ,s_{i_p})}|d)]\rangle_j \right).\ ] ] where by @xmath133 we denote the average with respect to the random variables appearing in the clause .
this is a crucial step in our analysis , in fact , similar considerations apply to the term in the second line of ( [ pippo ] ) , which can be written as @xmath134 = & & e \left\langle \log \omega \left(e^{\beta u s_i}\right ) \right\rangle_u \ ; .\end{aligned}\ ] ] the same kind of averages @xmath107 and @xmath46 appear in the two terms which can be therefore directly compared as we will do in the next section .
this property , linked to the poissonian character of the graph defined by the model would not hold for other ensembles of random graphs and the analysis would be technically more involved . substituting in ( [ pippo ] )
we find : @xmath135\rangle_h -\frac{p}{n}\sum_i
\langle \log \omega ( e^{\beta u s_i } ) \rangle_u \right ] \label{pippo2}\ ] ] rearranging terms and using ( [ quattordici ] ) we finally find that the free - energy @xmath108 can be written as @xmath136 + \int_0 ^ 1 dt\ ; r_{rs}[g , t]+o(1/n ) \label{fverars}\ ] ] where @xmath137 $ ] coincides the expression of the variational free - energy in the replica treatment under condition @xmath138 = p[h]$ ] @xmath139 at @xmath94 and @xmath140 $ ] is a remainder term . instead of writing the formulae for general clauses , in order to keep the notations within reasonable simplicity , we specialize now to the specific cases of the @xmath0-spin model and the k - sat .
notice that in all models @xmath141 = -\frac{1}{\beta } \langle \log ( 2 \cosh ( \beta h ) ) \rangle_h |_{t=0}\ ] ] in the case of the @xmath0-spin @xmath142 . substituting in eq.([pippo2 ] ) and rearranging terms
one immediately finds : @xmath143 & = & \frac{1}{\beta } \left [ \alpha \left(p \left\langle \log ( \cosh \beta u ) \right\rangle_u - \left\langle \log ( \cosh \beta j ) \right\rangle_j \right ) - \left\langle \log ( 2 \cosh \beta h ) \right\rangle_h + \right .
\nonumber \\ & & \left .
\alpha ( p-1 ) \left\langle \log \left ( 1 + \tanh ( \beta j)\prod_{t=1}^p \tanh ( \beta g_t ) \right ) \right\rangle_{\{g_t\ } , j } \right]\end{aligned}\ ] ] while the remainder is the @xmath105 integral of @xmath144= & & -\frac{\alpha}{\beta } \left[\frac{1}{n^p } \sum_{i_1, ...
,i_p } e \left\langle \log(1+\tanh(\beta j)\omega(s_{i_1} ... s_{i_p } ) ) \right\rangle_j -p e \left\langle \log(1+\tanh ( \beta u)\omega ( s_i ) ) \right\rangle_u + \nonumber \right . \\ & & \left .
( p-1 ) e \left\langle \log(1+\tanh ( \beta j)\prod_{t=1}^p \tanh ( \beta g_p ) ) \right\rangle_{\{g_t\},j } \right].\end{aligned}\ ] ] the expression for @xmath145 $ ] coincides with the rs free energy once extremized over the variational space of probability distributions @xmath101 @xcite .
terms have been properly added and subtracted in order to get a remainder which equal to zero if maximization over @xmath101 is taken , and the temperature is high enough for replica symmetry to be exact @xcite . as we will see , the remainder turns out to be positive .
@xmath145 $ ] is therefore , for all @xmath101 for which its expression makes sense , a lower bound to the free - energy . at saturation
@xmath146 = p[h ] |_{t=0 } \ ; \forall \ ; h\ ] ] should hold , which is simply the self - consistency rs equation . by using equation @xmath147
we can establish that the remainder is positive for even @xmath0 .
we expand the logarithm of the three terms in ( absolutely converging ) series of @xmath148 , and notice that thanks to the parity of the @xmath31 and the @xmath98 distributions , they will just involve negative terms .
we can then take the expected value of each terms and write @xmath149= \frac 1 \beta \sum_{n=0}^{\infty } \langle \tanh^{2 n } \beta j \rangle_{j } \frac{1}{n } \omega \left[(q^{(2n)})^p - p q^{(2n ) } \langle \tanh ^{2 n } \beta g \rangle_g^{p-1 } + ( p-1 ) \langle
\tanh ^{2 n } \beta g \rangle_g^p \right ] \label{sser}\ ] ] where we have introduced the overlap @xmath150 and the replica measure @xmath151 defined in section 2 .
the series in ( [ sser ] ) is an average of positive terms in the case of the viana - bray model @xmath16 , where we get perfect squares , and more in general for all even @xmath0 , as we can easily , starting from the observation that in this case @xmath152 is positive or zero for all @xmath153 , @xmath154 real . in the case of
@xmath0 odd , the same term is positive only if @xmath155 is itself positive or zero .
the bound of the free - energy would therefore be established if we were able to prove that the probability distributions of the @xmath156 has support on the positives . , but it is not clear its physical meaning . ] this property , which tells that anti - correlated states are not possible , is physically very sound whenever the hamiltonian is not symmetric under change of sign of all spins .
in fact , one expects the probability of negative values of the overlaps to be exponentially small in the size of the system for large @xmath3 .
unfortunately however we have not been able to prove this property in full generality .
notice that upon maximization on @xmath101 , the results of @xcite imply that the remainder is exactly equal to zero if the temperature is high enough for replica symmetry to hold . in the case of the @xmath2-sat , using def.([sat ] ) for the clause @xmath157 , we find relation : @xmath158 \ ; , \label{u - ksat}\ ] ] where @xmath159 . via direct inspection , the variational free - energy coincides with the rs expression @xcite @xmath160 & = & \frac{1}{\beta } \left [ \alpha ( p-1 ) \left\langle
\log \left ( 1 + ( e^{-\beta } -1)\prod_{t=1}^p \left ( \frac{1 + \tanh ( \beta g_t)}{2 } \right ) \right ) \right\rangle_{\{g_t\},\{j_t\ } } - \langle \log ( 2\cosh ( \beta h ) ) \rangle_h + \right .
\nonumber \\ & & \left .
\alpha p \langle \log ( 2\cosh ( \beta u ) ) \rangle_u - \alpha p \left\langle \log \left ( 1 + \frac{(e^{-\beta } -1)}{2}\prod_{t=1}^{p-1 } \left ( \frac{1 + \tanh ( \beta g_t)}{2 } \right ) \right ) \right\rangle_{\{g_t\},\{j_t\ } } \right ] \label{frsksat}\end{aligned}\ ] ] while the remainder is the @xmath105 integral of @xmath161 & = & -\frac{\alpha}{\beta } e \left [ \frac{1}{n^p } \sum_{i_1, ... ,i_p } \left\langle \log \left ( 1+(\e^{-\beta}-1)\omega(\prod_{t=1}^p \frac{1+j_{t } s_{i_t}}{2 } ) \right ) \right\rangle_{\{j_t\ } } - \right .
\nonumber \\ & & \frac{p}{n } \sum_i \left\langle \log \left ( 1 + \xi \omega\left(\frac{1+j s_i}{2 } \prod_{t=1}^{p-1}\frac{1+j_{t } \tanh ( \beta g_t)}{2 } \right ) \right ) \right\rangle_{\{g_t\ } , j , \{j_t\ } } + \nonumber \\ & & \left .
( p-1 ) \left\langle \log\left ( 1 + \xi \prod_{t=1}^{p}\frac{1+j_{r } \tanh ( \beta g_t)}{2}\right ) \right\rangle_{\{g_t\ } , \{j_t\ } } \right ] \ ; . \label{restoksat } \end{aligned}\ ] ] considerations analogous to the case of the @xmath0-spin , have led us to add and subtract terms from eq.([pippo2 ] ) to single out the proper remainder term .
expanding in series the logarithms , exploiting the symmetry of the probabilities distribution functions and taking the expectation of each term of the absolutely convergent series we finally obtain : @xmath162 = \frac{\alpha}{\beta } \sum_{n\ge 1 } \frac{(-1)^{n}}{n } ( \xi^*)^n \omega \left [ ( 1+q_n)^p - p(1+q_n)\langle(1+j \tanh ( \beta g))^n\rangle_{j , g}^{p-1 } + ( p-1 ) \langle(1+j \tanh ( \beta g))^n\rangle_{j , g}^{p } \right ] \label{resto2ksat}\ ] ] where we have defined @xmath163 and @xmath164 .
detailed calculations are given in the appendix .
as in the @xmath0-spin case , the previous sum is obviously positive for @xmath0 even . for @xmath0 odd we should again rely on the physical wisdom that all @xmath165 have positive support and so have the functions @xmath166
again , the variational free - energy coincides with the rs expression once extremized over @xmath101 at the condition @xmath167 at @xmath94 .
we establish here a more complex estimate , in a larger variational space of functional probability distributions
. the general strategy will be here to consider the same form for the auxiliary hamiltonian , but now with a more involved choice for the fields distribution .
the fields on different sites or different index @xmath168 will be still independent , but each site field distribution @xmath169 will be itself random i.i.d .
, chosen with a probability density functional @xmath102 $ ] , with support on symmetric distributions @xmath170
. it will be assumed that @xmath171 is such that all the expressions below make sense . in this case
, the variational approximation for the free - energy will be obtained from an estimate of @xmath172 = \frac{1}{mn } e_1 \log e_2(z^m(t))\ ] ] where we have denoted with : * @xmath173 the average w.r.t . @xmath88 for fixed distributions @xmath174 according to the measure @xmath175 where @xmath176 ensures the normalization . * @xmath177 the average with respect to the quenched clause variable , distributions the @xmath178 s and the poissonian variables @xmath83 s , which will be i.i.d . with probabilities @xmath179 , @xmath180 and @xmath181 respectively .
the number @xmath78 is real in the interval ( 0,1 ] .
the statistical ensemble of the auxiliary fields @xmath33 and @xmath99 will be now related to the one of the @xmath98 by : @xmath182 @xmath183 where : @xmath184\int dg_1\ ; g_1(g_1 ) ... dg_{p-1}\ ; g_1(g_{p-1 } )
b_j(g_1, ... ,g_{p-1})^m \delta ( u - u_j(g_1, ... ,g_{p-1 } ) ) \\ & & g(g|{q_1}, ... ,{q_k } ) = { \cal n}_{q ,
k}[{q_1}, ... ,{q_k } ] \ , \ , ( 2\cosh(\beta g))^m \int \prod_{l=1}^k d u_l \ ; \frac { { q_l } ( u_l)}{(2\cosh(\beta u_l))^m } \delta ( g - \sum_{l=1}^k u_l ) \label{innerprobs}\end{aligned}\ ] ] where
@xmath75 $ ] , @xmath76 $ ] and @xmath185 have been previously defined . with notations similar to the ones of the rs case ,
we can write @xmath186)= & & -\alpha \sum_m ( \pi(m,\alpha t n)-\pi(m-1,\alpha t n ) ) e'_1 \frac { 1}{nm } \log e_2 z(t)^m\nonumber \\ & & + \frac{\alpha p}{n } \sum_i \sum_{k_i } ( \pi(k_i,\alpha p(1-t))-\pi(k_i-1,\alpha p(1-t ) ) ) e''_{1,i }
\frac { 1}{nm } \log e_2 z(t)^m \label{der - rsb}\end{aligned}\ ] ] which , extracting explicitly the contribution from the @xmath11-th close in the first term and the @xmath83-th field @xmath33 in the second , following considerations similar to the rs case we find : @xmath186)= & & \alpha \sum_m ( \pi(m-1,\alpha t n ) ) \frac{1}{m}e'_1 \log \left [ \frac { e_2 z_{-m}^m
\omega(\e^{-\beta h_{j^{(\mu)}}(s_{i_1^m}, ... ,s_{i_p^m})})_{-m}^m } { e_2 z_{-m}^m } \right]\nonumber \\ & & -\frac{p \alpha}{n } \sum_i\sum_{k_i } \pi(k_i-1,\alpha p(1-t ) ) \frac{1}{m}e''_{1,i } \log \left [ \frac { e_2 z_{-u_i^{k_i}}^m \omega(\e^{\beta u_i^{k_i}s_i})_{-u_i^{k_i}}^m } { e_2 z_{-u_i^{k_i}}^m } \right ] . \label{pippo3}\end{aligned}\ ] ] again it can be recognized that the primed averages coincide with the averages over the original ensembles plus the averages on the variables appearing in the terms we extracted . finally we get : @xmath187)=\frac{\alpha}{m } e_1 \left [ \frac{1}{n^p}\sum_{i_1, ... ,i_p }
\left\langle \log \left ( \frac{e_2 z^m \omega ( e^{-\beta h_j(s_{i_1}, ...
,s_{i_p})})^m}{e_2 z^m } \right ) \right\rangle_j - \frac{p}{n}\sum_i \left\langle \log \left ( \frac{e_2 z^m \langle \omega ( e^{\beta u s_i})^m \rangle_u}{e_2 z^m }
\right ) \right\rangle_q \right ] \ ; . \label{pippo4}\ ] ] rearranging all terms one finds the estimate : @xmath188 + \int_0 ^ 1 dt\ ; r_{1rsb}[{\cal g},t]+o(1/n)\ ] ] where this time @xmath189 $ ] coincides with @xmath190 $ ] , the expression of the variational free - energy in the 1rsb treatment at the saddle point @xmath191 at @xmath94 , and @xmath192 $ ] is the remainder .
notice that the derivation immediately suggests how to generalize the analysis to more steps of replica symmetry breaking .
let us now specialize the formulae for the @xmath0-spin model and the k - sat .
again , in this case we will need the expression for @xmath193 $ ] : @xmath141 = \frac{1}{\beta m } \left [ \left\langle \log \left\langle \left ( \frac{1}{2 \cosh ( \beta h ) } \right)^m \right\rangle_h \right\rangle_p \right]_{|_{t=0 } } \ ; .\ ] ] in this case , plugging def.([spin ] ) in eq.([pippo4 ] ) rearranging , adding and subtracting terms one finds : @xmath194 & = & \frac{1}{\beta m } \left [ \left\langle \log \left\langle \left ( \frac{1}{2 \cosh ( \beta h ) } \right)^m \right\rangle_h \right\rangle_p - \alpha m \left\langle \log ( 2 \cosh ( \beta j ) ) \right\rangle_j - \alpha p \left\langle \log \left\langle \left ( \frac{1}{2 \cosh ( \beta u)}\right)^m \right\rangle_u \right\rangle_q \right .
\nonumber \\ & & \left .
+ \alpha ( p-1 ) \left\langle \log \left\langle \left ( 1 + \tanh ( \beta j ) \tanh ( \beta g_1 ) ...
\tanh ( \beta g_p ) \right)^m \right\rangle_{g_1, ... ,g_p } \right\rangle_{g_1, ... ,g_p ; j } \right ] \label{1rsb - variational - spin}\end{aligned}\ ] ] while the remainder is the @xmath105 integral of @xmath195= & & -\frac{\alpha}{\beta m } e_1 \left [ \frac{1}{n^p } \sum_{i_1, ... ,i_p } \left\langle \log \left ( \frac{e_2 z^m ( 1 + \omega ( s_{i_1} ... s_{i_p})\tanh ( \beta j))^m}{e_2 z^m } \right ) \right\rangle_j - \right .
\nonumber \\ & & \frac{p}{n } \sum_i \left\langle \log \left ( \frac{e_2 z^m \left\langle ( 1 + \omega(s_i)\tanh ( \beta u ) ) ^m
\right\rangle_u}{e_2 z^m } \right ) \right\rangle_q + \nonumber \\ & & \left .
( p-1 ) \left\langle \log \left\langle
\left ( 1 + \tanh ( \beta j ) \tanh ( \beta g_1 ) ...
\tanh ( \beta g_p ) \right)^m \right\rangle_{g_1, ... ,g_p } \right\rangle_{g_1, ... ,g_p ; j } \right ] \label{pippospin}\end{aligned}\ ] ] the expression for @xmath196 $ ] coincides with the 1rsb free - energy @xcite once maximized over the variational space of probability distribution functionals @xmath171 .
the maximization condition reads : @xmath197 = { \cal p}[p ] \ ; |_{t=0 } \ ; \forall \ ; p \ ; , \label{spersb}\ ] ] which is simply the self consistency 1rsb condition . for even @xmath0 ( and in particular for @xmath16 that corresponds to the viana - bray case )
, one can check that the remainder is positive just expanding the logarithm in series and exploiting the parity of the @xmath31 and the @xmath98 distributions . as
this is considerably more involved then in the rs case , we relegate this check to appendix a. in the @xmath2-sat case the expression for function @xmath198 reads : @xmath199 while the corresponding one for @xmath200 is the same as in the rs case .
the corresponding replica free - energy and remainder read @xmath201 & = & \frac{1}{m\beta } \left [ \alpha ( p-1 ) \left\langle
\log \left\langle \left ( 1 + \xi \prod_{t=1}^p \left ( \frac{1 + j_t \tanh ( \beta g_t)}{2 } \right ) \right)^m \right\rangle_{\ { g_t\ } } \right\rangle_{\{g_t\},\{j_t\ } } \right .
- \nonumber \\ & & \left .
\alpha p \left\langle \log \left\langle \left ( \frac{b(\{j_t\},\{g_t\})}{2 \cosh ( \beta u_j(\{j_t\},\{g_t\ } ) ) } \right)^m \right\rangle_{\{g_t\ } } \right\rangle_{\{g_t\},\{j_t\},j } + \left\langle \log \left\langle \left ( \frac{1}{2 \cosh ( \beta h ) } \right)^m \right\rangle_h \right\rangle_p \right ] \label{liuto}\end{aligned}\ ] ] the remainder is the @xmath105 integral of @xmath202 & = & -\frac{\alpha}{\beta m } e_1 \left [ \frac{1}{n^p } \sum_{i_1, ... ,i_p } \left\langle \log \left ( \frac{e_2 z^m \left ( 1 + \xi \omega \left ( \prod_{t=1}^p \frac{1 + j_t s_{i_t}}{2 } \right ) \right)^m}{e_2 z^m } \right ) \right\rangle_{\{j_t\ } } \right .
- \nonumber \\ & & \frac{p}{n } \sum_i \left\langle \log \left ( \frac{e_2 z^m \left\langle \left ( 1 + \xi \frac{1 + j \omega ( s_i)}{2 } \prod_{t=1}^{p-1 } \frac{1 + j_t \tanh ( \beta g_t)}{2 } \right)^m \right\rangle_{\ { g_t\}}}{e_2 z^m } \right ) \right\rangle_{\ { g_t\},\{j_t\},j } + \nonumber \\ & & \left . ( p-1 ) \left\langle \log \left\langle \left ( 1 + \xi \prod_{t=1}^p \left ( \frac{1 + j_t \tanh ( \beta g_t)}{2 } \right ) \right)^m \right\rangle_{\ { g_t\ } } \right\rangle_{\{g_t\},\{j_t\ } } \right ] \label{pipposat}\end{aligned}\ ] ] the expression for @xmath203 $ ] coincides with the 1rsb free energy once extremized under condition ( [ spersb ] ) , with the corresponding @xmath2-sat probability distribution functionals . notice that the proof of the positivity of ( [ pipposat ] ) for even @xmath0 is again dove via series expansion , all the detail are explained in appendix b. at this point we can take the zero temperature limit , finding that the resulting expression gives us a lower bound for the ground - state energy of the system , i.e. the minimal number of unsatisfied clauses .
notice that the @xmath204 limit of the replica free - energy is not trivial .
the necessary assumptions on the field distributions to get it correct are well known in the physical literature , and have been recently reviewed in @xcite .
recently mezard , parisi and zecchina @xcite have worked out the k - sat 1rsb solution for @xmath205 predicting a non zero ground - state energy for values of @xmath206 above a satisfiability threshold of @xmath207 , very well in agreement with the numerical simulations .
our results , together with the additional hypothesis of positivity of the support of the overlap functions imply that this value is an upper bound to the true threshold .
in this paper we have established that the free - energy of some families of diluted random spin models can be written as the sum of a term identical to the ones got in the cavity / replica plus an error term .
both the replica term and the remainder are different in different replica scheme , corresponding to the choice of statistical ensemble of the cavity fields .
we believe that the sign of the remainder is in general negative in the model we have considered , although we have been able to prove that only in the case of even @xmath0 . for odd @xmath0 our belief is supported by the physical wisdom that the overlap distributions are supported on the positives in the large @xmath3 limit .
we have considered the cases of replica symmetry and one step of replica symmetry breaking .
it is clear that the analysis could be extended to further levels of replica symmetry breaking , although the complexity of the analysis would greatly increase .
the 1rsb level is thought to give the exact scheme to treat the @xmath0 spin model and the k - sat problem for @xmath208 .
for the viana - bray model on the other hand it is believed that no finite rsb scheme furnish the exact solution , and one needs to consider the limit of infinite number of replica symmetry breaking .
it is not clear to us how to generalize the analysis to this case .
our analysis of the diluted models underlines a strong link between the guerra method and the cavity method which remained rather hidden in the fully connected case . in the cavity approach one considers incomplete graphs in which either sites or clauses are removed from the complete graph .
then , with the aid of precise physical hypothesis , consistency equations are written that allow to compute the free - energy from the comparison between the site and clause contributions . in the approach presented in this paper the removal of clauses
is compensated in average by the addition of some external fields which have precisely the statistics which is assumed with cavity .
the novelty of the approach is that it gives some control on the approximation involved , and proves the variational nature of the replica free - energies .
of course a complete control on the remainder in various situations would result in rigorous solutions .
although in this paper we have mainly worked at finite temperature , the zero temperature limit can be considered without harm .
this is particularly relevant in random satisfiability problem , where it is typically found a sat - unsat transition where the ground state energy passes from zero to non zero values .
our analysis shows that , the replica estimates for many of the models considered in the literature , and possibly some of the ones to be obtained in the future with the same method provide upper bounds for the satisfiability thresholds . in this paper
we have confined ourselves to spin models on graphs with poissonian connectivity .
the extension to more general diluted graph models will be presented in a forthcoming paper .
finally we would like to remark that despite the heavy formalism , our proofs to the bounds are conceptually simple .
they are issue of explicit computations and elementary positivity consideration .
we hope that this contributes to illustrate the elegance of the construction first introduced in @xcite .
we thank m. isopi for pointing us out the importance of ref . @xcite .
we thank g. gaeta for discussions .
in this appendix we will explicitly show that expression ( [ pippospin ] ) has positive sign .
we proceed expanding in absolutely convergent series each of the three addend in ( [ pippospin ] ) and showing , taking the expectation of each term that the resulting series is positive semidefinite .
the first term writes : @xmath210 where the term @xmath211 in the last line of eq.([medusa ] ) can be written as @xmath212 \ ; .\ ] ] where each @xmath213 ( @xmath214 ) is a product of @xmath215 gibbs measure with independent fields ( variables appearing in the @xmath216 averages ) , and same fields distributions and quenched disorder ( variables appearing in @xmath177 ) .
the quantities @xmath217 have been defined as : @xmath218 and in this case the averages are performed using a a generalized replica measure , defined as : @xmath219 = e_1 \left [ \frac{\prod_{s=1}^l e_2^{(s ) } z_{(s)}^m \omega_{(s ) } ( s_{i_1} ... s_{i_n})^{k_s } } { ( e_2 z^m)^l } \right ] \label{gen - repmeas}\ ] ] for any integer @xmath40 .
the average over @xmath31 selects the terms with even @xmath220 in ( [ medusa ] ) so that we finally find @xmath221\ ] ] notice that @xmath222 @xmath223 integer @xmath224 only in the current hypothesis that @xmath225 $ ] .
analogously , the term @xmath226 writes @xmath227\ ] ] or , making use of the definition of @xmath100 , @xmath228\ ] ] eventually , following analogous manipulations , the last term @xmath229 can be written as @xmath230 invoking [ probq2 ] and collecting all @xmath195 & = & \frac{\alpha}{\beta m } \sum_{l\ge 1}\frac{m^l}{l } \sum_{{k_1, ... ,k_l \atop \sum_{s=1}^l k_s { \rm even}}}^{1,\infty } \prod_{s=1}^l \left ( \frac{\prod_{r=1}^{k_s-1}(r - m)}{k_s ! }
\right ) \left\langle \left ( \tanh ( \beta j ) \right)^{\sum_{s=1}^l k_s } \right\rangle_j \ ; \cdot \nonumber \\ & & \omega^{(l ) } \left [ ( q^{(k_1, ... ,k_l)})^p - p a(k_1, ... ,k_l)^{p-1 } ( q^{(k_1, ... ,k_l ) } ) + ( p-1 )
a(k_1, ... ,k_l)l^{p } \right ] \label{final_monster}\end{aligned}\ ] ] where we have defined : @xmath231 each inner term of the series ( [ final_monster ] ) @xmath232 \label{final_term}\ ] ] is always positive semidefinite for @xmath0 even while we need the condition conditions @xmath233 for @xmath0 odd . for @xmath16
one retrieves the viana - bray result where ( [ final_term ] ) is a perfect square . as in the rs case
, one can now integrate eq.([final_monster ] ) and recognize that once more the total true free - energy can be written as variational term plus a positive extra one .
the variational term coincides with the 1rsb free - energy at stationarity and under condition @xmath234
the aim of this appendix is to show that the expression for the remainder @xmath236 $ ] in ( [ fverars ] ) for the @xmath2-sat model case as positive sign . for the @xmath2-sat
@xmath236 $ ] specializes to : @xmath161 & = & -\frac{\alpha}{\beta } e \left [ \left\langle \log \left ( \omega \left ( \exp^{-\beta \prod_{r=1}^p \frac{1+j_{r } s_{r}}{2 } } \right ) \right ) \right\rangle_{\{j_t\ } } - \right .
\nonumber \\ & & p
\left\langle \log \left ( 1 + \omega(s ) \tanh ( \beta u ) \right ) \right\rangle_u - - p \left\langle \log \left ( 1 + \frac{\xi}{2 } \prod_{t=1}^{p-1 } \left ( \frac{1 + j_t \tanh ( \beta g_t)}{2 } \right ) \right ) \right\rangle_{\{g_t\ } , \{j_t\ } } + \nonumber \\ & & \left . ( p-1 ) \left\langle \log\left ( 1 + \xi \prod_{t=1}^{p}\frac{1+j_{r } \tanh ( \beta g_t)}{2}\right ) \right\rangle_{\{g_t\ } , \{j_t\ } } \right ] \label{rem - ksat - rs-1}\end{aligned}\ ] ] which thanks to the relation between @xmath36 and @xmath100 , rewrites as @xmath161 & = & -\frac{\alpha}{\beta } e \left [ \left\langle \log \left ( 1+(\e^{-\beta}-1)\omega(\prod_{t=1}^p \frac{1+j_{t } s_{t}}{2 } ) \right ) \right\rangle_{\{j_t\ } } - \right .
\nonumber \\ & & p \left\langle \log \left ( 1 + \xi \omega\left(\frac{1+j s}{2 } \prod_{t=1}^{p-1}\frac{1+j_{t } \tanh ( \beta g_t)}{2 } \right ) \right ) \right\rangle_{\{g_t\ } , j , \{j_t\ } } + \nonumber \\ & & \left .
( p-1 ) \left\langle \log\left ( 1 + \xi \prod_{t=1}^{p}\frac{1+j_{r } \tanh ( \beta g_t)}{2}\right ) \right\rangle_{\{g_t\ } , \{j_t\ } } \right ] \label{rem - ksat - rs}\end{aligned}\ ] ] the last term has been added and subtracted from eq.([fverars ] ) in order to extract a remainder that would vanish if replica symmetry holds , and maximization is performed on @xmath100 . as in the @xmath0-spin case , we will proceed in a taylor expansion of expression ( [ rem - ksat - rs ] ) in powers of @xmath237 , and rely on absolute convergence to average each term of the series .
expanding the first term in ( [ rem - ksat - rs ] ) we can write @xmath238 = \nonumber \\ & & \sum_{n\ge 1 } \frac{(-1)^{n+1}}{n } ( \xi^*)^n e \left [ \left\langle \omega\left(\prod_{t=1}^p ( 1+j_t s_t)\right)^n \right\rangle_{\{j_t\ } } \right ] = \nonumber \\ & & \sum_{n\ge 1 } \frac{(-1)^{n+1}}{n } ( \xi^*)^n \omega \left [ \prod_{t=1}^p \left ( 1 + \sum_{l=1}^n \left\langle j^l_t \right\rangle_{j_t } \sum_{a_1< ...
<a_l}^{1,n } s_t^{a_1} ... s_t^{a_l}\right ) \right]=\nonumber \\ & & \sum_{n\ge 1 } \frac{(-1)^{n+1}}{n } ( \xi^*)^n \omega \left [ \prod_{t=1}^p \left ( 1 + \sum_{l=1}^n \left\langle j^l_t \right\rangle_{j_t } \sum_{a_1< ... <a_l}^{1,n } q^{a_1 ... a_l } \right ) \right]=\nonumber \\ & & \sum_{n\ge 1 } \frac{(-1)^{n+1}}{n } ( \xi^*)^n \omega[(1+q_n)^p]\end{aligned}\ ] ] where we have defined @xmath239 and @xmath240 .
notice that due to the negative sign of @xmath241 , the coefficients @xmath242 are all negative .
the analogous expansion of the second term is : @xmath243= \nonumber \\ & & \sum_{n\ge 1 } \frac{(-1)^{n+1}}{n } ( \xi^*)^n \omega \left [ \left ( 1 + \sum_{l=1}^n \langle j^l \rangle_j \sum_{a_1< ... <a_l}^{1,n } q^{a_1 ...
a_l } \right ) \left\langle \prod_{t=1}^{p-1 } \prod_{l=1}^{n } \left ( 1 + j_t \tanh ( \beta g_t ) \right ) \right\rangle_{\{j_t\},\{g_t\ } } \right ] = \nonumber \\ & & \sum_{n\ge 1 } \frac{(-1)^{n+1}}{n } ( \xi^*)^n\omega \left [ ( 1+q_n)\left\langle(1+j \tanh ( \beta g))^n\right\rangle_{j , g}^{p-1 } \right]\end{aligned}\ ] ] finally , the third terms in eq.([rem - ksat - rs ] ) immediately reads @xmath244 the sum of the three pieces in eq.([rem - ksat - rs ] ) gives : @xmath162 = \frac{\alpha}{\beta } \sum_{n\ge 1 } \frac{(-1)^{n}}{n } ( \xi^*)^n \omega \left [ ( 1+q_n)^p - p(1+q_n)\langle(1+j \tanh ( \beta g))^n\rangle_{j , g}^{p-1 } + ( p-1 ) \langle(1+j \tanh ( \beta g))^n\rangle_{j , g}^{p } \right]\ ] ] the previous sum is always positive semidefinite for @xmath0 even while we need @xmath245 for @xmath0 odd . we proceed in the same way as in the @xmath0-spin case .
the algebra is elementary but more tedious and involved , therefore we will only list the final results of the calculation . starting from eq.([pipposat ] ) , we again expand in series the first term , getting , with a treatment similar to the rs case : @xmath247 = \sum_{l\ge 1}\frac{m^l}{l } \sum_{k_1, ... ,k_l}^{1,\infty } ( -\xi^*)^{\sum_{s=1}^l k_s } \prod_{s=1}^l \left ( \frac{\prod_{r=1}^{k_s-1}(r - m)}{k_s ! }
\right ) \omega^{(l ) } \left [ ( 1 + { \bf q}(k_1, ... ,k_l))^p \right ] \label{primomostro}\end{aligned}\ ] ] where we have defined : @xmath248 analogous steps give for the second term in eq.([pipposat ] ) @xmath249 \left\langle \prod_{s=1}^l \left\langle \left ( 1 + j \tanh ( \beta g ) \right)^{k_l } \right\rangle_g \right\rangle_{g , j}^{p-1 } \label{terzomostro}\end{aligned}\ ] ] and for the third term @xmath250 where in the last two terms we can further expand @xmath251 with @xmath40 equal to @xmath87 and @xmath0 respectively . since @xmath252 it is easy to see how only positive terms of the series survive .
collecting all , we eventually find the complete power expansion for @xmath246 : @xmath253 \label{ultimoomostro}\end{aligned}\ ] ] where we have defined @xmath254 again , every term of the expansion is positive for even @xmath0 and for @xmath0 odd under condition @xmath255 .
let us briefly sketch the proof of the existence of the thermodynamic limit of free - energy of the @xmath0 spin model for @xmath0 .
let us define a model which interpolates between two non interacting systems with @xmath256 and @xmath257 spins respectively , and a system of @xmath258 spins .
each clause @xmath259 will belong to the total system with probability @xmath105 , to the first subsystem with probability @xmath260 and to the second subsystem with probability @xmath261 .
we chose the indices @xmath262 in the following way : for each clause the indices will be i.i.d .
with probability @xmath105 , the indices will be chosen uniformly in the set @xmath8 , with probability @xmath263 the indices will be chosen in @xmath264 and with probability @xmath265 in the set @xmath266 .
let us consider the free - energy @xmath267 .
a direct calculation of its @xmath105-derivative @xmath268 e \langle \log(1+\tanh(\beta j ) \omega(s_{i_1} ... s_{i_p}))\rangle_j.\ ] ] expanding the logarithm in series , observing that thanks to the symmetry of the @xmath31 distribution the odd term vanish , introducing the replica measure and using the convexity of the function @xmath269 for even @xmath0 one proves that @xmath270 which implies sub - additivity @xmath271 ; this is in turn is a sufficient condition to the existence of the free - energy density .
the same prove applies to the even @xmath0 random k - sat model . for odd
@xmath0 we face a difficulty similar to the one in the replica bounds .
we can not prove sub - additivity due to the need to consider negative values of the overlaps , and non convexity of @xmath269 for negative @xmath155 .
for a review on the work of mathematicians on spin glass systems see : a. bovier and p. picco ( eds . ) _ mathematical aspects of spin glasses and neural networks _ ,
progress in probability , vol .
41 , birkhauser , boston , 1997 .
+ m. talagrand , proceedings of the berlin international congress of mathematicians .
, extra volume i , ( 1998 ) 1 and references therein .
we refer to the web page of m. talagrand for the comprehensive list of references on the work of this author : http://www.math.ohio-state.edu/ talagran/ j. s. yedidia , w. t. freeman , and y. weiss , _ understanding belief propagation and its generalizations _ , 2001 .
merl technical report tr 2001 - 22 , available at http://www.merl.com/papers/tr2001-22 .
j. s. yedidia , w. t. freeman , y. weiss , in _ advances in neural information processing systems 13 _ , edited by t. k. leen , t. g. dietterich , and v. tresp , ( mit press , cambridge , ma , 2001 ) | in this paper we generalize to the case of diluted spin models and random combinatorial optimization problems a technique recently introduced by guerra ( cond - mat/0205123 ) to prove that the replica method generates variational bounds for disordered systems .
we analyze a family of models that includes the viana - bray model , the diluted @xmath0-spin model or random xor - sat problem , and the random k - sat problem , showing that the replica method provides an improvable scheme to obtain lower bounds of the free - energy at all temperatures and of the ground state energy . in the case of k
- sat the replica method thus gives upper bounds of the satisfiability threshold . |
the phase transition of strongly interacting matter has been intensively studied for many years . as early as 1960s ,
before the discovery of qcd , there was speculation about a possible new phase of strongly interacting matter , based on studies of the thermodynamics of a hadron gas .
particularly , in the statistical bootstrap model @xcite , the asymptotically exponential mass spectrum of hadrons implied the existence of a limiting temperature of about @xmath0 mev ( the hagedorn temperature ) above which hadrons can not exist . after the discovery of qcd and in particular asymptotic freedom , the discussion focused on the ideas of a quark gluon plasma ( qgp ) , a system of weakly interacting quarks and gluons , and a possible transition between a pion gas and a quark gluon plasma .
the physical picture for such a transition was that at the critical temperature the additional degrees of freedom carried by the quarks and gluons would be released leading to a rapid increase of the entropy , energy - density and pressure .
this was supported by first lattice qcd ( lqcd ) calculations @xcite . lately , however , the hadron resonance gas experienced a renaissance mainly due to its successful description of hadron yields in heavy ion and also elementary particle collisions @xcite .
in addition it was realized that the sharp rise of the entropy density near the transition temperature , as observed in lattice qcd calculations , could be accounted for by a hadron resonance gas as well @xcite .
the success of the hadron resonance gas below the transition temperature @xmath1 , however , changes the physical interpretation for the qcd phase transition . instead of releasing additional degrees of freedom at @xmath1 ,
the system has to reduce the number of active degrees of freedom at the transition temperature , because a hadron gas has many more than a quark gluon plasma . in case of a hagedorn exponential mass spectrum
, the entropy density actually diverges at the hagedorn temperature .
but even if the number of hadrons is restricted to those needed for a successful description of hadronic final states in heavy ion and elementary particle collisions , the entropy quickly exceeds that observed on the lattice .
meanwhile many papers have addressed this issue ranging from qcd - based approaches ( see for example @xcite ) to various generalizations of the hadron gas @xcite . in this article
we will study the transition region , and we will provide a simple and intuitive modification to the hadron gas model in order to qualitatively reproduce the crossover transition observed in lattice qcd @xcite .
our calculations are based on the ideas proposed in @xcite , where it was shown that under certain circumstances a gas of extended hadrons could produce phase transitions of the first or second order , and also a smooth crossover transition that might be qualitatively similar to that of lattice qcd .
we first observe that for the typical particle density at the transition the size of hadrons needs to be taken into account as it leads to a considerable suppression of the available phase space . as a result
the number of effective degrees of freedom is reduced .
this suppression alone , however , does not explain the ideal gas behavior of lattice qcd at high temperatures , above @xmath1 .
additional model assumptions have to be made .
for example one could explicitly introduce a deconfined phase and then match the two different phases .
for this exercise to work , however , rather detailed assumptions about the intrinsic nature of hadrons have to be made in order to avoid the occurrence of an actual phase transition @xcite . in this paper
, we follow a different route .
we adopt the philosophy that the same partition function should describe both `` phases '' . to this end
we need to introduce appropriate dynamics in order to model the crossover observed on the lattice .
this approach is similar in spirit to @xcite .
we find that the mit bag model @xcite of the hadrons is well suited for our purposes , because it embodies confined and deconfined phases from the very beginning .
thus , we will describe hadrons as extended bags of qgp and we will assume an infinite mass spectrum of the hagedorn s type .
the additional dynamics needed to describe the transition are simply the elastic interactions between hadrons .
they give rise to a _
kinetic _ pressure that in turn `` squeezes '' the bag - like hadrons .
we find that the behavior of our system depends sensitively on the choice of parameters for the mass spectrum .
one can obtain either a real phase transition @xcite or , as we shall show , a crossover .
in addition , in the latter case the specific parametrization for the mass spectrum affects the microscopic structure of the gas of `` compressible '' hadrons at high temperatures ( @xmath2 ) .
one finds either a high temperature phase that is populated by one or few infinitely large bags , consistent with the usual picture of a qgp . or
, for a different choice of parameters , one obtains a system of many , densely packed heavy hadrons , which nonetheless exhibit the thermodynamic properties of a qgp of massless quarks and gluons . on first sight
, our approach appears to be similar to the ideas of percolation models @xcite .
the finite size corrections to the statistical ensemble remove all the configurations with overlapping hadrons , resulting in large hadronic states dominating the partition function .
this is similar to percolation .
however , there are quite some differences in the specific implementation .
first , in our model , we do not consider an explicit coupling between the bags as it is done in the percolation model of @xcite .
instead , we take the effect of the kinetic pressure onto the bag sizes into account , resulting in a self - consistency relation for the effective bag pressure . this is more in the spirit of a mean field description , although we do not introduce an additional interaction but simply consider the kinetic pressure .
second , in contrast to purely geometric percolation , in our model the number of hadrons and their sizes are not independent quantities . in a given multihadron state ,
melting two or more hadrons together ( to form a bigger one ) results in a different kinetic pressure and , in turn , in the rearrangement of the sizes of all the hadrons . throughout this paper
, we will maintain a simple schematic approach to highlight the main features of the model leaving a more detailed quantitative analysis and further generalizations to future work .
we will confine ourselves to the simplest case of nonrelativistic boltzmann particles and we will neglect subtleties such as surface effects and van der waals type residual interaction among the bags . this paper is organized as follows : in section [ pf ] we introduce the main ideas of the model and derive the grand - canonical partition function for the gas of compressible hadrons . in section [ isobpf ]
we will use the corresponding isobaric partition function to perform a comprehensive numerical analysis .
we will study the pressure , the energy density and the entropy density of the system .
we will further analyze particles number , the filling fraction and the average mass of particles in the system .
in this section we will set up the general formalism for our model .
let us start with the partition function @xmath3 of an ideal gas of boltzmann particles of mass @xmath4 and degeneracy @xmath5 in the nonrelativistic limit .
for the subsequent discussion it is advantageous to express the partition function @xmath3 in a multiplicity expansion , i.e. as a sum of partition functions @xmath6 for fixed particle numbers @xmath7 : @xmath8 with @xmath9}= \exp{\left[-\frac{m}{t}\right ] } \left(\frac{m t}{2 \pi}\right)^{3/2}\ ; . \label{1.2}\ ] ] here @xmath10 and @xmath11 are the volume and the temperature of the system , respectively .
the function @xmath12 is the _ grand - canonical _ partition function with vanishing chemical potentials . in the context of this paper
we shall refer to @xmath13 as the _ canonical _ partition function keeping in mind that this notation deviates from the conventions for relativistic hadron gases , where the canonical ensemble has fixed abelian charges ( such as electric charge , strangeness , baryon number ) , but no constraints on the number of particles .
( [ 1.1 ] ) can be easily generalized to a multispecies gas of particles .
if we label with @xmath14 the various particle species we have : @xmath15 = \exp{\left[v \sum_{j=1}^{k } g_j \phi(m_j , t)\right]}\ ; . \label{1.3}\ ] ] in case of @xmath16 , it is convenient to replace the discrete index @xmath17 with a continuous spectrum density @xmath18 so that the number of species in the mass interval @xmath19 $ ] is given by @xmath20 . formally , we make the substitution : @xmath21 by expanding the exponential in eq .
( [ 1.3 ] ) the partition function can then be written as : @xmath22 \ ; . \label{1.5}\ ] ] because a hadron gas , or , more precisely , a hagedorn gas is characterized by an exponential mass spectrum , we set @xmath23 where the parameters @xmath24 and @xmath25 will be determined from empirical data . in the case of gas - of - bags models , which also have an exponential mass spectrum , the parameters @xmath24 and @xmath25 will have to be determined from the underlying ( dynamical ) model parameters , such as bag pressure , and so on . we note that @xmath24 has dimensions of @xmath26^{\alpha-1}$ ] and @xmath25 typically ranges from @xmath27 to @xmath28 depending on the model . in eq .
( [ 1.7 ] ) , @xmath29 simply parametrizes the mass spectrum . in the context of the mit bag model
, @xmath29 can be interpreted as the effective `` temperature '' inside the bag , as will be discussed in section [ rpf ] . by substituting eq .
( [ 1.7 ] ) in eq .
( [ 1.5 ] ) we arrive at the following partition function for a hadron gas @xmath30 \\ \nonumber & & \times \exp{\left[\frac{\sum_{i=1}^n m_i}{t_0}-\frac{\sum_{i=1}^n m_i}{t}\right ] } \ ; .\end{aligned}\ ] ] this partition function @xmath3 ( and also @xmath6 ) is divergent for @xmath31 as already pointed out by hagedorn .
although an upper limit in the mass spectrum @xmath18 regulates the divergences , it will not prevent the system from having a much higher entropy density than that observed on the lattice .
an exponential mass spectrum without any cutoff may certainly be an oversimplification and a more realistic calculation may take into account discrete states as well as a mass spectrum that grows less than exponential above a certain mass .
however , empirically the known hadronic states do indeed grow exponentially up to a mass of @xmath32 . above that , very few states are known and it is not clear if this is an indication of a saturating density of states or simply the lack of experimental data on higher mass resonances .
therefore , working with an exponential mass spectrum without any cutoff appears to be an approximation as good as any other .
furthermore , because we are interested only in bulk thermodynamic quantities such as energy density and pressure , the use of a continuous mass spectrum should be a reasonable approximation as all these quantities represent integrals / sums over the mass spectrum .
therefore , in this paper we will assume that the mass spectrum is of the hagedorn type and will discuss a dynamical scenario in the framework of the mit bag model , which will regulate the partition function .
to develop the partition function of our model , we need to recall some of the basic features of the mit bag model @xcite . in its simplest formulation
, hadrons can be considered as _
bags _ of partonic fields confined in a spatial region with a constant potential energy per unit volume @xmath33 , where @xmath33 is commonly referred as the bag constant or bag pressure . the total energy , i.e. the mass @xmath4 of a bag with volume @xmath34 is then given by @xcite : @xmath35 where @xmath36 is the internal energy of the field inside the bag . when its linear extension is larger than the wavelengths of the partons ( the quanta of the inner field ) , we can approximate the bag by a gas of free massless particles confined to its volume @xcite . for a sufficiently large @xmath34 ,
the internal energy is then given by the relation : @xmath37 where @xmath38 is the pressure of the gas .
for a single hadron the stability condition requires : @xmath39 which then gives @xmath40 the effective temperature of the bag is related to the bag pressure by the relation @xmath41 , where @xmath42 is a dimensionless constant whose value depends on the number of internal degrees of freedom of the gas inside the bag . for large @xmath34 , the entropy @xmath43 of a bag , is the entropy of a massless gas with internal energy @xmath36 and pressure @xmath33 @xcite , therefore ) one should subtract a constant @xmath44 which corresponds to the entropy at @xmath45 . here
, this term has been omitted as it is immaterial for our purposes . ] , @xmath46 from eq .
( [ entrobag ] ) one can derive the level density @xmath47 note that the generic spectrum introduced in eq .
( [ 1.7 ] ) , has an additional contribution : @xmath48 .
this factor can be interpreted as a logarithmic correction to the entropy of the bag . in what follows
, we will retain the spectrum in eq .
( [ 1.7 ] ) and we will analyze different values of @xmath25 . of course , eq .
( [ baglevdens ] ) will correspond to the case @xmath27 .
once the temperature @xmath11 of the gas of bags approaches @xmath29 , the average masses and hence the volume ( see eq .
( [ massbagm ] ) ) of the bags grow very fast . therefore
, the bag - like hadrons tend to occupy more and more of the available space and , eventually , they will overlap . to avoid multiple counting of the phase space ,
configurations with overlapping bags need to be excluded from the partition function . for a finite system of volume @xmath10
this can be achieved with an excluded volume correction , where the total volume @xmath10 is replaced by the available volume @xmath49 , where @xmath50 is the volume of the @xmath51th particle .
in addition the volume of all bags @xmath52 should not exceed the total volume @xmath10 .
following ref .
@xcite , this leads to the modified @xmath7-particle phase space integral @xmath53 \rightarrow \left[\prod_{i=1}^n \frac{1}{(2 \pi)^3 } \int \d^3 { { \rm p}}\right ] \left(v-\sum_{i=1}^n v_i\right)^n \theta{\left ( v-\sum_{i=1}^n v_i \right ) } \ ; .
\label{nphsp}\ ] ] this modified phase space integral results in a well - defined and finite partition function at every temperature @xmath54 \\ \nonumber & & \times \exp{\left[\frac{\sum_{i=1}^n m_i}{t_0}-\frac{\sum_{i=1}^n m_i}{t}\right]}\left(v-\sum_{i=1}^n
v_i\right)^n \theta{\left ( v-\sum_{i=1}^n v_i \right ) } \ ; .\end{aligned}\ ] ] it can be shown that such a system of extended hadrons leads to a constant value for the energy - density @xmath55 , in contradiction with lattice
qcd , where the energy density is found to increase with the fourth power of the temperature , @xmath56 .
in fact , for @xmath57 , the most favorite configurations are those where the hadrons occupy all the available space . in this case , the energy density of the system correspond to @xmath58 , i.e. the inner density of the hadrons ( see eq .
( [ massbagm ] ) ) .
the underlying reason for this behavior is that the system is not able to pick up additional kinetic energy once the entire volume is filled with bags , as the bags have no more room to move .
obviously some additional dynamics needs to be included to allow for the system to pick up more energy as the temperature is increased .
to this end we adopt the idea of compressible bags @xcite .
more precisely , we will allow the volume of the hadrons to vary under the effect of the pressure generated by their own thermal motion in a self - consistent way .
consequently , as the temperature and hence the pressure increase , the bags will be compressed and acquire a higher internal mass / energy density .
we will show that the system does not exhibit any limiting value of the energy density , and , under appropriate conditions , exhibits the desired increase of the energy density and entropy . to illustrate the underlying mechanism ,
let us consider a gas of many hadrons . for small temperatures , @xmath59 ,
the system is dilute ( @xmath60 ) and behaves like a gas of noninteracting point - particles .
with increasing temperature , the average mass , and hence the spatial extent of the hadrons , increases and as @xmath11 approaches @xmath29 the dilute - gas approximation seizes to be valid .
the pressure exerted by the other particles becomes sizable and its effect on the hadrons properties , such as the size , can no longer be ignored . in other words , in addition to the bag pressure @xmath33 , every particle in the system will feel an additional _ kinetic _ pressure @xmath61 that is generated by the thermal motion of the other hadrons in the gas . in this situation , the stability condition , eq .
( [ dum ] ) , needs to be modified by taking onto account the contribution of the kinetic pressure @xmath61 .
microscopically , the pressure @xmath61 can be interpreted as the consequence of elastic collisions .
the effect of inelastic collisions , which are certainly present in a hadron gas , in our approach are accounted by the infinite mass spectrum of hadrons without enforcing any constraint on the number of particles @xmath7 . in this way
all the possible configurations with few large hadrons or many small ones are included .
this is analogous to the hadron - resonance gas model , where a large part of the inelastic interaction is taken into account by adding resonances as free particles in the gas . neglecting any surface effect , the simplest generalization of the stability condition , eq .
( [ dum ] ) , is @xmath62 a pictorial illustration of the pressure balance in the last equation is given in fig .
( [ bp ] ) .
each hadron in the gas is characterized by the same internal pressure @xmath38 .
but instead of being a constant ( as in the case of a single hadron in the vacuum where @xmath63 ) , @xmath38 now depends on @xmath11 and @xmath10 , and must be evaluated in a self - consistent fashion from the partition function itself .
it is clear that the number of hadrons , their sizes and the kinetic pressure are all connected by the above self - consistency relation .
accordingly , as we have already pointed out , if we split or combine two or more hadrons , the pressure , and thus the volume of all the hadrons in the system , changes .
this new state corresponds to a distinct state of the ensemble that can not be obtained by a simple geometrical clustering procedure as usually done in percolation models @xcite . to account for this additional dynamics , we need to generalize the partition function , eq .
( [ 1.7b ] ) . to this end
, we write the explicit dependence of volumes and the masses of the bags on the pressure @xmath38 @xmath64 we also rewrite the exponential mass spectrum , @xmath65 in terms of the general expression for the entropy @xmath66 leading to the substitution @xmath67 finally , because the bag masses depend on @xmath38 , instead of integrating on @xmath68 as in eq .
( [ 1.7b ] ) we will perform the integral over the internal energies , i.e. we substitute @xmath69 the factor @xmath70 in the previous formula ensures that we recover the partition function , eq .
( [ 1.7b ] ) in the limit of @xmath71 , i.e. in the dilute gas limit . with the replacements in eq .
( [ replacements ] ) to ( [ intvar ] ) , the modified grand - canonical partition function can now be written on the basis of eq .
( [ 1.7b ] ) and reads : @xmath72 \exp{\left\{\left [ \frac{4}{3kp_r^{1/4}}-\frac{1}{t}\left(1+\frac{b}{3p_r}\right ) \right]\sum_{i=1}^n
u_i\right\}}\\ \nonumber \ ; & & \times \left(v-\sum_{i=1}^n \frac{u_i}{3p_r } \right)^n \theta{\left ( v-\sum_{i=1}^n\frac{u_i}{3p_r } \right)}\ ; .\end{aligned}\ ] ] the new partition function @xmath3 is identical ( by construction ) to eq .
( [ 1.7b ] ) when @xmath63 .
conversely , for a given set @xmath73 , the effect of a finite kinetic pressure @xmath74 is to squeeze each bag to a smaller size ( see eq .
( [ replacements ] ) ) , and as a result , the effective bag temperature @xmath75 increases . the eq .
( [ gcpartfunupr ] ) can be made more familiar by substituting @xmath76 leading to @xmath77\\ \nonumber \ ; & \times & \exp{\left\{\left [ \frac{1}{kp_r^{1/4}}-\frac{1}{t}\left(\frac{3}{4}+\frac{b}{4p_r}\right ) \right]\sum_{i=1}^n \eta_i\right\ } } \\ \nonumber \ ; & & \times \left(v-\sum_{i=1}^n \frac{\eta_i}{4p_r } \right)^n \theta{\left ( v-\sum_{i=1}^n\frac{\eta_i}{4p_r } \right)}\ ; .\end{aligned}\ ] ] obviously , in the dilute gas limit , @xmath71 @xmath78 .
we further introduce a lower bound @xmath79 for the integrals over @xmath80 .
this is needed because , for @xmath81 , the spectrum in eq .
( [ 1.7 ] ) has a pole in @xmath82 , resulting in a divergent partition function for @xmath83 .
because there are no hadrons lighter than the pion , we will set @xmath84 gev .
because the pressure @xmath61 is thermally generated , it must be calculated from the partition function itself , resulting in a self - consistency relation .
this is best achieved by introducing the isobaric partition function , which is defined as the laplace transform of @xmath3 over the variable @xmath10 : @xmath85 } \ ; . \label{ibpf1}\ ] ] the quantity @xmath86 in eq .
( [ ibpf1 ] ) plays the role of a constant external pressure .
accordingly , the equilibrium condition requires @xmath87 the integral in eq .
( [ ibpf1 ] ) can be solved analytically ( see appendix a ) and gives : @xmath88^n=\frac{1}{s - f(t , s ) } \label{ibpf2}\ ] ] with @xmath89 } \ ; .\end{aligned}\ ] ] in the limit @xmath90 the asymptotic behavior of @xmath3 is defined by the singularity of @xmath91 with the largest real part @xcite .
we have two distinct cases : @xmath92 and @xmath93 .
for @xmath92 the transform @xmath91 in eq .
( [ ibpf2 ] ) has two kinds of singularities : the first , @xmath94 , is given by the pole of @xmath95 , i.e. @xmath96 whose solution is the pressure of the system according to eq .
( [ pkst ] ) . the second singularity , @xmath97 , corresponds to a divergence of the function @xmath98 itself .
this happens if the exponent of the integrand of @xmath99 in eq .
( [ ibpf3a ] ) vanishes , i.e. @xmath100 the situation is schematically represented by the leftmost curve in fig .
( [ fts ] ) .
the solid line represents the function @xmath99 for a given temperature and the @xmath101 s dashed line corresponds to @xmath102 .
the intersection between the dashed and the solid line corresponds to the solution @xmath103 of eq .
( [ ibpf4 ] ) and is denoted by a black dot .
the function @xmath99 , is a positive function of @xmath102 that goes to infinity for @xmath104 and tends to zero as @xmath105 .
consequently , there is always a solution for eq .
( [ ibpf4 ] ) and the pole @xmath106 corresponds to the rightmost singularity . thus the pressure has always a solution .
for @xmath93 the situation is different , however . in this case
the function @xmath98 has an essential discontinuity at @xmath97 : it is finite at @xmath104 , and diverges for @xmath107 .
since @xmath97 increases with temperature ( see eq .
( [ sing ] ) ) and @xmath108 as @xmath105 , for sufficiently large @xmath11 , @xmath109 , and , consequently , eq .
( [ ibpf4 ] ) does not have a solution .
this situation is illustrated by the rightmost curve in fig .
( [ fts ] ) where @xmath110 ( denoted by a x ) lies below the diagonal since the singularity @xmath111 for @xmath112 ( see eq .
( [ sing ] ) ) . ] .
both these cases , have been discussed in @xcite where the absence of the solution @xmath94 was identified with the onset of a phase transition . here , we analyze in detail the case of a crossover transition , i.e. @xmath92 .
before we proceed , let us fix the model parameters .
in what follows , we set @xmath113 and we keep the product @xmath114 fixed .
this yields a bag pressure @xmath115 gev@xmath116 ( i.e. @xmath117 mev ) which is a plausible value for the bag model parameter .
the constant @xmath42 has been chosen in order to obtain roughly the same value for @xmath118 as lqcd for large @xmath11 ( see fig .
( [ isob1 ] ) ) .
its value can also be estimated by counting the degrees of freedom of a thermal system of independent quarks and gluons .
our choice lies between the values for the lightest quark doublet u , d ( @xmath119 ) and for u , d and s quark ( @xmath120 ) .
the values for the remaining parameter @xmath24 have been fixed by fitting the shape of the actual hadron mass spectrum over the mass range of @xmath121 .
they are given in table [ tab1 ] . .
[ cols="<,^,^,^,^,^,^",options="header " , ] of course a different mass range or a different choice for @xmath29 would affect these fits .
however , for our schematic considerations here , a fine tuning of the model parameters is rather meaningless . in the same spirit
we also ignore the @xmath122 deviation of the lqcd result for @xmath118 from the free gas ( stefan - boltzmann ) limit @xcite . in fig .
( [ isob1]a ) we plot the pressure @xmath123 evaluated with the isobaric partition function , eq .
( [ ibpf4 ] ) .
as one can see , the results depend on the choice of @xmath25 . for @xmath27 and @xmath124 , the curves grow with the temperature with larger slopes for smaller @xmath25 s . instead , for @xmath125
, they settle onto constant asymptotic values ( as we have verified numerically up to @xmath126 gev ) . as @xmath25 change from @xmath127 to @xmath128
the asymptotic value converges very fast to @xmath129 ( the stefan - boltzmann limit .
accordingly , the energy density is given by @xmath130 . ] ) from above .
as shown in the plot , the curves practically coincide with the stefan - boltzmann limit already for @xmath131 .
this behavior can be understood by inspecting the solution @xmath132 for the pressure . because @xmath94 is always larger than @xmath97 , we have @xmath133 and for @xmath134 @xmath135 the pressure @xmath136 converges to @xmath137 only for sufficiently large values of @xmath25 , when the solutions @xmath94 and @xmath97 get closer and closer as @xmath134 , to obtain the pressure of an ideal gas the difference @xmath138 must decrease faster than @xmath139 . for the ratio @xmath140 it is sufficient that the difference @xmath138 grows slower than @xmath141 . ] . in fig .
( [ isob1]b ) , we plot the ratio @xmath118 , where @xmath55 has been evaluated numerically by using the relation : @xmath142 again this quantity converges to a finite asymptotic limit only for @xmath125 and , as before , coincides with the massless gas limit @xmath143 for @xmath131 and @xmath144 .
the overall behavior is roughly the same as lqcd except for a small `` horn '' right above @xmath29 .
a closer look reveals that this is due to the constant bag pressure @xmath33 that produces a contribution @xmath145 to the energy density of the system .
being a constant term its contribution to @xmath146 becomes negligible at high temperatures . in fig .
( [ isob2]a ) , the ratio @xmath147 is plotted up to @xmath148 gev .
this quantity corresponds to the trace of the energy - momentum tensor @xmath149 , which is actually the fundamental quantity calculated in lqcd @xcite . in fig .
( [ isob2]b ) , we plot the ratio @xmath150 , where @xmath102 is the entropy density .
again , for @xmath27 and @xmath124 , the curves do not converge to a constant value , as expected from the previous results for @xmath55 and @xmath61 . in our scheme , the pressure @xmath61 ( and thus the energy density and the entropy ) exceeds the corresponding stefan - boltzmann limit , except for @xmath151 and @xmath128 where it converges to it .
to obtain a lower pressure ( at least at finite temperatures ) one needs to further reduce the effective degrees of freedom of the system .
this could be possibly achieved by introducing a surface energy term .
in addition to the fact that part of the energy of the system would be spent to create the bags surface , such a contribution would favor spherically shaped bags , resulting in a further suppression of the accessible phase space .
probably , a similar effect could be also obtained including some residual repulsive interaction of the van der waals type .
we stress that the general behavior of our model for the pressure , energy- and entropy - density can not be obtained by simply introducing an upper mass - cutoff on the exponential spectrum in eq .
( [ 1.7a ] ) : the entropy - density would considerably exceed that obtained from lqcd even if we kept only masses up to @xmath152 gev .
in addition , the system would reach the massless gas limit only at temperatures much higher than the cutoff itself . on the other hand , in our model
it is absolutely essential to assume an infinite mass spectrum .
otherwise the flat behavior in fig .
( [ isob1 ] ) and on fig . ( [ isob2]b ) would be spoiled and all these quantities would decrease with the temperature .
it is also interesting to plot the average bag effective `` temperature '' @xmath153 . as shown in fig .
( [ tb ] ) , for @xmath131 and @xmath144 this converges to @xmath11 very quickly above @xmath29 . for @xmath154 or smaller
the effective bag temperature is always larger than the system temperature .
this fact is a direct consequence of the inequality in eq .
( [ presslim ] ) that gives @xmath155 notice also that in the region @xmath112 , @xmath156 for any value of @xmath25 .
in other words , the pressure @xmath61 is negligible with respect to @xmath33 and the system behaves as a standard hadron gas .
it is worth mentioning that in our framework the compressible hadrons can exchange energy only through a mechanical work ( compression ) .
they are , therefore , thermally insulated from the rest of the system and the bag temperature , as well as the entropy , must be understood as quantities that measure the degeneracy of the hadronic states . finally , for @xmath125 , our model seems to produce a smooth crossover transition toward a new regime whose features are very similar to those of a gas of massless particles , even though no deconfined states are included in the partition function . to better understand this behavior it is useful to study the particles density @xmath157 ( fig .
( [ new1]a ) ) and the _ filling fraction _ ( @xmath158 ) ( fig .
( [ new1]b ) ) which is defined as : @xmath159 where @xmath160 is the average volume occupied by the hadrons ( for a rigorous definition and formula see appendix b ) .
the particles density can be calculated from the isobaric partition function by introducing a fictitious fugacity @xmath161 ( to be set to 1 afterward ) for each particle in the system , i.e. replacing @xmath3 with @xmath162 accordingly , the isobaric partition function in eq .
( [ ibpf2 ] ) becomes @xmath163 and the corresponding solution for the pressure @xmath164 .
in the infinite volume limit , the particles density can be then obtained as : @xmath165 where in the last equality we have used the known relation @xmath166 . as shown in fig .
( [ new1]a ) , the particles density grows very rapidly as @xmath11 approaches @xmath29 from below .
this is qualitatively what one expects for the hadron gas , where the average number of particles shows a monotonically growing behavior .
conversely , for @xmath167 this quantity depends very strongly on the choice of the parameter @xmath25 . for @xmath27 , @xmath124 and @xmath127
we observe a change in the slope , but the curves still grow monotonically . for @xmath168 and @xmath152
the particles density has a local minimum at @xmath169 gev and @xmath170 gev , respectively ( the latter lies outside the plotted region ) , and then it starts growing again with smaller slopes for larger values of @xmath25 . for @xmath144
the local minimum has disappeared , and after a sharp maximum at @xmath171 gev the particles density goes to zero .
this is the effect of the finite size of hadrons , which tends to saturate the available system volume .
in fact , as shown in fig .
( [ new1]b ) , when @xmath172 the filling fraction , eq .
( [ 3.5a ] ) , is relatively small , whereas for higher temperatures , the system is almost totally filled by extended particles , i.e. @xmath173 . in this scenario ,
the space and the phase space available is strongly suppressed , and the system tends to be populated by a smaller number of heavy particles .
this effect is strongest for @xmath144 ( the filling fraction converges very fast to 1 ) and therefore @xmath174 . for smaller values of @xmath25
this saturation effect becomes slightly less pronounced .
a closer inspection reveals that for @xmath175 the filling fraction has a maximum at very high temperature and then decreases with a very small slope , the rate of decrease of the filling fraction for large temperatures is maximum , but still only @xmath176 going from @xmath177 gev to @xmath178 gev ] , whereas for @xmath179 it seems to settle to a constant value @xmath180 .
the phase space , therefore , is never entirely suppressed . the behavior of the system changes continuously by varying @xmath25 .
a numerical analysis indicates that there exist a value @xmath181 between @xmath182 and @xmath183 such that , at high @xmath11 , the particles density vanishes for any @xmath184 . in this range ,
the system is populated by one or few infinitely extended hadrons that occupy the entire space , filling the system with their inner qgp matter , which is a possible scenario for the deconfined phase .
conversely , for @xmath185 , we find many , rather heavy,``squeezed '' hadrons , which nonetheless mimic an ideal gas of massless particles .
the number of particles , however , might be affected by the introduction of a surface energy term in the spectrum in eq . ( [ 1.7 ] ) .
such a term would result in an energy cost associated with the splitting of a large hadron into many smaller ones and might then widen the range of values for @xmath25 that lead to @xmath186 at temperatures above @xmath29 .
note that , even though @xmath187 can have a minimum or even vanish , the entropy of the system always increases monotonically with the temperature ( see fig .
( [ isob2]b ) ) .
this is due to the fact that the dominant contribution ( already at @xmath188 ) comes from the bags entropies .
this has been checked by using the classical expression for a gas with @xmath189 particles ( with the excluded volume correction ) and degeneracy @xmath5 : @xmath190 the contribution @xmath191 corresponds to the sum of the bag entropies in eq .
( [ entropy ] ) , and it is dominant at high @xmath11 . another interesting quantity is the average hadron mass @xmath192 , shown in fig .
( [ new2 ] ) that has been evaluated according to @xmath193 as one can see , smaller values of @xmath25 correspond to higher @xmath192 at low temperature , whereas the situation is reversed at high @xmath11 .
this follows from eq .
( [ amass ] ) as a direct consequence of the behavior of @xmath187 ( fig .
( [ new1]a ) ) . for @xmath144
the average mass @xmath192 grows with the temperature with the maximum slope and diverges when @xmath174 , i.e. when the system is populated by a finite number of infinite `` hadrons '' . in
the vicinity of the transition region @xmath192 varies from @xmath194 gev ( for @xmath27 ) to @xmath195 gev ( for @xmath144 ) at @xmath196 gev , and from @xmath197 gev to @xmath198 at @xmath199 gev , for @xmath27 and @xmath128 , respectively .
these values will somewhat depend on the choice of the model parameters .
they would depend even more on an eventual upper mass cutoff .
in fact , the high value of @xmath192 for @xmath27 results from our assumption of an infinite mass spectrum and already at @xmath196 gev it falls outside the region of the known hadrons .
our estimates are , however , lower than what one obtains for an exponential spectrum of point - like hadrons as in eq .
( [ 1.7a ] ) .
in such a case , for @xmath27 the average mass is @xmath200 gev at @xmath196 gev and infinity at @xmath201 gev ( the partition function itself is divergent ) . a mass cutoff at @xmath202
gev would reduce these numbers to @xmath203 gev and @xmath204 gev at @xmath196 gev and @xmath199 gev , respectively . as a final remark
, we want to discuss the consistency of the model . to check this point
we must make sure that quantities such as the energy density ( that we have evaluated from the pressure by using eq .
( [ isobeps ] ) ) correspond to a thermal average of the form @xmath205 a first hint in this direction , is given , _ a posteriori _
, by the results shown in this section , particularly , by quantities such as the filling fraction .
the @xmath158 has been evaluated by making use of the relation in eq .
( [ amass ] ) ( see appendix b ) that implicitly relies on a form like eq .
( [ concheck1 ] ) for the energy density . indeed ,
a wrong thermodynamical interpretation of @xmath55 would likely have lead to dramatic consequences on the filling fraction , which contrarily assumes only physical values in the interval @xmath206 $ ] .
however , to make a more direct test , we will provide an approximate expression for the grand - canonical partition function and we will compare @xmath55 in fig .
( [ isob1]b ) with its corresponding value obtained as in eq .
( [ concheck1 ] ) . this can be done for the case @xmath154 . for this particular value of @xmath25 the multiple integrals over @xmath80 in eq .
( [ partfunmpr ] ) can be reduced to an unidimensional integral ( see appendix c ) that greatly facilitates the numerical treatment .
although , as demonstrated in this section , results do depend on the choice of @xmath25 , for the range of @xmath207 , one can hope that the following arguments will be still valid .
using @xmath208 in eq .
( [ partfunmpr ] ) we get for the partition function @xmath209 with @xmath210^n \frac{1}{n ! } \left [ \prod_{i=1}^n \int_{m_c}^{\infty } \d \eta_i \right]\nonumber \\ \
; & & \times \exp{\left\{\left [ \frac{1}{kp_{r , n}^{1/4}}-\frac{1}{t}\left(\frac{3}{4}+\frac{b}{4p_{r , n}}\right ) \right]\sum_{i=1}^n \eta_i\right\}}\\ \nonumber \ ; & & \times \left(v-\sum_{i=1}^n \frac{\eta_i}{4p_{r , n } } \right)^n \theta{\left ( v-\sum_{i=1}^n\frac{\eta_i}{4p_{r , n } } \right)}\ ; .\end{aligned}\ ] ] here we have added the suffix @xmath7 to @xmath38 to indicate the dependence of the pressure @xmath211 on the particle number in the canonical ensemble .
the pressure @xmath211 is now just a parameter of the model . to determine its value
, we need to find the value of @xmath211 that maximizes the logarithm of the integrand in eq .
( [ gcpartfun ] ) : @xmath212\sum_{i=1}^n \eta_i+n
\log \left(v-\sum_{i=1}^n \frac{\eta_i}{4p_{r , n } } \right ) \ ; , \label{integrand}\ ] ] where ( for the moment ) we omit the @xmath213 function .
the above expression depends on @xmath214 . to simplify the following derivation
, we will introduce an approximation . instead of solving @xmath215 for a generic set @xmath216
, we find a solution for the most important configurations defined by the value @xmath217 that maximizes the integrand .
in other words , we solve the system of equations : @xmath218 note that the second condition ensures that @xmath219 has the form of eq .
( [ concheck1 ] ) as any implicit dependence on @xmath11 of @xmath211 do not contribute to @xmath220 .
in fact , if we denote with @xmath221 the derivative performed only on the explicit @xmath11 dependence of @xmath222 we have @xmath223 where the second term vanishes because of the second condition in eq .
( [ system ] ) .
( [ system ] ) results in ( see appendix d ) : @xmath224 the interpretation of the above expression is straightforward .
writing @xmath225 one obtains the equivalent equation : @xmath226 here , the right - hand side of eq .
( [ eq2a ] ) is simply the canonical pressure of an ideal gas of @xmath7 particles with the total volume @xmath10 replaced by the available volume @xmath227 .
it is then natural to identify @xmath228 with the kinetic pressure in the canonical ensemble . of course ,
once averaged over @xmath7 and over @xmath229 , for a sufficiently large @xmath10 , this pressure must coincide with @xmath136 evaluated with the isobaric partition function .
( [ eq2a ] ) has the form of a self - consistency relation , as @xmath230 appears also in the right - hand side in the excluded volume term .
( [ eq2a ] ) is a quadratic form and has two solutions : a negative and positive one , and the negative corresponds to the situation where volume of hadrons exceeds the total volume , @xmath231 .
the positive solution on the other hand ensures @xmath232 and therefore the condition for the @xmath213 function in eq .
( [ gcpartfun ] ) is always fulfilled . in what follows
, we will adopt the positive solution of the eq .
( [ eq2a ] ) for any @xmath233 ( not only for the most probable set @xmath234 ) and we will integrate numerically @xmath6 in the variables @xmath235 . the partition function @xmath3 is then evaluated by summing over the particle number @xmath7 up to a cutoff @xmath236 @xmath237 where @xmath236 is sufficiently large to ensure the accuracy of our calculations .
finally we test the consistency of our picture by comparing the ratios @xmath140 and @xmath118 with the results obtained with the isobaric partition function . in fig .
( [ check ] ) , the ratios @xmath140 ( left panel ) and @xmath118 ( right ) have been evaluated with the isobaric partition function ( solid line ) and with the grand canonical partition function ( dashed line ) for a volume @xmath238 gev@xmath239 , which is , as we checked , a good approximation of the infinite volume limit .
as one can see , in both cases the two curves are very close , actually for the energy density they are practically coincident .
the small ( expected ) difference between the isobaric and the grand - canonical result reflects the quality of our approximation .
the same test has been also performed on the entropy density , the particles density and the filling fraction .
for all these quantities we have observed an equally good agreement .
in this article , we have studied the crossover transition of the gas of bags @xcite .
we have found that the behavior of the system depends sensitively on the parameter @xmath25 of the hagedorn - like mass spectrum @xmath240 .
the system exhibits a crossover transition for @xmath207 and an actual phase transition for larger values . in the range @xmath207 we made a coarse scan of @xmath25 , setting @xmath27 , @xmath124 , @xmath127 , @xmath241 , @xmath152 and @xmath128 . for @xmath242
the gas of bags undergoes a sharp ( yet , continuous ) transition qualitatively similar to lattice qcd . in this range , the asymptotic values of @xmath243 , @xmath118 , @xmath244 ,
coincide with the stefan boltzmann limit for @xmath131 and @xmath144 , whereas they settle to slightly larger values for @xmath154 and @xmath127 . for @xmath245 and @xmath124 , these quantities grow indefinitely with the temperature with small , decreasing , slopes going from @xmath27 to @xmath246 .
we have also studied the ( strong ) dependence of the particles density @xmath247 ( where @xmath248 ) on @xmath25 .
we have found that there exist a limiting value @xmath181 between @xmath182 and @xmath183 such that for @xmath249 the particles density vanishes at high temperature .
the system is then populated by one ( or few ) infinite bag(s ) that occupies the entire volume .
conversely , for @xmath250 , @xmath187 grows with the temperature . in the range @xmath251 ,
the ideal gas behavior is mimicked by a number of heavy extended bags that saturates the phase space forming a dense system . a pictorial representation summarizing the various high - temperature phases of the model
is given in fig .
( [ schem ] ) . in this work
we have explored a simple , intuitive model for a gas of hadrons in the vicinity of the transition at vanishing baryochemical potential . to this end
, we have adopted the idea of the mit bag model and we have described hadrons as extended qgp bags . we have shown that , in the vicinity of @xmath29 ( which is directly related to the transition temperature of our model ) , elastic interactions among hadrons play a fundamental role . in our schematic approach , they are quantified by the thermal pressure @xmath61 .
the effect of @xmath61 is to squeeze the hadrons , and for a certain set of model parameters , @xmath125 , the ideal gas behavior at high temperature can be reproduced . at the same time , the effective inner `` temperature '' , or rather degeneracy parameter , @xmath252 , of the bags increases with @xmath61 , resulting in a temperature - dependent mass spectrum . above @xmath29 , the physical picture of
the qgp phase corresponds to a number @xmath253 of infinite bags that occupy the entire space .
this is indeed the situation for @xmath249 . on the other hand
, a large number of independent qgp bags ( such as for @xmath250 ) would contradict the lattice findings of vanishing flavor - flavor correlations @xcite .
however , in order to define precisely the range of values of @xmath25 that lead to a consistent qgp scenario , it is fundamental to study the effect of a surface energy term .
this contribution disfavors configurations with a large number of particles ( they are more `` expensive '' in terms of surface energy ) and might reduce the number of bags in the high - temperature phase , widening the range of possible values for @xmath25 .
future work will likely concentrate on the effect of surface energy terms in addition to the study of van der waals type residual interactions .
it would be also interesting to analyze in detail the behavior of the system at finite baryochemical potential .
this could be done starting from the formalism developed in @xcite .
the authors thank m. i. gorenstein for a critical reading of the manuscript and for useful suggestions .
this work is supported by the director , office of energy research , office of high energy and nuclear physics , divisions of nuclear physics , of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 .
by using the equations ( [ simplass ] ) and ( [ pkst ] ) , the isobaric partition function in eq . ( [ ibpf1 ] ) can be conveniently written as : @xmath254 h_n(\{\eta_i \},t , s ) \\ \nonumber \ ; & & \times \int_{0}^{\infty}\d v \exp{\left[-sv\right ] } \left(v-\sum_{i=1}^n \frac{\eta_i}{4 ( b+st ) } \right)^n \theta{\left ( v-\sum_{i=1}^n\frac{\eta_i}{4(b+st ) } \right ) } \end{aligned}\ ] ] where @xmath255 \nonumber\\ \ ; & & \times \exp{\left\{\left [ \frac{1}{k(b+st)^{1/4}}-\frac{1}{t}\left(\frac{3}{4}+\frac{b}{4(b+st)}\right ) \right]\sum_{i=1}^n \eta_i\right\ } } \;.\end{aligned}\ ] ] the integral over @xmath256 can be carried out @xmath257 } \left(v-\sum_{i=1}^n \frac{\eta_i}{4 ( b+st ) } \right)^n \theta{\left ( v-\sum_{i=1}^n\frac{\eta_i}{4(b+st ) } \right ) } \\
\nonumber & & = \frac{n!}{s^{n+1 } } \exp{\left\ { -\sum_{i=1}^n \eta_i \frac{s}{4(b+st ) } \right\ } } \ ; .\end{aligned}\ ] ] and eq . ( [ appa1 ] )
becomes : @xmath258 \frac{h_n(\{\eta_i \},t , s)}{s^n } \exp{\left\ { -\sum_{i=1}^n \eta_i \frac{s}{4(b+st ) } \right\ } } \ ; .\ ] ] the multiple integral in the last equation , can be factorized as : @xmath259 h_n(\{\eta_i \},t , s ) \exp{\left\ { -\sum_{i=1}^n \eta_i \frac{s}{4(b+st ) } \right\ } } \equiv f(t , s)^n\ ; .\ ] ] one then obtains eq .
( [ ibpf2 ] ) with @xmath99 given by eq .
( [ ibpf3a ] ) .
by using eq .
( [ replacements ] ) , we write the average volume occupied by hadrons as @xmath260 where we used the stability condition @xmath261 . in turn , the average mass @xmath192 reads ( see also eq .
( [ amass ] ) ) : @xmath262 the filling fraction @xmath158 can then be expressed in terms of @xmath55 , @xmath61 and the particles density @xmath187 as : @xmath263
the partition function of @xmath7 particles can be reduced to a unidimensional integral when @xmath154 as the factor @xmath264 reduces to 1 .
let us begin by making the substitution : @xmath265 that gives for @xmath13 in eq .
( [ gcpartfun ] ) : @xmath266^{n } \frac{1}{n ! } \left [ \prod_{i=1}^n \int_{0}^{\infty } \d y_i \ ; y_i \right]\left(v-\frac{n m_c}{4 p_{r , n}}- \sum_{i=1}^n \frac{y^2_i}{4 p_{r , n } } \right)^n \nonumber \\ \ ; & & \times \exp{\left\{\left [ \frac{1}{kp_{r , n}^{1/4}}-\frac{1}{t}\left(\frac{3}{4}+\frac{b}{4p_{r , n}}\right ) \right]\left(\sum_{i=1}^n y^2_i+n m_c\right)\right\}}\ ; . \label{a1}\end{aligned}\ ] ] it is now convenient to rewrite our integral by using @xmath7-dimensional hyperspherical coordinates by setting : @xmath267 and @xmath268 with @xmath269 we now note that , apart from @xmath270 , the integrand in eq .
( [ a1 ] ) depends only on @xmath271 .
this is true also for the pressure @xmath272 , where @xmath228 is given by the equations ( [ eq2a ] ) and , in the new variables , reads : @xmath273 where @xmath274 eq .
( [ a1 ] ) can now be written as : @xmath275^{n}\nonumber \\ \nonumber \ ; & & \times \frac{1}{n ! } \int_{0}^{\infty } \d r \ ; r^{n-1 } \int_{0}^{\frac{\pi}{2 } } \d \phi_1 \ldots \int_{0}^{\frac{\pi}{2 } } \d \phi_{n-1 } \sin^{n-2}{\phi_1}\sin^{n-3}{\phi_2 } \ldots \sin{\phi_{n-2}}\\ \nonumber \ ; & & \times r^n \left [ \prod_{i=1}^{n-1 } \cos{\phi_i } \right ] \sin^{n-1}{\phi_1}\sin^{n-2}{\phi_2 } \ldots \sin{\phi_{n-1 } } \left(v-\frac{n m_c}{4 p_{r , n}}- \frac{r^2}{4 p_{r , n } } \right)^n \nonumber \\ \ ; & & \times \exp{\left\{\left [ \frac{1}{kp_{r , n}^{1/4}}-\frac{1}{t}\left(\frac{3}{4}+\frac{b}{4p_{r , n}}\right ) \right]\left(r^2 +
n m_c\right)\right\}}\ ; .\end{aligned}\ ] ] next we separate eq .
( [ a6 ] ) into an angular integral , @xmath276 , and a radial integral , @xmath277 , such that : @xmath278^{n}\frac{1}{n ! } a_n i_n \label{a7}\ ] ] where : @xmath279 \sin^{n-1}{\phi_1}\sin^{n-2}{\phi_2 } \ldots \sin{\phi_{n-1 } } \ ; , \end{aligned}\ ] ] and @xmath280\left(r^2 + n m_c\right)\right\}}\ ; .\end{aligned}\ ] ] the angular integral @xmath276 can be solved by applying the following recursion relation : @xmath281 yielding @xmath282 accordingly , by using @xmath283 one obtains : @xmath284 the eq .
( [ a6 ] ) then reduces to : @xmath285^{n}\frac{1}{n!}\frac{1}{2^{n-1}(n-1 ) ! } \int_{0}^{\infty } \d r \ ; r^{2n-1 } \\ \nonumber \ ; & & \times\left(v-\frac{n m_c}{4 p_{r , n}}- \frac{r^2}{4 p_{r , n } } \right)^n \exp{\left\{\left [ \frac{1}{kp_{r , n}^{1/4}}-\frac{1}{t}\left(\frac{3}{4}+\frac{b}{4p_{r , n}}\right ) \right]\left(r^2 +
n m_c\right)\right\}}\ ; .\end{aligned}\ ] ]
performing the derivatives , the system in eq .
( [ system ] ) reads @xmath286 by subtracting the first equation from the second equation , one easily obtains eq .
( [ cccheck ] ) .
r. hagedorn , in _ erice 1972 , proceedings of the study institute on high energy astrophysics _
( cambridge university press , cambridge , 1974 ) , pp .
255 - 296 .
m. creutz , l. jacobs and c. rebbi , phys .
d * 20 * , 1915 ( 1979 ) .
braun - munzinger , j. stachel , j. p. wessels and n. xu , phys .
b * 365 * , 1 ( 1996 ) .
j. cleymans , d. elliott , h. satz and r. l. thews , z. phys .
c * 74 * , 319 ( 1997 ) .
w. broniowski , a. baran and w. florkowski , acta phys . polon .
b * 33 * , 4235 ( 2002 ) .
f. becattini , j. manninen and m. gazdzicki , phys .
c * 73 * , 044905 ( 2006 ) .
a. andronic , p. braun - munzinger , k. redlich and j. stachel , j. phys .
g * 35 * , 104155 ( 2008 ) .
j. manninen and f. becattini , phys .
c * 78 * , 054901 ( 2008 ) .
f. becattini , z. phys .
c * 69 * , 485 ( 1996 ) .
f. becattini , in _ proceedings of xxxiii eloisatron workshop on universality features in multihadron production and the leading effect _ edited by a. zichichi ( 1996 ) , pp .
74 - 104 , hep - ph/9701275 f. becattini and u. w. heinz , z. phys .
c * 76 * , 269 ( 1997 ) .
f. becattini and g. passaleva , eur .
j. c * 23 * , 551 ( 2002 ) .
f. karsch , k. redlich and a. tawfik , eur .
j. c * 29 * , 549 ( 2003 ) .
m. cheng _ et al .
_ , phys . rev .
d * 77 * , 014511 ( 2008 ) .
d. h. rischke , h. stoecker , w. greiner and b. l. friman , j. phys .
g * 14 * , 191 ( 1988 ) .
m. xu , m. yu and l. liu , phys .
lett . * 100 * , 092301 ( 2008 ) .
r. hagedorn and j. rafelski , _ from hadron gas to quark matter .
1 _ cern th 2947 ( 1980 ) .
j. rafelski and r. hagedorn , _ from hadron gas to quark matter . 2 _ cern th 2969 ( 1980 )
. j. i. kapusta , phys .
d * 23 * , 2444 ( 1981 ) .
m. i. gorenstein , v. k. petrov and g. m. zinovev , phys .
b * 106 * , 327 ( 1981 ) .
s. kagiyama , a. minaka and a. nakamura , prog .
phys . * 89 * , 1227 ( 1993 ) .
m. i. gorenstein , w. greiner and s. n. yang , j. phys .
g * 24 * , 725 ( 1998 ) .
s. kagiyama , s. kumamoto , a. minaka , a. nakamura , k. ohkura and s. yamaguchi , eur .
j. c * 25 * , 453 ( 2002 ) , and references therein .
m. i. gorenstein , m. gazdzicki and w. greiner , phys . rev .
c * 72 * , 024909 ( 2005 ) . i. zakout , c. greiner and j. schaffner - bielich , nucl .
phys . a * 781 * , 150 ( 2007 ) .
k. a. bugaev , phys .
c * 76 * , 014903 ( 2007 ) .
i. zakout and c. greiner , phys .
c * 78 * , 034916 ( 2008 ) .
y. aoki , g. endrodi , z. fodor , s. d. katz and k. k. szabo , nature * 443 * , 675 ( 2006 ) .
a. chodos , r. l. jaffe , k. johnson , c. b. thorn and v. f. weisskopf , phys . rev .
d * 9 * , 3471 ( 1974 ) . f. becattini and l. ferroni , eur . phys .
j. c * 35 * , 243 ( 2004 ) .
d. stauffer , a. aharony , _ introduction to percolation theory _ ,
taylor & francis , london 1994 .
a. coniglio and w. klein , j. phys .
a * 13 * , 2775 ( 1980 ) .
s. fortunato and h. satz , phys .
b * 475 * , 311 ( 2000 ) .
s. fortunato and h. satz , nucl .
a * 681 * , 466 ( 2001 ) .
s. fortunato , f. karsch , p. petreczky and h. satz , nucl .
. proc .
suppl . * 94 * , 398 ( 2001 ) .
s. fortunato , f. karsch , p. petreczky and h. satz , phys .
b * 502 * , 321 ( 2001 ) .
v. koch , a. majumder , and j. randrup _ phys .
_ * 95 * 182301 ( 2005 ) . c. r. allton _
* d71 * 054508 ( 2005 ) .
r. v. gavai and s. gupta _ phys .
rev . _ * d73 * 014004 ( 2006 ) . | we formulate a simple model for a gas of extended hadrons at zero chemical potential by taking inspiration from the compressible bag model .
we show that a crossover transition qualitatively similar to lattice qcd can be reproduced by such a system by including some appropriate additional dynamics . under certain conditions , at high temperature , the system consist of a finite number of infinitely extended bags , which occupy the entire space .
in this situation the system behaves as an ideal gas of quarks and gluons . |
tidal disruption events ( tde ) of stars by massive black holes @xcite provide one of the ways to detect an inactive massive black hole in the center of a galaxy . stars whose orbit brings them too close to a black hole can be disrupted and a large fraction of their mass can be accreted .
this sudden accretion event is expected to produce a luminous uv and x - ray tde flare @xcite .
it has recently been argued that tdes can also produce jets @xcite .
these jets are expected to be detectable due to their non - thermal emission , and in radio they are likely to appear as transients that peak on the timescales of about a year @xcite after the tidal disruption
. swift j1644 + 57 ( swj1644 in the rest of the article ) is a non - thermal transient detected on march 28th.2011 @xcite .
it is thought to be prototype of a jetted tde ( jtde ) .
its radio emission @xcite is long - lived and complex .
one of the first attempts to explain the radio emission of swift j1644 + 57 using one - dimensional models and relativistic hydrodynamic simulations@xcite assumed that the emission is coming from the shocks that form when the jet interacts with the circumnuclear medium ( cnm ) .
this model is more analogous to a grb afterglow than to an agn .
while the model was successful in explaining the first few weeks of swj1644 radio emission , it turns out that the long - term emission is not easily accounted for by a radio afterglow from a single ultrarelativistic , narrow jet .
the reason for this is that the source rebrightened after @xmath0 days , the time at which the afterglow model predicted a steady decline ( * ? ? ?
* figure 1 ) .
the late time rebrightening implies that additional energy is injected in the external shock months after the burst@xcite .
this may happen if , for instance , the jtde is accompanied by a powerful mildly relativistic component@xcite .
in this contribution we focus on the numerical and technical aspects of jtde modeling . in section [ sec : num ] we describe the numerical codes used in the calculations and how they depend on each other .
the section [ sec : sim ] gives an overview of computational requirements and their consequences on the kind of models that we were able to produce .
the discussion and summary is given in the section [ sec : sum ] .
swj1644 is a distant object ( its redsfhit is @xmath1 @xcite ) , and is thus too small to be resolved in great detail .
thus , the only direct observations that we have are the total fluxes for different times and frequencies ( i.e. , multiwavelength light curves ) .
nevertheless , when simulating this source we need to construct a detailed , resolved model of the swj1644 jet , simulate its dynamics , compute its resolved emission , and only in the final stage integrate the resolved emission to compute the total flux to be compared to observations . in order to achieve this
we use two numerical codes so that the output of the first code ( _ mrgenesis _ ) is used as an input to the second one ( _ spev _ ) .
the former simulates the jet dynamics , while the latter computes its emission .
_ mrgenesis _ is a code developed to solve the equations of relativistic numerical ( magneto ) hydrodynamics _ in the laboratory frame ( i.e. , a frame at rest with respect to the _ distant source_)_. the code is massively parallelized using mpi and openmp librararies , and uses the parallel hdf5 library for file input and output .
we employ a method of lines for our simulations in which the spatial reconstruction is a based on the third order ppm reconstruction routines@xcite , and the time update is performed by means of a third order tvd runge - kutta scheme@xcite .
numerical fluxes are evaluated with the marquina flux formula@xcite .
we simulate swj1644 in two dimensions assuming axisymmetry and discretize the computational domain using spherical coordinates .
the black hole is assumed to be located at the origin and the tde occurs at scales that are much smaller than the ones we consider in our work ( we simulate neither the tde nor the jet launching ) .
we set the beginning of the numerical grid at @xmath2 cm .
the jet is injected through the inner boundary in an angular interval @xmath3 $ ] , assuming a constant lorentz factor @xmath4 , and a luminosity whose time dependence is @xmath5^{-5/3}$ ] .
@xmath6 is the time during which the jet luminosity is constant , and the peak lumiosity @xmath7 is computed from the total jtde energy @xmath8 through @xmath9 .
the jet interacts with a cnm that is assumed to be at rest with a density profile @xmath10 , where @xmath11 is the proton mass , @xmath12 is the particle number density at @xmath13 cm and @xmath14 is the power - law index of the density profile .
we assume that initially both the jet and the external medium are cold . during the simulation we periodically store the snapshots of the computational grid in order to obtain a sufficient coverage of the jet - cnm interaction to be able to compute the resulting emission with sufficient precision ( see section [ sec : sim ] ) .
once the relativistic hydrodynamics simulation has finished , the snapshots it produced are used by the code _ spev _ to compute the emission .
the spatial distribution of the hydrodynamics variables needs to be stored with sufficient frequency because of the space - time transformations that need to be performed to compute the observable radiation in the ( distant ) _ observer frame_. _ _ spev__@xcite is designed to process an arbitrary number of pre - computed sets of hydrodynamic states of the fluid ( i.e. , density , pressure , positions , velocities , magnetic field ) and to compute the resulting non - thermal emission .
the procedure can be summarized as follows : 1 .
* injection * : _ spev _ uses a set of lagrangian particles ( lps ) as a representation of the emitting volumes .
each lp has a defined position , shape and velocity .
attached to it there is a representation of the non - thermal electron ( nte ) energy distribution . in _
, new lps are created each time a snapshot is processed .
the position of the shock fronts that form as a consequence of the jet - cnm interaction is detected , and lps are injected into the shocked fluid immediately behind the front . since the actual process of particle acceleration occurs on temporal and spatial scales far below the ones we are able to simulate ( see e.g. , @xcite for a recent review ) , we use a phenomenological model to link the fluid state with nte distribution .
the model we use is a standard one @xcite : the energy contained in the ntes and in the stochastic magnetic field are assumed to be fractions @xmath15 and @xmath16 of the shocked fluid thermal energy .
the injection spectrum is assumed to be a power - law in energy @xmath17 , where @xmath18 is the power - law index and @xmath19 is the nte lorentz factor .
we determine the normalization and energy cutoffs of the distribution using a numerical procedure described in detail in sections 3.2 and 3.3 of @xcite .
transport , evolution and emission * : as it processes the snapshots , _ spev _ transports the injected lps , assuming that after they have been injected they move with the fluid .
an exception are the newly injected lps : between the snapshot in which the particle is injected and the subsequent one the two edges parallel to the shock front are assumed to move with different velocities ( one with the velocity of the shocked fluid and the other one with the faster velocity of the shock front ) .
this ensures that new lps correctly represent the increase in the volume due to newly shocked fluid , regardless of the numerical resolution or the frequency with which snapshots are stored .
simultaneously , the nte energy losses due to synchrotron radiation , as well as the energy gains(losses ) due to adiabatic compression(expansion ) are taken into account : @xmath20 where @xmath21 , @xmath22 , @xmath23 and @xmath24 are the time , fluid density , magnetic field energy density and external radiation field energy density ( all measured in the lp frame ) .
section 3.2 of @xcite describes an efficient numerical algorithm for solving eq .
[ eq : losses ] for each nte distribution in each lp .
at the same time that the particles are evolved , their non - thermal emission is computed ( see section 4 of @xcite for a description of an efficient numerical algorithm for computing synchrotron emission and absorption coefficients , and section a.2.1 of the same paper for a description of the way particles in two - dimensional axisymmetric models can be used to represent three - dimensional emitting volumes ) .
the coefficients computed in this step are stored in a global array that is used in the final step .
3 . * solving the radiative transfer equation * : after all the snapshots have been processed and the emission from all lps has been computed for all times , the time- and frequency - dependent emission can be computed .
we assume that there is a virtual screen ( `` virtual detector '' , vd ) located in front of the source , and the goal is to compute the intensity in each of the virtual pixels of the vd .
this is done by solving the radiative transfer equation @xmath25 where @xmath26 , @xmath27 and @xmath28 are the specific intensity , emissivity and absorption coefficient .
@xmath29 is the distance along the line of sight from the jet to a particular pixel and it relates the simulation time @xmath21 to the time of observation @xmath30 by @xmath31 , where @xmath32 is the distance from the vd to the source . in other words , for a given @xmath30 and @xmath21 , @xmath29 is the distance of a lp from which radiation is observed at time @xmath30 if that lp is emitting at a time @xmath21 .
we note that eq .
[ eq : intens ] needs to be solved in such a way that @xmath29 is monotonously increasing from the jet towards the vd .
therefore , after all the snapshots have been processed , the emission and absorption coefficients need to be sorted according to their @xmath29 and then eq . [ eq : intens ] is solved for each pixel . in practice
the positions of lps in steps ( i ) and ( ii ) are stored in a set of ( large ) intermediate files ( `` preprocessor '' files ) .
a postprocessor executes the radiation part of step ( ii ) by reading the preprocessor files and storing the coefficients in memory .
after all preprocessor files have been postprocessed , the step ( iii ) is executed .
both preprocessor and postprocessor parts of _ spev _ are paralellized using openmp .
the postprocessor , in particular , needs to be run on a machine with enough shared memory because all information produced in step 2 needs to be held in memory for the step 3 to correctly compute the intensities .
as mentioned in section [ sec : hydro ] , we model swj1644 jet in two - dimensions .
however , as will be explained in this section , the total cost in terms of computing time , disk storage and working memory of each two - dimensional simulation is high .
therefore , we first perform a number of much less costly tests using a one - dimensional approximation to two - dimensional models . the goal of those tests is to determine the values of a number of the parameters of the numerical simulation for which the results are satisfactory .
afterward , the two - dimensional simulations are performed using the values determined in one - dimensional tests . using _ mrgenesis _
we simulate one - dimensional spherically symmetric jets . for the purpose of computing their emission with _
spev _ we assume that their opening angles are @xmath33 .
we fix all the physical parameters of the jet ( @xmath34 erg , @xmath35 , @xmath36 rad , @xmath37 s ) and of the external medium ( @xmath38 @xmath39 , @xmath40 ) .
this corresponds to the inner relativistic component of the two - component jet in @xcite .
the numerical parameters that will be varied are : * @xmath41 : time interval between snapshots ( _ mrgenesis _ ) * @xmath42 : radial resolution ( _ mrgenesis _ ) * @xmath43 : angular resolution ( _ _ mrgenesis _ _ as angular cell size . ] ) * @xmath44 : number of nte energy distribution bins ( _ spev _ ) * @xmath45 : resolution of the virtual detector ( _ _ spev _ _ degrees ) , so that its emission is axially - symmetric and we only need to compute the emission along vd x - axis . for off - axis observations a ( much more costly ) two - dimensional vd calculation
is performed ( * ? ? ?
* section 5.2 ) .
we comment in detail on technical challenges of 2d simulations in next section . ] ) we choose as a default set of parameters @xmath46 s , @xmath47 cm , @xmath48 rad , @xmath49 , @xmath50 .
we follow the jet evolution until @xmath51 seconds . during the first @xmath52 seconds the snapshots are output every @xmath41 , and
afterward the time interval increases by a factor @xmath53 after every snapshot is output .
[ fig : var1d ] shows the light curves for the default model evaluated at 1.4ghz ( red line ) and the consequence of variations in the resolution of the hydrodynamic models , i.e. , as a result of changing @xmath42 .
as can be seen , relatively small differences of the order of a percent are visible at early times and after the light curve maximum .
the top right panel in fig .
[ fig : var1d ] shows the result of changing angular resolution . here
the differences are somewhat larger , especially when comparing the lowest and highest resolution ( @xmath54% at peak ) .
the middle left panel in fig .
[ fig : var1d ] shows how the results depend on the number of bins in which nte energy distribution of each lp is discretized .
we see that the convergence is only achieved for @xmath55 . in the middle right panel
we can see that the oscillations at early times are due to poor vd resolution .
the angular size of the jet is very small at those times and it is only properly resolved for @xmath56 for a linear pixel distribution ( covering the interval [ @xmath57 , @xmath58 cm ] ) , and @xmath59 for a logarithmic one ( covering the interval [ @xmath60 cm , @xmath58 cm ] ) . in the next section
we will explain how we solved the problem of noise for 2d simulation light curves , especially the off - axis ones where it is not feasible to distribute the vd pixels logarithmically .
finally , the bottom left panel shows the ( weak ) dependence on @xmath41 .
[ fig : var1d ] ghz radio light curves of one - dimensional tests ( section [ sec:1d ] ) .
the default model is shown in red .
the insets show the zoom - in around the light curve maximum .
_ top left panel _ : variations in @xmath42 . _ top right panel _ : variations in @xmath43 . _ middle left panel _ : variations in the number of bins @xmath44 .
note that @xmath61 and @xmath62 curves overlap .
_ middle right panel _ : variations in the number of vd pixels .
note that the curves for @xmath63 and @xmath64 with a logarithmic distribution of pixels overlap .
_ bottom left panel : _ variations in @xmath41 .
_ bottom right panel : _ emission from two dimensional two - component jet for different effective angular resolutions ( see section [ sec:2d]).,title="fig : " ] ghz radio light curves of one - dimensional tests ( section [ sec:1d ] ) .
the default model is shown in red .
the insets show the zoom - in around the light curve maximum . _
top left panel _ : variations in @xmath42 . _
top right panel _ : variations in @xmath43 .
_ middle left panel _ : variations in the number of bins @xmath44 .
note that @xmath61 and @xmath62 curves overlap .
_ middle right panel _ : variations in the number of vd pixels .
note that the curves for @xmath63 and @xmath64 with a logarithmic distribution of pixels overlap .
_ bottom left panel : _ variations in @xmath41 .
_ bottom right panel : _ emission from two dimensional two - component jet for different effective angular resolutions ( see section [ sec:2d]).,title="fig : " ] ghz radio light curves of one - dimensional tests ( section [ sec:1d ] ) .
the default model is shown in red .
the insets show the zoom - in around the light curve maximum . _
top left panel _ : variations in @xmath42 . _
top right panel _ : variations in @xmath43 .
_ middle left panel _ : variations in the number of bins @xmath44 .
note that @xmath61 and @xmath62 curves overlap .
_ middle right panel _ : variations in the number of vd pixels .
note that the curves for @xmath63 and @xmath64 with a logarithmic distribution of pixels overlap .
_ bottom left panel : _ variations in @xmath41 .
_ bottom right panel : _ emission from two dimensional two - component jet for different effective angular resolutions ( see section [ sec:2d]).,title="fig : " ] ghz radio light curves of one - dimensional tests ( section [ sec:1d ] ) .
the default model is shown in red .
the insets show the zoom - in around the light curve maximum . _
top left panel _ : variations in @xmath42 . _
top right panel _ : variations in @xmath43 .
_ middle left panel _ : variations in the number of bins @xmath44 .
note that @xmath61 and @xmath62 curves overlap .
_ middle right panel _ : variations in the number of vd pixels .
note that the curves for @xmath63 and @xmath64 with a logarithmic distribution of pixels overlap .
_ bottom left panel : _ variations in @xmath41 .
_ bottom right panel : _ emission from two dimensional two - component jet for different effective angular resolutions ( see section [ sec:2d]).,title="fig : " ] ghz radio light curves of one - dimensional tests ( section [ sec:1d ] ) .
the default model is shown in red .
the insets show the zoom - in around the light curve maximum . _
top left panel _ : variations in @xmath42 . _
top right panel _ : variations in @xmath43 .
_ middle left panel _ : variations in the number of bins @xmath44 .
note that @xmath61 and @xmath62 curves overlap .
_ middle right panel _ : variations in the number of vd pixels .
note that the curves for @xmath63 and @xmath64 with a logarithmic distribution of pixels overlap .
_ bottom left panel : _ variations in @xmath41 . _
bottom right panel : _ emission from two dimensional two - component jet for different effective angular resolutions ( see section [ sec:2d]).,title="fig : " ] ghz radio light curves of one - dimensional tests ( section [ sec:1d ] ) .
the default model is shown in red .
the insets show the zoom - in around the light curve maximum . _
top left panel _ : variations in @xmath42 . _
top right panel _ : variations in @xmath43 .
_ middle left panel _ : variations in the number of bins @xmath44 .
note that @xmath61 and @xmath62 curves overlap .
_ middle right panel _ : variations in the number of vd pixels .
note that the curves for @xmath63 and @xmath64 with a logarithmic distribution of pixels overlap .
_ bottom left panel : _ variations in @xmath41 .
_ bottom right panel : _ emission from two dimensional two - component jet for different effective angular resolutions ( see section [ sec:2d]).,title="fig : " ] in the previous section we showed how the light curves change as the five principal numerical parameters are changed .
most importantly for the 2d simulations , with the previous tests we have already obtained a very good proxy for the largest values of @xmath42 and @xmath65 we can use in order to reduce the computational cost without compromising the accuracy of the results . in the final simulations showed in @xcite we used @xmath66 cm and @xmath67 rad . in this section
we discuss the technical challenges of computing the emission even from these relatively low - resolution hydrodynamic simulations .
the complete two - dimensional evolution is stored in @xmath68 snapshots .
it was fortunate that hdf5 format permits data compression so that the compressed snapshots occupy @xmath69 gbytes . running the preprocessor with the default parameters of section [ sec:1d ] produces @xmath70 ( compressed ) intermediate files .
each of the first @xmath71 files contains the positions , geometry , velocity , magnetic field and the @xmath72 bins of the nt energy distibution of @xmath73 lps , while the last file only contains @xmath74 lps .
this means that in total there are @xmath75 two - dimensional emitting lps in step ( ii ) of the _ spev _ algorithm ( see section [ sec : rad ] ) .
since each of the lps has to be converted to a three - dimensional emitting volume , it may contribute to multiple vd pixels and also to multiple observing times , we soon realized that the realistic upper limit of @xmath76 tbyte of ram on our shared memory machine would not be enough to compute off - axis images at full resolution .
a remedy to this problem was to artificially reduce the angular resolution when preprocessing the hydrodynamic snapshots by a factor of @xmath77 or @xmath78 ( this would roughly correspond going from the second - best to third- of fourth - best angular resolution in top left panel in figure [ fig : var1d ] . ) .
this reduces the number of emitting volumes so that even off - axis radiative transfer calculations could be performed . for the purposes of the numerical tests we computed the on - axis light curves ( in complete analogy to the tests presented in section [ sec:1d ] ) .
bottom right panel in figure [ fig : var1d ] shows on - axis light curves using the full angular resolution , as well as the result of degrading it by factors of 2 and 4 . in figure
[ fig : observer ] we show a subset of the data that _ spev _ stores when postprocessing the intermediate files . in this particular case
we show all the contributions to the emission at @xmath79 seconds and at 15.4ghz .
the data shown originates from lps emitting at different times from different positions ( the lower the value of @xmath80 the earlier the lp had to emit for it to be observed at @xmath81 seconds ) . in other words ,
what is shown does not correspond to one particular snapshot of the jet evolution .
nevertheless , we know that the edge of the emitting region corresponds to the position of the bow shock of the jet .
we also see in the bottom right panel in figure [ fig : observer ] that a small region behind the shock is optically thin and most of the contribution to the emission comes from this region and from the surface at which the emission becomes optically thin ( the transition between light and dark blue regions ) .
using this type of analysis it is possible to understand , among other things , why the inverse compton cooling of the electrons is not very efficient at low frequencies ( though it may be important for higher frequency observations of swj1644 , see @xcite ) : the electrons cool after being accelerated by the shock , but since they are advected behind it , once they enter the opaque region ( dark blue ) , they stop contributing to the observed emission .
therefore the observer does not see a large effect of the inverse compton cooling ( * ? ? ?
* section 4.1.2 ) .
seconds at @xmath82 ghz .
_ top left _ : synchrotron emissivity in erg s@xmath83 @xmath39 hz@xmath83 . _
top right _ : synchrotron self - absorption in cm@xmath83 .
_ bottom left _ : specific intensity in erg s@xmath83 @xmath85 hz@xmath83 .
the intensity is the result of integrating equation [ eq : intens ] from bottom to top for each vertical line . _
bottom right _ : optical depth measured from the observer ( integral of the absorption from top to bottom along vertical lines ) .
the light blue colored region denotes the part of the emitting volume that is transparent , while the emission in the dark blue region is absorbed . as explained in section [ sec : onaxis ] , the information displayed here is a subset of the data stored in the global array used by _
spev _ to compute the intensity in each pixel ( see section [ sec : rad ] ) .
the luminosity is obtained by integrating the intensity along vertical lines in the top right panel ( taking into account that the emission is axially symmetric ) . ] here we comment on the procedure we use to reduce the numerical noise that shows at early times ( figure [ fig : var1d ] ) . as was commented in section [ sec:1d ] , the jet is small at early times , so that at a typical resolution of @xmath86 points per spatial direction it is very poorly resolved by the vd . due to memory limitations we were not able to use vd resolution of @xmath87 .
however , the resolutions of @xmath88 pixels were feasible .
therefore , we adopted the following procedure : 1 .
pre - determine the jet viewing angle @xmath89 and the times @xmath90 ( @xmath91 ) at which it is to be observed .
2 . compute the radio images using @xmath92 vd for each @xmath90 in a single calculation .
3 . for each @xmath90 , determine the rectangle that contains @xmath93 of the emission . then expand the rectangle by a safety margin .
4 . for each @xmath90 , independently re - compute the radio image by limiting the vd to the rectangle computed in the previous step , but using the resolution @xmath92 .
the above procedure enables us to compute the emission from early times without numerical artifacts times smaller than in the lower panels , but all are computed using the same vd resolution . ] .
the tradeoff is that we need to postprocess the intermediate files multiple times : once in step ( ii ) and then additional @xmath94 times in step ( iv ) .
the computing time does not necessarily scale as @xmath94 since , for each @xmath90 , the emission and absorption coefficients are only computed from those emitting volumes that contribute to the emission , while the rest ( typically corresponding to most of the volume of the model ) are ignored .
this is made efficient by indexing the intermediate files . for a given observing angle the calculations at @xmath95 different observing frequencies last @xmath96 days .
in this contribution we have discussed the technical details of the one and two dimensional simulations of the swj1644 jet . in section [ sec : num ] we explain in detail how the relativistic hydrodynamic simulations are coupled to the radiative transfer code to produce the synthetic observations ( light curves ) that can be directly compared to the observations . in section [ sec:1d ] we show that many one - dimensional tests can be used to probe the sensitivity of the computed light curves on the resolution of the hydrodynamic simulations , on the frequency of stored snapshots , on the resolution of the discretization in energy space and of the virtual detector . the results obtained helped to determine the resolution of the two dimensional simulation so that its computational cost is acceptable , but also so that the final light curves are as free of numerical artifacts as possible .
this procedure has been successful in enabling us to produce a viable two - component jet model of the swj1644 jet .
pm and maa acknowledge the support from the european research council ( grant camap-259276 ) , and the partial support of grants aya2013 - 40979-p , prometeo - ii-2014 - 069 and saf2013 - 49284-exp .
dg acknowledges support from the nasa grant nnx13ap13 g .
bdm acknowledges support from the nsf grant ast-1410950 and the alfred p. sloan foundation .
we thankfully acknowledge the computer resources , technical expertise and assistance provided by the `` centre de clcul de la universitat de valncia '' through the use of _ llus vives _ cluster and _ tirant _ , the local node of the spanish supercomputation network .
we use the open source software packages _ _ grace _ _ and _ _ ggplot2 _ _ to produce the graphics in this paper .
9 lacy j h , townes c h and hollenbach d j _ apj _ * 262 , 120 * rees m j 1988 _ nature _ * 333 * , 523 evans c r and kochanek c s 1989 _ apj _ * 346 * , 13 strubbe l e and quataert e 2009 _ mnras _ * 400 * , 2070 lodato g and rossi e m 2011 _ mnras _ * 410 * 359 giannios d and metzger b d 2011 _ mnras _ * 416 * , 20102 krolik j h and piran t 2012 _ apj _ * 749 * , 92 tchekhovskoy a , metzger b d , giannios d and kelley l z 2014 _ mnras _ * 437 * , 2744 metzger b d , williams p k g and berger e 2015 _ apj _ * 806 * , 224 bloom j s _ et al .
_ 2011 _ science _ * 333 * , 203 burrows d n _ et al .
_ 2011 _ nature _ * 476 * , 421 levan a j _ et al .
_ 2011 _ science _ * 333 * , 109 zauderer b a _ et al .
_ 2011 _ nature _ * 476 * , 425 berger e _ et al .
_ 2012 _ apj _ * 748 * , 36 zauderer b a _ et al .
_ 2013 _ apj _ * 767 * , 152 metzger b d , giannios d and mimica p 2012 _ mnras _ * 420 * , 3528 van velzen s , krding e and falcke h 2011 _ mnras _ * 417 * , l51 mimica p , giannios d , metzger b d and aloy m a 2015 _ mnras _ * 450 * , 2824 aloy m a , ibez j m , mart j m and mller e 1999 _ apjs _ * 122 * , 151 leismann t _
_ 2005 _ a&a _ * 436 * , 503 mimica p , giannios d and aloy m a 2009 _ a&a _ * 494 * , 879 colella p and woodward p r 1984 _ jcp _ * 54 * , 174 shu c w and osher s j 1984 _ jcp _ * 77 * , 439 donat r , font j a , ibez j m and marquina a 1988 _ jcp _ * 146 * , 58 mimica p _
_ 2009 _ apj _ * 696 * , 1142 sironi s , keshet u and lemoine m 2015 _ space science reviews _ * 191 * , 519 sari r , piran t and narayan r 1998 _ apj _ * 49 * , l17 mimica p and aloy m a 2012 _ mnras _ * 421 * , 2635 aloy c. , 2013 , master s thesis , univ .
de valencia . , https://riunet.upv.es/handle/10251/35350 kumar p , barniol duran r , bonjak and piran t 2013 _ mnras _ * 445 * , 3919 | in this work we focus on the technical details of the numerical simulations of the non - thermal transient swift j1644 + 57 , whose emission is probably produced by a two - component jet powered by a tidal disruption event . in this context
we provide details of the coupling between the relativistic hydrodynamic simulations and the radiative transfer code .
first , we consider the technical demands of one - dimensional simulations of a fast relativistic jet , and show to what extent ( for the same physical parameters of the model ) do the computed light curves depend on the numerical parameters of the different codes employed . in the second part
we explain the difficulties of computing light curves from axisymmetric two dimensonal simulations and discuss a procedure that yields an acceptable tradeoff between the computational cost and the quality of the results . |
the discovery of thousands of extrasolar planet systems ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) and the ongoing efforts to characterize the atmospheres on many of these planets greatly expand the scope of atmospheric science from the limited sample of planets in the solar system .
new regimes of thermal structure , dynamics , and escape have all been subject to intense scrutiny , especially on close - in extrasolar giant planets ( egps ) or hot jupiters that typically orbit within 0.1 au of their host stars , the objects for which we currently have the best observational constraints .
more recently , ionization of these atmospheres has been studied because of the recognition that it may strongly affect the temperatures and dynamics of hot exoplanet atmospheres ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these studies have employed a number of simplifying assumptions , primarily using the induction equation for resistive mhd , and reached often contradictory conclusions .
there is , as yet , no general agreement on the importance of ion drag or resistive heating in exoplanet atmospheres .
our study here expands on the previous analyses by considering photoionization as well as thermal ionization of the atmosphere .
we find that photoionization dominates on the dayside and is responsible for the primary effects of ion drag on the neutral atmosphere .
we also examine in detail the role of collisions and anisotropic resistivity , effects usually neglected or treated approximately in resistive mhd , and show that ion drag strongly couples different atmospheric regions and may dominate the dynamics of the middle and upper atmosphere of close - in egps bar . ] . because our knowledge of exoplanet atmospheric structure and dynamics is still rudimentary , detailed predictions of electrodynamics on egps are not yet possible .
nevertheless , our results show that ion drag may dominate the dynamics and temperature structure in some regions of the atmosphere and therefore can not be neglected .
transit observations have revealed the presence of escaping ions and neutral atoms in the upper atmospheres of close - in egps such as hd209458b , hd189733b , and wasp-12b ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as a result , previous studies of photoionization have mostly concentrated on the thermosphere ( @xmath0 10@xmath1 bar ) where its role in ionizing hydrogen and helium , heating the atmosphere and powering mass loss is well recognized ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the effective temperatures of many close - in egps , however , are high enough for alkali metals such as na and k to remain in the atmosphere as atoms instead of condensing or forming molecules .
both na and k have been detected on different hot jupiters ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and the detection of na on the well known transiting planet hd209458b constituted the first detection of an exoplanet atmosphere @xcite .
the ionization potentials of these metals are relatively low and thus they can be effectively ionized both thermally and by photoionization throughout the atmosphere .
as we will demonstrate , this leads to higher electron densities in the middle and lower atmospheres than in the thermosphere . as a results , the conductivities in close - in egp atmospheres are higher than in any planetary ionosphere of the solar system , and in fact closer to the conductivities in the outer atmospheres of stars like the sun .
we use a photochemical model to calculate the thermal ionization and photoionization rates between 100 bar and 10@xmath2 bar , and identify the basic conductivity regimes in close - in egp atmospheres based on the resulting electron densities and assumed magnetic field strengths .
the photochemical calculations are based on a prescribed temperature profile and properly include the shielding of the relevant metals from ionizing radiation by other atoms and molecules .
the details of these simulations are presented in a companion paper @xcite .
we use the results to constrain the composition of the ionosphere and , together with a generalized ohm s law , to study the basic mechanisms of electrodynamics in hot egp atmospheres .
we treat hd209458b ( @xmath3 1.32 @xmath4 , @xmath5 0.69 @xmath6 , @xmath7 0.047 au ) as a prototypical hot jupiter and use its properties in all of the calculations below .
the general results of this study , however , are not limited to any specific hot jupiter . as usual , we also assume that the rotation rate of hd209458b is consistent with tidal locking , given that the synchronization timescale of its rotation period is much shorter than the age of the system @xcite .
the magnetic fields of egps are currently unconstrained by observations , neither do we possess a robust theory for planetary dynamos with definite predictive capabilities . given these uncertainties , we assume a basic untilted dipole field with the jovian magnetic moment of @xmath8 1.56 @xmath9 10@xmath10 a m@xmath11 in all calculations below , unless otherwise explicitly indicated .
the importance of electrodynamics and ion drag varies greatly between different regions of the atmosphere .
in addition to the plasma density , it depends on the magnetization of the plasma that affects the conductivities .
magnetization is the coupling of the electrons and ions to the magnetic field in the atmosphere that is quantified by : @xmath12 where @xmath13 is the gyrofrequency of species @xmath14 and @xmath15 is its collision frequency with other species @xmath16 .
high magnetization of @xmath17 , where @xmath18 , implies strong coupling to the magnetic field whereas low values of @xmath19 indicate that collisions dominate .
electron densities on close - in egps are much higher than the electron densities in the ionospheres of the solar system . to illustrate this , figure [ fig : ionosphere ] shows the dominant ion and electron densities in the dayside ionosphere of hd209458b based on our model @xcite .
we used the same t - p profiles as @xcite to calculate the recombination and thermal ionization rates , and included neutral photochemistry and photoionization in the model .
previous models of ion drag on close - in egps have only considered thermal ionization of the atmosphere ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as shown by figure [ fig : ionosphere ] , photoionization of abundant metals such as sodium , potassium and magnesium produce electron densities above the 1 bar level that are 10100 times higher than the electron densities based on the saha equation @xcite .
this means that the upper atmosphere above the 110 mbar level can not be treated as an insulator , as suggested by @xcite .
naturally , photoionization does not take place in the night side . some of the electrons are , however , transported from the dayside to the night side by circulation even on tidally locked planets .
in addition to thermal ionization based on the night side t - p profile , we estimated the night side electron densities from a simple balance of advection and recombination between the dusk terminator and the anti - stellar point along the equator : @xmath20 where @xmath211 km s@xmath22 is the zonal wind speed ( e.g. , * ? ? ? * ) can be derived from scaling laws based on the thermal wind equation and plausible estimates of the energy balance .
they have also been predicted by numerous other circulation models . ] and @xmath23 .
the effective recombination rate that is calculated by using the night side t - p profile is @xmath24 , where @xmath25 are the recombination rates of the dominant ions and @xmath26 are the dayside ion fractions .
the night side electron densities are also shown in figure [ fig : ionosphere ] .
most of the dayside electrons above the 0.2 bar level are released by photoionization . in line with previous results from @xcite , the electron density in the thermosphere ( @xmath0 10@xmath1 bar )
is lower by about an order of magnitude in the night side than in the dayside . above the 10@xmath27 bar level , in the extended thermosphere
, the line of sight optical depth is low enough to allow the stellar xuv radiation to penetrate to the night side and significant differences between day and night side electron densities are not expected .
between 10@xmath1 bar and 0.2 bar , however , the dayside electron densities are orders of magnitude higher than the night side electron densities .
below the 0.2 bar level , where thermal ionization dominates , there are again practically no diurnal differences .
ion drag and resistive heating arise from electric currents in the atmosphere that are driven by perturbations to the magnetic field .
these perturbations , in turn , arise from variable electric fields induced by plasma motions .
the generalized ohm s law provides a relationship between the currents and the electric fields that depends on atmospheric conductivities . with faraday s law
, the ohm s law can also be converted into an induction equation that describes the evolution of the magnetic field perturbations . here
we outline the derivation of a generalized ohm s law that is valid for both weakly and strongly ionized media and thus sufficiently flexible to be valid in most planetary atmospheres .
we use this ohm s law in section [ sc : emdyn ] to model electrodynamics .
we note that the generalized ohm s law differs from the usual ohm s law used in studies of planetary ionospheres in the solar system that typically assume weak ionization .
the derivation is based on combining the ion , electron , and neutral momentum equations : @xmath28 where @xmath29 is the angular velocity of planetary rotation , and eliminating the electron velocity by using the definition of current density : @xmath30 the result for partly ionized media composed of a neutral species with mass @xmath31 and a single ion species with mass @xmath32 and charge @xmath33 is ( e.g. , * ? ? ?
* ) : @xmath34 \mathbf{j } \nonumber \\ & + & \frac{b}{e n_p } \left [ 1 - y_n \frac{k_{in } - ( m_e / m_i ) k_{en}}{k_{en}+k_{in } } \right ] \mathbf{j } \times \mathbf{b } \label{eqn : ion_ohm}\end{aligned}\ ] ] where @xmath35 is the plasma density , @xmath36 is magnetization ( for species @xmath14 colliding with @xmath16 ) , @xmath37 is the electric field in the rest frame of the ions , @xmath38 is the mass mixing ratio of the neutral atmosphere , and @xmath39 is a unit vector in the direction of the magnetic field lines . in this work
we ignore the inertia ( @xmath40 ) , coriolis , and pressure gradient terms in the ohm s law .
the inertia terms can be neglected because the typical advection timescale @xmath41 10@xmath42 s , where @xmath43 1 km s@xmath22 is the characteristic wind speed on egps ( e.g. , * ? ? ?
* ) , is much longer than the plasma - neutral collision timescale of 10@xmath441 s. by similar logic , the coriolis force term is also negligible .
the pressure gradient terms can be neglected if @xmath45 where @xmath46 is the speed of light , @xmath47 is the permeability of free space , @xmath48 is the pedersen resistivity ( see below ) , @xmath49 is the plasma beta , @xmath50 is the plasma alfven speed and @xmath51 is the ion plasma frequency .
we find that inequality ( [ eqn : prescond ] ) is true everywhere in our atmosphere for a planetary magnetic moment @xmath52 .
when @xmath53 , however , inequality ( [ eqn : prescond ] ) does not hold in the upper atmosphere .
it is convenient to write the ohm s law in terms of the electric field in the center of mass frame @xmath54 . in this frame
the velocity is given by : @xmath55 where @xmath56 , which is obtained by combining the ion and neutral momentum equations : @xmath57.\ ] ] thus the ohm s law in the center of mass frame is given by : @xmath58 the last term is negligible , and the result is @xmath59 where the parallel ( @xmath60 ) , cowling ( @xmath61 ) and hall ( @xmath62 ) resistivities are given by : @xmath63 .
\end{aligned}\ ] ] taking the curl of equation ( [ eqn : ohm1 ] ) and using faraday s law yield the induction equation : @xmath64 .
\label{eqn : find1}\ ] ] another way to write equation ( [ eqn : ohm1 ] ) is : @xmath65 where the conductivity tensor is given by : @xmath66 the conductivities are related to the resistivities by : @xmath67 where @xmath68 .
contrary to the fully ionized mhd equations that are used in many astrophysical applications or the weakly ionized approximations of ionospheric electrodynamics , equations ( [ eqn : find1 ] ) and ( [ eqn : ohm2 ] ) are simultaneously valid for both cases , thus providing a useful connection between the two regimes .
we note that as @xmath69 1 , the conductivities in the center of mass frame are equivalent to the conductivities in the neutral frame that we will use in section [ sc : emdyn ] .
in addition to the conductivities and the ohm s law , parameters such as the electron plasma frequency @xmath70 provide basic insights to the dynamic regime in the ionospheres of close - in egps .
in figure [ fig : pfreqs ] we compare the dayside plasma frequency with the electron - neutral collision frequency .
we have calculated the collision frequencies by including the neutrals h , h@xmath71 , and he , with rate expressions from @xcite and @xcite .
the results indicate that @xmath72 exceeds @xmath73 below the 0.01 bar level on the dayside and below the 10@xmath74 bar level in the night side , independently of the planetary magnetic field strength . in this regime electron plasma waves
are suppressed , the electrons ( and ions ) equilibrate with the neutral atmosphere and the medium does not behave as a plasma ( e.g. , * ? ? ?
the lundquist number measures the degree to which the background ( planetary ) magnetic field is perturbed by plasma motion , and it is given by @xcite : @xmath75 where @xmath76 is the alfven speed determined by using the total mass density @xmath77 .
the lundquist number is derived from the ratio of the advection term to the diffusion term in the basic induction equation : @xmath78 where @xmath79 is the perturbation to the background ( planetary ) magnetic field @xmath80 .
contrary to the standard form of the magnetic reynolds number @xmath81 ( e.g. , * ? ? ?
* ) , the lundquist number accounts for collisions with the neutral atmosphere and the stability of the planetary magnetic field that is generated in the interior of the planet .
figure [ fig : lundq ] shows @xmath82 as a function of pressure in the dayside atmosphere of hd209458b .
the results indicate that @xmath82 reaches unity near the 10@xmath110@xmath83 bar level whereas in the lower atmosphere the values of @xmath82 are very low .
this is not a coincidence .
the lundquist number is @xmath84 below the 10@xmath85 bar level where the electron and ion magnetizations are both less than unity .
this suggests that significant perturbations to the planetary magnetic field are unlikely in the deep atmosphere , mostly because collisions with the neutral atmosphere suppress plasma behavior .
we note that this result is in contrast to the recent work by @xcite . in the upper and middle atmosphere ,
however , plasma dynamics can alter the planetary magnetic field and a solution to the full induction equation may be required to characterize the time evolution of the magnetic field .
the conductivities depend on the collision frequencies that also determine the degree of magnetization within the plasma .
since the ohm s law ( [ eqn : ohm2 ] ) is given for a three fluid system and our ionosphere consists of several ions , we calculated the conductivities based on a mean ion species with the mass and collision frequencies given by : @xmath86 we have found that the difference between this approach and fully accounting for several ion species in the ohm s law is small .
figure [ fig : regimes ] shows the magnetization regimes and conductivities in the dayside atmosphere of hd209458b .
the atmosphere can be divided into four different regions based on the magnetizations @xmath87 , @xmath88 , and @xmath89 of the electrons and ions . at the highest altitudes in the m4 region @xmath90 and
the plasma behaves as if it were fully ionized . as a general rule
this is the case when the electron volume mixing ratio is @xmath91 10@xmath85 , independent of the magnetic field strength . in our model of hd209458b the transition to the fully ionized regime occurs above the 4 @xmath9 10@xmath27 bar level in the dayside and above the 5 @xmath9 10@xmath92 bar level in the night side .
the m4 region is not the focus of this work and hereafter we mostly concentrate on the m1 , m2 , and m3 regions . in the lower atmosphere , below the 1 mbar level ,
electrons and ions are not coupled to the magnetic field lines and both @xmath88 and @xmath93 1 .
thus the hall and cowling resistivities are both small and the parallel resistivity is dominated by the @xmath94 term . under these circumstances conductivity
is isotropic , and the induction equation ( [ eqn : find1 ] ) reduces to : @xmath95 where @xmath96 is the neutral flow velocity and the isotropic resistivity is : @xmath97 this result is also illustrated by figure [ fig : regimes ] where the pedersen conductivity is equal to the parallel conductivity in the m1 region and the hall conductivity rapidly decreases with increasing pressure .
we do not expect large scale currents and ion drag to be important in the m1 region . since both the electrons and ions are coupled to the neutral atmosphere
, there is no physical mechanism to enable charge separation that is required to support currents ( section [ subsc : m2m3 ] ) . in order to see this
, we summed the electron and ion momentum equations to obtain : @xmath98 \approx \mathbf{j } \times \mathbf{b } \nonumber \\ & \rightarrow & ( \mathbf{u}_p - \mathbf{u}_n ) \approx \frac{\mathbf{j } \times \mathbf{b}}{n_p m_i \nu_{in}}.\end{aligned}\ ] ] where @xmath99 is the plasma velocity .
if @xmath100 in the m1 region , we have to an order of magnitude @xmath101 which is much less than unity , implying that @xmath102 and thus that @xmath103 0 . in the m1 region , this relationship also explains the small values of @xmath82 that we discussed in section [ subsc : basics ] . in general
, it confirms our physical intuition that ion drag is not significant in the deep atmosphere
a result that arises partly from the proportionality of the isotropic conductivity @xmath104 to electron magnetization @xmath88 ( see also section [ sc : emdyn ] ) . in the m2 region the electrons are coupled to the magnetic field ( @xmath105 1 ) while ions are coupled to the neutral atmosphere by collisions ( @xmath106 1 ) .
the partial decoupling of first the electrons in the m2 region and then both the electrons and ions in the m3 region provides a natural mechanism of charge separation that is based on the interaction of the electron and ion gyromotion and collisions with the neutral wind . in the absence of collisions electrons and ions
rotate around the magnetic field lines with gyroradii that depend on their initial velocities .
collisions with the neutral wind cause the ions and electrons to drift in opposite directions , due to their opposite sense of rotation around the magnetic field lines , and generate a current . in the m1 region , however , collisions are so frequent that the gyromotion and the generation of significant currents are largely suppressed . under these circumstances
both the ions and the electrons simply follow the neutral wind . in terms of magnetization ,
the m2 region is similar to the e layer in the earth s ionosphere . on hd209458b , however , the peak pedersen conductivity is more than 5 orders of magnitude higher than the corresponding conductivity anywhere in the earth s ionosphere , and actually comparable to the corresponding m2 region conductivity in the solar chromosphere . the conductivity tensor in the m2 region is anisotropic and the hall conductivity is higher than the pedersen conductivity ( figure [ fig : regimes ] ) .
the parallel conductivity is generally higher than the perpendicular conductivities .
as we show in section [ sc : emdyn ] , the anisotropy of the conductivity tensor and high parallel conductivity enhance ion drag and resistive heating in the upper atmosphere .
again in terms of magnetization , the m3 layer is equivalent to the f layer in the earth s ionosphere , but the conductivities on hd209458b are still much higher than the conductivities in any planetary ionosphere .
both the electrons and ions are strongly coupled to the magnetic field ( @xmath107 1 ) , and the conductivity tensor remains anisotropic . typically the hall conductivity is smaller than the pedersen conductivity and the ratio @xmath108 is again much larger than unity . as ions begin to separate from the neutral flow and drift in the opposite direction to the electrons , the m3 region , similarly to the m2 region , can also support strong currents . finally , we briefly comment on the approximations of the ohm s law ( [ eqn : ohm1 ] ) that are valid in the m2 and m3 regions .
we find that @xmath109 throughout the atmosphere .
below the fully ionized regime it is also true that @xmath110 and @xmath111 1 .
thus , to a fairly good approximation the parallel , cowling and hall resistivity reduce to : @xmath112 which are roughly valid below the 10@xmath113 bar level on hd209458b .
these simplifications are typical in models of solar system ionospheres ( e.g. , * ? ? ?
* ) and they imply that the generalized ohm s law in the center of mass frame reduces to the ohm s law in the neutral frame for much of the atmosphere .
it is important to note that magnetization does _ not _ depend on the electron density or conductivities it only depends on the strength of the magnetic field and pressure ( see equation [ eqn : kst1 ] ) . for this reason ,
the boundaries of the m1 , m2 , and m3 regions are roughly the same in the night side as they are on the dayside .
the conductivities , however , are significantly different . for example , figure [ fig : night_pedc ] compares the pedersen conductivities between the day and night sides . due to the differences in the electron densities , the conductivities below the 10@xmath1 bar level are orders of magnitude lower in the night side . as a result ,
ion drag directly modifies dynamics below the 10@xmath1 bar level only on the dayside ( see section [ subsc : magnetic2 ] ) , although the effects of strong ion drag will be felt indirectly in the night side .
the boundaries of the m1 , m2 , and m3 regions obviously depend on the planetary magnetic field strength . in order to illustrate this , we calculated magnetization for different dipole moments ranging from 0.01 @xmath114 to 10 @xmath114 ( equatorial surface fields @xmath115 from 1.8 @xmath9 10@xmath1 t to 1.8 @xmath9 10@xmath85 t ) .
figure [ fig : bfield ] shows the lower boundaries of the m2 and m3 regions as a function of dipole moment .
it implies that for a reasonable range of dipole moments the m2 and m3 regions are always located above the 0.01 bar level .
given that scaling relations such as equation ( [ eqn : veldif ] ) are valid regardless of the electron density , this is the pressure range where ion drag is potentially significant .
for very low magnetic moments of @xmath116 0.01 @xmath114 ( e.g. , * ? ? ?
* ) , on the other hand , the m2 region is located entirely in the thermosphere and the m3 region merges into the m4 region ( i.e. , the ` fully ionized ' regime ) .
higher magnetic fields expose more of the atmosphere to ion drag , but the general effect of the surface field strength on the conductivities deserves further scrutiny .
clearly @xmath117 does not depend on @xmath115 .
pedersen conductivity , on the other hand , is always equal to @xmath117 in the m1 region and less than @xmath117 in the m2 region .
thus @xmath118 _ decreases _ with increasing @xmath115 in the m2 and m3 regions ( figure [ fig : night_pedc ] ) .
similarly , the hall conductivity @xmath119 decreases with increasing @xmath115 in the m2 and m3 regions while it increases with @xmath115 in the m1 region .
the overall effect of increasing @xmath115 is thus to enhance the degree of anisotropy in the conductivity tensor .
we note , though , that increasing @xmath115 leads to stronger ion drag despite the reduced conductivities ( and perpendicular currents ) because ion drag also depends directly on @xmath115 , and this compensates for the reduced conductivities ( section [ subsc : magnetic2 ] ) .
in this section we use a steady state ohm s law to confirm that currents and ion drag can significantly affect the dynamics and energy balance in the m2 and m3 regions .
our main focus is on the anisotropic conductivity that has not been considered in previous studies of exoplanet atmospheres . under the general assumption that @xmath120 0 , equation ( [ eqn : ohm2 ] ) can be written as : @xmath121 .
\label{eqn : estatic1}\ ] ] this equation can be solved self - consistently during each time step within a gcm to obtain a realistic description of ion drag . to the best of our knowledge ,
gcms that include ion drag in this manner have not been developed for any other planet than the earth ( e.g. , * ? ? ?
the coupling of electrodynamics to realistic circulation models is indeed a complex undertaking that we do not pursue here .
instead we concentrate on a few simple examples to demonstrate the qualitative effect of electrodynamics on planetary atmospheres . in all cases below
we work in the reference frame of the neutral atmosphere .
mid - latitude zonal jets arise naturally on rotating planets from geostrophic balance .
the best known example is the jet stream on the earth ( e.g. , * ? ? ?
we consider a highly idealized eastward jet in the northern hemisphere with a peak zonal wind speed of @xmath122 1 km s@xmath22 that is constant in longitude and follows a gaussian profile in latitude and a pressure profile shown in figure [ fig : box_wind ] .
although this wind profile is not predicted by any exoplanet circulation model , the pressure dependency of the jet is motivated by such models and the wind speed is designed to reach maximum at 0.01 bar , reduce to zero at 10 bar ( e.g. , * ? ? ?
* ) and decrease to about 500 m s@xmath22 near the 10@xmath1 bar level @xcite .
we note that these properties are typical of an equatorial jet on egps , but here we adapt them to a mid - latitude jet for illustration purposes .
we consider an equatorial jet with the same pressure and latitude dependency in section [ subsc : eq_jet ] . at mid - latitudes
the magnetic dipole field lines are almost vertical and the magnetic field coordinates can roughly be reduced to cartesian coordinates axis lies in the direction of the magnetic field , which defines the directions of the @xmath123 and @xmath124 axes . ] .
the components of the current density are : @xmath125 where @xmath123 is the meridional dimension and @xmath124 is the zonal dimension ( figure [ fig : box_geometry ] ) .
we assume that the meridional wind ( @xmath126 ) is negligible , the zonal wind ( @xmath127 ) is constant in longitude , and the conductivities only change with altitude . in solving equations ( [ eqn : currents ] )
we also assume that the magnetic field strength @xmath128 1.8 @xmath9 10@xmath74 t ( 1.8 g ) is constant with latitude and altitude .
thus @xmath120 0 yields : @xmath129 despite the fact that @xmath130 1 in the upper atmosphere , we use the static approximation with @xmath131 0 and write : @xmath132 where @xmath133 is the electrostatic potential , @xmath134 and @xmath135 is the source term .
we note that the symmetry of this example means that @xmath133 is constant with longitude and thus @xmath136 0 .
we solve equation ( [ eqn : estatic2 ] ) numerically for @xmath133 at pressures ranging from 10 bar to 10@xmath2 bar with zero current boundary conditions .
such boundary conditions are not appropriate in all circumstances .
the aim of this section , however , is to provide useful qualitative insight to atmospheric electrodynamics and these boundary conditions are sufficient for this purpose . as a result , figure [ fig : efields1 ] shows the polarization electric field @xmath137 at @xmath138 45@xmath139 for three cases that highlight different aspects of electrodynamics . in the first case we have assumed that the zonal jet is constant in pressure , and set @xmath140 1 s m@xmath22 everywhere ( dotted line )
this case shows that currents vanish and there is no ion drag when both the wind speed @xmath127 and @xmath141 are constant with @xmath142 because under these circumstances equation ( [ eqn : estatic2 ] ) yields @xmath143 everywhere .
this leads to the important conclusion that ( @xmath144 ) must change along the magnetic field lines for there to be any appreciable ion drag .
the second case demonstrates the general result that ion drag attempts to eliminate any variation of @xmath145 with altitude ( i.e. , along the magnetic field lines ) . in this case
we used the full wind profile ( figure [ fig : box_wind ] ) and the dayside conductivities ( figure [ fig : regimes ] ) .
the resulting polarization field @xmath146 , which is nearly constant with altitude , is shown by the solid line in figure [ fig : efields1 ] . in this case
( @xmath147 ) changes with altitude and thus the potential @xmath133 must also change with altitude . due to the diffusive effect of the left hand side in equation ( [ eqn : estatic2 ] ) the polarization field @xmath148 . as a result ,
the meridional current at @xmath138 45@xmath139 is positive between 8 @xmath9 10@xmath1 bar and 0.04 bar and negative at other pressure levels .
thus ion drag decelerates the zonal wind around the peak altitude of the wind profile and accelerates it at other altitudes .
if @xmath141 is constant in altitude , @xmath127 will tend to a constant value with altitude at each latitude .
if @xmath141 is not constant , the vertical wind profiles tend to the structure of the magnetic field based on equation ( [ eqn : estatic2 ] ) .
we note that although @xmath136 0 , there is a zonal current that arises from the hall effect with @xmath149 ( equation [ eqn : currents ] ) , which is constant with longitude . given that @xmath150 in the m2 region , this current is often comparable to or stronger than the meridional current .
the hall current flows perpendicular to both the magnetic field and the electric field and thus it does not contribute to resistive heating of the atmosphere ( section [ subsc : joule ] ) .
it does , however , lead to ion drag on the meridional winds .
since we consider the meridional wind negligible and generally do not concern ourselves with the meridional momentum balance in this example , we do not discuss the hall currents any further here . the third example ( dashed line in figure [ fig : efields1 ] )
demonstrates the error of ignoring the anisotropy of the conductivity tensor .
this case is otherwise identical to the second example above but we set @xmath151 .
the resulting electric field is up to 10 times lower in the thermosphere and 23 times lower throughout the atmosphere .
interestingly the effect of anisotropic conductivity is not limited to the m2 and m3 regions because the lower atmosphere has to adapt to the electric fields in the upper atmosphere in regions where the magnetic field lines traverse through the whole atmosphere .
this effect is potentially large and it can not be ignored , even when it complicates the electrodynamics significantly . for the remainder of this section we concentrate on the second example above . as we stated above , ion drag attempts to make @xmath127 constant with altitude at each latitude within the jet .
the current loop responsible for this is shown by figure [ fig : jx1a ] while the corresponding current density at @xmath138 45@xmath139 is shown by figure [ fig : jx1b ] . with upward vertical field lines in the northern hemisphere ,
the current flows anti - clockwise in the upper atmosphere and clockwise in the lower atmosphere .
the current density at @xmath138 45@xmath139 varies from about 0.1 ma m@xmath152 to a maximum of 30 ma m@xmath152 . in terms of latitude , the current is strongest near the peak of the wind profile and decreases away from the peak ( not shown ) .
thus deceleration around the 0.01 bar level and acceleration in the upper and lower atmospheres is strongest at @xmath138 45@xmath139 . as a result
, we expect a flattened version of the original zonal jet from the 0.01 bar level to appear over a wide range of altitudes if ion drag is sufficiently strong to affect the momentum balance .
as we argued before , ion drag can dominate the dynamics on close - in egps in the m2 and m3 regions . to see this we used the ratio of the ion drag term @xmath153 in the neutral momentum equation to the pressure gradient to define an ion drag length : @xmath154 where @xmath155 .
this is the _
shortest _ length scale at which acceleration by ion drag overtakes acceleration by the pressure gradients or advection .
the definition can be generalized by assuming that @xmath156 ( e.g. , * ? ? ?
* ) , in which case : @xmath157 figure [ fig : box_mom ] shows @xmath158 based on equations ( [ eqn : ld1 ] ) and ( [ eqn : ld2 ] ) .
as expected , it indicates that @xmath159 above the 10@xmath85 bar level . in this region
the neutral momentum balance is dominated by pressure gradients ( that arise from variations in stellar heating ) and ion drag .
the altitude peak of the jet , however , is at the 0.01 bar level where @xmath160 150 @xmath161 , suggesting that ion drag forces are less than 10@xmath152 of the pressure gradient forces .
this means that circulation driven by stellar irradiation in the lower atmosphere can strongly interfere with circulation in the upper atmosphere while being little affected itself .
the generalized version of @xmath158 tends to overestimate the importance of ion drag , although there can also be regions where it underestimates ion drag .
nevertheless , in this case the agreement between the two @xmath158 is quite good .
we note that figure [ fig : box_mom ] also confirms the conclusion in section [ subsc : m1 ] that ion drag is not important in the m1 region where @xmath162 .
finally , we discuss an important simplification that arises from the near - constancy of @xmath146 along the magnetic field lines when @xmath163 . under such conditions equation ( [ eqn : estatic1b ] )
can be integrated along the field lines and solved directly for the polarization field ( e.g. , * ? ? ?
* ) : @xmath164 this formula yields an electric field of @xmath165 -130 mv m@xmath22 at @xmath138 45@xmath139 , which is the same as the actual electric field in figure [ fig : efields1 ] .
this justifies our use of equation ( [ eqn : estatic3 ] ) below in section [ subsc : eq_jet ] where we consider the currents associated with an equatorial zonal jet . in this section
we consider an equatorial zonal jet , which is a more realistic example on close - in egps and can be used to highlight the effects of a more complicated magnetic field geometry ( figure [ fig : geom1 ] ) .
we assume the same pressure and latitude dependency as before ( figure [ fig : box_wind ] ) , but with the peak at the equator .
we reduce the problem to two dimensions and assume again that @xmath127 is constant with longitude while ignoring vertical and meridional winds .
we write equation ( [ eqn : estatic1b ] ) in a right - handed coordinate system defined by the direction of the magnetic field lines ( @xmath39 ) as : @xmath166 where the @xmath123 direction is always perpendicular to the local magnetic field line and @xmath124 is the zonal direction .
the path length element @xmath167 along the magnetic dipole field lines is : @xmath168 where @xmath169 is the radial distance to the field line at the equator and @xmath170 is latitude . _
it is important to note that , in contrast to section [ subsc : mid_lat_jet ] , we assume a south to north orientation for the magnetic field in this example_. if the field - aligned currents vanish at the footpoints of the magnetic field lines , we can use equation ( [ eqn : estatic4 ] ) to obtain the same simplification for @xmath146 as in equation ( [ eqn : estatic3 ] ) , with @xmath171 replaced by @xmath167 .
the boundary conditions in this case are somewhat more defensible than the zero current boundary conditions in the mid - latitude case .
the use of equation ( [ eqn : estatic3 ] ) , however , leads to some limitations on our results .
section [ subsc : mid_lat_jet ] indicates that this equation is roughly valid along magnetic field lines that pass to the region where @xmath172 . at mid - latitudes
this is the case for all field lines but at the equator only the field lines with l values ( @xmath173 ) in the m2 or m3 regions qualify ( e.g. , figure [ fig : geom1 ] ) .
thus we only show results above the 10@xmath85 bar level in this section .
figure [ fig : efields2 ] shows the net electric field perpendicular to @xmath39 in the neutral rest frame ( @xmath174 ) at the equator with the associated current density @xmath175 for three different cases that again highlight different aspects of electrodynamics .
the corresponding values of @xmath158 are shown in figure [ fig : eq_mom ] .
all of these examples are based on the dayside conductivity profiles from figure [ fig : regimes ] . in the first case the zonal wind @xmath176 1 km
s@xmath22 is constant everywhere ( dotted line ) . in the second case
the zonal wind speed varies with latitude but remains constant in pressure ( dashed line ) .
finally , in the third case we use the full wind profile ( figure [ fig : box_wind ] with the peak at the equator ) that varies with both latitude and pressure ( solid line ) . similarly to section [ subsc : mid_lat_jet ] , all of these cases produce zonal hall currents in the m2 region .
the hall currents lead to ion drag on the meridional and vertical winds that we ignore in this section .
the first case demonstrates that spatial variations in the dipole magnetic field alone can drive significant ion drag in the upper atmosphere . for constant @xmath127 , the polarization field @xmath146 along the magnetic field lines
is given by : @xmath177 where @xmath178 is the conductivity - weighted mean magnetic field strength along the field line with @xmath173 .
the dipole magnetic field strength along the field lines with a given l value is : @xmath179 we note that @xmath178 is heavily weighted towards values around the peak of @xmath118 near @xmath180 2 @xmath9 10@xmath74 bar ( figure [ fig : regimes ] ) .
for example , field lines with @xmath181 ( @xmath182 10@xmath1 bar ) reach the conductivity peak at @xmath138 14.5@xmath139 . at this point @xmath183 1.7
@xmath9 10@xmath74 t and thus @xmath184 -170 mv m@xmath22 .
the equatorial field strength is @xmath185 1.3 @xmath9 10@xmath74 t and thus the net equatorial electric field at the 10@xmath1 bar level is @xmath186 -40 mv m@xmath22 , which agrees reasonably well with the value of @xmath187 -30 mv m@xmath22 in figure [ fig : efields2 ] . in this case
@xmath188 is negative and thus a downward current flows at the equator with peak @xmath189 12 ma m@xmath152 near the lower boundaries of the m2 and m3 regions .
the current is strong enough to dominate the neutral momentum balance ( figure [ fig : eq_mom ] ) , and ion drag accelerates the zonal wind in the upper atmosphere until @xmath190 or until a balance with the local pressure gradients is achieved . the magnitude of the current density @xmath175 is roughly constant with latitude within 20@xmath139 of the equator but beyond 20@xmath139 it slightly decreases with latitude ( not shown ) .
this means that acceleration is most effective at low latitudes , possibly leading to the creation of an equatorial jet in the upper atmosphere instead of constant @xmath127 .
the second case demonstrates that ion drag does not in general act as a diffusive ( drag ) force . to show this
, we assumed that the zonal wind only varies with latitude but remains constant in pressure . as a result , @xmath188 is positive at the equator and drives an upward current with a peak @xmath189 34 ma m@xmath152 .
the polarization field @xmath146 along each magnetic field line depends on the conductivity - weighted mean @xmath191_l$ ] .
considering again the magnetic field line with @xmath192 1.124 , the wind speed at @xmath138 14.5@xmath139 is @xmath176 360 m s@xmath22 and @xmath191_l \approx$ ] 61 mv m@xmath22 .
thus the estimated @xmath193 bar@xmath194 69 mv m@xmath22 , which is not too different from the actual value of 52 mv m@xmath22 ( figure [ fig : efields2 ] ) .
the positive @xmath188 and upward current arise because the zonal wind speed decreases with latitude and the equatorial magnetic field lines in the upper atmosphere connect to higher latitudes near the boundary of the m1 region ( figure [ fig : geom1 ] ) .
the upward equatorial current decelerates the zonal wind and introduces a vertical gradient into the wind profile .
figures [ fig : exarrow ] and [ fig : efields3 ] illustrate the latitude dependency of the current by showing the direction and magnitude , respectively , of @xmath175 around the equator .
the initial zonal wind profile drives current loops on both sides of the equator .
the current density decreases with latitude and thus ion drag flattens the zonal jet in the upper atmosphere .
the return current near the 10@xmath85 bar level is supported by a negative @xmath188 ( figure [ fig : exarrow ] ) that accelerates the zonal wind in the stratosphere .
thus ion drag does act like a diffusive drag force at each pressure level within the zonal jet .
this diffusion , however , is limited by the structure of the dipole field and in the vertical direction ion drag introduces gradients into the flow instead of smoothing them .
the third case shows that the pressure profiles predicted by existing close - in exoplanet gcms for equatorial zonal jets are not necessarily realistic , given the strong influence of ion drag .
the pressure profile of the mean zonal wind shown by figure [ fig : box_wind ] that is motivated by these models leads to the lowest net electric field and current density in the upper atmosphere out of our examples ( figure [ fig : efields2 ] ) .
the resulting current loops are similar in morphology to the second example above , but the current density is lower . even in this case , however , @xmath158 is much less than @xmath161 in the upper atmosphere and this demonstrates the remarkable sensitivity of the m2 and m3 regions to ion drag . as we noted before ,
night side ion drag is not significant below the 10@xmath1 bar level ( section [ subsc : night_cond ] ) . in order to explore this further
, we used the night side pedersen conductivities ( figure [ fig : night_pedc ] ) to calculate @xmath188 and @xmath175 based on the equatorial jet that varies with latitude but is constant with pressure ( section [ subsc : eq_jet ] ) .
we find that the night side @xmath188 is not significantly different from the dayside , and in fact higher below the 10@xmath1 bar level ( figure [ fig : exnight ] ) . because of
the lower conductivities , however , the equatorial @xmath175 is substantially lower and as a result @xmath195 only above the 3 @xmath9 10@xmath1 bar level .
this agrees with the results in section [ subsc : night_cond ] . the strong influence of ion drag on the dayside ,
however , can still modify night side circulation in the middle atmosphere . also , in all of the examples ( sections [ subsc : mid_lat_jet ] and [ subsc : eq_jet ] ) we assumed that conductivity is constant with latitude and longitude . while this may be defensible for estimating the pedersen currents on the dayside , the drop in conductivities in the night side leads to peculiar consequences for the zonal hall current . under the artificial symmetry that we assumed earlier
the hall current is constant with longitude and thus @xmath136 0 . in the night side , however , the conductivity and thus the hall current decreases significantly below the 10@xmath1 bar level and this should lead to a non - zero @xmath196 that may point from dusk to dawn on the dayside .
our approach here can not be used to investigate this or other similar aspects of the complex dynamics between the day and night side that can only be studied properly by circulation models with self - consistent electrodynamics . in general , the consequences of the reduced conductivity in the night side are somewhat similar to reducing the magnetic field strength .
this is because @xmath158 and magnetization depend on the conductivities , which are reduced either by lowering the electron density or the magnetic field strength .
for example , a dipole moment of @xmath197 0.04 @xmath114 is consistent with the lower boundary of the m2 region near the 3 @xmath9 10@xmath1 bar in the dayside , in which case @xmath195 only above the 10@xmath1 bar level .
we note , however , that the actual boundary of the m2 region in the night side does not depend on the electron density and is relatively insensitive to the changes in the t - p profile .
we can further conclude from this that while there is always potential for strong ion drag in the m2 region , this potential can only be realized if the electron density is sufficiently high . a higher magnetic field strength ,
on the other hand , enhances ion drag in the lower atmosphere , but in a way that is not necessarily straightforward . as we stated before
, @xmath118 decreases in the upper atmosphere with increasing magnetic moment @xmath198 as the m2 region moves to higher pressures ( section [ subsc : magnetic ] ) .
for example , if we increase the dipole moment from @xmath114 to 10 @xmath114 , the magnitude of the equatorial @xmath188 increases substantially while the magnitude of the current density @xmath175 does not ( figure [ fig : exstrongb ] ) . ion drag , however , also increases with the magnetic field strength through @xmath199 and as a result the region where @xmath195 shifts down to 3 @xmath9 10@xmath85 bar on the dayside and 10@xmath83 bar in the night side , despite the fact that @xmath175 does not increase substantially .
heating of the neutral atmosphere by electric currents is often mislabeled as ` joule heating ' or ` ohmic dissipation ' . in reality , of course , neutrals do not carry charge and there is no reason for them to be directly affected by electric fields or currents
. instead the neutral atmosphere is heated by mechanical friction that is caused by collisions with the electrons and ions @xcite . by coincidence
, the combined heating rate of the plasma and the neutral atmosphere reduces to an expression that has the appearance of conventional joule heating @xmath200 under the same simplifying assumptions that lead to the ion drag term @xmath199 in the momentum equation .
actual joule heating or ohmic dissipation , however , only appears in the combined energy equation of the electrons and ions , and it is typically less significant than frictional heating by collisions . estimating the frictional heating rate in actual atmospheres is complicated because the partitioning of ion drag into kinetic and thermal energies can not be properly determined without self - consistent models of the momentum and energy balance .
in particular , ion drag strongly modifies the dynamics in the upper atmosphere , and the steady state circulation is not known _ a priori_. in the end the frictional heating rate depends on the degree to which the atmosphere is prevented from reaching equilibrium with respect to ion drag by other forces .
more sophisticated models are required to properly address this question in future studies . here
we concentrate on order of magnitude estimates of @xmath201 based on the mid - latitude and equatorial jets in sections [ subsc : mid_lat_jet ] and [ subsc : eq_jet ] , respectively , with the caveat that these estimates may turn out to be substantially different from the actual heating rates .
we used the mid - latitude jet to calculate the magnitude of frictional heating , and to explore the relative contribution of the field - aligned and perpendicular currents . in this case
the heating rate is : @xmath202 figure [ fig : joule1 ] shows the heating rates from field - aligned ( vertical ) and perpendicular currents at @xmath138 45@xmath139 and figure [ fig : joule2 ] shows a contour plot of @xmath201 based on the current loop in figure [ fig : jx1b ] .
these figures show that @xmath201 peaks near the maximum of the zonal jet , both in pressure and in latitude . in general , field - aligned currents
do not contribute significantly to the heating rate , other than in narrow regions where @xmath175 changes sign . in the upper atmosphere @xmath201
is generally very high .
for example , the bolometric stellar flux at hd209458b is @xmath203 9.8 @xmath9 10@xmath42 w m@xmath152 .
if this flux is absorbed by the atmosphere and distributed globally , the column - integrated frictional heat flux @xmath204 at @xmath138 45@xmath139 amounts to about 2.9 % of the effective stellar flux .
most of the frictional heating occurs above the 0.01 bar level where the volume heating rate @xmath201 is comparable or higher than the stellar heating rate .
remarkably , above the 10@xmath1 bar level @xmath204 is more than 1,000 times higher than the stellar x - ray and euv ( xuv ) heating rate @xcite . in reality , however , these estimates of @xmath204 are likely to be too high .
this is because the assumed zonal jet will be modified by ion drag above the 10@xmath85 bar level .
nevertheless , the frictional heating rates based on such a jet demonstrate the potential of ion drag to modify the energy balance in the upper atmosphere .
the influence of the zonal wind profile on the heating rate is further illustrated by figure [ fig : joule3 ] , which shows @xmath201 based on the three different equatorial jets in section [ subsc : eq_jet ] .
the highest @xmath204 above the 10@xmath85 bar level , about 5 % of the stellar flux , is obtained with the constant zonal wind of 1 km s@xmath22 everywhere . in this case
the energy is mostly deposited in the thermosphere where @xmath204 is over 10,000 times higher than the xuv heating rate .
such a high frictional heating rate is clearly unrealistic , and it demonstrates that the zonal wind will be modified by ion drag instead . in the second case of a constant zonal wind with pressure , @xmath204 is about 0.5 % of the stellar flux and still over 1,000 times higher in the thermosphere than the stellar xuv heating rate .
the more realistic mean zonal wind profile in the third case leads to @xmath204 that is about 0.02 % of the stellar flux and about 30 times higher than the stellar xuv flux in the thermosphere .
we have shown how the source of free electrons on close - in egps varies both horizontally and with pressure . in the thermosphere above the 10@xmath1 bar level
most of the electrons come from photoionization of h on the dayside . due to horizontal transport ,
the long lifetime of h@xmath205 , and the penetration of the stellar xuv radiation to the night side , the diurnal electron density difference in the thermosphere is at most an order of magnitude @xcite .
we note though that any estimates of the electron density in the thermosphere should be treated with caution .
basic models of the h / he ionosphere typically overestimate the peak electron density and fail to explain the relatively large diurnal differences observed in the ionospheres of jupiter and saturn ( e.g. , * ? ? ?
* ; * ? ? ?
these difficulties point to possible deficiencies in our general understanding of h@xmath71 physical chemistry ( e.g. , * ? ? ?
* ) that can only be resolved by a detailed study of the solar system giant planets .
a major difference from the solar system giant planets is that alkali metals such as na and k do not condense on hot close - in egps and are present in the atmosphere as atoms .
these metals have a relatively low ionization potential and they are effectively ionized both thermally and by stellar uv radiation . in a clear atmosphere the relevant uv radiation ( @xmath206 286 nm ) penetrates well past the thermosphere and photoionization dominates over thermal ionization down to the 0.2 bar level @xcite . as a result , there is a lower ionospheric peak near the 1 mbar level where the electron density is in fact much higher than in the thermosphere and far higher than the electron density in any ionosphere of the solar system .
the recombination time of the alkali metals , however , is relatively short and the electron density in the lower atmosphere , above the 0.2 bar level , reduces by many orders of magnitude in the night side . in this regard
our results differ from previous studies of ion ( magnetic ) drag on extrasolar giant planets that have all ignored photoionization and only included electron densities from thermal ionization . based on our results for the dayside , the assumption that the atmosphere can be treated as an insulator above the 0.01 bar level ( e.g. , * ? ? ?
* ; * ? ? ?
* ) is clearly invalid .
further , the dayside electron densities are generally much higher than expected from thermal ionization .
thus currents flow in the upper atmosphere and do not necessarily have to close in the interior . in this way the prevalence of photoionization in the upper atmosphere significantly complicates electrodynamics on close - in egps .
it also reduces the dependency of the atmospheric electron density on temperature only ( e.g. , * ? ? ?
* ) and makes the conductivities directly dependent on the stellar uv radiation .
the selection of appropriate heavy elements in calculating the electron densities is also important .
for example , @xcite included k only and missed an important contribution to the conductivities from the more abundant na .
@xcite , on the other hand , included h , he , na , k , li , rb , fe , cs , and ca while @xcite included the first 28 elements of the periodic table .
our calculations show that the inclusion of h , he , na , k , mg , and c is sufficient .
the inclusion of these elements is also defensible because , with the exception of he , there is at least some evidence for their presence in the atmospheres of close - in egps ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . with regard to the other species
, it is particularly important to exclude elements that are likely to condense in the deep atmosphere .
for example , ca and fe have strong absorption lines that should be detectable in the observations of hd209458b but are not seen @xcite . with the t - p profile for hd209458b
, these species are expected to condense below the 10 bar level and thus their lack in the data is probably not a coincidence .
related to this , none of the heavy elements in our calculations are fully ionized at the pressure levels probed by the observations and thus ionization itself can not explain the lack of detection .
this , however , means that the non - detection of k on hd209458b is very surprising , and we discuss this problem extensively in a companion paper @xcite .
in general , future observations of metal line absorption on different egps can potentially constrain the composition of the ionosphere and thus the role of ion drag in the atmosphere .
electrodynamics is strongly dependent on the electron density and magnetic field strength that vary with pressure .
we used the calculated electron densities to constrain the conductivity regimes on close - in egps and provide the appropriate form of the generalized ohm s law in each regime ( section [ subsc : conds ] ) .
the basic regimes of parallel ( ohmic ) , hall , and cowling ( ambipolar ) resistivity that depend on the electron and ion magnetization @xmath207 are well recognized in the solar atmosphere community ( e.g. , * ? ? ?
* ) , and have recently also been applied to magnetized circumstellar disks ( e.g. , * ? ? ?
the same regimes have also been studied extensively in the terrestrial ionosphere where they are separated based on the parallel , hall , and pedersen conductivities instead of resistivity ( e.g. , * ? ? ?
they have not , however , received any attention in the extrasolar planet community until now .
models of stellar atmospheres and planetary ionospheres have traditionally relied on two different approaches to electromagnetic effects that , as we have shown , are not as clearly separable on close - in egps .
the main differences between the solar chromosphere and planetary ionospheres are the higher ionization fraction and magnetic field strength on the sun . as a result ,
the solar magnetic field is susceptible to instability and perturbations from the plasma motion while perturbations to the magnetic dipole field of the earth by ionospheric electrodynamics are negligible . in the latter ( static )
case the magnetic field is fixed and atmospheric electrodynamics is described in terms of currents and electric fields ( the e , j paradigm ) while in the former ( dynamic ) case the magnetic field evolves continuously and magnetohydrodynamics is described in terms of the plasma flow and magnetic field ( the b , v paradigm ) .
in contrast to many astrophysical mhd applications , however , in both cases the conductivity tensor is anisotropic .
the ohm s law is commonly used in electrodynamic models while the induction equation is used by mhd models .
these equations describe the same physics and they can easily be converted from one to the other by using faraday s law . in section [ subsc : ohms ]
we present the generalized three - fluid ohm s law and induction equation in the center of mass rest frame that connect the ( e , j ) and ( b , v ) paradigms and are valid for any degree of ionization in the atmosphere .
such equations are necessary in the atmospheres of close - in egps where the ionization fraction and magnetic field strength change with both pressure and horizontal location . previous models of ion drag on egps ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) are limited to the lower atmosphere below the 1 mbar level ( the m1 region ) . in this region
the electron / ion gyrofrequencies are much lower than the electron / ion - neutral collision frequencies ( @xmath208 1 ) .
the mechanism that drives the current in the atmosphere ( section [ subsc : m2m3 ] ) relies on the right combination of gyromotion and collisions , and once @xmath209 1 this mechanism is largely suppressed .
further , in our case the plasma frequency @xmath210 below the 0.01 bar level on the dayside and below the 10@xmath74 bar level in the night side .
thus plasma behavior is suppressed in the lower atmosphere and both electrons and ions simply equilibrate with the neutral atmosphere ( e.g , * ? ? ?
our derivation of the ohm s law , however , appears to contradict this simple physical intuition . in particular ,
in the m1 region the generalized ohm s law reduces to equation ( [ eqn : perna ] ) with isotropic conductivity @xmath211 that has been used in practically all of the previous exoplanet studies .
there is no reason to doubt the general validity of this equation in the m1 region . in agreement with our basic physical intuition ,
however , the neutral momentum equation and the ohm s law can be used to show that ion drag is not important in the m1 region .
physically speaking , the length scale @xmath158 ( section [ sc : emdyn ] ) over which the current ( or ion drag ) becomes significant increases rapidly as @xmath88 decreases . as a general rule , @xmath162 in the m1 region , even if one assumes the maximum current generation by neutral winds with zero polarization electric fields .
it can be shown that @xmath158 is related to the lundquist length scale , which is the length scale over which the lundquist number @xmath212 1 .
the lundquist number plays the same role in predominantly neutral atmospheres as the magnetic reynolds number @xmath81 plays in fully ionized plasmas by quantifying the degree to which the background magnetic field is perturbed by plasma motions .
in contrast to the standard @xmath81 ( e.g. , * ? ? ?
* ) , the lundquist number accounts for plasma - neutral collisions and the stability of the magnetic dipole field that is continuously regenerated in the deep interior of the planet .
our results show that @xmath213 1 in the m1 region , indicating that perturbations to the dipole field due to plasma motions in the lower atmosphere should be small .
interestingly , our results provide some justification for studies that use equation ( [ eqn : perna ] ) in steady state to estimate frictional heating rates in the deep atmosphere based on predetermined wind patterns ( e.g. , * ? ? ?
this approach is technically acceptable because ion drag is negligible and the magnetic field should be largely unperturbed by dynamics in the atmosphere although it is limited by the neglect of the interior currents that give rise to the planetary magnetic field and currents in the upper atmosphere . from a physical perspective , there is always some degree of friction , no matter how small , between the neutral atmosphere and the plasma that attempts to separate from the neutrals .
our treatment of the electrodynamics in the m2 and m3 regions of close - in egps is entirely new . in these regions
the conductivity tensor becomes anisotropic , with the degree of anisotropy increasing with the magnetic field strength . in the m2 region
the electrons partly decouple from the neutrals while ions remain coupled to the neutrals ( @xmath105 1 , @xmath106 1 ) and in the m3 region both ions and electrons decouple from the neutrals ( @xmath107 1 ) .
the parallel ( magnetic field - aligned ) conductivity in the m2 and m3 regions is typically much larger than the perpendicular hall and pedersen conductivities . by design ,
the hall conductivity is always higher than the pedersen conductivity in the m2 region while the pedersen conductivity is much higher than the hall conductivity in the m3 region .
the partial decoupling of the plasma from the neutral atmosphere promotes current generation over short length scales ( section [ subsc : m2m3 ] ) and thus @xmath159 on the dayside above the 1 mbar level ( section [ sc : emdyn ] ) . with magnetic moments close to @xmath114 or higher , ion drag is so strong that it is likely to dominate the momentum balance in the m2 and m3 regions together with the pressure gradients arising from stellar irradiation .
in addition , the lundquist number approaches and exceeds unity in the upper atmosphere , implying that the dipole magnetic field is susceptible to perturbations by the currents in the atmosphere .
the short recombination time of the alkali metals , however , means that the influence of ion drag is much less prominent in the night side .
effectively , the reduced conductivity in the night side has the same effect as lowering the magnetic field strength on the dayside by a factor of 10100 .
this leads to an interesting dichotomy between the dynamics in the dayside and the night side . given that large scale dynamics on different exoplanets can potentially be constrained by observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , this dichotomy is worthy of further study by mhd models with anisotropic conductivity that can capture the turbulent transition across the terminator .
in contrast to the conductivities , magnetization does not change significantly from the dayside to the night side .
this means that while the m2 and m3 regions are always potentially affected by ion drag , this potential is only realized if the conductivities are sufficiently high . obviously , magnetization depends on the assumed magnetic field strength . with a higher magnetic moment of 10 @xmath114 ,
the m2 region extends down to the 0.01 bar level , thus also extending the region of anisotropic conductivity deeper into the atmosphere . with a magnetic moment of 0.1 @xmath114 ,
on the other hand , ion drag is only important in the thermosphere even on the dayside .
finally , in the m4 region above the 10@xmath27 bar level the ionization fraction is sufficiently high ( @xmath214 ) for the plasma to become completely decoupled from the neutrals . in this regime
the ohm s law in the plasma frame ( equation [ eqn : ion_ohm ] ) reduces to the fully ionized limit .
this region is important because on close - in egps it is believed to be escaping hydrodynamically , with detectable signatures in uv transit observations @xcite .
we find that ion drag on the neutral atmosphere in the upper atmosphere is relatively strong , raising the possibility that the capture of the plasma on closed field lines can also affect neutral mass loss rates @xcite .
the combined thermal and dynamical pressure of the upper atmosphere , however , exceeds the magnetic pressure at the equator for @xmath215 and this opens the possibility that the magnetic field lines follow the escaping plasma , in the same manner as they do in the solar wind .
further , even if ion drag is important in the upper atmosphere , it is not clear to what degree it inhibits mass loss from close - in egps as the capture of the escaping gas on closed field lines leads to significant heating of the upper atmosphere that can possibly restore the escape rate @xcite .
the development of a two fluid model to simulate the plasma and the neutral atmosphere with the induction equation for the magnetic field is a useful future avenue to addressing these questions . in order to better understand the basic mechanisms of electrodynamics
, we studied highly idealized mid - latitude and equatorial zonal jets .
these jets have a gaussian shape in latitude and a pressure dependency that is designed to mimic the zonal mean circulation on close - in egps such as hd209458b @xcite .
we relied on the static approximation ( equation [ eqn : estatic1 ] ) although , as we have explained , this approximation needs to be revised in future work .
in contrast to some of the previous studies ( e.g. , * ? ? ? * ; * ? ? ?
* ) , we showed that ion drag does not always act like a diffusive drag force and it can not be modeled as rayleigh drag .
instead , we find that ion drag attempts to eliminate variations of ( @xmath144 ) along the magnetic field lines and in many cases this can lead to the acceleration of the winds and the introduction of both vertical and meridional gradients into the flow .
we also showed that ion drag does not generally vanish at the equator . at mid - latitudes the dipole magnetic field lines are nearly vertical and
penetrate practically through the whole atmosphere .
thus the consequences of anisotropic conductivity in the m2 and m3 regions are also felt in the m1 region , and the polarization electric field is roughly constant and equal to the conductivity - weighted mean of @xmath216 along the magnetic field lines ( section [ subsc : mid_lat_jet ] ) . with a constant magnetic field ,
ion drag attempts to remove variations of the wind speed along the field lines .
the degree to which this is possible depends on the strength of the ion drag forces compared to the other terms in the momentum equation . because ion drag is much less important below the 1 mbar level than it is in the upper atmosphere , circulation in the lower atmosphere can drive the dynamics at higher altitudes . at the equator the horizontal magnetic field lines with l ( @xmath217 ) values in the m2 and m3 regions connect to relatively low latitudes .
there is only a narrow region in the equatorial m1 region where the field lines do not pass through the m2 and m3 regions that we ignore in this work .
the detailed structure of the currents in the upper atmosphere arising from the equatorial jet depends sensitively on the assumed vertical and meridional wind profiles but typically symmetric loops appear on both sides of the equator ( section [ subsc : eq_jet ] ) . we find that zonal winds predicted by current exoplanet gcms are likely to be significantly modified above the 1 mbar level on planets where the magnetic moment is comparable to @xmath114 .
we also find that the magnetic dipole field geometry alone drives significant ion drag in the upper atmosphere even when the zonal wind is initially constant with altitude and latitude .
we note that the zonal jets give rise to ion drag on the meridional and vertical winds by a zonal hall current that we ignore in this work .
this current does not contribute to frictional heating because under the ( artificial ) symmetry of our examples @xmath136 0 and the hall current is perpendicular to the net electric field in the @xmath123 direction .
in reality the hall current in the m2 region can not be constant in longitude because of the large drop in conductivity in the night side .
this means that a non - zero zonal electric field with either dusk to dawn or dawn to dusk orientation can develop on the dayside , depending on the direction of the zonal current ( that can vary with pressure ) .
this complication should be studied by more comprehensive models in the future that can properly address the diurnal differences on close - in egps .
finally , ion drag heats the atmosphere mostly through mechanical friction from plasma - neutral collisions @xcite .
we show that the frictional heating rate is strongly dependent on the assumed vertical and horizontal wind profile . whether or not the currents contribute to the kinetic energy of the atmosphere through ion drag or to the thermal energy through friction
depends on the momentum balance that needs to be worked out by self - consistent models .
nevertheless , our results demonstrate the potential for significant frictional heating in the upper atmosphere .
the column - integrated frictional heat flux @xmath204 based on the mid - latitude jet in section [ subsc : mid_lat_jet ] amounts to about 2.9 % of the globally averaged stellar bolometric flux .
the different equatorial jets in section [ subsc : eq_jet ] produce values of @xmath204 between 0.02 and 5 % .
these heat fluxes on the dayside are small but not negligible , and if they are deposited deep enough they could affect the radius evolution of the planet @xcite .
our calculations at the equator are limited to the upper atmosphere above the 1 mbar level , but the mid - latitude solution extends to 10 bar . in this latter case
we find that most of the frictional heat is actually deposited in the high altitude current loop above the 0.01 bar level . the largest relative effect from frictional heating is felt in the thermosphere where the heat flux is 1010,000 times higher than the stellar xuv heating rate , depending on the assumed zonal wind profile .
clearly , some of these heat fluxes are unrealistically high due to the lack of self - consistent dynamics .
nevertheless , frictional heating of the upper atmosphere by electric currents can be important because it effectively uses the energy from circulation that is powered by stellar insolation at lower altitudes .
even a small fraction of the bolometric flux that heats the lower atmosphere is large compared to the stellar xuv flux . in the thermosphere , additional heating can broaden the absorption lines of atoms and ions and affect the interpretation of uv transit observations @xcite .
frictional heating from ion drag can also enhance the energy - limited mass loss rate from close - in egps . in the solar system , ion drag can help to explain the abnormally high temperatures in the thermospheres of the giant planets @xcite .
self - consistent models of ion drag on giant planets , both in the solar system and extrasolar planets , are therefore an interesting direction for future research .
the aim of this work was to use electron densities calculated by a new photochemical model @xcite to constrain electrodynamics and ion drag in the atmospheres of close - in egps .
the electron densities are based on a solar composition of atoms such as h , mg , na and k that have been detected in hot jupiter atmospheres ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we used the properties of the transiting planet hd209458b in these calculations , although the results have general validity and are not necessarily intended to represent any specific planet .
we showed that close - in egps have ionospheres that extend to the lower atmosphere with electron densities that are much higher than the electron densities in any ionosphere of the solar system . as a result
, the electrical conductivity in the upper atmospheres of close - in egps can be comparable to the corresponding conductivity in the solar chromosphere . in line with @xcite
, we find that photoionization of h and he dominates in the thermosphere above the 10@xmath1 bar level . in this region
diurnal variations in electron density are also relatively small .
photoionization of metals such as mg , na , and k , however , creates a lower ionospheric peak near the 1 mbar level in the dayside where the electron density is higher than in the thermosphere .
photoionization dominates over thermal ionization on the dayside down to about 0.2 bar while thermal ionization produces most of the free electrons at deeper pressures . as a result ,
future observations of species like mg , na , and k can be used to characterize the ionosphere further , particularly if their abundances can be better constrained .
previous studies of ion drag on close - in egps ignored photoionization @xcite and thus severely underestimated the dayside electron densities .
phase curve observations of hot jupiters ( e.g. , * ? ? ?
* ) can potentially be used to constrain the effect of ion drag on global dynamics .
due to the neglect of photo - ionization , however , previous models have underestimated the day - night contrast in conductivity in the middle atmosphere .
they have also relied on rayleigh drag to simulate the global effects of ion drag . in our examples
ion drag attempts to eliminate variations of ( @xmath144 ) along the magnetic field lines .
sometimes this leads to behavior that is similar to a frictional drag force at other times ion drag actually introduces vertical and horizontal structure to the wind profiles .
we also find that ion drag does not vanish at the equator ( e.g. , * ? ? ?
in fact vertical currents that lead to ion drag at the equator are generally required to close the current loops generated by an equatorial zonal jet . in light of these findings
, we feel that the consequences of ion drag on global dynamics may have to be re - evaluated
. another potentially interesting consequence of photo - ionization is that it makes the conductivities and thus ion drag and resistive heating dependent on the stellar uv fluxes rather than temperature only .
this raises the possibility of correlations between the stellar uv output and observable properties of transiting planets such as brightness temperatures .
these correlations might be difficult to recognize because the atmospheric temperatures obviously also depend on the stellar flux , and much of the uv radiation responsible for ionizing , say , na and k is in the blackbody continuum .
nevertheless , both fuv radiation and x - rays , that do not necessarily correlate strongly with the stellar continuum , penetrate to the middle atmosphere and could form the basis for looking for such correlations in the future .
in addition , our results point to particularly strong effects of ion drag in the upper atmosphere .
we divided the atmosphere into four different regimes ( m1 , m2 , m3 , and m4 ) based on conductivities and magnetization , and provided the three - fluid non - ideal mhd equations that are valid in these regimes .
previous models have not included the atmosphere above the 1 mbar level ( the m2 , m3 , and m4 regions ) . in these regions
the electrons and ions partly decouple from the neutral atmospheres and the combination of gyromotion and less frequent collisions leads to charge separation and strong currents .
we demonstrate that , with the jovian magnetic moment , currents and ion drag in the middle and upper atmosphere are sufficiently strong to dominate the momentum balance , together with the neutral pressure gradients that arise from stellar insolation and subsequent dynamics .
ion drag also affects the thermal energy balance of the atmosphere mostly through frictional heating arising from plasma - neutral collisions @xcite .
based on our examples , we find column - integrated local maximum heat fluxes that amount to 0.45% of the globally averaged stellar bolometric flux , deposited mostly in the upper atmosphere above the 0.01 bar level .
the potential for frictional heating is particularly strong in the thermosphere where ion drag may enhance mass loss rates and affect the interpretation of uv transit observations ( e.g. , * ? ? ?
* ; * ? ? ?
. a magnetic field can also affect the morphology of the escaping plasma around planets like hd209458b , and this opens the possibility for future observations of escaping atmospheres as a means to place some constraints on the magnetic field strengths , even though the interpretation of such observations is likely to remain ambiguous for a long time to come . ion drag and frictional heating
may also prove to be important on the giant planets in the solar system .
a recent study by @xcite shows in principle that heating arising from electrodynamic coupling of jupiter s thermosphere and stratosphere can help to explain the relatively high temperatures in the jovian thermosphere .
in addition to direct heating , other studies on saturn raise the question of whether wind - driven ion drag could change the predicted circulation and help to redistribute energy from the polar auroral region to low and middle latitudes more efficiently ( e.g. , * ? ? ? * ; * ? ? ?
in the absence of direct constraints on the winds in the upper atmosphere , however , these questions can only be properly addressed with self - consistent models of the momentum and energy balance .
ttk and rvy acknowledge support by the national science foundation ( nsf ) grant ast 1211514 .
jykc acknowledges the hospitality of the kavli institute for theoretical physics ( kitp ) , santa barbara .
pl acknowledges financial support from the programme national de plantologie ( pnp ) . we thank j. erwin , r. jokipii , j. giacalone , and v. vasyliunas for useful discussions and correspondence .
adams , f. c. 2011 , apj , 730 , 27 bai , x - n . , & stone , j. m. 2013 , apj , 769 , 76 batalha , n. m. , et al .
2013 , apjs , 204 , 24 batygin , k. , & stevenson , d. j. 2010 , apjl , 714 , 238 batygin , k. , stevenson , d. j. , & bodenheimer , p. h. 2011 , apj , 738 , 1 baumjohann , w. , & treumann , r. a. 1997 , basic space plasma physics .
( london , england : imperial college press ) ben - jaffel , l. , & hosseini , s. s. 2010 , apj , 709 , 1284 ben - jaffel , l. , & ballester , g. e. 2013 , a&a , 553 , a52 charbonneau , d. , brown , t. m. , noyes , r. w. , & gilliland , r. l. 2002 , apj , 568 , 377 cho , j. y - k .
2008 , phil .
a , 366 , 4477 fossati , l. , et al .
2010 , apjl , 714 , 222 garcia munoz , a. 2007 , plan .
space sci .
, 55 , 1426 griemeier , j .- m . , et al .
2004 , a & a , 425 , 753 guillot , t. , burrows , a. , hubbard , w. b. , lunine , j. i. , & saumon , d. 1996 , apjl , 459 , l35 hallett , j. t. , shemansky , d. e. , & liu , x. 2005 , apj , 624 , 448 kliore , a. j. , et al .
2009 , j. geophys .
res . , 114 ,
a04315 knutson , h. a. , et al .
2007 , nature , 447 , 183 koskinen , t. t. , cho , j. y - k . , achilleos , n. , & aylward , a. d. 2010a , apj , 722 , 178 koskinen , t. t. , yelle , r. v. , lavvas , p. , & lewis , n. 2010b , apj , 723 , 116 koskinen , t. t. , harris , m. , yelle , r. v. , & lavvas , p. 2013a
, icarus , 226 , 1678 koskinen , t. t. , harris , m. , yelle , r. v. , & lavvas , p. 2013b
, icarus , 226 , 1695 koskinen , t. t. , et al . 2013c , icarus , 226 , 1318 lavvas , p. , et al .
2014 , submitted to apj leake , j. e. , et al .
2013 , arxiv:1310.0405v2 lecavelier des etangs , a. , et al .
2012 , a & a , 543 , l4 linsky , j. l. , et al .
2010 , apj , 717 , 1291 menou , k. 2012 , apj , 745 , 138 moses , j. i. , et al .
2011 , astrophys .
j. , 737 , 15 .
mller - wodarg , i. c. f. , moore , l. , galand , m. , miller , s. , & mendillo , m. 2012 , icarus , 221 , 481 perna , r. , menou , k. , & rauscher , e. 2010a , apj , 719 , 1421 perna , r. , menou , k. , & rauscher , e. 2010b , apj , 724 , 313 perna , r. , heng , k. , & pont , f. 2012 , apj , 751 , 59 rauscher , e. , & menou , k. 2013 , apj , 764 , 103 richmond , a. d. 1995 , j. atmos .
phys . , 57 , 1103 richmond , a. d. , & thayer , j. p. 2000 , in magnetospheric current systems , geophysical monograph ( american geophysical union ) , 118 , 131 rogers , t. m. , & showman , a. p. 2014 , apjl , 782
, 4 salby , m. l. 1996 , fundamentals of atmospheric physics ( san diego , ca , usa : academic press ) schunk , r. w. , & nagy , a. f. 2000 , ionospheres : physics , plasma physics , and chemistry ( cambridge , england : cambridge university press ) sing , d. k. , vidal - madjar , a. , desert , j .- m . , lecavelier des etangs , a. , & ballester , g. 2008a , apj , 686 , 658 sing , d. k. , et al . 2008b , apj , 686 , 667 sing , d. k. , et al .
2011 , a & a , 527 , a73 showman , a. p. , & guillot , t. 2002 , a & a , 385 , 166 showman , a. p. , et al .
2009 , apj , 699 , 564 showman , a. p. , fortney , j. j. , lewis , n. k. , & shabram , m. 2013 , apj , 762 , 24 smith , c. g. a. 2013 , icarus , 226 , 923 snellen , i. a. g. , de kok , r. j. , de mooij , e. j. w. , & albrecht , s. 2010 , nature , 465 , 1049 tenenbaum , p. , et al .
2014 , apjs , 211 , 6 trammell , g. b. , arras , p. , & li , z .- y .
2011 , apj , 728 , 152 udry , s. , & santos , n. c. 2007 , annu .
astrophys . , 45 , 397 vasyli@xmath218nas , v. m. , & song , p. 2005 , j. geophys .
, 110 , a02301 vidal - madjar , a. , et al .
2003 , nature , 422 , 143 vidal - madjar , a. , et al .
2004 , apjl , 604 , 69 wardle , m. 2007 , astrophys .
space sci .
, 311 , 35 yelle , r. v. 2004 , icarus , 170 , 167 yelle , r. v. , & miller , s. 2004 , in jupiter . the planet , satellites and magnetosphere ( cambridge , england : cambridge university press ) , 185 | strong ionization on close - in extrasolar giant planets suggests that their atmospheres may be affected by ion drag and resistive heating arising from wind - driven electrodynamics .
recent models of ion drag on these planets , however , are based on thermal ionization only and do not include the upper atmosphere above the 1 mbar level .
these models are also based on simplified equations of resistive mhd that are not always valid in extrasolar planet atmospheres .
we show that photoionization dominates over thermal ionization over much of the dayside atmosphere above the 100 mbar level , creating an upper ionosphere dominated by ionization of h and he and a lower ionosphere dominated by ionization of metals such as na , k , and mg .
the resulting dayside electron densities on close - in exoplanets are higher than those encountered in any planetary ionosphere of the solar system , and the conductivities are comparable to the chromosphere of the sun . based on these results and assumed magnetic fields , we constrain the conductivity regimes on close - in egps and use a generalized ohm s law to study the basic effects of electrodynamics in their atmospheres .
we find that ion drag is important above the 10 mbar level where it can also significantly alter the energy balance through resistive heating .
due to frequent collisions of the electrons and ions with the neutral atmosphere , however , ion drag is largely negligible in the lower atmosphere below the 10 mbar level for a reasonable range of planetary magnetic moments .
we find that the atmospheric conductivity decreases by several orders of magnitude in the night side of tidally locked planets , leading to a potentially interesting large scale dichotomy in electrodynamics between the day and night sides . a combined approach that relies on uv observations of the upper atmosphere , phase curve and doppler measurements of global dynamics , and visual transit observations to probe the alkali metals
can potentially be used to constrain electrodynamics in the future . |
dissipation of energy from a physical system to a thermal bath takes place as a result of a fluctuating force that acts on the system because of its coupling to the bath .
the dissipative friction force is proportional to the correlation function of the fluctuating force .
the force resulting from coupling to a chaotic system is also fluctuating .
the question that will be studied in the present paper is on what time scales it leads to friction and what is the relation of this friction to the autocorrelation function of the fluctuating force .
an example is slow particle coupled to a fast particle so that the motion of the fast particle is chaotic for each position of the slow particle .
various models for dissipation of energy from a slow particle by a fast one have been developed .
to our knowledge the first models of this type were introduced in the context of nuclear physics @xcite . in particular , a model where particles move within a region bounded by a deforming boundary , modeling the nuclear surface was studied @xcite .
the energy transferred between the boundary and particles enclosed inside was calculated classically and quantum mechanically in the framework of some approximations .
recently some detailed numerical simulations were performed along these lines and the regime of validity of various approximations was tested @xcite .
the dissipation for a wide class of model systems was explored by wilkinson and austin [ 4 - 6 ] in the framework of random matrix theory ( rmt ) , relevant for a situation where the motion of the fast particles is chaotic .
the relation to landau - zener tunneling was also studied in these works .
a different rmt model was studied by mizutori and aberg @xcite . in all these studies dissipation
was found .
in addition , a different approach , aimed at emphasizing the relation to many body problems was introduced @xcite . a unified picture of many of these models has recently been presented by cohen @xcite .
a systematic investigation of the interaction of a slow system with a fast one is possible with the help of multiple scale analysis . under such conditions ott demonstrated @xcite that the phase space volume enclosed by the energy surface of the fast particle is an adiabatic invariant , namely its change is much slower than that of the fast particle hamiltonian .
it has been demonstrated for various conditions that it is indeed an adiabatic invariant @xcite . in the present paper
we study the behavior of a slow particle that is coupled to a fast chaotic system .
a model for such a system , that is quite general , and has been studied by berry and robbins ( br ) @xcite and by jarzynski @xcite in the framework of multiple scale analysis , is defined by the hamiltonian : @xmath0 the phase space coordinates of the slow particle are @xmath1 and its mass is @xmath2 . for simplicity
it is coupled only through its position to the fast system whose phase space coordinates are @xmath3 .
the latter system has the property that if @xmath4 is kept fixed it is fully chaotic .
the crucial feature of the system we wish to study in this work is that it exhibits a wide separation of time scales- the evolution of the fast system , characterized by the time scale @xmath5 , is so rapid that it explores all of the phase space accessible to it energetically before the slow particle , characterized by the time scale @xmath6 , moves appreciably .
the adiabaticity parameter is @xmath7 .
one way to realize this is to couple two particles with a mass ratio of @xmath8 , as one can see by rescaling the equations of motion .
we now turn to analyze the dynamics generated by the hamiltonian ( [ brjhamiltonian ] ) with the approximation that the slow particle evolves under the influence of the _ average _ force exerted on it by the fast system , which can be treated as a system described by a slowly varying hamiltonian .
the time dependence of this hamiltonian is determined by the dynamics of the slow particle .
first the classical dynamics is outlined and later the quantum mechanical behavior is summarized . in the case of ( [ brjhamiltonian ] )
the average force is given by : @xmath9 where @xmath10 is a normalized probability density in the fast particle phase space .
a formalism that includes _ fluctuations _ was developed by jarzynski @xcite .
the results of this paper do not depend on these fluctuations and therefore the formalism of br will be used .
the probability density satisfies the liouville equation : @xmath11 written in a way that emphasizes that the evolution of the fast system is indeed on a much shorter time scale than the time scale on which the fast hamiltonian changes .
@xmath12 denotes poisson brackets taken with respect to @xmath13 . with the aid of the multiple scale expansion : @xmath14 berry and robbins @xcite were able to calculate the force acting on the slow particle up to first order in @xmath15 : @xmath16 to leading order , the force is given by the classical analogue of the born - oppenheimer force : @xmath17 where @xmath18 is chosen such that the phase space volume enclosed by the energy surface of the fast particle , @xmath19 , is constant .
the leading correction to @xmath20 includes a velocity dependent force : @xmath21_{e = e({\bf r})},\ ] ] where @xmath22 and : @xmath23 in the last equation : @xmath24 denotes the microcanonical average and @xmath25 .
finally , @xmath26 in ( [ brcij ] ) is the classical trajectory obtained by integrating the equations of motion generated by @xmath27 with fixed @xmath4 backwards from @xmath13 for the time @xmath28 .
in addition to the velocity dependent force ( [ f1 ] ) , and to the same order in @xmath15 , there is a force that does not depend on velocity @xcite .
this force can be expressed as a gradient of a time dependent potential , and therefore is a correction to the born - oppenheimer force ( [ f0 ] ) . because of the form of @xmath29 , it describes two qualitatively different forces .
the first of these is geometric magnetism , and it is related to the antisymmetric part of @xmath30 .
this force has been studied analytically in the systems under discussion by br @xcite and numerically by berry and sinclair @xcite .
the second force is associated with the symmetric part of @xmath30 and is related to deterministic friction .
this force has been studied in @xcite .
a central question that can be addressed at this point is under which conditions does the slow particle feel friction due to the velocity dependent force @xmath31 . in order for this to happen @xmath30 of ( [ f1 ] )
has to have a positive definite symmetric part . the behavior when the fast system is quantum mechanical , to first order in @xmath32 , has also been studied by br @xcite .
they found that in this case @xmath30 is an antisymmetric tensor , meaning that the system exhibits _ only _ geometric magnetism and _ no _ friction .
this difference is a result of the discreteness of the quantum spectrum .
the quantum correlation function corresponding to the symmetric part of @xmath33 of ( [ brcij ] ) is : @xmath34,\end{aligned}\ ] ] and the infinite time integral over it vanishes @xcite . in order to understand how the crossover between the classical and quantum behavior occurs , it is instructive to calculate the integral of the correlation function over a finite time .
since the discordance between the quantum and the classical models appears in the symmetric part of @xmath30 it can be studied for the case where @xmath4 is replaced be a scalar , time dependent parameter , @xmath35 . in this case
@xmath36 will be denoted by @xmath37 for simplicity ( @xmath38 and @xmath39 on the rhs of ( [ brcijq ] ) will then be replaced by the scalar parameter @xmath35 ) .
following br we assume in the calculation that the initial state is an eigenstate of the hamiltonian , @xmath40 .
it was verified by br that their result holds also if the initial state is a mixture .
our calculation can also be extended to a mixture leading to the same results .
if the initial state is a pure state but not an eigenstate of the hamiltonian one can check that within the assumptions of the paper the results are similar to the ones found if the initial state is an eigenstate of the hamiltonian .
the finite time integral that should be calculated then is : @xmath41 the correlation function and its integral may depend on the initial state @xmath42 .
this dependence has been suppressed in the notation for simplicity . in what follows
, it will become clear that it is not important for the results of the present paper .
taking the classical limit @xmath43 for any finite @xmath44 and then the limit @xmath45 should result in a non - vanishing value of @xmath46 , while for any finite value of @xmath47 , @xmath46 should vanish .
the friction on the time scale @xmath44 is proportional to @xmath48 as can easily be inferred from ( [ f1 ] ) and ( [ briij ] ) . the experimental meaning of this statement will be clarified in what follows . in order to understand the mechanism of this discordance , br studied a model correlation function where the levels were equally spaced .
they found that the function is periodic in time with period @xmath49 , where @xmath50 is the level spacing .
moreover , in the classical limit , which in their model corresponds to taking @xmath51 , @xmath37 approaches the classical correlation function . for systems
whose classical dynamics is chaotic , the energy levels are not equally spaced , but rather are distributed according to rmt @xcite .
the long time behavior of @xmath48 is determined by the levels nearest to @xmath42 , namely @xmath52 , as can be seen from ( [ brcijq ] ) .
therefore , one may expect behavior different from the one found for the equally spaced spectrum .
the natural question to ask is whether there is a characteristic time scale for the crossover between the quantum behavior of the integral @xmath48 and its classical behavior .
the most nave answer to this question is that the characteristic time scale is the heisenberg time , because it is the only time scale in the problem , and it is on this time scale that the quantum to classical crossover usually takes place . on the other hand
, one can argue that there is no time scale for this crossover at all @xcite .
in rmt the probability for two consecutive levels to be separated by a distance @xmath53 behaves like @xmath54 for small spacings [ 19 - 22 ] .
consequently , @xmath55 for long times , where here @xmath56 denotes the rmt ensemble average , and @xmath57 is the integral ( [ indef ] ) for some @xmath58 .
the answer given by the analysis presented in this paper is surprising . for the gaussian orthogonal ensemble ( goe ) ( @xmath59 ) one
indeed finds that @xmath60 decays like @xmath61 , but for the gaussian unitary ensemble ( gue ) ( @xmath62 ) one finds that it decays like a gaussian with a characteristic time proportional to the heisenberg time or vanishes after the heisenberg time depending of the parametric dependence on @xmath35 .
the integral @xmath63 was also calculated for other values of @xmath58 .
why is the nature of the decay of @xmath64 important ?
there is the quantum - classical discordance that has already been mentioned , and one would like to analyze the scale that is required to observe the crossover between the regimes .
it is relevant for some experiments , that will be mentioned below .
in addition , there is the issue of the relevance of such an effect .
the time over which the correlation function decays should be compared with other time scales present in the specific system studied .
one such time scale is @xmath65 , which is the time scale for the breakdown of the first order of the multiple scale analysis .
non - perturbative effects , such as landau - zener tunneling , become important on a time scale of @xmath66 . in realistic experiments
there is also the time scale for quantum decoherence @xmath67 . in order to observe the classical to quantum crossover discussed in the present work
@xmath60 should exhibit substantial decay for @xmath68 and of the order of @xmath67 .
the model discussed in the present work is relevant for some experimental situations .
consider for example a molecular beam prepared in a classical configuration , where initially many levels are substantially populated .
the beam travels in a slowly varying field @xcite .
consequently the internal dynamics in the molecules is in a slowly varying potential . on short time
scales the behavior is classical , the integral of the correlation function is positive and energy is absorbed in the motion of the internal degrees of freedom .
on longer time scales the integral of the correlation function decays to zero and therefore one realizes that actually no energy is absorbed by the molecules .
the outcome of the experiment depends on the time scale of decoherence , @xmath67 , that is , the energy absorption by the internal degrees of freedom is proportional to @xmath69 .
another example is of quantum dots where parameters are varied adiabatically , like in pumping experiments , but with dots that are closed , so that their spectrum is discrete @xcite . in section [ secrmt ] a specific rmt model
is defined . for this model
the ensemble average of the integral of the correlation function ( [ indef ] ) is calculated analytically .
it is demonstrated that most of the contribution for long times originates from the nearest neighbor levels . in section [ nnd ]
the integral of the correlation function ( [ indef ] ) , predicted by the nearest neighbor level spacing distribution , is calculated for various distributions , including some that are not related to rmt models . in section [ secdiscussion ]
the results of this work are analyzed and discussed .
the main purpose of this paper is to study ( [ brcijq ] ) and its finite time integral ( [ indef ] ) in the framework of rmt .
the reason for this is that random matrices describe many characteristic properties of realistic quantum - chaotic systems [ 17 , 25 - 28 ] . for simplicity
our random matrices will depend on one external parameter @xmath35 , so that we shall study the one dimensional version of ( [ brcijq ] ) , where the vector of parameters @xmath70 is replaced by the scalar @xmath35 . for each random matrix
we shall be able to calculate both @xmath37 and @xmath48 .
we wish to construct a random matrix model for some of the levels of a system whose quantum hamiltonian depends on some parameter .
the @xmath71 levels we wish to simulate by the random matrix lie within an energy strip of width @xmath72 , that depends on @xmath71 . later on we shall be interested in studying the semiclassical limit .
the meaning of taking this limit in the present context is to increase the density of levels in the @xmath73-strip : in the classical limit the spectrum becomes continuous .
we shall work with the well studied gaussian ensembles @xcite .
these are defined through four parameters : @xmath58 which defines the symmetry of the random matrices , their dimension @xmath71 , the mean value of their elements and their variance ( given through the parameter @xmath74 ) .
all of these need to be chosen carefully in terms of parameters of the physical system , being simulated by the random matrix .
the symmetry of the ensemble should be chosen to correspond to the real system . if the latter exhibits time reversal symmetry then the ensemble is the orthogonal one @xmath75 .
if the system does not exhibit this symmetry then the ensemble is unitary @xmath76 .
the mean value of the matrix elements can be chosen to be zero , which corresponds to setting the ensemble average of the reference level @xmath77 .
the mean level density satisfies the semi - circle law @xcite : @xmath78 for large @xmath71 , where we have used the definition : @xmath79 . in subsection [ secshort ]
the relation between the parameters of the rmt model and the ones of the physical system will be discussed .
we shall use the hamiltonian introduced by austin and wilkinson @xcite and model a parameter dependent system by the @xmath80 random matrix : @xmath81 where @xmath82 are @xmath80 random matrices from the same goe or gue ensemble .
there are three advantages to working with @xmath83 : ( a ) it belongs to the same ensemble that @xmath82 belong to , ( b ) the derivatives of its matrix elements belong to the same ensemble too because : @xmath84 ( c ) the matrices @xmath83 and @xmath85 are statistically independent . if we insert @xmath83 and @xmath85 into eq .
[ brcijq ] , and then perform the ensemble average , we obtain : @xmath86\bigg>,\ ] ] where @xmath87 and @xmath56 denotes rmt ensemble averaging .
the subscript @xmath58 will denote the symmetry : @xmath59 for goe and @xmath62 for gue .
the correlation function @xmath88 and its finite time integral @xmath57 are ensemble averaged .
the @xmath56 will be dropped from these quantities for notational simplicity .
the statistical independence of @xmath85 and @xmath83 implies @xmath89\bigg>,\ ] ] while the fact that @xmath85 belongs to the same ensemble as @xmath83 implies @xmath90 for @xmath91 , leading to : @xmath92\bigg>.\ ] ] we would like to make the connection between @xmath93 and the form factor : @xmath94e^{i2\pi\epsilon \tau}d\epsilon,\ ] ] where @xmath95 is the density of states and @xmath96 is the smoothed density of states .
the variable @xmath97 is the energy measured in units of the mean level spacing @xmath98 and @xmath99 is time in units of the heisenberg time , @xmath100 .
. [ cnnocorr ] can be written in the following form : @xmath101e^{i2\pi\eps\tau}d\epsilon,\ ] ] where the @xmath102 results from the omission of the term @xmath103 in the sum ( [ cnnocorr ] ) . comparing the last equation with ( [ formfactor ] )
one can see that : @xmath104 in this work we are mainly interested in the time integral of the correlation function ( [ indef ] ) : @xmath105.\ ] ] in the limit @xmath106 the term in the square brackets is just @xmath107 , the two point spectral correlation function at zero energy separation .
it vanishes as a result of level repulsion . in order to perform actual calculations we make use of the well known form factor for goe and gue @xcite .
it is standard to define @xmath108 for goe it is given for example in mehta s book ( see @xcite p. 137 ) : @xmath109 & \tau\leq1 \\ -1+\tau\ln\left[\frac{2\tau+1}{2\tau-1}\right ] & \tau\geq1 \end{array } \right.,\ ] ] from which one obtains : @xmath110\right ] & \tau\leq1 \\
\frac12-\left[\frac12\left(1-\tau\right)+\frac12\left(\tau^2-\frac14\right)\ln \left[\frac{2\tau+1}{2\tau-1}\right]\right ] & \tau\geq1 \end{array } \right .. \ ] ] for @xmath111 @xmath112 falls off asymptotically as @xmath113 for gue ( see @xcite p. 95 ) : @xmath114 from which one obtains : @xmath115 & \tau\leq1 \\ 0 & \tau\geq1 \end{array } \right .. \ ] ] in order to compare the analytical results that hold in the infinite @xmath71 limit with results for finite @xmath71 , ensembles of the @xmath80 matrices @xmath116 and @xmath117 of ( [ model ] ) , belonging to goe or gue were generated numerically .
the integral of the correlation function @xmath57 was then calculated numerically , by ensemble averaging .
the results are presented in figs .
[ figgoe ] & [ figgue ] for goe and gue respectively and compared with ( [ igoe ] ) and ( [ igue ] ) .
units where @xmath118 were used ( see subsection [ secshort ] ) . the numerical errors in the figures
were calculated according to @xmath119 where @xmath120 is the number of matrices used in the ensemble average . for completeness
the results for gse are obtained with the help of ( see @xcite p. 166 ) : @xmath121 following the calculation performed for the other ensembles one finds : @xmath122 & \tau\leq 2 \\ 0 & \tau\geq 2 \end{array } \right .. \ ] ] a crucial approximation in the long time regime is ( [ cnapproxlong0 ] ) , where only the contribution of the nearest neighbors is taken into account .
this approximation is tested in figs .
[ figgoebruno ] & [ figguebruno ] for the model ( [ model ] ) . in the numerical test
the nearest and next - nearest neighbor spacings are taken from the center of the matrix .
we see from these figures that the approximation ( [ cnapproxlong0 ] ) is quite reasonable and it improves as time increases .
the reason is that as time grows the oscillations of the various terms in ( [ cndef ] ) with energy become stronger , enhancing the dominance of the nearest neighbor contributions .
the approximation is better for goe than for gue because the small @xmath53 weight in the integral ( [ inapproxlong ] ) is larger . the asymptotic behavior of @xmath143}/i$ ] for large real @xmath144 is @xmath145}/(\sqrt{\pi } y)$ ] @xcite .
therefore , we find for goe : @xmath146 and there is _ no _ characteristic time for the crossover from classical to quantum behavior .
this decay is extremely close to ( [ igoea ] ) hence for large @xmath134 the contribution of the nearest neighbor spacings accounts for nearly all of @xmath147 .
for gue the falloff is much faster since in the large @xmath134 limit the gaussian in ( [ inapproxlong1gue ] ) dominates , and there _ is _ a characteristic time @xmath148 ( in units of the heisenberg time ) , for the crossover to quantum behavior . for long time @xmath149 is extremely small , while it is exactly zero according to ( [ igue ] ) .
therefore also for gue ( [ cnapproxlong0 ] ) is an excellent approximation for the long time limit . for completeness let us present the results for the gaussian symplectic ensemble ( gse ) ( for which @xmath150 ) , analogous to ( [ inapproxlong1goe ] ) and ( [ inapproxlong1gue ] ) . for this
we need to substitute the appropriate @xmath136 into ( [ inapproxlong ] ) .
according to eq . 202 in @xcite : @xmath151},\ ] ]
leading to : @xmath152}.\ ] ] one can immediately see that there is a characteristic time scale for the quantum to classical crossover , as there was in the gue case . as for the gue case
@xmath153 is extremely small for long times , while ( [ igse ] ) gives zero . in figs .
[ figgoefluc ] & [ figguefluc ] it is demonstrated that the long time results of the simulation tend to the analytical formulas as more and more members are included in the ensemble .
in the previous section the integral of the correlation function @xmath183 was studied for rmt models .
one conclusion was that it is dominated by the nearest neighbor level spacings .
it was found also that there is a big difference between goe on the one hand and gue and gse on the other . in this section
we shall study the decay of the correlation function if the nearest neighbor level spacing distribution is given and will not rely on the assumption of an invariant rmt ensemble .
we will also assume that the fluctuations of @xmath184 are not important and this quantity can be replaced by the constant @xmath185 .
it will be assumed that the distribution of the nearest neighbor level spacings is of the form : @xmath186}\ ] ] where @xmath187 and @xmath188 are constants .
the rmt distributions ( [ psgoe ] ) , ( [ psgue ] ) and ( [ psgse ] ) for @xmath189 are of this form .
the integral for the correlation function ( [ inapproxlong ] ) takes the form : @xmath190 } \sin(2 \pi \tau s)=\frac{c}{\pi a ^{\beta/2}}{\cal i}_\beta ( y)\ ] ] where @xmath191}\sin{s y},\ ] ] with @xmath192 .
the decay of this function will be explored in what follows , for arbitrary @xmath58 .
one can verify that this function satisfies the ordinary differential equation : @xmath193 .
\ \end{array}\right.\ ] ] making the substitution : @xmath194}\ ] ] and changing to the variable @xmath195 , one arrives at a new differential equation : @xmath196f(x)=0\\ f_\beta(0)=0;~~ f'_\beta(0)=\frac12\gamma\left[\frac{\beta+1}2\right ] .
\ \end{array}\right.\ ] ] eq .
[ diffeqn3 ] is a well known equation , and its solutions are parabolic cylinder functions @xcite : @xmath197 , \func{v}\left[1/2 -\beta , x\right]$ ] . for arbitrary @xmath58 ,
the solution of ( [ diffeqn3 ] ) is a linear combination of these two functions : @xmath198 + b_\beta~\func{v}\!\left[\frac12-\beta , x\right],\ ] ] in which @xmath199 are constants to be determined from the initial conditions . in particular , if @xmath58 is an odd natural number it turns out that @xmath200 , if it is an even natural number @xmath201 while for non - integer @xmath58 both @xmath202 and @xmath203 are non - vanishing .
this fact is very important for the asymptotic behavior of @xmath204 , which is determined by the large @xmath205 behavior of @xmath206 , v\left[\frac12 -\beta , x\right]$ ] @xcite : @xmath207&\sim&x^{\beta-1}\exp{\left[-x^2/4\right]}\\
\label{vasymp } \func{v}\!\left[\frac12 -\beta , x\right]&\sim&\sqrt{\frac2\pi}~x^{-\beta}\exp{\left[x^2/4\right]},\end{aligned}\ ] ] as @xmath208 .
one can immediately deduce that the first solution is subdominant for all @xmath58 except for the special case when it is an even natural number , for which @xmath201 in ( [ solgen ] ) . since @xmath209 is proportional to @xmath183 and @xmath144
is proportional to @xmath134 , for any fixed @xmath210 ( @xmath211 ) : @xmath212 as @xmath111 , while for @xmath213 ( @xmath211 ) : @xmath214}\tau^{\beta-1},\ ] ] as @xmath111 .
this is precisely the type of behavior found for the random matrix ensembles treated explicitly ( [ inapproxlong2goe ] , [ inapproxlong1gue ] & [ inapproxlonggse ] ) .
finally , one wonders what would the analogous results be in the case of the poisson distribution . returning to ( [ cnapproxlong0 ] )
, one obtains for the correlation function : @xmath215 where @xmath216 is the variance of the off - diagonal matrix elements between nearest neighbor levels , which we shall leave unspecified , as we are only interested in the behavior as a function of @xmath134 .
an approximation similar to the one leading to ( [ cnapproxlong1 ] ) was made .
for the poisson case the nearest neighbor spacing distribution is : @xmath217,\ ] ] leading to : @xmath218 the integral over this expression is : @xmath219 the asymptotic behavior of which is given by : @xmath220 as @xmath111 . in the absence of level repulsion one
indeed finds that the integral of the correlation function does not vanish .
this does not result in any contradiction with the classical limit where for an integrable system @xmath221 .
the reason is that for integrable systems the eigenfunctions of neighboring energy levels typically have an exponentially small overlap ( in @xmath222 ) .
this small overlap is the physical reason for the poisson distribution .
therefore in the classical limit @xmath223 exponentially fast in @xmath222 .
many systems that are neither integrable nor chaotic were found to have a semi - poisson distribution @xcite .
for this case , that shows linear level repulsion , the nearest neighbor spacing distribution is : @xmath224,\ ] ] the integral of the correlation function was calculated along the lines of the calculation for the poisson distribution .
the result is : @xmath225 it decays like @xmath226 in the long time limit . for the semi - poisson distribution
there is level repulsion and indeed the integral of the correlation function decays with time .
the correlation function of the force applied by a fast quantum system on a slow classical one is calculated within the leading order correction to the adiabatic approximation following berry - robbins and jarzynski .
its finite time integral @xmath48 of ( [ indef ] ) is proportional to the dissipation rate on the time scale @xmath44 . in the present work @xmath48
was calculated under various statistical assumptions . in section [ secrmt ]
it was studied in the framework of rmt . for the specific dependence of the model ( [ model ] ) on the parameter @xmath35 , the hamiltonian @xmath227 and @xmath228 are statistically independent @xcite . for this case
it was shown that up to a proportionality constant , the integral of the correlation function is simply related to the integral of the form factor ( by eq .
[ iandk ] ) . since the form factor is known in rmt , the integral of the correlation function was calculated and found to vanish for times beyond the heisenberg time for gue and to fall off as a power law for goe .
this is a remarkable and surprising difference .
the result for gse is similar to the one found for gue , except that the integral of the correlation function vanishes after twice the heisenberg time .
the properties of the model ( [ model ] ) are satisfied approximately by many systems @xcite , therefore it is expected that the results of this work are relevant for a wide range of problems . for the model ( [ model ] ) we have shown that for long times the results are dominated by the nearest neighbor spacings . if only these spacings are taken into account one finds that for long times the integral falls off as a power law for goe and as a gaussian , where the characteristic time is proportional to the heisenberg time for gue and gse .
these results do not require all the properties of the model ( [ model ] ) .
they require only that the contribution of the fluctuations in the absolute value of the matrix elements between nearest neighboring states ( neighboring in energy ) is negligible , namely that the contribution of @xmath229 of ( [ dcn ] ) can be ignored compared to the one of @xmath230 .
this is clearly a weaker assumption than complete statistical independence between @xmath227 and @xmath228 .
therefore we expect the difference found between goe and gue ( as well as gse ) to be generic for rmt models with various dependencies on the parameter @xmath35 .
the long time behavior of the rmt models is expected to provide a faithful representation of the behavior of chaotic systems , since it is dominated by the small level spacings .
for short times , on the other hand , the behavior depends on the specific properties of each system .
the short time behavior of the rmt model was presented here only to set the relation between the constants of the rmt model and the ones of the chaotic system . the assumption of the dominance of the contribution of nearest neighbor level spacings , together with the assumption that @xmath229 of ( [ dcn ] ) is negligible enables the calculation of the integral @xmath48 of ( [ indef ] ) for various distributions of nearest neighbor level spacings even if these do not necessarily originate from rmt models . for the distribution ( [ psbeta ] ) , that is a generalization of the distributions found for goe , gue and gse , one can calculate @xmath64 for various values of @xmath58 . in section [ nnd ]
it is found to decay as @xmath231 for all values of @xmath58 , except when @xmath58 is a positive even integer , for which it decays like a gaussian with a characteristic time that is proportional to the heisenberg time .
what is special when @xmath58 is a positive even integer ?
for these values the integral ( [ incbeta ] ) can be extended to the range @xmath232 $ ] .
the integrand is an entire function , the contour of integration can be deformed in the complex plane and the integral is dominated by a saddle point . for other values of @xmath58 an extension of the integral to negative @xmath44 , so that the integrand is analytic , is impossible .
the point @xmath233 is an end point and for large @xmath44 the integral is dominated by it , leading to power law decay . for non - integral @xmath58 ,
the point @xmath233 is also a singular point .
it would be nice to find a more physical explanation for this difference between the various ensembles . for completeness the integral of the correlation function
was calculated for the poisson and the semi - poisson distributions .
the various rmt formulas ( eqs .
[ igoe ] & [ igue ] ) are developed for the limit of infinite matrices .
this limit is approached extremely fast , as can be seen in figs .
[ figgoe ] & [ figgue ] . the convergence to the average , as a function of @xmath120 , the number of members of the ensemble , is slow ( see figs . [ figgoefluc ] & [ figguefluc ] )
. the crucial approximation that was made generalizing the results beyond the model ( [ model ] ) was neglecting @xmath234 of ( [ dcn ] ) .
although reasonable , its validity for chaotic systems should be checked .
the results may hold also for mixed systems if sticking to integrable regions does not take place on time scales relevant for the calculation .
for chaotic systems corrections of order higher than the leading one , in the adiabatic approximation , may lead to different behavior after some time ( @xmath235 or @xmath66 ) .
dephasing , as a result of the coupling to the environment will destroy the quantum correlations on a time scale @xmath67 . in experiments of the type mentioned in the end of the introduction , dephasing is always present , and if @xmath236 , the system will dissipate energy at a rate proportional to @xmath69 .
we have benefited from discussions with m. berry , e. bogomolny , d. cohen , b. eckhardt , j. feinberg , f. izrailev , c. jarzynski , c. marcus , e. ott , r. prange , j. robbins , h .- j .
stckmann , m. wilkinson and m. zirnbauer .
we thank in particular b. simons for extremely illuminating remarks .
this research was supported in part by the u.s.israel binational science foundation ( bsf ) , by the minerva center for non - linear physics of complex systems , by the israel science foundation , by the niedersachsen ministry of science ( germany ) and by the fund for promotion of research at the technion .
one of us ( sf ) would like to thank r. prange for the hospitality at the university of maryland where this work was completed .
j. blocki , y. boneh , j. r. nix , j. randrup , m. robel , a. j. sierk , and w. j. swiatecki . , 113:330 , 1978 ; s. e. koonin , r. l. hatch , and j. randrup .
, a283:87 , 1977 ; s. e. koonin and j. randrup . ,
a289:475 , 1977 .
d. cohen . , 82:4951 , 1999 ; lecture notes , to be published in proceedings of session cxliii of the enrico fermi summer - school on `` new directions in quantum chaos '' , varenna 1999 ; `` chaos and energy spreading for time - dependent hamiltonians , and the various regimes in the theory of quantum dissipation '' , cond - mat/9902168 .
o. bohigas , m. j. giannoni , and c. schmidt .
, 52:1 , 1984 ; o. bohigas , in _ chaos and quantum physics , _ proc . of les - houches summer school , session lii , 1989 _ _ , edited by m. j. giannoni , a. voros , and j. zinn - justin ( elsevier , 1991 ) .
l. landau and y. smorodinsky , _ lectures on the theory of the atomic nucleus _
( state tech .- theoret .
press , moscow , 1955 ) , p. 93
, translation , pp .
plenum press ( consultants bureau ) , new york , 1958 .
reproduced in @xcite .
a. v. andreev and b. l. altshuler . , 75:902 , 1995 ; o. agam , b. l. altshuler , and a. v. andreev . , 75:4389 , 1995 ; a. v. andreev , o. agam , and b. l. altshuler .
, 76:3947 , 1996 ; a. v. andreev , o. agam , b. d. simons , and b. l. altshuler . , 482:536 , 1996 ; , 79:1778 , 1997 .
e. bogomolny and j. keating .
, 77:1472 , 1996 .
m. r. zirnbauer , in _ suppersymmetry and trace formulae , chaos and disorder _ , edited by i. v. lerner , j. p. keating , and d. e. khmelnitskii ( kluwer academic / plenum publishers , new york , 1999 ) , p. 193 .
m. feingold , r. g. littlejohn , s. b. solina , and j. s. pehing . , 146:199 , 1990 ; m. feingold , d. m. leitner , and m. wilkinson . , 66:986 , 1991 ; m. wilkinson , m. feingold , and d. m. leitner .
, 24:175 , 1991 . | the autocorrelation function of the force acting on a slow classical system , resulting from interaction with a fast quantum system is calculated following berry - robbins and jarzynski within the leading order correction to the adiabatic approximation .
the time integral of the autocorrelation function is proportional to the rate of dissipation .
the fast quantum system is assumed to be chaotic in the classical limit for each configuration of the slow system .
an analytic formula is obtained for the finite time integral of the correlation function , in the framework of random matrix theory ( rmt ) , for a specific dependence on the adiabatically varying parameter .
extension to a wider class of rmt models is discussed . for the gaussian unitary and symplectic ensembles for long times the time integral of the correlation function vanishes or falls off as a gaussian with a characteristic time that is proportional to the heisenberg time , depending on the details of the model .
the fall off is inversely proportional to time for the gaussian orthogonal ensemble .
the correlation function is found to be dominated by the nearest neighbor level spacings .
it was calculated for a variety of nearest neighbor level spacing distributions , including ones that do not originate from rmt ensembles .
the various approximate formulas obtained are tested numerically in rmt .
the results shed light on the quantum to classical crossover for chaotic systems .
the implications on the possibility to experimentally observe deterministic friction are discussed .
# 1 pacs : 05.45.mt , 05.45.ac , 05.45.-a |
distance - redshift relation plays an important role in the astronomy and observational cosmology .
the standard distance , which has been used in the most of previous studies , is based on a postulate that a distribution of matter in the universe is homogeneous ( e.g. , weinberg 1972 ) .
it has been , however , well recognized that our universe is highly inhomogeneous on small scales . since the inhomogeneities of the mass distribution focus ( defocus ) the bundle of light rays ( the gravitational lensing effects ) , the distance in an inhomogeneous universe deviates from that in the homogeneous friedmann universe
it is , therefore , obvious that a detail understanding of the propagation of light rays in the inhomogeneous universe is necessary for correct studies of objects in a distant universe . since the pioneering work by gunn ( 1967 )
, there has been a lot of progress in this subject .
babul & lee ( 1991 ) , among others , studied the effects of the ricci focusing by weak inhomogeneities .
they found that the dispersion in image magnifications due to large scale ( @xmath0mpc , where @xmath1km / sec / mpc ) structures is negligible even for sources at redshift of 4 .
they also pointed out that the dispersion is very sensitive to the nature of the matter distribution on small scales .
the same result was also obtained by frieman ( 1996 ) .
he improved babul & lee s study to reflect the recent developments in numerical and observational studies of the large scale structure . in these studies ,
the effects of the weyl focusing which induce a shear of light ray bundle were neglected .
nakamura ( 1997 ) examined the effects of the shear on the image magnification in the cold dark matter model universe with linear density perturbation .
he found that the effect is sufficiently small and concluded that the weyl focusing can be safely neglected for a light ray passing through a linear density inhomogeneities .
jaroszyski et al .
( 1990 ) and wambsganss et al .
( 1997 ) used the multiple gravitational lens theory with a large n - body simulation of the cold dark matter universe .
their result agreed well with analytical studies .
the above studies mainly focused on the large scale inhomogeneities , whereas the effects of small scale objects , such as galaxies and clusters of galaxies , have not been fairly taken into account .
kayser & refsdal ( 1988 ) investigated the gravitational lensing effects due to randomly distributed king model galaxies .
they paid a special attention to a high magnification part of the magnification probability distribution .
recently wambsganss , cen & ostriker ( 1998 ) studied the gravitational lensing effects by using the large n - body simulation with an effective resolution of comoving @xmath2kpc .
they , first , shoot the light ray through the lens planes by using the multiple gravitational lens equation , then the magnification matrix is determined from the mapping of the light ray positions between the image and source plane . in this procedure , the ricci and weyl focusing can not be treated independently , therefore no discussion is given for the weyl focusing effect in their paper .
however , the magnification factor as a function of position in the source plane and image plane are presented in figure 4 and 6 of their paper .
those figures show that there is quite a large region that is demagnified by a small amount , and a few relatively small spots that are quite highly magnified by the small scale inhomogeneities .
the purpose of this paper is to examine the weyl focusing effect due to the small scale inhomogeneities , and we are not concerned with the effect due to the large scale structure .
the density distribution in the universe is modeled as a randomly distributed isothermal lenses .
the isothermal lens is approximated by virialized objects , such as galaxy and cluster of galaxies .
this model is similar to the one studied by kayser & refsdal ( 1988 ) .
however our emphasis is different from theirs , i.e. we pay special attention to the gravitational lensing effects on a majority of light rays .
although this model is a very simplified and unrealistic one , we believe that the model is good enough to investigate the essential points of the weyl focusing effect due to the small scale inhomogeneities .
this paper is organized as follows : in section 2 , we describe our models and method of simulating the light propagation .
the results of the simulation are summarized in section 3 .
the paper concludes with discussion in section 4 .
we use the multiple lens equations to trace the propagation of infinitesimal bundles of light rays .
schneider , ehlers & falco ( 1992 ) deals with the theory of multiple gravitational lensing in detail .
here we simply describe only the aspects which are directly relevant to this paper . as was done in previous studies ( see e.g. , blandford & narayan , 1986 and kovner 1987 ) , we consider @xmath3 screens ( lens planes ) between an observer @xmath4 and source @xmath5 , located at redshifts @xmath6 with @xmath7 runs from 1 to @xmath8 with @xmath9 . in the following ,
the quantities on the lens and source planes are described by indices @xmath10 .
the position vector of a light ray on the @xmath7-th lens plane is denoted by @xmath11 .
let @xmath12 denotes the deflection angle of a light ray at position @xmath11 on the @xmath7-th lens plane .
the multiple gravitational lens equation and the evolution equation of the lensing magnification matrix @xmath13 are written as , @xmath14 @xmath15 for @xmath16 , where @xmath17 @xmath18 denotes the standard angular diameter distance between redshifts of @xmath6 and @xmath19 ( 0 and @xmath6 ) with @xmath20 , the comma denotes differentiation with respect to the components of @xmath21 and @xmath22 .
the deflection angle @xmath23 is determined by the following equation : @xmath24 where @xmath25 is the surface mass density and @xmath26 is its average value . for a notation convenience
, we introduce the following quantities : @xmath27 the optical tidal matrix @xmath28 is decomposed into the ricci and weyl focusing terms , respectively : @xmath29 in general , equation ( [ mab ] ) is not an explicit equation for @xmath13 , since the equation involves a summation over @xmath30 evaluated on the light ray path , such that one first has to solve the multiple gravitational lens equation ( [ le ] ) .
however for the light rays traveling in regions where @xmath31 and @xmath32 , one can expand @xmath33 in powers of @xmath31 and @xmath30 about its value when the light ray is unperturbed .
we rewrite equation ( [ le ] ) as @xmath34@xmath35@xmath36 and equation ( [ mab ] ) as @xmath37@xmath38@xmath39@xmath38@xmath40@xmath38@xmath41@xmath38@xmath42 , where @xmath43 is the first term of the right hand side of equation ( [ le ] ) and @xmath44 is the second term , but the deflection angle is evaluated at the unperturbed light ray position . expanding equation ( [ mab ] ) in terms of @xmath31 and @xmath30 , one finds @xmath45 @xmath46 @xmath47,\ ] ] for @xmath48 . in the above expressions , @xmath30 and @xmath31 are evaluated at the unperturbed light ray position .
the image magnification factor of a point like source is given by the inverse of the determinant of the magnification matrix , i.e. @xmath49 . up to the order of @xmath50 , the determinant is @xmath51 ^ 2 - \left| \sum_{i=1}^{j-1 } { \cal{f } } ( z_i , z_j ) \right|^2 \nonumber \\ & & + 2 \sum_{i=2}^{j-1 } \sum_{k=1}^{i-1 } \left\ { { \cal{r}}(z_i , z_j ) { \cal{r}}(z_k , z_i ) + \mbox{re } \right [ { \cal{f}}^{\ast}(z_i , z_j ) { \cal{f}}(z_k , z_i ) \left ] + { \cal{r}}_{,a}(z_i , z_j ) \alpha_a ( z_k , z_i ) \right\},\end{aligned}\ ] ] for @xmath48 .
this is our principal equation . in general , the effects of ricci and weyl focusing on image magnification are coupled .
however up to this order , they are not coupled .
we call the terms in the equation ( [ recdetm ] ) which involve ricci focusing terms as `` ricci contribution '' and that involve weyl focusing terms as `` weyl contribution '' .
we adopt the truncated singular isothermal sphere as the lens model .
its surface mass density is written as @xmath52 where @xmath53 is the one - dimensional velocity dispersion and @xmath54 is the half mass radius ( pei , 1993 ) .
the corresponding 3-dimensional mass density runs as @xmath55 @xmath56 and as @xmath57 @xmath58 , and the mass within the radius @xmath59 is @xmath60 then the total mass is @xmath61 .
this models should provide a fair approximation to virialized objects , such as a galaxy and cluster of galaxies with isothermal dark halos .
for the distribution of matter in the universe , we assume that the isothermal lenses are randomly distributed with the average mass density @xmath62 and the rest of the matter has the uniform distribution .
we also assume that the comoving density of lenses is constant in time , thus @xmath63 .
furthermore , we approximate the lensing effects of the isothermal objects , except the nearest one , to a form of uniform surface mass density . under the above assumptions and the circularly symmetric mass distribution of the lens model , the deflection angle
is given by @xmath64 \nonumber \\ & = & a_{cr } ( z_i , z_j ) { { d_j } \over { d_i } } \left [ { { r_g } \over { r_g+r } } -{{r_g r } \over { r_c ( r_g+r_c)}}\right],\end{aligned}\ ] ] with @xmath65 where @xmath59 is the distance between a light ray position and center of the nearest lens object in the @xmath7-th lens plane .
@xmath66 is a half of the mean separation length between the lens objects defined by @xmath67^{-{1 \over 2}},\end{aligned}\ ] ] where @xmath68 is the average surface mass density of the lens objects in the @xmath7-th lens plane , and @xmath69 is the redshift interval between the @xmath70-th and @xmath7-th lens planes , and @xmath71 is the density parameter of lens objects .
the first term in square bracket of the second line in equation ( [ al ] ) describes the deflection due to the nearest lens object , the second term is due to the others , and the last term is due to the background average mass density of lens objects . from the deflection angle ( [ al ] ) , the ricci and weyl focusing terms are immediately given by @xmath72 , \\
% \label{weylcon } \left| { \cal{f}}(z_i , z_j ) \right| & = & a_{cr } ( z_i , z_j ) \left[{1\over 2 } { { r_g ( r_g+ 2 r ) } \over { r ( r_g+r)^2}}\right ] .
\end{aligned}\ ] ] we introduce a compactness parameter @xmath73 as follows ; @xmath74 this parameter measures the effectiveness of an isothermal object as a gravitational lens . as was pointed out by kayser & refsdal ( 1988 ) , the model with randomly distributed isothermal lens objects
is completely described by two parameters . the parameters they used depend on the distance between source and observer . on the other hand ,
we characterize the model by the compactness parameter @xmath73 and density parameter of lens object @xmath75 ( hereafter we shall simply denote it as @xmath76 ) which are independent of redshifts of the source and lenses .
we should note that the optical tidal matrix @xmath30 becomes larger than unity at the very central region of the lens object ( the inner region of @xmath77 ) , therefore the validity of assumptions used in deriving the equation ( [ recdetm ] ) breaks down for light rays passing through that region .
we now estimate the strong lensing effect on our study in two ways .
the first is based on an order - of - magnitude estimate ( section 14 of peebles , 1993 , and also futamase & sasaki , 1989 ) , we examine a magnitude of the tidal matrix @xmath30 .
suppose the lensing objects are randomly distributed and each with mass @xmath78 , where @xmath79 is a characteristic comoving size of a lens object and is of order @xmath54 .
hence the mean comoving number density of the lens objects is @xmath80 , so the mean comoving separation distance is @xmath81 .
then for a geodesic affine comoving distance of @xmath82 , the light ray gravitationally encounters such objects @xmath83 times in average . at each encounter , the contribution to the tidal matrix is @xmath84 , where @xmath85 is a comoving angular diameter distance with the subscript @xmath86 @xmath87 stands for a lens ( source ) , @xmath88 is the comoving impact parameter , and we have assumed that the mean comoving impact parameter is of order @xmath89 .
since the sign of each contribution will be random , the total contribution to the optical tidal matrix will be @xmath90^{1\over 2 } \langle{{{h_0 } \over c } { { \hat{d}_d \hat{d}_{ds } } \over { \hat{d}_s } } } \rangle \left[{{h_0 } \over c } \lambda \right]^{1\over 2}\nonumber\\ \label{magoft } & \sim & \sqrt{\omega_l } \sqrt{\nu } \langle{{{h_0 }
\over c } { { \hat{d}_d \hat{d}_{ds } } \over { \hat{d}_s } } } \rangle \left[{{h_0 } \over c } \lambda \right]^{1\over 2}.\end{aligned}\ ] ] the contribution from the direct encounters can be similarly estimated by noting that the average number of encounters is @xmath91 , with each encounter contributing @xmath92 with random sign .
the result turns out to be the same as that of gravitational distant encounters , equation ( [ magoft ] ) . in the case of einstein - de sitter universe model ,
the comoving distance @xmath82 becomes @xmath93 at the source redshift @xmath94 and the averaged value of distance combination over the lens redshifts is @xmath95 for any redshifts of source .
accordingly , we find that the magnitude of the optical tidal matrix scales as @xmath96 for the source redshift of 3 . in the following we only consider lens models with @xmath97 and @xmath98 , therefore a typical value of the optical tidal matrix can be expected to be of order @xmath99 or lower .
thus , we can conclude that @xmath30 is less than unity for a majority of random lines of sight .
next we examine the lensing optical depth defined by turner , ostriker & gott , ( 1984 ) to quantify the probability of the light rays being affected by strong lensing . the differential optical depth for the truncated singular isothermal sphere model
is given by @xmath100 where @xmath101 ^ 2.\ ] ] we numerically integrate the last equation for cases of the lens models with @xmath102 , @xmath103 , @xmath104 and @xmath105 , and for the einstein - de sitter universe model .
the results are presented in table 1 . from the table 1
, it can be found that the probabilities of strong lensing events are very small except for an extreme model with @xmath106 . even for the extreme model ,
the probability is not significantly large .
we thus conclude that the strong lensing effects do not significantly alter our results .
it can be said from the above two estimations that we can safely use the perturbative equation ( [ recdetm ] ) .
.the lensing optical depth . [ cols="^,^,^,^,^ " , ] observationally , the strong lensing events among the random lines of sight are very rare . for example , the probability of the multiply imaged quasars in a quasar sample is at most @xmath107 ( e.g. , claeskens , jaunsen & surdej , 1996 ) . combining this fact with our results summarized in table 2 ,
it may be reasonably concluded that , as far as our simple matter distribution model is concerned , the weyl focusing has a negligible effect on the majority of light rays even for sources at the redshift of 5 .
this result can be naturally explained by the following two reasons : first , the weyl focusing is a second order effect on the image magnification , therefore it becomes important only for the light rays passing through a very high non - linear ( relatively rare ) region .
secondly , since we have assumed the random distribution for the isothermal lenses , the rays coherently affected by the weyl focusing are very rare , consequently , the majority of light rays are only weakly influenced by the weyl focusing .
in this paper , we restricted our study to the gravitational lensing effects due to the randomly distributed isothermal lenses .
we found that the weyl focusing effect is small , @xmath108 , for a majority of light rays .
lee & paczyski ( 1990 ) examined the gravitational lensing effects in randomly distributed clumps with gaussian surface mass density profile .
they found that the image magnifications are dominated by ricci focusing and the weyl focusing has no significant effect .
they only considered a case of a source redshift of @xmath109 , in an @xmath110 and @xmath111 universe .
however we have found that , as far as the random distribution of lens objects is concerned , the weyl focusing is rather small even for higher redshift . in this study , a correlation of lens objects and large scale structure are not taken into account .
the study of the influences of the correlation on the weyl focusing lies outside the scope of this paper , and will be examined in future works . on the other hand ,
the weyl focusing effect due to the large scale structures is investigated by using n - body simulation ( jaroszyski et al . , 1990 ) and
analytically ( nakamura , 1997 ) .
these two studies show that , although there is an uncertainty of a normalization of the density power spectrum , the magnitude of weyl focusing due to the large scale structure is of the order @xmath112 for the source redshifts of @xmath113 ( figure 3 of jaroszyski et al . , 1990 and figure 3 of nakamura , 1997 ) .
consequently it can be said that , comparing the results of the above mentioned studies with our results , the weyl focusing effect due to the large scale structures is comparable with or larger than that due to the small scale inhomogeneities .
the author wishes to thank the referee for valuable comments on the first version of the manuscript which significantly improved the quality of this paper .
he would like to thank professor p. schneider , professor t. futamase , dr .
m. hattori and m. takada for fruitful discussions .
he also would like to thank dr .
p. premadi for carefully reading and commenting the manuscript
. 99 babul a. , lee m. h. , 1991 , mnras , 250 , 407 blandford r. , narayan r. , 1986 , apj , 310 , 568 claeskens j. -f .
, jaunsen a. o. , surdej j. , 1996 , in kochanek c. s. , hewitt j. n. , eds , proc .
173 , astrophysical applications of gravitational lensing , p. 13 frieman j. a. 1997 , comments astrophys .
, 18 , 323 futamase t. , sasaki m. , 1989 , phys . rev .
d. 40 , 1993 gunn j. e. , 1967 , apj , 150 , 737 jaroszyski m. , park c. , paczyski b. , gott j. r. , 1990 , apj , 365 , 22 kayser r. , refsdal s. , 1988 , a&a , 197 , 63 kovner i. , 1987 , apj , 316 , 52 lee m. h. , paczyski b. , 1990 , apj , 357 , 32 nakamura t. t. , 1997 , publ .
japan , 49 , 151 peebles p. j. e. , 1993 , principles of physical cosmology , princeton univ . press , princeton pei y. c. , 1993 , apj , 404 , 436 schneider p. , ehlers j. , falco e. e. , 1992 , gravitational lenses , springer - verlag , new york turner e. l. , ostriker j. p. , goot j. r. , 1984 , apj , 284 , 1 wambsganss j. , cen r. , xu g. , ostriker j. p. , 1997
, apj , 475 , l81 wambsganss j. , cen r. , ostriker j. p. , 1998
, apj , 494 , 29 weinberg s. , 1972 , gravitation and cosmology , new york , wiley | weyl focusing effects on the image magnifications are investigated by using the multiple gravitational lens theory .
we focus on the gravitational lensing effects due to the small scale virialized objects , such as galaxies and clusters of galaxies .
we consider a simple model of an inhomogeneous universe .
the matter distribution in the universe is modeled by randomly distributed isothermal objects .
we found that , for the majority of the random lines of sight , the weyl focusing has no significant effect and the image magnification of a point like source within redshift of 5 is dominated by the ricci focusing .
[ firstpage ] cosmology gravitational lensing . |
the attractive hubbard model represents an unvaluable tool to understand properties of pairing and superconductivity in systems with attractive interactions .
the simplifications introduced in this model allow a comprehensive study of the evolution from the weak - coupling regime , where superconductivity is due to bcs pairing in a fermi liquid phase , and a strong coupling regime , in which the system is better described in terms of bosonic pairs , whose condensation gives rise to superconductivity ( bose einstein ( be ) superconductivity ) @xcite .
it has been convincingly shown that such an evolution is a smooth crossover and the highest critical temperature is achieved in the intermediate regime where none of the limiting approaches is rigorously valid@xcite .
a realization of such a crossover scenario has been recently obtained through the development of experiments on the condensation of ultracold trapped fermionic atoms@xcite . in these systems
the strength of the attraction can be tuned by means of a tunable fano - feschbach resonance , and the whole crossover can be described@xcite . in the context of high - temperature superconductivity , the intermediate - strong coupling regime in which incoherent pairs are formed well above the critical temperature has been invoked as an interpretation of the pseudogap phase@xcite . moreover , since the early days of the discovery of these materials , the evolution with the doping level of both the normal- and the superconducting - phase properties induced some authors@xcite to recognize the fingerprints of a crossover between a relatively standard bcs - like superconductivity in the overdoped materials and a strong - coupling superconductivity associated to bose - einstein condensation ( be ) in the underdoped materials .
indeed at optimal doping the zero - temperature coherence length is estimated to be around 10@xmath220 @xcite , i.e. , much smaller than for conventional superconductors but still large enough to exclude the formation of local pairs.@xcite it is understood that the attractive hubbard model has not to be taken as a microscopic model for the cuprates , since a realistic description of the copper - oxygen planes of these materials unavoidably requires a proper treatment of strong coulomb repulsion .
this simplified model represents instead an ideal framework where the evolution from weak to strong coupling can be studied by simply tuning the strength of the attraction .
the main aim of the present work is to identify if , and to which extent , at least some aspects of the phenomenology of the cuprates can be interpreted simply in terms of a crossover from weak to strong coupling .
the main simplifications introduced by the attractive hubbard model can be summarized as _ ( i ) _ neglect of repulsion .
even if some attraction has to develop at low energy , the large short - range coulomb repulsion implies that the interaction must become repulsive at high - energy in real systems . in some sense , an attractive hubbard model picture can at
most be applied to the low - energy quasiparticles . _
( ii ) _ the model naturally presents s - wave superconductivity , as opposed to the d - wave symmetry observed in the cuprates _
( iii ) _ neglect of retardation effects .
the hubbard model describes instantaneous interactions , while every physical pairing is expected to present a typical energy scale .
the model is written as @xmath3 where @xmath4 ( @xmath5 ) creates ( destroys ) an electron with spin @xmath6 on the site @xmath7 and @xmath8 is the number operator ; @xmath9 is the hopping amplitude and @xmath10 is the hubbard on - site attraction ( we take @xmath11 , with an explicit minus sign in the hamiltonian ) .
notice that , with this notations , the hamiltonian is explicitly particle - hole symmetric for @xmath12 , which therefore corresponds to @xmath13 ( half - filling ) . despite its formal simplicity , this model can be solved exactly only in @xmath14 , while in larger dimensionality analytical calculations are typically limited to weak ( @xmath15 ) or strong ( @xmath16 ) coupling , where the bcs and the be approaches are reliable approximations .
it is anyway known that for @xmath17 , the ground state of ( [ hubbard ] ) is superconducting for all values of @xmath10 and all densities @xmath18 , with the only exception of the one - dimensional half - filled case . at half - filling the model
has an extra - symmetry and the superconducting and the charge - density - wave order parameters become degenerate . a reliable description of the evolution of the physics as a function of @xmath10 requires to treat the two limiting regimes on equal footing overcoming the drawbacks of perturbative expansions .
quantum monte carlo ( qmc ) simulations represent a valuable tool in this regard , and they have been applied to the two@xcite and three@xcite dimensional attractive hubbard model .
even if the sign problem does not affect these simulations , finite size effects and memory requirements still partially limit the potentiality of this approach .
a different non perturbative approach is the dynamical mean - field theory ( dmft ) , that neglects the spatial correlations beyond the mean field level in order to fully retain the local quantum dynamics , and becomes exact in the limit of infinite dimensions@xcite . due to the local nature of the interaction in the attractive hubbard model
, we expect that the physics of local pairing is well described in dmft . moreover , this approach is not biased toward metallic or insulating states , and it is therefore particularly useful to analyze the be - bcs crossover . on the other hand ,
the simplifications introduced by the dmft are rigorously valid only in the infinite dimensionality limit , and even if the dmft has obtained many successes for three dimensional systems , its relevance to lower dimensionality like @xmath19 is much less established , and represents a fourth limitation of our study in light of a comparison with the physics of the cuprates
. in particular , the role of dimensionality in determining the pseudogap properties of the attractive hubbard model has been discussed in refs .
@xcite .
the study of the attractive hubbard model can greatly benefit of a mapping onto a repulsive model in a magnetic field .
the mapping is realized in a bipartite lattice@xcite by a staggered particle - hole transformation on the down spins @xmath20 .
the attractive model with a finite density @xmath18 transforms into a half - filled repulsive model with a finite magnetization @xmath21 .
the chemical potential is transformed , accordingly , into a magnetic field @xmath22 . in the @xmath13 case ( half - filling ) the two models are therefore completely equivalent .
we notice that the above mapping does not only hold for the normal phases , but extends to the broken symmetry solutions .
the three components of the antiferromagnetic order parameter of the repulsive hubbard model are in fact mapped onto a staggered charge - density - wave parameter ( @xmath23 component of the spin ) and an s - wave superconducting order parameter ( @xmath24 components ) .
the above mapping is extremely useful , since it allows to exploit all the known results for the repulsive model and for the mott - hubbard transition to improve our understanding of the attractive model . in recent works
the dmft has been used to study the normal phases of the attractive hubbard model . in particular
, a phase transition has been found both at finite@xcite and at zero temperature@xcite between a metallic solution and an pairing phase of pairs .
the insulating pairs phase is nothing but a realization of a superconductor without phase coherence , i.e. , a collection of independent pairs . as it has been discussed in ref.@xcite , this phase is the negative-@xmath10 counterpart of the paramagnetic mott insulator found for the repulsive hubbard model .
we notice that the insulating character of the pairing phase is a limitation of the dmft approach , in which the residual kinetic energy of the preformed pairs is not described .
the pairing transition has been first identified in ref .
@xcite by means of a finite temperature qmc solution of the dmft .
the @xmath25 study of ref .
@xcite has clarified that the pairing transition is always of first order except for the half - filled case , and that it takes place with a finite value of the quasiparticle weight @xmath26 , associated to a finite spectral weight at the fermi level . in the latter paper
, it has also been shown that the pairing transition gives rise to phase separation . for what concerns
the onset of superconductivity , a dmft calculation of the critical temperature @xmath27 has been performed for the case of @xmath28 in the same ref .
the @xmath27 curve , extracted from the divergence of the pair - correlation function in the normal phase , displays a clear maximum at intermediate coupling and reproduces correctly both the bcs and the be predictions in the asymptotic limits , remaining finite for all @xmath29 . in this work
we complement the analysis of ref .
@xcite , by extending our phase diagram to finite temperature , still using exact diagonalization ( ed ) to solve the impurity model associated with the dmft of the hubbard model@xcite .
we also compare the normal state solutions with the superconducting solutions which are stable at low temperatures .
the use of ed allows us to reach arbitrarily small temperatures which are hardly accessible by means of qmc .
quite naturally , the extension of ed to finite temperature requires a more severe truncation of the hilbert space .
we have checked that all the thermodynamical quantities we show are only weakly dependent on the truncation .
the plan of the paper is the following : in sec .
ii we briefly introduce the dmft method and its generalization to the superconducting phase ; in sec .
iii we discuss the finite temperature phase diagram in the normal phase characterizing the low - temperature pairing transition ; in sec .
iv we analyze the superconducting solutions ; in sec .
v we compare different estimators of the pseudogap temperature in the high - temperature normal phase .
vi contains our concluding remarks .
the dmft extends the concept of classical mean - field theories to quantum problems , by describing a lattice model in terms of an effective dynamical local theory .
the latter can be represented through an impurity model subject to a self - consistency condition , which contains all the information about the original lattice structure through the non - interacting density of states ( dos)@xcite . starting from the hubbard model ( [ hubbard ] ) , we obtain an attractive anderson impurity model @xmath30 the self - consistency is expressed by requiring the identity between the local self - energy of the lattice model and the impurity self - energy @xmath31 where @xmath32 is the local green s function of ( [ aim ] ) , and @xmath33 is the dynamical weiss field , related to the parameters in ( [ aim ] ) by @xmath34 by expressing the local component of the green s function in terms of the lattice green s function , namely @xmath35 , eq .
( [ sigma ] ) implies @xmath36 where @xmath37 is the non interacting density of states of the original lattice .
we consider the infinite - coordination bethe lattice , with semicircular dos of half - bandwidth @xmath38 ( i.e. , @xmath39 ) , for which eq . ( [ selfconsistence ] ) is greatly simplified and becomes @xmath40 in this work we also consider solutions with explicit s - wave superconducting order , by allowing for local anomalous green s functions @xmath41 .
the whole dmft formalism can then be recast in nambu - gorkov spinorial representation@xcite , and eqs .
( [ sigma ] ) and ( [ selfconsistence ] ) must be read as matrix identities in the nambu space . as far as the impurity model is concerned
, we need to describe an anderson impurity model with a superconducting bath or , equivalently , with an anomalous hybridization in which cooper pairs are created and destroyed in the electronic bath , i.e. , a term @xmath42 is added to ( [ aim ] ) .
the heaviest step of the dmft approach is to compute @xmath32 for the anderson model ( [ aim ] ) .
this solution requires either a numerical approach or some approximation .
here we use exact diagonalization .
namely , we discretize the anderson model , by truncating the sums over @xmath43 in eqs .
( [ aim ] ) and ( [ g0 ] ) to a finite number of levels @xmath44 .
it has been shown that extremely small values of @xmath44 provide really good results for thermodynamic properties and reliable results for spectral functions . in this work
we use the ed approach at finite temperature , where it is not possible to use the lanczos algorithm , which allows to find the groundstate of extremely large matrices . to obtain the full spectrum of the hamiltonian , needed to compute the finite - temperature properties ,
we are forced to a rather small value of @xmath44 , up to 6 .
all the results presented here are for @xmath45 , and we always checked that changing @xmath44 from 5 to 6 does not affect the relevant observables we discuss in the present work , except for the real - frequency spectral properties .
in this section we limit our analysis to normal phase paramagnetic solutions in which no superconducting ordering is allowed . even if the s - wave superconducting solution is expected to be the stable one at low temperatures , our normal state solutions are representative of the normal phase above the critical temperature .
the region in which the normal state is stable may of course be enlarged by frustrating superconductivity through , e.g. , a magnetic field .
moreover , the nature of the normal phase gives important indications on the nature of the pairing in the different regions of the phase diagram .
as mentioned above , it has been shown that the normal phase of the attractive hubbard model is characterized by a `` pairing '' transition between a fermi - liquid phase and a phase in which the electrons are paired , but without any phase coherence among the pairs .
the pairing transition has been first discussed at finite temperature in ref .
@xcite , and a complete characterization at @xmath25 has been given in ref .
@xcite . in this paper
we complete the finite temperature study of the transition and connect it to the zero - temperature phase diagram , finally drawing a complete phase diagram in the attraction - temperature plane for a density @xmath46 , taken as representative of a generic density ( except for the peculiar particle - hole symmetric @xmath13 case ) .
this situation would correspond to a repulsive model at half - filling in an external magnetic field tuned to give a finite magnetization @xmath47 .
the @xmath25 dmft solution of the attractive hubbard model is characterized by the existence of two distinct solutions , a metallic one with a finite spectral weight at the fermi level and an insulating solution formed by pairs , with no weight at the fermi level .
the previous study has also clarified that the quasiparticle weight @xmath48 , which may be used as a sort of order parameter for the mott transition at half - filling , loses this role for the doped attractive hubbard model , being it finite both in the metallic and pairing phases . at @xmath25 ,
the metallic solution exists only for @xmath49 , and the insulating one for @xmath50 , with @xmath51 .
in other words , a coexistence region is present where both solutions exist , and where the actual ground state is determined minimizing the internal energy . .
in each panel are shown the metallic ( @xmath52 ) and insulating ( @xmath53 ) solutions .
( the chosen value of the attraction lies in the coexistence region).,width=245 ] the clear - cut @xmath25 characterization of the two solutions based by the low - energy spectral weight is lost at finite temperature , where both solution have finite weight at the fermi level .
nonetheless , two families of solutions can still be defined , each family being obtained by continuous evolution of the different @xmath25 phases .
the two solutions are still clearly identified at relatively low temperatures , further increasing the temperature , the differences between the two solutions is gradually washed out , as shown in fig .
[ img ] , where we plot the temperature evolution of the imaginary part of the green s function in imaginary frequency for @xmath54 , which lies in the @xmath25 coexistence region .
while at @xmath55 and @xmath56 the difference in the two solutions is still clear , at @xmath57 , the two solutions become basically indistinguishable .
this result suggests that , as intuitively expected , the temperature reduces the difference between the solutions and consequently , the size of the coexistence region , which is expected to close at some finite temperature critical point ( the attractive counterpart of the endpoint of the line of metal - insulator transitions in the repulsive model @xcite ) . a similar information is carried by the analysis of the average value of double occupancy @xmath58 .
this quantity naturally discriminates between an pairing phase with a large value of @xmath59 and a metal with a smaller value . as shown in fig .
[ doubocc ] , at low temperature we have two solutions with a different value of @xmath59 in the coexistence region , and a jump in this quantity at the transition . upon increasing the temperature ,
the two solutions tend to join smoothly one onto the other , signaling again the closure of the coexistence region , which is substituted by a crossover region .
analogous behavior is displayed by the quasiparticle weight @xmath60 . repeating the same analysis for a wide range of coupling constants and temperatures , we are able to construct a finite - temperature phase diagram for the pairing transition , shown in fig .
for temperatures smaller than a critical temperature @xmath61 , we compute the finite temperature extensions of @xmath62 and @xmath63 , which mark the boundary of the coexistence region .
the two lines ( depicted as dashed lines in fig .
3 ) converge into a finite temperature critical point at @xmath64 and @xmath65 . despite the closure of the coexistence region , a qualitative difference between weak coupling and strong coupling solutions
can still be identified for @xmath66 , determining a crossover region in which the character of the solution smoothly evolves from one limit to the other as the attraction is tuned . at this stage ,
the crossover region is `` negatively '' defined as the range in which the green s function does not resemble any of the two low temperature phases .
the crossover lines are estimated as the points in which it becomes impossible to infer from the matsubara frequency green s function whether the low - energy behavior is metallic or insulating .
it has been shown for the repulsive hubbard model that this kind of crossover is accompanied by a qualitative difference in transport properties . in the region on the left of the crossover ,
the conduction is metallic and the resistivity increases with temperature . in the intermediate crossover region the system behaves like a semiconductor with a resistivity which decreases upon heating , and finally in the phase on the right of the crossover region the system behaves like a heated insulator@xcite . coming from the left ,
the first crossover occurs when @xmath32 has no longer a clear metallic behavior with a finite value at zero frequency , while the second crossover line delimits the region in which the gap of the paired solution is closed by thermal excitations .
we will come back later to the crossover region and compare the above defined lines with physically sensible estimators of the pseudogap temperature , like the specific heat and the spin susceptibility .
-@xmath67 plane .
at low temperature two critical lines @xmath68 and @xmath69 individuate the coexistence region .
the two lines converge in a finite temperature critical point . at higher temperatures we can still define two crossover lines .
the superconducting critical temperature is also drawn as a solid line ( cfr .
4).,width=321 ] turning to the coexistence region , we can also ask ourselves which is the stable phase .
this requires a comparison between the gibbs free energies of the two phases . at half - filling , where the attractive and the repulsive model are equivalent , it has been shown that at @xmath25 the metallic solution is stable in the whole coexistence region@xcite . at finite temperature
it has been shown numerically that the insulator becomes stable in a large portion of the coexistence region due to its large entropy@xcite .
the transition is therefore of first order for all temperature below the critical temperature , except for the two second - order endpoints at @xmath25 and @xmath70 . for densities out of
half - filling it has been shown in ref .
that the transition is of first order already at @xmath25 and it is accompanied by a small phase separation region . for @xmath46 ,
the @xmath25 first - order transition occurs quite close to @xmath63 .
analogously to the half - filling case , the finite temperature almost immediately favors the pairing phase .
indeed , computing the free energy following , e.g. , ref .
@xcite , we find the pairing phase stable for almost every point in the coexistence region .
we had to use an extremely dense mesh of points in the @xmath10 direction to identify a small section where the metallic phase is stable at finite temperature .
therefore the finite temperature first - order transition occurs extremely close to the @xmath62 line for finite temperature and rapidly moves closer to @xmath63 only at really small temperatures .
the above stability analysis has been restricted to normal phase solutions .
indeed the superconducting solution is expected to be the stable one at @xmath25 for all densities and values of the interaction @xmath10 .
the critical temperature @xmath27 is obtained directly as the highest temperature for which a non - vanishing anomalous green s function @xmath71 exists .
the dmft critical temperature @xmath27 for @xmath46 as a function of @xmath10 is reported in fig .
[ tc ] ( full dots ) and it qualitatively reproduces the limiting behavior , with an exponential bcs - like behavior for small u s and a @xmath72 decrease at large @xmath10 according to the expression for the be condensation temperature hard - core boson system@xcite . as a result
, @xmath27 assumes its maximum value of about @xmath73 for an intermediate coupling strength @xmath74 .
interestingly , the maximum @xmath27 occurs almost exactly at the coupling for which the pairing transition in the normal phase would take place in the absence of superconductivity .
it might be noticed however that , while the be result ( open triangles in fig .
[ tc ] ) basically falls on top of the dmft results , the bcs formula ( open circles ) only qualitatively follows the full solution .
this `` asymmetry '' in recovering the bcs behavior arises from the partial screening of the bare attraction due to second order polarization terms@xcite . because of these corrections
the attraction is renormalized as @xmath75 , so that @xmath76 .
when this correction is plugged in the bcs formula for @xmath27 , it results in a correction to the prefactor .
if we simply extract the rescaling factor for a given small value of @xmath10 ( @xmath77 ) and we simply scale the whole weak - coupling curve by this factor , we obtain the points marked with asterisks , whose agreement with the dmft results does not require further comments .
it is interesting , instead , to compare the dmft estimations for @xmath27 with the qmc results : despite the presence of many factors ( such as the exact shape of the d.o.s . of the model or the finite dimension effects ) which are capable to introduce relevant variations in the values of @xmath27 ,
some general similarities appear clearly .
indeed , while simple rescaling the data in terms of the half - bandwidth @xmath38 , both @xmath27 and @xmath78 estimations with the two@xcite and three@xcite dimensional qmc are lower than the dmft evaluation ( i.e. , @xmath79 , @xmath80 for the @xmath19 case , and @xmath81 , @xmath82 in @xmath83 , even if for a lower density of @xmath28 ) , one can observe , quite surprisingly , that the ratio between @xmath27 and @xmath78 is around @xmath84 in both the dmft and the two qmc cases . at @xmath46 : the dmft data ( black circle ) are compared with the bcs -both the bare ( empty circles ) and the rescaled one ( stars)- and the be mean field predictions ( empty triangles ) for an hard - core boson systems ( see refs . ) .,width=321 ] coming back to our dmft results , the simplest and most important observation is that the critical temperature is always higher than the critical temperature for the pairing transition .
for example in our @xmath46 case , @xmath85 is about @xmath86 , against a @xmath61 of @xmath87 . as a result ,
the whole phase transition is hidden by superconductivity , which remains as the only real instability of the system ( cfr .
nevertheless , the crossover lines at higher temperature survive the onset of superconductivity .
therefore the normal phase we reach for @xmath1 is really different according to the regime of coupling we are in . for weak - coupling ,
the normal phase is substantially a regular fermi - liquid and superconductivity occurs as the standard bcs instability . in the strong - coupling regime , the normal phase is instead a more correlated phase which presents a pseudogap in the spectrum . at intermediate coupling , where the superconducting critical temperature reaches its maximum , the normal phase is in a crossover region between the two limiting behaviors .
even if the onset of superconductivity completely hides the pairing transition , the fingerprints of the low - temperature normal phase are still visible in the high - temperature phase diagram , in which a crossover from a metallic phase to a gapped phase is still present .
it is tempting to associate the region in which the system behaves as a collection of incoherent pairs to the pseudogap regime of the cuprates .
it is important to underline that , in this framework , the definition of the pseudogap phase is somewhat tricky , and it implies a certain degree of arbitrariness . in this section we come back to this region and compute various observable whose anomalies have been used to identify the pseudogap phase and compare the related estimates of the pseudogap temperature @xmath88 . our first estimate is based on the evaluation of the uniform spin susceptibility @xmath89 as a function of temperature for different attraction strengths . the opening of a gap in the spin excitation spectrum , not associated with any long - range order , represented in fact one of the first indications of existence of the pseudogap phase in high - temperature superconductors .
the dmft calculation of @xmath89 can be performed by evaluating the derivative of the magnetization @xmath90 with respect to a uniform magnetic field in the limit of vanishing @xmath91 . in terms of the local green functions @xmath92}{h}.\ ] ]
this calculations has been performed by varying the temperature in a wide range ( @xmath93 ) for four different values of the pairing interaction ( @xmath94 and @xmath95 ) and represent an extension of the results reported in ref .
the results of our calculation are summarized in fig .
[ chispin ] . .
these values of @xmath10 are representative of all the interesting region of the phase diagram in fig .
[ phd ] , moving from the metallic to the paired side .
the values of the superconducting critical temperature are marked by small black arrows.,width=321 ] in the weak - coupling side ( @xmath96 ) we find a conventional metallic behavior of @xmath89 , which increases monotonically with decreasing temperature .
the interaction reduces the zero - temperature extrapolated value with respect to the non - interacting result @xmath97 . on the opposite side of the phase diagram , in the strong - coupling regime ( @xmath98 )
the standard high - temperature behavior of @xmath89 extends only down to a certain temperature @xmath99 , where a maximum of @xmath89 is reached .
when the temperature is further reduced @xmath89 starts to decrease , exponentially approaching zero in the @xmath25 limit , signaling the opening of a gap in the spin - excitation spectrum .
a qualitatively similar behavior is found also in the intermediate - coupling regime , at least as long as the value of @xmath10 stays larger than @xmath100 ( e.g. , @xmath101 ) , or , in other words , as long as the left line defining the crossover region in fig .
[ phd ] is not crossed .
the behavior of @xmath89 becomes richer for @xmath102 . at high temperatures @xmath89
closely resembles the insulating case , displaying a clear maximum at a temperature @xmath99 . by approaching @xmath25 ,
@xmath89 no longer vanishes , but it rises at small temperatures displaying a minimum for a temperature lower than @xmath99 : a metallic behavior is therefore recovered , associated to the narrow resonance at the fermi level@xcite .
such a behavior naturally defines a different temperature scale @xmath103 , which is associated to the minimum of @xmath89 and represents the lower border of the pairing zone , or in a sense , of the `` pseudogap '' region .
conversely , this low - temperature behavior has not been observed to our knowledge in finite - dimensional qmc simulation @xcite . in practice , the system displays a pseudogap behavior in the region between @xmath103 and @xmath99 , whose boundary , labeled as @xmath104 , is represented in fig .
we finally mention that the temperature @xmath99 for which @xmath89 is maximum scales with @xmath10 .
this finding is in a qualitative agreement with a qmc simulation@xcite , where the @xmath105 is taken as a definition of the temperature below which the pseudogap appears . from a more quantitative point of view , as happens for @xmath27 and @xmath78 , the values of @xmath105 of the qmc simulations are lower than our dmft results ( i.e. , @xmath106 when @xmath19 and @xmath107 for @xmath83 , against the dmft estimate of @xmath108 ) .
however , also in this case the ratio between @xmath109 and @xmath78 has a more universal value around @xmath110 .
another relevant quantity is the specific heat @xmath111 , that we obtain by differentiating a fit to the dmft internal energy @xmath112 for the same attraction strengths and report in fig .
. also for this quantity the weak coupling case ( @xmath113 ) behaves as a regular metal , with a linear behavior at small temperatures ( @xmath114 , with @xmath115 , @xmath116 being the effective mass ) .
followed by a smooth decrease when the temperature exceeds the typical electronic energy scale .
the same qualitative result is found for the noninteracting system , where the low-@xmath67 slope is smaller since the interacting system has a larger effective mass . in the opposite strong coupling limit
we observe the typical activated behavior of gapped systems for small temperatures , with an exponential dependence of @xmath117 which extends up to a temperature @xmath118 large enough to wipe out the effect of the gap .
it is therefore natural to associate such a temperature to the closure of the pseudogap . in the most interesting @xmath119 case ,
two features are clearly present in the @xmath117 curve .
the first , low - temperature feature is the evolution of the small-@xmath10 metallic feature , which acquires a larger slope as @xmath120 is increased due to the enhancement of the effective mass , and shrinks as a consequence of the reduced coherence temperature of the metal .
the second feature is instead the evolution of the large-@xmath10 insulating one , and would show an activated behavior partially hidden by the low-@xmath67 metallic peak .
thus , the system behaves like a metal in the small temperature range , while it has a pseudogap for intermediate temperature .
we estimate the lower boundary of the pseudogap region in this intermediate coupling regime through the maximum of the low - temperature feature , which is controlled by the effective coherence scale of the metal .
the upper bound is naturally defined as the temperature in which the activated behavior disappears . as a result
, the specific heat analysis determines a pseudogap region with a very similar shape than the one determined through the spin susceptibility , with a re - entrance of metallic behavior in the intermediate coupling regime at low temperatures . .
all the @xmath121 lines are obtained by differentiating the the internal energy @xmath122 .
the expression of @xmath122 is computed directly by fitting the dmft data.,width=321 ] for three different values of the interaction : @xmath123 and @xmath95 ( from the upper to the lower row ) both at the low - temperature ( @xmath124 , left panels ) and at the high - temperature ( @xmath125 , right panels).,width=302 ] an inspection to the spectral function can strengthen our insight on the pseudogap phase . in principle the ed algorithm allows to directly compute finite frequency spectral functions @xmath126 , avoiding the problems and ambiguities intrinsic with analytic continuation techniques .
unfortunately , the discretization of the hilbert space which allows for an ed solution results in `` spiky '' local spectral functions formed by a collection of @xmath127-functions . in this light
, we find it useful to compute @xmath128 by analytically continuing eq.([sigma ] ) , analytically computing the local retarded green function @xmath129 .
this procedure provides more `` realistic '' descriptions of both the non - interacting dos and of the strong - coupling pairing phase .
however , even though the spectral functions are smoothed by this procedure , we can only extract informations about the gross features of the spectra as , e.g. , the amplitude of the gap @xmath130 . keeping these limitations in mind ,
some results for @xmath128 are plotted in fig .
[ spectral ] : for the weak- ( @xmath96 ) , the intermediate- ( @xmath101 ) and strong - coupling ( @xmath131 ) case a low and high - temperature set of data are shown .
apart from the obvious appearing and enlarging of a gap in @xmath128 with increasing @xmath10 , which is evident in the low - temperature data , it should be noticed that both in the intermediate and the strong - coupling regime there is apparently no tendency to a closure of such a gap when the temperature is raised . indeed , as it is shown in the second and the third row of fig .
[ spectral ] , the gap starts to fill at some temperature ( @xmath132 for @xmath133 and @xmath134 for @xmath131 ) , but for these values of @xmath10 much of the spectral weight remains in the high - energy hubbard bands , and the gapped structure does not completely vanish up to the highest temperature reached in our calculation ( @xmath135 ) .
on the other hand , qmc results in @xmath19@xcite obtained through maximum entropy show a closure of the gap in @xmath128 at a temperature lower that our threshold .
further investigation is needed to understand whether the discrepancy is due to a different behavior between @xmath19 and the infinite dimensionality limit , or it is determined by the technical difficulties involved in the calculation of real frequency spectra in both approaches .
the persistence of the gap structure at high temperature that we find in dmft is also obtained within a perturbative analysis of superconducting fluctuations at strong coupling in @xmath19@xcite . in fig .
[ tstar ] we compare our estimates of the pseudogap temperature obtained through different physical quantities .
we draw the borders of the pseudogap region as determined from the spin susceptibility ( @xmath136 ) and the specific - heat behavior ( @xmath137 and the value of the superconducting gap @xmath138 at zero temperature : the upper borders of the spin and the specific - heat `` pseudogap '' region scale roughly with @xmath10 , as @xmath138 does , so that both @xmath137 and @xmath139 are proportional to @xmath138 , as the experimentally determined pseudogap . at low temperature ,
the pseudogap region boundary as extracted from thermodynamic response functions displays a clear re - entrance , which can be associated with the onset of the low - temperature quasi - particle peak .
we also notice that the low - temperature curve qualitatively follows the behavior of the @xmath69 line .
as mentioned above , the slope of @xmath69 is easily interpreted in terms of entropy balance between the two phases , which favors the preformed pairs phase .
our phase diagram also represents a warning regarding attempts to extrapolate the low - temperature behavior from the high - temperature data in order to compare with finite - dimensional qmc calculations . if one , as , e.g. , in ref .
, extrapolated the high - temperature behavior down to @xmath25 in order to estimate the metal - insulator point , would have obtained an estimate of @xmath140 significantly lower than the real @xmath63 .
this finding emphasize how the high - temperature properties of the attractive hubbard model are only weakly dependent on dimensionality , as indicated by the similarity between dmft and finite - dimension qmc , while the low - temperature behavior may well be dependent on the dimensionality , as well as on the details of the bandstructure of the underlying lattice . , i.e. , the temperatures of both the maxima and the minima of @xmath141 , while the empty circle marked @xmath142 that is the temperature associated with the maxima and the minima of @xmath117 .
the regions on the left of these two lines can be interpreted as the zone of `` pseudogap '' behavior for the spin and the specific heat respectively .
these lines are then compared with the behavior of the anomalous part of the self - energy at zero temperature ( @xmath138 , empty triangles).,width=321 ]
in this paper we have investigated the finite temperature aspects of pairing and superconductivity in the attractive hubbard model by means of dmft , considering both normal and superconducting solutions . in the normal phase
we have identified two families of solutions , a fermi - liquid metallic phase and a preformed - pair phase with insulating character . the latter phase is formed by local pairs without phase coherence .
a finite region of the coupling - temperature phase diagram is characterized by the simultaneous presence of both solutions . in the low temperature regime
a first - order transition occurs within this region when the free energies of the two solutions cross , and the region closes at a certain temperature ( @xmath143 ) in a critical point .
interestingly , some trace of the two solutions survives even for temperature larger than the critical temperature , and two crossover lines can be defined separating a normal metal , a sort of semiconductor in which the gap is closed by temperature , and the preformed - pair phase with a well defined gap .
when superconductivity is allowed , the superconducting solution is stable for all values of the attraction and the critical temperature is always larger than the pairing transition temperature in the normal phase . in the superconducting state , we find an evolution from a weak - coupling bcs - like behavior , with exponentially small @xmath27 from a normal metal to the superconductor , and a strong - coupling regime in which superconductivity is associated to the onset of the phase - coherence among the preformed pairs that occurs at @xmath144 .
the highest @xmath27 is obtained in the intermediate region between this two limiting cases , namely for @xmath145 , which is extremely close to the zero - temperature critical point of the normal phase .
the presence of the pairing transition affects the normal phase above @xmath27 also when superconductivity establishes . in particular
, one could be tempted to identify the phase of preformed pairs obtained at strong coupling with the pseudogap behavior observed in cuprates . in order to test the adequacy of such an identification
, we computed different observables , whose anomalies can identify the appearance of the pseudogap , like the spin susceptibility , the specific heat and the single particle spectral functions . in the intermediate region of coupling , where the pairing transition occurs and the superconducting critical temperature reaches its maximum , the pseudogap region presents a re - entrance at low temperatures associated with a small coherent peak in the spectral function . at temperatures smaller than this coherence temperature
the system behaves like a normal metal with renormalized effective mass . on the other hand ,
the high - temperature boundary of the pseudogap region scales with @xmath10 regardless the criterion we use to estimate it .
the estimate of the pseudogap temperature from specific heat and spin susceptibility both scale with the zero - temperature gap , as in the cuprates .
the most striking difference between our pseudogap phase - diagram and the experiments in the cuprates is that the pseudogap phase in the attractive hubbard model is much larger than the experimental one , as it is measured by the large value of @xmath146 at the optimal value of the attraction .
the experimental @xmath88 around optimal doping is instead very close to @xmath27 , and , according to some authors , the pseudogap line tends to zero at optimal doping .
moreover , the pseudogap temperature observed in the cuprates is definitely much smaller than the one found within our dmft of the attractive hubbard model .
this inadequacy of the attractive hubbard model in describing some features of the pseudogap phase descend from the above mentioned strong simplifications of the model ( neglect of retardation effects , coulomb repulsion and d - wave symmetry of the gap ) and of our dmft treatment which is exact only in the infinite dimensionality limit .
one could be tempted in maintaining an attractive hubbard model description for the quasiparticles alone , but it is important to point out that this interpretation can not be pushed too far . as an example , it is clear that such a description would fail for temperatures larger than the quasiparticle renormalized bandwidth .
a better description of the pseudogap phase would require models in which both an attraction and a repulsion are present .
this is for instance the case of the models introduced in refs . , where the superconducting phenomenon only involves heavy quasiparticles which experience an unscreened attraction and a richer behavior of the pseudogap ( which in this case closes around optimal doping ) is found .
this work is also supported by miur cofin 2003 .
we acknowledge useful discussions with s. ciuchi , m. grilli , and g. sangiovanni . | the finite - temperature phase diagram of the attractive hubbard model is studied by means of the dynamical mean field theory .
we first consider the normal phase of the model by explicitly frustrating the superconducting ordering . in this case
we obtain a first - order pairing transition between a metallic phase and a paired phase formed by strongly coupled incoherent pairs .
the transition line ends in a finite temperature critical point , but a crossover between two qualitatively different solutions still occurs at higher temperature . comparing the superconducting and the normal phase solutions , we find that the superconducting instability always occurs before the pairing transition in the normal phase takes place , i.e. , @xmath0 . nevertheless , the high - temperature phase diagram at @xmath1 is still characterized by a crossover from a metallic phase to a preformed pair phase .
we characterize this crossover by computing different observables that can be used to identify the pseudogap region , like the spin susceptibility , the specific heat and the single - particle spectral function . |
estimating the density @xmath12 of an unknown measure @xmath1 is a well - known problem in statistical analysis , physics or engineering . in a statistical context , one is usually given observations in the form of a sample of independent or dependent identically distributed random variables obtained from the unknown measure @xmath1 .
and so there has been extensive research on estimating the density based on these observations .
for instance , in one of the most popular approaches , the kernel density estimation @xcite , the density @xmath12 is estimated via a linear combination of kernel functions - each of them being identified with exactly one observation .
the crucial step in this method is to choose an appropriate bandwidth for which minimizing the integrated or the mean - integrated squared error between @xmath12 and its estimate is most common .
another very popular approach uses wavelets @xcite , an example of approximating a density by a truncated orthonormal series .
the coefficients in the truncated wavelet expansion are moment estimates derived from the given identically distributed observations . again
, the approximation accuracy is often measured by the mean - integrated squared error and depends on the number of observations and the degree of the truncation .
this approach provides a global density estimate satisfying both good local and periodic approximation properties . for further details the interested reader
is referred to @xcite and the many references therein . in another context - arising in challenging fields such as image recognition , solving nonlinear differential equations ,
spectral estimation or speech processing - no direct observation is available , but rather finitely many moments of the unknown measure @xmath1 are given .
then the issue is to reconstruct or approximate the density @xmath12 based on the only knowledge of finitely many moments , ( say up to order @xmath4 ) , an _ inverse _ problem from moments .
a simple method due to @xcite approximates the density @xmath12 by a polynomial @xmath8 of degree at most @xmath5 , so that the moments of the measure @xmath14 matches those of @xmath1 , up to order @xmath5 .
however , and in contrast with more sophisticated approaches , the resulting polynomial approximation @xmath8 is not guaranteed to be a density ( even though @xmath12 is ) as it may takes negative values on the domain of integration .
one classical approach to the moment problem is the pad approximation @xcite which is based on approximating the measure by a ( finite ) linear combination of dirac measures .
the dirac measures and their weights in the decomposition are determined by solving a nonlinear system of equations . in the maximum entropy estimation ( another classical approach )
one selects the best approximation of @xmath12 by maximizing some functional entropy , the most popular being the boltzmann - shannon entropy . in general
some type of weak convergence takes place as the degree increases as detailed in @xcite .
alternatively the norm of the approximate density is chosen as an objective function @xcite , which allows to show a stronger convergence in norm . in @xcite , maximum entropy and pad
approximates have been compared on some numerical experiments .
finally , piecewise polynomial spline based approaches have also been proposed in @xcite
. * motivation . *
our main motivation to study the ( inverse ) moment problem arises in the context of the so - called generalized problem of moments ( gpm ) .
the abstract gpm is a infinite - dimensional linear program on some space of borel measures on @xmath15 and its applications seem endless , see e.g. @xcite and the many references therein .
for instance , to cite a few applications , the gpm framework can be used to help solve a weak formulation of some ordinary or partial differential equations , as well as some calculus of variations and optimal control problems .
the solution @xmath12 of the original problem ( or an appropriate translate ) is interpreted as a density with respect to the lebesgue measure @xmath3 on some domain and one computes ( or approximates ) finitely many moments of the measure @xmath16 by solving an appropriate finite - dimensional optimization problem .
but then to recover an approximate solution of the original problem one has to solve an inverse problem from moments .
this approach is particularly attractive when the data of the original problem consist of polynomials and basic semi - algebraic sets . in this case one
may define a hierarchy ( as the number of moments increases ) of so - called semidefinite programs to compute approximations of increasing quality .
in this paper we consider the following inverse problem from moments : let @xmath1 be a finite borel measure absolutely continuous with respect to some reference measure @xmath3 on a box @xmath17 of @xmath15 and whose density @xmath12 is assumed to be in @xmath10 , with no continuity assumption as in previous works .
the ultimate goal is to compute an _ approximation _
@xmath18 of @xmath12 , based on the only knowledge of finitely many moments ( say up to order @xmath5 ) of @xmath1 . in addition , for consistency , it would be highly desirable to also obtain some convergence " @xmath19 as @xmath11 .
\(a ) firstly , we approximate the density @xmath12 by a polynomial @xmath20 of degree @xmath5 which minimizes the mean squared error @xmath21 ( or equivalently the @xmath10-norm @xmath22 ) over all polynomials @xmath8 of degree at most @xmath5 . we show that an unconstrained @xmath23-norm minimizer @xmath20 exists , is unique , and coincides with the simple polynomial approximation due to @xcite ; it can be determined by solving a system of linear equations .
it turns out that @xmath20 matches all moments up to degree @xmath5 , and it is even easier to compute if it is expressed in the basis of polynomials that are orthonormal with respect to @xmath3 .
no inversion is needed and the coefficients of @xmath20 in such a basis are just the given moments . moreover we show that @xmath24 in @xmath10 as @xmath11 , which is the best we can hope for in general since there is no continuity assumption on @xmath12 ; in particular @xmath25 almost - everywhere and almost - uniformly on @xmath17 for some subsequence @xmath26 , @xmath27 .
even though both proofs are rather straightforward , to the best of our knowledge it has not been pointed out before that not only this mean squared error estimate @xmath20 is much easier to compute than the corresponding maximum entropy estimate , but it also converges to @xmath12 as @xmath11 in a much stronger sense . for the _ univariate case _ , in references @xcite and @xcite the authors address the problem of approximating a _ continuous _ density on a compact interval by polynomials or kernel density functions that match a fixed number of moments . in this case , convergence in supremum norm is obtained when the number of moments increases .
an extension to the noncompact ( stieltjes ) case is carried out in @xcite . notice that in @xcite it was already observed that the resulting polynomial approximation also minimizes the mean square error and its coefficients solve a linear system of equations . in @xcite the minimum - norm solution ( and not the minimum distance solution ) is shown to be unique solution of a system of linear equations . in @xcite the minimal distance solution is considered but it is obtained as the solution of a constrained optimization problem and requires an initial guess for the density estimate .
\(b ) however , as already mentioned and unlike the maximum entropy estimate , the above unconstrained @xmath23-norm minimizer @xmath20 may _ not _ be a density as it may take negative values on @xmath17 .
of course , the nonnegative function @xmath28 $ ] also converges to @xmath12 in @xmath23 but it is not a polynomial anymore .
so we next propose to obtain a _ nonnegative polynomial _
approximation @xmath20 by minimizing the same @xmath23-norm criterion but now under the additional constraint that the candidate polynomial approximations should be nonnegative on @xmath17 . in principle
such a constraint is difficult to handle which probably explains why it has been ignored in previous works .
fortunately , if @xmath17 is a compact basic semi - algebraic set one is able to enforce this positivity constraint by using putinar s positivstellensatz @xcite which provides a nice positivity certificate for polynomials strictly positive on @xmath17 . importantly , the resulting optimization problem is convex and even more , a semidefinite programming ( sdp ) problem which can be solved efficiently by public domain solvers based on interior - point algorithms .
moreover , again as in the unconstrained case , we prove the convergence @xmath24 in @xmath10 as @xmath11 ( and so almost - everywhere and almost - uniform convergence on @xmath17 as well for some subsequence @xmath29 , @xmath27 ) which is far stronger than the weak convergence obtained for the maximum entropy estimate .
notice , in @xcite methods for obtaining some non - negative estimates are discussed , however these estimates do not satisfy the same properties in terms of mean - square error minimization and convergence as in the unconstrained case . in the kernel density element method @xcite a nonnegative density estimate for the univariate case is obtained by solving a constrained convex quadratic optimization problem . however , requiring each coefficient in the representation to be nonnegative as presented there seems more restrictive than the nonnegative polynomial approximation proposed in this paper .
\(c ) our approach is illustrated on some challenging applications . in the first set of problems we are concerned with recovering the shape of geometrical objects whereas in the second set of problems we approximate solutions of nonlinear differential equations
moreover , we demonstrate the potential of this approach for approximating densities with jump discontinuities , which is harder to achieve than for the smooth , univariate functions discussed in @xcite .
the resulting @xmath23-approximations clearly outperform the maximum entropy estimates with respect to both running time and pointwise approximation accuracy . moreover , our approach is able to handle sets @xmath17 more complicated than a box ( as long as the moments of the measure @xmath30 are available ) as support for the unknown density , whereas such sets are a challenge for computing maximum entropy estimates because integrals of a combination of polynomials and exponentials of polynomials must be computed repeatedly . in section [ notations ]
we introduce the notation and we state the problem to be solved . in section [ l2 ]
we present our approach to approximate an unknown density @xmath12 by a polynomial @xmath20 of degree at most @xmath5 via unconstrained and constrained @xmath23-norm minimization , respectively ; in both cases we also prove the convergence @xmath24 in @xmath10 ( and almost - uniform convergence on @xmath17 as well for some subsequence ) as @xmath5 increases . in section [ secnumex ]
we illustrate the approach on a number of examples - most notably from recovering geometric objects and approximating solutions of nonlinear differential equations - and highlight its advantages when compared with the maximum entropy estimation . finally , we discuss methods to improve the stability of our approach by considering orthogonal bases for the functions spaces we use to approximate the density . and
we discuss the limits of approximating discontinuous functions by smooth functions in connection with the well - known gibbs effect .
let @xmath31 $ ] ( resp .
@xmath31_d$ ] ) denote the ring of real polynomials in the variables @xmath32 ( resp .
polynomials of degree at most @xmath5 ) , whereas @xmath33 $ ] ( resp .
@xmath33_d$ ] ) denotes its subset of sums of squares ( sos ) polynomials ( resp .
sos of degree at most @xmath34 ) . with @xmath2 and a given reference measure @xmath3 on @xmath17 ,
let @xmath10 be the space of functions on @xmath17 whose square is @xmath3-integrable and let @xmath35 be the convex cone of nonnegative elements .
let @xmath36 ( resp .
@xmath37 ) be the space of continuous functions ( resp . continuous nonnegative functions ) on @xmath17
let @xmath38 be the space of polynomials nonnegative on @xmath17 .
for every @xmath39 the notation @xmath40 stands for the monomial @xmath41 and for every @xmath4 , let @xmath42 whose cardinal is @xmath43 .
a polynomial @xmath44 $ ] is written @xmath45 and @xmath46 can be identified with its vector of coefficients @xmath47 in the canonical basis @xmath48 , @xmath39 .
denote by @xmath49 the space of real @xmath50 symmetric matrices , and by @xmath51 the cone of positive semidefinite elements of @xmath49 . for any @xmath52 the notation @xmath53 stands for positive semidefinite .
a real sequence @xmath54 , @xmath55 , has a representing measure if there exists some finite borel measure @xmath1 on @xmath56 such that @xmath57 given a real sequence @xmath54 define the riesz linear functional @xmath58\to\r$ ] by : @xmath59.\ ] ] given @xmath60 , the moment matrix of order @xmath5 associated with a sequence @xmath54 , @xmath55 , is the real symmetric matrix @xmath61 with rows and columns indexed by @xmath62 , and whose entry @xmath63 is @xmath64 , for every @xmath65 . if @xmath66 has a representing measure @xmath1 then @xmath67 because @xmath68 with @xmath66 as above and @xmath69 $ ] ( with @xmath70 ) , the localizing matrix of order @xmath5 associated with @xmath66 and @xmath71 is the real symmetric matrix @xmath72 with rows and columns indexed by @xmath73 , and whose entry @xmath63 is @xmath74 , for every @xmath65 .
if @xmath66 has a representing measure @xmath1 whose support is contained in the set @xmath75 then @xmath76 because @xmath77 we consider the following setting . for @xmath2 compact , let @xmath1 and @xmath3 be @xmath78-finite borel measures supported on @xmath17 .
assume that the moments of @xmath3 are known and @xmath1 is absolutely continuous with respect to @xmath3 ( @xmath79 ) with radon - nikodm derivative ( or density ) @xmath80 , with respect to @xmath3 .
the density @xmath12 is unknown but we know finitely many moments @xmath81 of @xmath1 , that is , @xmath82 for some @xmath83 . the issue is to find an estimate @xmath84 for @xmath12 , such that @xmath85 we briefly describe the maximum entropy method due to @xcite as a reference for later comparison with the mean squared error approach .
if one chooses the boltzmann - shannon entropy @xmath86 , the resulting estimate with maximum - entropy is an optimal solution of the optimization problem @xmath87 it turns out that an optimal solution @xmath20 is of the form @xmath88 for some vector @xmath89 .
hence , computing an optimal solution @xmath20 reduces to solving the finite - dimensional convex optimization problem @xmath90 where @xmath81 is the given moment information on the unknown density @xmath12 .
if @xmath91 , @xmath4 , is a sequence of optimal solutions to ( [ concaveopt ] ) , then following weak convergence occurs : @xmath92 for all bounded measurable functions @xmath93 continuous almost everywhere . for more details the interested reader is referred to @xcite .
since the estimate @xmath20 is an exponential of a polynomial , it is guaranteed to be nonnegative on @xmath17 and so it is a density .
however , even though the problem is convex it remains hard to solve because in first or second - order optimization algorithms , computing the gradient or hessian at a current iterate @xmath94 requires evaluating integrals of the form @xmath95 which is a difficult task in general , except perhaps in small dimension @xmath96 or @xmath97 .
in this section we assume that the unknown density @xmath12 is an element of @xmath10 , and we introduce our mean squared error , or @xmath23-norm approach , for density approximation .
we now restrict @xmath18 to be a polynomial , i.e. @xmath98 for some vector of coefficients @xmath99 .
we first show how to obtain a polynomial estimate @xmath100_d$ ] of @xmath12 satisfying ( [ momconditions ] ) by solving an unconstrained optimization problem .
let @xmath101 denote the sequence of moments of @xmath3 on @xmath17 , i.e. , @xmath102 , @xmath39 , with @xmath103 and let @xmath104 denote the moment matrix of order @xmath5 of @xmath3 .
this matrix is easily computed since the moments of @xmath3 are known .
consider the unconstrained optimization problem @xmath105_d } \ : \displaystyle\vert u - u_d\vert_2 ^ 2 \:\left(= \int_{\omega } ( u - u_d)^2 d\lambda\right ) .
\label{unconopt}\ ] ] [ l2minprop ] let @xmath2 have a nonempty interior and let @xmath106 for some open set @xmath107 .
a polynomial @xmath108_d$ ] is an optimal solution of problem ( [ unconopt ] ) if and only if its vector of coefficients @xmath109 is an optimal solution of the unconstrained quadratic optimization problem @xmath110 then @xmath111 is the unique solution of ( [ unconqp ] ) , and @xmath108_d$ ] satisfies : @xmath112 observe that for every @xmath113_d$ ] with vector of coefficients @xmath114 , @xmath115 the third term on the right handside being constant , it does not affect the optimization and can be ignored .
thus , the first claim follows .
the second claims follows from the well - known optimality conditions for unconstrained , convex quadratic programs and the fact that @xmath104 is nonsingular because @xmath116 for all @xmath4 . indeed , if @xmath117 for some @xmath118 then necessarily the polynomial @xmath119_d$ ] with coefficient vector @xmath120 vanishes on the open set @xmath121 , which implies that @xmath122 , in contradiction with @xmath123 .
finally , let @xmath124 be the vector of coefficients associated with the monomial @xmath40 , @xmath125 .
from @xmath126 we deduce @xmath127 which is the desired result .
thus the polynomial @xmath108_d$ ] minimizing the @xmath23-norm distance to @xmath12 coincides with the polynomial approximation due to @xcite defined to be a polynomial which satisfies all conditions ( [ momconditions ] ) .
note that this is not the case anymore if one uses an @xmath128-norm distance with @xmath129 .
next , we obtain the following convergence result for the sequence of minimizers of problem ( [ unconopt ] ) , @xmath4 .
[ propconv ] let @xmath17 be compact with nonempty interior and let @xmath3 be finite with @xmath106 for some open set @xmath107 .
let @xmath91 , @xmath4 , be the sequence of minimizers of problem ( [ unconopt ] ) . then @xmath130 as @xmath11 .
in particular there is a subsequence @xmath26 , @xmath27 , such that @xmath25 , @xmath3-almost everywhere and @xmath3-almost uniformly on @xmath17 , as @xmath131 .
since @xmath17 is compact , @xmath132 $ ] is dense in @xmath10 .
hence as @xmath0 there exists a sequence @xmath133 $ ] , @xmath27 , with @xmath134 as @xmath131 . observe that if @xmath135 then as @xmath136 solves problem ( [ unconopt ] ) , it holds that @xmath137 for all @xmath138 , which combined with @xmath134 yields the desired result .
the last statement follows from ( * ? ? ?
* theorem 2.5.2 and 2.5.3 ) .
note that computing the @xmath23 norm minimizer @xmath20 is equivalent to solving a system of linear equations , whereas computing the maximum entropy estimate requires solving the potentially hard convex optimization problem ( [ concaveopt ] ) .
moreover , the @xmath23-convergence @xmath139 in proposition [ propconv ] ( and so almost - everywhere and almost - uniform convergence for a subsequence ) is much stronger than the weak convergence ( [ weak ] ) . on the other hand , unlike the maximum entropy estimate , the @xmath23-approximation @xmath108_d$ ] is not guaranteed to be nonnegative on @xmath17 , hence it is not necessarily a density .
methods to overcome this shortcoming are discussed in [ consubsec ] .
[ framermk ] in general the support @xmath140 of @xmath1 may be a strict subset of @xmath141 . in the case
where @xmath142 is not known or its geometry is complicated , one chooses a set @xmath143 with a simple geometry so that moments of @xmath3 are computed easily . as demonstrated on numerical examples in section [ secnumex ] ,
choosing an enclosing frame @xmath17 as tight as possible is crucial for reducing the pointwise approximation error of @xmath20 when the degree @xmath5 is fixed . in the maximum entropy method of [ subsecmee ]
no enclosing frame @xmath143 is chosen . in the @xmath23-approach ,
choosing @xmath143 is a degree of freedom that sometimes can be exploited for a fine tuning of the approximation accuracy . from the beginning
we have considered a setting where an exact truncated moment vector @xmath66 of the unknown density @xmath144 is given
. however , usually one only has an approximate moment vector @xmath145 and in fact , it may even happen that @xmath145 is not the moment vector of a ( nonnegative ) measure . in the latter case ,
the maximum of the convex problem ( [ concaveopt ] ) is unbounded , whereas the @xmath23-norm approach always yields a polynomial estimate @xmath108_d$ ] .
if @xmath145 is a slightly perturbed version of @xmath66 , the resulting numerical error in @xmath18 and side effects caused by ill conditioning may be reduced by considering the regularized problem @xmath146 where @xmath147 is the euclidean norm of the coefficient vector @xmath148 of @xmath113_d$ ] , and @xmath149 ( fixed ) is a regularization parameter approximately of the same order as the noise in @xmath145 . the coefficient vector of an optimal solution @xmath150 of ( [ unconoptregular ] ) is then given by @xmath151 the effect of small perturbations in @xmath66 on the pointwise approximation accuracy of @xmath20 for @xmath12 is demonstrated on some numerical examples in section [ secnumex ] . however , a more detailed analysis of the sensitivity of our approach for noise or errors in the given moment information is beyond the scope of this paper .
as we have just mentioned , the minimizer @xmath20 of problem ( [ unconopt ] ) is not guaranteed to yield a nonnegative approximation even if @xmath144 on @xmath17 . as we next see
, the function @xmath152 $ ] also converges to @xmath12 for the @xmath23-norm but * it does not satisfy the moments constraints ( i.e. , does not match the moments of @xmath153 up to order @xmath5 ) ; * it is not a polynomial anymore ( but a piecewise polynomial ) . in the sequel , we address the second point by approximating the density with a polynomial nonnegative on @xmath17 , which for practical purposes , is easier to manipulate than a piecewise polynomial .
we do not address explicitly the first point , which is not as crucial in our opinion .
note however that at the price of increasing its degree and adding linear constraints , the resulting polynomial approximation may match an ( a priori ) fixed number of moments .
adding a polynomial nonnegativity constraint to problem ( [ unconopt ] ) yields the constrained optimization problem : @xmath105_d } \:\ { \displaystyle\vert u - u_d\vert_2 ^ 2 \::\ : u_d \geq 0 \:\text { on } \omega\ } \label{nonnegopt}\ ] ] which , despite convexity , is untractable in general .
we consider two alternative optimization problems to enforce nonnegativity of the approximation ; the first one considers necessary conditions of positivity whereas the second one considers sufficient conditions for positivity .
consider the optimization problem : @xmath105_d}\:\ { \displaystyle\vert u - u_d\vert_2 ^ 2 \::\ :
\m_d(u_d\,\z ) \succeq 0\,\ } , \label{noliftopt}\ ] ] where @xmath102 is the moment sequence of @xmath3 and @xmath154 is the localizing matrix associated with @xmath18 and @xmath101 .
observe that problem ( [ noliftopt ] ) is a convex optimization problem because the objective function is convex quadratic and the feasible set is defined by linear matrix inequalities ( lmis ) .
the rationale behind the semidefiniteness constraint @xmath155 in ( [ noliftopt ] ) follows from a result in @xcite which states that if @xmath156 and @xmath157 for all @xmath5 then @xmath158 on @xmath17 .
[ denselemma ] if @xmath17 is compact then @xmath38 is dense in @xmath159 with respect to the @xmath23-norm . as @xmath17 is compact
the polynomials are dense in @xmath10 .
hence there exists a sequence @xmath160 $ ] , @xmath4 , such that @xmath161 as @xmath11 .
but then the sequence @xmath162 , @xmath4 , with @xmath163 $ ] for all @xmath164 , also converges for the @xmath23-norm . indeed , @xmath165 so let @xmath27 and @xmath166 be such that @xmath167 and so @xmath168 . as @xmath169 is continuous and
@xmath17 is compact , by the stone - weierstrass theorem there exists a sequence @xmath170 $ ] , @xmath171 , that converges to @xmath169 for the supremum norm .
hence @xmath172 for all @xmath173 ( for some @xmath174 ) .
therefore , the polynomial @xmath175 is positive on @xmath17 and @xmath176 therefore we have found a sequence @xmath177 , @xmath27 , such that @xmath178 as @xmath131 . [ propconvnolift ]
let @xmath17 be compact with nonempty interior .
then problem ( [ noliftopt ] ) has an optimal solution @xmath20 for every @xmath4 , and @xmath179 as @xmath11 .
fix @xmath5 and consider a minimizing sequence @xmath180_d$ ] with @xmath181 monotonically decreasing and converging to a given value as @xmath182 .
we have @xmath183 for all @xmath184 .
therefore as @xmath185 defines a norm on the finite dimensional space @xmath31_d$ ] ( @xmath17 has nonempty interior ) the whole sequence @xmath186 is contained in the ball @xmath187 . as the feasible set is closed , problem ( [ noliftopt ] ) has an optimal solution .
let @xmath188 $ ] be a sequence of optimal solutions of problem ( [ noliftopt ] ) .
by lemma [ denselemma ] , @xmath38 is dense in @xmath189 .
thus there exists a sequence @xmath133_{\geq 0}$ ] , @xmath27 , with @xmath190 as @xmath131 . as @xmath191 on @xmath17
then necessarily @xmath192 for all @xmath5 and all @xmath138 .
in particular @xmath193 where @xmath194 .
therefore @xmath195_d$ ] is a feasible solution of problem ( [ noliftopt ] ) which yields @xmath196 for all @xmath5 . combining with @xmath197 yields the desired result .
let @xmath198 denote the standard monomial basis of @xmath132_d$ ] .
let @xmath17 be a basic compact semi - algebraic set defined by @xmath199 for some polynomials @xmath200 $ ] , @xmath201 with @xmath202 . then , with @xmath203 , consider the optimization problem : @xmath204_d } & \displaystyle\vert u - u_d\vert_2 ^ 2\\ \text{s.t .
} & u_d(\x ) = \b_d(\x)^t \a_0\b_d(\x ) + \sum_{j=1}^m \b_{d - d_j}(\x)^t \a_j\b_{d - d_j}(\x)g_j(\x ) , \\ & \a_0 \in { \mathbb s}_{+}^{s(d)},\:\a_j \in { \mathbb s}_{+}^{s(d - d_j ) } , \quad j=1,\ldots , m . \end{array } \label{putinaropt}\ ] ] since the equality constraints are linear in @xmath18 and the entries of @xmath205 , @xmath206 , the feasible set of ( [ putinaropt ] ) is a convex lmi set .
moreover , the objective function is convex quadratic .
note that whereas the semidefinite constraint @xmath155 in ( [ noliftopt ] ) was a relaxation of the nonnegativity constraint @xmath207 on @xmath17 , in ( [ putinaropt ] ) a feasible solution @xmath18 is necessarily nonnegative on @xmath17 because the lmi constraint on @xmath18 is a putinar certificate of positivity on @xmath17 .
however , in problem ( [ putinaropt ] ) we need to introduce @xmath208 auxiliary matrix variables @xmath205 , whereas ( [ noliftopt ] ) is an optimization problem in the original coefficient vector @xmath209 and does not require such a lifting .
thus , problem ( [ noliftopt ] ) is computationally easier to handle than problem ( [ putinaropt ] ) , although both are convex sdps , which are substantially harder to solve than problem ( [ unconqp ] ) .
[ propconvlift ] let @xmath17 be compact with nonempty interior .
for every @xmath210 , problem ( [ putinaropt ] ) has an optimal solution @xmath20 and @xmath211 as @xmath11 .
with @xmath210 fixed , consider a minimizing sequence @xmath188_d$ ] for problem ( [ putinaropt ] ) .
as in the proof of proposition [ propconvnolift ] one has @xmath212 .
as @xmath20 is feasible , @xmath213 for some real symmetric matrices @xmath214 , @xmath206 . rewriting this as @xmath215 and integrating with respect to @xmath3 yields @xmath216 with @xmath217 . hence ,
for every @xmath5 , @xmath218 as @xmath219 , @xmath220 , and @xmath214 , @xmath201 , we conclude that all matrices @xmath221 are bounded
. therefore the minimizing sequence @xmath222 belongs to a closed bounded set and as the mapping @xmath223 is continuous , an optimal solution exists . from lemma [ denselemma ]
there exists @xmath224_{\geq0}$ ] , @xmath27 , such that @xmath225 as @xmath131 .
using properties of norms , @xmath226 , and so @xmath227 as @xmath131 .
moreover , as @xmath228 is strictly positive on @xmath17 , by putinar s positivstellensatz @xcite , there exists @xmath229 such that @xmath230 for some real matrices @xmath231 , @xmath206 . so letting @xmath232 $ ] , the polynomial @xmath228 is a feasible solution of problem ( [ putinaropt ] ) whenever @xmath233 and with value @xmath234 . hence @xmath235 as @xmath131 , and by monotonicity of the sequence @xmath236 , @xmath4 , the result follows .
since @xmath17 is compact , proposition [ propconvnolift ] and [ propconvlift ] imply that minimizers of the two constrained @xmath23 norm minimization problems ( [ noliftopt ] ) and ( [ putinaropt ] ) converge almost uniformly to the unknown density @xmath12 as in the unconstrained case .
[ rmkdom ] our approach can handle quite general sets @xmath17 and @xmath142 as support and frame for the unknown measure in the unconstrained and constrained cases , the only restriction being that ( i ) @xmath17 and @xmath142 are basic compact semi - algebraic sets , and ( ii ) one can compute all moments of @xmath3 on @xmath17 .
in contrast , to solve problem ( [ concaveopt ] ) by local minimization algorithms using gradient and possibly hessian information , integrals of the type @xmath237 must be evaluated .
such evaluations may be difficult as soon as @xmath238 .
in particular in higher dimensions , cubature formulas for approximating such integrals are difficult to obtain if @xmath17 is not a box or a simplex .
in this section , we demonstrate the potential of our method on a range of examples .
we measure the approximation error between a density @xmath12 and its estimate @xmath18 by the average error @xmath239 and the maximum pointwise error @xmath240 in some examples we also consider @xmath241 and @xmath242 , the respective errors on particular segments of the interior of @xmath17 .
we compare the performance of our approach to the maximum entropy estimation from [ subsecmee ] .
both methods are encoded in matlab .
we implemented the mean squared error ( @xmath23 ) minimization approach for the standard monomial basis , which results in solving a linear system in the unconstrained case . in the constrained case we apply sedumi to solve the resulting sdp problem . in both cases , the numerical stability can be improved by using an orthogonal basis ( such as legendre or chebychev polynomials ) for @xmath243 $ ] . in order to solve the unconstrained , concave optimization problem ( [ concaveopt ] ) , we apply the matlab optimization toolbox command fminunc as a black - box solver .
the observed performance of the maximum entropy estimation ( mee ) method may be improved when applying more specialized software . ,
title="fig:",scaledwidth=45.0% ] , title="fig:",scaledwidth=45.0% ] [ mnatsaex ] first we consider the problem of retrieving the density @xmath244 on @xmath245 ^ 2 $ ] given its moments @xmath246 , which was considered as a test case in @xcite .
this example is a priori very favorable for our technique since the desired @xmath12 is a polynomial itself . when solving problem ( [ unconopt ] ) for @xmath247 we do obtain the correct solution @xmath248 in less than @xmath249 secs
thus , the pointwise approximation is much better than the one achieved in @xcite .
moreover , @xmath250 without adding lmi constraints .
+ the question arises of how the polynomial approximation behaves if the moment vector contains some noise , i.e. if it does not exactly coincide with the moment vector of the desired @xmath12 .
we solve problem ( [ unconopt ] ) for @xmath251 where the maximal relative componentwise perturbation @xmath252 between @xmath66 and @xmath145 is less than @xmath253 .
the pointwise error between @xmath254 and the solution for the perturbed moment vector @xmath255 is sufficiently small , as pictured in figure [ mnatsafig ] . .estimating the density @xmath256 from an exact moment vector.[absfunresultsexact ] [ cols="^,^,^,^",options="header " , ] , green : @xmath257 unconstrained , red : @xmath257 constrained , cyan : @xmath258 unconstrained , magenta : @xmath258 constrained.[indfunputinarfig],scaledwidth=60.0% ] [ indfunputinarex ] as a first simple example where solving one of the constrained problems is required , consider @xmath259\rightarrow\r , u(x)=i_{[0.5 , 1]}(x)$ ] and its given vector of moments .
we solve problems ( [ unconopt ] ) and ( [ putinaropt ] ) for @xmath260 .
as illustrated in figure [ indfunputinarfig ] , solving problem ( [ putinaropt ] ) provides globally nonnegative estimates for @xmath12 .
this comes at the price of decreasing approximation accuracy compared to solving the unconstrained problem , since the feasible set of problem ( [ putinaropt ] ) is a subset of the feasible set of ( [ unconopt ] ) , cf .
table [ indfunputinarres ] .
density estimation via @xmath23 norm minimization over the truncated space of polynomials is a special case of approximating a function belonging to a complicated functional space by a linear combination of basis elements of an easier , well - understood functional space .
this general setting has been considered for both real and complex functional spaces . in the following
we discuss the particular case when the basis of the easier functional space is orthogonal , and also effects resulting from truncating the infinite dimensional series expansion of the unknown function .
the classic , complex analogue of the real , inverse moment problem discussed in the previous sections is the problem of approximating a periodic function @xmath12 by a trigonometric polynomial .
an orthonormal basis for the space of trigonometric polynomials is given by @xmath261 .
it provides the fourier series expansion : @xmath262 where @xmath263 in the univariate case , or @xmath264 where @xmath265 in the bivariate case .
one obtains a trigonometric polynomial approximation for @xmath12 when truncating this series expansion at some @xmath4 , @xmath266 it is a well - known fact in fourier theory that @xmath267 the fourier approximation for a periodic function is therefore the trigonometric analogue of the real polynomial approximation for a function obtained by mean squared error minimization in our approach . as in the real case
we have almost uniform convergence .
however , @xmath18 does not converge to @xmath12 uniformly if @xmath12 piecewise continuously differentiable with jump discountinuities , due to an effect known as the gibbs phenomenon .
it states that the truncated approximation shows a near constant overshoot and undershoot near a jump discontinuity .
this overshoot does not vanish , it only moves closer to the jump for increasing @xmath5 .
since the truncated fourier series is the trigonometric analogue of unconstrained mean squared norm minimizer of problem ( [ unconopt ] ) , the question arises of how the trigonometric estimate behaves for a periodic functions when adding nonnegativity constraints as in problems ( [ noliftopt ] ) and ( [ putinaropt ] ) which aim at preventing the estimate from over- or undershooting near jump discontinuities . in order to derive a tractable sdp problem ,
the nonnegativity constraints for the trigonometric polynomial need to be relaxed or tightened to lmi constraints .
the difference with the real case will be the moment and localizing matrices being of toeplitz type in contrast with the hankel type matrices in the constraints of problem ( [ noliftopt ] ) . in our discussions in the previous sections we approximated an unknown density by a linear combination of elements of the monomial basis of the space of real polynomials .
this choice of a basis for @xmath268 $ ] has several theoretical and practical shortcomings .
for once , it is not an orthogonal basis with respect to the lebesgue measure , and moreover the moment matrix @xmath104 in problem ( [ unconopt ] ) , whose inverse needs to be computed to determine @xmath20 , is severely ill - conditioned @xcite . as already pointed out in @xcite , when choosing an orthogonal basis such as legendre or chebychev polynomials for @xmath268 $ ] , the @xmath23 norm minimizer @xmath20 can be determined according to a closed - form formula .
thus , we do not even have to solve a linear system of equations .
this is essentially the real analogue of the closed form formula for the coefficients in the fourier series approximation of periodic functions . in [ subseccomplex ]
we discussed the gibbs phenomenon . for reasons outlined there
, we expect to observe this effect in the real case as well , when approximating an @xmath23 integrable function with jump discontinuities .
in fact , the function from example [ indfunputinarex ] illustrates that .
as shown in figure [ indfunputinarfig ] , we observe an over- and undershoot on both sides of the discontinuity
. the amplitude of this overshoot does not decrease for increasing @xmath5 , but the overshoot moves closer to the jump .
when adding the global nonnegativity constraint for the polynomial estimate , the undershoot at the left side disappears .
however , it is compensated for by a weaker , overall pointwise approximation accuracy of the estimate .
this observation is a first , partial answer to the question raised in the complex case .
we introduced an approach for estimating the density of a measure given a finite number of its moments only , in the multivarite case and with no continuity assumption on the density .
as an estimate we choose the polynomial minimizing the @xmath23 norm distance , or mean squared error to the unknown density .
we have shown that this estimate is easy to determine by solving a linear system of equations . moreover
, it converges almost uniformly towards the desired density when the degree increases . by minimizing the mean
squared error subject to additional linear matrix inequality constraints , which translates to solving a semidefinite program , we obtain a density estimate guaranteed to be nonnegative on the support of the measure .
also in the constrained case , we have shown almost uniform convergence of the nonnegative mean squared error minimizer towards the unknown density , when the degree increases .
mean squared error minimization is often superior to maximum entropy estimation in terms of approximation accuracy and computation time , as demonstrated for a number of examples .
moreover , it is able to handle general , basic , compact semialgebraic sets as support for the unknown measure , which present a challenge to maximum entropy estimation in higher dimension .
_ real analysis and probability _ , academic press inc . , boston , 1972 .
athanassoulis , p.n .
the truncated hausdorff moment problem solved by using kernel density functions , probabilistic engineering mechanics 17:273 - 291 , 2002 . g. a. baker , p. graves - morris .
pad approximants .
addison - wesley , reading , ma , 1981 .
m. bertero , c. de mol , e. r. pike .
linear inverse problems with discrete data .
i : general formulation and singular system analysis .
inverse problems 1(4):301330 , 1985 .
j. m. borwein , a. s. lewis . on the convergence of moment problems .
soc . 325(1):249271 , 1991 .
d. l. donoho , i. m. johnstone , g. kerkyacharian , d. picard .
density estimation by wavelet thresholding .
statist . 24(2):508539 , 1996 .
a. eloyan , s. k. ghosh .
smooth density estimation with moment constraints using mixture distributions
. j. nonparamet .
stat . 23(2):513531 , 2011 .
gavriliadis , g.a .
the truncated stieltjes moment problem solved by using kernel density functions , journal of computational and applied mathematics 236:4193 - 4213 , 2012 .
r. k. goodrich , a. steinhardt .
@xmath269 spectral estimation , siam j. appl . math . 46(3):417426 , 1986 .
d. henrion , j. b. lasserre , m. mevissen .
the generalized problem of moments for nonlinear differential equations .
work in progress , 2012 .
a. j. izenman .
recent developments in nonparametric density estimation . j. amer .
assoc . 86(413):205224 , 1991 .
e. t. jaynes .
information theory and statistical mechanics , phys .
series ii 106(4):620630 , 1957 .
e. t. jaynes .
information theory and statistical mechanics ii .
phys . rev .
series ii 108(2):171190 , 1957 .
v. john , i. angelov , a. a. ncl , d. thvenin .
techniques for the reconstruction of a distribution from a finite number of its moments . chemical engineering science 62(11):28902904 , 2007 .
l. k. jones , v. trutzer , on extending the orthogonality property of minimum norm solutions in hilbert space to general methods for linear inverse problems , inverse problems 6(3):379388 , 1990 .
g. kerkyacharian , d. picard .
density estimation by kernel and wavelet methods : optimality of besov spaces , stat .
letters 18(4):327336 , 1993 .
h. j. landau .
moments in mathematics .
appl . math .
37 , amer .
soc . , 1987
. j. b. lasserre .
moments , positive polynomials and their applications .
imperial college press , london , uk , 2009 . j. b. lasserre . a new look at nonnegativity on closed sets and polynomial optimization .
siam j. optim . 21:864885 , 2011 . s. x. liao , m. pawlak . on image analysis by moments
ieee trans .
pattern analysis and machine intelligence 18(3):254266 , 1996 .
l. r. mead , n. papanicolaou .
maximum entropy in the problem of moments .
j. math . phys . 25(8):24042417 , 1984 . m. mevissen , j. b. lasserre , d. henrion .
moment and sdp relaxation techniques for smooth approximations of problems involving nonlinear differential equation .
ifac world congress on automatic control , milan , italy , 2011 .
m. mimura .
asymptotic behaviors of a parabolic system related to a planktonic prey and predator model .
siam j. appl . math .
37(3):499 - 512 , 1979 .
r. m. mnatsakanov .
moment - recovered approximations of multivariate distributions : the laplace transform inversion .
letters 81(1):17 , 2011 .
e. parzen . on estimation of a probability density function and mode .
ann . math .
stat . 33:1065 - 1076 , 1962 .
m. putinar .
positive polynomials on compact semi - algebraic sets .
indiana univ .
j. 42:969984 , 1993 .
s.b . provost .
moment - based density approximations , the mathematica journal 9(4):727 - 756 , 2005 .
g. talenti .
recovering a function from a finite number of moments .
3(3):501 - 517 , 1987 . m. r. teague .
image analysis via the general theory of moments .
70(8):920 - 930 , 1980 . m. vannucci .
nonparametric density estimation using wavelets .
discussion paper 95 - 26 , isds , duke university , 1998 . | we consider the problem of approximating the unknown density @xmath0 of a measure @xmath1 on @xmath2 , absolutely continuous with respect to some given reference measure @xmath3 , from the only knowledge of finitely many moments of @xmath1 .
given @xmath4 and moments of order @xmath5 , we provide a polynomial @xmath6 which minimizes the mean square error @xmath7 over all polynomials @xmath8 of degree at most @xmath5 . if there is no additional requirement , @xmath6 is obtained as solution of a linear system .
in addition , if @xmath6 is expressed in the basis of polynomials that are orthonormal with respect to @xmath3 , its vector of coefficients is just the vector of given moments and no computation is needed .
moreover @xmath9 in @xmath10 as @xmath11 . in general nonnegativity of @xmath6 is not guaranteed even though @xmath12 is nonnegative .
however , with this additional nonnegativity requirement one obtains analogous results but computing @xmath13 that minimizes @xmath7 now requires solving an appropriate semidefinite program .
we have tested the approach on some applications arising from the reconstruction of geometrical objects and the approximation of solutions of nonlinear differential equations . in all cases
our results are significantly better than those obtained with the maximum entropy technique for estimating @xmath12 .
* keywords : * moment problems ; density estimation ; inverse problems ; semidefinite programming . |
measurements of cosmic microwave background ( cmb ) temperature anisotropies by the wilkinson microwave anisotropy probe ( wmap ) have ushered in the era of precision cosmology , confirming that the universe is spatially flat , with a matter budget dominated by dark matter and a baryonic mass fraction @xmath11 @xcite in agreement with the measured ratio of deuterium - hydrogen abundances ( d / h ) @xcite .
wmap measurements of large - scale cmb polarization also yield the optical depth @xmath12 to the surface of last scattering ( sls ) , meaningfully constraining cosmological reionization . together with surveys of supernovae @xcite , galaxies @xcite , and galaxy clusters @xcite
, wmap measurements build the case that the universe s expansion is accelerating , due to dark energy " or modifications of general relativity @xcite , and constrain other physical parameters ( such as the sum of neutrino masses @xmath13 @xcite and the effective number of massless neutrino species @xmath14 ) .
cmb temperature observations ( wmap , boomerang @xcite , cbi @xcite and acbar @xcite ) probe properties of the primordial density field , such as the amplitude @xmath15 , slope @xmath16 , and running @xmath17 of its power spectrum .
these observations constrain deviations from the adiabatic , nearly scale free and gaussian spectrum of perturbations predicted by the simplest models of inflation , but also offer controversial hints of deviations from these models ( see refs . @xcite and references therein ) .
experimental upper limits to b - mode polarization anisotropies ( e.g. dasi @xcite and bicep @xcite ) impose constraints to the energy density of relic primordial gravitational waves @xcite .
the _ planck _ satellite , launched in may 2009 , will obtain extremely precise measurements of the cmb temperature anisotropy power spectrum ( @xmath18 ) up to @xmath0 and the e - mode polarization anisotropy power spectrum ( @xmath19 ) up to @xmath20 @xcite .
robust measurements of the acoustic horizon and distance to the sls will break degeneracies in dark energy surveys @xcite .
polarization measurements will yield the optical depth @xmath12 to the sls @xcite , further constraining models of reionization and breaking the degeneracy between @xmath16 and @xmath12 @xcite .
cosmological parameters will be determined with much greater precision .
more precise values of @xmath16 and @xmath17 will be obtained from cmb data alone , helping to robustly constrain inflationary models and alternatives to inflation @xcite .
the advent of _ planck _ , ongoing ( spt @xcite and act @xcite ) experiments at small scales , and a future space based polarization experiment like cmbpol @xcite all require predictions of primary anisotropy multipole moments @xmath21 with @xmath22 accuracy . during atomic hydrogen
( h ) recombination , the thomson scattering opacity drops , decoupling the baryon - photon plasma and freezing in acoustic oscillations .
the phases of acoustic modes are set by the peak location of the visibility function @xcite , while damping scales @xcite and the amplitude of polarization @xcite are set by its width .
small - scale cmb anisotropies are also smeared out by free electrons along the line of sight , suppressing power on small scales so that @xmath23 , where @xmath12 is the total optical depth of this w @xcite .
an accurate prediction of the time - dependent free - electron fraction @xmath24 from cosmological recombination is thus essential to accurately predict cmb anisotropies .
recent work has highlighted corrections of @xmath25 to the standard recombination history computed by recfast @xcite .
these corrections will propagate through to predictions of anisotropies , and neglecting them would lead to biases and errors in _ planck
_ measurements of cosmological parameters @xcite .
the use of the cmb as a probe of the first ionizing sources and of physics at energy scales greater than @xmath26 thus requires an accurate treatment of the @xmath27 atomic physics of recombination @xcite .
direct recombination to the hydrogen ground state is ineffective because of the high optical depth to photoionization @xcite .
recombination proceeds indirectly , first through recombination to a @xmath28 state of h , and then by cascades to the ground state .
because of the optical thickness of the lyman-@xmath4 ( ly@xmath4 ) lines , the resulting radiation may be immediately absorbed , exciting atoms into easily ionized states
. there are two ways around this bottleneck @xcite . in the first
, the sequence of decays from excited h levels ends with a two - photon decay ( usually @xmath29 ) .
the emitted photons may have a continuous range of energies , allowing escape off resonance and a net recombination . in the second , photons emitted in the @xmath30 transition
redshift off resonance due to cosmological expansion , preventing re - excitation and yielding some net recombination .
the dominant escape channel is from the @xmath31 lyman-@xmath32 line .
these resonant transitions give off line radiation and distort the cmb @xcite .
peebles , sunyaev , kurt , and zeldovich modeled recombination assuming that all net recombination resulted from escaping the @xmath33 bottleneck @xcite .
this three - level - atom ( tla ) treatment included recombinations to excited states , under the assumption of equilibrium between energy levels @xmath4 and angular momentum sublevels @xmath34 for all @xmath35 ( note the use of @xmath34 for atomic angular momentum and @xmath36 for cmb multipole number ) .
this sufficed until the multi - level - atom ( mla ) model of seager et al .
@xcite , which included hydrogen ( h ) and helium ( he ) , separately evolved excited states assuming equilibrium between different @xmath34 , accurately tracked the matter / radiation temperatures @xmath37/@xmath38 @xcite , accounted for line emission using the sobolev approximation @xcite , and included @xmath39 chemistry .
this treatment underlies the recfast module used by most cmb anisotropy codes , including those used for wmap data analysis @xcite .
the higher precision of _ planck _ requires new physical effects to be considered , among them two - photon transitions from higher excited states in h and he @xcite , other forbidden and semiforbidden transitions in he @xcite , feedback from ly@xmath4 lines @xcite , and corrections to the sobolev approximation due to a host of radiative transfer effects in h and he resonance lines @xcite .
most recent work on recombination has focused in one way or another on the radiative transfer problem .
here we direct our attention to the populations of very high-@xmath4 states .
one important effect is the breakdown of statistical equilibrium between states with the same value of the principal number @xmath4 but different angular momenta @xmath34 .
this effect is dramatic at late times .
when @xmath34 sublevels of a level @xmath4 are resolved , increases in @xmath24 of @xmath40 at late times result @xcite .
this changes predicted @xmath21 s at a statistically significant level for _
planck_. highly excited states in hydrogen also change the recombination history at a level significant for _ planck_. while levels as high as @xmath41 were included in the treatment of ref .
@xcite underlying recfast , @xmath34 sublevels were not resolved .
it is thus important to update cosmological recombination histories to include high - n states of h while resolving @xmath34 sublevels , in order to predict the @xmath21 s as well as cmb spectral distortions from recombination . simultaneously including very high @xmath4 and resolving the @xmath34 sublevels is computationally expensive , taking nearly a week on a standard workstation for @xmath42 @xcite , using a conventional multilevel - atom recombination code .
this becomes prohibitively expensive for higher values of @xmath6 , unless considerable resources are devoted to the problem .
to date , this has prevented a determination of how @xmath24 converges with @xmath6 and how high @xmath6 must be to predict @xmath21 s for _ planck_. the existence of electric dipole selection rules @xmath43 means the relevant rate matrices are sparse , and we have used this fact to develop a fast code , recsparse , to explore convergence with @xmath6 . while the computation time @xmath44 for standard @xmath34-resolving recombination codes scales as @xmath45 , with recsparse
the scaling is @xmath46 , where @xmath47 .
with recsparse , we can calculate recombination histories for @xmath48 in @xmath49 days on a standard work - station ; this would likely take weeks using a conventional code .
for the first time , we have calculated recombination histories for @xmath6 as high as @xmath50 _ with @xmath34 sublevels resolved_. while previous computations have included some forbidden transitions , none have included optically thick electric quadrupole ( e2 ) transitions in atomic hydrogen .
we include e2 transitions , and find that the resulting correction to cmb anisotropies is negligible .
we find that the correction to cmb @xmath21 s due to extremely excited levels is @xmath51 or less if @xmath52 , in the purely radiative case .
this paper is not the final word on recombination ; atomic collisions must be properly included and the effect of levels with @xmath53 must be included to conclusively demonstrate absolute convergence .
the end goal of the present recombination research program is to include all important effects in a replacement for recfast , as the interplay of different effects is subtle . in sec .
[ mla ] , we review the formalism of the multilevel atom ( mla ) , and follow by explaining how we extend the mla to include very high - n states ( sec . [ highnsec ] ) and electric quadrupole transitions ( sec .
[ quadtheorysec ] ) .
state populations , recombination histories , and effects on the @xmath21 s are presented in sec .
[ res_highn ] .
we conclude in sec .
[ concl ] .
we use the same fiducial cosmology as in ref .
@xcite : total matter density parameter @xmath54 , @xmath55 , @xmath56 , @xmath57 , and helium mass fraction @xmath58 .
we now review the elements of the standard multilevel - atom ( mla ) treatment of cosmological recombination . for fundamental constants , we use nist ( national institute of standards and technology )
codata ( committee on data for science and technology ) values everywhere @xcite .
unless explicitly noted otherwise , we make the substitution @xmath59 in all expressions for the bohr radius @xmath60 and the ground - state hydrogen ionization potential @xmath61 to correctly account for reduced - mass effects .
cgs units are used except where explicitly noted otherwise .
we follow the abundance @xmath62 , where @xmath63 is the total number density of hydrogen nuclei and @xmath64 is the density of hydrogen in a state with principal quantum number @xmath4 and angular momentum @xmath34 ( we denote the state @xmath65 $ ] ) .
we evolve these abundances including bound - bound and bound - free radiative , single photon , dipole transitions , as well as the @xmath29 two - photon transition , which has rate @xmath66 @xcite .
focusing on the effect of single - photon dipole processes at high @xmath6 , we neglect higher @xmath4 two - photon processes but note that their effects are large enough that they must be included in a final recombination code @xcite .
note that we also neglect collisional transitions .
we comment on how this may change our conclusions in sec .
[ lresec ] .
bound - bound electric dipole processes are described by the equation @xcite @xmath67 with @xmath68 where @xmath69 is the downward einstein rate coefficient for decays from @xmath70 $ ] to @xmath65 $ ] and @xmath71 is the probability that a photon emitted in the @xmath70\to[n , l]$ ] line escapes the resonance without being reabsorbed .
this probability is calculated in the sobolev approximation , described in sec .
[ rad_trans ] . for lower @xmath34 states easily described using the @xmath72 orbital notation , we will sometimes use the notation @xmath73 , @xmath74 , and so on to simplify the discussion .
the degeneracy of @xmath65 $ ] is @xmath75 .
we explicitly keep track of the angular momentum quantum number @xmath34 , as this will simplify discussion of our sparse - matrix technique in sec .
[ sparse_sec ] . the photon occupation number blueward / redward of a line transition ( @xmath70\to [ n , l]$ ] )
is denoted @xmath76 where @xmath77 is the photon occupation number at photon energy @xmath78 and radiation temperature @xmath38 . here
@xmath79 is an infinitesimal line width and @xmath80 is the energy of a photon produced in the transition @xmath70\to[n , l]$ ] .
the simplest possible assumption for @xmath77 is a blackbody ; we discuss further subtleties in sec .
[ rad_trans ] : @xmath81 here @xmath82 is the usual boltzmann constant .
the @xmath83 term accounts for stimulated and spontaneous emission .
the two - photon term is @xcite @xmath84,\label{mla_c}\end{aligned}\ ] ] where @xmath85 and the second term describes two - photon absorption with a rate coefficient obtained by requiring that forward / backward rates satisfy the principle of detailed balance .
the bound - free term is @xcite @xmath86de_{e}\label{bfratenet},\end{aligned}\ ] ] with@xmath87p_{\rm m}\left(e_{e},t_{\rm m}\right)\end{aligned}\ ] ] and @xmath88 this integral is over the total energy @xmath89 of a recombining electron .
the energy of a recombination photon is @xmath90 , where @xmath91 is the bound - state energy of the recombined electron .
the recombination rate in @xmath92 of such an electron to the bound state @xmath93 $ ] is @xmath94 and is discussed in sec .
[ bfratesec ] .
the ionization rate in @xmath95 is @xmath96 , and easily shown by detailed balance considerations to be @xcite @xmath97 the free - electron abundance is @xmath98 , where @xmath99 is the free - electron density . we restrict our attention to times after helium recombination , and so the free proton abundance @xmath100 .
the net bound - free rate [ eq . ( [ bfratenet ] ) ] includes both spontaneous and stimulated recombination .
the electron energy distribution is a maxwellian with matter temperature @xmath37 : @xmath101 numerically solving the radiative transfer problem is computationally intensive , but tremendous simplification can be achieved with the sobolev escape probability formalism , also known as the sobolev approximation @xcite .
the hubble flow can be used to define a lengthscale over which the bulk flow induces a velocity change equal to the thermal velocity : @xmath102 , where @xmath103 is the value of the hubble expansion parameter when the radiation has temperature @xmath38 and @xmath104 is the mass of an atom @xcite .
the conditions of the sobolev approximation are @xcite : ( i ) @xmath105 is much smaller than the typical length scales over which cosmological quantities vary , ( ii ) @xmath106 is much smaller than the typical time scales over which cosmological quantities vary , ( iii ) complete frequency distribution the rest - frame frequency of an outgoing scattered photon @xmath107 does not depend on the incoming frequency @xmath108 and ( iv ) no other emission , absorption , or scattering processes occur in the vicinity of the line .
corrections to the sobolev approximation result from diffusion around resonance lines @xcite , atomic recoil @xcite , thomson scattering near resonances @xcite , and overlap of the higher ly series lines , leading to important corrections to cosmological recombination calculations . in this work , however , we work in the sobolev approximation to focus on other physical effects . in the sobolev approximation ,
the escape probability for photons produced in the downward transition @xmath70\to[n , l]$ ] is @xcite @xmath109 where the sobolev optical depth is given by @xmath110 with transition frequency @xmath111 correct expressions for @xmath112 are obtained by reversing arguments . during cosmological recombination , transitions between excited states are optically thin ( @xmath113 ) @xcite , and so we set @xmath114 in our calculations for non - lyman lines .
transitions in the lyman ( ly ) series ( @xmath115 , @xmath116 , @xmath117 ) are optically thick ( @xmath118 ) @xcite , and so @xmath119 .
ly transitions can not , however , be ignored in the recombination calculation , as the rate at which atoms find their way to the ground state through the redshifting of resonance photons , @xmath120 is comparable to @xmath121 and other two - photon rates @xcite . strictly speaking ,
@xmath122 depends on @xmath123 , and so one should solve for @xmath124 and then iteratively improve the solution .
the populations of the excited states , however , are very small and the maximum resulting correction to the optical depth is @xmath125 ( for @xmath126 ) @xcite .
we thus drop the second term in eq .
( [ sobdepth ] ) , simplifying our computation by working in the approximation where the lyman-@xmath4 ( ly@xmath4 ) line optical depth depends only on the ground - state population and not on the excited - state populations .
another aspect of the lyman - series lines is feedback : a photon that escapes from the ly@xmath4 ( @xmath127 ) line will redshift into the ly@xmath128 line and be reabsorbed .
recsparse has the ability to implement the resulting feedback , using the iterative technique of ref .
this slows down the code by a factor of a few , however , and so to efficiently focus on the @xmath6 problem , we turned feedback off .
for the high lyman lines , feedback is almost instantaneous : the universe expands by a factor of @xmath129 during the time it takes to redshift from ly@xmath4 to ly@xmath128 . in the instantaneous - feedback limit
, the ly@xmath4 lines do not lead to a net flux of h atoms to the ground state . to approximate
this net effect we turned off lyman transitions with @xmath130 ; this leads to a smaller error than would result from leaving these transitions on but disabling feedback .
previous tests using the code of ref .
@xcite show resulting errors in the recombination history at the @xmath131% level ; in any case , this should only weakly be related to the @xmath6 problem .
all of the recombination histories and plots in this paper were produced by running recsparse with both feedback and lyman transitions from @xmath130 disabled .
electrons , though nonrelativistic during recombination , interact with photons through thomson scattering . as a result , they do not follow the simple adiabatic scaling @xmath132 , where @xmath133 is the cosmological scale factor . the matter temperature is set using the asymptotic solution of ref .
@xcite for @xmath134 , after which the relevant ordinary differential equation ( ode ) is solved numerically ; this transition point occurs in the regime of mutual validity for the numerical and asymptotic solutions .
we neglect subdominant processes , such as free - free , line , photorecombination and collisional ionization cooling , as well as photoionization and collisional recombination heating @xcite .
the wide range of disparate time scales in this problem would naively necessitate a stiff differential equation solver .
this computational expense can be avoided by repackaging eqs .
( [ mla_a ] ) , ( [ mla_b]),and ( [ mla_c])-([mla_d ] ) .
these equations may be rewritten for excited states as ( @xmath65\neq[1,0]$ ] ) @xmath135 with @xmath136 where the integrated photoionization rate from @xmath65 $ ] is @xmath137 and @xmath138 is defined in eq .
( [ mla_b ] ) .
the downward flux to the ground state is @xmath139 where the first term describes ly@xmath4 series transitions ( stimulated and spontaneous ) while the second accounts for the @xmath140\to [ 1,0]$ ] two - photon transition .
kronecker delta symbols ( @xmath141 and @xmath142 ) are employed throughout to enforce @xmath65=[n^{\prime},l^{\prime}]$ ] and @xmath143 ) .
the source term @xmath144 includes flux from the ground state and direct recombination into the state @xmath65 $ ] : @xmath145 this can also be rewritten in matrix notation : @xmath146 , where @xmath147 is the matrix of rates with components given by eq .
( [ stiffset_b ] ) . the left - hand side of eq .
( [ stiffset_a ] ) is associated with the recombination time scale , while both terms on the right - hand side are associated with much shorter atomic time scales . for example , the longest lifetimes in the recombination problem are those of the @xmath148 and @xmath149 states ( @xmath150 and @xmath151 when ly-@xmath32 optical depth peaks at @xmath152 ) , considerably shorter than the recombination time scale of @xmath153 . thus we make a steady - state approximation , @xmath154 , which is formally valid because the reciprocal of the minimum eigenvalue of @xmath147 peaks at @xmath155 , which is @xmath156 of the duration of recombination .
thus the excited - state abundances are given by @xmath157 the rates in @xmath147 and @xmath158 depend on @xmath159 , @xmath160 , @xmath37 , @xmath38 , and @xmath161 .
the ground - state population is given by @xmath162\neq [ 1,0]}x_{n , l}$ ] , but since excited - state populations are small ( @xmath163 ) , @xmath160 can be eliminated from eq .
( [ stiffset_e ] ) using the approximation @xmath164 .
we can then solve for the evolution of @xmath159 , leaving out ineffective direct recombinations to the ground state : @xmath165\neq [ 1,0]}\left(\gamma_{nl}x_{n , l}-\frac{g_{l}}{2}a_{1,n}^{0,1}p_{1,n}^{0,1}\mathcal{n}_{1
n } x_{1s}\delta_{l,1}\right).\label{xe_evol } \end{array}\end{aligned}\ ] ] the steady - state approximation thus allows us to convert a stiff system of ordinary differential equations into a large system of coupled linear algebraic equations , along with a single ordinary differential equation .
the original effective 3-level atom " ( tla ) treatments of cosmological recombination in refs .
@xcite were built on the assumption that the primary bottlenecks to effective recombination are the slow @xmath29 transition rate and the reabsorption of @xmath31 resonance photons by the optically thick plasma .
other crucial assumptions included radiative equilibrium between excited states , @xmath166 and statistical equilibrium between angular momentum sublevels : @xmath167 recombination to higher excited states was included through an effective case b " total recombination constant @xmath168 ( recombinations to the ground state are omitted ) @xcite .
as the radiation field cools and the baryon density falls at late times , the transitions coupling high - n to low - n become inefficient , as do those coupling different @xmath34 sublevels with the same @xmath4 .
this leads to a breakdown of statistical equilibrium ( note however that the steady - state approximation is still valid ) , and so eqs . ( [ rad_eq ] ) and ( [ l_eq ] ) cease to apply . in ref .
@xcite , eq .
( [ rad_eq ] ) is relaxed while eq .
( [ l_eq ] ) is still imposed , and @xmath169 corrections to the tla prediction for @xmath24 result . at late times ,
nonequilibrium effects cause a net flux downward from states with quantum number @xmath4 to the ground state , accelerating recombination .
the inclusion of progressively more shells increases the number of downward cascade channels to the ground state for continuum electrons .
thus higher @xmath6 leads to faster recombination and lower @xmath24 .
reference @xcite reports results for @xmath6 as high as @xmath170 .
the lyman ( @xmath30 ) transitions from very high-@xmath4 states overlap with the lyman continuum , motivating ref .
@xcite s claim that there is no need to go past @xmath41 .
the real question as to whether the different values of @xmath4 are well defined , however , is whether the broadening of the state , @xmath171 ( where @xmath12 is the lifetime ) is larger than the splitting of adjacent energy levels , @xmath172 .
the intrinsic broadening for a typical level with @xmath173 is @xmath174 @xcite .
thus @xmath175 and so these extremely high-@xmath4 energy levels are well defined ; indeed , transitions between highly excited states in such nonequilibrium plasmas are seen in interstellar h ii regions and are a useful diagnostic of physical conditions @xcite . for extremely large @xmath4 ,
the above physical argument may break down because of additional broadening contributed by interactions with the radiation field and the plasma .
for example , the broadening due to stimulated emission and absorption scales as @xmath176 ( the spontaneous @xmath177 times the phase space density for photons in the @xmath178 transitions ) and that due to electron - impact collisions scales as @xmath179 @xcite ; at sufficiently high @xmath4 these will dominate over @xmath180 and the atomic energy levels will become blended .
however , the orders of magnitude of the collisional coefficients @xcite suggest that this occurs at values of @xmath4 larger than those considered in this paper .
we have also verified that for conditions of interest for the recombining cosmological plasma , the plasma debye length is greater than the average bound electron radius @xmath181 as long as @xmath182 .
more recent work @xcite shows that additional @xmath183 corrections to @xmath24 arise when eq .
( [ l_eq ] ) is not imposed and the populations of @xmath34 sublevels are followed separately . bottlenecks to decays from high @xmath34 imposed by @xmath184 slow down cascades to the ground state , and thus lead to slower recombination . in this case , the sidelength of @xmath147 is @xmath185 . since the number of computational steps needed to invert a matrix is generically a @xmath186 process , the computational time needed for a single ode time step in the recombination time will be proportional to @xmath187 . as noted in ref .
@xcite , a recombination calculation with @xmath42 already takes @xmath188 days on a standard workstation .
it this thus difficult to explore how quickly @xmath189 converges for progressively higher values of @xmath6 .
even between @xmath190 and @xmath42 , @xmath183 changes are seen in the tt and ee multipole moments ( @xmath21 s ) of the cmb on cmb @xmath21 s . in that work
, it is noted that the fractional difference between the @xmath21 s for @xmath191 and @xmath192 falls within a heuristic _
planck _ performance benchmark .
higher values of @xmath6 come even closer to the fiducial case of @xmath192 , a fact used to argue that even @xmath191 recombination is adequate for _ planck _ data analysis . from the cauchy convergence criterion , however , we know that a meaningful convergence test requires a comparison between successive members in a sequence . using the results of ref .
@xcite alone , the question of convergence with @xmath6 thus remains open . ]
@xcite . in spite of the computational challenge
, it is thus crucial to push the calculation to sufficiently high @xmath6 that corrections to @xmath24 from remaining @xmath53 are so small that they do not effect @xmath18 or @xmath19 at a level statistically significant compared to the predicted _ planck _ sample variance ( e.g. , several parts in @xmath193 for @xmath194 ) @xcite .
there are two challenges in treating such a big multilevel atom .
the first is the calculation of atomic transition rates at extremely high @xmath4 ; this is tractable because of some convenient recursion relations .
the second is simultaneously evolving the populations of @xmath195 states .
we discuss these in turn below . here
we discuss the einstein coefficients for dipole bound - bound and bound - free transitions in atomic hydrogen , which are used in our recombination computation .
we omit reduced - mass corrections to make a consistent comparison with refs .
@xcite , but include them when calculating actual recombination histories .
the spontaneous electric dipole transition rate @xmath196 for a nonrelativistic hydrogen atom is given by @xcite @xmath197 , \label{ab}\end{aligned}\ ] ] where @xmath198 is the charge of an electron , @xmath199 is the _
planck _ constant , and @xmath200 denotes the radial matrix element between the states @xmath93 $ ] and @xmath201 $ ] at order @xmath202 in the multipole expansion .
for example , @xmath203 denotes the quadrupole rate , and so on.the restriction @xmath184 enforces electric dipole selection rules . here @xmath204 is the radial wave function of an electron in a hydrogen atom , with principal quantum number @xmath4 and angular momentum quantum number @xmath34 , at a dimensionless distance @xmath205 .
all dimensionless distances are measured in terms of @xmath60 . for coulomb wave functions ,
this integration yields the gordon formula @xcite : @xmath206 where @xmath207 , @xmath208 with @xmath209 , @xmath210 , and @xmath211 . here
@xmath212 is gauss s hypergeometric function for integer @xmath213 and @xmath214 , evaluated using the recursion relationship @xmath215+\left(a+bx - c\right ) { \ensuremath{{}_{2}\!{f_{1}}}}\left(a , b , c;x\right),\end{aligned}\ ] ] with initial conditions @xmath216 we use eqs .
( [ ab])-([hgc ] ) to calculate bound - bound transition rates at the beginning of a mla computation , storing them for easy and repeated access .
we compared the resulting radial matrix elements with several values for high @xmath4 in ref .
@xcite and found agreement to all @xmath217 published digits .
we calculated oscillator strengths and compared with ref .
@xcite ( all transitions with @xmath4 and @xmath218 were evaluated , as was the entire balmer series for @xmath219 ) and found agreement to all @xmath220 published digits . we also compared with the results in ref .
@xcite ( in which oscillator strengths were computed up to @xmath221 for @xmath222 ) and found agreement to @xmath223 digits .
we attribute the difference in oscillator strengths to the fact that a polynomial expansion of @xmath224 was used in ref .
@xcite , rather than the more stable recursion relationship .
we also compared with the dipole one - photon rates used for the @xmath225 mla computation of ref .
most rates agreed to @xmath226 or more significant figures .
transition rates between @xmath227 and @xmath202 orbitals only agreed to @xmath228 significant figures .
we ran our mla model using the rates of ref .
@xcite and verified that these small disagreements do not lead to any differences in @xmath189 at the desired level of accuracy .
given the high quantum numbers considered , it was important to verify that no numerical instability plagues our numerical implementation of these recursions .
we thus checked matrix elements computed using eqs .
( [ hga])-([hgc ] ) against values estimated using the wkb approximation , as detailed in the appendix . [ bbratesec ] bound - free rates are evaluated using the same principle , but one of the two states used to evaluate matrix elements must be a continuum coulomb wave function .
the resulting matrix element is @xcite@xmath229 where @xmath230 is the continuum coulomb wave function for a recombining photoelectron with angular momentum quantum number @xmath231 and dimensionless energy @xmath232 .
the energy of the outgoing photon is @xmath233 .
this integral may also be evaluated in terms of hypergeometric functions , which in turn yields a recursion relationship for @xmath234 @xcite : @xmath235\nonumber g_{n,\kappa}^{l-1,l}\\ & -4n^{2}\left(n^{2}-l^{2}\right)\left[1+\left(l+1\right)^{2}\kappa^{2}\right]g_{n,\kappa}^{l , l+1}\nonumber , \\
g_{n,\kappa}^{l-1,l-2}&=\left[4\left(n^{2}-l^{2}\right)+l\left(2l+1\right)\left(1+n^{2}\kappa^{2}\right)\right]\nonumber g_{n,\kappa}^{l , l-1}\\ & -4n^{2}\left[n^{2}-\left(l+1\right)^{2}\right]\left(1+l^{2}\kappa^{2}\right)g_{n,\kappa}^{l+1,l}\label{befmel } .\end{aligned}\ ] ] the initial conditions of the recursion are @xcite
@xmath236 these matrix elements are tabulated at the beginning of each mla run for all @xmath237 , and @xmath238 ; this range of @xmath239 is partitioned into @xmath240 logarithmically spaced bins , with each bin containing @xmath241 equally spaced @xmath239 values .
bound - free matrix elements were compared with tabulated values for low @xmath4 in ref .
@xcite and agreed to all @xmath242 listed digits .
matrix elements were also compared with those used in ref .
@xcite ; we found agreement to one part in @xmath243 , aside from @xmath244 transitions , as already discussed .
the recombination rate to @xmath93 $ ] as a function of energy is then @xmath245 with @xmath246 at each value of @xmath247 , the tabulated matrix elements , eqs .
( [ bfratenet ] ) and ( [ recratenet ] ) are used to calculate thermally averaged recombination rates , using an @xmath241-point newton - cotes @xcite formula for the integration and neglecting stimulated emission .
large bins are added into the integral until it has converged to a fractional precision of @xmath248 .
we compared our values with integrated rates tabulated in ref .
@xcite and found agreement to all @xmath242 listed digits . comparing with the rates used in ref .
@xcite , we found agreement to one part in @xmath243 , aside from s - p transitions . + + in saha equilibrium , @xmath249p_{\rm m}\left(e_{e},t_{\rm m}\right)\nonumber\\=\eta_{\rm h}x_{n , l}\mathcal{n}\left(e_{\gamma},t_{\rm r}\right)\beta_{nl}\left(e_{\gamma}\right),\end{aligned}\ ] ] and so by the principle of detailed balance , @xmath250}{\mathcal{n}\left(e_{\gamma},t_{\rm r}\right ) } \right|_{\rm eq } p_{\rm m}\left(e_{e}\right).\end{aligned}\ ] ] we verified that our computed thermally averaged recombination and ionization rates satisfied this equality to machine precision .
we also checked bound - free matrix elements computed using eq .
( [ befmel ] ) against values estimated using the wkb approximation , as detailed in the appendix [ appendix_wkb ] . [ bfratesec ] the key to making the recombination problem tractable for high values of @xmath6 is the sparsity of eqs .
( [ stiffset_a ] ) and ( [ stiffset_b ] ) .
dipole selection rules only allow coupling of states with angular momentum quantum numbers @xmath34 and @xmath251 if @xmath184 .
it is easiest to understand how sparsity simplifies the problem with a slight change of notation .
we can compose the vector @xmath252 ( with components @xmath253 ) of excited - state populations , as @xmath254 where @xmath255 and @xmath256 denotes a vector of the populations of all states with angular momentum @xmath34 , except for the @xmath257 state .
specifically , @xmath258 where @xmath259 the source vector @xmath158 can similarly be written by concatenating source vectors @xmath260 ; each @xmath260 feeds all states with angular momentum @xmath34 .
the rate matrix may be similarly built of submatrices @xmath261 , as illustrated in fig .
the complete rate matrix is block tridiagonal , and the blocks decrease in dimension as @xmath34 increases .
the matrix @xmath261 has components @xmath262 in the steady - state approximation , eq . ( [ stiffset_a ] ) can be rewritten as a system of matrix equations . if @xmath117 , @xmath263 if @xmath264 , @xmath265 to close the system , we must truncate the hierarchy by excluding states with @xmath53 as both sources and sinks , which is equivalent to setting @xmath266 for @xmath267 .
then for @xmath268 , @xmath269 it might be possible to approximate the correction due to this truncation error , using asymptotic expressions for @xmath270 and saha equilibrium abundances for @xmath53 .
this will only work if @xmath6 is sufficiently high for nearly perfect equilibrium saha equilibrium to hold between states with @xmath53 and the continuum . at any given time step ,
the actual quantity of interest is not the inverse @xmath271 of the rate matrix but the solution set @xmath272 to the steady - state rate equations .
the closed form solution to eqs .
( [ bottom])-([top ] ) is @xmath273 , \label{ssola}\ ] ] if @xmath274 . if @xmath268 , then @xmath275.\label{ssolb}\ ] ] here @xmath276 and @xmath277 our new mla code , recsparse , operationally implements this solution at each time step as follows : 1 . using the values of @xmath38 and @xmath159 , @xmath37 is calculated using the results of sec .
[ rad_trans ] .
all relevant @xmath261 and @xmath260 are computed using eqs .
( [ submat_comp ] ) and ( [ stiffset_e ] ) and stored .
all @xmath278 and @xmath279 are computed using eqs .
( [ saa])-([sab ] ) and stored .
4 . equation ( [ ssolb ] ) is applied to obtain the solution for @xmath280 .
5 . equation ( [ ssola ] ) is iterated to obtain the solutions for all @xmath256 .
the free - electron fraction @xmath159 is then evolved forward in time using @xmath272 and eq .
( [ xe_evol ] ) .
it would also be interesting to compute the cumulative spectral distortion emitted by the line and continuum processes responsible for recombination @xcite .
this fractional perturbation of @xmath281 to the blackbody intensity of the cmb could be detectable with future experiments and would offer a test both of our understanding of recombination and of new physics behind the surface of last scattering ( e.g. , time variation of fundamental constants , energy injection by decaying / annihilating dark matter ) @xcite .
this and the development of a fast code for _ _ planck _ _ data analysis including all the relevant physical effects will be the subject of future work .
recsparse begins at @xmath282 , assuming saha equilibrium to compute the initial value of @xmath159 and setting @xmath37 as discussed in sec .
[ rad_trans ] .
excited - state populations are obtained using the method of sec .
[ sparse_sec ] .
submatrix inversions are implemented using the double precision routine dgesvx from the lapack library @xcite .
time evolution of @xmath24 with eq .
( [ xe_evol ] ) is implemented using the @xmath283-order runge - kutta - cash - karp ( rkck ) implementation in _ numerical recipes _ @xcite .
the rapid time scale for return to saha equilibrium introduces a stiff mode into the equations at early times , necessitating care in the choice of a stepsize for the integrator .
we were able to achieve relative precision of @xmath284 by placing @xmath285 time steps at @xmath286 and @xmath50 steps in the range @xmath287 , partitioning each interval into equally sized steps in @xmath288 ; relative errors were estimated by halving step size and comparing values of @xmath24 at identical time steps .
the computation time @xmath44 for recsparse scales as @xmath46 , where @xmath47 .
this is an empirical estimate for the range of @xmath6 that we have explored , and may not extend to higher @xmath6 values .
in contrast , for standard mla codes , @xmath289 .
we can calculate recombination histories for @xmath48 in @xmath49 days on a standard workstation ; this would likely take weeks using a conventional mla code .
[ num_meth ]
early work on recombination highlighted the importance of forbidden transitions , as half of the hydrogen atoms in the universe form by way of the @xmath290 forbidden " transition @xcite .
recent work has included additional forbidden " transitions in the mla treatment , namely , two - photon transitions ( @xmath291 and @xmath292 ) in h @xcite , two - photon and spin - forbidden transitions in he @xcite , as well as electric quadrupole ( e2 ) transitions in he @xcite . until this work , the impact of e2 transitions in h on recombination has not been considered , even though they are optically thick for transitions to / from the ground state . for optically thick lines ,
the overall transition rate is proportional to @xmath293 .
since @xmath294 , the overall transition rate is independent of the rate coefficient .
transitions such as electric quadrupoles , which seem weaker " judging from rate coefficients alone , can thus be as important as stronger " transitions , like the ly@xmath4 lines .
for example , this is why the semiforbidden he i 591 line is important in cosmological recombination @xcite .
we thus include e2 quadrupole transitions in our mla computation to properly assess their relevance for cosmological recombination .
m1 ( magnetic dipole ) transition rates in h are typically suppressed by an additional factor of @xmath295 , and are thus negligible @xcite .
the electric quadrupole ( e2 ) einstein a - coefficient for transitions from states @xmath93 $ ] to states @xmath296 $ ] is @xcite : @xmath297 where the quadrupole matrix element is @xmath298 the matrix elements of the reduced angular tensor operator @xmath299 are given by @xmath300 where the last factor is the well - known wigner-@xmath301 symbol .
this operator is defined as @xmath302 the dimensionless radial quadrupole integral is @xmath303 the radial matrix element for the @xmath304 transition is a special case of eq .
( b.13 ) of ref .
@xcite with @xmath305 : @xmath306^{1/2}\frac{\left(n-1\right)^{n-4}}{\left(n+1\right)^{n+4}}.\label{crap}\end{aligned}\ ] ] the obvious way to include quadrupole transitions into our mla code would be to generalize eq .
( [ mid ] ) to include @xmath307 transitions : @xmath308 the resulting system is obviously not as sparse as in the dipole case , and solving for all @xmath309 would be computationally more expensive , slowing down the whole mla code .
since the contribution from even the largest quadrupole rates may turn out to be small , we pursue a computationally less expensive approach .
higher energy e2 transitions will proceed much faster than lower energy ones , since e2 rates scale as @xmath310 .
in particular , transitions to and from the @xmath257 ground state will dominate any other quadrupole contributions to the recombination problem , since @xmath311^{5}{\mathrel{\hbox{\rlap{\lower.55ex\hbox{$\sim$ } } \kern-.3em \raise.4ex \hbox{$>$}}}}10^{3}~\mbox{if $ q\geq 2$}.\ ] ] moreover , the @xmath312 lines are optically thick for small @xmath4 .
we thus restrict our consideration to @xmath313 transitions , since other quadrupole transitions are corrections to a correction . "
a further simplification follows if we recall that the ly@xmath4 lines are all optically thick @xcite .
thus , the transition @xmath292 is highly probable to be immediately followed by a transition @xmath314 .
this yields a net @xmath315 transition , analogous to an @xmath34-changing collision , which occurs with forward rate @xmath316 .
the reverse process occurs with rate @xmath317 , where @xmath318 is a factor relating forward and backward rates .
if the @xmath202 and @xmath319 states were in equilibrium , the two rates would cancel , so by the principle of detailed balance , @xmath320 , where @xmath321 " denotes an equilibrium value .
the net @xmath322 transition rate due to e2 transitions is thus @xmath323 since this overall rate obeys the @xmath43 selection rule , it can be numerically implemented within the same framework as the dipole rates .
we ran the recsparse code for a variety of @xmath6 values . here
we omitted e2 transitions to focus on the effect of deviations from statistical equilibrium and increasing @xmath6 .
we begin by discussing deviations from equilibrium , and proceed to discuss the recombination history and numerical convergence with @xmath6 .
[ res_highn ] the assumptions of statistical equilibrium between different @xmath34 sublevels within the same @xmath4 shell and boltzmann equilibrium between different @xmath4 states fail at late times , as discussed in sec .
[ highnsec ] . furthermore ,
as reactions become inefficient on the hubble time scale and @xmath24 freezes out , saha equilibrium between the continuum and excited states of h also fails .
below , we discuss each of these failures quantitatively . at early times , the populations of hydrogen atoms in states with the same @xmath4 but different angular momentum @xmath34 are in statistical equilibrium [ see eq.@xmath324([l_eq ] ) ] .
radiative transitions do not include reactions that are @xmath34 changing but @xmath4 conserving .
the @xmath34 sublevels must thus be kept in equilibrium by a combination of sequences of allowed radiative transitions and atomic collisions .
these processes become inefficient at later times , leading the different @xmath34 sublevels to fall out of equilibrium . both the tla treatment of peebles and the later mla treatment of seager et al . rely on the statistical equilibrium assumption @xcite .
our recsparse code relaxes this assumption and follows the populations of all @xmath34 sublevels separately . for @xmath325 ,
the resulting populations are marked by several features , shown in figs .
[ res_state_gas_lneq_boltz_fine ] and [ res_state_gas_lneq_boltz ] at early and late times , respectively .
we use @xmath326 to compare actual with equilibrium populations , where @xmath327 deviations begin modestly at early times ( @xmath328 for @xmath329 ) but are quite large by late times ( @xmath330 by @xmath331 ) .
lower @xmath34 states depopulate efficiently , and are significantly underpopulated relative to statistical equilibrium expectations . states with @xmath117 can only make downward dipole transitions in @xmath4 if @xmath116 . these rates are several order of magnitude lower than lyman - series rates with the same @xmath332 , and so @xmath117 states depopulate less efficiently than other low-@xmath34 states .
this explains the upturn at the lowest @xmath34 values .
the @xmath43 selection rule implies that higher @xmath34 states couple efficiently to neighboring bound states ( @xmath184 ) with a limited range of accessible @xmath333 , since @xmath334 .
these states thus depopulate less efficiently than states with lower @xmath34 due to this bottleneck .
the recombination rate @xmath335 peaks in the range @xmath336 .
together , these facts imply the presence of a peak in @xmath337 , which turns out to occur in the range @xmath338 for a wide range of @xmath4 at all times .
the transition to @xmath339 occurs in the range @xmath340 , also for a wide range of @xmath4 at all times . at very high @xmath34 ,
recombination rates are so slow that these states are again underpopulated relative to statistical equilibrium , though less dramatically than they are at low @xmath34 .
the observed amplitude and shape of the curves in figs . [ res_state_gas_lneq_boltz_fine]-[res_state_gas_lneq_boltz ] qualitatively agree with the results in refs .
@xcite-@xcite , including the upturn near the lowest @xmath34 and sharp minimum at @xmath341 .
the minimum is due to fast balmer transitions out of the @xmath341 state .
when we computed a recombination history with these rates ( @xmath342 for @xmath35 ) artificially set to zero , the minimum moved to @xmath343 , as shown in fig .
[ balmerfig ] .
it is interesting that the curves in figs .
[ res_state_gas_lneq_boltz_fine]-[res_state_gas_lneq_boltz ] exhibit the same behavior with @xmath34 as the departure coefficients of ref .
@xcite , which describe neutral hydrogen ( also in the steady - state approximation ) in interstellar h ii regions .
recsparse only takes into account radiative transitions , and omits @xmath34 and @xmath4-changing collisions .
these rates would flatten all the curves in figs .
[ res_state_gas_lneq_boltz_fine]-[balmerfig ] , lessening deviations from statistical equilibrium between the different @xmath34 sublevels @xcite .
indeed , the assumption of statistical equilibrium between these states at all times is formally equivalent to the limit of infinite @xmath34-changing collision rates .
theoretical estimates for collisional rates all depend on different assumptions and tabulated rates disagree by factors of two or more ( see , e.g. , ref .
@xcite ) . as a function of redshift @xmath344
, we estimate the ratio @xmath345 of collisional to radiative transition rates out of the state @xmath93 $ ] , where @xmath346 is the collisional rate coefficient ( in @xmath92 ) , and @xmath347 is the total radiative lifetime of the state , including stimulated emission and absorption . using the rate coefficients in ref .
@xcite , we estimate that collisional rates ( per unit time ) are of the same order of magnitude as radiative rates for @xmath348 at @xmath349 , @xmath350 at @xmath351 , @xmath352 at @xmath353 , and @xmath354 at @xmath355 . in other words , as the primordial gas cools , collisions come to only influence the highest h energy levels , which contain the least bound electrons .
this occurs because of the exponential decrease in the free - electron density @xmath356 in the early stages of recombination , which drives down collision rates accordingly . near @xmath349 and shortly thereafter , radiative rates alone are high enough to keep the excited states in @xmath34-equilibrium .
collisions thus have little effect on @xmath24 at early times .
there may , however , be a window at some intermediate redshift , when collision rates are still relatively high , but departures from @xmath34-equilibrium are large enough to warrant including collisions in the recombination model .
a full calculation is necessary to understand the actual effect . a final answer on the effect of resolving @xmath34 sublevels on both the recombination history @xmath24 and the recombination spectrum awaits a robust theoretical calculation of the relevant collisional rates .
this is an area of future investigation . at fixed @xmath4 and @xmath6 , as computed by recsparse.,width=312 ] we may also compare the total population of the @xmath357 energy level to values in boltzmann equilibrium with @xmath33 : @xmath358 the recombination rate to states with @xmath359 is greater than the downward cascade rate , creating a bottleneck to depopulating these states .
this bottleneck causes an over - population of the excited states compared to the equilibrium values of eq .
( [ boldef ] ) , as shown in fig .
[ res_state_gas_nneq_boltz ] .
the ratio @xmath360 is @xmath361 at early times but grows as high as @xmath362 by @xmath363 .
the ratio approaches a constant at high @xmath4 , as energy levels get closer to the continuum and the energy differences between successive levels shrink .
relative to @xmath33 , excited states are over - populated , but there is no population inversion or cosmic maser .
excited states are still less populated than the @xmath33 energy level , just not as dramatically as they would be if eq .
( [ boldef ] ) held . among highly excited states , some pairs of levels do exhibit population inversion . for effective maser action , inversion must occur between pairs of radiatively connected levels , and the coherence of the radiation field must not be destroyed along the line of sight .
this effect will be explored in detail in future work . in extremely dense structure - forming regions ,
more dramatic population inversion may result and lead to local masing ; if these masers were observed , they could offer interesting new probes of structure formation near @xmath364 as well as the physics of reionization @xcite .
recombination becomes inefficient at late times ; i.e. , the recombination time scale @xmath365^{-1}$ ] becomes longer than the age of the universe .
saha equilibrium expressions for @xmath159 and @xmath366 fail dramatically at late times .
the free - electron fraction @xmath159 freezes out and is higher than the saha equilibrium value , and thus @xmath366 is lower than the saha equilibrium value .
excited states are overpopulated relative to the ground state , but still not enough to be in saha equilibrium with the continuum .
the tower of excited states is thus also underpopulated relative to saha equilibrium , as shown in figs .
[ res_state_gas_nneq_saha ] and [ res_state_gas_nneq_sahaz ] .
lower energy levels fall out of saha equilibrium faster than higher energy levels .
higher energy levels are closest to saha equilibrium , but at late times ( @xmath367 ) , even the population of the @xmath368 level is nearly @xmath369 below its saha equilibrium value . modeling the effect of states with @xmath53 may require saha equilibrium abundances to hold in the regime past the cutoff . to this end , it is important to properly model atomic collisions ( which would push atoms towards saha equilibrium at a lower transitional value of @xmath6 ) , and apply even greater computational resources to obtain @xmath24 for even higher @xmath6 .
shell compared to its population in boltzmann equilibrium with @xmath33 , as computed by recsparse at a variety of times through the recombination process.,width=312 ] shell compared to the saha equilibrium population , as computed by recsparse at a variety of times through the recombination process .
, width=312 ] values as an explicit function of cosmological redshift @xmath344.,width=312 ] to explore the relative convergence of @xmath24 over a wide logarithmic range of @xmath6 values , we computed @xmath24 for @xmath370 , and @xmath50 .
we define a relative error : @xmath371 here @xmath372 is the @xmath373 @xmath6 value .
we show the resulting recombination histories and @xmath374 in fig .
[ hist ] . as @xmath6 increases
, the larger number of pathways to the ground state makes recombination more efficient , decreasing @xmath375 and making @xmath374 positive . the relative error @xmath374 shrinks with @xmath6 , indicating that relative convergence is taking place , as demonstrated in fig .
note , however , that the relative error may not be a good proxy for the absolute error .
suppose that the absolute error is given by @xmath376 , where @xmath377 , for some normalization @xmath378 and power - law index @xmath379 .
then it is easy to show that for @xmath380 , @xmath381 . in other words
, the relative error will underestimate the absolute error . to demonstrate absolute convergence
, one should demonstrate that the physics neglected by ignoring transitions to @xmath53 does not cause large changes in @xmath24 .
we also calculated recombination histories for @xmath382 , and @xmath383 .
we may also assess the effect of the computed changes in @xmath24 on the cmb @xmath21 s .
to this end , we replace the usual table generated and used in the recfast module of cmbfast with a table of our own output for different @xmath6 values , smoothly stitching our history onto the usual recfast history at the boundaries @xmath282 and @xmath384 .
we tried a variety of smoothing schemes including no smoothing at all , and determined that the resulting error was at most @xmath369 the change already induced by varying @xmath6 .
the choice of smoothing scheme is thus a correction to a correction " and does not alter the conclusions of our analysis . in particular , the number of sigmas at which power spectra corrected and uncorrected for higher-@xmath4 levels can be distinguished will change by at most @xmath369 of itself as a result of changing the smoothing scheme .
the statistical significance of higher-@xmath4 shells will thus be essentially unchanged by the choice of smoothing scheme .
the results for temperature and e - mode polarization anisotropy power spectra ( @xmath18 and @xmath19 ) are shown in figs .
[ tt ] and [ ee ] , respectively . here
we also define a relative error : @xmath385 here xx denotes the tt or ee label of the power spectrum under consideration .
the relative error @xmath386 is always positive , indicating that increasing @xmath6 also increases @xmath387 , as shown in figs .
[ tt ] and [ ee ] .
the common ( tt and ee ) origin for this effect is clear from fig .
higher @xmath6 makes recombination more efficient , driving down the freeze - out value of @xmath24 and the residual optical depth @xmath12 , leading to the high-@xmath34 plateaus seen in fig .
[ tt ] and [ ee ] . as a result ,
the smearing out of primary cmb anisotropies by relic free electrons , @xmath23 @xcite , is less dramatic when @xmath6 is increased .
the relative error @xmath386 shrinks with increasing @xmath6 . taken as a proxy for the absolute error ,
@xmath386 may be compared to a crude ( cosmic variance ) estimate of the required accuracy of @xmath388 predictions in the damping tail : @xmath389 here @xmath390 is the fraction of the sky covered by a cmb experiment . for @xmath391 , results are shown in figs .
[ tt ] and [ ee ] and we see that only for @xmath5 does the relative error shrink to a level comparable with the cosmic variance .
the ultimate aim is for _ the total _ correction from recombination physics to be less than statistical errors , so any individual contribution such as the truncation error at @xmath6 should be @xmath392 . in any case , collisions must be properly included to show absolute convergence , and so this should be a key focus of future work on highly excited states in hydrogen recombination . to more
realistically assess the importance of high-@xmath4 states , @xmath393 should be compared with a realistic error estimate for _
planck_. different values of redshift @xmath344 .
errors shrink with @xmath6 , indicating relative convergence .
note , however , that this figure gives no scale for the absolute error.,width=312 ] computed using cmbfast , modified to include successively more accurate recsparse recombination histories .
pairs of @xmath6 values used for the comparison are indicated in the legend . @xmath18
increases with @xmath6 , as discussed in sec .
[ resultsn ] .
the correction shrinks with increasing @xmath6 .
the long dashed line indicates the cosmic variance target for @xmath394 , as discussed in the text.,width=312 ] computed using cmbfast , modified to include successively more accurate recsparse recombination histories .
pairs of @xmath6 values used for the comparison are indicated in the legend of fig .
[ tt ] . @xmath19
increases with @xmath6 , as discussed in sec .
[ resultsn ] .
the correction shrinks with increasing @xmath6 .
the long dashed line indicates the cosmic variance target for @xmath394 , as discussed in the text.,width=312 ] as a test of the importance of the modified recombination history for _ planck _ , we have compared our corrections to the power spectrum @xmath395 with the forecast _ planck _ error bars .
the comparison is done by means of the statistic @xmath396 where @xmath397 is the fisher matrix for the cmb power spectrum .
for the temperature - only case , @xmath36 ranges from 2 to @xmath398 and hence @xmath399 is an @xmath400 matrix ; when polarization is included , @xmath399 expands to a @xmath401 matrix incorporating tt , ee , and te spectra .
the @xmath402 statistic is the number of sigmas at which the corrected and uncorrected power spectra could be distinguished assuming perfect knowledge of the cosmological parameters , and hence represents the largest possible bias ( in sigmas ) on any combination of cosmological parameters in any fit that incorporates the cmb @xcite .
we use the forecast noise and beam curves for _ planck _ data 70 ghz ( low - frequency instrument ) and 100 and 143 ghz ( high - frequency instrument ) channels in the blue book @xcite , and assume a usable sky fraction of @xmath403 .
the computation considering the difference between the @xmath404 and 250 curves gives a @xmath402 value of @xmath405 .
however , the actual error in the @xmath10 calculation is somewhat greater because even the @xmath5 calculation is not completely converged .
if the error in the @xmath406s scales as @xmath407 and has a shape that varies slowly with @xmath6 , then our value of @xmath402 should be increased by a factor of @xmath408^{-1}$ ] ; for @xmath409 ( as suggested by fig .
[ conv ] ) this is 1.39 .
thus if the power - law extrapolation is to be trusted there is a @xmath410 error ( @xmath411 ) in the cmb power spectrum if one restricts attention to @xmath10 , and a @xmath412 times smaller error ( @xmath413 ) at @xmath5 .
a similar comparison between @xmath7 and 250 implies an error of @xmath414 at @xmath7 .
this suggests that in the _ purely radiative _ problem the cmb power spectrum is converged ( in the sense that our remaining errors are small compared to projected _ planck _ errors ) at @xmath415 ; however this issue will have to be reconsidered in future work when collisions are included . using the treatment of sec .
[ quadtheorysec ] and an integration stepsize fine enough to obtain a fractional accuracy of @xmath416 in @xmath159 , we compute the effect of e2 quadrupole transitions on cosmological hydrogen recombination for several values of @xmath6 .
we can parametrize this effect using @xmath417 and @xmath418 note that unlike the case of varying @xmath6 , these are the absolute errors induced by ignoring e2 transitions .
the results are shown in fig .
[ quad_hist ] .
the maximum effect of e2 transitions occurs at @xmath419 with a fractional enhancement of @xmath420 , and the calculation seems well converged by @xmath225 .
corrections due to higher excited states would be a correction to a correction , and so we ignore them .
although the correction from e2 transitions is small , it extends over a broad epoch at late times after reaching its maximum . to determine if this could affect cmb anisotropies in an observable way , we modify and run cmbfast @xcite using recombination histories computed with / without e2 transitions .
we incorporated recsparse recombination histories including e2 transitions into cmbfast by applying the same method employed in sec .
[ resultsn ] . running the recombination histories including e2 quadrupole transitions through cmbfast gives a maximum change @xmath421 in both temperature and polarization , negligible compared to cosmic variance .
thus e2 transitions in hydrogen are negligible for cmb applications . .
the net effect is always to speed up recombination.,width=312 ]
we have developed a new recombination code , recsparse , optimized for tracking the populations of many energy shells of the hydrogen atom while resolving angular momentum sublevels .
the code runs more quickly than would be anticipated using simple scaling arguments , which would yield the the scaling @xmath289 . using recsparse
, we find empirically that for the range of @xmath6 values used , computation time scales as @xmath46 , where @xmath47 . with this code
, we have computed cosmological hydrogen recombination histories for a series of @xmath6 values going as high as @xmath5 and explored the highly nonequilibrium state of the resulting atomic hydrogen .
the resulting correction @xmath422 satisfies @xmath423 for @xmath424 when @xmath5 and converges with @xmath425 .
the correction to the @xmath21 s becomes of order the cosmic variance when @xmath5 . in light of realistic error
estimates for _ planck _ , the resulting cmb anisotropy spectra @xmath387 are converged to @xmath51 at fisher - matrix level for @xmath404 in the purely radiative case , assuming error extrapolations may be trusted . to definitively answer the question of absolute convergence
, collisions must be included to speed the approach to saha equilibrium at high @xmath4 , allowing a conclusive treatment of states beyond the truncation limit , with @xmath53 .
future work should also properly account for the overlap of the lyman resonance line series at high @xmath4 .
it will also be interesting to determine if there is coherent stimulated emission between excited states , given its relevance for the detectability of faint cmb spectral distortions from the epoch of recombination .
finally , the sparse - matrix methods applied here or similar techniques could be profitably applied in the development of fast recombination codes for cmb data analysis , even at early times in recombination , when only lower values of @xmath6 are relevant .
the authors acknowledge useful conversations with y. ali - hamoud , n. bode , a. pullen , t. l. smith , j. chluba , j. a. rubio - martn , and the participants of the july 2009 paris workshop on cosmological recombination .
is supported by the dan david foundation and the gordon and betty moore foundation .
is supported by doe de - fg03 - 92-er40701 , the national science foundation under contract no .
ast-0807337 , and the alfred p. sloan foundation .
the development of laser spectroscopy of high-@xmath4 states in hydrogen and other atoms , along with the study of nonlinear and multiphoton ionization , required the computation of dipole radial matrix elements for high and even fractional quantum numbers in a coulomb or perturbed coulomb potential @xcite .
until adequate algorithms for these computations were ultimately developed , the wentzel , kramers , brillouin , and jeffreys ( wkbj ) semiclassical approximation ( quite accurate for @xmath426 ) @xcite proved a useful tool for estimating @xmath427 . at high
@xmath4 , radial wave functions in the coulomb potential have a large number of nodes and thus a short wavelength @xmath428 . for the wkb approximation to be valid ,
it is necessary that @xmath429 . because of the large number of nodes in the coulomb wave functions at high @xmath4
, the wkb approximation is ideally suited to estimating matrix elements for transitions between high @xmath4 .
[ appendix_wkb ] in the classically allowed region , the nonrelativistic wkb radial wave function for a hydrogen atom is @xmath430}\end{aligned}\ ] ] with @xmath431^{1/2},\label{wkb_bb_wfb}\end{aligned}\ ] ] where the inner classical turning point @xmath432 is a solution of the equation @xmath433 . substituting eq .
( [ wkb_bb_wfb ] ) into eq .
( [ mel_bb ] ) for the dipole matrix element , and making several additional approximations , the following expression is obtained if @xmath434 and @xmath435 @xcite : @xmath436\label{bb_low_wkb } .\end{aligned}\ ] ] with @xmath437 , @xmath438 , @xmath439 , @xmath440 , and @xmath441 here @xmath79 is the eccentricity of a keplerian orbit with the quantum numbers @xmath442 and @xmath443 , and @xmath444 is a bessel function of the first kind .
these estimates agree with matrix elements computed using eq .
( [ hga ] ) to a precision of @xmath445@xmath446 ; the agreement worsens as @xmath447 . if @xmath448 and @xmath449 , then @xcite @xmath450 with @xmath451 . here
@xmath452 is a modified bessel function of the second kind .
these estimates agree with matrix elements computed using eq .
( [ hga ] ) to a precision of @xmath453@xmath454 ; the agreement worsens as @xmath227 shrinks , at which point eq .
( [ bb_low_wkb ] ) becomes more accurate .
a wkb estimate of bound - free matrix elements is obtained by making the substitution @xmath455 in eq .
( [ wkbfar ] ) @xcite .
the resulting estimate is reasonable if @xmath456 and agrees with matrix elements computed using eq .
( [ befmel ] ) to a precision of @xmath446 .
this analysis confirms that the high @xmath4 and @xmath34 values under consideration do not afflict our evaluation of eqs .
( [ hga ] ) or ( [ befmel ] ) with any instability that would throw computed rates off by orders of magnitude . | calculations of cosmological hydrogen recombination are vital for the extraction of cosmological parameters from cosmic microwave background ( cmb ) observations , and for imposing constraints to inflation and reionization .
the _ planck _
mission and future experiments will make high precision measurements of cmb anisotropies at angular scales as small as @xmath0 , necessitating a calculation of recombination with fractional accuracy of @xmath1 .
recent work on recombination includes two - photon transitions from high excitation states and many radiative transfer effects .
modern recombination calculations separately follow angular momentum sublevels of the hydrogen atom to accurately treat nonequilibrium effects at late times ( @xmath2 ) .
the inclusion of extremely high - n ( @xmath3 ) states of hydrogen is then computationally challenging , preventing until now a determination of the maximum @xmath4 needed to predict cmb anisotropy spectra with sufficient accuracy for _ planck_. here , results from a new multi - level - atom code ( recsparse ) are presented .
for the first time , ` forbidden ' quadrupole transitions of hydrogen are included , but shown to be negligible .
recsparse is designed to quickly calculate recombination histories including extremely high-@xmath4 states in hydrogen .
histories for a sequence of values as high as @xmath5 are computed , keeping track of all angular momentum sublevels and energy shells of the hydrogen atom separately .
use of an insufficiently high @xmath6 value ( e.g. , @xmath7 ) leads to errors ( e.g. , @xmath8 for _ planck _ ) in the predicted cmb power spectrum . extrapolating errors ,
the resulting cmb anisotropy spectra are converged to @xmath9 at fisher - matrix level for @xmath10 , in the purely radiative case . |
the dark cloud l1287 contains the galactic nebulae gn 00.34.0 ( associated with the young f - type star rno 1 ) and gn 00.33.9 which harbors the iras point source iras 00338 + 6312 ( @xmath0 pc , * ? ? ?
two young stars , rno 1b and rno 1c , lie slightly south - west of the catalog position of the iras source . those objects show properties of fu orionis objects ( fuors ) : ( 1 ) @xcite found that rno 1b brightened by at least 3 mag over a period of 12 years and that it showed a variable and blueshifted p cygni profile in h@xmath1 and additional broad and double peaked absorption lines in its optical spectrum ; ( 2 ) @xcite obtained near infrared spectroscopic data for both objects and found the fuor typical strong 2.3@xmath2 m co absorption bands . as the error ellipse of
the iras source includes both fuors it was long debated whether there is still another deeply embedded source close to the fuor objects or whether iras just measured the integrated flux from these two objects . from high resolution near
infrared ( nir ) polarimetric maps @xcite concluded that an additional embedded object should be present close to the location of the iras source .
this idea was supported by the discovery of a 3.6 cm continuum peak @xcite and an h@xmath3o maser @xcite , both nearly coinciding with the iras position .
a bipolar outflow in the region was found by @xcite and later confirmed by @xcite . as the positions of the iras source and the fuors line up along the outflow axis it has long been uncertain which source was driving the outflow . from interferometric observations in cs @xcite suggested that the iras source was most likely the driving source .
recently , @xcite mapped the outflow in co and came to the same conclusion .
however , @xcite presented ( sub)millimeter observations favoring rno 1c as the outflow driving source .
in this paper we present mid infrared ( mir ) imaging and spectroscopy data taken with the _ irac _ and _ irs _ instruments onboard the _
spitzer space telescope_. for the first time we resolve the mir point source associated with the iras source and find additional , partly deeply embedded , objects .
the two _ irs _ spectra probe the composition of the dense circumstellar environment in the vicinity of rno 1b/1c and bear additional traces of shocked h@xmath3 gas .
all data were part of the gto program by r. gehrz and publicly available from the _ spitzer _ data archive .
an overview of the observational setup and the datasets is provided in table [ journal ] .
the _ irac _
images were obtained in sub
array mode leading to an effective field of view of 40@xmath4 centered on the position given in table [ journal ] .
the spectra taken with _
irs _ cover the wavelength range @xmath5 @xmath2 m .
overplotting the spectral slits of _ irs _ on the 2mass ks - filter image reveals that apparently rno 1b and rno 1c were not centered within the slits ( fig . [ slits ] ) .
the short low resolution spectrum ( @xmath6 @xmath2 m , r @xmath7 ) and the short high resolution spectrum ( @xmath8 @xmath2 m , r @xmath9 ) close to rno 1b seems to probe mainly flux coming from in between the two objects .
the spectrum close to rno 1c contains presumably also less flux than expected due to the slight mispointing .
the long high resolution part in either spectrum ( @xmath10 @xmath2 m , r @xmath9 ) includes flux from both components and also additional flux from the iras source as all three objects lie within the slit of the spectrograph .
the _ irac _ images were reduced with the mopex package provided by the _
spitzer science center_. interpolation , outlier detection and co addition of the images were carried out for each filter individually .
the astrometry was refined by comparing the position of detected sources to known 2mass objects .
the photometry was carried out with the daophot package provided within the iraf environment . as there was only a limited number of sources available within the small field covered by the camera we measured the point - spread function ( psf ) of rno 1c and used it as reference to do psf photometry for all the other sources .
following the _ irac _ data handbook 5.1.1 we converted the pixel values from mjysr@xmath11 to dns@xmath11 and computed the corresponding magnitudes via @xmath12 with @xmath13 denoting the flux measured in dns@xmath11 and @xmath14 being the zero point for each filter ) , 18.94 ( 4.5@xmath15 ) , 16.88 ( 5.8@xmath15 ) , 17.39 ( 8@xmath15 ) . ] .
for the initial psf fit we used a psf size of 2 pixel and applied aperture corrections as described in the data handbook to obtain the final magnitudes .
our final _
irs _ spectra are based on the intermediate droopres ( for the low resolution data ) and rsc ( for the high resolution data ) products processed through the s13.2.0 version of the _ spitzer _ data pipeline .
partially derived from the smart software package @xcite these intermediate data products were further processed using spectral extraction tools developed by the `` formation and evolution of planetary systems '' ( feps ) _ spitzer _ science legacy team ( see , explanatory supplement v. 3.0 within the feps data deliveries . for the short low
resolution observations , the spectra where extracted using a 6.0 pixel fixed - width aperture in the spatial dimension resulting in a 39.96 arcsec@xmath16 extraction aperture on the sky .
the background was subtracted using associated pairs of imaged spectra from the two nodded positions along the slit .
this process also subtracts stray light contamination from the peak - up apertures and adjusts pixels with anomalous dark current relative to the reference dark frames .
pixels flagged by the _ spitzer _
data pipeline as being `` bad '' were replaced with a value interpolated from an 8 pixel perimeter surrounding the errant pixel .
the high resolution spectra were extracted with the full aperture size ( 53.11 arcsec@xmath16 for the short high and 247.53 arcsec@xmath16 for the long high module ) .
the sky contribution was estimated by fitting a continuum to the data . both , the low and high resolution spectra , are calibrated using a spectral response function derived from irs spectra appropriate for the two nod positions in the slit at which we extract the specta , and cohen or marcs stellar models for a suite of calibrators provided by the
_ spitzersciencecentre_. the multiple orders in our spectra match to within 10% .
these small flux offsets are likely related to small pointing offsets .
we computed correction factors for possible flux loss due to telescope mispointing based on the psf of the _ irs _ instrument .
however , accurate results can only be obtained if the flux is dominated by one source .
the offsets also result in low level fringing at wavlengths longer than 20 micron in the low resolution spectra and at all wavelengths in the high resolution spectra .
we removed these fringes using the irsfringe package developed by f. lahuis .
the relative errors between spectral points within one order are dominated by the noise on each individual point and not by the calibration . based on our experience with _ spitzer irs _ data
we estimate a relative flux calibration error across a spectral order of @xmath17 % and an absolute calibration error between orders / modules of @xmath18 % .
these values , however , are based on point source observations with accurate telescope pointing .
the data presented here are more complicated as at least part of the emission is moderately extended , multiple sources contribute to the observed fluxes , each contribution is wavelength dependent , and the telescope slits were not centered on the main objects .
thus , an accurate flux calibration between the different modules is difficult and apparent offsets are further discussed in section 3.2 .
within the limited field of view of the _ irac _ subarray mode we identified 8 sources that were detected in at least three of the four _ irac _ bands .
figure [ fig1 ] provides a comparison between the rno 1b/1c region as seen in the 2mass ks - filter and the 5.8@xmath2 m _ irac _ filter . for the first time the mir point source related to iras 00338 + 6312 is detected in the _ irac _ band .
additional fainter objects are also present , some of which are detected for the first time .
[ fig9 ] shows a color composite image of the rno 1b/1c complex based on three _ irac _ filters .
table [ irac_fluxes ] lists all objects that were identified in at least three _ irac _ bands and summarizes the derived fluxes .
the errors are based on the results from the psf - photometry . since we used the psf of rno 1c as reference
psf the corresponding errors are relatively small compared to the other objects .
iras 00338 + 6312 and rno 1 g were not detected in the shortest _ irac _ band at 3.6@xmath2 m .
the 2mass image in fig .
[ fig1 ] shows that at least the iras source is deeply embedded in a dense dusty environment explaining the non - detection at this wavelength .
[ fig2 ] shows two color color plots based on the four _ irac _ bands for all objects listed in table [ irac_fluxes ] .
following @xcite who analyzed a large sample of pre - main sequence stars in the taurus star - forming region we use these plots to classify the different objects .
the colors of the reddest objects ( iras 00338 + 6312 and rno 1 g ) are consistent with very young protostars .
the colors of the newly discovered objects rno1 irac1 and rno1 irac3 are consistent with class 0/i systems although it has to be confirmed that these are indeed nearby objects and not highly reddened background sources . while the colors of the fuor object rno 1c fit in the region of class ii objects from @xcite in both plots ( fig .
[ fig2 ] ) , rno 1b fits only in one color - color diagram into this regime . in the right plot of fig .
[ fig2 ] rno 1b is too red for a class ii source in the [ 4.5]-[5.8 ] color , but not red enough in the [ 3.6]-[4.5 ] color to be a class 0/i object . finally ,
rno1 irac 2 and rno 1f , which both show up in the 2mass ks - band image , can also be interpreted as class ii objects .
however , the colors of rno1 irac2 , but partly also of rno 1f , are possibly altered due to local extinction effects . in figure [ fig3 ]
we show the _ spitzer irs _ spectra without any extinction correction .
as pointed out earlier the spectral slits of the short wavelength modules were apparently not directly centered on the objects rno 1b and rno 1c and thus the measured fluxes can not be attributed to these objects with very high accuracy .
also , there seems to be an offset in the flux in the high resolution regime compared to the low resolution part of the spectrum . to see to which extent this effect is caused by different aperture sizes we plotted the spectra in intensities (
i.e. , flux density per solid angle ) rather then in jansky per wavelength .
it shows , however , that even after this correction significant offsets remain : around 13@xmath2 m the difference between the low resolution spectrum and the short high resolution spectrum amounts to a factor of @xmath19@xmath20 and @xmath19@xmath21 for rno 1b and rno 1c , respectively , with the high resolution part showing higher intensities .
at least for the rno1 b spectrum this offset is larger than normally expected from the calibration accuracy .
we thus believe that part of this offset can be attributed to the different orientations of the spectral slits on the sky probing different regions of the extended emission .
the offset between the short and long wavelength range of the high resolution data corresponds to factors of @xmath19@xmath22 and @xmath19@xmath23 at 20@xmath2 m for rno 1b and rno 1c , respectively .
also here the short high resolution spectrum shows higher intensities .
this can be explained by the approximately five times smaller aperture in the short high resolution module .
although this aperture is probing a significantly smaller regions on the sky these regions do presumably contribute in total to flux in the above mentioned wavelength regime .
the large aperture of the long high resolution module on the other hand is certainly also probing regions without any significant flux ( see also , fig . [ slits ] )
. at 6@xmath2 m both spectra show clear water ice absorption bands . at slightly longer wavelength ( @xmath246.85@xmath2 m )
additional absorption possibly arising from ch@xmath25oh ice is present @xcite .
the 10@xmath2 m silicate band is also seen in absorption in both spectra ( stronger close to rno 1c ) , although the shape of the feature differs from the absorption feature caused by typical interstellar medium silicate grains . to analyze the differences in more detail we fitted a continuum to the 10@xmath2 m region of the spectra and computed the optical depth ( fig [ 10mu_absorption ] ) . in both cases
the peak of the absorption is slightly shifted towards shorter wavelengths indicating a non - ism like dust composition .
additional absorption ( rno 1c ) or possibly additional dust emission on top of the absorption feature ( rno 1b ) is seen at longer wavelengths . at 15.2@xmath2 m
co@xmath3 ice is creating a prominent absorption band and at @xmath2418@xmath2 m additional silicate absorption seems to be present .
all these features indicate the existence of a dense dusty and icy environment in which the two fuors are embedded . a further analysis of these features also for a larger sample of fuors will be presented in an upcoming paper ( quanz et al . in preparation ) .
in addition to the ice and silicate features , h@xmath3 emission lines from purely rotational quadrupole transitions are present in both spectra ( fig .
[ fig3 ] ) . while in the spectrum close to rno 1b all transitions from s(1 ) to s(7 ) can be identified , the spectrum close to rno 1c shows only the lines from s(1 ) to s(5 ) .
this can be explained with the higher continuum flux close to the s(6 ) and s(7 ) line in the latter spectrum and with apparently lower excitation temperatures ( see below ) .
the lowest transition s(0 ) near 28.22@xmath2 m is not detected in either spectrum .
this is due to the strongly rising continuum at longer wavelengths .
here , the spectral slit contained flux from both rno objects and also from the deeply embedded iras source so that the continuum emission and the related flux errors completely dominate a possible weak emission line .
a detailed analysis is provided in the appendix . keeping in mind the existence of a molecular outflow that is powered by
iras 00338 + 6312 and directly oriented in the direction of rno 1c and rno 1b ( e.g. , * ? ? ?
* ) , the detection of h@xmath3 emission lines in both spectra hints towards shock induced emission related to the outflow .
the spectra consequently bear information about the circumstellar material close to rno 1b/1c and about the outflow coming from the iras source .
since we observe h@xmath3 lines even in the 10@xmath2 m silicate absorption bands the outflow appears to lie in front of the dusty environment as otherwise the high extinction ( a@xmath26 mag and a@xmath27 mag for rno 1b and rno1c , respectively , * ? ? ?
* ) would have prevented a detection .
the measurement of the relative strengths of multiple h@xmath3 lines allows an analysis of the physical conditions of the shocked material . for this
we assume local thermal equilibrium ( lte ) and optically thin line emission which is supported by the low einstein coefficients of the involved quadrupole transitions ( table [ hydrogen_lines ] ) .
following @xcite the column density of an upper energy level @xmath28 is then given by @xmath29 where @xmath30 denotes the observed line intensity in erg s@xmath11 @xmath31 sr@xmath11 , @xmath32 is the energy difference of the two states involved in the transition , @xmath33 is the einstein coefficient for the transition and @xmath34 is the optical depth at the observed wavelength . in fig .
[ fig4 ] we show gaussian fits to the observed emission lines in the spectrum close to rno 1b and fig .
[ fig5 ] presents corresponding fits to the lines detected close rno 1c . before fitting the emission lines we subtracted the underlying continuum which was fitted with a second order polynomial
the finally derived integrated line fluxes and column densities are shown in table [ hydrogen_lines ] .
we corrected the line fluxes for extinction effects using the results of @xcite and assuming @xmath35 mag .
this value represents the extinction found by @xcite towards the nearby star rno 1 and should be a better estimate for the line - of - sight extinction than the above mentioned extinction values towards rno 1b and rno 1c . in any case
the influence of the extinction on the derived shock temperatures and column densities ( see below ) is negligible .
the 1sigma errors in the line flux , the line intensity and the column density ( table [ hydrogen_lines ] ) were derived from generating 500 spectra where we added a gaussian noise to each measured flux point based on the initial individual 1sigma uncertainty . from these spectra
we computed the mean values and related errors of the listed parameters . especially at the short wavelength end
it shows that not all lines were detected with a 3sigma confidence level .
however , since most lines were convincingly measured we decided to keep also the tentative detections in our analyses .
since we assume lte the rotational energy levels shall be populated following boltzmann statistics with a unique temperature for several lines .
the involved temperatures @xmath36 of the shocked material can be derived from a so called `` rotational diagram '' . by using the results from eq .
[ eq1 ] , plotting the logarithm of @xmath37 against @xmath38 ( i.e. the formal temperature corresponding to the absolute upper energy level of the respective rotational transition ) and fitting a straight line to the data one finds that the slope of the straight line is proportional to @xmath39 .
here , @xmath40 denotes the spin degeneracy of each energy level ( 1 for even j , 3 for odd j ) and @xmath41 is the rotational degeneracy .
these numbers assume that the ortho to para ratio of the involved hydrogen is close to its lte value of 3 at @xmath36 .
a deviation from this lte assumption ( i.e. , ortho to para @xmath42 3 ) would result in a downwards displacement of the data points with an odd j - number ( ortho h@xmath3 ) relative to the points with even j - numbers ( para h@xmath3 ) and thus create a `` zig zag '' pattern in the rotational plot ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . in fig .
[ fig6 ] , [ fig7 ] , and [ fig8 ] we show rotational plots for the measured lines near rno 1b and rno 1c . in all plots
the errors for the data points denote the 1sigma uncertainty in the measured column density . the rotational diagram of rno 1b in fig .
[ fig6 ] shows a clear curvature in the data points .
this departure from a single straight line is indicative of several temperature components in the shocked material .
we thus fitted the data with two superimposed temperature regimes ( hot and warm ) that rather represent the minimum and maximum of the involved temperatures than any distinct intermediate value .
the dash - dotted line fits the high energy regime and represents the hot component with a temperature of 2991@xmath43596 k. the dotted line corresponds to a temperature of 1071@xmath43121 k designating the warm component .
the uncertainties in the temperature denote the 1sigma confidence level of the fits .
as the derived temperature is extremely sensitive to the slope of the fitted straight line the corresponding errors are rather large .
in addition to the temperatures one can also derive the total h@xmath3 column density of the observed shock within the given aperture . from the y - intersects of the fits we estimate the hot component to have @xmath44(h@xmath3)@xmath45 and for the warm component we find @xmath46(h@xmath3)@xmath47 . for the column densities derived from the spectrum close to rno 1c we plotted two rotational diagrams shown in fig . [ fig7 ] and [ fig8 ] .
first , we fitted the data with a single temperature component of 1754@xmath43321 k. however , the shocked material is apparently not in lte as the data points with odd and even j - numbers are difficult to fit simultaneously with a straight line or a curve .
as mentioned above this displacement is indicative of a departure from the lte ortho to para ratio of 3 .
[ fig8 ] shows the same data if one assumes an ortho to para ratio of 1 .
now , a two component fit , similar to that in fig . [ fig6 ] , is possible .
for the hot component we derive a temperature of 2339@xmath43468 k , while the warm component is significantly colder with 466@xmath43264 k. the total h@xmath3 column densities amount to @xmath48(h@xmath3)@xmath49 and @xmath50(h@xmath3)@xmath51 , respectively .
table [ components ] summarizes the temperatures and total column densities of the different components for both spectra .
the _ irac _ images reveal for the first time the embedded mir point source associated with iras 00338 + 6312 .
however , already @xcite concluded from polarimetric observations that an additional source should be present close to the iras point source position .
furthermore , many authors analyzing the bipolar outflow related to this region concluded that a deeply embedded object was most likely the driving source @xcite instead of one of the rno objects @xcite .
previous ground based mir observations @xcite were possibly not sensitive enough to detect iras 00338 + 6312 .
the observed properties of iras 00338 + 6312 convincingly classify the object as a young protostar and support the idea that it is indeed driving the molecular outflow . from their co measurements
@xcite derived a total mass for this bipolar outflow of 1.4 m@xmath52 , indicating that iras 00338 + 6312 is an intermediate mass protostar .
the detection of the iras source solves also the problem that up to now the seds of rno 1b and rno 1c were unusually steep longward of 25@xmath2 m for fuor objects @xcite .
the images presented here have sufficient sensitivity and spatial resolution to show that the newly revealed source associated with iras 00338 + 6312 is most probably responsible for this flux excess .
in addition to the iras object @xcite also suspected another embedded source which we identify with rno 1 g .
furthermore , the detection of rno 1f @xcite , which was initially assumed to be a density enhancement rather than an independent self - luminous source @xcite , is confirmed with our _ irac _ images . only rno 1d , which was thought to be located between rno 1b and rno 1c @xcite , does not show up in our images .
this source was , however , also not detected by @xcite and only tentatively seen in the data from @xcite .
it is not clear whether all detected objects are physically confined to the rno 1b/1c region .
especially the membership of the newly discovered objects rno1 irac1 , rno1 irac2 , and rno1 irac3 needs to be confirmed .
however , the _ irac _ observations suggest that rno 1b and rno 1c belong to a small cluster of ( partly very ) young objects . to our knowledge , although some fuors are known or suspected to be in binary or multiple systems ( e.g. , * ? ? ?
* ) , the existence of fuors in a cluster - like environment is a so far unique finding . only the fuor candidate v1184 tau ( cb34 ) belongs also to a small cluster @xcite but its classification as fuor is far less certain than for the rno objects @xcite . in any case
the still deeply embedded protostars in the close vicinity of rno 1b/1c put strong constraints on the age of the objects if one assumes coeval evolution .
the magnitudes we derive for the two fuor objects at 3.6@xmath2 m can be compared to previous ground based measurements at 3.8@xmath2 m by @xcite and @xcite
. however , one certainly has to keep in mind that the _ irac _ filter is not only different in terms of central wavelength but also the spectral width differs from that of the ground based instruments . while @xcite found @xmath53mag and @xmath54mag for rno 1b and rno 1c , respectively , @xcite observed rno 1c to be slightly brighter than rno 1b ( @xmath55mag vs. @xmath56mag ) .
also in our measurements rno 1c seems to be the brighter component .
however , both objects appear to be fainter in our observations compared to @xcite . apart from the differences in the filter properties intrinsic variations in the luminosity of the fuors and/or changing local extinction effects can account for the apparent variability in these objects .
the detection of h@xmath3 rotational line emission can in general be attributed to collisional excitation from c- or j - type shocks ( e.g. , * ? ? ?
* ; * ? ? ?
c - type shocks are magnetic and less violent towards molecules in comparison to the mostly hydrodynamic j - type shocks that can dissociate h@xmath3 molecules even at lower shock velocities . from our rotational diagrams we derived shock temperatures that are similar to those found in the outflow from cepheus a @xcite . these authors could fit a two component c - shock model to their data and derived shock velocities between 25 and 30 kms@xmath11 for a cold and a hot component , respectively . by comparing our results directly to theoretical shock models
we find that the c - shock model 1 from @xcite can explain both the hot and warm component in the spectrum close to rno 1b .
considering our derived temperatures as an upper and lower limit for the shocked material , shock velocities between 15 and 30 kms@xmath11 are required .
however , model 1 of @xcite predicts on ortho to para ratio of @xmath192 at the low velocity / low temperature limit which is not directly evident from our observations .
more recent models by @xcite predict lower shock temperatures in c - type shocks for the same shock velocities in comparison to @xcite .
following their models , our hot component in fig.[fig6 ] requires velocities exceeding 40 kms@xmath11 .
slightly slower velocities ( @xmath1935 kms@xmath11 ) are already needed to explain the hot temperature component in the shock close to rno 1c ( fig .
[ fig8 ] ) . the warm component in this diagram represents shock velocities between 15 kms@xmath11 and 20 kms@xmath11 .
although it is clear that some uncertainties between the observations and theory still remain , c - type shocks seem to provide a solid explanation for the observed shock temperatures . the measurement of an ortho to para ratio smaller than 3 in fig .
[ fig8 ] implies that despite the high temperatures the gas has not yet reached equilibrium between the ortho and para states .
thus , the observed ratio is the legacy of the temperature history of the gas ( see , e.g. , * ? ? ? * ; * ? ? ? * ) and transient heating by a recently passing shock wave already caused higher temperatures .
as the corresponding region on the sky lies closer to the iras source than the region probed in fig .
[ fig6 ] , the observations suggest that we see the signatures of at least two shock waves : the first shock wave is probed close to rno 1b ( fig .
[ fig6 ] ) where we observe several temperature components that have apparently reached their equilibrium state .
a more recent shock wave is seen close to rno 1c ( fig .
[ fig8 ] ) where the ratio of the line column densities hints also towards high shock velocities and temperatures but the gas is not yet in thermal equilibrium .
our conclusions can be summarized as follows : * we detected and resolve for the first time the mir point source associated with iras 00338 + 6312 which appears to be an embedded intermediate
mass protostar driving the known molecular outflow in the rno 1b/1c region . *
the detection of additional ( partly previously unknown ) point sources suggests that the fuor objects rno 1b/1c belong to a young small stellar cluster . to our knowledge ,
rno 1b/1c are the only well studied and confirmed fuors that apparently belong to a cluster like environment . * all but
two objects were detected in all four _ irac _ bands and their colors are consistent with class 0/i - ii objects .
the two objects that were not detected at 3.6@xmath2 m ( including iras 00338 + 6312 ) are still very deeply embedded protostars .
* having apparently extremely young objects in the direct vicinity of the fuors confirms the suspected young age for this type of objects although their mir colors are consistent with class ii objects .
* the two mir spectra of the region bear clear signs of a dense icy and dusty circumstellar environment as solid state features are seen in absorption . *
the spectra show also h@xmath3 emission lines from purely rotational transitions .
we presume that these lines arise from shocked material within the molecular outflow .
the derived shock temperatures and velocities are in agreement with c - type shock models . *
the observations of the h@xmath3 lines suggest that the outflow lies in front of rno 1b/1c as otherwise the high optical depth towards these objects would have prevented the detection . *
while in one spectrum the gas probed by the h@xmath3 line emission seems to be in lte , the other spectrum shows a deviation from the expected lte ortho to para h@xmath3 ratio .
this indicates the presence of at least two shock waves , the most recent one responsible for the non lte line ratios .
the results presented here advantageously combine space - based mir imaging and spectroscopy .
while the images give insights into the photometric properties of several young embedded objects and also fu orionis type stars the spectra provide information on the dusty and icy circumstellar environment of the young cluster as well as on shocked gas within a molecular outflow . in this context
the existence of fuors in a cluster like environment needs to be pointed out .
it is still debated whether all young low mass stars undergo an fuor phase or whether these objects form a special sub group of ysos ( e.g. , * ? ? ?
our results suggests that the fuor phenomenon occurs also in young ( small ) stellar clusters making it thus possibly more common than so far expected .
in addition , the co existence of fuors and very young protostars in the same environment strengthens the idea that the fuor phase of ysos is linked to the early stages of the star formation process .
possible future investigations of the rno 1b/1c region should address whether the observed objects do indeed belong to the same region and whether all of them are young stars .
furthermore , a detailed study of the shocked material including high resolution narrow band imaging and spectroscopic mapping ( e.g. , with _ spitzer irs _ ) will provide deeper insights into its extension , physical conditions , and its relation to the molecular outflow .
a complete analysis of the ice and dust features in the observed spectra is currently underway and will be presented in an upcoming paper ( quanz et al . in preparation ) .
quanz kindly acknowledges support from the german _ friedrich - ebert - stiftung_. we thank h. beuther , r. mundt , a. carmona and s. birkmann for useful discussions and conversations .
an anonymous referee provided valuable criticism that helped improve the paper .
this work is based on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
this research has made use of the simbad database , operated at cds , strasbourg , france .
_ facilities : _
in the following we analyze whether the non - detection of the 28.22@xmath2 m in either spectrum is in agreement with the models . the two component models in fig .
[ fig6 ] and [ fig8 ] can be used to predict the column densities of the s(0 ) emission line .
one expects @xmath57\approx 18.3\,\rm{cm}^{-2}\ ] ] @xmath57\approx 18.8\,\rm{cm}^{-2}\ ] ] for the spectrum close to rno 1b and rno 1c , respectively . with @xmath40=1 and @xmath58=5 and applying eq .
( [ eq1 ] ) the following line intensities are derived : @xmath59 @xmath60 the @xmath61-coefficient for the transition is @xmath62s@xmath11 @xcite .
taking into account the aperture size for the long wavelength high resolution module of the spectrograph of 247.53 arcsec@xmath16 ( @xmath63 ) the predicted integrated line fluxes amount to : @xmath64 @xmath65 to estimate the peak of the 28.22@xmath2 m line in jansky we assume that the spectral resolution in the high resolution modules is constant and that the fwhm of the s(0 ) line can be extrapolated from the fwhm of the s(1 ) line at 17.03@xmath2 m .
since we fit a gaussian profile to the emission lines and we want to derive the peak flux of this gaussian , we also have to take into account the relation between the fwhm we measure for the line and the @xmath66 of the profile .
this relation is fwhw=@xmath67 . with the expected fwhm of @xmath68 @xmath69 we find thus corresponding peak fluxes of : @xmath70 @xmath71 these values have to be compared to the measured uncertainties in the spectra .
the mean 1@xmath66 level in the spectral range between 28.0 and 28.4@xmath2 m is , however , @xmath72 @xmath73 so that even in the more favorable case of rno 1c the emission line is not expected to be detected with a confidence level greater than @xmath191.8@xmath74 .
llllcccc spitzer irac & 00:36:45.8 & + 63:28:56 & 5027072 & 3.6 , 4.5 , 5.8 , 8.0@xmath2 m & - & 0.4 / 1 / 4 & 2003 - 12 - 20 + spitzer irs & 00:36:46.34 & + 63:28:53.76 & 6586624 & low resolution & 6 sec / 3 & - & 2004 - 01 - 07 + & 00:36:46.89 & + 63:28:58.44 & & & & + spitzer irs & 00:36:46.34 & + 63:28:53.76 & 6586624 & high resolution & 6 sec / 5 & - & 2004 - 01 - 07 + & 00:36:46.89 & + 63:28:58.44 & & & & clcccccc 1 & rno 1b & 00:36:46.05 & + 63:28:53.29 & [email protected] & [email protected] & [email protected] & [email protected] + 2 & rno 1c & 00:36:46.65 & + 63:28:57.90 & [email protected] & [email protected] & [email protected] & [email protected] + 3 & rno 1f & 00:36:45.74 & + 63:29:04.09 & [email protected] & [email protected] & [email protected] & [email protected] + 4 & rno 1 g & 00:36:47.14 & + 63:28:49.95 & - & [email protected] & [email protected] & [email protected] + 5 & iras
00338 + 6312 & 00:36:47.34 & + 63:29:01.61 & - & [email protected] & [email protected] & [email protected] + 6 & rno1 irac1 & 00:36:48.44 & + 63:28:39.98 & [email protected] & [email protected] & [email protected] & [email protected] + 7 & rno1 irac2 & 00:36:47.90 & + 63:28:36.30 & [email protected] & [email protected] & [email protected] & [email protected] + 8 & rno1 irac3 & 00:36:47.85 & + 63:28:41.23 & [email protected] &
[email protected] & 10.74@xmath43 0.10 & [email protected] + lccccccc + s(1 ) j=3 - 1 & 17.0348 & 1015.08 & 4.76(-10 ) & 53.11 & 7.91(-21)@xmath433.8(-22 ) & 6.34(-5)@xmath433.0(-6 ) & 1.44(19)@xmath436.8(17 ) + s(2 ) j=4 - 2 & 12.2786 & 1681.63 & 2.75(-9 ) & 53.11 & 1.33(-20)@xmath435.2(-22 ) & 1.06(-4)@xmath434.1(-6 ) & 3.00(18)@xmath431.2(17 ) + & & & & 39.96 & 6.75(-21)@xmath431.2(-21 ) & 7.18(-5)@xmath431.3(-5 ) &
2.03(18)@xmath433.7(17 ) + s(3 ) j=5 - 3 & 9.6649 & 2503.73 & 9.83(-9 ) & 39.96 & 2.00(-20)@xmath432.7(-21 ) & 2.13(-4)@xmath432.9(-5 ) & 1.33(18)@xmath431.8(17 ) + s(4 ) j=6 - 4 & 8.0251 & 3474.48 & 2.64(-8 ) & 39.96 & 9.12(-21)@xmath431.9(-21 ) & 9.71(-5)@xmath432.1(-5 ) & 1.87(17)@xmath434.0(16 ) + s(5 ) j=7 - 5 & 6.9095 & 4585.94 & 5.88(-8 ) & 39.96 & 2.13(-20)@xmath432.6(-21 ) & 2.26(-4)@xmath432.8(-5 ) &
1.68(17)@xmath432.1(16 ) + s(6 ) j=8 - 6 & 6.1086 & 5829.66 & 1.14(-7 ) & 39.96 & 7.04(-21)@xmath431.3(-21 ) & 7.50(-5)@xmath431.4(-5 ) & 2.54(16)@xmath434.8(15 ) + s(7 ) j=9 - 7 & 5.5112 & 7196.20 & 2.00(-7 ) & 39.96 & 1.32(-20)@xmath436.1(-21 ) & 1.41(-4)@xmath436.5(-5 ) & 2.46(16)@xmath431.1(16 ) + + + s(1 ) j=3 - 1 & 17.0348 & 1015.08 & 4.76(-10 ) & 53.11 & 3.26(-21)@xmath432.7(-22 ) & 2.61(-5)@xmath432.1(-6 ) & 5.92(18)@xmath434.8(17 ) + s(2 ) j=4 - 2 & 12.2786 & 1681.63 & 2.75(-9 ) & 53.11 & 5.17(-21)@xmath437.7(-22 ) & 4.14(-5)@xmath436.2(-6 ) & 1.17(18)@xmath431.7(17 ) + s(3 ) j=5 - 3 & 9.6649 & 2503.73 & 9.83(-9 ) & 39.96 & 7.52(-21)@xmath432.0(-21 ) & 8.00(-5)@xmath432.1(-5 ) & 4.98(17)@xmath431.3(17 ) + s(4 ) j=6 - 4 & 8.0251 & 3474.48 & 2.64(-8 ) & 39.96 & 1.34(-20)@xmath434.1(-21 ) & 1.43(-4)@xmath434.4(-5 ) & 2.74(17)@xmath438.5(16 ) + s(5 ) j=7 - 5 & 6.9095 & 4585.94 & 5.88(-8 ) & 39.96 & 1.10(-20)@xmath433.3(-21 ) & 1.17(-4)@xmath433.5(-5 ) & 8.69(16)@xmath432.6(16 ) + lcc + hot & 2991@xmath43596 & [email protected] + warm & 1071@xmath43121 & [email protected] + + + hot & 2339@xmath43468 & [email protected] + warm & 466@xmath43264 & [email protected] + + single & 1754@xmath43321 & [email protected] + + m _ irac _ filter ( right ) .
the images are inverted and the contrast is set to make also fainter objects visible .
the _ irac _
image contains more sources and especially the iras source clearly shows up at 5.8@xmath2 m while it seems hidden behind a dust lane in the 2mass image .
the magnitudes of the individual sources are listed in table [ irac_fluxes ] .
, width=642 ] .
the dashed boxes indicate the regions corresponding to class ii objects in @xcite .
the dashed - dotted region in the right plot defines the position of class 0/i sources in @xcite . in the left plot
the class 0/i objects from @xcite have the same [ 5.8]-[8.0 ] colors as the class ii objects but redder [ 3.6]-[4.5 ] colors .
thus , they lie `` above '' the dashed box .
the reddening vectors correspond to @xmath75 mag and are based on a vega - like spectrum and the reddening law given by @xcite . ] | we present _ spitzer irac _ and _ irs _ observations of the dark cloud l1287 .
the mid infrared ( mir ) _ irac _ images show deeply embedded infrared sources in the vicinity of the fu orionis objects rno 1b and rno 1c suggesting their association with a small young stellar cluster .
for the first time we resolve the mir point source associated with iras 00338 + 6312 which is a deeply embedded intermediate
mass protostar driving a known molecular outflow .
the _ irac _
colors of all objects are consistent with young stars ranging from deeply embedded class 0/i sources to class ii objects , part of which appear to be locally reddened .
the two _ irs _ spectra show strong absorption bands by ices and dust particles , confirming that the circumstellar environment around rno 1b/1c has a high optical depth .
additional hydrogen emission lines from pure rotational transitions are superimposed on the spectra .
given the outflow direction , we attribute these emission lines to shocked gas in the molecular outflow powered by iras 00338 + 6312 .
the derived shock temperatures are in agreement with high velocity c - type shocks . |
observations of rotational co transitions are the primary method used to determine molecular masses , particularly in systems outside the milky way . for an expanded discussion of using co to determine masses , i refer the reader to the review by bolatto , wolfire , & leroy ( @xcite ) . here
i briefly introduce some of the concepts and focus on work published since the review .
the relationship between molecular column density or mass and @xmath0co emission is determined through two equations : @xmath1 } & = & { \rm x_{co}\,i_{co}\,[k\,km\,s^{-1 } ] } \\ { \rm m_{mol}\,[m_\odot ] } & = & { \rm \alpha_{co}\,l_{co}\,[k\,km\,s^{-1}\,pc^2].}\end{aligned}\ ] ] in the disk of the milky way the value of is @xmath2 , which in turn implies @xmath3 including a contribution to the mass of 36% due mostly to he according to its cosmological abundance .
the @xmath0co transition is usually very optically thick ( in the milky way disk the typical @xmath0co/@xmath5co ratio is @xmath6 , suggesting @xmath7 ) . for gravitationally - bound clouds , its integrated intensity encodes information about mass through the line width , so the use of these equations requires that co and are coextensive and that the velocity dispersion of the gas is a reflection of its self - gravity .
the peak temperature of the line carries information about the product of the temperature of the gas and its beam - filling fraction .
it is easy to show that for self - gravitating entities will have mild dependencies on density and temperature , such that @xmath8 where @xmath9 is the gas density and @xmath10 is its kinetic temperature ( see bolatto et al .
@xcite for a more detailed discussion ) .
note that since @xmath11 , the relevant @xmath10 is that at the @xmath12 surface .
the fact that co is optically thick makes this method robust to changes in the co abundance and to the details of the excitation ( [ sec : breakdown ] ) , which affect methodologies that rely on optically thin tracers . by comparison , in the optically thin regime @xmath13\,t_{ex}\,\exp(h\nu/(kt_{ex}))$ ] ( a similar equation can be written for any optically thin rotation transition ) .
determining masses using optically thin emission requires accurate knowledge of abundances and excitation temperatures , either of which may change over a very large range ( for example , @xmath14 is a very strong function of density and @xmath10 ) .
using co to estimate molecular masses of entire galaxies requires further assumptions ( dickman , snell , & schloerb @xcite ) : the co emission must arise from an ensemble of self - gravitating clouds with a narrow range of @xmath15 . as long as the velocity dispersion in each cloud in the ensemble is a reflection of its gravity , and clouds
do not shadow each other , it can be shown that the total luminosity will reflect the sum of the cloud masses .
the assumptions behind the standard proportionality between co luminosity and molecular mass break down in three circumnstances : 1 ) at low metallicities , 2 ) in environments where the gas motions do not reflect self - gravity , and 3 ) in places where @xmath15 is very different from the disk of the milky way .
although ( 3 ) may be important in some cases , it is probably not the dominant driver of changes . because @xmath10 depends on the radiation field
mostly driven by star formation and star formation is enhanced at high density , there is likely a large degree of compensation in their ratio .
moreover , photodissociation regions ( pdrs ) are self - regulated to keep an approximately constant temperature for the @xmath12 surface ; increasing the radiation field impinging on a pdr drives its surface temperature higher , but it also pushes the @xmath16 transition further into the cloud and into cooler regions .
realistic simulations incorporating cloud structure and chemistry provide key insights on how is affected by changes in the radiation field and the distribution of column densities , velocity dispersion , and dust extinction ( e.g. , clark & glover @xcite ) .
cloud structure has an impact on the effects of increasing the incident radiation field , but changes little in large , dense , and bound clouds . at low metallicities the hypothesis of coextensivity between h@xmath4 gas and co emission breaks down .
high co abundance requires a minimum extinction ( @xmath17 , weakly dependent on the incident radiation field ) . by comparison ,
most of the atomic hydrogen becomes molecular at a much lower extinction ( @xmath18 ) .
extinction is more difficult to build up at low metallicity , where the dust - to - gas ratio is lower , resulting in a increasing fraction of the h@xmath4 gas not being associated with bright co emission .
physical models that incorporate the relevant physics qualitatively agree on this picture ( maloney & black @xcite ; rllig et al .
@xcite ; wolfire , hollenbach , & mckee @xcite ) , as do numerical calculations incorporating cloud structure and time - dependent chemistry ( glover & mac low @xcite ) . resolved observations of the magellanic clouds further validate this picture ; a strong correlation between co intensity and inferred is seen ( lee et al .
@xcite ) , and global values rapidly increase for decreasing metallicity ( e.g. , jameson et al .
@xcite ) . indeed ,
the co emitting regions of clouds at low metallicity are increasingly smaller for decreasing metallicity , while the fraction of the cloud encompassed by the molecular envelope faint in co increases ( leroy et al .
@xcite , @xcite ) .
this is strikingly demonstrated in the recent observations of individual molecular clouds in the very low metallicity wlm galaxy , which show very small co emitting regions ( rubio et al .
@xcite ) .
thus at low metallicity the amount of co emission per molecule is set by the distribution of column densities in the cloud , mediated by the dust - to - gas ratio .
it is , in fact , fairly insensitive to the gas metallicity itself , but it depends on metallicity through the dust - to - gas ratio .
this highlights the importance of better understanding dust production and destruction mechanisms in galaxies .
it is increasingly clear that there is not a one - to - one correspondence between gas metallicity and dust - to - gas ratio ( e.g. , fisher et al .
@xcite ) , with observations finding a large dispersion in dust - to - gas ratio at fixed metallicity for low metallicity and high specific star formation rate ( remy - ruyer et al .
@xcite ) . understanding how the distribution of cloud column densities is established is also important , since it plays a key role at determining the fraction of the gas that reaches the critical necessary for bright co emission . in very gas - rich galaxies
it is possible to maintain an extended intercloud medium that is molecular and emits in co. this medium experiences the combined gravitational potential of the stellar , gas , and dark matter components , not just its own self - gravity . in these circumnstances
the hypothesis that the velocity dispersion of the medium is a reflection of its mass is broken , and a `` diffuse '' intercloud medium may dominate the co luminosity .
a simple - minded application of the co - to- conversion factor yields the harmonic mean between the molecular and stellar masses .
this is likely the reason for the well - established observation that local ultra - luminous ir galaxies are disproportionally luminous in co ( downes & solomon @xcite ) , although a dense phase could still dominate their gas mass ( papadopoulos et al .
@xcite ) .
a particularly interesting case to study is the ngc 253 starburst , close enough to observe in great detail with the atacama large millimeter telescope ( alma ) .
indeed , alma observations can simultaneously be used to measure the global properties and also break up the emission into clouds .
leroy et al . (
@xcite ) present observations of the nuclear region where high density tracer molecules ( hcn , hco@xmath19 , cs , and isotopologues ) are used together to identify ten giant molecular clouds in the starburst .
the authors then use six optically thin tracers ( which include both molecules and dust ) to determine masses for each cloud : the different tracers have a factor of 3 scatter , but a well defined mean . using the surface densities so determined together with the size and velocity dispersion measurements , they are able to place these clouds in a virial diagram ( figure [ fig : one ] ) .
the result is that the clouds in the ngc 253 starburst are self - gravitating , in the sense that their velocity dispersion reflects the expectation for virialized clouds of their surface density .
cloud surface densities are very large , hovering around @xmath20 ( clouds in the milky way disk are typically @xmath21 ) .
nonetheless , the derived conversion factor is @xmath22 .
the conversion factor in the large clouds is even higher , with the four largest clouds yielding @xmath23 .
the uncertainties are large ( 0.5 dex ) , due to the systematic uncertainties in the optically thin mass determination , and the nave application of the @xmath15 scaling would lead us to expect a factor @xmath24 higher conversion factor .
these clouds are also embedded in an extended co `` diffuse '' medium , which has most of the luminosity .
the total measured is uncertain but likely @xmath25 .
this same picture of a mix of phases probably accounts for the bulk of the discrepancy between estimated using the virial theorem in resolved clouds in some galaxy centers ( donovan - meyer et al .
@xcite ) and estimated from dust modeling in the same objects ( sandstrom et al .
@xcite ) . in summary
, what matters in these circumnstances is to understand what fraction of the total co emission originates from self - gravitating complexes , and what is the source of the observed velocity dispersion .
bound , self - gravitating complexes will likely have a conversion factor that is close to the `` galactic '' value ( modulated by changes due to density and temperature effects ) , while the intercloud medium could contribute significantly to the luminosity .
in bolatto et al . ( @xcite ) we propose the following expression to encompass the regimes of the conversion factor : with @xmath27 for @xmath28 and @xmath29 otherwise .
the gas metallicity relative to the milky way , @xmath30 , is used here as an `` observable '' proxy of the dust - to - gas ratio , but we have pointed out the shortcomings of that approximation in [ sec : lowmet ] . the `` typical '' surface density of clouds in the milky way disk is assumed to be 100 , hence @xmath31 is the typical surface density of molecular clouds in the system relative to that in the milky way disk .
conversely , @xmath32 stands for the total surface density that is exerting gravitational attraction in units of 100 .
the first factor in equation [ eq : aco ] incorporates the physics of the - transition relative to the c@xmath19-co transition , as we understand them . in systems of lower metallicity will increase exponentially unless the decrease in metallicity is compensated by an increase in the typical surface density of clouds .
the exponential character of the relation is not a fitting choice : it reflects the theoretical expectation that co emission is rapidly confined to the highest column density regions of the cloud ( wolfire et al .
@xcite ) . in that sense ,
co acts more as a column density tracer than as a bulk mass tracer at low metallicity . the second factor in equation [ eq : aco ] is considerably more uncertain , and reflects the expectation that the co emitting gas , while not necessarily self - gravitating , still has to be bound to the overall potential of the system .
therefore its velocity dispersion will be a function of the total surface density .
the crucial unknown , which we repeat , is the fraction of co emission that arises from self - gravitating clouds : a system where most co emission originates in such clouds will likely have close to galactic .
the choice of 100 as the break point between purely self - gravitating gas and emission from gas bound to the overall potential is likely conservative ( it could be higher ) , and put forward based on the results for galaxy centers . similarly , the @xmath33 exponent implicitly assumes that the co luminosity is dominated by the `` diffuse '' intercloud phase .
galaxy centers may be peculiar : they can have high apparent velocity dispersions due to bulk gas motions ( streaming due to bars , for example ) which would result in a disproportionately large fraction of `` diffuse '' co emission .
so it is possible that this second factor incorrectly predicts too large a correction in .
conversely , it would be hard to push under the value of this prediction unless there is a large increase in the gas temperature .
undoubtedly , future observations of optically thin tracers and multi - transition excitation studies will help clarify some of these unknowns .
99 bolatto , a. d. , wolfire , m. , & leroy , a. k. 2013 , araa , 51 , 207 clark , p. c. , & glover , s. c. o. 2015 , mnras , 452 , 2057 dickman , r. l. , snell , r. l. , & schloerb , f. p. 1986
, apj , 309 , 326 donovan meyer , j. , koda , j. , momose , r. , et al .
2013 , apj , 772 , 107 downes , d. , & solomon , p. m. 1998 , apj , 507 , 615 field , g. b. , blackman , e. g. , & keto , e. r. 2011 , mnras , 416 , 710 fisher , d. b. , bolatto , a. d. , herrera - camus , r. , et al .
2014 , nature , 505 , 186 glover , s. c. o. , & mac low , m .-
2011 , mnras , 412 , 337 jameson , k. e. , bolatto , a. d. , leroy , a. k. , et al .
2015 , apj submitted ( arxiv:1510.08084 ) lee , c. , leroy , a. k. , schnee , s. , et al .
2015 , mnras , 450 , 2708 leroy , a. k. , bolatto , a. d. , ostriker , e. c. , et al . 2015 , apj , 801 , 25 leroy , a. k. , bolatto , a. , bot , c. , et al .
2009 , apj , 702 , 352 leroy , a. , bolatto , a. , stanimirovic , s. , et al .
2007 , apj , 658 , 1027 maloney , p. , & black , j. h. 1988 , apj , 325 , 389 papadopoulos , p. p. , van der werf , p. , xilouris , e. , isaak , k. g. , & gao , y. 2012 , apj , 751 , 10 rmy - ruyer , a. , madden , s. c. , galliano , f. , et al .
2014 , a&a , 563 , a31 rllig , m. , ossenkopf , v. , jeyakumar , s. , stutzki , j. , & sternberg , a. 2006 , a&a , 451 , 917 rubio , m. , elmegreen , b. g. , hunter , d. a. , et al .
2015 , nature , 525 , 218 sandstrom , k. m. , leroy , a. k. , walter , f. , et al .
2013 , apj , 777 , 5 wolfire , m. g. , hollenbach , d. , & mckee , c. f. 2010 , apj , 716 , 1191 | with an increased appreciation for the role of gas in galaxy evolution , there is renewed interest in measuring gas masses for galaxies .
i review some of the basic concepts in using co to determine molecular masses , and discuss some of the recent work . |
the main goal of our ongoing araucaria project is an accurate calibration of the extragalactic distance scale in the local universe . since the first publications resulting from this project ( pietrzyski & gieren 2002 , pietrzyski et al .
2002 ) we have investigated numerous commonly used standard candles , including the cepheid period - luminosity relation , the rr lyrae period - luminosity - metallicity relation , the mean brightness of the red clump stars and the tip of the red giant branch ( trgb ) magnitude in optical and near - infrared bands ( gieren et al . 2005 ; szewczyk et al .
2008 ; pietrzyski et al . 2009 , 2010 ) .
each of these methods of distance measurement have their own and unique potential of application , arising from different distance ranges , observational difficulties and sources of systematic error limiting the achievable accuracy .
measurements are disturbed by numerous effects , like reddening or differences of the chemical composition and age of stellar populations .
each of these factors affects distance measurements in a different way , depending on the technique that was used , as well as the band in which the observations were carried out . as a result ,
discrepancies between the distances obtained for the same galaxy from different methods can reach 20% , as in the case of the m33 galaxy ( gieren et al .
2013 ) . identifying and correctly assessing the impact of population effects is an indispensable step in setting up an accurate distance scale of the universe .
in this paper we focus attention on the trgb method .
the concept of using the trgb as a distance indicator was proposed by baade in 1944 , and later by sandage in 1971 .
the first attempts to measure distances with this technique were made by mould , kristian & da costa ( 1983 , 1984 ) and graham ( 1982 ) .
subsequent studies showed that the absolute brightness of the trgb in the optical i band depends very little on metallicity .
this dependence was calibrated with extensive observations of red giants in galactic globular clusters ( da costa & armandroff 1990 ) . in 1993 lee ,
freedman & madore investigated the accuracy of distance measurement methods , by comparing measurement results for 10 local group galaxies .
they used three different standard candles : cepheids , rr lyrae stars and the trgb .
the general conclusion from this comparison was that all three methods yield results which are consistent within 0.1 mag .
recent studies confirm that distance measurements based on the trgb do not depend on the shape of the star formation history ( sfh ) , or the age of the stars if the mean metallicity of the red giant branch ( rgb ) is smaller than -0.3 dex and the stellar population is older than 1.5 gyr ( kennicutt et al . 1998 ; ferrarese et al . 2000 ; barker et al .
2004 ) . to date
, the distances to more than 40 galaxies nearer than 10 mpc have been obtained with the i band trgb brightness ( karachentsev et al .
2003 ; jacobs et al .
2009 ) .
however , like many other techniques , the trgb measurements can be affected by reddening .
this effect is identified as one of the most important systematic errors on distance measurements .
important progress was made in 2004 when valenti , ferraro & origlia calibrated the trgb infrared j , h and k absolute brightnesses in terms of metallicity .
near - infrared photometry minimizes the influence of interstellar reddening .
unfortunately , however , the trgb brightness in the near - infrared bands is very sensitive to metallicity ( a 0.1 dex change in metallicity changes the brightness of the trgb in the k band by 0.058 mag ) . at present , the accuracy of distance measurements based on the j and k band photometry of the trgb is not sufficiently well studied , and distances to only a few galaxies have been obtained using infrared bands .
recently , our group extended the list of near - infrared based trgb distance measurements to the carina , fornax , sculptor dwarf spheroidal galaxies ( pietrzyski et al .
2009 ) , and ngc 6822 , ngc 3109 , ic 1613 , and wlm ( grski , pietrzyski , gieren 2011 ) .
it is certain that the trgb method has a great potential of application .
the reliability of the method in the optical i band has been shown by many studies and the near - infrared measurements offer the opportunity to minimize the reddening problem .
however , there is an open question about the limitation and uncertainty on the distances obtained from these measurements .
the trgb method was successfully used to measure the distances to galaxies with red giant branches dominated by a homogeneous and old stellar population .
salaris and girardi ( 2005 ) pointed out that in case of composite stellar populations , distance measurements with the trgb technique can be affected by systematic errors . by employing theoretical stellar population synthesis techniques , they studied galaxies with complex star formation histories ( sfhs ) and age - metallicity relationships ( amrs ) , namely the large and small magellanic clouds ( respectively lmc and smc ) .
they concluded that the trgb distance measurements to magellanic cloud - type galaxies can be affected by up to 0.1 - 0.2 mag when using calibrations based on galactic globular clusters .
the lmc and smc are the two most extensively studied galaxies , and with more than 500 distance determinations from a variety of techniques ( de grijs et al .
2014 ) they are a natural choice for studies aimed at testing the accuracy of the trgb method .
the recent very accurate 2.2% distance measurement to the lmc from eclipsing binaries ( pietrzyski et al .
2013 ) gives us a unique opportunity to compare different methods , and use the lmc as a very accurate zero - point for distance calibrations .
measurements of the trgb brightness in the optical i band were performed by reid et al .
( 1987 ) , sakai et al .
( 2000 ) and cioni et al .
the reported brightness determinations differ from each other at the level of 0.3 mag .
studies of the near - infrared trgb brightness in the lmc were performed by kato et al .
( 2007 ) using the infra - red survey facility ( irsf ) , nikolaev & weinberg ( 2000 , using 2mass ) , cioni et al .
( 2000 ) , and macri et al .
( 2015 ) . in the cases of nikolaev & weinberg , kato et al . and
macri et al . , the mean brightness of the trgb is reported without distance determination . in the case of cioni ( denis survey ) ,
an advanced approach was adopted for the trgb detection , and the distance determination is in excellent agreement with the 18.50 mag unreddened distance to the lmc coming from the late - type eclipsing binary studies of our group .
however , these authors obtained the distance with a trgb bolometric brightness calibration .
bolometric corrections in this case have been calculated from the k band and @xmath0 color with blackbody fits . the reported result are certainly valuable , but they can not serve as the basis for the analysis of trgb population effects . since the reported measurements are inconclusive in terms of statistical and systematic error assessment , we decided to thoroughly re - investigate the large and small magellanic clouds .
expected population effects should significantly differ in optical and infrared bands , which was the reason to focus on the optical i band , the near - infrared j , k bands , and bolometric brightness .
additionally , separate measurements of the trgb magnitudes in 17 lmc fields make it possible to evaluate the trgb intrinsic brightness variations for all analyzed bands .
our paper is organized as follows .
the data sources and detection methodology are described in the following section .
next , we present calibrations , and the resulting distance determinations to the lmc and smc .
finally , we present a discussion of our results and a summary .
for our trgb near - infrared analysis we used the irsf magellanic clouds point source catalog ( kato et al .
the irsf is a 1.4 m telescope , located at the south african astronomical observatory ( saao ) in sutherland .
the telescope is equipped with the sirius camera ( 0.45 arcsec pixel scale ) .
the photometric system consists of three near - infrared filters ( j : 1.25 @xmath1 , h : 1.63 @xmath1 , k@xmath2 : 2.14 @xmath1 ) , similar to the 2mass and ukirt photometric systems . while there are other available surveys ( e.g. denis and 2mass ) it is noticeable that irsf provides a better angular resolution and deeper photometry than denis or 2mass . for the optical analysis we used photometric maps of the ogle - iii survey ( udalski et al .
observations were carried out on the 1.3 m warsaw telescope equipped with an eight - chip mosaic camera with 0.26 arcsec pixel scale .
the warsaw telescope is located at the las campanas observatory , operated by the carnegie institution of washington . in order to secure the most precise and reliable measurements possible , with minimal contributions from reddening and geometrical effects ,
we have selected fields in the central regions of the lmc and smc , with internal reddening lower than e(v - i)=0.1 mag . the location of each of the 17 selected fields in the lmc and the five fields chosen in the smc is shown in figure [ fig : lmc ] and figure [ tab : smc ] , respectively .
the size of each field , both in the lmc and smc , is 35 arcmin @xmath3 35 arcmin .
starting with the first publications on the trgb method , many different techniques have been adopted in order to measure the brightness of the trgb . in this paper
we are using the sobel filter , which was introduced by lee et al . in 1993 .
it is based on the convolution of a luminosity function histogram with a discrete sobel kernel [ -2 , -1 , 0 , 1 , 2 ] .
subsequently this method was adopted for a continuous luminosity function ( sakai et al .
a discrete histogram is replaced by a corresponding gaussian - smoothed luminosity function @xmath4 , which follows the expression : @xmath5 \mathrm , \label{eq : sobel}\ ] ] where m@xmath6 is the magnitude of the @xmath7-th star , @xmath8 is the @xmath9 star photometric error , and @xmath10 is the total number of stars in the sample .
sobel filter answer @xmath11 is defined as : @xmath12 where @xmath13 is the mean photometric error for all the stars with magnitudes between @xmath14 and @xmath15 mag . in this method , the brightness of the highest value of the sobel filter answer is adopted as the observed brightness of the trgb .
other detection techniques offer some advantages , in particular when the rgb luminosity function is weakly populated ( makarov et al .
these advantages are unessential for our data , because in each analyzed field at least 700 stars populate the range of 1 mag below the trgb .
furthermore , one can not exclude offsets between different methods . in order to estimate the statistical uncertainty of detection we set up sobel filter with a bootstrap resampling method .
figures [ fig : lmc106 ] , [ fig : lmc127 ] and [ fig : smc162 ] show exemplary luminosity functions and answers of the sobel filter with marked trgb detection for the lmc and smc .
tables [ tab : lmc ] and [ tab : smc ] present the trgb brightness measurements in the i , j and k bands in each of the 17 lmc fields and five smc fields .
the mean magnitudes of the measured trgb in the lmc are 14.62 @xmath16 0.03 mag , 13.27 @xmath16 0.04 mag and 12.13 @xmath16 0.04 respectively in the i , j and k bands .
the mean magnitudes of the trgb we measured in the smc are 15.04 @xmath16 0.07 , 13.89 @xmath16 0.04 and 12.91 @xmath16 0.04 in the i , j and k bands .
in order to accurately determine the distance one has to take into account the reddening . extinction has a relatively small effect in the k infrared band , while it is crucial for the optical i band .
schlegel et al . (
1998 ) give values of the galactic foreground reddening toward the lmc and smc : e(b - v)=0.057 mag and e(b - v)=0.035 mag , respectively . in case of the trgb measurements , internal reddening
is frequently neglected , due to the lack of knowledge of precise values . for the lmc and smc , values of internal reddening
are known from the color excess of red clump stars ( haschke , grebel & duffau 2011 ) . like stars arriving toward the trgb ,
red clump stars belong to population ii and it seems to be appropriate to adopt those values . for our investigation
we decided to take into account only fields with internal reddening lower than e(v - i)=0.1 mag .
the adopted values of e(v - i ) from haschke , grebel & duffau ( 2011 ) are reported in tables [ tab : lmc ] and [ tab : smc ] .
based on the cardelli et al .
( 1989 ) @xmath17 reddening law , we calculated values of extinction for our near - infrared and optical bands .
the mean values for all our fields are : @xmath18 mag , @xmath19 mag , @xmath20 mag , @xmath21 mag for the lmc , and @xmath22 mag , @xmath23 mag , @xmath24 mag , @xmath25 mag for the smc .
for the near - infrared bands we use the empirical trgb absolute brightness - metallicity calibrations provided by valenti , ferraro & origlia ( 2004 ) .
these calibrations were obtained from the homogeneous photometry of 24 galactic globular clusters spanning a metallicity range from -2.12 to -0.49 dex .
absolute magnitudes are expressed in the 2mass photometric system : @xmath26 } - 5.67 \mathrm , \ ] ] @xmath27 } - 6.98 \mathrm .\ ] ] valenti , ferraro & origlia ( 2004 ) also provided a calibration for the absolute bolometric brightness of the trgb in terms of metallicity : @xmath28 } - 3.87\mathrm .\ ] ] to obtain this calibration , the authors transformed the observed k band magnitudes onto a theoretical plane using bolometric corrections for population ii giants ( montegriffo et al .
1995 ) . in the case of our data
, we decided to transform the observed trgb i band magnitudes to bolometric magnitudes with the bolometric corrections calculated from the unreddened ( v - i ) colors of the red giants ( da costa & armandroff 1990 ) : @xmath29 all of the above calibrations use metallicity on the carretta & gratton ( 1997 ) scale .
in the subsequent part of this paper we are also using metallicities calculated from the ( v - i ) color of the red giant branch that is expressed on the zinn & west ( 1984 ) scale .
for this reason , in this case we employ the trgb absolute bolometric brightness calibration of bellazzini & ferraro ( 2001 ) in which metallicity is also on the zinn & west ( 1984 ) scale , and covers the range from -2.2 to -0.2 dex : @xmath30 } - 3.76\mathrm .\ ] ] bellazzini & ferraro ( 2001 ) also provided an optical i band trgb calibration , that follows the formula : @xmath31}^2 + 0.48 \mathrm{[fe / h ] } - 3.66 \mathrm{.}\ ] ] it is worth noting that the bolometric brightness is much less sensitive to metallicity than the k band brightness , but both brightnesses tend to increase with rising metallicity .
the optical i band brightness follows the opposite trend .
the metallicity of the red giant branches in the lmc and smc were the subject of many spectroscopic studies .
olszewski et al . (
1991 ) and da costa et al .
( 1991 ) conducted measurements of the lmc clusters in the outer disc of the galaxy .
they found old clusters with metallicity as low as -1.9 dex , while most of the examined objects had ages less than 3 gyr , and metallicities in the vicinity of -0.5 dex .
surprisingly , there is an evident lack of clusters with intermediate age between 3 and 13 gyr , and a corresponding metallicity from -1.7 to -1.0 dex .
cole at al .
( 2000 ) studied 39 red giants in the proximity of the lmc bar . in contradiction to da costa et al .
( 1991 ) and olszewski et al .
( 1991 ) , they found significantly fewer objects with metallicities lower than -1.0 dex , which hints at different chemical evolution scenarios in the central part of the lmc and its outer disc .
though the younger and more metal abundant part of the metallicity distribution seems to be similar for the da costa et al .
( 1991 ) measurements and the cole at al .
( 2000 ) results .
they both tend to have a metallicity maximum at -0.57 @xmath16 0.04 dex and a mean metallicity of -0.64 @xmath16 0.02 dex . on this basis
, we adopt a metallicity of -0.6 dex with 0.1 dex error due to the standard deviation and the uncertainty of the calibration of the ca ii ir triplet for young , metal - rich stars . since spectroscopic studies are time consuming , and for many galaxies are scarce , the common approach for the trgb distance determinations is to adopt the metallicity calculated from the photometric calibration of the unreddened ( v - i ) color of the red giant branch ( lee at al .
1993 ) : @xmath32 } = -12.64 + 12.6(v - i)_{-3.5}-3.3(v - i)^2_{-3.5 } \mathrm{. } \label{eq : mett}\ ] ] the @xmath33 color is measured for the absolute brightness range @xmath34 to assure sufficient number of stars ( at least 100 ) . in order to determine these values we assumed 18.49 mag distance to the lmc ( pietrzyski et al .
2013 ) and 18.87 mag distance to the smc ( graczyk et al .
reddening correction was applied to the i band brightness and ( v - i ) color of all the stars on the red giant branch .
measured mean @xmath33 colors and their spread for all our fields are reported in tables [ tab : lmc ] and [ tab : smc ] . as pointed out by salaris and girardi ( 2005 ) , in the case of the lmc
, the photometric calibration leads to a much lower metallicity than the spectroscopic measurements . for our lmc fields , the mean metallicity calculated from the ( v - i ) color is -1.23 @xmath16 0.07 dex .
a similar value was also reported by sakai et al .
the cause of this discrepancy is related to the age distribution of the lmc stars and a different metal mixture than in globular clusters from which the above calibration was obtained ( salaris and girardi 2005 ) .
the same discrepancy in metallicity is observed in the smc .
dobbie et al .
( 2014 ) measured spectroscopic metallicities for 3000 red giants in the smc , and obtained a mean metallicity value of [ fe / h ] = -1.0 @xmath16 0.1 dex .
kamath et al .
( 2014 ) obtained a similar value ( [ fe / h ] = -1.14 @xmath16 0.2 dex ) from spectral observations of 42 post - rgb stars .
our calculations , based on photometric calibrations , lead to a metallicity -1.51 @xmath16 0.03 dex . as in the case of the lmc ,
the metallicity calculated from photometric calibrations is lower by about 0.5 dex compared with the spectroscopic values .
the calibrations introduced in the previous section are based on spectroscopic metallicities of stars in the galactic globular clusters .
for this reason we are using spectroscopic metallicity , while metallicity calibrated from the @xmath35 color will be discussed in the final part of this paper .
our measured mean trgb brightnesses , adopted reddenings and metallicities lead to the following distance moduli for the lmc : @xmath36 @xmath37 @xmath38 @xmath39 for the smc we obtain : @xmath40 @xmath41 @xmath42 @xmath43 tables [ tab : lmc_err ] and [ tab : smc_err ] summarize the contributions to total statistical and systematic uncertainties of our distance determinations . the systematic uncertainty of our distance determinations is composed of a reddening systematic error ( 0.05 mag ) and a metallicity systematic error ( 0.1 dex ) .
the statistical error is estimated as the standard deviation of the trgb measurements for all fields .
identified contributors are photometric and detection errors , reddening errors , and variations of metallicity ( 0.08 dex for the lmc and 0.03 for the smc ) .
metallicity variations were estimated from the mean color of the rgb with equation 8 .
it is important to state at this moment , that all mentioned errors are not independent .
both photometric and reddening errors are correlated with estimated metallicity variations .
reddening is the main contributor to the statistical and systematic uncertainty of our distance determination for the i band and for the bolometric brightness , while it is negligible for the k band distance determination .
the k band distance is mainly affected by the metallicity error , which is negligible for the bolometric brightness - based trgb distance estimation .
our trgb brightness measurements are in good agreement with results obtained by other authors , though some comment is necessary .
nikolaev & weinberg ( 2000 ) measured the trgb brightness in the k band from 2mass data , and obtained @xmath44 mag .
this value was estimated from the luminosity function histogram , which is biased by binning confines .
in fact , direct examination of the presented histogram suggests that the value measured with sobel filter should yield @xmath45 mag and would be consistent with our result ( @xmath46 mag ) , and also with the recent result obtained by macri et al .
( 2015 ) ( @xmath47 mag ) .
cioni et al .
( 2000 ) measured the trgb brightness from the denis survey , and obtained @xmath48 mag in the denis photometric system .
this estimation is in significant disagreement with nikolaev & weinberg ( 2000 ) , macri et al .
( 2015 ) and our results .
the reason for this disagreement , as indicated by cioni et al . ( 2000 ) , can be related to the non - standard detection technique used by the authors or , even more explicitly , to the non - standard photometric system of the denis survey .
the i band trgb brightness was obtained by reid et al .
( 1987 ) ( @xmath49 mag ) and sakai et al .
( 2000 ) ( @xmath50 mag with reddening corrected individually for each star , with mean @xmath51 mag ) .
our i band trgb brightness measurement ( @xmath52 mag ) is in a good agreement with these results . nonetheless , reported distance moduli vary from 18.4 mag to 18.6 mag , due to different values of reddening and metallicity adopted by each author .
the distance determinations presented in our paper from the measurements of the trgb brightness are inconsistent with the distances to the lmc and smc obtained from the eclipsing binaries ( lmc : 18.49 @xmath16 0.02 mag , pietrzyski et al .
2013 ; smc : 18.97 @xmath16 0.03 mag , graczyk et al .
our optical i band distance to the lmc is underestimated at the level of 0.2 mag .
the infrared k band distance is overestimated by 0.2 mag , both for the lmc and smc .
the discrepancy between i and k band distance moduli is 0.4 mag for the lmc and 0.3 mag for the smc .
although there are significant deviations for all bands , the differential distance ( @xmath53 ) between the lmc and smc is surprisingly consistent for the j , k bands and for the bolometric brightness , i.e. @xmath54 mag , @xmath55 mag and @xmath56 mag.it is also important that these values are in good agreement with the differential distance obtained from the eclipsing binaries ( @xmath57 mag ; graczyk et al .
however , the optical i band differential distance between the lmc and smc ( @xmath58 mag ) is not consistent with the j , k bands , bolometric brightness , and the eclipsing binaries .
although a deep analysis of the observed discrepancies is beyond the main scope of this paper , we briefly discuss possible sources of systematic errors .
one of the origins of systematic error may arise from the calibration of valenti , ferraro & origlia ( 2004 ) .
it affects distance estimations from infrared @xmath59 and @xmath60 band trgb brightness , and from the bolometric brightness . in the lmc and smc
we expect that the sobel filter method should indicate the trgb brightness with a precision within the reported uncertainties , thanks to the well populated bright part of the rgb .
in contrast to this , the calibrations of valenti , ferraro & origlia ( 2004 ) were obtained with * the trgb * brightness measurements in 24 galactic globular clusters . for the majority of these globular clusters
it is impossible to apply the sobel filter , because in the close proximity of the trgb there are only a few stars .
instead , the authors assumed that the brightest star on the rgb corresponds to the trgb .
this can lead to a significant systematic error of the absolute brightness of the trgb in their calibration , which leads to an underestimation of our distance moduli based on the detection of the trgb with the sobel filter .
viaux et al .
2013 estimated the probability distribution of the offset between the brightest star in the m5 cluster and the trgb . in the case of the m5 cluster , with a 65% confidence level ,
the offset between the brightest star and the trgb is smaller than 0.05 mag . to correctly asses this effect ,
a calculation for all clusters from valenti , ferraro & origlia ( 2004 ) should be applied , but it should not be larger than 0.05 mag , as in the case of the m5 cluster .
we note at this point , that taking into account the offset between the brightest star on the rgb and the brightness of the trgb will lead to even larger discrepancies of our distance measurements with the 18.49 mag and 18.97 mag value of the distance moduli to the lmc and smc coming from the eclipsing binaries .
another possible explanation of the disagreement of our results with the classical distances to the magellanic clouds is connected with the value of the zero - point in both calibrations ( valenti , ferraro & origlia 2004 and bellazzini & ferraro 2001 ) .
the error in the zero - point value may be caused by adopting incorrect values of the distances to the globular clusters which were used to obtain those calibrations .
it would explain why the differential distance between the lmc and smc is consistent in the j and k bands and for the bolometric brightness with the results from eclipsing binaries , but it does not explain why the distance obtained for each of these galaxies differs in the j and k bands and for the bolometric brightness . even taking into account that the calibration of valenti , ferraro & origlia ( 2004 ) has a different error in the zero - point than the calibration of bellazzini & ferraro ( 2001 ) , it is unlikely for it to reach 0.4 mag difference between i and k bands .
while rather unlikely , one could suspect that the disagreement between bolometric brightness and k band is caused by adopting incorrect values of metallicity or reddening in our calculations . a different value of metallicity ( e.g. [ fe / h ] = -1.0 dex ) can bring the calculated distance to the lmc in the k band close to 18.50 , but the discrepancies for the other bands remain at the level of 0.2 mag . even adopting different values for both metallicity and reddening can not bring all estimations together ( i , j , k band and bolometric brightness ) to the expected distance .
the same applies to the smc .
this implies systematic errors , strongly depending on the band which is used for the measurements .
the most convincing explanation of the dominating systematic error is provided by salaris and girardi ( 2005 ) .
their analysis of artificial synthesis populations based on star formation histories confirms the discrepancies obtained in this paper .
the authors are unable to predict precise corrections , but they show that young stars should produce an effect that is imitating a lower metallicity of the stars
. the observed underestimation of the distance in the optical i band and overestimation in the near - infrared bands exactly corresponds to this effect . as mentioned in the previous sections , the trgb distance determinations are in many cases based on metallicities calculated from the ( v - i ) color of the red giant branch ( equation 8) .
the metallicity calculated with this method yields values of -1.23 dex for the lmc , and -1.51 dex for the smc , being 0.5 dex underestimated when comparing to the spectroscopic results . the absolute brightness obtained from adopting this metallicity leads to distance moduli of 18.51 mag for the lmc , and 19.04 for smc , and is in good agreement with the distance to the lmc and smc obtained from the eclipsing binaries .
this rather peculiar accuracy of distance estimation is caused by the opposing effects of stellar population to both calculated metallicity and bolometric brightness of the trgb , which tend to compensate each other .
this effect will be discussed in detail in a forthcoming paper .
we have measured the brightnesses of the trgb in the large and small magellanic clouds in the optical @xmath61 band and in the near - infrared @xmath59 and @xmath60 bands .
measurements were conducted separately in 17 fields in the central region of the lmc and in five fields in the center of the smc .
variations of the trgb brightness between different fields in the lmc are at the level of 0.03 - 0.04 mag and are caused by differences in the internal reddening and metallicity . with the calibrations presented in section 4 , we were able to calculate absolute brightnesses of the trgb in the @xmath61 , @xmath59 , @xmath60 bands , and for the bolometric brightness .
the distances calculated on this basis are inconsistent between each other , and differ from the generally accepted 18.50 mag distance to the lmc , and 19.0 mag to the smc . for the @xmath61 band , the distance is underestimated by 0.2 mag , and for the k band it is overestimated by 0.2 mag .
the detected systematic errors are caused by population effects in the large and small magellanic clouds mainly by the mean age of the stars on the red giant branch .
the reported differences can be taken as an upper limit on the systematic error of the trgb distance measurements
. they should not affect distance measurements to galaxies with red giant branches dominated by a homogeneous and old stellar population .
this systematic population effect is smaller for the @xmath59 band and for the bolometric brightnesses , as compared to the k band .
moreover , distances obtained from the bolometric brightness in tandem with a metallicity calculated from the @xmath35 color of red giants agree with the 18.50 mag eclipsing binary distance to the lmc , and with the 19.00 mag distance to smc . however , the adopted value of the metallicity is incorrect , and this fortuitous agreement should be considered as due to the compensatory effect of oppositely directed systematic errors of bolometric absolute brightness and metallicity .
we gratefully acknowledge financial support for this work from the polish national science center grant preludium 2012/05/n / st9/03846 and team subsidy from the foundation for polish science ( fnp ) .
we also gratefully acknowledge financial support from the basal centro de astrofisica y tecnologias afines ( cata ) pfb-06/2007 , and from the millenium institute of astrophysics ( mas ) of the iniciativa cientifica milenio del ministerio de economia , fomento y turismo de chile , project ic120009 .
m.c . acknowledges additional support from fondecyt grant 1141141 .
this work is based on observations made by the ogle project and the irsf magellanic clouds point source catalog , and it is a pleasure to thank for the excellent photometric data provided by those teams .
baade , w. 1944 , , 100 , 137 barker , m.k . ,
sarajedini , a. , harris , j. 2004 , , 606 , 869 bellazzini , m. , ferraro , f. r. , pancino , e. 2001 , , 556 , 635 cardelli , j. a. , clayton , g. c. , mathis , j. s. 1989 , , 345 , 245 carretta , e. , & gratton , r. 1997 , _ _
a&a__s , 121 , 95 cioni , m.r .
, van der marel , r.p . ,
loup , c. and habing , h.j .
2000 , _ a&a _ , 359 , 601 cole , a. a. , smecker - hane , t. a. , gallagher , j. s. 2000 , , 120 , 1808 da costa , g. s. , & armandroff , t. e. 1990 , , 100 , 162 da costa g. s. 1991 , in haynes r. , milne d. , proc . iau symp . 148 , 183 de grijs r. , wicker j. e. , bono g. , 2014 , , 149 , 179 dobbie , p. d. , cole , a. a. , subramaniam , a. , keller , s. 2014 , , 442 , 1680 ferrarese , l. , ford , h. c. , huchra , j. , et al . 2000 , , 128 , 431 gieren , w. , grski , m. , pietrzyski , g. , et al .
2013 , , 773 , 69 gieren , w. , pietrzyski , g. , soszyski , i. , bresolin , f. , kudritzki , r. p. , minniti , d. , storm , j. 2005 , , 628 , 695 grski , m. , pietrzyski , g. , gieren , w. 2011 , , 141,194 graczyk , d. , pietrzyski , g. , thompson , i. b. , gieren , w. , pilecki , b. 2014 , , 780 , 59 graham , j. a. 1982 , , 252 , 474 haschke , r. , grebel , e. k. , duffau , s. 2011 , , 141 , 158 jacobs , b. a. , rizzi , l. , tully , r. b. , et al .
2009 , , 138 , 332 kamath , d. , wood , p. r. , van winckel , h. 2014 , , 439 , 2211 karachentsev , i. d. , grebel , e. k. , sharina , m. e. , et al .
2003 , 404 , 93 kato , d. , nagashima , c. , nagayama , t. , et al .
2007 , pasj , 59 , 615 kennicutt , r. c. , jr . , stetson , p. b. , saha , a. , et al .
1998 , , 498 , 181 lee , m.g . ,
freedman , w.l .
, madore , b.f .
1993 , , 417 , 553 macri , l.m .
, ngeow , c.c . ,
kanbur , s.m . , mahzooni , s. , smitka , m.t .
2015 , , 149 , 117 makarov , d. , makarova , l. , rizzi , l. , et al .
2006 , , 132 , 2729 mould , j. r. , kristian , j. , da costa , g. s. 1983 , 270 , 471 mould , j. , kristian , j. , da costa , g. s. 1984 , , 278 , 575 montegriffo , p. , ferraro , f. r. , fusi pecci , f. , origlia , l. 1995 , mnras , 276 , 739 nikolaev , s. , weinberg , m. d. 2000 , , 542 , 804 olszewski , e. w. , schommer , r. a. , suntzeff , n. b. , harris , h. c. 1991 , 101 , 515 pietrzyski , g. , & gieren , w. 2002 , , 124 , 2633 pietrzyski , g. , gieren , w , fouqu , p. pont , f. 2002 , , 123 , 789 pietrzyski , g. , grski , m. , gieren , w. , ivanov , v. d. , bresolin , f. , kudritzki , r .- p .
2009 , , 138 , 459 pietrzyski , g. , grski , m. , gieren , w. , laney , d. , and udalski , a. 2010 , , 140 , 1038 pietrzyski , g. , graczyk , d. , gieren , w. , et al .
2013 , nature , 495 , 76 reid n. , mould j. , thompson i. 1987 , 323 , 433 sakai , s. , madore , b. , freedman , w.l .
1996 , , 461 , 713 sakai , s. , zaritsky , d. , kennicutt , r. c. jr .
2000 , , 119 , 1197 salaris , m. , girardi , l. 2005 , , 357 , 669 sandage , a. 1971 , 166 , 13 schlegel , d. j. , finkbeiner , d. p. , davis , m. 1998 , 500 , 525 szewczyk , o. , pietrzynski , g. , gieren , w. , et al .
2008 , , 136 , 272 udalski , a. , soszynski , i. , szymanski , m. k. , et al .
2008 , aca , 58 , 89 valenti , e. , ferraro , f.r . ,
origlia , l. 2004 , , 354 , 815 viaux , n. , catelan , m. , stetson , p. b. , raffelt , g. g. , redondo , j. , valcarce , a. a. r. , weiss , a. 2013 , _ a&a _ , 558 , 12 zinn , r. , & west , m.j .
1984 , apjs , 55 , 45 lmc100 & 5:19:02.2 & -69:15:07 & 14.61 & 13.22 & 12.12 & 0.06 & 1.71 & 0.18 + lmc103 & 5:19:02.9 & -69:50:26 & 14.62 & 13.24 & 12.11 & 0.06 & 1.69 & 0.18 + lmc108 & 5:13:01.9 & -67:28:40 & 14.57 & 13.31 & 12.14 & 0.09 & 1.68 & 0.17 + lmc110 & 5:12:43.6 & -68:39:42 & 14.58 & 13.25 & 12.11 & 0.08 & 1.70 & 0.18 + lmc112 & 5:12:21.5 & -69:50:21 & 14.59 & 13.27 & 12.10 & 0.07 & 1.69 & 0.14 + lmc117 & 5:06:55.3 & -68:03:58 & 14.64 & 13.27 & 12.17 & 0.05 & 1.63 & 0.13 + lmc118 & 5:06:25.4 & -68:39:25 & 14.61 & 13.27 & 12.14 & 0.07 & 1.65 & 0.15 + lmc119 & 5:06:02.5 & -69:15:02 & 14.65 & 13.27 & 12.13 & 0.07 & 1.63 & 0.15 + lmc120 & 5:05:39.8 & -69:50:28 & 14.65 & 13.35 & 12.15 & 0.08 & 1.67 & 0.12 + lmc126 & 5:00:02.4 & -68:39:31 & 14.61 & 13.29 & 12.18 & 0.08 & 1.65 & 0.16 + lmc127 & 4:59:33.6 & -69:14:54 & 14.63 & 13.29 & 12.21 & 0.09 & 1.66 & 0.15 + lmc128 & 4:59:03.6 & -69:50:24 & 14.68 & 13.25 & 12.11 & 0.08 & 1.67 & 0.12 + lmc161 & 5:25:32.5 & -69:14:59 & 14.62 & 13.34 & 12.17 & 0.09 & 1.67 & 0.15 + lmc162 & 5:25:43.3 & -69:50:24 & 14.57 & 13.22 & 12.08 & 0.05 & 1.65 & 0.16 + lmc163 & 5:25:52.2 & -70:25:55 & 14.64 & 13.23 & 12.07 & 0.07 & 1.73 & 0.12 + lmc169 & 5:32:22.8 & -69:50:26 & 14.69 & 13.30 & 12.15 & 0.08 & 1.72 & 0.16 + lmc170 & 5:32:48.1 & -70:25:53 & 14.60 & 13.24 & 12.10 & 0.06 & 1.66 & 0.18 + [ tab : lmc ] smc100 & 0:50:06.4 & -73:08:19 & 15.20 & 13.87 & 12.75 & 0.04 & 1.50 & 0.11 + smc103 & 0:50:08.6 & -73:43:44 & 15.15 & 13.84 & 12.82 & 0.04 & 1.50 & 0.11 + smc106 & 0:58:06.7 & -73:20:21 & 14.99 & 13.89 & 12.86 & 0.04 & 1.48 & 0.07 + smc108 & 0:57:31.5 & -72:09:29 & 15.02 & 13.91 & 12.98 & 0.04 & 1.48 & 0.09 + smc113 & 1:05:02.8 & -72:09:32 & 14.98 & 13.92 & 12.90 & 0.04 & 1.49 & 0.11 + [ tab : smc ] + std & 0.03 & 0.03 & 0.04 & 0.04 + photometric & 0.02 & 0.02 & 0.02 & 0.02 + detection & 0.02 & 0.02 & 0.02 & 0.02 + std reddening & 0.02 & 0.01 & 0.006 & 0.001 + std metallicity & 0.01 & 0.03 & 0.02 & 0.05 + + reddening & 0.07 & 0.07 & 0.04 & 0.02 + metallicity & 0.02 & 0.04 & 0.03 & 0.06 + + std & 0.07 & 0.07 & 0.04 & 0.04 + photometric & 0.02 & 0.02 & 0.02 & 0.02 + detection & 0.02 & 0.02 & 0.02 & 0.02 + std reddening & 0.02 & 0.01 & 0.006 & 0.001 + std metallicity & 0.005 & 0.01 & 0.01 & 0.02 + + reddening & 0.07 & 0.07 & 0.04 & 0.02 + metallicity & 0.02 & 0.04 & 0.03 & 0.06 + | we present a precise optical and near - infrared determination of the tip of the red giant branch ( trgb ) brightness in the large and small magellanic clouds ( respectively lmc and smc ) .
the commonly used calibrations of the absolute magnitude of the trgb lead to an overestimation of the distance to the lmc and smc in the k band , and an underestimation of the distance in the optical i band for both galaxies . reported discrepancies are at the level of 0.2 mag , with respect to the very accurate distance determinations to both clouds based on late - type eclipsing binaries .
the differential distances between the lmc and smc obtained in the j and k bands , and for the bolometric brightness are consistent with each other , and with the results obtained from eclipsing binaries and other distance indicators . |
there are several different approaches to the fundamental concept of fractional poisson process ( fpp ) on the real line .
the renewal approach consists of considering the characterization of the poisson process as a sum of independent non - negative random variables , and instead of the assumption that these random variables have an exponential distribution , we assume that they have the mittag - leffler distribution ( see @xcite ) . in @xcite , the renewal approach to the fractional poisson process is developed and it is proved that its one - dimensional distributions coincide with the solution to fractionalized state probabilities .
another equivalent approach is using inverse subordinator , a kind of fractional poisson process can be constructed @xcite . in @xcite , following this method , the fpp is generalized and defined , obtaining a fractional poisson random field ( fprf ) parametrized by points of the euclidean space @xmath0 , as it has been done for fractional brownian fields , see , e.g. , @xcite .
the starting point of our extension will be the set - indexed poisson process which is a well - known concept , see , e.g. , @xcite . in this paper , we first present a martingale characterization of the fractional poisson process .
we extend this characterization to fprf using the concept of increasing path and strong martingales .
this characterization permits us to give a definition of a set - indexed fractional poisson process .
we study the fractional differential equation for fprf . finally , we study mixed - fractional poisson processes .
the paper is organized as follows . in the next section ,
we collect some known results from the theory of subordinators and inverse subordinators , see @xcite among others . in section [ sect:3 ] ,
we prove a martingale characterization of the fpp , which is a generalization of the watanabe theorem . in section [ sect:6 ] , another generalization called `` mixed - fractional poisson process ''
is introduced and some distributional properties are studied as well as watanabe characterization is given .
section [ sect:4 ] is devoted to fprf .
we begin by computing covariance for this process , then we give some characterizations using increasing paths and intensities .
we present a gergely - yeshow characterization and discuss random time changes .
fractional differential equations are discussed on section [ sect:5 ] .
finally , we present some simulations for the fprf . other different generalizations of fpp and fields can be found in @xcite .
this section collects some known resuts from the theory of subordinators and inverse subordinators @xcite . consider a non - decreasing lvy process @xmath1 starting from @xmath2 which is continuous from the right with left limits ( cadlag ) , continuous in probability , with independent and stationary increments . such a process is known as a lvy subordinator with laplace exponent @xmath3 where @xmath4 is the drift and the lvy measure @xmath5 on @xmath6 satisfies @xmath7 this means that @xmath8 consider the inverse subordinator @xmath9 which is given by the first - passage time of @xmath10@xmath11the process @xmath9 is non - decreasing and its sample paths are a.s .
continuous if and its sample paths are a.s .
right continuous if @xmath12 is strictly increasing .
we have @xmath13 and for any @xmath14 @xmath15 which implies that for any @xmath16 , since @xmath17 let @xmath18 be the renewal function . since @xmath19then @xmath20 characterizes the inverse process @xmath21 , since @xmath22 characterizes @xmath23 we get a covariance formula @xcite @xmath24 the most important example is considered in the next section , but there are some other examples .
let @xmath25 , be an @xmath26stable subordinator with @xmath27 .
the density of @xmath28 is of the form @xcite @xmath29 here we use the wright s generalized bessel function ( see , e.g. , @xcite ) @xmath30where @xmath31 and @xmath32 .
the set of jump times of @xmath33 is a.s .
the lvy subordinator is strictly increasing , since the process @xmath34 admits a density .
then the inverse stable subordinator @xmath35 has density @xcite @xmath36 the laplace transform of the density @xmath37 is @xmath38 its paths are continuous and nondecreasing . for @xmath39
the inverse stable subordinator is the running supremum process of brownian motion , and for @xmath40 $ ] this process is the local time at zero of a strictly stable lvy process of index @xmath41 let @xmath42be the mittag - leffler function @xcite , and recall the following : \i ) the laplace transform of function @xmath43 is of the form @xmath44
\(ii ) the function @xmath45 is an eigenfunction at the the fractional caputo - djrbashian derivative @xmath46 with eigenvalue @xmath47 @xcite @xmath48where @xmath46 is defined as ( see @xcite ) @xmath49 [ prop:2.1 ] the @xmath50-stable inverse subordinators satisfy the following properties : 1 .
@xmath51 2 .
both processes @xmath52 and @xmath53 are self - similar @xmath54 3 . for @xmath55 , @xmath56 ^{1-\alpha } } .\ ] ] + in particular , + @xmath57^{\nu } = \frac{\gamma ( \nu + 1)}{\gamma ( \alpha \nu + 1)}t^{\alpha \nu } , \quad\nu > 0;\ ] ] + @xmath58 this subsection collects some results from the theory of inverse subordinators , see @xcite
. different kinds of inverse subordinators can be considered .
let @xmath59 and @xmath60 be two independent stable subordinators .
the mixture of them @xmath61 is defined by its laplace transform : for @xmath62 , @xmath63 it is clear that @xmath64 is not self - similar , unless @xmath65 since @xmath66 this expression is equal to @xmath67 if and only if for any @xmath68 @xmath69 in which case the process @xmath70 can be reduced to the classical stable subordinator ( up to constant ) .
the inverse subordinator is defined by @xmath71 we assume that @xmath72 without loss of generality ( the case @xmath73 reduces to the previous case of single inverse subordinator ) .
it was proved in @xcite that @xmath74where @xmath75 is the two - parametric generalized mittag - leffler function ( @xcite ) @xmath76also for the laplace transform of the density @xmath77 @xmath78 of inverse subordinator @xmath79 we have the following expression @xcite : @xmath80e^{-u[c_{1}s^{\alpha _ { 1}}+c_{2}s^{\alpha _ { 2}}]},\quad s\geq 0 , \label{m2}\]]and the laplace transform of @xmath81 is given by @xmath82 from ( * ? ? ?
* theorem 2.3 ) we have the following expression for @xmath83 : @xmath84@xmath85 one can also consider the tempered stable inverse subordinator , the inverse subordinator to the poisson process , the compound poisson process with positive jumps , the gamma and the inverse gaussian lvy processes
. for additional details see @xcite .
the first definition of fpp @xmath86 was given in @xcite ( see also @xcite ) as a renewal process with mittag - leffler waiting times between the events @xmath87 where @xmath88 @xmath89 are iid random variables with the strictly monotone mittag - leffler distribution function @xmath90 the following stochastic representation for fpp is found in @xcite : @xmath91 where @xmath92 is the classical homogeneous poisson process with parameter @xmath93 which is independent of the inverse stable subordinator @xmath94 one can compute the following expression for the one - dimensional distribution of fpp : @xmath95where @xmath96 is given by ( [ d1 ] ) , @xmath97 is the mittag - leffler function ( [ ml1 ] ) , @xmath98 is the @xmath99th derivative of @xmath97 , and @xmath100 is the three - parametric generalized mittag - leffler function defined as follows @xcite : @xmath101 where @xmath102 is the pochhammer symbol .
finally , in @xcite it is shown that the marginal distribution of fpp satisfies the following system of fractional differential - difference equations:@xmath103with initial conditions : @xmath104 and @xmath105 where @xmath46 is the fractional caputo - djrbashian derivative ( [ fd ] ) .
see also @xcite .
note that @xmath106\big]= \int_{0}^{\infty } [ \mathrm{e}n(u)]f_{\alpha } ( t , u)du=\lambda t^{\alpha } /\gamma ( 1+\alpha ) , \]]where @xmath107 is given by ( [ d1 ] ) , and @xcite showed that @xmath108where @xmath109 is given in while @xmath110 . in particular , @xmath111+\frac{\lambda t^{\alpha } } { \gamma ( 1+\alpha ) } \\ & = & \frac{\lambda ^{2}t^{2\alpha } } { \gamma ^{2}(1+\alpha ) } \big(\frac{\alpha \gamma ( \alpha ) } { \gamma ( 2\alpha ) } -1\big)+\frac{\lambda t^{\alpha } } { \gamma ( 1+\alpha ) } , \quad t\geq 0.\end{aligned}\]]the definition of the hurst index for renewal processes is discussed in @xcite . in the same spirit , one can define the analogous of the hurst index for the fpp as @xmath112 to prove the formula , one can use the conditional covariance formula ( * ? ? ?
* exercise 7.20.b ) : @xmath113 where @xmath114 and @xmath21 are random variables , and @xmath115 really , if @xmath116 then @xmath117 , and @xmath118 since , for example , if @xmath119 , then @xmath120 , and @xmath121 for more than one random variable in the condition , the conditional covariance formula becomes more complicated , it can be seen even for the conditional variance formula : @xmath122 | y_1 \big ) + \mathrm{var}\big ( \mathrm{e } ( z | y_1 ) \big).\ ] ] the corresponding formulas can be found in @xcite .
that is why for random fields we develop another technique , see appendix .
let @xmath123 be a complete probability space .
recall that the @xmath124adapted , @xmath125-integrable stochastic process @xmath126 is an @xmath124martingale ( sub - martingale ) if @xmath127 @xmath128 a.s .
, where @xmath129 is a non - decreasing family of sub - sigma fields of @xmath130 .
the following theorem is known as the watanabe characterization for homogeneous poisson processes ( see , @xcite and @xcite ) : let @xmath131 be a @xmath124adapted , simple locally finite point process . then @xmath132 is a homogeneous poisson process iff there is a constant @xmath133 , such that the process @xmath134 is an @xmath124martingale .
we extend the well - known watanabe characterization for fpp .
let @xmath135 be a simple locally finite point process .
then @xmath136 is a fpp iff there exist a constant @xmath93 and an @xmath137-stable subordinator @xmath138 @xmath139 such that , denoted by @xmath140 its inverse stable subordinator , the process @xmath141 is a martingale with respect to the induced filtration @xmath142 .
if @xmath136 is a fpp , then @xmath143 where @xmath144 is the inverse of an @xmath137-stable subordinator and @xmath132 is a poisson process with intensity @xmath145 therefore @xmath146 is a martingale .
notice that @xmath53 is continuous increasing and adapted ; therefore it is the predictable intensity of the sub - martingale @xmath147 conversely , it is enough to prove that @xmath143 where @xmath132 is a poisson process , independent of @xmath94 consider the inverse of @xmath148 @xmath149notice that in general @xmath150 does not equal to @xmath151 since for example @xmath150 is continuous , but @xmath152 is not .
@xmath153 can be seen as a family of stopping times .
then , by the doob optional sampling theorem ( * ? ? ?
* theorem 7.29 ) , @xmath154is still a martingale .
then @xmath155 is a martingale .
since @xmath150 is increasing , @xmath156 is a simple point process . following the classical watanabe characterization
, @xmath156 is a classical poisson process with parameter @xmath133 . call this process @xmath157 then @xmath158 is a fpp .
in this section , we consider a more general mixed - fractional poisson process ( mfpp ) @xmath159where the homogeneous poisson process @xmath160 with intensity @xmath93 and the inverse subordinator @xmath161 given by ( [ ism ] ) are independent .
we will show that @xmath162 is the stochastic solution of the system of fractional differential - difference equations : for @xmath163 , @xmath164 with initial conditions : @xmath165where @xmath46 is the fractional caputo - djrbashian derivative ( [ fd ] ) , and for @xmath166 , @xmath167 , @xmath168 using the formulae for laplace transform of the fractional caputo - djrbashian derivative ( see , @xcite):@xmath169one can obtain from ( [ me1 ] ) with @xmath170 the following equation @xmath171for the laplace transform @xmath172 thus @xmath173and using the formula for an inverse laplace transform ( see , @xcite ) , for @xmath174 and @xmath175 : @xmath176 one can find an exact form of the @xmath177 in terms of generalized mittag - leffler functions ( [ ml3 ] ) : @xmath178for @xmath179we obtain from ( [ me1 ] ) : @xmath180where @xmath181thus from ( [ me1 ] ) we obtain the following expression for the laplace transform of @xmath182 @xmath183 @xmath184 on the other hand , one can compute the laplace transform from the stochastic representation ( [ sr ] ) .
if @xmath185where @xmath186 is given by ( [ md2 ] ) , then using ( [ m2]),([m3 ] ) we have for @xmath187@xmath188dx\]]@xmath189note that @xmath190thus@xmath191the same expression as ( [ for1 ] )
. we can formulate the result in the following form : the mfpp @xmath162 defined in ( [ sr ] ) is the stochastic solution of the system of fractional differential - difference equations ( [ me1 ] ) with initial conditions ( [ me2 ] ) .
note that in @xcite one can find some other stochastic representations of the mfpp ( [ sr ] ) .
also , some analytical expression for @xmath192 is given by ( [ f ] ) , while the analytical expression for @xmath182for @xmath179 are given by ( [ md3 ] ) .
moreover , @xmath182for @xmath179 can be obtained by the following recurrent relation : @xmath193where @xmath194 and from : @xmath195 from ( * ? ? ?
* theorem 2.1 ) and ( [ ui ] ) , we have the following expressions for moments in form of the function@xmath196\\ & \qquad + \lambda \frac{1}{c_{2}}t^{\alpha _ { 2}}e_{\alpha _ { 2}-\alpha _ { 1},\alpha _ { 2}+1}(-c_{1}t^{_{\alpha _ { 2}-\alpha _ { 1}}}/c_{2 } ) , \\ \mathrm{cov}(n^{\alpha _ { 1},\alpha _ { 2}}(t),n^{\alpha _ { 1},\alpha _ { 2}}(s ) ) & = \lambda u(\min ( t , s ) ) + \lambda ^{2}\big\ { \int_{0}^{\min ( t , s)}\big(u(t_{1}-\tau ) \\ &
\qquad \qquad + u(t_{2}-\tau ) \big)du(\tau ) -u(t_{1})u(t_{2})\big\ } .\end{aligned}\ ] ] we extend the watanabe characterization for mfpp .
let @xmath197 be a non - negative right - continuous non - decreasing deterministic function such that @xmath198 , @xmath199 , and @xmath200 for any @xmath201
. such a function will be called _
consistent_. the mixed - fractional non - homogeneous poisson process ( mfnpp ) is defined as @xmath202where the homogeneous poisson process @xmath160 with intensity @xmath203 , and the inverse subordinator @xmath161 given by ( [ ism ] ) are independent .
let @xmath135 be a simple locally finite point process .
@xmath136 is a mfnpp iff there exist a consistent function @xmath204 and a mixed stable subordinator \{@xmath205
@xmath206 @xmath207 defined in ( [ m1 ] ) , such that @xmath208is a martingale with respect to the induced filtration @xmath209 , where @xmath210 is the inverse mixed stable subordinator .
if @xmath136 is a mfnpp , then @xmath211 where the homogeneous poisson process @xmath160 and the inverse subordinator @xmath161 given by ( [ ism ] ) are independent . therefore @xmath212 is a martingale .
conversely , it is enough to prove that @xmath211 where @xmath132 is a poisson process , independent of @xmath94 consider the pseudo inverse of @xmath213 @xmath214 @xmath153 can be seen as a family of stopping times . then , by the doob optional stopping theorem , @xmath215is still a martingale .
then @xmath216 is a martingale .
since @xmath150 is increasing , @xmath156 is a simple point process . following the extension of watanabe characterization given in @xcite
, @xmath156 is a non - homogeneous poisson process with deterministic intensity @xmath217 .
call this process @xmath157 then @xmath218 is a mfnpp .
this section collects some known results from the theory of two - parameter poisson processes and homogeneous poisson random fields ( prf ) ( see , e.g. , @xcite , among the others ) .
let @xmath219 be a complete probability space and let @xmath220 be a family of sub-@xmath221-fields of @xmath130 such that \(i ) @xmath222 for any @xmath223 \(ii ) @xmath224 contains all null sets of @xmath225 \(iii ) for each @xmath226 @xmath227 where @xmath228 denotes the partial order on @xmath229 which means that @xmath230 given @xmath231 we denote by @xmath232 the increments of the random field @xmath233 over the rectangle @xmath234 $ ] .
in addition , we denote @xmath235 a strong martingale is an integrable two - parameter process @xmath136 such that @xmath236 for any @xmath237 .
let @xmath238 be a family of sub-@xmath221-fields of @xmath130 satisfying the previous conditions ( i ) , ( ii ) , ( iii ) for all @xmath239 an @xmath240 prf is adapted , a cadlag field @xmath241 , such that , \(1 ) @xmath242 a.s .
\(2 ) for all @xmath231 the increments @xmath243 are independent of @xmath244 and has a poisson law with parameter @xmath245 , that is , @xmath246 where @xmath247 $ ] , @xmath133 , and @xmath248 is the lebesgue measure of @xmath249 .
if we do not specify the filtration , @xmath238 will be the filtration generated by the field itself , completed with the nulls sets of @xmath250 it is known that then there is a simple locally finite point random measure @xmath251 such that for any finite @xmath252 and for any disjoint bounded borel sets @xmath253@xmath254 while @xmath255 a random simple locally finite counting measure @xmath132 is a two - parameter prf iff @xmath256 is a strong martingale .
let @xmath257 and @xmath258 be two independent inverse stable subordinators with indices @xmath259 and @xmath260 which are independent of the poisson field @xmath261 in @xcite it is defined the fractional poisson field ( fprf ) as follows : @xmath262 we obtain the marginal distribution of fprf : for @xmath163 , @xmath263where @xmath37 is given by ( [ d1 ] ) . in other words , for @xmath264 @xmath265where the wright generalized bessel function is defined by ( [ wr ] ) , and @xmath266 is defined by . using the laplace
transform given by ( [ l1 ] ) one can obtain an exact expression for the double laplace transform of ( p2 ) : for @xmath163 , @xmath267 it is known ( see , e.g. , ( * ? ? ?
* theorems 1.1 . and 1.2 ) ) that a single point random field @xmath268 can be regarded as a random set of points and could be characterized by its system of finite - dimensional distributions as @xmath269 for all integers @xmath270 and all compacts @xmath271 .
note that @xmath272 \big ] \notag \\ & = \int_{0}^{\infty } \int_{0}^{\infty } \mathrm{e}n(u_{1},u_{2})f_{\alpha _ { 1}}(t_{1},u_{1})f_{\alpha _ { 2}}(t_{2},u_{2})du_{1}du_{2 } \notag \\ & = \lambda t_{1}^{\alpha _ { 1}}t_{2}^{\alpha _ { 2}}/[\gamma ( 1+\alpha _ { 1})\gamma ( 1+\alpha _ { 2 } ) ] \label{exp1}\end{aligned}\]]and , for @xmath273 , @xmath274 \\ & \times \big[\frac{1}{\gamma ( 1+\alpha _ { 2})\gamma ( \alpha _ { 2})}\int_{0}^{\min ( t_{2},s_{2})}(t_{2}-\tau _ { 2})^{\alpha _ { 2}}+(s_{2}-\tau _ { 2})^{\alpha _ { 2}})\tau _ { 2}^{\alpha _ { 2}-1}d\tau _ { 2}-\frac{(s_{2}t_{2})^{\alpha _ { 2}}}{\gamma ^{2}(1+\alpha
_ { 2})}\big ] \\ & + \big[\frac{(s_{1}t_{1})^{\alpha _ { 1}}}{\gamma ^{2}(1+\alpha _ { 1})}-\frac{2}{\gamma ( 2+\alpha _ { 1})}(\min ( t_{1},s_{1}))^{1+\alpha _ { 1}}\big ] \\ & \times \big[\frac{1}{\gamma ( 1+\alpha _ { 2})\gamma ( \alpha _ { 2})}\int_{0}^{\min ( t_{2},s_{2})}(t_{2}-\tau _ { 2})^{\alpha _ { 2}}+(s_{2}-\tau _ { 2})^{\alpha _ { 2}})\tau _ { 2}^{\alpha _ { 2}-1}d\tau _ { 2}- \\ & \qquad \frac{2}{\gamma ( 2+\alpha _ { 2})}(\min ( t_{2},s_{2}))^{1+\alpha _ { 2}}\big ] + \big[\frac{(s_{2}t_{2})^{\alpha _ { 2}}}{\gamma ^{2}(1+\alpha _ { 2})}-\frac{2}{\gamma ( 2+\alpha _ { 2})}(\min ( t_{2},s_{2}))^{1+\alpha _ { 2}}\big ] \\ & \times \big[\frac{1}{\gamma ( 1+\alpha _ { 1})\gamma ( \alpha _ { 1})}\int_{0}^{\min ( t_{1},s_{1})}(t_{1}-\tau _ { 1})^{\alpha _ { 1}}+(s_{1}-\tau _ { 1})^{\alpha _ { 1}})\tau _ { 1}^{\alpha _ { 1}-1}d\tau _ { 1}- \\ & \qquad \frac{2}{\gamma ( 2+\alpha _ { 1})}(\min ( t_{1},s_{1}))^{1+\alpha _ { 1}}\big]\big\ } + \lambda \frac{(\min ( t_{1},s_{1}))^{\alpha _ { 1}}(\min ( t_{2},s_{2}))^{\alpha _ { 2}}}{\gamma ( 1+\alpha _ { 1})\gamma ( 1+\alpha _ { 2 } ) } ; \end{aligned}\end{gathered}\ ] ] in particular , for @xmath273 , @xmath275where @xmath276 we can summarize our results in the following [ prop:4.2 ] let @xmath277 be a fprf defined by ( [ fprf ] )
. then 1 .
@xmath278 @xmath279 is given by ( [ lp ] ) ; 2 .
@xmath280 and @xmath281 are given by ( [ exp1 ] ) , ( [ var ] ) , ( [ covf ] ) , respectively .
the proof of the theorem is given in the appendix .
note that formulae ( [ exp1]),([var ] ) and ( [ covf ] ) are corrected versions of the formulae ( 26 ) and ( 27 ) of the paper @xcite . following the ideas of this paper
, the hurst index of the fractional poisson random field in @xmath282 can be defined as follows : @xmath283 any random field @xmath284 defined on the positive quadrant @xmath285 can be extended in the whole space @xmath286 in the following way : let @xmath287 be independent copies of the random field @xmath288
then one can define @xmath289therefore , modifying the cadlag property we obtain a poisson like random field @xmath290 which has a similar covariance structure ( replacing @xmath291 by @xmath292 ) .
let @xmath25 , be an @xmath50-stable subordinator , and @xmath293 be its inverse ( @xmath294 ) .
recall that @xmath152 is a cadlag strictly increasing process @xmath201 , while @xmath144 @xmath295 is nondecreasing and continuous . as a consequence ,
the latter defines a random nonnegative measure @xmath296 on @xmath297 such that @xmath298)=y_{\alpha } ( t)$ ] .
the @xmath221-algebra @xmath299 contains all the information given by @xmath296 : @xmath300 now , let @xmath301 be a fpp , where @xmath132 has intensity @xmath47 .
we denote by @xmath302 its natural filtration .
we note that each @xmath298)$ ] is @xmath299-measurable , while @xmath303))$ ] is independent of @xmath304 for any @xmath305)$ ] .
hence , for any bounded @xmath306-measurable random variable @xmath307 , we have @xmath308 } dx_v\big ) & = { \rm e}\big ( y(s ) { \rm e}\big ( \int_0^\infty { { \mathbbm{1}}}_{(s , t ] } n(\mu _ { \alpha } ( dv))\big| \sigma(\mathcal{f}_s^x,\mathcal{g } ) \big ) \big ) \\ & = { \rm e}\big ( y(s ) \int_0^\infty { { \mathbbm{1}}}_{(\mu _ { \alpha } ( [ 0,s]),\mu _ { \alpha } ( [ 0,t ] ) ] } { \rm e } ( dn_w | \sigma(\mathcal{f}_s^x,\mathcal{g } ) ) \big ) \\ & = { \rm e}\big ( y(s ) \lambda \mu _ { \alpha } ( [ s , t ] ) \big ) \\ & = { \rm e}\big ( \int_0^\infty y(s ) { { \mathbbm{1}}}_{(s , t ] } \lambda \mu _ { \alpha } ( dv ) \big ) .\end{aligned}\ ] ] in other words , by ( * ? ? ? * theorem t4 ) , the fpp @xmath136 is a doubly stochastic poisson process with respect to the filtration @xmath309 .
therefore a first characterization of a fpp may be written in the following way .
a process @xmath310 is a fpp iff it is a doubly stochastic poisson process with intensity @xmath311 , with respect to the filtration @xmath309 . in other words , whenever @xmath312 are disjoint bounded borel sets and @xmath313 are non - negative integers , then @xmath314 an analogous result may be found for fprf .
in fact , let @xmath257 and @xmath258 be two independent inverse stable subordinators with indices @xmath259 and @xmath315 .
let @xmath316 and @xmath317 , @xmath318 and @xmath319 as above .
if @xmath320 is the product measure and @xmath321 , we can mimic the same reasoning as above once we have noted that @xmath322),\mu_{\alpha_2}([0,s_{2 } ] ) } n ( w_{1},w_{2 } ) $ ] and @xmath323 are conditionally independent , given @xmath324 .
in fact @xmath325),\mu_{\alpha_2}([0,s_{2 } ] ) } n ( \mu_{\alpha_1}([0,t_{1}]),\mu_{\alpha_2}([0,t_{2 } ] ) ) \big| \sigma(\mathcal{f}^x_{\infty , s_2 } \vee \mathcal{f}^x_{s_1 , \infty } , \mathcal{g } ) ) \big ) \\ & = { \rm e}\big ( \delta_{\mu_{\alpha_1}([0,s_{1}]),\mu_{\alpha_2}([0,s_{2 } ] ) } n ( \mu_{\alpha_1}([0,t_{1}]),\mu_{\alpha_2}([0,t_{2 } ] ) ) \mathcal{g } ) \\ & = \lambda \mu_{\boldsymbol{\alpha } } ( ( ( s_1,s_2),(t_1,t_2 ) ] ) .
\end{aligned}\end{gathered}\ ] ] in other words , with the notation given in @xcite , the fprf @xmath136 is a @xmath326-doubly stochastic poisson process with respect to the filtration @xmath327 by ( * ? ? ?
* theorem 1 ) .
again , we may summarize this result in the following statement .
[ thm : char1 ] a process @xmath328 is a fprf iff it is a @xmath326-doubly stochastic poisson process with intensity @xmath329 , with respect to the filtration @xmath327 . in other words , whenever @xmath312 are disjoint bounded borel sets in @xmath330 and @xmath331 are non - negative integers , then @xmath332 now , let @xmath333 be fixed .
the process @xmath334 is the trace of the fprf along the increasing @xmath201-indexed family of sets @xmath335 $ ] . as a consequence of the previous results ,
we obtain : [ thm : along1and2 ] a random simple locally finite counting measure @xmath328 is a fprf iff @xmath336 are independent , and fixed @xmath337 , the process @xmath338 , conditioned on @xmath318 , is a fpp @xmath339 , the process @xmath340 , conditioned on @xmath319 , is a fpp @xmath341 , and the two processes @xmath342 are conditionally independent given @xmath343 .
assume that @xmath328 is a fprf and @xmath333 fixed .
denote by @xmath344 and note that @xmath345 .
let @xmath312 be disjoint bounded borel sets and @xmath313 non - negative integers .
we have @xmath346 \times b_i)=x_i \ }
\big| & \sigma(\mathcal{g}_1,\sigma ( \{y_{\alpha_{2}}(t),t\geq 0\}))\big ) = { \rm p}\big ( \bigcap_{i=1}^n \ { n_{\alpha _ { 1},\alpha _ { 2}}([0,t_1 ] \times b_i)=x_i \ } \big| \mathcal{g } \big ) \\ & = \prod_{i=1}^n \frac { \exp(-\lambda \mu_{\boldsymbol{\alpha}}([0,t_1 ] \times b_i ) ) ( \lambda \mu_{\boldsymbol{\alpha}}([0,t_1 ] \times b_i))^{x_i } } { x_i ! } \\ & = \prod_{i=1}^n \frac { \exp(-\lambda y_{\alpha_{1}}(t_1 ) \cdot \mu_{\alpha_2 } ( b_i ) ) ( \lambda y_{\alpha_{1}}(t_1 ) \cdot \mu_{\alpha_2 } ( b_i ) ) ^{x_i } } { x_i ! } , \end{aligned}\ ] ] and hence @xmath347 , where , conditioned on @xmath318 , @xmath268 is a poisson process with intensity @xmath348 .
the conditional independence follows by similar arguments , and hence the first implication is proved .
conversely , by @xcite , to prove proposition [ thm : char1 ] it is sufficient to prove .
denote by @xmath349 so that @xmath350 and @xmath351
. then , denoting by @xmath352 the conditional independence of @xmath136 and @xmath21 , given @xmath353 , we have by hypothesis that @xmath354 for any @xmath355 .
thus , * @xmath356 , @xmath357 , @xmath358 , @xmath359 , then @xmath360 and hence @xmath361 * note that @xmath362 , where @xmath363 .
in addition , @xmath364 , and @xmath365 . hence @xmath366 * now , note that both @xmath367 and @xmath368 belong to @xmath369 , while @xmath370 . hence @xmath371 combining , and we finally get : @xmath372 theorem [ thm : along1and2 ] shows an example of characterizations of fprf . in @xcite
, the authors proved a characterization of the inhomogeneous poisson processes on the plane thorough its realizations on increasing families of points ( called increasing path ) and increasing families of sets , called increasing set ( see also @xcite ) .
we are going to characterize an fprf in the same spirit .
let @xmath373 be the collection of the closed rectangles @xmath374 , where @xmath375 .
the family @xmath373 generates a topology of closed sets @xmath376 , which is closed under finite unions and arbitrary intersections , called a _
lower set _ family ( see , e.g. , @xcite ) .
in other words , when a point @xmath377 belongs to a set @xmath378 , all the rectangle @xmath379 is contained in @xmath380 : @xmath381 a function @xmath382 is called an _ increasing set _ if @xmath383 , it is continuous , it is non - decreasing ( @xmath384 ) , and the area it underlies is finite for any @xmath201 and goes to infinity when @xmath201 increases ( @xmath385 ) . note that , for a nonnegative measure @xmath386 on @xmath387 , it is well - defined the non - decreasing right - continuous function : @xmath388given an increasing path @xmath389 and a random nonnegative measure @xmath132 ( in @xcite , it is an increasing and additive process ) , we may define the one - parameter process @xmath390 as the trace of @xmath132 along @xmath389 : @xmath391 [ thm : charl1 ] a random simple locally finite counting measure @xmath328 is a fprf iff , conditioned on @xmath299 , @xmath392 is an inhomogeneous poisson process with intensity @xmath393 , for any increasing set @xmath389 , independent of @xmath299 .
assume that @xmath328 is a fprf .
then , for any @xmath394 , the sets @xmath395 are disjoint . by , @xmath396 = x_i \ } \big| \mathcal{g } \big ) & = { \rm p}\big ( \bigcap_{i=1}^n \ { n_{\alpha _ { 1},\alpha _ { 2}}(b_i)=x_i \ } \big| \mathcal{g } \big ) \\ & = \prod_{i=1}^n \frac { \exp(-\lambda \mu_{\boldsymbol{\alpha}}(b_i ) ) ( \lambda \mu_{\boldsymbol{\alpha}}(b_i))^{x_i } } { x_i!}\\ & = \prod_{i=1}^n \frac { \exp\big(-\lambda \cdot ( \mu_{\boldsymbol{\alpha } } \circ \gamma ) ( s_i , t_i ] \big ) \big(\lambda \cdot ( \mu_{\boldsymbol{\alpha } } \circ \gamma ) ( s_i , t_i ] \big)^{x_i } } { x_i ! } .\end{aligned}\ ] ] conversely ,
note that that may be checked only on disjoint rectangles @xmath397 ( see also @xcite ) . after ordering partially the rectangles with respect to @xmath398 ,
one can build an increasing sets @xmath399 such that @xmath400 where @xmath394 . by hypothesis
, @xmath392 is an inhomogeneous poisson process with intensity @xmath401 .
then , @xmath402 = x_i \ } \big| \mathcal{g } \big ) \\ & = \prod_{i=1}^n \frac { \exp\big(-\lambda \cdot ( \mu_{\boldsymbol{\alpha } } \circ \gamma ) ( s_i , t_i ] \big ) \big(\lambda \cdot ( \mu_{\boldsymbol{\alpha } } \circ \gamma ) ( s_i , t_i ] \big)^{x_i } } { x_i ! } \\ & = \prod_{i=1}^n \frac { \exp(-\lambda \mu_{\boldsymbol{\alpha}}(b_i ) ) ( \lambda \mu_{\boldsymbol{\alpha}}(b_i))^{x_i } } { x_i!}. \qedhere\end{aligned}\ ] ] now , a function @xmath403 is called an _ increasing path _ if @xmath404 , it is continuous , it is non - decreasing ( @xmath405 ) , and the area it underlies goes to infinity ( @xmath406 ) . in other words ,
an increasing path is an increasing set where , for each @xmath201 , @xmath407 is a rectangle . given an increasing path @xmath389 and a process @xmath408 , we define the one - parameter process @xmath390 as the trace of @xmath132 along @xmath389 : @xmath409 when dealing with the laws of the traces of a process along increasing paths , one can not hope to prove , for instance , the conditional independence of two filtrations as @xmath410 and @xmath411 , since the event that belong to those filtrations are generated by the increments of the process on rwgions that are not comparable with respect to the partial order @xmath398 . as an example , there is no increasing path that separates the three rectangles @xmath412 , @xmath413 and @xmath414 and hence we can not give the joint law of @xmath415 and @xmath416 . on the other hand , proposition [ thm : char1 ] suggests that , if we assume the independence of @xmath417 and @xmath418 conditioned on @xmath419 , may be proved for @xmath420 , @xmath421 and @xmath422 via increasing paths ( as in @xcite ) .
this consideration has suggesed the following definition .
we say that the filtration satisfies the conditional independence condition or the cairoli - walsh condition ( ( f4 ) in @xcite , see also @xcite ) if for any @xmath130-measurable integrable random variable @xmath423 and for any @xmath424@xmath425 thus , following the same ideas as in @xcite , one can prove the following result .
[ thm : charl2 ] a random simple locally finite counting measure @xmath328 is a fprf iff , conditioned on @xmath299 , the cairoli - walsh condition holds and @xmath390 is an inhomogeneous poisson process with intensity @xmath426 , for any increasing path @xmath389
. let @xmath427 be a metric space equipped with a radon measure on its borel sets .
we assume existence of an indexing collection @xmath428 on @xmath427 , as it is defined in @xcite .
we are interested to considering processes indexed by a class of closed sets from @xmath427 .
@xmath429 is called an increasing path if it is continuous and increasing : @xmath430 ( called a flow in @xcite ) we can now define set - indexed fractional poisson process . a set - indexed process @xmath431 is called a set - indexed fractional poisson process(sifpp ) , if for any increasing path @xmath389 the process @xmath432 is an fpp . following results of @xcite ,
we can state any sifpp is a set - indexed lvy process .
details and martingale characterizations will be presented elsewhere .
let @xmath433 be a sequence of i.i.d .
@xmath434-uniform distributed random variables , independent of the processes @xmath435 , @xmath436 .
the random indexes associated to the ` records ' @xmath437 are inductively defined by @xmath438 it is well known that @xmath439 , and hence the @xmath440-th record @xmath441 of the sequence is well defined : @xmath442 , @xmath443 . since @xmath444 , then @xmath445 . moreover , the number of @xmath446 s that realize the maximum by time @xmath447 is almost surely asymptotic to @xmath448 as @xmath449 . in other words ,
the sequence @xmath450 growths exponentially fast . now , given a increasing set @xmath399 , we define @xmath451 a random simple locally finite counting measure @xmath328 is a fprf iff @xmath392 is distributed as @xmath452 , for any increasing set @xmath399.[thm : gy - c ] in the proof we assume that @xmath453 almost surely . when this is not the case , the proof should be changed as in @xcite , where generalized random variables are introduced exactly when @xmath454 . by theorem [ thm : charl1 ]
, we must prove that , conditioned on @xmath299 , @xmath455 is an inhomogeneous poisson process with intensity @xmath401 .
conditioned on @xmath456 , let @xmath457 be the continuous deterministic cumulative distribution function .
let @xmath458 be its pseudo - inverse , and define @xmath459 , for each @xmath447 .
then @xmath460 is a sequence of i.i.d .
random variables with cumulative function @xmath461 . as in @xcite , put @xmath462 , ( @xmath463 ) omitting in the increasing sequence @xmath464 all the repeating elements except one , we come to the strictly increasing sequence ( * ? ? ?
* eq . ( 3 ) )
@xmath465 now , since @xmath458 is monotone , it is obvious by definition that @xmath466 .
again , @xmath458 is monotone , and hence @xmath467 that is the process @xmath468 defined in ( * ? ? ? * eq .
the thesis is now an application of ( * ? ? ?
* theorem 1 ) and theorem [ thm : charl1 ] .
the process @xmath469 may be used to reparametrize the time of the increasing paths and sets .
in fact , for any increasing path @xmath470 , let @xmath471 be the first time that the intensity is seen to be bigger than @xmath472 on the increasing path , and define @xmath473 the reparametrization of @xmath399 made by @xmath469 .
analogously , for any increasing set @xmath399 , let @xmath474 we note that , for any fixed @xmath472 and @xmath378 @xmath475 where @xmath476 we recall that a random measurable set @xmath477 is called a @xmath478-stopping set if @xmath479 for any @xmath480 . as a consequence , the reparametrization given in transforms @xmath481 into @xmath482 , a family of continuous increasing stopping set by .
such a family is called an optional increasing set . the random time change theorem ( which can be made an easy consequence of the characterization of the poisson process given in @xcite ) together with theorem [ thm : charl1 ] and theorem [ thm : charl2 ]
give the following corollaries , that can be seen as extensions of some results in @xcite . a random simple locally finite counting measure @xmath328 is a fprf iff , conditioned on @xmath299 , @xmath483 is a standard poisson process , for any increasing set @xmath389 . a random simple locally finite counting measure @xmath328 is a fprf iff , conditioned on @xmath299 , the cairoli - walsh condition holds @xcite and @xmath484 is a standard poisson process , for any increasing path @xmath389 .
a direct calculation may be applied to show that the marginal distribution of the classical poisson random field @xmath485@xmath486satisfy the following differential - difference equations:@xmath487and the initial conditions:@xmath488 we are now ready to derive the governing equations of the marginal distributions of fprf @xmath489@xmath490given by ( [ p2 ] ) or ( [ lp ] ) .
these equations have something in common with the governing equations for the non - homogeneous fractional poisson processes @xcite . for a function
@xmath491 the caputo - djrbashian mixed fractional derivative of order @xmath492 is defined by@xmath493 assuming that @xmath494is integrable as function of four variables @xmath495,@xmath496 the double laplace transform@xmath497where @xmath498 is the double laplace transform of the function @xmath499 .
note that the laplace transform of @xmath500 given by ( [ l1 ] ) as @xmath501 is of the form @xmath502 and its inverse is the delta distribution @xmath503 .
accordingly , as @xmath504 , @xmath500 converges weakly to @xmath503 , and we denote it by @xmath505 . the proof of ( [ 5.6 ] ) is standard and we omit it ( see @xcite for the one - dimensional case ) .
let @xmath506 be the fprf defined by ( [ fprf ] ) .
* 1 ) * then its marginal distribution given in ( [ md ] ) satisfy the following fractional differential - integral recurrent equations:@xmath507 f_{\alpha _ { 1}}(t_{1},x_{1})f_{\alpha _ { 2}}(t_{2},x_{2})dx_{1}dx_{2 } ; \label{5.8 } \\ \mathrm{d}_{t_{1},t_{2}}^{\alpha _
{ 1},\alpha _ { 2}}p_{k}^{\alpha _ { 1},\alpha _ { 2}}\left ( t_{1},t_{2}\right ) & = \int_{0}^{\infty } \int_{0}^{\infty } \big [ \left ( -\lambda + \lambda
^{2}x_{1}x_{2}\right ) p_{k}^{\alpha _ { 1},\alpha _ { 2}}\left ( x_{1},x_{2}\right ) \notag \\ & \qquad + \left ( \lambda -2\lambda ^{2}x_{1}x_{2}\right ) p_{k-1}^{\alpha _ { 1},\alpha _ { 2}}\left ( x_{1},x_{2}\right ) + \lambda ^{2}x_{1}xp_{k-2}^{\alpha _
{ 1},\alpha _ { 2}}\left ( x_{1},x_{2}\right ) \big ] \notag \\ & \qquad \qquad \times f_{\alpha _ { 1}}(t_{1},x_{1})f_{\alpha _ { 2}}(t_{2},x_{2})dx_{1}dx_{2 } , \qquad\qquad\qquad\qquad k\geq 2 ; \label{5.9}\end{aligned}\ ] ] with the initial conditions:@xmath508 * 2 ) * for @xmath509 hence ( [ 5.7 ] ) , ( [ 5.8 ] ) and ( [ 5.9 ] ) become ( [ 5.1 ] ) , ( [ 5.2 ] ) and ( [ 5.3 ] ) correspondingly . *
1 ) * the initial conditions are easily checked using the fact that @xmath510 a.s .
let @xmath511 , be defined as in equations ( [ p2 ] ) or ( [ lp ] ) .
then the characteristic function of the fprf , for @xmath512:@xmath513 taking the double laplace transform of ( [ 5.11 ] ) and using ( [ l1 ] ) and ( [ l2 ] ) yields@xmath514and@xmath515 using an integration by parts for a double integral @xcite:@xmath516we get from ( [ 5.6 ] ) , ( [ 5.12 ] ) and ( [ 5.12 ] ) with @xmath517 .\end{aligned}\]]thus@xmath518 using ( [ 5.6 ] ) , ( [ l1 ] ) we can invert the double laplace transform as follows:@xmath519and finally , by inverting the characteristic function ( [ 5.11 ] ) , we obtain @xmath520 f_{\alpha _ { 1}}(t_{1},x_{1})f_{\alpha _ { 2}}(t_{2},x_{2})dx_{1}dx_{2}.\ ] ] using ( [ 5.1 ] ) , ( [ 5.2 ] ) and ( [ 5.3 ] ) we arrive to ( [ 5.7 ] ) , ( [ 5.8 ] ) and ( [ 5.9 ] ) correspondingly . *
2 ) * finally , as @xmath521 we have @xmath522 , and their laplace inversions are delta function : @xmath523 .
thus , 2 ) is proven .
in this section we show some simulations of fprf made with matlab based on the @xmath50-stable random number generator function ` s`tblrnd . 0.45 0.45 + 0.45 0.45 the subordinators @xmath34 are simulated exactly at times @xmath524 , where @xmath525 till they reach a defined value @xmath526 .
more precisely , @xmath527 where @xmath136 is independently simulated with ` s`tblrnd(@xmath50 , 1 , @xmath528{\delta}$ ] , 0 ) .
accordingly , @xmath529{\delta})^{\alpha } \ } = \exp \{-\delta s^{\alpha } \ } , \qquad s\geq 0,\ ] ] and hence @xmath530 the simulation of the inverse stable subordinators @xmath531 $ ] are thus made at times @xmath532 with values @xmath533 . to simulate a fprf @xmath534 on the window @xmath535
, we first simulate two independent inverse stable subordinators @xmath536 and @xmath537 . by proposition
[ thm : char1 ] , the value of @xmath538 on each rectangle @xmath539 is a poisson random variable with mean @xmath540 . as @xmath541 , we approximate it with a bernoulli random variable @xmath21 of parameter @xmath540 . when @xmath542 , we add a point at random inside the rectangle . in figure [ fig:1 ] the simulations of the inverse stable subordinators @xmath543 and @xmath544 and the corresponding fprf @xmath538 for different values of @xmath545 and @xmath546 are shown .
the simulations of @xmath538 are plotted twice : we have rotated each figure in order to underline the spatial dependence of the spread of the points of the process @xmath538 in connection with the marginal intensities @xmath543 and @xmath544 .
for example , in figure [ fig:1](c ) two different marginal distribution are expected since @xmath547 and @xmath548 .
while @xmath549 produces a quite uniform distribution of points , @xmath550 generates clusters in correspondence of its steeper slopes . with a montecarlo procedure
, we also compute @xmath551 for different values of @xmath552 and @xmath546 .
in fact , in , we approximate @xmath553 with @xmath554 where @xmath555 and @xmath556 are independent sequences of i.i.d
. distributed as @xmath557 and @xmath558 , respectively .
the simulations with @xmath559 are presented in figure [ fig:2 ] .
we underline the variety of the shape of distributions that can be generated with this two - parameter model in addition to its flexibility to include , for example , different cluster phenomena .
0.45 0.45 + 0.45 0.45 * acknowledgements . *
n. leonenko and e. merzbach wish to thank g. aletti for two visits to university of milan .
n. leonenko was supported in particular by cardiff incoming visiting fellowship scheme and international collaboration seedcorn fund and australian research council s discovery projects funding scheme ( project number dp160101366 ) .
the authors thank anonymous reviewers , for their helpful comments .
in this appendix , we prove a general result that can be used to compute the covariance structure of the parameter - changed poisson random field : @xmath560 where @xmath561 and @xmath562 are independent non - negative non - decreasing stochastic processes , in general non - markovian with non - stationary and non - independent increments , and @xmath563 is a prf with intensity @xmath133 .
we also assume that @xmath564 and @xmath565 are independent of @xmath566 \2 ) if @xmath564 and @xmath565 have second moments , so does @xmath572 and @xmath573 ^{2}\left\ { \mathrm{e}y_{1}^{2}(t_{1})\mathrm{e}y_{2}^{2}(t_{2})-\left ( \mathrm{e}y_{1}^{2}(t_{1})\right ) ^{2}\left ( \mathrm{e}y_{2}^{2}(t_{2})\right ) ^{2}\right\ } \\ + \mathrm{var}n(1,1)\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{2}(t_{2 } ) \end{gathered}\ ] ] and its covariance function @xmath574 for @xmath575 is given by : @xmath576 ^{2}\left\ { \mathrm{cov}\left ( y_{1}(t_{1}),y_{1}(s_{1})\right ) \mathrm{cov}\left ( y_{2}(t_{2}),y_{2}(s_{2})\right ) + \right .
\notag \label{3.4 } \\ & \qquad+\left ( \mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})-\mathrm{e}y_{2}^{2}(s_{2})\right ) \left ( \mathrm{cov}\left ( y_{1}(t_{1}),y_{1}(s_{1})\right ) -\mathrm{e}y_{1}^{2}(s_{1})\right ) \notag \\ & \qquad+\left ( \mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})-\mathrm{e}y_{1}^{2}(s_{1})\right ) \left ( \mathrm{cov}\left ( y_{2}(t_{2}),y_{2}(s_{2})\right ) -\mathrm{e}y_{2}^{2}(s_{2})\right ) \ } \notag \\ & \qquad+\mathrm{var}n(1,1)\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2 } ) \notag\end{aligned}\ ] ] and for any @xmath577 and @xmath578 from @xmath579 @xmath576 ^{2}\big\ { \mathrm{cov}\left ( y_{1}(t_{1}),y_{1}(s_{1})\right ) \mathrm{cov}\left ( y_{2}(t_{2}),y_{2}(s_{2})\right ) \\ & \quad + \left ( \mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})-\mathrm{e}y_{2}^{2}(\min ( s_{2},t_{2}))\right ) \left ( \mathrm{cov}\left ( y_{1}(t_{1}),y_{1}(s_{1})\right ) -\mathrm{e}y_{1}^{2}(\min ( s_{1},t_{1})\right ) \\ & \quad+\left ( \mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})-\mathrm{e}y_{1}^{2}(\min ( s_{1},t_{1})\right ) \left ( \mathrm{cov}\left ( y_{2}(t_{2}),y_{2}(s_{2})\right ) -\mathrm{e}y_{2}^{2}(\min ( s_{2},t_{2})\right ) \big\ } \\ &
\quad+\mathrm{var}n(1,1)\mathrm{e}y_{1}(\min ( s_{1},t_{1}))\mathrm{e}y_{2}(\min ( s_{2},t_{2})).\end{aligned}\ ] ] these formulae are valid for any lvy random field @xmath580,@xmath581 , with finite expectation @xmath582 and finite variance @xmath583 for prf @xmath584 and to apply these formulae one needs to know @xmath585 and @xmath586 which are available for many non - negative processes @xmath587 and @xmath588 induction inverse subordinators . we know that for a prf @xmath594 @xmath595 ^{2}=\mathrm{e}n(1,1 ) \left ( t_{1}-s_{1}\right ) \left ( t_{2}-s_{2}\right ) + \left [ \mathrm{e}n(1,1)\left ( t_{1}-s_{1}\right ) \left ( t_{2}-s_{2}\right ) \right ] ^{2}.\ ] ] for the variance
, we have @xmath597 g_{t_{1}}^{(1)}(du_{1})g_{t_{2}}^{(2)}(du_{2 } ) \\ & & -\left ( \mathrm{e}n(1,1)\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{2}(t_{2})\right ) ^{2 } \\ & = & \left ( \mathrm{e}n(1,1)\right ) ^{2}\mathrm{e}y_{1}^{2}(t_{1})\mathrm{e}y_{2}^{2}(t_{2})+\mathrm{var}n(1,1)\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{2}(t_{2 } ) \\ & & -\left ( \mathrm{e}n(1,1)\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{2}(t_{2})\right ) ^{2 } \\ & = & \left ( \mathrm{e}n(1,1)\right ) ^{2}\left\ { \mathrm{e}y_{1}^{2}(t_{1})\mathrm{e}y_{2}^{2}(t_{2})-(\mathrm{e}y_{1}(t_{1}))^{2}(\mathrm{e}y_{2}(t_{2}))^{2}\right\ } \\ & & -\mathrm{var}n(1,1)\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{2}(t_{2}).\end{aligned}\ ] ] then @xmath599 ^{2}s_{1}s_{2}+\mathrm{e}\left\ { n(t_{1},s_{2})-n(t_{1},s_{1})+n(s_{1},s_{2})\right\ } n(s_{1},s_{2 } ) \\ & \quad+\mathrm{e}\left\ { n(t_{2},s_{1})-n(s_{1},s_{2})+n(s_{1},s_{2})\right\ } n(s_{1},s_{2})-\mathrm{e}n^{2}(s_{1},s_{2 } ) \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left ( t_{1}-s_{1}\right ) \left ( t_{2}-s_{2}\right ) s_{1}s_{2}+\mathrm{e}n^{2}(s_{1},s_{2})+\mathrm{e}n^{2}(s_{1},s_{2})-\mathrm{e}n^{2}(s_{1},s_{2 } ) \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left ( t_{1}-s_{1}\right ) \left ( t_{2}-s_{2}\right ) s_{1}s_{2}+\mathrm{var}n^{2}(s_{1},s_{2})+\left [ \mathrm{e}n(s_{1},s_{2})\right ] ^{2 } \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left ( t_{1}-s_{1}\right ) \left ( t_{2}-s_{2}\right ) s_{1}s_{2}+\mathrm{var}n(1,1)s_{1}s_{2}+\left [ \mathrm{e}n(1,1)\right ] ^{2}s_{1}^{2}s_{2}^{2 } \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left [ \left ( t_{1}-s_{1}\right ) \left ( t_{2}-s_{2}\right ) s_{1}s_{2}+s_{1}^{2}s_{2}^{2}\right ] + s_{1}^{2}s_{2}^{2}\mathrm{var}n(1,1).\end{aligned}\ ] ] since the processes @xmath601 are independent , a conditioning argument yields @xmath602 ^{2}\left\ { \int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } u_{1}u_{2}v_{1}v_{2}g_{t_{1},s_{1}}^{(1)}(du_{1},dv_{1})g_{t_{2},s_{2}}^{(2)}(du_{2},dv_{2})\right . \\ & \qquad -\int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } u_{1}v_{1}v_{2}^{2}g_{t_{1},s_{1}}^{(1)}(du_{1},dv_{1})g_{t_{2}}^{(2)}(du_{2 } ) g_{s_{2}}^{(2)}(dv_{2 } ) \\ & \qquad \ -\int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } u_{2}v_{1}^{2}v_{2}g_{t_{2},s_{2}}^{(1)}(du_{2},dv_{2})g_{t_{1},}^{(1)}(du_{1 } ) g_{s_{1 } } ^{(1)}(dv_{1 } ) \\ & \qquad + 2\left .
\int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{2\infty } v_{1}^{2}v_{2}^{2}g_{t_{1},}^{(1)}(du_{1})g_{s_{1 } } ^{(1)}(dv_{1})g_{t_{2 } } ^{(2)}(du_{2})g_{s_{2 } } ^{(2)}(dv_{2})\right\ } \\ & \qquad + \mathrm{var}n(1,1)\int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } \int_{0}^{\infty } v_{1}v_{2}g_{t_{1 } } ^{(1)}(du_{1})g_{s_{1 } } ^{(1)}(dv_{1})g_{t_{2 } } ^{(2)}(du_{2})g_{s_{2 } } ^{(2)}(dv_{2 } ) \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left\ { \mathrm{e}y_{1}(t_{1})y_{1}(s_{1})\mathrm{e}y_{2}(t_{2})y_{2}(s_{2})-\mathrm{e}y_{1}(t_{1})y_{1}(s_{1})\mathrm{e}y_{2}^{2}(s_{2})\right .
\\ & \qquad -\mathrm{e}y_{2}(t_{2})y_{2}(s_{2})\mathrm{e}y_{2}^{2}(s_{1 } ) + \left .
2\mathrm{e}y_{2}^{2}(s_{1})y_{2}^{2}(s_{2})\right\ } + \mathrm{var}n(1,1)\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2}),\end{aligned}\ ] ] and @xmath603 \\ & = \mathrm{e}n(y_{1}(t_{1}),y_{2}(t_{2}))n(y_{1}(s_{1}),y_{2}(s_{2}))-\left [ \mathrm{e}n(1,1)\right ] ^{2}\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2 } ) \\ & \quad= \left [ \mathrm{e}n(1,1)\right ] ^{2}\left\ { \mathrm{e}y_{1}(t_{1})y_{1}(s_{1})\mathrm{e}y_{2}(t_{2})y_{2}(s_{2})-\mathrm{e}y_{1}(t_{1})y_{1}(s_{1})\mathrm{e}y_{2}^{2}(s_{2})-\mathrm{e}y_{2}(t_{2})y_{2}(s_{2})\mathrm{e}y_{2}^{2}(s_{1})\right .
\\ & \quad + \left .
2\mathrm{e}y_{1}^{2}(s_{1})\mathrm{e}y_{2}^{2}(s_{2})-\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})\right\ } + \mathrm{var}n(1,1)\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2 } ) \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left\ { \left [ \mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))+\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})\right ] \left [ \mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))+\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})\right ] \right .
\\ & \quad -\left [ \mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))+\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})\right ] \mathrm{e}y_{2}^{2}(s_{2})- [ \mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2 } ) ) \\ & \quad+\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2 } ) ] \mathrm{e}y_{1}^{2}(s_{1 } ) \\ & \quad + \left .
2\mathrm{e}y_{1}^{2}(s_{1})\mathrm{e}y_{2}^{2}(s_{2})-\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})\right\ } + \mathrm{var}n(1,1)\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2 } ) \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left\ { \mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))\mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))+\mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})\right .
\\ & \quad + \mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1 } ) \\ & \quad -\mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))\mathrm{e}y_{2}^{2}(s_{2})-\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}^{2}(s_{2 } ) \\ & \quad-\mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))\mathrm{e}y_{1}^{2}(s_{1})-\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})\mathrm{e}y_{1}^{2}(s_{1 } ) \\ & \quad + \left .
2\mathrm{e}y_{1}^{2}(s_{1})\mathrm{e}y_{2}^{2}(s_{2})-\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})\right\ } + \mathrm{var}n(1,1)\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2 } ) \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2 } \ { \mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))\mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))+\mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1 } ) ) [ \mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2 } ) \\ & \quad-\mathrm{e}y_{2}^{2}(s_{2 } ) ] + \mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))\left [ \mathrm{e}y_{1}(t_{1})\mathrm{e}y_{2}(s_{2})-\mathrm{e}y_{1}^{2}(s_{1})\right ] \\ & \quad + \mathrm{e}y_{1}^{2}(s)(\mathrm{e}y_{2}^{2}(s_{2})-\mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2 } ) ) \\ & \quad + \left .
\mathrm{e}y_{2}^{2}(s_{2})(\mathrm{e}y_{1}^{2}(s_{1})-\mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1}))\right\ } + \mathrm{var}n(1,1)\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2 } ) \\ & = \left [ \mathrm{e}n(1,1)\right ] ^{2}\left\ { \mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))\mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))\left [ \mathrm{e}y_{2}(t_{2})\mathrm{e}y_{2}(s_{2})-\mathrm{e}y_{2}^{2}(s_{2})\right ] \right . \\ & \quad \times \left ( \mathrm{cov}(y_{1}(t_{1}),y_{1}(s_{1}))-\mathrm{e}y_{1}^{2}(s_{1})\right ) \\ & \quad+\left ( \mathrm{e}y_{1}(t_{1})\mathrm{e}y_{1}(s_{1}))-\mathrm{e}y_{1}^{2}(s_{1})\right ) \left ( \mathrm{cov}(y_{2}(t_{2}),y_{2}(s_{2}))-\mathrm{e}y_{2}^{2}(s_{2})\right)\ } \\ & \quad + \mathrm{var}n(1,1)\mathrm{e}y_{1}(s_{1})\mathrm{e}y_{2}(s_{2}).\end{aligned}\ ] ] | we present new properties for the fractional poisson process and the fractional poisson field on the plane . a martingale characterization for fractional poisson processes
is given .
we extend this result to fractional poisson fields , obtaining some other characterizations .
the fractional differential equations are studied .
we consider a more general mixed - fractional poisson process and show that this process is the stochastic solution of a system of fractional differential - difference equations .
finally , we give some simulations of the fractional poisson field on the plane . |
recently there has been increasing interest in the point to point resistance of finite resistor networks . for an up to date list of references
see [ 1 ] .
wu [ 2 ] expressed the resistance of a finite graph in terms of the eigenvalues and eigenvectors of the kirchhoff matrix [ 3 ] avoiding the zero eigenvalue .
he also used his method to obtain the resistance of a rectangular grid with various boundary conditions .
recently chair [ 4],[5 ] has used wu s formula to obtain exact formulae for the complete graph minus n edges and the n - cycle graph with first and second neighbours .
essam and wu [ 6 ] used one of the results of [ 2 ] to find an asymptotic expansion for the resistance between opposite corners of a rectangular network with free boundary conditions .
this was extended by izmailian and huang [ 1 ] to other boundary conditions .
izmailian , kenna and wu [ 7 ] have recently introduced a modification of a the method of wu [ 2 ] .
their new method also avoids the zero eigenvalue of the kirchhoff matrix by expressing the resistance in terms of the eigenvalues and eigenvectors of a principal cofactor .
this method will be referred to here as the ikw method .
current interest lies in rectangular networks with zero resistance along one edge . here the case of free boundary conditions along the other three edges will be considered .
contracting the zero resistance branches leads to a network which can be displayed as a fan ( see figure 1 ) .
the @xmath2 fan will be supposed to have n radial lines and m transverse arcs . the main topic of this article is the derivation of the resistance between two nodes of a fan network in general position . the result is shown in equation ( 2 ) .
the special case with one node at the apex and the other on the boundary has appeared in [ 8 ] .
tan and coworkers [ 9 ] , [ 10 ] , [ 11 ] introduced a cobweb network which has periodic boundary conditions in the direction parallel to the zero resistance edge .
the case @xmath4 of an @xmath2 cobweb network is shown in figure 2 .
the cobweb network is related to the fan network by joining corresponding nodes on the first and last radial lines of the fan .
the formula for the symmetric case of the fan with n odd and both input and output nodes on the central radial line yields the result for the cobweb with both nodes on any single radial line .
the cobweb network may also be thought of as a network embedded on the surface of a cylinder with resistors along the length and round the circumference and one end closed .
izmailian , kenna and wu [ 7 ] used their kirchhoff matrix method to obtain the resistance of the cobweb network between two nodes in general position .
the ikw method requires the solution of two eigenvalue problems and results in a double summation . for the fan resistance considered here we use the method of tan , zhou and yang [ 10 ] which requires the solution of just one eigenvalue problem together with the solution of a linear recurrence relation with constant coefficients .
the final formula ( 2.2 ) involves only a single summation . by a simple change of boundary condition the cobweb formula of [ 7 ]
is reproduced .
tan , zhou and yang [ 10 ] considered the resistance of an @xmath2 cobweb network between a node at the center and a node on the boundary . on the basis of results for general n and m = 1 , m = 2 [ 9 ] and m = 3 [ 10 ] they conjectured the result for general m and n ( see section 2.2.1 ) .
the case m = 4 was proved by tan , zhou and lou [ 11 ] .
the conjecture for general m and n has recently been proved by izmailian , kenna and wu [ 7 ] as a special case of their formula for the resistance between an arbitrary pair of nodes on the cobweb .
further analysis was required to reduce their double summation to the single summation formula of the conjecture . a proof of the above conjecture for odd n by the present authors [ 8 ] results from restricting the calculation to the resistance between the apex and a general node on the boundary of the fan . as noted above the resistance when the node of the fan is in the midpoint of the boundary
is the same as that of the cobweb .
the method used was a simplification of the one used here .
the proof [ 8 ] was shorter than that of izmailian et al [ 7 ] due to the simpler method and restricted domain .
formulae for both the fan ( 2 ) and cobweb network ( 6 ) with input and output nodes in general position are presented in section 2 .
also simplifications of the formulae are considered when both nodes are on the same radial or transverse line . in section 3 the details of the method
are given in the context of the fan .
section 4 gives the change in boundary condition required to obtain the cobweb formula .
the results are discussed in section 5 where the method used here [ 10 ] is summarised and compared with the ikw method [ 7 ] .
in general the fan has @xmath6 radial lines labelled from @xmath7 to @xmath8 .
its resistance @xmath9 will be determined between two nodes , one , @xmath10 , distant @xmath11 up the radial line @xmath12 and the other , @xmath13 , distant @xmath14 up the radial line @xmath15 ( see figure 1 ) .
let @xmath16 , @xmath17 and @xmath18 .
further suppose that @xmath19 are the greater and lesser solutions of @xmath20 with @xmath21 the result is @xmath22 where @xmath23 and @xmath24 for the cobweb network the @xmath7 and @xmath8 lines are joined by additional branches to create cylindrical boundary conditions .
the result is independent of @xmath25 and @xmath26 and depends only on @xmath14 , @xmath11 and the transverse distance @xmath27 between @xmath13 and @xmath10 as expected . in section 4 the fan proof is simply extended to the cobweb by changing only the boundary condition with the result @xmath28 .
\end{aligned}\ ] ] equation ( 6 ) is equivalent to equation ( 30 ) of izmailian , kenna and wu [ 7 ] .
their summation runs from @xmath29 to @xmath30 and in their notation @xmath31 , @xmath32 and @xmath33 .
@xmath34 so that @xmath35 and hence @xmath9 and @xmath36 are determined by the resistance ratio and @xmath37 apart from the parameters @xmath38 .
the formulae may be simplified in the following two special cases which for the cobweb already appear in [ 7 ] but with different notation . without loss of generality
we take the line to be @xmath39 and set @xmath40 . in this case
@xmath41 with the result @xmath42 ^ 2 \frac{\cosh[(2s+1)l_i]\cosh[(2t+1)l_i]}{\sinh[2l_i]\sinh[(2n l_i ] } .\ ] ] note : @xmath43 means that @xmath44 and @xmath45 are both on the @xmath46 radial line but should not be taken to mean @xmath47 as the notation suggests .
if @xmath44 is the apex then @xmath48 and @xmath49 .
setting @xmath50 in ( 6 ) also gives the the resistance between two nodes on the same radial line for the cobweb .
@xmath51 ^ 2 \frac{\coth[nl_i]}{\sinh[2l_i ] } , \ ] ] which is equal to @xmath52 .
this is because in this symmetric case of the fan the potentials on the left and right radial boundaries are equal .
thus when the nodes on the left radial line are joined to the corresponding nodes on the right radial line no current flows through the connecting branches .
the resistance of the cobweb network so formed is therefore the same as for the fan .
tan et al [ 10 ] conjectured that the resistance from center to the boundary of the cobweb network is given by @xmath53 where @xmath54 $ ] .
since @xmath55 and @xmath56 , @xmath57^{2}=(\cos\theta_{i})^{2}$ ] .
thus from ( 10 ) @xmath58}{\sinh[2l_i ] } .\ ] ] but @xmath59 and @xmath60 which proves the conjecture .
this is the proof of izmailian , kenna and wu [ 7 ] .
setting @xmath61 implies @xmath62 so that the numerator of the summand in ( 2 ) becomes @xmath63 .
further if we set @xmath64 the distance between the input and output nodes is @xmath65 . in this case @xmath66 + \sinh(dl_i ) \cosh [ 2(t - s)l_i]).\ ] ] if further @xmath67 then the input and output nodes are symmetrically placed relative to the radial line boundaries and @xmath68}{\sinh(2l_i)\cosh(nl_i ) } c_i(y)^2 .\ ] ] setting @xmath61 in ( 3 ) @xmath69}{\sinh(2l_i)\sinh(nl_i ) } c_i(y)^2 .\ ] ] in this case @xmath70 since the symmetry is destroyed by the net current flow from source to sink in the transverse direction .
firstly the formula for the fan will be derived and that for the cobweb will follow immediately by a simple change of boundary condition . to find @xmath9 we inject current @xmath71 at node @xmath10 and remove it at @xmath13 .
let @xmath72 be the resulting current in the @xmath73 resistor from the edge of the @xmath74 radial line flowing towards the center ( see figure 3 ) .
using ohm s law the potential difference may be measured along a path from @xmath10 to @xmath75 and then to @xmath13 with the result @xmath76 to determine the radial currents consider the voltage loop @xmath77 , shown in figure 1 , centered on the @xmath73 resistor of the @xmath74 radial line .
if current @xmath71 enters at the node of height @xmath78 on the radial line @xmath79 charge conservation gives @xmath80 when @xmath81 in ( 17 ) , @xmath82 .
the sum of the voltage differences round the loop is zero so using ohm s law @xmath83 where @xmath84 for @xmath85 and is zero for @xmath86 . combining these equations @xmath87 where @xmath88 , @xmath89 .
equation ( 20 ) may be written in matrix form @xmath90i_k - i_{k-1 } -hj\delta_{k , x}\epsilon_{i , y } , \ ] ] where @xmath91 is an @xmath1-dimensional unit matrix , @xmath92 is a column matrix with @xmath73 element @xmath93 and @xmath94 _ the boundary conditions _ for @xmath8
we only use the loop @xmath95 in figure 3 to obtain the boundary equations @xmath96 or in matrix form @xmath97i_t\end{aligned}\ ] ] with a similar equation for @xmath7 .
@xmath98 has eigenvalues @xmath99 and eigenvectors @xmath100 .
the @xmath101 component of @xmath102 is given by [ 7 ] @xmath103 = \sin(2j\theta_i ) .\ ] ] let @xmath104 be the matrix with @xmath73 row @xmath102 and define @xmath105 .
inverting this relation and using ( 16 ) @xmath106 where @xmath107}{\sin \theta_i } .\ ] ] multiplying ( 21 ) on the left by @xmath104 , noting that @xmath108 is diagonal with diagonal elememts @xmath109 , and taking the @xmath73 component @xmath110 where @xmath111 and @xmath112 .\ ] ] applying @xmath104 to ( 25 ) and taking the @xmath73 component @xmath113 for @xmath114 the general solution of ( 29 ) is a linear combination of @xmath115 and @xmath116 where where @xmath117 and @xmath118 are solutions of ( 1 ) in terms of which @xmath119 and @xmath120 .
the coefficients depend on the region .
@xmath121 matching the solutions at @xmath12 and @xmath15 @xmath122 subsituting in the boundary equations ( 31 ) @xmath123 the final two relations arise from the @xmath12 and @xmath15 radial lines where the current @xmath71 is input and output . using ( 29 ) with @xmath124 and secondly with @xmath125 , in the second case replacing @xmath71 by @xmath126 @xmath127 solving equations ( 35 ) , ( 36 ) and ( 37 ) for @xmath128 and @xmath129 and substituting in ( 32 ) gives @xmath130}{(\lambda_i-\bar\lambda_i)d_i}\quad \hbox{and}\quad x_{-p}(i ) = -\frac { hj [ \gamma\zeta_i(y_p)-\beta\zeta_i(y_q)]}{(\lambda_i-\bar\lambda_i)d_i } , \ ] ] where @xmath131 .
@xmath132 substituting ( 38 ) into ( 27 ) gives the required result ( 2 ) .
all that needs to be done to obtain the cobweb formula is to replace equations ( 36 ) by @xmath133 solving equations ( 35 ) , ( 37 ) , ( 42 ) and ( 43 ) for @xmath128 and @xmath129 and substituting in ( 32 ) gives ( 38 ) where now , using @xmath134 , @xmath135 and @xmath136 substituting ( 38 ) , with these values of @xmath137 and @xmath138 , into ( 27 ) gives the required result ( 6 )
the method of tan , zhou and yang [ 10 ] has been used to derive a formula ( 2 ) for the resistance of a fan network ( see fig.1 ) .
the formula is for the resistance between two nodes in general position . by changing only the boundary condition ( 36 )
we also reproduce the result ( 6 ) of izmailian , kenna and wu [ 7 ] for the related cobweb network .
their method is quite different from ours in that it requires the diagonalisation of two matrices instead of one .
also significant further work is required to reduce their formula from two to a single summation .
a further advantage of the present method is that the current distribution is also determined ( see for example figure 4 ) .. thus @xmath139 where @xmath140 is given by ( 32),(33 ) and ( 34 ) .
the coefficients are determined at the same time as solving for @xmath128 and @xmath129 .
when both input and output nodes are on the same radial line and @xmath67 it is found that the fan and cobweb networks have the same resistance .
the two networks are related by connecting corresponding nodes in the radial boundaries .
it is argued that because of the symmetry there will be no flow in the connecting bonds and hence the resistance equality .
the method of tan et al [ 10 ] used here splits the derivation into three parts .
the first creates a matrix relation between the current distributions on three successive radial lines .
the second part diagonalises the matrix relation to produce a recurrence relation involving only variables on the same transverse line .
the relation is second order with constant coefficients and is solved in the third part .
basically the method reduces the problem from two dimensions to one dimension .
equation ( 20 ) can be written in the form @xmath141 which when @xmath114 reduces to the discrete laplace equation .
the method described here is therefore applicable to any problem involving the laplacian .
this work is supported by jiangsu province education science plan project ( no .
d/2013/01/048 ) , the research project for higher education research of nantong university ( no .
2012gj003 ) .
11 izmailian n sh and huang m - c 2010 expansion of the resistance between two maximally separated nodes on an @xmath142 by @xmath143 resistor network , phys . rev .
e * 82 * 011125 wu f y 2004 theory of resistor networks : the two - point resistance , j. phys . a : math
. gen . * 37 * 6653 kirchhoff g 1847 ann
. phys . chem . *
72 * 497 - 508 chair n 2012 exact two - point resistance and simple random walk on the complete graph minus n edges , ann .
phys . * 327 * no.12 , 2899 - 3206 chair n 2014 the effiective resistance of the n - cycle graph with four nearest neighbours , j. stat . phys .
* 154 * no.4 , 1177 - 1190 essam j w and wu f y 2009 j. phys .
a : math . theor .
* 42 * , 025205 izmailian n sh , kenna r and wu f y 2014 the two point resistance of a resistor network : a new formulation and application to the cobweb network , j phys a : math.theor . *
47 * 035003 ( 10pp ) essam j w , tan z - z and wu f y 2013 proof and extension of the resistance formula for an @xmath2 cobweb network conjectured by tan , zhou and yang .
* arxiv:1312.6727*[cond - mat.stat - mech ] tan z - z 2011 resistance network model , ( xian : xidian university press ) pp9 - 216 tan z - z , zhou l and yang j - h 2013 the equivalent resistance of @xmath144 cobweb network and its conjecture of an @xmath2 cobweb network " , j phys a : math
. theor . * 46 * 195202 ( 12pp ) tan z - z , zhou l and luo d - f 2013 resistance and capacitance of a @xmath145 cobweb network and two conjectures , int . j. circ
appl , to appear doi : 10 . | the resistance between two nodes in general position on a fan network with @xmath0 radial lines and @xmath1 transverse lines is determined . also
a similar result of izmailian , kenna and wu [ 7 ] for an @xmath2 cobweb network is reproduced but the method used here is significantly different .
it avoids the use of the kirchhoff matrix , requires the solution of just one instead of two eigenvalue problems and results directly in only a single summation .
further the current distribution is given explicitly as a biproduct of the method .
the method is the same as that used by tan , zhou and yang [ 10 ] to find the cobweb resistance between center and perimeter for @xmath3 and general @xmath0 .
proof of their conjecture for general @xmath1 is discussed .
+ + * key words : * resistor network ; two - point resistance ; cobweb ; fan * pacs numbers : * 84.30.bv , 01.55.+b , 02.10.yn , 05.50+q |
genetic interaction measures how different genes collectively contribute to a phenotype , and can reveal functional compensation and buffering between pathways ( or protein complexes ) under genetic perturbations .
recently , genome - wide screening of genetic interactions have become possible for different species via high - throughput methods @xcite in which the phenotypic effect of the double - knockout of each pair of genes are compared with the aggregated effects of the two individual knockouts under an assumption of independence . an extreme case , called synthetic lethality , occurs when a double knockout results in the death of a cell even when the single knockouts have no effect .
the genetic interaction networks revealed by these experiments provide novel insights both when analyzed by themselves @xcite and when integrated with other molecular networks or genomic datasets @xcite , such as physical interaction @xcite , gene expression @xcite and chemical - genetic interaction data @xcite . for higher eukaryotes such as human ,
these reverse - genetics approaches have not been as straightforward due to both less amenable genetics and more complex phenotypes of interest such as disease onset and survival , which are difficult to study in cell based assays . despite slow progress in mapping genetic interactions using reverse genetics approaches in human cells
, we are accumulating a wealth of data on individual genetic variations in human populations .
genome - wide association studies ( gwas ) capturing single - nucleotide polymorphisms ( snps ) or copy number variations @xcite that have been widely applied for studying genetic differences between disease samples ( cases ) , and normal samples ( controls ) @xcite , offer an alternative means to map genetic interactions .
for example , if two genetic variants have weak individual association with a disease but very strong joint association , the genes controlled by the two variants may have compensating functions that can buffer variations in each other , but yield a much higher risk of the disease phenotype of interest for a joint mutation .
indeed , the genetic interactions between genetic variants ( such as snps and cnvs ) have been extensively studied as statistical epistasis ( used interchangeable with genetic interaction in this paper ) @xcite in both genome - wide @xcite and targeted studies @xcite .
however , existing approaches mostly consider pairs ( or high - order combinations ) of genetic variants as separate interaction candidates , and hence estimate their statistical significance and study their biological interpretations in an isolated manner @xcite .
while many statistically significant and biologically relevant instances of epistasis are discovered , these approaches may have the following drawback .
several pairs of genetic variants may not have statistically significant genetic interactions when considered as individual interaction candidates , but nevertheless , can be collectively significant if they are highly enriched for two pathways that have complementary functions ( a between pathway model , bpm @xcite ) .
this limitation motivates the design of approaches that can model the collective significance of snp pairs based on their network structure .
specifically , we aim to construct human disease - specific ( cases vs. controls ) genetic interactions from gwas case - control datasets and then discover bpms from the constructed network , i.e. , two sets of genetic variants that have many genetic interactions across the two sets but none or very few within either set . for the community doing research on genetic interaction , the novelty is the exploration of whether the biologically interesting structures ( e.g. , bpms ) discovered from the genetic interaction networks of lower eukaryotes such as yeast also exist in complex diseases , such as cancer and neurological diseases , for higher eukaryotes such as humans . for the community doing research on gwas data analysis ,
this work provides additional evidence to support the shift from the analysis of single genes ( or snps ) to sets of genes ( e.g. , gene sets @xcite or protein interaction subnetworks @xcite ) .
existing approaches that focus on the discovery of individually significant pathways or subnetworks ignore those pathways or subnetworks that are individually insignificant but are compensatory to each other as a combination ( e.g. , pairs of pathways or subnetworks ) in their strong association with a disease phenotype . in this paper , we propose a general framework for constructing human disease - specific genetic interaction networks with gwas data . because different types of genomic data have unique characteristics that need to be addressed , we focus on genome - wide case - control snp data and its accompanying linkage disequilibrium ( ld ) structure .
we discuss the challenges in the construction of genetic interaction networks due to ld structure and propose a general approach with three steps : ( 1 ) estimating snp - snp genetic interactions , ( 2 ) identifying ld blocks and summarizing snp - snp interactions to ld block - block genetic interactions , and ( 3 ) functional mapping ( e.g. gene mapping ) for each ld block . to illustrate how the constructed genetic interaction network can be used to obtain both known and novel biological insights about disease phenotype of interest in the case - control study we designed two sets of functional analyses ( @xmath0 and @xmath1 ) to analyze the genetic interaction networks constructed on each of the six case - control snp datasets used in this study . for functional analysis @xmath0
, we study whether a constructed human genetic interaction network has functional significance with respect to independent biological databases .
specifically , we compare the ld block - block genetic interaction network and the genetic - interaction - profile - based ld block - block similarity network with the human functional network integrated in @xcite .
interestingly , we find that the pairs of ld blocks that have high genetic interaction and those pairs that have high similarity of genetic interaction profiles have significantly higher functional similarity .
this motivates the potential utility of the constructed genetic interaction network for revealing both known and novel biological insights into the disease phenotype of interests in a case - control study . for functional analysis @xmath1 , we study how to use the constructed human genetic interaction network to provide detailed insights about the compensation between pathways in their joint association with a disease phenotype .
specifically , we discover between pathway models ( bpm ) from the block - block genetic interaction network .
a bpm contains two sets of ld blocks , which have many cross - set genetic interactions but very few within - set genetic interactions .
the experiments on the six snp datasets demonstrate that the discovered bpms have statistically significant properties ( supported by permutation tests of case - control groupings ) such as across - set densities of genetic interactions and functional enrichments based on three sets of biological databases .
the significant bpms may provide indications of the compensation between pathways ( or protein complexes ) in their association with the disease phenotypes , and serve as a novel type of biomarker for complex diseases .
comprehensive experimental results on the six case - control snp datasets support several points : ( i ) from the perspective of genetic interaction analysis , the constructed human genetic interaction network has functional significance , and the biologically interesting motifs such as bpm that are common in lower eukaryotes also exist in human with respect to complex diseases such as cancer and parkinson s disease ; ( ii ) from the perspective of gwas data analysis and biomarker discovery , discovering bpms from the constructed human genetic interaction network can help reveal novel biological insights about complex diseases , beyond existing approaches for gwas data analysis that either ignore interactions between snps , or model different snp - snp genetic interactions separately rather than studying global genetic interaction networks as in this study .
we first present a general framework for constructing genetic interaction networks from genome - wide case - control snp datasets , and then describe two approaches for the functional analysis of the resulting networks that can be used to discover novel biological insights about complex diseases .
for each step in this process , we selected a particular approach .
various alternatives are possible , but due to the limitation of space and the desire to clearly explain our general approach , we do not discuss those alternatives in any detail .
however , given the significant results obtained by the current approach consistent over six datasets ( section [ sec : exp ] ) , some of these alternatives should be explored to see if additional improvements in the results are possible .
there are three steps in the network construction framework : ( i ) measuring all pairwise snp - snp genetic interaction with respect to the case - control grouping in a gwas dataset , ( ii ) summarizing snp - to - snp interactions into a block - level genetic interaction network , and ( iii ) functional mapping for each block .
the principal goal of measuring genetic interactions between two snps is to capture the non - additive effect between the two snps in their combined association with the phenotype of interest . for this purpose ,
we leverage the extensive research on statistical epistasis @xcite that has been recently reviewed by h. cordell @xcite . among the different measures for epistasis ,
we selected the information theoretic _ synergy _
. the _ synergy _ between two snps with respect to a binary class label variable ( cases vs. controls ) , i.e. , @xmath2 is defined in @xcite as follows : @xmath3 where @xmath4 denotes the mutual information between the class variable @xmath5 and a variable @xmath6 @xmath7 in this paper , _ synergy _ and mutual information are always normalized by @xmath8 , after which , @xmath2 ranges from @xmath9 to @xmath10 .
the focus of this paper is on positive _ synergy _ since we want to measure the interaction between two snps beyond the independent additive effect of their joint association with a case - control phenotype .
the interpretation of a positive _ synergy _ , e.g. 0.2 , is that the two snps as a combination provide 20% more information about @xmath5 beyond the summation of information provided independently by the two snps . in this paper , we say two snps have a positive genetic interaction if they have a positive _ synergy _ ( beyond - additive effect ) to keep the sign of genetic interaction _ synergy _ consistent . note that , in reverse - genetics based yeast genetic interaction , negative genetic interaction is used to denote beyond - additive effect .
we chose to compute synergy for all pairs of snps rather than just those pairs for which one snp has sufficiently large marginal effects @xcite since we did not want to risk missing snp pairs that have weak ( or no ) marginal effect but a strong combined effect as discussed in @xcite since these pairs are essential for building an interaction network that may have even better statistical power than other approaches .
after the _ synergy _ calculation for all pairwise snps , we get a full weighted snp - snp network .
we denote this matrix as @xmath11 , as shown in figure [ fig : flowchar_construct ] ) . this network can not be directly interpreted because of the ld structure in the snp data . in the next section
, we discuss this challenge in detail and present an approach to address it . ;
( 2 ) identifying ld blocks ; and ( 3 ) summarizing a snp - snp network to a block - block genetic interaction network ( @xmath12 ) . ]
due to the ld structure in snp data , nearby snps tend to have correlated genotypes over the samples .
therefore , if a pair of snps ( @xmath13 and @xmath14 ) have large _ synergy _ , the snps close to @xmath13 probably also have large _ synergy _ with the snps close to @xmath14 .
this can result in a trivial type of local motif ( approximate bicliques ) in the snp - snp network as illustrated in figure [ fig : ld_biclique].this biases the functional analysis since such bicliques do not reflect the functional similarity between the snps in the same ld block . in order to gain non - trivial insights from the genetic interaction network ,
we propose to summarize the snp - snp interaction network by an ld block - block network . * identifying ld blocks * : different measures for estimating ld such as @xmath15 and @xmath16 @xcite are specifically designed for measuring the non - random associations between polymorphisms at different loci .
these measures capture the difference between observed and expected allelic frequencies ( assuming random distributions ) , which depend on the phase information and define an ld block within a genomic region . with a related but different purpose , our goal in this study is to identify a set of snps on the same chromosome having similar genotype profiles and use a single block to represent them .
we use hamming similarity to measure the correlation between the genotype profiles of snps .
( snps are columns in the case - control data as illustrated in figure [ fig : flowchar_construct ] . )
this similarity serves better for our purpose of measuring mathematical similarity between two snps rather than the ld .
we take such a conservative approach in order to make sure that we do not create two separate ld blocks for snps that have similar genotype profiles .
this also avoids the estimation of phase information , which adds additional uncertainty that may confound the analysis .
we do not restrict an ld block to be within a local genomic region .
this is because snps that are far from each other can also have high genotype correlation , at least from the mathematical perspective as shown in @xcite .
again , we take such a conservative approach in order to make sure that we do not create two separate ld blocks for snps that have similar genotype profiles , which may happen if a window - size constraint is used . for simplicity , we perform a greedy search of ld blocks .
specifically , we randomly take a snp and combine it with all the other snps ( on the same chromosome ) with hamming similarity above a threshold @xmath17 ( @xmath18 is used in this study ) as an ld block .
a snp will only be assigned to one ld block .
* snp - snp network to ld block - block network * : after identifying all the ld blocks in a dataset , we have a many - to - one mapping from all the snps to a set of blocks . given this mapping ,
we summarize the snp - snp _ synergy _ network to a block - block _ synergy _ network using the following general function to estimate the _ synergy _ between two blocks @xmath19 and @xmath20 , @xmath21 where @xmath22 denotes a general aggregation function , e.g. @xmath23 or @xmath24 . in this paper
, we adopt the @xmath23 function based on the following reasons and observations : ( 1 ) biologically , it is likely that only one pair of snps across two ld blocks are truly causative in the case - control phenotype , in which case @xmath23 is the ideal aggregation function .
( 2 ) based on multiple datasets used in the experiment section , the @xmath23 function consistently yields coherence with existing biological knowledge gained from yeast genetic interaction networks .
( 3 ) in the sanity check , the pairs with top _ synergy _ values have similar ld - block sizes as the null distribution , and thus are not due to the bias of large ld block sizes .
there are other aggregation approaches @xcite that we will explore in future work .
after this step , we have a block - block genetic interaction network , which we denote as @xmath12 as illustrated in figure [ fig : flowchar_construct ] . * functional mapping for each ld block * : after the construction of an ld block - block genetic interaction network , a functional mapping for each block is required to interpret the structure of the interaction network in functional terms that have biological meaning . for this purpose
, we first assign each snp to the closest gene based on its genome location .
then , the genes of an ld block are obtained from the snps that were assigned to that block in the ld identification step .
we will explore more advanced gene mapping strategies in future work .
interestingly , even with this simple gene mapping approach , the functional analysis in section [ sec : exp ] shows that the constructed ld block - block interaction network still appears to have functional structure .
note that the gene mapping does not result in a gene - gene interaction network .
this is because an ld block can span multiple genes so that gene - gene network derived from the blocks would contain many trivial biclique patterns .
from this perspective , the genetic interaction network constructed for yeast in @xcite from eqtl data may have included a large number of false positive gene - gene edges since it connected all the gene pairs from the two ld blocks .
in contrast , the block - block network constructed in this study has the least amount of bias from trivial bicliques . in section [ sec : construction ] , we presented a framework for constructing a block - block genetic interaction network from a human case - control genomic dataset . in this subsection , we present two sets of functional analyses : @xmath0 , which comparing the constructed ld block - block genetic interaction network and the corresponding similarity network derived from the human functional network , and @xmath1 , which discovers between pathway models ( bpm ) from the ld block - block genetic interaction network for functional enrichment analysis @xcite . both types of analysis have been used in the systematic interpretation of double - knockout - based yeast genetic interaction networks @xcite .
here we use a similar approach to reveal novel biological insights from the genetic interaction networks constructed from human case - control genomic datasets .
figure [ fig : flowchart_bpm_net_analysis ] shows the overall design of the two types of functional analysis and will be referenced extensively in the rest of this section . an important point in understanding following discussion
is that we first threshold the block - block genetic interaction matrix ( @xmath25 ) to a binary matrix ( @xmath26 ) .
specifically , we binarize this network with a quantile threshold @xmath27 ( e.g. 1% ) , such that those block - block edges with _ synergy _ in the top @xmath27 quantile ( those with large beyond - addtive effect interactions ) are kept in the binary network . we denote the matrix representation of this network as @xmath28 as illustrated in figure [ fig : flowchart_bpm_net_analysis ] . both of the analyses make use of binarized matrix @xmath28 . . there are two types of analysis : ( @xmath0 ) constructing the block - block similarity network ( @xmath29 ) from @xmath28 and enrichment analysis , and ( @xmath1 ) discovering bpms ( between pathway models ) from @xmath28 and enrichment analysis .
both enrichment analyses are with respect to three sources of biological knowledge , namely ( i ) the block - block functional similarity network ( @xmath30 ) derived from the human functional network @xmath31 , ( ii ) the human go annotations and ( iii ) the human molecular signature database ( msigdb ) . ] in the first approach , we study whether the constructed human genetic interaction network has functional significance supported by independent biological databases . specifically , we first use jaccard similarity to measure the similarity between the profiles of genetic interactions of two blocks in @xmath26 . this results in matrix @xmath32
as shown in figure [ fig : flowchart_bpm_net_analysis ] .
the motivation is that the analysis of global yeast genetic interaction networks @xcite has shown that the similarity between such interaction profiles of two genes is correlated with the functional similarity of two genes . in the experimental section
, we will use the constructed block - block similarity matrix @xmath32 to compare the similarity of block interaction profiles with the human gene - gene functional network integrated in @xcite . functional analysis approach @xmath1 , takes a complementary approach to discover between pathway models ( bpm ) .
( this approach was demonstrated to be effective in the analysis of yeast genetic interaction networks @xcite . ) using insights gained from yeast genetic interaction network , a bpm contains two sets of genes that have many across - set genetic interactions and within - set protein - protein interactions @xcite .
the two sets of genes in a bpm may correspond to two biological pathways ( or protein complexes ) that have redundant ( or complementarity ) biological functions with respect to the case - control grouping . in the context of this study , a bpm contains two sets of ld blocks , which have many large cross - set genetic interactions ( synergy values ) in @xmath26 and very few within - set genetic interactions .
discovering bpms from the binarized disease - specific genetic interaction matrix ( @xmath26 ) can provide novel insights beyond existing approaches designed for analyzing case - control snp datasets .
recently , existing approaches for analyzing case - control datasets have shifted from discovering single genes ( or snps etc ) to sets of genes ( e.g. pathways @xcite or protein interaction subnetworks @xcite ) .
while many statistically significant and biologically relevant gene sets or subnetworks are discovered , existing approaches may ignore those pathways or subnetworks that are individually insignificant but are compensatory to each other as a combination in their strong association with a disease .
from this perspective , a bpm captures the compensation between two pathways in their combined association with complex diseases such as cancer that may be caused by the perturbation of multiple ( e.g. , a pair of ) pathways but not individual pathways .
this may lead to the discovery of a new type of complex biomarker , i.e. pairs of compensating pathways or protein complexes .
many algorithms have been proposed to discover bpms from yeast genetic interaction networks .
these algorithms mostly depend on integration of both physical interaction network and genetic interaction network @xcite .
such integrative approaches have the advantage of better interpretability because of the integration with physical interaction data .
however , given our goal of discovering bpms from human case - control datasets and the fact that human protein interaction network is not as complete and lacks reproducibility @xcite , an integrative approach may miss many bpms that are not yet well - supported by the existing limited functional knowledge of the human genome .
for the above reason , in this study , we took an exploratory approach in which , we search for bpms only from the genetic interaction matrix @xmath26 . given a binary symmetric matrix ( @xmath26 ) , the bpm discovery problem can be reformulated as a quasi - biclique discovery problem @xcite . a quasi - biclique is defined as a non - weighted bipartite graph @xmath33 ( where @xmath6 and @xmath34 are two sets of ld blocks and @xmath35 is a set of ld block - block genetic interaction edges ) such that , for a given parameter @xmath36 .
@xmath37 where @xmath38 denotes the number of edges between @xmath39 and all the nodes in @xmath34(similarly for @xmath40 ) . in this paper
, we adopt a greedy algorithm with polynomial time complexity @xcite to efficiently search for quasi - blciques from the binary block - block network ( @xmath26 ) .
note that , several other algorithms that are designed for quasi - biclique discovery or bpm discovery @xcite can also be applied for the same purpose . these will be explored in future work as this paper focuses on presenting the overall framework .
based on the definition , a quasi - biclique may also have many edges within each of the two sets ( @xmath6 and @xmath34 ) , while bpms that with no or very few within - set edges are relatively more interesting .
therefore , after the discovery of a set of quasi - bicliques , a postprocessing will be applied to further select a subset of bicliques with a small fraction of within - set genetic interactions as bpm candidates for the following functional analysis .
we will design experiments to interpret the discovered bpms with human functional network @xcite and conduct enrichment analysis with human go annotations @xcite and human molecular signature database ( msigdb@xcite ) in section [ sec : bpmresults ] .
in this section , we present the experimental results on the two sets of functional analyses ( described in section [ sec : analysis ] ) on the network constructed with the framework presented in section [ sec : construction ] . in the experiments , we use six case - control snp datasets , one from a genome - wide association study on parkinson s disease ( disease vs. normal ) @xcite , and the others from targeted studies ( with around 3000 snps over about 1000 genes ) on myeloma long vs. short survival @xcite , myeloma cases vs. controls @xcite , lung cancer cases vs. controls ( both heavy smokers ) @xcite , rejection vs. no - rejection after kidney transplant @xcite , bone disease ( large vs. small number of bone lesions ) @xcite .
we denote these 6 datasets as parkinson , myeloma - survival , myeloma , lung , kidney , bone in this section . for the genome - wide parkinson data
, we selected 8994 non - synonymous snps for this study for the concern of computational efficiency , given the huge number of all pairwise snp - snp genetic interaction to compute , especially in a large number ( 100 ) of permutations for each dataset ( functional analysis ii ) .
table [ tab : datasets ] displays the information about the six datasets , including the number of cases , controls , snps , ld blocks identified and edges in an ld block - block genetic interaction network ( @xmath26 ) .
.information about the datasets .
the number of blocks depends on the hamming similarity threshold @xmath17(@xmath41 used in this paper ) .
the number of ld block - block ( b - b ) edges depends on the bianrization threshold @xmath27 ( 1.5% for all the datasets except parkinson , for which we used 0.25% to speed up the bpm discovery in functional analysis ii . ) [ cols= " < , < , < , < , < , < " , ] several observations can be made from table [ tab : fdrtable ] . 1 .
* statistical significance of the discovered bpms * : many bpms are significant with respect to the @xmath42 on the the density of genetic interactions across the two sets .
specifically , there are significant bpms discovered with respect to @xmath42 on all the six datasets .
this indicates that the existence of genetically buffered functional modules as captured by the bpm structure is evident in genome - wide case - control snp datasets .
these bpms generally have larger sizes and interaction densities than the random bpms discovered from permutation tests .
* biological significance of the discovered bpms * : many bpms are significant with respect to the last three measures , i.e. @xmath43 , @xmath44 and @xmath45 , which suggests that they are not only statistically significant , but are also supported by independent genomic / proteomic evidence .
more specifically , the density of edges in the functional network , within both compensatory modules ( @xmath43 , @xmath44 ) as well as between them ( @xmath45 ) , is frequently higher for the bpms derived from the real data as compared to the random permutations of case - control groupings .
while table [ tab : fdrtable ] gives an overall summary of the bpms discovered from each of the datasets , figure [ fig : bpm_examples ] shows two illustrative examples of bpms discovered from parkinson and myeloma - survival .
many bpms from the other four datasets are available on the supplementary website .
p - values ( @xmath46 test ) for all the individual snps in the blocks of each side of a bpm is also shown .
a dashed line is shown in each histogram , which corresponds to the @xmath47 p - value without bonferroni correction .
all the snps have insignificant individual p - value supports the highlight of the proposed framework to discover significant bpms containing snps with weak individual association , that are mostly ignored by existing approaches . ]
several observations can be made from figure [ fig : bpm_examples ] .
we will use the bpm discovered from parkinson as a running illustrative example of their interpretation . 1 . * the statistical significance of the bpm discovered from parkinson : * the table on the left shows the fdrs for each of the six bpm measures . as shown , the six fdrs are all below @xmath48 with several below @xmath49 ( the latter four ) .
a @xmath42 of @xmath50 ( fdr @xmath51 ) shows the dense genetic interactions between the two sets of ld blocks in this bpm .
an @xmath44 of 0.98 ( fdr 0.05 ) indicates that the right set of ld blocks have strong ld block - block functional similarity , which agrees the concept of bpm , i.e. each side of the bpm is likely to be involved in a common pathway or a process .
an @xmath45 of @xmath52 ( fdr 0.03 ) suggests that the two sets of ld blocks in this bpm may control two functions respectively , which may have compensatory function under genetic variations . 2 .
* most snps in the bpm have insignificant individual association * : the histograms of individual snp association ( @xmath53 p - value ) ) indicate that , almost all of the snps in the significant bpm have insignificant individual p - value ( below the green dashed line which corresponds to @xmath47 even before bonferroni correction ) .
this supports the utility of the proposed framework for discovering significant bpms that contain snps with weak individual association , which would be mostly ignored by existing approaches .
* all cross - set snp pairs in the bpm have insignificant genetic interaction when considered separately * : we also compute fdr for each snp pair in a bpm across the two sets of ld blocks . for the two examples shown in the figure , the lowest fdrs of individual snp pairs are @xmath54 and @xmath55 in the parkinson example and the myeloma - survival example respectively .
this indicates that none of the snp pairs in the two datasets would be considered as significant epistasis if they are considered in an isolated manner .
in contrast , discovering them as a bpm with the proposed approach yields their significant fdrs .
this demonstrates the effectiveness of the proposed functional analysis for genome - wide case - control snp datasets .
* the two sets in the bpm contain ld blocks either from different chromosomes or from the same chromosome but with different genotype profiles * : the two heat maps of all pair s hamming similarities for the all the snps ( grouped by ld blocks ) indicate that the snps within each ld block have correlated genotype profiles while those from different ld blocks have different genotype profiles .
this indicates that the bpm is not a trivial biclique that is due to ld structure .
note that , this is an illustration of the effectiveness and necessity of constructing and analyzing a block - level genetic interaction network instead of either a snp - level network or a gene - level network as discussed in the methods section .
the diversity of chromosomes in both sets of a bpm indicates that the two compensating functions ( e.g. pathways ) are from different chromosomes , showing the complexity of the mechanisms underlying the disease phenotype .
* an illustrative interpretation on the bpm discovered on the parkinson s dataset * : given the statistical significance of the bpm structures discovered across many of the gwas studies , we further asked whether the genes on both sides were enriched for known pathways or processes using gene sets defined by gene ontology terms or msigdb .
consistent with their overlap with the functional network , a number of the modules involved in bpms did show significant enrichment ( see figure [ fig : bpm_examples ] for examples on the parkinson and myeloma - survival datasets or the supplementary website for a complete list ) .
the bpm shown in figure [ fig : bpm_examples ] that is associated with parkinson s disease was a pair of modules , one enriched for ion channel activity and the other enriched for protein kinase activity .
the ion channel activity enrichment is driven by three genes @xmath56 , @xmath57 , @xmath58 , each of which comes from a separate ld block on three different chromosomes ( chromosomes 9 , 17 , and 21 , respectively ) .
this is potentially interesting as potassium channels were recently suggested as a possible new target for therapeutics @xcite . it was hypothesized that such ion channels may affect the progressive loss of dopamine neurons , which is the main cause of parkinson s disease .
it is also interesting to note that mouse knock - out mutants of @xmath57 , a potassium voltage - gated channel protein , have been associated with the so - called shaker / waltzer phenotype , which is characterized by rapid bilateral circling during locomotion @xcite .
the complementary module in this parkinson s disease bpm was enriched for protein kinase activity due to the presence of the two protein kinases @xmath59 ( c - mer proto - oncogene tyrosine kinase ) and @xmath60 ( eukaryotic translation initiation factor 2-alpha kinase 3 ) , suggesting that combined mutations affecting ion channel activity and one of these signaling pathways may be causal determinants of parkinson s . while the specific link is unclear , it is interesting to note that @xmath60 is one of the key regulators of the eukaryotic translation initiation factor and thus , it controls global rates of protein synthesis in the cell .
it is certainly conceivable that mutations in such a protein with relatively global influence on protein levels could modify , and in this case aggravate , the effects of other mutations .
in fact , mutations in another translation initiation factor , @xmath61 , were recently associated with vanishing white matter disease , a disorder that causes rapid deterioration of the central nervous system @xcite .
given the large number of genes involved in the ld blocks associated with each bpm , identifying the genes functionally responsible could be quite difficult and is one of the main caveats of this type of analysis
. however , this process can be aided by simple enrichment analysis , which in this case appears to implicate processes whose link to parkinson s disease seems plausible .
in this paper , we target the construction and functional analysis of disease - specific human genetic interaction networks from genome - wide association data designed for case - control studies on complex diseases . specifically , we focused on genome - wide case - control snp data , which has its linkage disequilibrium ( ld ) structure .
we discussed the challenges in the detection of genetic interactions due to ld structure and propose a general approach with three steps : ( 1 ) estimating snp - snp genetic interactions , ( 2 ) identifying genome segments in linkage disequilibrium ( ld ) and mapping of snp interactions to ld block - block interactions , and ( 3 ) mapping for ld blocks to genes .
we performed two sets of functional analyses on six case - control snp datasets to study if the constructed human genetic interaction network has functional significance supported by independent biological evidence by comparing with a human functional networks .
we also demonstrated how the constructed interaction network can provide high - resolution insights about the compensation between pathways in their joint association with a disease phenotype by discovering between - pathway models .
comprehensive experimental results on six case - control datasets demonstrated that ( i ) from the perspective of genetic interaction analysis , the constructed human genetic interaction network has functional significance , and that biologically interesting motifs such as bpm that are common in lower eukaryotes also exist in the genetic interaction network discovered from human genetic variations associated with complex diseases such as cancers and parkinson s disease ; ( ii ) from the perspective of gwa data analysis , discovering bpms from the constructed human genetic interaction network can help reveal novel biological insights about complex diseases , beyond existing approaches for gwas data analysis that either ignore interactions between snps , or model different snp - snp genetic interactions separately rather than studying global genetic interaction networks as done in this study .
this paper focused on the presentation of the overall framework of constructing and analyzing human disease - specific genetic interaction network with gwas data .
there are a number of interesting and necessary directions for future work such as exploring the effect of different epistasis measures , the effect of different ld block identification approaches , the effect of different aggregation functions , and the effect of different gene mapping approaches . in conclusion
, we want to highlight that , even though we chose to use some relatively simple and conservative options in the framework which needs further exploration as discussed above , the significant statistical and biological evidence obtained from the two sets of functional analyses demonstrate the effectiveness of the current framework in revealing several consistent observations over six case - control snp datasets .
g. fang , m. haznadar , w. wang , m. steinbach , b. vanness , and v. kumar .
efficient discovery of high - order snp combinations with strong phenotype association .
technical report 013 , dept . of computer science , univ . of minnesota , 2010 .
b. v. ness , c. ramos , m. haznadar , a. hoering , j. c. jeff haessler , s. jacobus , m. oken , v. rajkumar , p. greipp , b. barlogie , b. durie , m. katz , g. atluri , g. fang , r. gupta , m. steinbach , v. kumar , r. mushlin , d. johnson , and g. morgan . .
, 6:66 , 2008 . | genetic interaction measures how different genes collectively contribute to a phenotype , and can reveal functional compensation and buffering between pathways under genetic perturbations .
recently , genome - wide investigation for genetic interactions has revealed genetic interaction networks that provide novel insights both when analyzed independently and when integrated with other functional genomic datasets . for higher eukaryotes such as human ,
the above reverse - genetics approaches are not straightforward since the phenotypes of interest for higher eukaryotes such as disease onset or survival , are difficult to study in a cell based assay . in this paper , we propose a general framework for constructing and analyzing human genetic interaction networks from genome - wide single nucleotide polymorphism ( snp ) datasets used for case - control studies on complex diseases .
specifically , we propose a general approach with three major steps : ( 1 ) estimating snp - snp genetic interactions , ( 2 ) identifying linkage disequilibrium ( ld ) blocks and mapping snp - snp interactions to ld block - block interactions , and ( 3 ) functional mapping for ld blocks .
we performed two sets of functional analyses for each of the six case - control snp datasets used in the paper , and demonstrated that ( i ) genes in ld blocks showing similar interaction profiles tend to be functionally related , and ( ii ) the network can be used to discover pairs of compensatory gene modules ( between - pathway models ) in their joint association with a disease phenotype .
the proposed framework should provide novel insights beyond existing approaches that either ignore interactions between snps or model different snp - snp pairs with genetic interactions separately .
furthermore , our study provides evidence that some of the core properties of genetic interaction networks based on reverse genetics in model organisms like yeast are also present in genetic interactions revealed by natural variation in human populations . |
we discuss the direct photon production in two hadron collision where one hadron is transversely polarized .
we present the hadron tensor for this process and study the effects which lead to the soft breaking of factorization ( or the universality breaking ) through the qed and qcd gauge invariance @xcite . in a similar manner as in @xcite , the special role
is played by the contour gauge for gluon fields . moreover , we present the results of the detailed analysis of hadron tensor in the feynman gauge with the particular emphasis on the qed gauge invariance @xcite
. moreover , the results in the feynman and contour gauges coincide if the gluon poles in the correlators @xmath0 are absent .
this is in agreement with the relation between gluon poles and the sivers function which corresponds to the `` leading twist '' dirac matrix @xmath3 .
this also suggests that the gluonic poles appear only in `` physical '' components of gluon field , coinciding with total ones in the axial gauge .
we study the semi - inclusive process where the hadron with the transverse polarization collides with the other unpolarized hadron to produce the direct photon in the final state in : @xmath4 for ( [ process ] ) ( also for the drell - yan process ) , the gluonic poles manifest .
we perform our calculations within a _ collinear _ factorization and , therefore , it is convenient to fix the dominant light - cone directions as @xmath5 accordingly , the quark and gluon momenta @xmath6 and @xmath7 lie along the plus dominant direction while the gluon momentum @xmath8 along the minus direction .
the final on - shell photon and quark(anti - quark ) momenta can be presented as @xmath9 and , the mandelstam variables for the process and subprocess are defined as @xmath10 the amplitude ( or the hadron tensor ) of ( [ process ] ) constructed by the contributions from ( i ) the leading ( lo ) in particle number diagrams : two diagrams with a radiation of the photon before ( @xmath11 ) and after ( @xmath12 ) the quark - gluon vertex with the gluon going to the lower blob , see the right side of fig .
[ fig - dirphot ] ; ( ii ) the next - to - leading order ( nlo ) diagrams : eight diagrams constructed from the lo diagrams by insertion of all possible radiations of the additional gluon which together with the quark goes to the upper blob , see the left side of fig .
[ fig - dirphot ] .
so , we have the hadron tensor related to the corresponding asymmetry : @xmath13 here , we will mainly discuss the hadron tensor rather than the asymmetry itself .
so , the hadron tensor as an interference between the lo and nlo diagrams , @xmath14 , can be presented by fig .
[ fig - dirphot ] where the upper blob determines the matrix element of the twist-3 quark - gluon operator while the lower blob the matrix element of the twist-2 gluon operator related to the unpolarized gluon distribution . as a result of factorization
, we should reach the factorized form for the considered subject : @xmath15 we now dwell on the qcd gauge invariance of the hadron tensor for the direct photon production ( dpp ) .
first of all , let us remind that having used the _ contour gauge conception _
@xcite , one can check that the representation @xmath16 belongs to the gauge defined by @xmath17=1 $ ] , while the representation @xmath18 corresponds to the gauge that defined by @xmath19=1 $ ] . in both ( [ b - plus ] ) and ( [ b - minus ] ) , the function @xmath20 related to the following prametrization : @xmath21 further , to check the gauge invariance , we have to consider all typical diagrams , depicted in fig.[fig - dirphot ] , that correspond to the certain @xmath22-process ( see , @xcite ) . notice that , for the qcd gauge invariance , we have to assume that all charged particles are on its mass - shells .
that is , we will deal with only the physical gluons . as a result of our calculation
, the qcd gauge invariance ( or the ward identity ) takes place provided only the presence of the different complex prescriptions in gluonic poles dictated by the final or initial state interactions : @xmath23=1 \rightarrow \frac{t(x_1,x_2)}{x_1-x_2 - i\epsilon}\,\\ \\
\hspace{-2cm}{\rm \bf isi}&\rightarrow & \frac{1}{-\ell^+ + i\epsilon } \rightarrow { \rm gauge}\,\,\,[z^-,\ , -\infty^-]=1 \rightarrow \frac{t(x_1,x_2)}{x_1-x_2 + i\epsilon}\,\\ \end{array}\,\,\ , \bigg\ } \rightarrow \textbf{qcd gauge invariance}\,,\end{aligned}\ ] ] where @xmath24 and @xmath25 imply the final and initial state interactions , respectively .
we now calculate the full expression for the hadron tensor which involves both the standard and new contributions to the gluon pole terms . after computing the corresponding traces and performing simple algebra within the frame we are choosing
, it turns out that the only nonzero contributions to the hadron tensor come from the following terms @xmath26[x_1 y s + i\epsilon]^2}\ , \frac{\varepsilon^{q_\perp + s_\perp -}}{p_1^+}\ , b^{v}_-(x_1,x_2)\,,\end{aligned}\ ] ] @xmath27[x_1 t + i\epsilon]^2}\ , \frac{\varepsilon^{q_\perp + s_\perp -}}{p_1^+}\ , b^{v}_+(x_1,x_2)\,,\end{aligned}\ ] ] @xmath28 ^ 2}\ , \frac{\varepsilon^{q_\perp + s_\perp -}}{2x_1p_1^+ + i\epsilon}\ , \int dx_2\,b^{v}_+(x_1,x_2)\,,\end{aligned}\ ] ] and @xmath29[x_2 t + i\epsilon][(x_1-x_2 ) y s + i\epsilon]}\ , \frac{\varepsilon^{q_\perp + s_\perp -}}{p_1^+}\ , b^{v}_+(x_1,x_2)\,.\end{aligned}\ ] ] here , @xmath30 , @xmath31 , @xmath32 .
the other diagram contributions disappear owing to the following reasons : ( i ) the @xmath33-algebra gives @xmath34 ; ( ii ) the common pre - factor @xmath35 goes to zero , ( iii ) cancelation between different diagrams . analysing the results for the terms h1 , h7 , d4 and h10 ( see eqns .
( [ diah1])([diah10 ] ) ) , we can see that @xmath36 in other words , as similar to the drell - yan process , the new ( non - standard " ) contributions generated by the terms h1 , h7 and d4 result again in the factor of @xmath37 compared to the standard " term h10 contribution to the corresponding hadron tensor .
we now discuss the hadron tensor which contributes to the single spin ( left - right ) asymmetry measured in the drell - yan process with the transversely polarized nucleon ( see fig .
[ fig - dy ] ) : @xmath38 here , the virtual photon producing the lepton pair ( @xmath39 ) has a large mass squared ( @xmath40 ) while the transverse momenta are small and integrated out .
the left - right asymmetry means that the transverse momenta of the leptons are correlated with the direction @xmath41 where @xmath42 implies the transverse polarization vector of the nucleon while @xmath43 is a beam direction @xcite .
since we perform our calculations within a _ collinear _ factorization , it is convenient to fix the dominant light - cone directions as @xmath44 so , the hadron momenta @xmath45 and @xmath46 have the plus and minus dominant light - cone components , respectively .
accordingly , the quark and gluon momenta @xmath6 and @xmath7 lie along the plus direction while the antiquark momentum @xmath8 along the minus direction .
the photon momentum reads ( see fig .
[ fig - dy ] ) @xmath47 we work within the feynman gauge for gluons . after calculation of all relevant traces in the factorized hadron tensor and after some algebra , we arrive at the following contributions for the unintegrated hadron tensor ( which involves all relevant contributions except the mirror ones ) : the standard diagram depicted in fig . [
fig - dy](a ) gives us @xmath48 x_1\int dx_2 \frac{b^{(2)}(x_1,x_2)}{x_1-x_2+i\epsilon } + \frac{p_{1\,\mu}}{y } \
, \varepsilon_{\nu s^t - p_2}\ , \int dx_2 \frac{b^{(\perp)}(x_1,x_2)}{x_1-x_2+i\epsilon } \bigg\}\ , , \nonumber\end{aligned}\ ] ] while the non - standard diagram presented in fig .
[ fig - dy](b ) contributes as @xmath49 in eqns .
( [ dy - st ] ) and ( [ dy - nonst ] ) , all indices have to be treated as contravariant ones independently of the real position in the formulae .
we also introduce the shorthand notation : @xmath50 .
moreover , the parametrizing functions are associated with the following correlators : @xmath51\ , , \\ \label{parfunb2 } & & b^{(2)}(x_1,x_2 ) \longleftarrow { \cal f}_2\big[\langle p_1 , s^t| \bar\psi(\eta_1)\ , \gamma^\perp \ , a^+(z)\ , \psi(0 ) | s^t , p_1 \rangle \big]\ , , \\ \label{parfunbperp } & & b^{(\perp)}(x_1,x_2 ) \longleftarrow { \cal f}_2\big[\langle p_1 , s^t| \bar\psi(\eta_1)\ , \gamma^+\ , \big(\partial^\perp\ , \ , a^+(z)\big)\ ,
\psi(0 ) | s^t , p_1
\rangle \big]\,,\end{aligned}\ ] ] where @xmath52 denotes the corresponding fourier transformation . summing up all contributions from the standard and non - standard diagrams , we finally obtain the expression for the unitegrated hadron tensor .
we have @xmath53 \ , \varepsilon_{\nu s^t -p_2}\ ,
\int dx_2 b^{(1)}(x_1,x_2 ) + \nonumber\\ & & \frac{p_{2\,\mu}}{x_1 } \ , \varepsilon_{\nu s^t - p_2}\ , \int dx_2 b^{(2)}(x_1,x_2 ) - \big [ \frac{p_{2\,\nu}}{x_1 } \varepsilon_{\mu s^t - p_2 } + \frac{p_{2\,\mu}}{x_1 } \varepsilon_{\nu s^t - p_2 } \big ] x_1\int dx_2 \frac{b^{(2)}(x_1,x_2)}{x_1-x_2+i\epsilon } + \nonumber\\ & & \frac{p_{1\,\mu}}{y } \ , \varepsilon_{\nu s^t - p_2}\ , \int dx_2 \frac{b^{(\perp)}(x_1,x_2)}{x_1-x_2+i\epsilon } \bigg\}\,,\end{aligned}\ ] ] notice that the first term in eqn .
( [ dy - ht-1 ] ) coincides with the hadron tensor calculated within the light - cone gauge @xmath54 .
let us now discuss the qed gauge invariance of the hadron tensor . from eqn .
( [ dy - ht-1 ] ) , we can see that the qed gauge invariant combination is @xmath55 \
, \varepsilon_{\nu s^t -p_2},\ , \quad q_\mu { \cal t}_{\mu\nu } = q_\nu { \cal t}_{\mu\nu}=0.\end{aligned}\ ] ] at the same time , there is a single term with @xmath56 which does not have a counterpart to construct the gauge - invariant combination @xmath57 therefore , the second term in eqn .
( [ dy - st ] ) should be equal to zero .
this also leads to nullification of the second term in eqn .
( [ dy - nonst ] ) .
now , to get the qed gauge invariant combination ( see ( [ gi - comb ] ) ) one has no the other way rather than to combine the first terms in eqns .
( [ dy - st ] ) and ( [ dy - nonst ] ) which justifies the treatment of gluon pole in @xmath58 using the complex prescription . in addition
, we have to conclude that the only remaining third term in ( [ dy - st ] ) should not contribute to ssa .
moreover , we can conclude that , in the case with the substantial transverse component of the momentum , there are no any sources for the gluon pole at @xmath59 . as a result , the function @xmath60 has no gluon poles and , due to t - invariance , @xmath61 , obeys @xmath62 . on the other hand ,
if we have @xmath3 in the correlator , see ( [ parfunb1 ] ) , the transverse components of gluon momentum are not substantial and can be neglected .
that ensures the existence of the gluon poles for the function @xmath58 .
this corresponds to the fact that the sivers function , being related to gluon poles , contains the `` leading twist '' projector @xmath3 .
so , we may conclude that in the feynman gauge the structure @xmath63 does not produce the imaginary part and ssa as well .
working within the feynman gauge , we derive the qed gauge invariant ( unitegrated ) hadron tensor for the polarized dy process : @xmath64 \varepsilon_{\nu s^t -p_2 } \hspace{-1.5mm}\int dx_2 b^{(1)}(x_1,x_2).\end{aligned}\ ] ] further , after calculation the imaginary part ( addition of the mirror contributions ) , and after integration over @xmath65 and @xmath66 , the qed gauge invariant hadron tensor takes the form @xmath67 \ , \varepsilon_{\nu s^t -p_2}\ , t(x_b , x_b)\,.\end{aligned}\ ] ]
this expression fully coincides with the hadron tensor which has been derived within the light - cone gauge for gluons .
moreover , the factor of @xmath37 in the hadron tensor , we found within the axial - type gauge @xcite , is still present in the frame of the feynman gauge .
the different prescriptions in the gluonic poles .
moreover , the different prescriptions are needed to ensure the qcd gauge invariance .
this situation can be treated as a soft breaking of the universality condition resulting in factorization breaking .
we find that the non - standard " new terms , which exist in the case of the complex twist-@xmath2 @xmath68-function with the corresponding prescriptions , do contribute to the hadron tensor exactly as the standard " term known previously .
this is another important result of our work .
we also observe that this is exactly similar to the case of drell - yan process studied in @xcite .
we argue the absence of gluon poles in the correlators @xmath0 based on the light - cone dynamics . at the same time
, we also show that the lorentz and qed gauge invariance of the hadron tensor calculated within the feynman gauge requires the function @xmath60 to be without any gluon poles .
this property seems to be natural from the point of view of gluon poles relation @xcite to sivers functions as the latter is related to the projection @xmath3 .
note also that this may be understood as appearance of gluonic poles due to the physical components @xcite of gluonic fields .
another support of this picture is provided by the fact , that imaginary phase in [ qcd - illust ] allows to derive @xcite burkardt sum rules from conservation laws .
we would like to thank v. braun , i.o .
cherednikov , a.v .
efremov , d. ivanov and l. szymanowski for useful discussions and correspondence .
this work is partly supported by the hl program .
99 i. v. anikin and o. v. teryaev , eur .
j. c * 75 * ( 2015 ) 5 , 184 [ arxiv:1501.04536 [ hep - ph ] ] .
i. v. anikin and o. v. teryaev , phys . lett .
b * 690 * ( 2010 ) 519 [ arxiv:1003.1482 [ hep - ph ] ] ; i. v. anikin and o. v. teryaev , j. phys . conf .
* 295 * ( 2011 ) 012057 [ arxiv:1011.6203 [ hep - ph ] ] .
i. v. anikin and o. v. teryaev , arxiv:1508.00227 [ hep - ph ] ( in press in plb ) .
s. v. ivanov , g. p. korchemsky and a. v. radyushkin , yad .
fiz . * 44 * , 230 ( 1986 ) [ sov .
j. nucl .
phys . * 44 * , 145 ( 1986 ) ] ; s. v. ivanov and g. p. korchemsky , phys .
b * 154 * , 197 ( 1985 ) ; s. v. ivanov , fiz .
chast . atom .
yadra * 21 * , 75 ( 1990 ) .
n. n. bogolyubov and d. v. shirkov , `` introduction to the theory of quantized fields , '' intersci .
astron . * 3 * , 1 ( 1959 ) . v. barone , a. drago and p. g. ratcliffe , phys . rept .
* 359 * , 1 ( 2002 ) [ arxiv : hep - ph/0104283 ] .
d. boer , p. j. mulders and f. pijlman , nucl .
b * 667 * ( 2003 ) 201 [ hep - ph/0303034 ] .
x. ji , j. w. qiu , w. vogelsang and f. yuan , phys .
* 97 * ( 2006 ) 082002 [ hep - ph/0602239 ] .
p. g. ratcliffe and o. v. teryaev , hep - ph/0703293 ; mod .
a * 24 * ( 2009 ) 2984 [ arxiv:0910.5348 [ hep - ph ] ] . | we discuss the role of gluon poles and the gauge invariance for the hadron tensors of drell - yan and direct photon production processes with the transversely polarized hadron .
these hadron tensors are needed to construct the corresponding single spin asymmetries . for the drell - yan process
, we perform our analysis within both the feynman and axial - type ( contour ) gauges for gluons . in both the feynman and contour gauges ,
we demonstrate that the gauge invariance leads to the need of the new ( non - standard ) diagrams .
moreover , in the feynman gauge , we argue the absence of gluon poles in the correlators @xmath0 related traditionally to @xmath1 . as a result ,
these terms disappear from the final qed gauge invariant drell - yan hadron tensor . for the direct photon production , by using the contour gauge for gluon fields
, we find that there are new twist-@xmath2 terms present in the hadron tensor of the considering process in addition to the standard twist-@xmath2 terms . |
the decoupling of the self - diffusion constant from the viscosity or the structural relaxation time also referred to as the breakdown of the stokes - einstein ( se ) relation that occurs for temperatures @xmath0 near the glass transition temperature @xmath1 is among the most prominent features of fragile glass formers @xcite .
the decoupling has been considered as one of the signatures of spatially heterogeneous dynamics or `` dynamical heterogeneities '' @xcite . on the other hand ,
it is also well recognized that the onset temperature @xmath2 of the decoupling is close to a crossover temperature at which transport properties change their characters @xcite and below which the dynamics is thought to be dominated by activated hopping processes over barriers @xcite .
then , a natural question arises as to possible connections among the decoupling , dynamical heterogeneities , and the hopping processes . in this paper ,
such connections are explored by extending the idealized mode - coupling theory ( mct ) for glass transition @xcite .
the idealized mct has been known as the most successful microscopic theory for glass transition . indeed , extensive tests of the theoretical predictions carried out so far against experimental data and computer - simulation results suggest that the theory deals properly with some essential features of glass - forming liquids @xcite . on the other hand ,
a well - recognized limitation of the idealized mct is the predicted divergence of the @xmath3-relaxation time at a critical temperature @xmath4 also referred to as the nonergodic transition which is not observed in experiments and computer simulations .
an extended version of mct developed in ref .
@xcite aims at incorporating activated hopping processes which smear out the sharp nonergodic transition and restore ergodicity for @xmath5 , but its applicability has been restricted to schematic models .
this is because of the presence of the subtraction term in the expression for the hopping kernel , which violates the positiveness a fundamental property of any correlation spectrum .
there have been relatively few other attempts to go beyond the idealized mct @xcite , and incorporating the hopping processes into the theory for glass transition has been a major unsolved problem .
we present here a new formulation which is motivated by ideas from the dynamical theory originally developed to describe diffusion - jump processes in crystals @xcite .
the dynamical - theory approach adapted in this work to glass - forming liquids treats hopping as arising from vibrational fluctuations in quasi - arrested state where particles are trapped inside their cages , and the hopping rate is formulated in terms of the debye - waller factors characterizing the structure of the quasi - arrested state .
the resulting expression for the hopping rate takes an activated form , and the barrier height for the hopping is `` self - generated '' in the sense that it is present only in those states where the dynamics exhibits a well defined plateau .
it will be discussed how such a hopping rate can be incorporated to develop a new extended version of mct .
we will then investigate whether the developed theory accounts for the mentioned decoupling for @xmath6 .
such an investigation makes sense since @xmath4 is also found to be close to @xmath7 @xcite , i.e. , the onset temperature of the decoupling @xcite .
it will also be examined whether our theory reproduces characteristic features of dynamical heterogeneities revealed by recent computer simulations .
more specifically , we shall study whether the theory predicts a substantial increase of the non - gaussian parameter @xcite , double - peak structure in the probability distribution of particle displacements @xcite , and the presence of a growing dynamic length scale @xcite , which the idealized mct failed to reproduce .
the paper is organized as follows . in sec .
[ sec : theory ] , we formulate our new extended mct .
numerical results of the theory are presented in sec .
[ sec : results ] for a model of the lennard - jones system , and connections of the hopping processes with the breakdown of the se relation and with aspects of dynamical heterogeneities are discussed .
the paper is summarized in sec .
[ sec : conclusion ] . appendix [ appendix : gs - theory ] outlines a novel derivation of the hopping kernel formulated in ref .
@xcite , and appendix [ appendix : msd - ngp ] is devoted to the derivation of the extended - mct equations for the mean - squared displacement and the non - gaussian parameter .
we start from surveying basic features of the idealized mct @xcite ( see also appendix [ appendix : gs - theory ] ) .
a system of @xmath8 atoms of mass @xmath9 distributed with density @xmath10 shall be considered .
structural changes as a function of time @xmath11 are characterized by coherent density correlators @xmath12 . here
@xmath13 with @xmath14 referring to @xmath15th particle s position denotes density fluctuations for wave vector @xmath16 ; @xmath17 the liouville operator ; @xmath18 the canonical averaging for temperature @xmath19 ; @xmath20 the static structure factor ; and @xmath21 . within the zwanzig - mori formalism @xcite
one obtains the following exact equation of motion : [ eq : gle - phi ] @xmath22 here @xmath23 with boltzmann s constant @xmath24 , and the memory kernel @xmath25 describes correlations of fluctuating forces . introducing the laplace transform with the convention @xmath26 ( @xmath27 ) , eq .
( [ eq : gle - phi - a ] ) is equivalent to the representation @xmath28 \}.
\label{eq : gle - phi - b}\ ] ] under the mode - coupling approximation , the fluctuating forces are approximated by their projections onto the subspace spanned by pair - density modes @xmath29 .
the factorization approximation for dynamics of the pair - density modes yields the following idealized - mct expression for the memory kernel to be denoted as @xmath30 : [ eq : ideal - mct - m ] @xmath31 here @xmath32 , and the vertex function is given by @xmath33^{2},\ ] ] in terms of @xmath34 and the direct correlation function @xmath35 .
the idealized - mct equations ( [ eq : gle - phi ] ) and ( [ eq : ideal - mct - m ] ) exhibit a bifurcation for @xmath36 also referred to as the nonergodic transition at a critical temperature @xmath4 @xcite . for @xmath37 , the correlator relaxes towards @xmath38 as expected for ergodic liquid states . on the other hand , density fluctuations for @xmath5 arrest in a disordered solid , quantified by a debye - waller factor @xmath39 .
it is the factorization approximation [ _ cf . _
( [ eq : dynamic - factorization ] ) ] that leads to the nonergodic transition at @xmath4 .
therefore , one has to consider corrections , @xmath40 , to go beyond the idealized mct , which shall be quantified via the hopping kernel defined by [ eq : extended - memory ] @xmath41 the correction term reads @xmath42 , and the memory kernel @xmath43 can be expressed as @xmath44 .
\label{eq : extended - memory - b}\ ] ] as demonstrated in appendix [ appendix : gs - theory ] , one can derive based on eq .
( [ eq : extended - memory - a ] ) an expression for @xmath45 which is essentially the same as that of the extended mct of gtze and sjgren @xcite by applying the zwanzig - mori formalism to @xmath25 and then introducing the corresponding mode - coupling approximation for the `` memory kernel '' to @xmath25 . such a hopping kernel , however , retains the same problem mentioned in sec .
[ sec : introduction ] . instead , our attempt for the extension of the idealized mct is motivated by the following observation : substituting eq .
( [ eq : extended - memory - b ] ) into eq .
( [ eq : gle - phi - b ] ) yields for small @xmath46 @xcite @xmath47
\}. \label{eq : alternative - gle}\ ] ] dropping @xmath45 , this equation reduces to the one of the idealized mct : approaching @xmath4 from above , @xmath48 for small @xmath46 becomes larger , and so does @xmath49 , leading to the nonergodic transition at @xmath50 . in the presence of @xmath45 , on the other hand , the transition is cutoff since the third term in the denominator of eq .
( [ eq : alternative - gle ] ) becomes unimportant when @xmath48 becomes large .
the long - time dynamics of @xmath51 in this case is thus determined by @xmath45 for small @xmath46 .
this observation raises a possibility of constructing a new approximate theory for @xmath45 described below .
we first derive a rate formula for a hopping process in which an atom at @xmath14 jumps to a nearby site @xmath52 separated by an interparticle distance .
the presence of such a process at low temperatures has been revealed by computer simulations @xcite .
this will be done via the dynamical theory originally developed to describe diffusion - jump processes in crystals @xcite .
the approach adapted here to glass - forming liquids treats hopping as arising from vibrational fluctuations ( phonons ) in quasi - arrested state where particles are trapped inside their cages .
let us suppose that the quasi - arrested state characterized by the debye - waller factors @xmath53 can be described as a frozen , irregular lattice @xcite .
each particle then has a well defined equilibrium position @xmath54 within the lifetime of the quasi - arrested state , and we introduce the displacement from the equilibrium position via @xmath55 .
the essential feature of the hopping process is that a jumping atom passes over a barrier formed by neighbors which block a direct passage to the new site .
the criterion that determines whether or not a given fluctuation is sufficient to cause a jump is therefore concerned with the relative displacements of the atom and the saddlepoint .
we thus employ as a `` reaction coordinate '' @xcite , @xmath56 \cdot \hat{\vec s}$ ] , and assume that a hopping occurs when @xmath57 exceeds a critical value @xmath58 , which measures the size of fluctuation needed to cause a jump .
here @xmath59 denotes the saddlepoint position with respect to @xmath54 , and the scalar product selects only those fluctuations directed towards @xmath60 .
each phonon displaces a hopping atom towards the saddlepoint .
the phonon phases are random , but the displacements may occasionally coincide in such a way that a hopping process occurs .
the hopping rate @xmath61 can then be calculated from such a probability per unit time , and one obtains along the line described in ref .
@xcite with the isotropic debye approximation , @xmath62 $ ] , in terms of the sound velocity @xmath63 .
here @xmath64 with the debye wave number @xmath65 , and @xmath66 .
notice that the sound velocity here refers to the one in the quasi - arrested state , which is renormalized by the debye - waller factors @xcite . to emphasize this ,
the sound velocity shall be expressed as @xmath67 in terms of the ( longitudinal ) elastic modulus @xmath68 consisting of the equilibrium value @xmath69 , where @xmath70 , and an additional contribution for the quasi - arrested state , for which mct yields @xcite @xmath71 with @xmath72 c_{k } + \frac{1}{5 } [ k c_{k}^{\prime}]^{2 } \right\}.\ ] ] the hopping rate is then given by @xmath73 .
\label{eq : w - kac - debye-2}\ ] ] the pre - exponential factor represents a mean attack frequency while the exponential term of the activated form gives the probability that the system is found at the critical displacement @xmath58 .
in addition , the barrier height for the hopping is `` self - generated '' in the sense that it is determined by the plateau heights @xmath53 of the coherent density correlators , and is present only in those states where the dynamics exhibits a well defined plateau .
we next relate the hopping rate @xmath61 to the hopping kernel @xmath45 . our discussion becomes simpler if the tagged - particle density correlator @xmath74 the self part of @xmath51 is considered , so this case shall be considered first .
hereafter , quantities referring to the tagged particle shall be marked with the superscript or subscript `` @xmath75 '' .
the zwanzig - mori equation for @xmath74 has the same form as eq .
( [ eq : gle - phi - a ] ) with @xmath76 , @xmath77 , and @xmath78 replaced by @xmath79 , @xmath80 , and @xmath81 , respectively ; the idealized - mct kernel corresponding to eq .
( [ eq : ideal - mct - m - a ] ) is given by @xmath82 with @xmath83 ^{2 } / [ ( 2 \pi)^{3 } q^{4 } ] $ ] @xcite ; and eq .
( [ eq : extended - memory - b ] ) holds with @xmath77 , @xmath84 , and @xmath85 replaced by @xmath80 , @xmath86 , and @xmath87 , respectively .
the arrested part @xmath88 of the correlator @xmath74 is referred to as the lamb - mssbauer factor . in the absence of the hopping kernel ,
the idealized kernel @xmath89 for @xmath5 arrests at a plateau for long times whose height is given by @xmath90 @xcite , i.e. , there holds @xmath91 for small @xmath46 . substituting this into eq .
( [ eq : alternative - gle ] ) for @xmath92 yields for small @xmath46 @xmath93 , \label{eq : phis - small - z-1}\ ] ] which determines the @xmath3 relaxation of @xmath74 the decay from the plateau @xmath88 to zero in the presence of @xmath94 .
on the other hand , when the @xmath3 relaxation is dominated by hopping processes characterized by a rate @xmath95 , the van hove self correlation function @xmath96 , related to @xmath74 via the inverse fourier transform @xmath97 and proportional to the probability of finding the tagged particle at @xmath98 and @xmath11 @xcite , obeys a simple rate equation @xmath99.\end{aligned}\ ] ] we assume that only hoppings with @xmath100 are relevant , in which @xmath101 $ ] denotes the weighted average of interparticle distances . here
@xmath102 denotes the first minimum of the radial distribution function @xmath103 defining the first shell ; @xmath104 with @xmath105 gives the mean number of particles at distance between @xmath106 and @xmath107 ; and @xmath108 is the coordination number of the first shell . the quantity @xmath109 serves as an analogue of the lattice spacing in crystals .
since there is no site dependence in the hopping rate we formulated [ _ cf .
_ eq . ( [ eq : w - kac - debye-2 ] ) ]
, there holds @xmath110.\ ] ] fourier transforming this yields @xmath111 \ , \phi_{q}^{s}(t).\ ] ] assuming that @xmath112 are oriented at random , the summation @xmath113 is given by the orientational average multiplied by the number of sites satisfying @xmath100 , which is approximated by the coordination number @xmath114 of the first shell .
this leads to @xmath115 \phi_{q}^{s}(t).\ ] ] noticing that the @xmath3 relaxation of @xmath74 starts from the plateau @xmath88 , the laplace transform of this equation reads @xmath116 \ , \}. \label{eq : phis - small - z-2}\ ] ] by comparing eqs .
( [ eq : phis - small - z-1 ] ) and ( [ eq : phis - small - z-2 ] ) , one arrives at the following expression for the hopping kernel : @xmath117 / f_{q}^{s}. \label{eq : delta - self}\ ] ] the collective hopping kernel @xmath45 consists of the self and distinct parts , @xmath118 . in the present work
, we shall adopt a simple model for @xmath45 in which the distinct part describing possible correlated jumps is neglected : @xmath119 this model for @xmath45 looks oversimplified , but nontrivial theoretical predictions follow from such a simple model as will be demonstrated in sec .
[ sec : results ] .
equations ( [ eq : gle - phi - a ] ) and ( [ eq : extended - memory - b ] ) with eqs . ( [ eq : ideal - mct - m - a ] ) , ( [ eq : w - kac - debye-2 ] ) , ( [ eq : delta - self ] ) , and ( [ eq : delta - collective ] ) constitute our new extended - mct equations for the coherent density correlator @xmath51 ; corresponding equations hold for the tagged - particle density correlator @xmath74 with the aforementioned replacement of @xmath76 , @xmath77 , @xmath78 , @xmath84 , and @xmath85 by @xmath79 , @xmath80 , @xmath120 , @xmath121 , and @xmath87 , respectively . the extended - mct equations for the mean - squared displacement and the non - gaussian parameter , which are required for our discussion in sec .
[ sec : results ] , are derived in appendix [ appendix : msd - ngp ] .
all these equations can be solved provided @xmath34 and @xmath122 are known as input .
( we notice that @xmath53 and @xmath88 can be obtained based on the knowledge of @xmath34 @xcite . )
in the following , numerical results of the extended theory will be presented for the lennard - jones ( lj ) system in which particles interact via the potential @xmath123 .
@xmath34 shall be evaluated within the percus - yevick approximation @xcite .
this model has been studied in ref .
@xcite based on the idealized mct . from here on , all quantities are expressed in reduced units with the unit of length @xmath124 , the unit of energy @xmath125 ( setting @xmath126 ) , and the unit of time @xmath127 .
the dynamics as a function of @xmath19 shall be considered for a fixed density @xmath128 , for which the critical temperature of the idealized mct is found to be @xmath129 @xcite . for @xmath122 entering into the extended - mct equations , we set @xmath130 estimated in ref .
@xcite from migration properties of crystals .
we notice that this value of @xmath122 is consistent with the lindemann length @xcite .
we also confirmed that the results to be presented below do not rely on the specific value of @xmath130 : nearly the same results were obtained with other values of @xmath122 , as far as those values consistent with the lindemann length are chosen .
as a function of @xmath131 for reduced temperatures @xmath132 , @xmath133 , @xmath134 , @xmath135 , @xmath136 , @xmath137 , and @xmath138 ( from left to right ) .
the wave numbers are @xmath139 ( upper panel ) and @xmath140 ( lower panel ) , which correspond to the first - peak and first - minimum positions of @xmath34 , respectively .
the solid and dashed curves denote the results from the extended and idealized mct , respectively .
the arrows in the upper panel refer to the peak positions of the non - gaussian parameter @xmath141 for @xmath142 , @xmath137 , and @xmath138 ( _ cf .
[ subsec : dh ] ) . ]
figure [ fig : phi ] shows the coherent density correlators @xmath51 for representative reduced temperatures @xmath143 whose values are specified in the caption .
the wave numbers shown are @xmath139 ( upper panel ) and @xmath140 ( lower panel ) , which correspond to the first - peak and first - minimum positions of @xmath34 , respectively .
the dashed curves refer to the idealized - mct results which exhibit the ergodic ( @xmath144 ) to nonergodic ( @xmath145 ) transition at @xmath50 ( @xmath146 ) @xcite .
the solid curves denote the results from the extended mct .
it is seen from fig .
[ fig : phi ] that the solid curves for @xmath147 and @xmath133 are hardly affected by the hopping processes , but a slight deviation from the dashed curve is discernible in the @xmath3-relaxation regime of the solid curve for @xmath148 .
the solid curve for @xmath149 exhibits the same decay up to @xmath150 as the corresponding dashed curve , but the relaxation thereafter is considerably accelerated .
the effects from the hopping processes are drastic for @xmath151 where the idealized mct predicts the arrested dynamics at the plateau @xmath53 , whereas the corresponding solid curves from the extended theory relax to zero for long times
. it would be interesting to analyze these extended - mct results based on various scaling laws developed in ref .
such an analysis , however , shall be deferred to subsequent publications , and in the following , we will focus on the connections of the hopping processes with the breakdown of the se relation and with aspects of dynamical heterogeneities . here
we investigate the breakdown of the se relation based on the extended mct , and compare our theoretical prediction with related computer - simulation results and experimental data .
a convenient experimental measure of the breakdown is a product @xmath152 formed with the diffusion coefficient @xmath153 and the viscosity @xmath154 , which grows as the se relation fails @xcite . in the present study , the @xmath3-relaxation time @xmath155 of the coherent density correlator at the peak position @xmath156 of @xmath34 , defined via the convention @xmath157 , shall be used as a substitute for @xmath158 .
this is justified since the @xmath19 dependence of the @xmath3-relaxation time at the structure factor peak is known to track that of @xmath158 @xcite .
the diffusion coefficient is determined from the long - time asymptote @xmath159 of the mean - squared displacement @xmath160^{2 } \rangle$ ] , whose extended - mct equations are derived in appendix [ appendix : msd ] . as a function of the reduced temperature @xmath161 .
the solid and dashed curves , both referring to the left scale , denote the results from the extended and idealized mct , respectively .
circles ( right scale ) denote the @xmath162 data of salol taken from ref .
@xcite and plotted versus @xmath163 with @xmath164 k @xcite .
the arrow marks @xmath165 k of salol . both the left and right ordinates range over 1.3 decades .
the inset exhibits the ratio @xmath166 of the product @xmath167 to the one at a reference temperature where the se relation holds .
the prediction from the extended mct ( solid curve ) is compared with simulation results for a binary mixture of lennard - jones particles @xcite ( filled squares ) and a binary mixture of dumbbell molecules of elongation @xmath168 @xcite ( filled triangles ) .
the procedure for the comparison is detailed in ref .
@xcite . ]
figure [ fig : dtau ] shows the theoretical prediction for the product @xmath167 as a function of the reduced temperature @xmath161 .
the solid and dashed curves , both referring to the left scale , denote the results from the extended and idealized mct , respectively .
the idealized - mct result for @xmath167 varies little for @xmath144 ( the increase is only about 10% for the @xmath163 range shown in fig .
[ fig : dtau ] ) , and does not account for the breakdown of the se relation .
this reflects the universal @xmath3-scale coupling predicted by the idealized mct @xcite , according to which both the @xmath3-relaxation time @xmath155 and the inverse of the diffusivity @xmath169 exhibit a universal power - law behavior @xmath170 for small @xmath163 , and hence , the product @xmath167 approaches a constant for @xmath171 ( @xmath172 ) . on the other hand ,
the extended mct predicts the increase of the product @xmath167 for @xmath173 .
figure [ fig : dtau ] thus shows one of the main results of this paper that the hopping processes are responsible for the breakdown of the se relation .
also presented in fig .
[ fig : dtau ] is the comparison of the theoretical result with the @xmath162 data of salol taken from ref .
@xcite ( circles , right scale ) .
the experimental data are also plotted versus @xmath163 with @xmath164 k of salol determined in ref .
@xcite . a direct comparison between the theoretical result and the experimental data can be made by plotting them on logarithmic scales of the same range as done in fig . [
fig : dtau ] .
it is seen that the extended - mct result is consistent with the experimental data concerning the degree of the breakdown of the se relation . the inset of fig .
[ fig : dtau ] compares the extended - mct result ( solid curve ) with simulation results for a binary mixture of lennard - jones particles @xcite ( filled squares ) and a binary mixture of dumbbell molecules of elongation @xmath168 @xcite ( filled triangles ) . here
the comparison is done in terms of the ratio @xmath166 of the product @xmath167 to the one at a reference temperature where the se relation holds .
( see ref .
@xcite concerning the details of the comparison . ) by definition , the ratio @xmath166 is unity if the se relation holds , whereas it exceeds unity as the se relation fails .
again , the theoretical prediction is consistent with the simulation results . on the other hand ,
recent measurements @xcite indicate that the self - diffusion coefficient is about 100 times faster near @xmath1 than that predicted by the se relation .
this is about a factor of 10 larger compared to our theoretical prediction near @xmath1 ( _ cf . _
[ fig : dtau ] ) , assuming that @xmath1 of the present system is located at @xmath174 estimated from the reduced temperature at @xmath1 of salol .
this implies that our model for the hopping kernel might be too primitive to be applicable near @xmath1 . in the following
, we shall therefore focus mainly on the regime @xmath175 where our theoretical prediction is consistent with the simulation results and experimental data .
we next explore a connection of the hopping processes with aspects of dynamical heterogeneities .
we start from the discussion on the non - gaussian parameter @xmath141 which characterizes deviations from the gaussian behavior of the van hove self correlation function @xcite : @xmath176 - 1 .
\label{eq : ngp - def}\ ] ] here @xmath177^{4 } \rangle$ ] .
general properties of @xmath141 revealed by computer simulations can be summarized as follows @xcite : ( i ) on the time scale at which the motion of particles is ballistic , @xmath141 is zero ; ( ii ) upon entering the time scale of the @xmath178 relaxation where the density correlators are close to their plateaus , @xmath141 starts to increase ; and ( iii ) on the time scale of the @xmath3 relaxation , @xmath141 decreases to zero , reflecting diffusive dynamics at long times which is a gaussian process .
it is also observed that the maximum value of @xmath141 and the time at which this maximum is attained both increase with decreasing @xmath19 @xcite .
positive @xmath141 means that the probability for a particle to move very far is enhanced relative to the one expected for a random - walk process .
the peak height of @xmath141 has therefore been interpreted as a measure of the dynamical heterogeneity that reflects different local environments around an individual particle . as a function of @xmath179 based on the extended mct for reduced temperatures @xmath180 , @xmath133 , @xmath134 , @xmath135 , @xmath136 , @xmath181 , and @xmath137 ( from left to right ) .
the corresponding results from the idealized mct , but for @xmath144 ( i.e. , @xmath37 ) only , are presented as dashed curves in the inset to avoid the overcrowding of the figure .
notice about an order of magnitude difference in the ordinate scales .
] the extended - mct equations for determining the non - gaussian parameter are derived in appendix [ appendix : ngp ] , and the resulting @xmath141 are plotted in fig .
[ fig : ngp ] for reduced temperatures @xmath161 specified in the caption .
the corresponding results from the idealized mct , but for @xmath144 ( i.e. , @xmath37 ) only , are presented as dashed curves in the inset to avoid the overcrowding of the figure . as noticed at the end of appendix [ appendix : ngp ] , the peculiar behavior @xmath182 for short times as predicted by our theory simply reflects that the ideal - gas contribution to the memory kernel is discarded in the mode - coupling approach .
such an ideal - gas contribution is responsible for the short - time ballistic regime , but is irrelevant as far as the long - time dynamics is concerned .
it is seen from the inset of fig .
[ fig : ngp ] that the peak height of @xmath141 predicted by the idealized mct does not grow with decreasing @xmath19 , and it is underestimated by almost an order of magnitude compared to that reported in computer simulations @xcite .
this defect has already been known from the theoretical work in ref .
@xcite , and has been demonstrated more explicitly in ref .
the main panel of fig .
[ fig : ngp ] , on the other hand , indicates that a substantial improvement is achieved by the extended mct in that the peak height grows upon lowering @xmath19 to an extent as observed in simulations @xcite .
( notice an order of magnitude difference in the ordinate scales of the main panel and the inset . )
thus , the extended mct reproduces a feature of the dynamical heterogeneity characterized by the non - gaussian parameter , and this is accomplished via the inclusion of the hopping processes . in terms of the non - gaussian parameter @xmath141 ,
the dynamics is most heterogeneous in the late-@xmath178 or early @xmath3 regime where the density correlators start to decay from the plateau ( see the upper panel of fig . [ fig : phi ] ) .
recently , it has been recognized from studies of four - point density correlation functions that there exists dynamical heterogeneity on a much longer time scale comparable to the @xmath3-relaxation time @xmath155 @xcite . recognizing that @xmath141 is dominated by those particles which move farther than expected from a gaussian distribution of particle displacements , flenner and szamel introduced a new non - gaussian parameter @xmath183^{-1}(t ) - 1 $ ] with @xmath184^{-1}(t ) \equiv
\langle 1/ [ { \vec r}_{s}(t ) - { \vec r}_{s}(0 ) ] ^{2 } \rangle$ ] , which weights strongly the particles which have not moved as far as expected from the gaussian distribution @xcite .
it is found that the peak position of @xmath185 is located on the time scale of @xmath155 , and hence , characterizes the longer - time dynamical heterogeneity .
it is also observed that the peak position of @xmath185 corresponds to the time at which two peaks in the probability distribution of particle displacements , reflecting populations of mobile and immobile particles ( see below ) , are of equal height .
this implies that the presence of the longer - time dynamical heterogeneity can be examined also through such a probability distribution .
the mentioned probability distribution @xmath186 of the logarithm of particle displacements @xmath187 at time @xmath11 can be obtained from the van hove self correlation function @xmath188 by the transformation @xcite @xmath189 the probability distribution is defined such that the integral @xmath190 is the fraction of particles whose value of @xmath191 is between @xmath192 and @xmath193 .
if the motion of a particle is diffusive with a diffusion coefficient @xmath153 , there holds @xmath194 \exp ( - \delta r^{2 } / 4 d t)$ ] @xcite . as argued in ref .
@xcite , the shape of the corresponding probability distribution under the diffusion approximation becomes time independent with the peak height @xmath195 . of the logarithm of single particle displacements based on the extended mct at the reduced temperature @xmath196 ( solid curves ) .
the times shown are @xmath197 , @xmath198 , @xmath199 , @xmath200 , @xmath201 , and @xmath202 ( from left to right ) .
the @xmath3-relaxation time @xmath155 at this reduced temperature is @xmath203 ( _ cf .
[ fig : phi ] ) .
the dashed curves refer to the probability distribution under the diffusion approximation for @xmath204 and @xmath202 .
the inset exhibits the time evolution of the mean - squared displacement @xmath205 at @xmath142 on double logarithmic scales ( solid curve ) .
the dotted curve refers to the diffusion asymptote , @xmath206 . ]
figure [ fig : psr ] shows the extended - mct result for the probability distribution @xmath207 ( solid curves ) at the reduced temperature @xmath196 .
the times shown are @xmath197 , @xmath198 , @xmath199 , @xmath200 , @xmath201 , and @xmath202 .
the @xmath3-relaxation time @xmath155 at this reduced temperature is @xmath203 ( _ cf .
[ fig : phi ] ) .
the dashed curves refer to the probability distribution under the diffusion approximation for @xmath204 and @xmath202 .
the inset exhibits the mean - squared displacement @xmath205 at @xmath142 on double logarithmic scales , from which one understands that the times chosen in the main panel range from the late-@xmath178 , plateau regime to the final @xmath3 regime where @xmath208 .
the appearance of the plateau in @xmath205 is due to particles being caged @xcite , and the peak of the probability distribution @xmath207 for @xmath197 reflects populations of such `` immobile '' particles . at later times
, the second peak develops in the probability distribution at @xmath209 ( i.e. , @xmath210 ) , reflecting `` mobile '' particles hopping over interparticle distances .
one infers from fig .
[ fig : psr ] that the time at which the two peaks in @xmath207 become of equal height is located on the time scale of @xmath155 , and this is consistent with the simulation result @xcite .
subsequently , the double - peak structure disappears , and the probability distribution approaches the one well described by the diffusion approximation .
coexistence of mobile and immobile particles is a direct indication of the dynamical heterogeneity @xcite .
our results shown in fig . [ fig : psr ] indicate that the hopping processes are responsible for such a double - peak structure in the probability distribution @xmath207 occurring on the time scale @xmath155 of the @xmath3 relaxation .
this also explains why the idealized mct failed to reproduce the double - peak structure in @xmath207 @xcite .
figure [ fig : psr ] also implies that the probability distribution approaches its diffusion asymptote only after its peak position exceeds those length scales where the double - peak structure in @xmath207 is observable .
thus , there is a certain length scale above which fickian diffusion sets in . in the following
, we shall quantify such a length scale characterizing the crossover from non - fickian to fickian diffusion , and investigate its temperature dependence .
to this end , let us introduce the ratio @xmath211 of the product @xmath212 to the one at some reference temperature , which is an analogue of the ratio @xmath166 studied in the inset of fig .
[ fig : dtau ] . here
the @xmath3-relaxation time @xmath213 of the tagged - particle density correlator is defined via the convention @xmath214 . under the diffusion approximation
, there holds @xmath215 @xcite .
thus , the product @xmath212 is constant , and hence , the ratio @xmath211 is unity , if the dynamics on the length scale @xmath216 is diffusive . on the other hand ,
the ratio @xmath211 exceeds unity if the dynamics on the length scale @xmath216 is non - fickian .
we shall therefore define the crossover wave number @xmath217 via @xmath218 , i.e. , as the wave number at which the ratio @xmath211 reaches 10% above unity .
the onset length scale of fickian diffusion shall then be defined via @xmath219 .
of the product @xmath212 to the one at a reference temperature @xmath220 as a function of the reduced temperature @xmath221 . here
@xmath220 is chosen such that @xmath222 .
the wave numbers shown are @xmath139 ( solid curve ) , 4.0 ( dashed curve ) , 2.0 ( dashed - dotted curve ) , and 1.0 ( dotted curve ) .
the horizontal dashed line marks @xmath223 chosen to determine the crossover wave number @xmath217 introduced in the text .
the inset exhibits the onset length scale of fickian diffusion @xmath219 as a function of the reduced temperature @xmath163 . ]
figure [ fig : length ] exhibits the ratio @xmath211 for wave numbers @xmath139 ( solid curve ) , 4.0 ( dashed curve ) , 2.0 ( dashed - dotted curve ) , and 1.0 ( dotted curve ) , as a function of the reduced temperature @xmath221 . the reference temperature @xmath220 for calculating @xmath211 is chosen such that @xmath222 .
the horizontal dashed line marks @xmath223 chosen to determine the crossover wave number @xmath217 introduced above . for reduced temperatures @xmath224 , all the ratios @xmath211 shown in fig .
[ fig : length ] are less than 1.1 , implying that the tagged - particle dynamics on the length scales comparable to and larger than @xmath225 can be well described by the fickian diffusion law . for @xmath226 , @xmath211 for @xmath139
exceeds 1.1 , meaning that the dynamics is diffusive only on length scales larger than 0.86 .
the onset length scale @xmath227 of fickian diffusion increases to @xmath228 at @xmath229 , and to @xmath230 at @xmath231 , as can be inferred from fig .
[ fig : length ] .
the length scale @xmath227 so obtained as a function of the reduced temperature @xmath163 is summarized in the inset of fig .
[ fig : length ] .
thus , the extended mct predicts the presence of a growing dynamic length scale .
it is anticipated that @xmath227 is intimately related to the mean size of the dynamic clusters observed in simulations @xcite , since fickian diffusion , i.e. , a random - walk process , is possible only over those length scales where the coherence length of such clusters is smeared out .
in this paper , we developed a new extended version of mct for glass transition .
the activated hopping processes are incorporated via the dynamical theory , originally formulated to describe diffusion - jump processes in crystals and adapted in this work to glass - forming liquids .
the dynamical - theory approach treats hopping as arising from vibrational fluctuations in quasi - arrested state where particles are trapped inside their cages , and the hopping rate is formulated in terms of the debye - waller factors characterizing the structure of the quasi - arrested state .
the resulting expression for the hopping rate takes an activated form , and the barrier height for the hopping is `` self - generated '' in the sense that it is present only in those states where the dynamics exhibits a well defined plateau .
it is discussed how such a hopping rate can be incorporated to develop a new extended mct . the extended mct deals with the interplay of two effects . nonlinear interactions of density fluctuations ,
as described by the idealized memory kernel @xmath48 , lead to the cage effect with a trend to produce arrested states for sufficiently low temperatures .
phonon assisted hoppings , taken into account via the hopping kernel @xmath45 , lead to the @xmath3 relaxation at all temperatures and restore ergodicity .
the interplay of these two effects is described by the memory kernel @xmath43 via eq .
( [ eq : extended - memory - b ] ) which has the form of a dyson equation . as demonstrated in sec .
[ sec : results ] for a model of the lennard - jones system , it leads to nontrivial theoretical predictions concerning the breakdown of the se relation and characteristic features of dynamical heterogeneities , which the idealized version of the theory failed to reproduce .
in dense liquids , relaxation is necessarily connected with rearrangements of large complexes of particles .
we have seen in fig .
[ fig : ngp ] that the peak height of the non - gaussian parameter @xmath141 grows substantially in the late-@xmath178 regime as the temperature is decreased . this enhanced probability for a particle to move further is what one would expect as a result of the building of a backflow in the liquid , and it is anticipated that the string - like motions observed in the late-@xmath178 regime @xcite reflect such a backflow pattern .
the backflow was originally discussed by feynman and cohen and taken into account in their theory for roton spectrum in liquid helium @xcite .
subsequently , it was found that the quantum - mechanical analogue of the idealized memory kernel reproduces the feynman - cohen result for roton spectrum @xcite , i.e. , the backflow phenomenon is within the scope of the idealized mct @xcite .
however , the idealized memory kernel alone does not account for such a pronounced peak in @xmath141 as observed in computer simulations @xcite .
this implies that the mentioned interplay with the hopping kernel plays a relevant role for the substantial increase of @xmath141 and the building of the backflow .
the double - peak structure in the probability distribution @xmath207 of particle displacements , which is most pronounced on the time scale @xmath155 of the @xmath3-relaxation and disappears at longer times ( _ cf .
[ fig : psr ] ) , reflects coexistence of mobile and immobile particles with a life time @xmath232 , and is a direct indication of the dynamical heterogeneity @xcite .
the extended mct developed here provides a natural explanation for its origin in terms of the cage and hopping effects . as argued in sec .
[ subsec : length ] , the life time of the double - peak structure gives rise to a growing dynamic length scale @xmath227 , which is associated with the decoupling of the time scales of the @xmath3 processes occurring on length scales smaller than @xmath227 from the diffusion coefficient .
the breakdown of the se relation discussed in sec .
[ subsec : breakdown ] can also be understood in terms of the decoupling of the @xmath3-relaxation time @xmath155 from the diffusivity which occurs when @xmath227 exceeds the average interparticle distance @xmath233 .
it is anticipated that the onset length scale @xmath227 of fickian diffusion is intimately related to the mean size of the dynamic clusters observed in simulations @xcite , since fickian diffusion , i.e. , a random - walk process , is possible only over those length scales where the coherence length of such clusters is smeared out .
then , our picture for the breakdown of the se relation is consistent with that of ref .
@xcite , where the connection between decoupling phenomena and the growing coherence length scale is discussed .
it is certainly necessary to improve further the theory developed here .
first of all , possible correlated hopping effects are discarded in our model for the hopping kernel .
this might explain why our theory underestimates the degree of the breakdown of the se relation near @xmath1 , which we referred to in connection with fig .
[ fig : dtau ] . that the development of the length scale @xmath227 seems suppressed for @xmath151 , as can be inferred from the inset of fig .
[ fig : length ] , might also be related to such a defect of the present theory .
second , we did not determine @xmath122 a square of the ratio @xmath234 formed with the critical size @xmath58 of the phonon - assisted fluctuation needed to cause a hopping and the average interparticle distance @xmath109 ( _ cf .
[ sec : theory ] ) microscopically , but simply took a value ( @xmath130 ) from the literature . regarding this ,
let us mention that nearly the same results as those presented in sec .
[ sec : results ] can be obtained with other values of @xmath122 , as far as those values consistent with the lindemann length @xcite are chosen .
thus , our theoretical results do not rely on the specific value of @xmath130 .
but , of course , it is desirable to determine @xmath122 consistently within the theory . third , the @xmath3-relaxation stretching is not enhanced at low temperatures .
for example , the stretching exponents @xmath235 , obtained via kohlrausch - law fits of the density correlators shown in fig .
[ fig : phi ] , for @xmath140 are 0.76 , 0.79 , and 0.80 for @xmath142 , @xmath137 , and @xmath138 , respectively .
one possibility to overcome this defect is to take into account distribution of hopping rates or barrier heights , which can be done , e.g. , by considering fluctuations in the debye - waller factors and @xmath122 .
four - point density correlators might be necessary for this purpose , though it is a difficult task to calculate such higher - order correlators .
but , in view of the significant results achieved by our theory , it is promising to pursue its further development .
the author is grateful to w. gtze and b. kim for discussions .
he also thanks h. sillescu for sending him the diffusivity data of salol presented in ref .
@xcite , and w. kob for the simulation results of ref .
this work was supported by grant - in - aids for scientific research from the ministry of education , culture , sports , science and technology of japan ( no . 20740245 ) .
in order to go beyond the idealized mct , one has to consider corrections to the idealized memory kernel , @xmath40 . the extended mct of gtze and sjgren @xcite can be considered as a theory for such corrections , and yields an expression for @xmath236 in terms of the hopping kernel @xmath45 that involves couplings to currents .
their extended theory is formulated with the generalized kinetic theory for phase - space density fluctuations , but essentially the same expression for the hopping kernel @xmath45 can be derived also with the standard projection - operator approach based on density and current - density fluctuations . in this appendix , we outline such a novel derivation .
we start from reviewing the derivation of the idealized memory kernel @xmath30 ( see ref .
@xcite for details ) .
let us introduce a projection operator @xmath237 onto the subspace spanned by the density fluctuations @xmath238 and the current - density fluctuations @xmath239
( @xmath240 ) . here
@xmath241 denotes the @xmath242 component of the velocity of @xmath15th particle . within the zwanzig - mori formalism @xcite ,
one obtains based on the operator @xmath237 the exact equation ( [ eq : gle - phi - a ] ) for the coherent density correlator @xmath51 .
the formal expression for the memory kernel entering there reads @xmath243 where @xmath244 , and the fluctuating force is given by @xmath245 here @xmath246 . under the mode - coupling approach ,
the fluctuating forces are approximated by their projections onto the subspace spanned by pair - density modes @xmath29 ; this is done by introducing the second projection operator @xmath247 and approximating @xmath248 , i.e. , @xmath249 notice here that the _ static _
factorization approximation @xmath250 is already used in the definition ( [ eq : p2-def ] ) of @xmath251 . within the convolution approximation for triple density correlations
, one finds for the projected fluctuating force @xmath252 substituting eq .
( [ eq : projected - f - a ] ) into eq .
( [ eq : m - before - factorization ] ) and then using the _ dynamical _ factorization approximation @xmath253 which factorizes averages of products evolving in time with the generator @xmath254 into products of averages formed with variables evolving with @xmath17 , one obtains the idealized kernel @xmath30 given in eqs . ( [ eq : ideal - mct - m ] ) . to extend the idealized theory , one has to avoid the use of the dynamical factorization approximation ( [ eq : dynamic - factorization ] ) .
we shall therefore start from eqs .
( [ eq : m - before - factorization ] ) and ( [ eq : projected - f - b ] ) for the memory kernel @xmath25 , to which the approximation ( [ eq : dynamic - factorization ] ) has yet to be applied . notice , however , that the static factorization approximation ( [ eq : static - factorization ] ) is still assumed in our approach ; this is employed in the definition ( [ eq : p2-def ] ) of @xmath251 , which is then used in deriving eq .
( [ eq : projected - f - b ] ) .
thus , there holds @xmath255 at @xmath256 . by introducing a new projection operator @xmath257 onto the subspace spanned by @xmath258 ,
i.e. , @xmath259 one obtains the following zwanzig - mori equation of motion for @xmath25 : @xmath260 here the factor @xmath261 in front of the convolution integral is just for later convenience , and the formal expression for the `` memory kernel '' @xmath262 is given by @xmath263 in terms of the `` fluctuating force '' @xmath264 where we have introduced @xmath265 .
substituting eq .
( [ eq : projected - f - b ] ) into eq .
( [ eq : r - def ] ) , one finds @xmath266 in which @xmath32 .
now , we apply the mode - coupling approximation to the kernel @xmath262 . because of
the odd time reversal symmetry of @xmath267 , the simplest mode - coupling approximation for @xmath262 can be introduced by defining the projection operator @xmath268 onto the subspace panned by the product @xmath269 of the density and current - density modes : @xmath270 where we have used the factorization approximation @xmath271 here and in the following , we use abbreviations @xmath272 and @xmath273 .
we thus obtain under the mode - coupling approximation @xmath274 , @xmath275 within the convolution approximation for triple density correlations , one finds @xmath276 in which @xmath277 is given by @xmath278 here @xmath279 refers to the @xmath242 component of the vector @xmath280 , and the function @xmath281 is given by @xmath282 .
when eq .
( [ eq : mct - r - a ] ) is substituted into eq .
( [ eq : mct - l-0 ] ) , the kernel @xmath262 is expressed in terms of four - mode correlators , for which we invoke the following dynamical factorization approximation : @xmath283 \nonumber \\ & & \quad - \ , \delta_{{\vec k } , { \vec p}^{\prime } } n^{2 } s_{k } s_{p } ( \hat{k}_{\mu } \hat{p}_{\lambda } /kp ) \partial_{t } \phi_{k}(t ) \partial_{t } \phi_{p}(t).\end{aligned}\ ] ] here @xmath284 and @xmath285 are longitudinal and transversal current correlators , respectively , which are normalized to unity at @xmath256 .
we then obtain the following form for the kernel @xmath262 that involves couplings to current modes : [ eq : mct - l ] @xmath286 .
\label{eq : mct - l - a}\end{aligned}\ ] ] the vertex functions @xmath287 , @xmath288 , and @xmath289 are expressed in terms of the thermal velocity , the static equilibrium quantities , and @xmath277 given in eq .
( [ eq : mct - r - b ] ) as @xmath290^{2},\ ] ] @xmath291^{2 } \bigr\},\end{aligned}\ ] ] @xmath292 \nonumber \\ & & \qquad \times \ , \bigl [ \sum_{\lambda } \frac{\hat{p}_{\lambda}}{p } v_{\lambda}({\vec q } ; { \vec k } , { \vec p } \ , ) \bigr].\end{aligned}\ ] ] finally , let us connect the laplace transform of @xmath262 to the hopping kernel @xmath45 using the definition given by eq .
( [ eq : extended - memory - a ] ) . the laplace transform of eq .
( [ eq : gle - m ] ) reads @xmath293 in view of this expression , let us formally introduce the function @xmath294 in terms of the laplace transform of the idealized memory kernel @xmath48 via @xmath295 here we have used the equality @xmath255 at @xmath296 noticed above .
substituting eqs .
( [ eq : laplace - gle - m ] ) and ( [ eq : laplace - gle - m - ideal ] ) into eq . ( [ eq : extended - memory - a ] ) , one obtains for the hopping kernel @xmath297 from the functional form of @xmath262 given in eq .
( [ eq : mct - l - a ] ) and interpreting that @xmath294 in eq .
( [ eq : delta - appendix ] ) subtracts those contributions already accounted for by the idealized memory kernel @xmath48 , one understands that the expression ( [ eq : delta - appendix ] ) for the hopping kernel is essentially the same as the one derived by gtze and sjgren @xcite .
the equation of motion for the mean - squared displacement @xmath205 can be obtained from eq .
( [ eq : gle - phi - a ] ) for @xmath74 by exploiting its relation to the small-@xmath298 behavior of @xmath299 @xcite : @xmath300 here we have introduced the @xmath301 limit of the memory kernel via @xmath302 . to obtain @xmath303 , it is more convenient to rewrite eq .
( [ eq : extended - memory - b ] ) in the form @xmath304 from which one finds [ eq : mct - msd ] @xmath305 here @xmath306 refers to the corresponding @xmath301 limit of the idealized memory kernel , and @xmath307 is defined via @xmath308 .
the former is given by @xcite @xmath309 whereas the latter reads after carrying out the @xmath301 limit in eq .
( [ eq : delta - self ] ) @xmath310 the non - gaussian parameter @xmath141 defined in eq .
( [ eq : ngp - def ] ) can be obtained from the mean - squared displacement @xmath205 and the mean - quartic displacement @xmath311 . the extended - mct equations for @xmath311
can be derived using the same method employed above for @xmath205 , but with higher order expansions in @xmath298 .
since @xmath311 is proportional to the fourth taylor coefficient in the small-@xmath298 expansion of @xmath312 @xcite , one can derive the following equation from the small-@xmath298 behavior of eq .
( [ eq : gle - phi - a ] ) for @xmath74 : @xmath313 = 0 .
\label{eq : gle - mqd}\end{aligned}\ ] ] here we introduced the memory kernel @xmath314 via the small-@xmath298 expansion of @xmath315 . using the corresponding expansion for the idealized memory kernel @xmath89 and the expansion @xmath316 for the hopping kernel , one obtains from eq .
( [ eq : extended - memory - for - appendix ] ) [ eq : mct - mqd ] @xmath317 the expression for @xmath318 reads @xcite @xmath319 , \label{eq : mct - mqd - b}\end{aligned}\ ] ] while one obtains from the small-@xmath298 expansion of eq .
( [ eq : delta - self ] ) @xmath320 / 3 .
\label{eq : mct - mqd - c}\ ] ] here , @xmath321 is defined via the small-@xmath298 expansion of the lamb - mssbauer factor @xmath322 @xcite
. equations ( [ eq : gle - mqd ] ) and ( [ eq : mct - mqd ] ) along with eqs .
( [ eq : mct - msd ] ) constitute the extended - mct equations for @xmath311 .
the non - gaussian parameter @xmath141 can then be obtained from eq .
( [ eq : ngp - def ] ) .
let us consider the short - time behavior of @xmath141 based on the derived equations . from eqs .
( [ eq : gle - msd ] ) and ( [ eq : gle - mqd ] ) we find for short times @xmath323 t^{4 } + o(t^{6}).\end{aligned}\ ] ] substituting these results into eq .
( [ eq : ngp - def ] ) , one obtains @xmath324 + o(t^{2 } ) .
\label{eq : ngp - short - times}\ ] ] since @xmath325 , we obtain from eq .
( [ eq : mct - mqd - b ] ) @xmath326 . using this result
, one can show that @xmath327 based on eqs .
( [ eq : mct - mqd - a ] ) and ( [ eq : mct - mqd - c ] ) . according to eq .
( [ eq : ngp - short - times ] ) , this means that the initial value of @xmath141 within the extended mct ( and also within the idealized mct which is based on newtonian dynamics ) is given by @xmath328 this is in disagreement with the exact initial behavior @xmath329 @xcite .
this discrepancy simply reflects that the ideal - gas contribution to the memory kernel is discarded in the mode - coupling approach . indeed , the ideal - gas contribution yields @xmath330 @xcite , which when substituted into eq .
( [ eq : ngp - short - times ] ) recovers @xmath329 .
the ideal - gas contribution to the memory kernel is responsible for the short - time ballistic regime , but is irrelevant as far as the long - time dynamics is concerned . | we develop a new extended version of the mode - coupling theory ( mct ) for glass transition , which incorporates activated hopping processes via the dynamical theory originally formulated to describe diffusion - jump processes in crystals . the dynamical - theory approach adapted here to glass - forming liquids treats hopping as arising from vibrational fluctuations in quasi - arrested state where particles are trapped inside their cages , and the hopping rate is formulated in terms of the debye - waller factors characterizing the structure of the quasi - arrested state .
the resulting expression for the hopping rate takes an activated form , and the barrier height for the hopping is `` self - generated '' in the sense that it is present only in those states where the dynamics exhibits a well defined plateau .
it is discussed how such a hopping rate can be incorporated into mct so that the sharp nonergodic transition predicted by the idealized version of the theory is replaced by a rapid but smooth crossover .
we then show that the developed theory accounts for the breakdown of the stokes - einstein relation observed in a variety of fragile glass formers .
it is also demonstrated that characteristic features of dynamical heterogeneities revealed by recent computer simulations are reproduced by the theory .
more specifically , a substantial increase of the non - gaussian parameter , double - peak structure in the probability distribution of particle displacements , and the presence of a growing dynamic length scale are predicted by the extended mct developed here , which the idealized version of the theory failed to reproduce .
these results of the theory are demonstrated for a model of the lennard - jones system , and are compared with related computer - simulation results and experimental data . |
top - condensation models @xcite are interesting candidates for a realistic theory of electroweak symmetry breaking ( ewsb ) .
they are a particular case of models of dynamical ewsb @xcite , where the electroweak ( ew ) gauge symmetry is broken due to the condensation of fermion - antifermion bilinears . in top - condensation models
the condensates are made of known fermions , mainly of the top quark . in the minimal models
no new particles are postulated .
the fundamental degrees of freedom are only fermions and gauge bosons .
no fundamental scalar fields such as the standard model ( sm ) higgs field are present . on the other hand the spectrum of the theory includes one or more composite higgs particles .
they are fermion - antifermion bound states and play a similar role as the fundamental higgs particles in the sm and its extensions . besides the well - established @xmath2 gauge interactions a new four - fermion interaction is postulated .
it provides the dynamics responsible of ewsb and the generation of sm fermion masses at the same time . note that in general two different sectors are necessary for these purposes . in technicolor , for example , technicolor gauge interactions trigger ewsb while extended technicolor interactions are required to generate fermion masses . due to the non - perturbative nature of the new interaction some approximation
is needed .
calculations at first order in the @xmath3 expansion , where @xmath4 is the number of colors are often made , obtaining a version of the nambu - jona - lasinio ( njl ) model @xcite .
next - to - leading order calculation are also available @xcite .
the four - fermion term postulated in top - condensation models is generally seen as an effective interaction . at some high energy scale
@xmath5 an asymptotically free gauge interaction becomes strongly - coupled . for energies below @xmath5 the new interaction
is effectively described by operators constructed with the fields corresponding to the light ( @xmath6 ) degrees of freedom of the theory . at low energies
the most important non - renormalizable operators are the ones having the lowest mass dimension .
therefore dimension - six four - fermion operators are considered .
normally only four - fermion operators made of ( pseudo)scalar fermion bilinears are taken into account .
they are the ones leading to ( pseudo)scalar composite fields .
axial vector and vector fermion bilinears are ignored .
note , however , that the distinction between ( pseudo)scalar and ( axial)vector bilinears is ambiguous due to fierz identities .
the new interaction must violate the flavor symmetry , i.e. must be non - universal , in order to generate the observed fermion mass pattern .
topcolor models @xcite are examples of a theory of this type . in a second scenario the four - fermion interaction term acquires a more fundamental status .
it is assumed that the sm with the higgs sector being replaced by a general dimension - six four - fermion interaction is a ( non - perturbatively ) renormalizable theory @xcite .
this is the case if one or more non - gaussian ultraviolet stable fixed points are established beyond the point - like approximation @xcite .
in this note the four - fermion interaction term is used as a starting point .
therefore , besides the four - fermion and the sm gauge couplings , the scale @xmath5 at which the whole lagrangian is defined , is also a parameter of the theory . for @xmath5
much bigger than the electroweak scale , complementary to the njl approach a perturbative renormalization group analysis can be made @xcite .
expansion is difficult to justify @xcite . ]
this method , which incorporates the sm gauge interactions , provides reliable values for the top - quark and higgs - boson masses . in this approach
the information that composite higgs doublets appear at the scale @xmath5 is encoded in the compositeness condition .
to have a very high scale @xmath5 is , however , not very attractive because the theory suffers from fine - tuning in exactly the same way as the sm .
another important point is the one related to the distinction between fundamental and composite higgs particles .
if the compositeness scale @xmath5 is very high , the composite nature of these particles can not be directly verified by experiments in the near future .
a more interesting possibility is to have a scale @xmath5 not very much higher than the ew scale , @xmath7 . in this case
no fine - tuning problem appears . besides
, the generation of the scale @xmath5 could be explained from dimensional transmutation .
this would solve or avoid the hierarchy problem .
a perturbative renormalization group analysis can not be justified in this case .
topcolor assisted technicolor @xcite , and top - quark see - saw @xcite are examples of theories of this type . in these theories the njl approach
is widely used . in this paper
, we present a top - condensation model in which together with the ew symmetry the discrete @xmath0 symmetry is spontaneously broken .
a @xmath0-conserving four - fermion interaction term including the three quark generations is considered . under certain conditions
the @xmath0 symmetry is spontaneously broken and the @xmath0-violating phase of the cabibbo - kobayashi - maskawa ( ckm ) matrix is generated . besides , realistic mixing angles and quark masses are obtained .
leading order approximation in the @xmath3 expansion is adopted .
a more detailed treatment of the model is presented in @xcite . the first model with spontaneous @xmath0 violation was considered by t. d. lee in 1973 @xcite , the same year that m. kobayashi and t. maskawa published their important paper @xcite . in the lee model ( or its generalization for the case of 3 quark families ) two fundamental higgs fields are considered . for certain values of the parameters of the model the @xmath0-symmetric
effective potential has a @xmath0-non - symmetric minimum . as a result
the model has a number of @xmath0-violating interactions , namely @xmath8 boson exchange ( as in the sm ) , charged and ( flavor violating and flavor conserving ) neutral higgs boson exchange . flavor changing neutral currents ( fcncs ) are present in this model already at tree level .
this requires a mechanism for their suppression in order to avoid conflict with experimental data .
there are two potential problems in models of spontaneous @xmath0-symmetry breaking at the ew scale .
the first is the one related with fcncs .
it can be shown @xcite that the requirements of spontaneous @xmath0 symmetry breaking at the ew scale , absence of fcncs at tree level , and a realistic ckm matrix can not be simultaneously satisfied .
therefore , models presenting spontaneous @xmath0-symmetry breaking at the ew scale have tree - level non - diagonal yukawa coupling matrices . in order to have a realistic model a suppression mechanism is needed .
the second potential problem is a domain wall problem which arises in a cosmological context .
the effective potential has in general degenerate minima corresponding to complex conjugate vacuum expectation values ( vevs ) . at the ew phase transition domains with different signs of the vev phases
are formed .
these domains are separated by walls with energy density much bigger than the closure energy of the universe ( after taking into account the effect of the universe expansion ) @xcite .
if one considers this problem to be a serious one , some solution must be found in order that spontaneous @xmath0 symmetry breaking at the ew scale is viable .
this paper is organized as follows : we begin in section [ secc : minscheme ] by presenting the minimal scheme in top - condensation models .
we continue , in section [ secc:1fam ] , with a model having a four - fermion interaction term which include only quarks of the third family .
section [ secc:3fam ] corresponds to the central part of this paper . here
we consider a model with a four - fermion interaction term including the three quark families . in this model
the @xmath0 symmetry is broken by the vacuum . in section [ secc : conclusions ] the conclusions are given . in appendix
[ app : eff_potential ] the calculation of the effective potential in the leading @xmath3 expansion is sketched . in appendix
[ app : auxfield_bases ] convenient basis changes for the auxiliary fields are given . in appendices [ app : two - point_1fam ] and [ app : two - point_3fam ] composite - field two - point functions are calculated for the models of sections [ secc:1fam ] and [ secc:3fam ] , respectively . in appendix
[ app : some_definitions ] some definitions used in section [ secc:3fam ] are given .
we start with the simplest lagrangian leading to ewsb in the context of top - condensation models .
@xmath0 is in this case a symmetry of the lagrangian and of the vacuum . in top - condensation models
the sm lagrangian without the higgs sector is considered . in its place a four - fermion interaction term made of sm fermions
is postulated : @xmath9 where the first sum is over all left- and right - handed fermions of the theory and the second contains the 3 yang - mills terms of the sm - symmetry group , @xmath10 .
the lagrangian @xmath11 is locally invariant under this symmetry group . in the simplest model @xcite the four - fermion interaction term @xmath12 ,
is given by @xmath13 where @xmath14 , and @xmath15 and @xmath16 are the top and bottom quark fields . the @xmath17 and @xmath18 indices are suppressed . a color - index contraction in each parenthesis and a @xmath18-index contraction between @xmath19 and @xmath20 are understood .
the coupling constant @xmath21 has mass dimension @xmath22 .
the theory is defined at the scale @xmath5 with all heavier degrees of freedom integrated out .
all momentum integrals of the theory are regularized using @xmath5 as a spherical cutoff . in order to study the vacuum of the theory
it is very convenient to use the auxiliary field formalism @xcite , especially if one goes beyond the minimal scheme as we do in the next sections .
the formalism is also useful for studying next - to - leading order corrections in the @xmath3 expansion @xcite .
let us introduce a scalar auxiliary field @xmath23 which possesses the same quantum numbers as the fermion bilinear @xmath24 , i.e. , the quantum numbers of the higgs doublet field in the sm .
the term @xmath12 is replaced by @xmath25 : @xmath26 with real mass parameter @xmath27 and yukawa coupling constant @xmath28 .
the lagrangian @xmath25 depends on the auxiliary field @xmath23 only quadratically .
one recovers @xmath12 , with @xmath29 , either integrating out the field @xmath23 from the generating functional in the path integral formulation , or imposing the constrains over the non - dynamical field @xmath23 ( euler - lagrange equations ) .
the vacuum of the theory is obtained minimizing the effective potential related to the field @xmath23 . from appendix [ app : eff_potential ]
we have the effective potential in the @xmath30 limit : @xmath31 with @xmath32 . from the minimum condition which involve the first derivative of @xmath33 with respect to @xmath34 one gets the gap equation @xmath35 this equation has in general two solutions . a symmetrical one located at @xmath36 , and , if @xmath37 , a second one with @xmath38 which breaks the ew symmetry .
evaluating the second derivative of @xmath39 at these two points it can be seen that the non - symmetrical solution corresponds to the minimum of the effective potential if @xmath40 . in this case the corresponding value of @xmath34 ( or equivalently of @xmath41 ) at the minimum is given by the solution of @xmath42 with @xmath43 .
note that in order to fulfill the last equation having @xmath44 , fine - tuning of @xmath21 is needed .
thus , if @xmath40 the electroweak symmetry is spontaneously broken to the @xmath45 electromagnetic one .
as in the sm , three goldstone bosons appear and the higgs mechanism provides the gauge bosons with masses . the fourth degree of freedom @xmath46 of the auxiliary field @xmath23 describes a scalar top - antitop bound state .
the inverse propagator of the field @xmath46 is given by @xmath47(k^2+m_t^2)}.\ ] ] the propagator has a pole located at @xmath48 .
the model predicts in this approximation a scalar bound state with a mass equal to twice the top - quark mass in eq .
( [ min1 ] ) in order to provide all the 6 quarks with masses in such a way that @xmath12 can be rewritten with help of only one composite higgs field ( see below ) , then the mass of the bound state remains almost unaffected @xmath49 . ] .
the auxiliary field @xmath23 plays a similar role as the higgs field in the sm . for this reason we call it ( composite ) higgs field in the following .
we generalize eq . ( [ min2 ] ) for the case in which the four chiral fields of the third family of quarks interact .
we consider the most general ( dimension 6 ) gauge - invariant four - quark interaction term involving the four chiral fields that can be written as a sum of products of fermion bilinears with the quantum numbers and lorentz structure of the sm higgs boson @xmath50,\ ] ] where @xmath51 .
+ due to the hermiticity of the lagrangian @xmath21 and @xmath52 are real .
one can set @xmath53 also real ( or positive ) by redefining one of the right - handed fermion fields . in this way the interaction term @xmath12 possesses only real coupling constants and
the complete lagrangian is invariant under a conventional @xmath0 transformation - term . ] . repeating the procedure followed in the previous section
we introduce @xmath54 spin - zero auxiliary fields @xmath55 .
the term @xmath12 is replaced by @xmath56 with @xmath57 where @xmath58 and @xmath59 are the real yukawa coupling constants and @xmath60 are mass parameters associated with the auxiliary fields .
the relations between the coupling constants in the two formulations of the model are given by @xmath61 in order to parameterize the space of couplings @xmath62 , it is enough to consider @xmath63 .
we restrict ourselves to @xmath63 in the following . from appendix [ app : eff_potential ]
we obtain the effective potential in the @xmath30 limit : @xmath64 with the @xmath65 matrix @xmath66 given by @xmath67 where summation over the indices @xmath68 and @xmath69 is understood .
the ground state of the theory is found by minimizing the effective potential with respect to the auxiliary fields @xmath70 and @xmath71 . due to the gauge invariance it is possible to gauge any field configuration into the following form : @xmath72 with @xmath34 , @xmath73 , @xmath74 . in the following @xmath34 ,
@xmath73 , @xmath75 , @xmath76 denote the classical fields and the corresponding non - primed symbols denote their vevs , @xmath77 in order to preserve the electromagnetic @xmath45 symmetry , the vev @xmath78 must be zero .
besides , if the phase @xmath79 is not a trivial one , the @xmath0 symmetry could be spontaneously broken .
this happens if , once non - trivial quark masses are generated , the coupling constant @xmath53 in the fermion mass basis is complex .
next we minimize the effective potential with respect to the variables @xmath34 , @xmath73 , @xmath76 , and @xmath75 .
we shall restrict ourselves to the parameter subspace with @xmath80 and search for local minima in this region .
it is possible to show @xcite that for @xmath81 there is no local minimum ( at least for @xmath82 ) .
the following conditions are sufficient in order to have a local minimum at a point with @xmath83 : @xmath84 the conditions a ) evaluated at the point @xmath85 , @xmath86 , @xmath87 , and @xmath80 are given by @xmath88 = 0,\ ] ] @xmath89 = 0,\ ] ] @xmath90 where @xmath91 the first derivative of the effective potential with respect to @xmath92 is given by @xmath93 using eq .
( [ 1fam12 ] ) , with @xmath94 , the last expression can be written as @xmath95 because we are studing the possibility of having spontaneous @xmath0 symmetry breaking , the interesting case is when both auxiliary field doublets condense ( @xmath96 ) and @xmath1 . below .
] we look for a local minimum of this form .
however , using @xmath1 we immediately note that eq .
( [ 1fam13 ] ) can only be fulfilled if @xmath97 thus , higgs vevs are all real and @xmath0 is a symmetry of the theory .
the first derivative conditions eqs .
( [ 1fam11 ] ) and ( [ 1fam12 ] ) determine the top and bottom quark masses .
these equations can be fulfilled by choosing the adequate values of the mass parameters @xmath98 and @xmath99 . in a similar way as for the minimal scheme , these conditions are fine - tuned if @xmath100 .
considering only terms of order @xmath101 in eq .
( [ 1fam16 ] ) , condition b ) is equivalent to @xmath102 for given couplings @xmath103 , the last inequality can be fulfilled by choosing the sign of @xmath104 properly .
finally we check condition c ) .
the hessian of the effective potential with respect to @xmath34 , @xmath73 , and @xmath76 is given by @xmath105 with @xmath106 .
to obtain eq .
( [ 1fam17 ] ) the first derivative conditions were used .
considering condition b ) which determines the sign of @xmath107 one can see that the last matrix has 3 positive eigenvalues . in summary
, we considered the model defined by ( [ 1fam1 ] ) with @xmath1 when both auxiliary fields condense breaking the ew symmetry but respecting the @xmath45 electromagnetic symmetry .
this occurs if eqs .
( [ 1fam11 ] ) and ( [ 1fam12 ] ) are fulfilled . for @xmath100
these two equations are fine - tuned .
the sign of @xmath107 is determined by condition ( [ 1fam16b ] ) .
furthermore spontaneous @xmath0 symmetry breaking does not occur .
the spectrum of the model is calculated in appendix [ app : two - point_1fam ] .
we found that only 4 of the 8 degrees of freedom related to the two auxiliary higgs doublets are relevant at low energies .
three of them are goldstone bosons and the fourth is a physical higgs particle with mass @xmath49 . at energies
much lower than @xmath5 this model can not be distinguished from the one which has only one auxiliary higgs doublet from the beginning ( in this case the condition @xmath108 must hold , see eq . ( [ 1fam4 ] ) ) .
the case of having @xmath109 is qualitatively very different because the lagrangian possesses a peccei - quinn @xmath45 symmetry . in this case
it is possible to introduce two auxiliary fields in such a way that one couples only to the field @xmath110 and the other only to the field @xmath111 ( 2 higgs doublet ( 2hd ) model type ii ) . if @xmath112 both auxiliary fields condense and one obtains a particle spectrum as follows : in the neutral sector two scalars with masses equal to @xmath113 and @xmath114 , a goldstone boson , and an axion appear . in the charged sector one
obtains a charged goldstone boson and a charged particle with mass @xmath115 .
we see that both , for the case of having @xmath109 and for @xmath1 ( at least when both auxiliary fields condense ) the number of parameters at low energies is two .
once one fixes them , e.g. the quark masses , the model is completely specified . for the case
@xmath1 one would like to understand why the number of parameters is only two and not three .
the reason is that in order to satisfy eqs .
( [ 1fam11 ] ) and ( [ 1fam12 ] ) simultaneously ( with @xmath1 ) , the following relation must hold : @xmath116 for a given @xmath53 ( @xmath117 ) , the values of @xmath21 and @xmath52 which fulfill the last equation describe a hyperbola .
in order to generate the @xmath0-violating phase of the ckm matrix and thus reproduce the observed mechanism of @xmath0 symmetry violation , quarks of the three generations must be included in the four - fermion interaction .
we are interested in the situation in which the @xmath0 symmetry is spontaneously broken by the vacuum .
therefore our starting point is a @xmath0-invariant lagrangian with the four - fermion interaction given by a generalization of eq .
( [ 1fam1 ] ) : @xmath118,\ ] ] where the coupling constants @xmath62 and the quark fields @xmath119 , @xmath120 , @xmath121 have indices @xmath68 , @xmath69 , @xmath122 , @xmath123 , which go from the first to the third quark generation .
the lagrangian @xmath12 includes four - fermion terms which mix quark fields of different quark families . because we demand the lagrangian to be @xmath0-invariant , all the coupling constants @xmath62 in @xmath12
are considered to be real . in this case
the number of parameters of eq .
( [ 3f1 ] ) is 171 .
, 81 couplings @xmath124 , and further 45 couplings @xmath125 . ] where we do not count the cutoff @xmath5 which is also a parameter of the model .
we now rewrite the interaction term , eq .
( [ 3f1 ] ) , in terms of auxiliary fields which have the quantum numbers of the sm higgs doublet field as we did in the previous sections .
we do not consider completely arbitrary couplings @xmath62 .
for simplicity we restrict ourselves to the subset for which the theory can be described by means of only two auxiliary fields , @xmath70 and @xmath71 . in terms of these auxiliary fields the four - fermion term @xmath12 is replaced by @xmath126 where @xmath127 with @xmath128 and @xmath129 @xmath130 real matrices .
the relations between the real coupling constants in the two formulations of the model are given by @xmath131 in consequence we restrict ourselves to a model with 36 , essentially the four @xmath130 yukawa matrices , parameters ( plus the cutoff scale @xmath5 ) .
we are now confronted with the following problem .
we want to find the values of the parameters of the model such that the vacuum of the theory breaks the ew symmetry in the observed way ( spontaneous @xmath0 symmetry violation keeping the electromagnetic @xmath45 symmetry unbroken ) . besides , the generated ckm matrix and the quark masses must correspond to their measured values .
however , the relation between the 36 parameters of our model and the quantities to be reproduced is rather complicated . in order to find an analytical solution
we introduce a self - consistent approach to the problem .
first , we assume that the minimum of the effective potential ( denoted by non - primed symbols ) is given by a field configuration with non - trivial values of @xmath132 , @xmath133 , and @xmath79 ( @xmath134 ) and with @xmath135 : @xmath136 vevs of this form are necessary in order to have a theory with spontaneous @xmath0 symmetry breaking and unbroken @xmath137 symmetry . inserting the higgs vevs in the yukawa interactions , eq .
( [ 3f3 ] ) , one gets the quark mass term : @xmath138 with @xmath139 .
the quark mass matrices for the up- and down - type quarks in the last equation are in general non - diagonal . in order to diagonalize them
we perform the following chiral rotations : @xmath140 where the primed fields denote the fermion fields in the mass basis and @xmath141 , @xmath142 , @xmath143 , @xmath144 are basis transformation matrices .
the ckm matrix is given by @xmath145 .
in the new basis the quark mass matrices , which are now diagonal and real are given by @xmath146 where the yukawa couplings in the mass basis @xmath147 , @xmath148 are defined by @xmath149 we emphasize that the higgs vevs in eq .
( [ 3f7 ] ) are still not determined .
using the last definitions , the relations between the composite fields @xmath55 and their constituent quark fields are given by : @xmath150 for @xmath151 . besides
, the interaction term @xmath152 is given in this basis by @xmath153 with @xmath154 .
combining eqs .
( [ 3f7 ] ) and ( [ 3f8 ] ) and using the fact that the matrices @xmath128 and @xmath129 are real , it is possible to write the yukawa couplings in the weak basis , @xmath128 , @xmath129 , as a function of the higgs vevs , the quark masses , and the basis transformation matrices : @xmath155,\\ g^{(2 ) } & = \frac{\sqrt{2}}{w\,\sin\eta } \ ; \mathcal{i}m(u^u m_u w^{u\dagger}),\\ h^{(1 ) } & = \frac{\sqrt{2}}{v}\;\big[\mathcal{r}e(u^d m_d w^{d\dagger } ) + \cot\eta \ ; \mathcal{i}m(u^d m_d w^{d\dagger})\big],\\ h^{(2 ) } & = -\frac{\sqrt{2}}{w\,\sin\eta } \ ; \mathcal{i}m(u^d m_d w^{d\dagger}).\\ \end{split}\ ] ] we have in this way transformed the original problem into one which can be solved in a self - consistent way : we must find values of @xmath132 , @xmath133 , and @xmath79 , to which we associate the yukawa couplings given in eqs .
( [ 2hd5 ] ) , such that the resulting effective potential has its minimum at the same values @xmath132 , @xmath133 and @xmath79 ( besides @xmath135 ) . for this purpose
we can vary the basis transformation matrices @xmath141 , @xmath142 , @xmath143 and @xmath144 .
these are arbitrary unitary matrices which must obey the condition @xmath156 . the effective potential related to the four - fermion interaction ( [ 3f1 ] )
is given in appendix [ app : eff_potential ] .
a detailed analysis of the local minimum conditions given in ( [ 1fam10 ] ) is presented in @xcite .
the conditions obtained in order to have a model presenting spontaneous @xmath0 symmetry violation , realistic quark masses , and the observed ckm matrix are : \i ) the first derivative conditions associated with the variables @xmath34 and @xmath73 which are given by @xmath157 where the @xmath158 matrices @xmath159 are given in appendix [ app : some_definitions ] .
these conditions can easily be fulfilled by choosing suitable mass parameters @xmath60 . for @xmath160
both conditions require fine - tuning .
\ii ) complex higgs vevs are obtained only if the factor @xmath161 , defined in appendix [ app : some_definitions ] , is different from zero . in this case , the first derivative condition associated with the variable @xmath76 leads to : @xmath162 where the real factors @xmath163 , @xmath16 , and @xmath161 are given in appendix [ app : some_definitions ] .
it can be seen that for @xmath164 this first derivative condition also requires from fine - tuning if @xmath100 .
\iii ) the alignment condition @xmath165 takes the following form : @xmath166 with the @xmath130 matrix @xmath167 and @xmath168 given in appendix [ app : some_definitions ] .
\iv ) the hessian matrix associated with the variables @xmath34 , @xmath73 , and @xmath76 must be positive definite .
the hessian matrix is given by @xmath169 with @xmath170 , and @xmath76 .
the first derivatives of the @xmath158 matrix @xmath66 are given in appendix [ app : some_definitions ] .
there are no quadratically divergent terms in eq .
( [ 3f25 ] ) ( @xmath171 ) .
they cancel after imposing the fine - tuned first derivative conditions . once one has a set of parameters @xmath128 , @xmath129 , @xmath172 which fulfills the previous conditions , the corresponding couplings @xmath62 are obtained using eqs .
( [ 1fam4 ] ) . for a detailed analysis of these conditions using a quark mass expansion
see @xcite .
the composite - field two - point functions , which allow us to find the composite higgs masses , are calculated for the charged and neutral sectors in appendix [ app : two - point_3fam ] . beside the three goldstone bosons
there are three neutral and one charged composite higgs bosons . in this approximation
we find that one neutral higgs mass is @xmath173 and that the rest of the higgs masses are much smaller .
we think this is a consequence of the crude approximation we adopted and not a property of the theory . for further comments
see our conclusions .
in this paper we considered a top - condensation model with spontaneous @xmath0 symmetry breaking .
we started with a @xmath0 invariant four - fermion interaction defined at some scale @xmath5 which involve quarks of the three generations .
we restricted ourselves to the case in which the four - fermion interaction can be rewritten with help of two auxiliary fields .
the minimum of the effective potential is obtained in a self - consistent manner .
this leads to the relevant conditions on the parameters of the model in order to obtain a realistic ckm matrix , including the @xmath0-violating phase , and the quark masses .
the following conditions are found : i ) first derivative conditions related to the variables @xmath34 and @xmath73 ( eqs .
( [ 2hd16 ] ) ) , ii ) the condition for spontaneous @xmath0 symmetry breaking @xmath164 , iii ) the preservation of the electromagnetic symmetry condition , and iv ) the second derivative conditions .
a more detailed treatment of these conditions including an analysis using a quark mass expansion can be found in @xcite .
the spectrum of the theory corresponds to the one of the 2hd extension of the sm .
the higgs particles are quark - antiquark bound states .
the related composite fields are given ( at the scale @xmath5 ) by eq .
( [ int11b ] ) .
thus , due to the hierarchy of the yukawa couplings the composite higgs particles are made mainly of the quarks of the third generation having their wave functions only a small light - quark component . in our approximation only one ( neutral ) composite higgs has an acceptable mass value ( @xmath49 ) . the other two
neutral and the charged composite higgs are unacceptably light .
we think , this could change beyond the @xmath3 expansion used here .
for example for @xmath174 a perturbative renormalization group approach where the sm gauge interactions are taken into account , leads to much bigger composite higgs masses @xcite .
spontaneous @xmath0 symmetry breaking at the ew scale presents two potential problems , namely , tree level fcncs @xcite and a domain wall problem @xcite .
we just comment ( in relation to the first potential problem ) that in the context of top - condensation models it is not understood why fcncs are suppressed ( if we simultaneously generate quark mixing angles ) independently of having spontaneous or explicit @xmath0 symmetry violation .
we also treated a simpler top - condensation model where only quark fields belonging to the third generation are included in the four - fermion interaction ( see eq .
( [ min2 ] ) ) .
this lagrangian is used as an intermediated step in theories like topcolor assisted technicolor @xcite or top - quark see - saw @xcite .
we found that in the general case ( without an extra peccei - quinn symmetry ) the spectrum of the theory corresponds to the one of the sm with the composite neutral higgs particle having a mass @xmath173 .
further modes associated with masses of order @xmath5 are present .
@xmath5 is , however , the upper limit of the validity range of the model .
for this reason these modes can not be interpreted as particles . in this simpler model
spontaneous @xmath0 symmetry breaking does not occur .
i would like to thank werner bernreuther for useful discussions and support during all stages of this work . +
this work was supported by the deutschen akademischen austauschdienst ( daad ) .
we present in this appendix the effective potential for @xmath54 scalar auxiliary fields @xmath55 coupled to 3 quark generations .
we consider the leading order contributions in the @xmath3 expansion , i.e. in the @xmath175 limit keeping @xmath176 fixed , which is equivalent to the fermionic determinant approximation .
the lagrangian is given by @xmath177 where @xmath178 contains the quark kinetic terms and @xmath179 with complex parameters @xmath180 , @xmath181 , and real mass parameters @xmath60 . no kinetic term for the auxiliary fields @xmath55 and no quartic term of the form @xmath182 are present .
the effective potential is given by @xmath183 where @xmath184 is the fermionic propagator in momentum space , and is a function of the scalar fields @xmath55 and the momentum @xmath122 .
after calculating the fermionic determinant one obtains @xmath185 with @xmath186 where summation over the indices @xmath68 and @xmath69 is understood .
the effective potential is of course gauge invariant .
in this appendix we define new bases for the auxiliary fields in the neutral and charged sectors . we choose the new bases in such a way that the goldstone bosons become basis vectors .
the advantage is that in each sector , neutral and charged , one of the basis vectors of the two - point proper - vertex matrix is already an eigenvector with associated eigenvalue equal to zero , i.e. a pole of the propagator .
the goldstone theorem tells us how to express the goldstone fields as a function of the scalar fields .
they are given by the infinitesimal displacements of the vacuum under the transformation generated by the broken generators .
the charged and neutral goldstone boson fields are given by @xmath187 where the fields @xmath188 , @xmath189 are components of the higgs fields @xmath139 . in the four - dimensional neutral sector
we define the new basis by @xmath190 where the orthogonal transformation matrix @xmath191 is given by @xmath192 the field @xmath62 is the normalized goldstone boson given in eq .
( [ hgb8 ] ) . in this new basis
the mass term of the neutral bosonic fields is given by @xmath193 with @xmath194 now we turn to the charged sector formed by two charged fields .
the new basis is defined by @xmath195 where the charged field @xmath196 is the normalized charged goldstone boson given in eq .
( [ hgb7 ] ) . in this new basis the mass term for the charged scalar fields is given by @xmath197 we also give the yukawa couplings , eq .
( [ a2 ] ) , in these bases .
the neutral boson interaction terms are given in the fermion mass basis by @xmath198 where the dots represent analogous terms for the down sector .
for the charged sector we found @xmath199 where @xmath200 and @xmath201 are the left and right projectors .
the matrices @xmath202 , @xmath203 , and @xmath204 are @xmath130 in flavor space .
@xmath205 is the ckm matrix , the matrices @xmath203 with @xmath206 are the diagonal fermion mass matrices from the up and down sector , and the matrices @xmath204 are given by @xmath207 with @xmath206 .
in this appendix we calculate the masses of the neutral and charged composite higgs for the case of one family of quarks and @xmath208 considered in section [ secc:1fam ] .
we do the calculation in the auxiliary field bases defined in appendix [ app : auxfield_bases ] .
let us first calculate the composite higgs masses in the neutral sector . as we saw in section [ secc:1fam ]
the theory is @xmath0-invariant . as a consequence the @xmath0-even ( @xmath209 and @xmath210 ) and the @xmath0-odd ( @xmath211 and @xmath62 ) fields do not mix and the @xmath212 two - point proper vertex matrix is a block diagonal matrix .
the feynman diagrams we have to calculate are shown in fig .
[ fig:2hd_1fam_neutral ] . using the first derivative conditions in the calculation , the non - vanishing two - point functions are @xmath213 in the @xmath0-odd sector , and @xmath214 in the @xmath0-even sector .
the integral @xmath215 is given by @xmath216}.\ ] ] the masses of the bound states are given by the values of @xmath217 at which the proper vertex matrix has vanishing eigenvalues . from eqs .
( [ h65 ] ) and ( [ h66 ] ) we see that at @xmath218 a zero eigenvalue with associated eigenvector @xmath62 , the neutral goldstone boson , appears . in the @xmath0-even sector
the @xmath219 entry of the @xmath65 matrix is of order in this expression were zero , we had @xmath220 . ]
@xmath101 and therefore , for @xmath221 , much bigger than the other matrix elements . in first approximation
the smaller eigenvalue of this matrix is given by @xmath222 . from eq .
( [ h68 ] ) we see that the propagator has a pole at @xmath223 .
the other two eigenvalues , associated with a @xmath0-even and a @xmath0-odd field , are of order @xmath101 .
+ now we treat the charged sector in an analogous way . for the two - point proper vertices , represented by the diagrams of fig .
[ fig:2hd_1fam_charged ] , we obtain @xmath224\ , i(m_t^2,m_b^2;p^2 ) \\ & \qquad\qquad\qquad\qquad\quad + ( k_t^2-k_b^2 ) \left(m_t^2 \log(\lambda^2/m_t^2)-m_b^2 \log(\lambda^2/m_b^2)\right ) \big\ } , \end{split}\ ] ] @xmath225\ , i(m_t^2,m_b^2;p^2)\\ & \qquad\qquad\qquad\qquad\qquad + ( m_t k_t - m_b k_b ) \left(m_t^2 \log(\lambda^2/m_t^2)-m_b^2 \log(\lambda^2/m_b^2)\right ) \big\ } \end{split}\ ] ] @xmath226\ , i(m_t^2,m_b^2;p^2 ) \\ & \qquad\qquad\qquad\qquad + ( m_t^2-m_b^2 ) \left(m_t^2 \log(\lambda^2/m_t^2)-m_b^2 \log(\lambda^2/m_b^2)\right ) \big\ } , \end{split}\ ] ] with @xmath227}.\ ] ] we have here a similar situation as in the neutral @xmath0-odd sector .
the @xmath228 element is of order @xmath101 and the other elements of the matrix vanish at @xmath218
. therefore the two poles are located at @xmath218 and at @xmath229 .
@xmath230 @xmath231\\ & \qquad\qquad -\sum_{i = u , c , t}\frac{m_i^2}{\lambda^2}\log(\lambda^2/m_i^2)\ : \mathcal{r}e\big(\ : w^{u\dagger}\:\mathcal{r}e(w^u m_u u^{u\dagger } ) \:\mathcal{i}m(u^u m_u w^{u\dagger})\:w^u\big)_{ii } \\ & \qquad\qquad -\sum_{i = d , s , b}\frac{m_i^2}{\lambda^2}\log(\lambda^2/m_i^2)\ : \mathcal{r}e\big(\ : w^{d\dagger}\:\mathcal{i}m(w^d m_d u^{d\dagger } ) \:\mathcal{r}e(u^d m_d w^{d\dagger})\:w^d\big)_{ii } , \end{split}\ ] ] @xmath232\\ & \qquad\qquad -\sum_{i = u , c , t}\frac{m_i^2}{\lambda^2}\log(\lambda^2/m_i^2)\ : \mathcal{r}e\big(\ : w^{u\dagger}\:\mathcal{i}m(w^u m_u u^{u\dagger } ) \:\mathcal{i}m(u^u m_u w^{u\dagger})\:w^u\big)_{ii } \\ & \qquad\qquad -\sum_{i = d , s , b}\frac{m_i^2}{\lambda^2}\log(\lambda^2/m_i^2)\ : \mathcal{r}e\big(\ : w^{d\dagger}\:\mathcal{i}m(w^d m_d u^{d\dagger } ) \:\mathcal{i}m(u^d m_d w^{d\dagger})\:w^d\big)_{ii } , \end{split}\ ] ] @xmath233 @xmath234 @xmath235
in this appendix we calculate the composite - field two - point functions of the three family case considered in section [ secc:3fam ] .
the situation is analogous to the one of appendix [ app : two - point_1fam ] where only one family of quarks was considered .
the values of the flowing momentum for which an eigenvalue of the proper vertex matrix becomes zero , correspond to the composite higgs masses we are looking for .
we use the neutral and charged auxiliary field bases defined in appendix [ app : auxfield_bases ] . in the calculation we also apply the first derivative conditions from the minimization of the effective potential . the non - vanishing neutral proper vertices are given by @xmath236 \\ & + \sum_{r , s=1}^3 \sum_{q = u , d}\ : i(m_r^2,m_s^2;p^2 ) \big[(p^2-m_r^2-m_s^2 ) ( k_q)_{rs}(k_q^\dagger)_{sr } \\ & \qquad\qquad\qquad\qquad \mp m_r m_s ( ( k_q)_{rs}(k_q)_{sr}+(k_q^\dagger)_{rs}(k_q^\dagger)_{sr } ) \big ] \bigg\ } , \end{split}\ ] ] @xmath241\ , i(m_r^2,m_s^2;p^2 ) , \label{2pf_neutral2 } \\ i\gamma_{\varphi^3,\varphi^3}(p^2)&= & \frac{2i n}{8\pi^2(v^2+w^2 ) } \sum_{i=1}^6 \ : ( p^2 - 4m_i^2)m_i^2\ , i(m_i^2,m_i^2;p^2 ) , \label{2pf_neutral3 } \\
i\gamma_{g , g}(p^2)&= & \frac{2i n}{8\pi^2(v^2+w^2)}\,p^2\ , \sum_{i=1}^6 \ : m_i^2\ , i(m_i^2,m_i^2;p^2 ) , \label{2pf_neutral4 } \\ i\gamma_{\varphi^1,\varphi^3}(p^2)&= & \frac{i n}{8\pi^2(v^2+w^2 ) } \sum_{i=1}^6 \ : ( p^2 - 4m_i^2)\,\sqrt{2}\,m_i\,\mathcal{r}e[(k_{u / d})_{ii}]\ , i(m_i^2,m_i^2;p^2 ) , \label{2pf_neutral5 } \\ i\gamma_{\varphi^2,g}(p^2)&= & \frac{i n}{8\pi^2(v^2+w^2)}\,p^2\ , \sum_{i=1}^6 \ : \sqrt{2}\,m_i\,\mathcal{r}e[(k_{u / d})_{ii}]\ , i(m_i^2,m_i^2;p^2 ) , \label{2pf_neutral6 } \\ i\gamma_{\varphi^1,g}(p^2)&= & \frac{i n}{8\pi^2(v^2+w^2)}\,p^2\ , \sum_{i=1}^6 \ : \sqrt{2}\,m_i\,\mathcal{i}m[(k_{u / d})_{ii}]\ , i(m_i^2,m_i^2;p^2 ) , \label{2pf_neutral7 } \\ i\gamma_{\varphi^2,\varphi^3}(p^2)&= & \frac{i n}{8\pi^2(v^2+w^2 ) } \sum_{i=1}^6 \ : ( 4m_i^2-p^2)\,\sqrt{2}\,m_i\,\mathcal{i}m[(k_{u / d})_{ii}]\ , i(m_i^2,m_i^2;p^2 ) , \label{2pf_neutral8}\end{aligned}\ ] ] where in eqs .
( [ 2pf_neutral5])-([2pf_neutral8 ] ) @xmath242 stands for @xmath243 for @xmath244 and for @xmath245 for @xmath246 .
the relevant diagrams of the neutral sector are shown in fig .
[ fig:3fam_masses_neutral ] .
@xmath247 \\ & \qquad\qquad\qquad\qquad\qquad + ( v_{ckm } k_d^\dagger)_{rs } ( k_d v_{ckm}^\dagger)_{sr } \,\big [ j(m_r^2,m_s^2;p^2 ) + 2 m_r^2 \big ] \\ & \qquad-\big [ ( k_u^\dagger v_{ckm})_{rs } ( k_d v_{ckm}^\dagger)_{sr } + ( v_{ckm } k_d^\dagger)_{rs } ( v_{ckm}^\dagger k_u)_{sr } \big ] 2 m_r m_s\ , i(m_r^2,m_s^2;p^2 ) \big\ } , \end{split}\ ] ] in all the two - point proper vertices the quadratic divergences cancel .
as expected , the neutral and charged goldstone bosons correspond to eigenvectors of the proper vertex matrix at @xmath218 with vanishing eigenvalues . besides these three goldstone bosons there are three neutral and one charged higgs particles .
the masses of these particles are roughly @xmath252 , with @xmath253 a general quark mass .
we find that in this approximation one neutral higgs mass is @xmath173 and the rest of the higgs masses are much smaller .
y. nambu , in _
new theories in physics , proceedings of the xi international symposium on elementary particle physics _ , kazimierz , poland , 1988 , edited by z. ajduk , s. pokorski and a. trautman ( world scientific , singapure 1989 ) , pp . 1 - 10
. v. a. miransky , m. tanabashi and k. yamawaki , phys .
b * 221 * ( 1989 ) 177 .
v. a. miransky , m. tanabashi and k. yamawaki , mod .
lett . a * 4 * ( 1989 ) 1043 .
w. a. bardeen , c. t. hill and m. lindner , phys .
d * 41 * ( 1990 ) 1647 .
m. suzuki , phys .
d * 41 * ( 1990 ) 3457 .
h. j. he and c. p. yuan , phys .
* 83 * ( 1999 ) 28 [ arxiv : hep - ph/9810367 ] . c. x. yue , l. wang and d. q. yu , phys . rev .
d * 70 * ( 2004 ) 054011 [ arxiv : hep - ph/0409047 ] .
f. larios , f. penunuri and m. a. perez , phys .
b * 605 * ( 2005 ) 301 [ arxiv : hep - ph/0409291 ] .
j. j. cao , g. l. liu and j. m. yang , phys .
d * 70 * ( 2004 ) 114035 [ arxiv : hep - ph/0409334 ] . c. x. yue and z. j. zong , j. phys .
g * 31 * ( 2005 ) 401 [ arxiv : hep - ph/0501131 ] .
x. l. wang , q. l. zhang and q. p. qiao , phys .
d * 71 * ( 2005 ) 014035 [ arxiv : hep - ph/0501145 ] .
c. t. hill and e. h. simmons , phys .
* 381 * ( 2003 ) 235 [ erratum - ibid .
* 390 * ( 2004 ) 553 ] [ arxiv : hep - ph/0203079 ] .
y. nambu and g. jona - lasinio , phys .
* 122 * ( 1961 ) 345 .
g. cveti , e. a. paschos and n. d. vlachos , phys .
d * 53 * ( 1996 ) 2820 [ arxiv : hep - ph/9501268 ] .
g. cveti and n. d. vlachos , phys .
b * 377 * ( 1996 ) 102 [ arxiv : hep - ph/9512333 ] .
c. t. hill , phys .
b * 266 * ( 1991 ) 419 .
h. gies , j. jaeckel and c. wetterich , phys .
d * 69 * ( 2004 ) 105008 [ arxiv : hep - ph/0312034 ] .
m. a. luty , phys .
d * 41 * ( 1990 ) 2893 . c. d. froggatt , i. g. knowles and r. g. moorhouse , phys .
b * 249 * ( 1990 ) 273 . c. d. froggatt , r. g. moorhouse and i. g. knowles , nucl . phys .
b * 386 * ( 1992 ) 63 .
u. mahanta , phys . lett .
b * 281 * ( 1992 ) 320 .
g. cveti and c. s. kim , nucl .
b * 407 * ( 1993 ) 290 [ arxiv : hep - ph/9303301 ] .
c. t. hill , phys .
b * 345 * ( 1995 ) 483 [ arxiv : hep - ph/9411426 ] .
b. a. dobrescu and c. t. hill , phys .
* 81 * ( 1998 ) 2634 [ arxiv : hep - ph/9712319 ] .
r. s. chivukula , b. a. dobrescu , h. georgi and c. t. hill , phys .
d * 59 * ( 1999 ) 075003 [ arxiv : hep - ph/9809470 ] . c. valenzuela , ph .
d. thesis , rwth aachen , arxiv : hep - ph/0503289 .
m. gronau , a. kfir , g. ecker , w. grimus and h. neufeld , phys . rev .
d * 37 * ( 1988 ) 250 .
g. c. branco , phys .
lett . *
44 * ( 1980 ) 504 . y. b. zeldovich , i. y. kobzarev and l. b. okun , zh . eksp .
* 67 * ( 1974 )
3 [ sov . phys .
jetp * 40 * ( 1974 ) 1 ] .
t. kugo , prog .
* 55 * ( 1976 ) 2032 .
g. cveti , annals phys .
* 255 * ( 1997 ) 165 [ arxiv : hep - ph/9605433 ] .
g. cveti and e. a. paschos , nucl .
b * 395 * ( 1993 ) 581 .
m. harada and n. kitazawa , phys .
b * 257 * ( 1991 ) 383 . | we present a top - condensation model in which the @xmath0 symmetry is spontaneously broken at the electroweak scale due to the condensation of two composite higgs doublets . in particular the @xmath0-violating phase of the ckm matrix
is generated .
a simpler model where only one quark family is included is also discussed . in this case , for a general four - fermion interaction ( @xmath1 ) , the particle spectrum is the one of the one higgs doublet model .
10 march 2005 1.5 cm cristin valenzuela .5 cm 2.5 cm |
the identification of an earth - like planet will have a great impact on astronomy and science in general . a growing number of discovered extrasolar planets are fostering anticipation for reaching the goal .
indeed , several extrasolar planets , such as gj 667c c @xcite and kepler-22 b @xcite , are suspected of being earth - like on the basis of their orbit and mass / size .
such discoveries have spurred a demand for a methodology for characterizing a candidate as earth - like . because the polarization spectrum of planetary reflected light contains physical and compositional information about the planet , the spectropolarimetry of extrasolar planets has potential to be a strong tool for characterizing a planet .
ground - based polarimetry is advantageous because the degree of polarization is , in principle , not affected by telluric absorption .
@xcite calculated the optical polarization spectra of light reflected by a modeled earth - like exoplanet .
the model showed that the light reflected by an earth - like exoplanet can be highly polarized by light scattering in the atmosphere , and by surface reflection .
single scattering generally produces higher polarization than multiple scattering or reflection from a rough surface .
the degree of polarization in the continuum of the spectra studied by @xcite decreased with increasing wavelength .
this was explained by an increasing contribution of surface reflection to the total planetary reflected light due to a decrease in the atmospheric absorption optical thickness .
in addition , @xcite derived the presence of enhanced polarization features at gaseous molecule absorption bands , e.g. , the o@xmath3-a band ( 760 nm ) .
the higher polarization was the result of a lower fraction of multiply scattered light ( and thus a higher fraction of singly scattered light ) in an absorption band as compared with the nearby continuum region .
@xcite also examined the effect of cloud coverage and found that it reduced the polarization degree ; this was attributed to an increased fraction of multiple scattering . different scattering or reflecting media exhibit different phase - angle dependences of the polarization .
hence , the phase angle at which the maximum polarization degree occurs ( @xmath4 ) can be used to characterize exoplanets @xcite .
for instance , @xmath4 is 90@xmath0 for molecular rayleigh single scattering @xcite .
in contrast , @xmath4 for a particulate surface can be larger than 90@xmath0 @xcite .
in fact , it is known that the lunar ( airless solid ) surface has a @xmath4 of 95 - 110@xmath5 @xcite .
the earth , the only `` known earth - like planet '' , should be observed carefully in advance to facilitate the identification of an earth - like exoplanet .
nonetheless there have been very few observational attempts to obtain the polarization spectrum of the earth .
the satellite - borne instrument polder conducted polarimetry of the earth in three optical bands @xcite .
however , its limited field of view ( swath : 2400 km ) should be accompanied by a challenge in reproducing the polarization of the disk - integrated earthlight , which can be used as a direct comparison with a distant ( and thus point - source ) exoplanet .
we conducted spectropolarimetry of earthshine on the moon as an alternative approach .
earthshine is earthlight reflected from the lunar surface .
it is commonly used by ground - based observers to obtain the disk - integrated light of the earth because the spatial information in earthlight is averaged out in the lunar surface reflection .
although photometry ( e.g. , @xcite & 2004 ; @xcite ) , regular intensity spectrometry ( e.g. , @xcite & 2006 ; @xcite ) , and monochromatic polarimetry @xcite of earthshine have been conducted ( as reviewed by @xcite ) , spectropolarimetry had never been reported until the recent publication of @xcite , who observed earthshine when the moon was at the waxing and waning quadratures .
our observations were performed independently , and we successfully monitored phase variations in the earthshine polarization spectra .
the results allow us to discuss both the wavelength and phase dependence of the polarization , namely the wavelength dependence of @xmath4 ; hence , they could provide more information than an observation at a single phase
. note that in this paper , the light of the earth as viewed from the space is referred to as _ earthlight _ , which is distinguished from _
earthshine _ on the moon as observed from the ground .
our observations were made on five different nights : 2011 march 9 - 13 ( ut ) .
the moon waxed from a new crescent to the first quarter .
the parameters of our earthshine observations are summarized in table [ tab : log ] .
the earth phase angle @xmath6 , which is defined as the sun - earth - moon angle in this paper , ranged from 49@xmath0 to 96@xmath0 .
the views of the earth as seen from the moon are reproduced in fig .
[ fig : earthviews ] .
the contributions of different earth scenes to earthshine are shown in table [ tab : surface ] .
the cloud coverage was virtually constant throughout the observation period .
the ratio of the contribution from the ocean to that from the land decreased from 1.05 to 0.51 .
we used the spectropolarimeter hbs ( @xcite ) mounted at the cassegrain focus ( f/18 ) of the 1.88 m telescope at okayama astrophysical observatory , located in okayama , japan .
the selected diaphragm was d2 , which has twin rectangular holes aligned in the east - west direction .
their size is 0.2 mm @xmath7 1.4 mm each , which corresponds to 1.2@xmath8 @xmath7 8.5@xmath8 .
hbs employs a half - wave plate and a wollaston prism .
a polarization spectrum is obtained by rotating the half - wave plate by 0@xmath0 , 22.5@xmath0 , 45@xmath0 , and 67.5@xmath0 . in our observations ,
hbs provided polarization spectra at wavelengths of 450 - 850 nm with a resolution of 6 nm .
the attached ccd was cooled to @xmath9c .
an observational sequence consisted of moonshine ( ms ) , earthshine ( es ) , near sky ( nsky ) , and far sky ( fsky ) observations .
we obtained 1 - 4 sets of effective sequences each night .
for the ms spectra , the telescope was pointed to 12.9@xmath10 west of the moon s center .
the use of ms is described in section [ sec : reduce ] .
then es spectra were observed at 11.3@xmath10 east of the center .
nsky and fsky spectra for background subtraction were obtained by pointing toward the vicinity of the es , which was located at 20@xmath10 and 30@xmath10 east of the moon s center , respectively .
the single exposure time was 60 - 300 seconds for es , 30 seconds for nsky / fsky , and 10 - 30 seconds for ms .
dark frames were obtained every night .
the dark count was evaluated to be less than 1 adu on the basis of a previous ccd performance test .
two types of daylight flat frames were acquired to correct for the effect of the difference in the sizes of the twin holes on the diaphragm and of non - uniformity along the spatial axis of the spectral images .
the ccd pixel sensitivity correction was processed using a prepared master flat frame .
stellar observations were made through a glan - taylor prism for instrumental depolarization correction .
unpolarized and strongly polarized standard stars were observed for instrumental polarization correction .
spectra from an hg - ne lamp were obtained once in the observation period for wavelength calibration .
the frame - by - frame wavelength shift was corrected using the absorption feature in the telluric o@xmath3 a - band in the object frames .
.parameters of earthshine observations .
integration time is expressed as @xmath11 .
@xmath6 is the earth phase angle ( sun - earth - moon angle ) . [
cols="^,^,^,<,^",options="header " , ] we compared our results for 2011 march 12 ( @xmath12 ) and march 13 ( @xmath13 ) with earthshine spectropolarimetry on 2011 april 25 ( @xmath14 ) by @xcite ( figs . [
fig : es.p.all]d , [ fig : es.p.all]e ) .
the polarization degrees of our spectra have virtually the same values at 550 - 600 nm .
our spectra decrease more slowly with increasing wavelength than those of @xcite .
the difference between the two groups is @xmath152 % at the highest , which is within the range of the possible variation attributed to the changing cloud coverage @xcite .
the spectrum on 2011 june 10 ( @xmath16 ) by @xcite exhibits higher polarization degrees than their data in april and ours on march 12 - 13 .
this can be attributed to differences in the scene types ( e.g. , cloud coverage ) in the earthshine contributing area ( sterzik et al .
they detected an enhanced polarization feature in the o@xmath3 a - band at 760 nm .
it has a peak height of 0.5%-1% , which is comparable to the statistical error in our data .
it is generally safe to say that the observational results from the two groups are consistent .
we compared stam s ( 2008 ) model of an earth - like planet with our observation near a quadrature . for this purpose
, we needed to consider the effect of lunar reflection on the polarized earthlight spectrum .
lunar reflection does not add polarization because the earth - moon - earth phase angle is zero @xcite .
instead , it depolarizes the polarized earthlight . @xcite
estimated the depolarization factor to be @xmath153.3 .
this evaluation was based on a comparison of his optical polarimetric observations of earthshine with the roughly estimated polarization degree of earthlight , as well as on laboratory measurements of lunar samples .
the wavelength dependence of depolarization has not been well determined .
we consider both a wavelength - independent factor of 3.3 and linearly dependent factors , i.e. , @xmath17 $ ] ( sterzik et al .
2012 ) .
[ fig : compmodel ] shows the earthshine polarization spectrum observed on march 12 ( @xmath12 ) , compared with the stam s ( 2008 ) model spectra for four sets of the parameters .
although both the observed and modeled spectra exhibit decreasing polarization degree with increasing wavelength , the model spectra do not match those based on the observation .
@xcite made a similar analysis and also found considerable differences between their observations and the model . because the earthshine observations by the two groups are consistent , further model sophistication and/or re - examination of the lunar depolarization factor is expected to resolve this discrepancy .
( 80mm,80mm)fig5.eps fig .
[ fig : es.p.all]f shows that the overall degree of polarization increases as the earth phase approaches quadrature .
interestingly , with a more detailed look , the phase dependence appears to differ with the wavelength . to express this
, we average the degree of polarization in the @xmath1 , @xmath18 , and @xmath19 bands and plot them against the earth phase angle ( fig .
[ fig : phasep ] ) .
the result suggests that @xmath4 ( the phase angle at which the maximum polarization degree occurs ) is larger for longer wavelengths : @xmath4 for the @xmath1 band is @xmath2 , whereas @xmath4 for the @xmath19 band seems to be larger than @xmath20 .
@xmath4 varies depending on the type of scattering / reflection process . as described in section [ sec : intro ] , the polarization of earthlight ( thus of earthshine ) is caused by scattering in the atmosphere and surface reflection .
polarization from rayleigh scattering by atmospheric molecules has @xmath4 at 90@xmath5 @xcite . on the other hand
, @xmath4 for surface reflection can be larger than 90@xmath5 .
the total light reflected from a particulate surface can be understood as the sum of three components : surface singly scattered rays , volume - scattered rays , and multiply scattered rays @xcite . because the volume- and multiply scattered rays are generally only weakly polarized , the primary source of polarization in the total surface reflection is the surface singly scattered light , which can be expressed approximately by the fresnel equations for specular reflection @xcite .
typical refractive indices for soils , vegetation , and liquid water are @xmath151.55 @xcite , @xmath151.38 @xcite , and @xmath151.33 @xcite , for which the fresnel equations yield @xmath4 values of @xmath15114@xmath0 , @xmath15108@xmath0 , and @xmath15106@xmath0 , respectively .
actual measurements proved that @xmath4 can be larger than 90@xmath5 for surface reflection : for soils , polarimetry of a terrestrial soil ( together with desert sand ) , the moon and mercury ( i.e. , airless solid bodies ) revealed that their @xmath4 values are roughly 100 - 110@xmath5@xcite , 95 - 110@xmath5 @xcite , and 93 - 103@xmath5 @xcite , respectively ; for vegetation , @xmath4 values of 90 - 120@xmath5 or beyond were measured by @xcite and @xcite ; for liquid water , a @xmath4 of @xmath15106@xmath5 was reported by @xcite .
these estimates and measurements suggest that the earthshine polarization at shorter wavelengths is dominated by rayleigh scattering by atmospheric molecules , whereas that at longer wavelengths has an increasingly effective contribution from reflection at the land and ocean surfaces of the earth .
indeed , considering the wavelength dependences of the scattering / reflectance processes , the fraction of the atmospheric rayleigh scattering in the total earthlight should be much larger at shorter wavelengths than it is at longer wavelengths : the rayleigh scattering radiance has a strong negative correlation ( proportional to @xmath21 ) , whereas the reflectances of soils and vegetation have generally positive correlations ; that of liquid water has a linearly ( i.e. , slower than rayleigh scattering ) negative correlation , according to the aster spectral library . as additional processes for producing the observed wavelength dependence of @xmath4
, we discuss the changes in scene ( ocean / land ratio , cloud coverage ) in the earthshine - contributing area on the earth during our observing period .
since the cloud coverage was virtually constant throughout the observation period ( table [ tab : surface ] ) , it should not be effective .
the ocean / land ratio decreased toward the end of the observations ( table [ tab : surface ] ) .
as described , the estimated or measured @xmath4 for both liquid water and land ( soils and vegetation ) were typically greater than 90@xmath5 .
hence , the @xmath4 of @xmath1590@xmath5 can not be explained without atmospheric rayleigh scattering .
a decrease in the ocean fraction will increase the ratio of the surface ( ocean and land ) reflection to the atmospheric rayleigh scattering in the total earthlight because the land is brighter than the ocean , and thus may produce a larger @xmath4 .
this effect should be more apparent at longer wavelengths , given the wavelength dependences of atmospheric rayleigh scattering ( strongly negative ) and surface reflection ( positive or weekly negative ) .
hence , a decrease in the ocean / land ratio is at least qualitatively capable of producing the observed wavelength dependence of @xmath4 .
note that this process is consistent with the interpretation described in the previous paragraph because the determining factor in both discussions is the same : the wavelength dependences of atmospheric rayleigh scattering and surface reflection . to summarize the discussion of the observed wavelength dependence of phase variations in earthshine polarized spectra ,
the results should be interpreted as an outcome of the wavelength - dependent ratio of atmospheric rayleigh scattering to surface reflection .
earthshine polarization at shorter wavelengths is dominated by atmospheric rayleigh scattering , whereas that at longer wavelengths has an increasingly effective contribution from surface reflection .
it is qualitatively possible that a decrease in the ocean fraction in the earthshine - contributing area was an additional process contributing to the observed wavelength dependence of the phase variation .
( 80mm,80mm)fig6.eps the polarization of earthlight , which is estimated from the earthshine polarized spectrum , has implications concerning the search for earth - like extrasolar planets . assuming a depolarization factor of @xmath153.3 at the lunar surface @xcite ,
the polarization degree of earthlight should be as much as @xmath1530% at 450 nm and @xmath1520% at 850 nm when the earth phase is near quadrature .
on the other hand , airless solid planetary bodies have a lower degree of the maximum polarization ; maximum polarizations of 6%-22% at 435 nm and 5%-9% at 840 nm were measured for the moon @xcite , and values of 9%-10% at 450 nm and 6%-7% at 850 nm were measured for mercury @xcite .
our observations of moonshine are also within the ranges measured by @xcite .
these results suggest that earth - like planets have significantly higher polarization than do airless solid bodies . the wavelength dependence of @xmath4 , the phase angle of the maximum polarization ( subsection .
[ sec : phase ] ) , may also be used as a signature of earth - like planets .
for earthshine , and thus for earthlight , @xmath4 increases with increasing wavelength : @xmath4 in the @xmath1 band is a @xmath2 phase angle , whereas that for longer wavelengths appears at larger phase angles .
in contrast , @xmath4 for airless solid bodies , namely , the moon and mercury , decreases with increasing wavelength ( @xcite ; fig.19 in @xcite ; table vi in @xcite ) . finally , we compare our results with those for other types of terrestrial planets than airless solid bodies .
venus , a terrestrial planet covered with thick atmospheric clouds , exhibits a completely different polarimetric behavior , which is explained by the mie theory for cloud particles with a refractive index of @xmath151.44 and a mean radius of @xmath15 1.1 @xmath22 @xcite : `` negative '' polarization ( i.e. , polarization in the direction parallel to the scattering plane ) is observed at most visible wavelengths and 30 - 140@xmath5 phase angles ; @xmath4 is located around 15 - 20@xmath5 phase angles in the 350 - 600 nm region @xcite .
therefore , venus - like planets should be distinguishable from earth - like planets . for mars , a terrestrial planet with a thin atmosphere
, we can not access any near - full - phase polarimetric measurements in multiple wavelengths .
@xcite presented polarization measurements by the spacecraft mars-5 at phase angles of 60@xmath5 and 90@xmath5 at a wavelength of 592 nm , together with ground - based observations for phase angles smaller than 43@xmath5 .
the results generally resembled the polarization of the moon and mercury : the maximum polarization was extrapolated to be @xmath157% at a phase angle of @xmath15100@xmath5 .
the maximum polarization of mars is significantly smaller than that of the earth .
its wavelength dependence of @xmath4 is inferred to be similar to that of the moon , which has an opposite dependence to that of the earth .
note that uncertainty in the lunar depolarization factor may affect the estimated polarization degree of earthlight .
however , it does not affect @xmath4 ; the reason is that the depolarization factor at a certain wavelength should be conserved regardless of the earth phase angles because the earth - moon - earth angle is constant ( always zero ) .
we conducted spectropolarimetry of earthshine for five earth phase angles ranging from 49@xmath0 to 96@xmath0 . the observed spectra exhibit a decreasing earthshine polarization degree with increasing wavelength at any phase .
the overall degree of polarization increases as the earth approaches a quadrature phase .
the observed phase dependence varies with the wavelength ; the maximum polarization for the @xmath1 band occurs at a phase angle of @xmath2 , whereas that for longer wavelengths is reached at larger phase angles .
further observations at phase angles larger than 100@xmath5 are desirable to establish the exact @xmath4 at longer wavelengths .
the observed wavelength dependence of @xmath4 is interpreted as indicating that the earthshine polarization at shorter wavelengths is dominated by atmospheric rayleigh scattering , whereas that at longer wavelengths has an increasingly effective contribution from the earth surface reflection .
the wavelength dependence of @xmath4 for the earth appears to be unique among the terrestrial planetary bodies in the solar system .
therefore it might be a signature expressing a distinctive characteristic of the earth : this planet has a scattering but transparent atmosphere above its surface .
we thank the staff members of okayama astrophysical observatory , national astronomical observatory of japan .
we are deeply grateful to dr .
enric pall , the referee , for giving helpful comments as well as kindly calculating the earth scenery data in table [ tab : surface ] .
we also acknowledge dr .
michael f. sterzik for providing us the observational data for fig .
[ fig : es.p.all ] .
part of this study was supported by the center for planetary science running under the auspices of the mext gcoe program entitled `` foundation of international center for planetary science '' , a joint project between kobe university and hokkaido university .
this work was also supported by jsps kakenhi grant number 24540231 . | we present the results of the optical spectropolarimetry of earthshine on the moon for earth phase angles ranging from 49@xmath0 to 96@xmath0 .
the observations were conducted on 2011 march 9 - 13 ( ut ) using the spectropolarimeter hbs installed on the 1.88 m telescope at okayama astrophysical observatory .
the wavelength coverage was 450 - 850 nm with a resolution of 6 nm .
the observed earthshine polarization degree spectra exhibit decreasing polarization degree with increasing wavelength at any phase .
the overall degree of polarization increases as the earth approaches a quadrature phase .
the phase dependence differs with the wavelengths ; the maximum polarization for the @xmath1 band occurs at a phase angle of @xmath2 , whereas that for longer wavelengths is reached at larger phase angles .
this is interpreted as indicating that earthshine polarization at shorter wavelengths is dominated by atmospheric rayleigh scattering , whereas that at longer wavelengths has an increasingly effective contribution from the earth surface reflection .
the wavelength dependence of the phase angle of the maximum polarization appears to be unique among the terrestrial planetary bodies in the solar system .
therefore this might constitute important evidence pointing toward a distinctive characteristic of the earth : the planet has a scattering but transparent atmosphere above its surface . |
the _ wide - field infrared survey explorer _ ( wise ) is a nasa medium - class explorer mission designed to survey the entire sky in four infrared wavelengths , 3.4 , 4.6 , 12 and 22 @xmath2 m ( denoted @xmath3 , @xmath4 , @xmath5 , and @xmath6 respectively ) @xcite .
the final mission data products are a multi - epoch image atlas and source catalogs that will serve as an important legacy for future research . while wise s primary science goals are to find the most luminous galaxies in the entire universe and to find the closest and coolest stars , the survey has yielded observations of over 157,000 minor planets , including near - earth objects ( neos ) , main belt asteroids ( mbas ) , comets , hildas , trojans , centaurs , and scattered disk objects @xcite .
this represents an improvement of nearly two orders of magnitude more objects observed than wise s predecessor mission , the _ infrared astronomical satellite _ ( iras ; * ? ? ?
* ; * ? ? ?
the wise survey began on 14 january , 2010 , and the mission exhausted its primary tank cryogen on 5 august , 2010 .
exhaustion of the secondary tank and the start of the neowise post - cryogenic mission occurred on 1 october , 2010 , and the survey ended on 31 january , 2011 .
infrared observations of all classes of minor planets are useful for determining size and albedo distributions , as well as thermophysical properties such as thermal inertia , the magnitude of non - gravitational forces , and surface roughness ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
of the more than 157,000 objects detected by neowise , some have physical properties such as diameter , albedo , and subsolar temperature measured by independent means such as radar observations , _ in situ _ spacecraft visits , or stellar occultations .
we can obtain diameters and albedos for the full sample of minor planets observed by wise , but this requires characterization of the systematic errors associated with using thermal models ( c.f . * ? ? ?
* ) . in this work
, we have used a set of objects with well - known , independently measured diameters and @xmath0 magnitudes to test the ability of thermal models created using wise data to accurately compute diameter and albedo .
figure [ fig : bandpasses ] shows model spectral energy distributions for minor planets with different effective temperatures with the wise bandpasses overplotted .
the color corrections given by @xcite allow the zero points to be corrected to account for the effect of flux generated by objects with varying effective temperature being observed through non - monochromatic bandpasses .
the absolute zero points and isophotal wavelengths of the wise passbands @xcite are calibrated using the particular kurucz photospheric spectrum for vega detailed by @xcite and validated absolutely to 1.1% by @xcite .
the width of the wise passbands ( particularly @xmath5 ) means that it is necessary to apply a color correction for sources with a different spectrum than vega . for the vast majority of wise extra - solar sources ,
these color correction terms are small ( only a few percent for a k2v star in @xmath5 , for example ) .
however , for objects such as minor planets with effective temperatures as low as 100 k , the flux correction factor for @xmath5 exceeds a magnitude @xcite , and readers are strongly encouraged to consult the wise explanatory supplement and @xcite for precise values .
@xcite also find a discrepancy between red and blue calibrators in bands @xmath5 and @xmath6 that require a -8% and + 4% adjustment to the zero point magnitudes in each band , respectively .
the color corrections derived in @xcite were produced by integrating over the system s relative spectral response throughput , which was measured during ground - based tests prior to launch @xcite .
we have compared the measured wise magnitudes to magnitudes derived from spherical thermal models created for 117 objects , all of which have diameters and/or albedos measured via _ in situ _ spacecraft measurements , radar studies , or stellar occultations . by comparing the predicted magnitudes to the as - measured magnitudes reported by the wise pipeline
, we have verified that the color corrections derived in @xcite as a function of an object s effective temperature correctly reproduce observed physical parameters .
we have used these objects to set limits on the systematic errors for diameters and albedos derived from wise observations of minor planets when using a spherical thermal model and the @xcite color corrections .
finally , we have determined an empirical relationship between an object s subsolar temperature and its effective temperature , which is necessary for selecting the appropriate color correction for solar system objects from @xcite .
we have assembled a list of objects with well - measured diameters and @xmath0 magnitudes that wise observed during the fully cryogenic portion of its mission ( table 1 ) .
these objects were chosen because their physical measurements were obtained using methods largely independent of infrared model parameters , such as radar imaging , direct measurements from spacecraft visits or flybys , or stellar occultations .
diameters and albedos derived from other infrared observatories such as iras , the _ spitzer space telescope _ , and ground - based observations were not used to verify the color corrections of @xcite due to the desire to produce an independent calibration without reference to other infrared observers model parameter assumptions . however , for some objects , such as those observed by @xcite , the radar diameter measurements were compared with iras diameters and were found to be consistent . of the @xmath1400 radar - observed asteroids ( _ http://echo.jpl.nasa.gov/index.html_ ) , a dozen or so spacecraft targets , and hundreds of occultation targets
, we identified wise observations for 117 objects . in order to reduce the potential difficulties associated with applying a spherical thermal model to non - spherical objects ,
we eliminated from further consideration all objects with peak - to - peak magnitude differences larger than 0.3 magnitudes in @xmath5 ; this reduced the sample to the 50 objects listed in table 1 .
the observations of these objects were retrieved by querying the minor planet center s ( mpc ) observation files to look for all instances of individual wise detections of the desired objects that were reported using the wise moving object processing system ( wmops ; * ? ? ?
the resulting set of position / time pairs were used as the basis of a query of wise source detections in individual exposures ( also known as level 1b " images ) using the infrared science archive ( irsa ) . in order to ensure that only observations of the desired moving object were returned from the query ,
the search radius was restricted to 0.3 arcsec from the position listed in the mpc observation file .
additionally , since wise collected a single exposure every 11 seconds and observes each part of the sky an average of 10 times , the modified julian date was required to be within 2 seconds of the time specified by the mpc .
artifacts were minimized by setting the following flags : cc_flags = 0 or p and ph_qual = a , b , or c ( this flag indicates that the source is likely to have been a valid detection ) .
objects brighter than @xmath7 and @xmath8 magnitudes were assumed to have flux errors equivalent to 0.2 magnitudes due to changes to the shape of the point spread function as the objects became saturated , and a linear correction was applied to the @xmath5 magnitudes in this brightness regime ( the wise explanatory supplement contains a more detailed explanation ) . per the explanatory supplement , objects brighter than @xmath9 and @xmath10 were not used .
each object had to be observed a minimum of three times in at least one wise band , and it had to be detected at least 40% of the time when compared to the band with the maximum number of detections ( usually , though not always , @xmath5 ) .
the wmops system is designed to reject inertially fixed objects such as stars and galaxies in bands @xmath5 and @xmath6 .
nonetheless , the individual images at all wavelengths were compared with wise atlas coadd and daily coadd source lists to ensure that inertially - fixed sources such as stars and galaxies were not coincident with the moving object detections . this check is particularly important in bands @xmath3 and @xmath4 where the density of background objects ( and hence the probability of a blended source ) is higher than at longer wavelengths .
any remaining blended sources in bands @xmath3 and @xmath4 were removed .
some objects were observed at multiple epochs , and observations separated by more than three days were modeled separately .
the thermal flux from an airless sphere is given by @xmath11 where @xmath12 is the object - to - observer distance , @xmath13 is the emissivity , @xmath14 is the object s radius , @xmath15 is the planck function , @xmath16 is the angle from the sub - observer point to a point on the asteroid such that @xmath16 is equal to the solar phase angle @xmath17 at the sub - solar point , and @xmath18 is an angle measured around the sub - observer point such that @xmath19 at the sub - solar point ( e.g , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
in order to compute this flux , the temperature distribution across the body must be computed . in the standard thermal model ( stm ) of @xcite , the temperature of an asteroid is assumed to be maximum at the subsolar point and zero on the point opposite of this ; this is the case of an object with zero thermal inertia .
in contrast , in the fast rotating model ( frm ) @xcite , the asteroid is assumed to be rotating much faster than its cooling time ( i.e. high thermal inertia ) , resulting in a constant surface temperature across all longitudes .
the so - called beaming parameter " was introduced by @xcite in the stm to account for the enhancement of thermal radiation observed at small phase angles .
the near - earth asteroid thermal model ( neatm ) of @xcite also uses the beaming parameter @xmath20 to account for cases intermediate between the stm and frm models , such that @xmath21^{1/4}\ ] ] where @xmath17 the solar phase angle , and the sub - solar temperature @xmath22 is given by : @xmath23^{0.25}.\ ] ] the emissivity , @xmath13 , is assumed to be 0.9 for all wavelengths ( c.f .
* ) , @xmath24 is the bond albedo , @xmath25 is the solar flux , @xmath26 is the stefan - boltzmann constant , and @xmath20 is the beaming parameter . in the stm , @xmath20 is set to 0.756 to match the occultation diameters of 1 ceres and 2 pallas , while in the frm , @xmath20 is equal to @xmath27 . with neatm , @xmath20 is a free parameter that can be fit when two or more infrared bands are available ( or with only one infrared band if diameter or albedo are known _ a priori _ as in this paper ) .
each object was modeled as a set of triangular facets covering a spherical surface with diameter equal to the ground - truth measurement ( c.f . * ? ? ?
model magnitudes were computed for each wise measurement , ensuring that the correct sun - observer - object distances were used for each one .
the temperature for each facet was computed , and the @xcite color corrections were applied to each facet .
the emitted thermal flux for each facet was calculated using neatm along with the bandcenters and zero points given in @xcite ; the temperature at the anti - subsolar point was set to 3 k , so the facets closest to this point contribute little flux .
the objects absolute magnitudes ( @xmath0 ) were taken from @xcite and @xcite when available ; otherwise , the values were taken from the mpc s orbital element files .
although @xcite , @xcite and @xcite have applied a 0.3 magnitude offset to the @xmath0 values for objects in their respective works , we did not apply the offset to the @xmath0 magnitudes @xcite .
the offset was caused by the dominance of unfiltered ccd photometry in the mpc datafiles , largely from the linear survey ; as more filtered measurements have become available from other surveys in recent years , the offset is no longer appropriate to use .
we will revisit these @xmath0 values as improved measurements become available .
diameters and albedos computed from the combination of diameter and @xmath0 from the radar , spacecraft , or occultation measurements are given in table 1 .
in addition , the saturnian moon phoebe had a measurement of subsolar temperature @xcite that was also used in the thermal model .
@xmath0 value was taken from @xcite , and its @xmath28 value from @xcite .
if an object had no prior independent measurement of geometric albedo ( @xmath29 ) from an _ in situ _ measurement , it was computed using the relationship @xmath30^{2},\ ] ] where @xmath0 is the object s absolute magnitude and @xmath31 is its diameter @xcite . for objects with three valid measurements in two or more wise bands dominated by thermal emission , @xmath20 was determined using a least squares minimization . in general , minor planets detected by neowise in bands @xmath3 and @xmath4 contain a mix of reflected sunlight and thermal emission .
thus , it was necessary to incorporate an estimate of reflected sunlight into the thermal model in order to use data from bands @xmath3 and @xmath4 . in order to compute the fraction of reflected sunlight in bands @xmath3 and
@xmath4 , it was also necessary to compute the ratio of the infrared albedo @xmath32 to the visible albedo @xmath33 .
we make the simplifying assumption that the reflectivity is the same in both bands @xmath3 and @xmath4 , such that @xmath34 ; the validity of this assumption is discussed below .
the geometric albedo @xmath33 is defined as the ratio of the brightness of an object observed at zero phase angle ( @xmath17 ) to that of a perfectly diffusing lambertian disk of the same radius located at the same distance .
the bond albedo ( @xmath24 ) is related to the visible geometric albedo @xmath33 by @xmath35 , where @xmath36 is the phase integral and is defined such that @xmath37 .
@xmath38 is the phase curve , and @xmath39 for @xmath40 .
@xmath28 is the slope parameter that describes the shape of the phase curve in the @xmath41 model of @xcite that describes the relationship between an asteroid s brightness and the solar phase angle . for @xmath42 , @xmath43 .
since @xmath36 is never larger than about 2/3 for any observed value of @xmath28 , the geometric albedo @xmath33 is not limited to @xmath441.0 ; it is possible to have a value as large as @xmath11.5 without violating conservation of energy .
conversely , values of @xmath33 approaching 1.0 still amount to integral reflectivity ( bond albedo ) of @xmath160% , less than newly fallen snow , or other analogs of very high albedo .
we make the assumption that @xmath32 obeys these same relationships , although it is possible that the phase integral described above varies with wavelength , so what we denote here as @xmath32 for convenience may not be exactly analogous to @xmath33 .
the flux from reflected sunlight was computed for each wise band using the iau phase curve correction @xcite : @xmath45,\ ] ] where @xmath46 is the predicted apparent magnitude ; @xmath14 is the heliocentric distance in au ; @xmath12 is the geocentric distance in au ; @xmath17 is the sun - observer - object angle ; @xmath28 is slope parameter for the magnitude - phase relationship ; and @xmath47,\ ] ] where @xmath48 unless a direct measurement of @xmath28 was available from @xcite or @xcite , we assumed a @xmath28 value of 0.15 .
finally , the weighted averages of the model magnitude , measured wise magnitude and @xmath22 were computed for each object by grouping together observations no more than a three day gap between them .
this was done to ensure that neos , which can have significant changes in distance over short times , were modeled accurately .
error bars on the model magnitudes and subsolar temperatures were determined for each object by running 100 monte carlo trials that varied the objects @xmath0 values by the error bars given in table 1 , their diameters by the error bars specified in table 1 , and the wise magnitudes by their error bars using gaussian probability distributions .
the minimum magnitude error for all wise measurements fainter than @xmath7 and @xmath49 magnitudes was 0.03 magnitudes , per the in - band repeatability measured in @xcite . for objects brighter than @xmath7 and @xmath8 ,
the error bars were increased to 0.2 magnitudes , as these magnitudes represent the onset of saturation . additionally , a linear correction was applied to objects with @xmath50 .
the effect is thought to be caused by changes in the point spread function observed when images start to enter saturation ( see the wise explanatory supplement for a discussion of the effects of saturation on photometry ) .
we have set the error bars to 0.2 magnitudes right at these limits in order to conservatively capture residual errors not fixed by the simple linear correction applied .
objects with @xmath51 and @xmath52 were too bright to be used .
the error bar for each object s model magnitude was equal to the standard deviation of all the monte carlo trial values .
if a measurement of @xmath0 was not available from @xcite or @xcite , the error in @xmath0 was taken to be 0.3 magnitudes @xcite .
figure [ fig : lightcurve ] shows an example of an asteroid s wise magnitudes as well as the resulting thermal model fit
. figures [ fig : calibrators ] show the difference between the average measured wise magnitudes and the model magnitudes for bands @xmath3 through @xmath6 using the color corrections given in @xcite .
in addition , we adjusted the @xmath5 effective wavelength blueward by 4% from 11.5608@xmath2 m to 11.0984 @xmath2 m , the @xmath6 effective wavelength redward by 2.5% from 22.0883 @xmath2 m to 22.6405 @xmath2 m , and we included the -8% and + 4% offsets to the @xmath5 and @xmath6 magnitude zeropoints ( respectively ) due to the red - blue calibrator discrepancy reported by @xcite .
the weighted means of the differences between observed and model magnitudes shown in figure [ fig : calibrators ] are given in table 2 , and they are near zero for most objects .
the apparent trend below zero for the objects with @xmath22 larger than @xmath1300 k could be due to any of the following causes : 1 ) imperfect knowledge of the system relative spectral response curves , particularly in @xmath6 ; 2 ) inaccuracies in either @xmath0 values or diameters for some of the objects ; 3 ) the temperature distribution may not precisely follow that given in equation 2 .
this is less important for objects observed at low phase angles .
the warmer objects observed by wise tend to be neos observed at higher phase angles , which could lead to the subsolar point contributing significantly more flux in the shorter wavelengths and the nightside of the object more flux at longer wavelengths .
nevertheless , most of the predicted magnitudes are in good agreement with the observed magnitudes , indicating that the procedure given in @xcite for color correction is adequate .
although many of the calibrator objects are known to be non - spherical ( e.g. from radar shape modeling or optical lightcurves ) , the wise observations generally consisted of @xmath110 - 12 observations per object uniformly distributed over @xmath136 hours @xcite , so on average , a wide range of rotational phases was sampled .
although the variation in effective spherical diameter resulting from rotational effects tends to be averaged out , caution must be exercised when interpreting effective diameter results using spherical models for objects that are known to have large - amplitude lightcurve variations .
figure [ fig : lightcurve_amplitudes ] shows the distribution of peak - to - peak @xmath5 amplitudes for the entire sample of 117 objects .
we have compared the mean differences between observed and model magnitudes for the entire sample of 117 objects to the sample of only those 50 objects with peak - to - peak @xmath5 amplitudes @xmath53 magnitudes , and there is no significant difference in the result given in table 2 .
nevertheless , to avoid any potential difficulties associated with applying spherical models to elongated objects , we excluded objects with peak - to - peak @xmath5 amplitudes from further analysis .
( figure [ fig : calibrators ] shows the difference between observed and model magnitudes for only the 50 low - amplitude objects ) . for the objects in table 1 for which pole orientations could be found in the lightcurve database of @xcite , we computed the visible lightcurve amplitude at the aspect observed by wise and found that for these objects , the apparent amplitudes were @xmath540.3 magnitudes .
the offsets and errors given in table 2 can be regarded as the minimum systematic errors in magnitude for minor planets observed by wise / neowise . since diameter is proportional to the square root of the thermal flux ( equation 1 ) , the minimum systematic diameter error due to uncertainties in the color correction is proportional to one - half the error in flux .
these magnitude errors result in a minimum systematic error of @xmath15 - 10% for diameters derived from wise data ; they are of similar magnitude to the diameter uncertainties of most of the underlying radar and spacecraft measurements , which are @xmath110% ( references are given in table 1 ) .
albedo is proportional to @xmath55 ( equation 4 ) , and so it is linearly proportional to flux to first order .
therefore , minimum systematic errors on albedos computed from wise observations are @xmath110 - 20% , subject to the assumption that spherical effective diameters can be computed for non - spherical shapes ( future work will model the objects as non - spherical shapes and will use shape models and rotational information produced by lightcurve and radar studies ) .
these should be regarded as minimum errors in cases of good signal - to - noise detections when the beaming parameter and the infrared albedo can be fit .
it should also be noted that these error estimates apply only to objects as distant as saturn , as the most distant object we have considered herein is saturn s moon phoebe .
objects observed by wise at greater distances ( and therefore lower temperatures ) may be subject to additional errors .
the color corrections specified in @xcite are given as a function of input spectra that are assumed to blackbodies of different effective temperatures ( @xmath56 ) . in order to use the @xcite color corrections by computing only the subsolar temperature rather than a faceted thermophysical model for each object ,
we have computed the relationship between @xmath22 and ( @xmath56 ) . the total flux ( @xmath57 , where @xmath58 ) was computed using neatm by applying a color correction to each facet on the sphere based on the facet s blackbody temperature .
the equivalent flux at the center wavelength of each bandpass is computed , @xmath59 ; since this is a monochromatic flux , the color corrections are unity .
we then derive an effective flux correction ( @xmath60 ) which is given by @xmath61 the final step is to find the @xmath56 that gives the blackbody flux correction @xmath62 .
figure [ fig : tss_vs_teff ] shows the relationship between @xmath22 and @xmath56 ; a least - squares fit to these points yields the relationship @xmath63 we excluded objects with more than 20% reflected sunlight in a given band because their color corrections will be dominated by the small corrections needed for a g star @xcite
. a single object can contribute up to four points on this plot , one for each band it is detected in with less than 20% reflected sunlight .
equation 9 provides a shortcut method to flux correct the wise magnitudes in lieu of performing the facet - by - facet correction described above with negligible additional error .
for objects that were detected according to the criteria described above in either @xmath3 or @xmath4 , we computed the albedo at these wavelengths ( @xmath32 , assuming @xmath64 ) in addition to @xmath33 .
although @xcite and @xcite assume that the albedo at 3.4 @xmath2 m is 1.4 times higher than @xmath33 , this result is based on seven s - type main belt asteroids observed at 3 @xmath2 m by @xcite .
@xcite use observations of a number of neos observed by _ spitzer _ and find @xmath65 consistent with 1.4 , using the relationship between @xmath20 and @xmath17 defined by @xcite .
we have not used the @xcite relationship to derive @xmath20 and instead have allowed it to vary .
since we know the diameter and can derive @xmath33 from equation 4 , we fit @xmath20 independently for each object in table 1 .
figure [ fig : beaming_hist ] shows the histogram of beaming parameters , and figure [ fig : phase_v_eta ] shows @xmath20 as a function of solar phase angle for the objects described in table 1 ; however , we expect to significantly improve upon this result in a future work by using the general population of asteroids observed by wise over a wide range of phase angles .
figure [ fig : pvpir ] shows @xmath65 for the objects detected in @xmath3 and @xmath4 ; as shown in table 1 , they represent a mix of neos , mbas , and irregular satellites . the weighted average of the ratio of @xmath65 is @xmath66 . in computing @xmath65 , we have made the simplifying assumption that @xmath28 does not vary with wavelength , although it is known that @xmath28 increases slightly , from 0.15 to 0.20 , when going from @xmath15 to @xmath14 bands @xcite .
however , in some cases , the variation in albedo with wavelength could be due to material and/or chemical properties ( c.f .
phoebe ; * ? ? ?
a future work comparing both the infrared and visible albedos with known taxonomic classifications is in progress .
the calibration methodology described in this work is not unique to the wise infrared data ; however , the uniquely large set of minor planet observations afforded by wise / neowise has permitted characterization of the systematic errors produced when applying spherical neatm models to a number of objects with previously measured diameters that span a wide range of populations .
we have studied neos , main belt asteroids , and irregular satellites . in particular , the selection of an effective temperature to use when applying a literature color correction can cause large changes in the resultant best - fit diameter unless the relation between @xmath22 and @xmath56 is well - understood or a complete faceted model is employed .
the color corrections described in @xcite have been used to produce an estimate of the minimum systematic errors in magnitudes for minor planets detected by wise / neowise @xcite .
we have shown that the minimum diameter errors resulting from residual uncertainties in the color corrections and assumptions regarding @xmath28 , @xmath0 , and albedo are comparable to the uncertainties in the diameters measured by radar or _ in situ _ spacecraft imaging ( i.e. @xmath110% ) ; albedos can be determined to @xmath120% when good signal - to - noise multi - band wise measurements and visible data are available . however , we note that objects more distant than the trojan asteroids may be subject to increased systematic errors , as the most distant object studied in this work was saturn s moon phoebe .
our model includes an estimate of reflected sunlight computed in all four bands , and we have computed albedos at 3.4 and 4.6 @xmath2 m as well as at visible wavelengths , which should yield interesting compositional insights when compared with spectroscopic taxonomic data .
areas for future improvement of thermal models include studying the effects of observing at high phase angles , modeling non - spherical shapes , and allowing both @xmath28 and albedo to vary as a function of wavelength .
all observations were processed with the pass 1 ( preliminary ) version of the wise data processing pipeline ; as the final version of the pipeline becomes available , which incorporates many improvements to instrumental calibration , we will revisit the thermal models for these objects .
we have also derived a simple linear relationship between subsolar temperature and the effective temperature used in @xcite , which should facilitate appropriate choices of color corrections for the wise bandpasses .
the thermal models computed herein are not intended to supplant the diameters measured by radar , in situ imaging or occultations , but rather they provide insight into the ability of spherical neatm thermal models to accurately determine diameter and albedo when applied to the general population of wise - observed minor planets for which these parameters have not been previously determined .
this demonstrates that the wise dataset offers a powerful new means of characterizing physical parameters of minor planets with great accuracy .
bauer , j. m. , grav . , t. , buratti , b. , hicks , m. 2006 icarus 184 , 181 busch , m. , giorgini , j. , ostro . , s. , benner , l. , jurgens , r. , rose , r. , hicks , m. , pravec , p. , kusnirak , p. , ireland , m. , scheeres , d. , broschart , s. , magri , s. , nolan , m. , hine , a. , margo , j .- l .
2007 icarus , 190 , 608 clark , r. , brown , r. , jaumann , r. , cruikshank , d. , nelson , r. , buratti , b. , mccord , t. , lunine , j. , baines , k. , bellucci , g. , bibring , j .-
capaccioni , f. , cerroni , p. , coradini , a. , formisano , v. , langevin , y. , matson , d. , mennella , v. , nicholson , p. , sicardy , b. , sotin , c. , hoefen , t. , curchin , j. , hansen , g. , hibbits , k. , matz , k .-
2005 nature 435 , 66 giorgini , j. d. , , s. j. ostro , l. a. m. benner , p. w. chodas , s. r. chesley , r. s. hudson , m. c. nolan , a. r. klemola , e. m. standish , r. f. jurgens , r. rose , a. b. chamberlin , d. k. yeomans , j .-
margot 2002 science 296 , 132 harris , a. , mommert , m. , hora , j. , mueller , m. , trilling , d. , bhattacharya , b. , bottke w. , chesley , s. , delbo , m. , emery , j. , fazio , g. , mainzer , a , penprase , b. , smith , h. , spahr , t. , stansberry , j. , thomas , c. 2011 aj 141 , 75 llllllll 5 & @xmath67 & 6.9 & @xmath68 & @xmath69 & @xmath70 & @xmath71 & + 8 & @xmath72 & 6.4 & @xmath73 & @xmath74 & @xmath75 & @xmath76 & + 13 & @xmath77 & 6.7 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & + 22 & @xmath82 & 6.5 & @xmath83 & @xmath84 & @xmath85 & @xmath86 & + 27 & @xmath87 & 7.0 & @xmath88 & @xmath89 & @xmath90 & @xmath91 & + 31 & @xmath92 & 6.7 & @xmath93 & @xmath94 & @xmath95 & @xmath96 & + 36 & @xmath97 & 8.5 & @xmath78 & @xmath98 & @xmath99 & @xmath100 & + 39 & @xmath101 & 6.1 & @xmath102 & @xmath103 & @xmath104 & @xmath105 & + 47 & @xmath106 & 7.8 & @xmath107 & @xmath98 & @xmath108 & @xmath109 & + 53 & @xmath110 & 8.8 & @xmath111 & @xmath112 & @xmath113 & @xmath114 & + 54 & @xmath115 & 7.7 & @xmath93 & @xmath78 & @xmath116 & @xmath117 & + 83 & @xmath118 & 8.7 & @xmath119 & @xmath120 & @xmath121 & @xmath122 & + 85 & @xmath123 & 7.6 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & + 94 & @xmath128 & 7.6 & @xmath129 & @xmath124 & @xmath130 & @xmath131 & + 97 & @xmath132 & 7.7 & @xmath133 & @xmath134 & @xmath135 & @xmath136 & + 105 & @xmath137 & 8.6 & @xmath129 & @xmath138 & @xmath139 & @xmath140 & + 114 & @xmath141 & 8.3 & @xmath142 & @xmath143 & @xmath144 & @xmath145 & + 114 & @xmath146 & 8.3 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & + 128 & @xmath151 & 7.5 & @xmath152 & @xmath129 & @xmath153 & @xmath154 & + 135 & @xmath155 & 8.2 & @xmath156 & @xmath68 & @xmath157 & @xmath158 & + 139 & @xmath159 & 7.9 & @xmath112 & @xmath98 & @xmath160 & @xmath161 & + 145 & @xmath162 & 8.1 & @xmath112 & @xmath112 & @xmath163 & @xmath164 & + 194 & @xmath165 & 7.7 & @xmath129 & @xmath124 & @xmath166 & @xmath140 & + 198 & @xmath167 & 8.3 & @xmath168 & @xmath169 & @xmath170 & @xmath171 & + 208 & @xmath172 & 9.3 & @xmath83 & @xmath84 & @xmath173 & @xmath174 & + 208 & @xmath172 & 9.3 & @xmath175 & @xmath84 & @xmath176 & @xmath177 & + 211 & @xmath178 & 7.9 & @xmath124 & @xmath78 & @xmath179 & @xmath180 & + 230 & @xmath181 & 7.3 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & + 266 & @xmath186 & 8.5 & @xmath98 & @xmath93 & @xmath187 & @xmath188 & + 308 & @xmath189 & 8.2 & @xmath93 & @xmath190 & @xmath191 & @xmath192 & + 345 & @xmath193 & 8.7 & @xmath98 & @xmath98 & @xmath194 & @xmath195 & + 350 & @xmath196 & 8.4 & @xmath190 & @xmath94 & @xmath197 & @xmath198 & + 444 & @xmath199 & 7.8 & @xmath200 & @xmath201 & @xmath202 & @xmath203 & + 488 & @xmath204 & 7.8 & @xmath98 & @xmath124 & @xmath205 & @xmath105 & + 522 & @xmath206 & 9.1 & @xmath207 & @xmath142 & @xmath208 & @xmath117 & + 566 & @xmath209 & 8.0 & @xmath124 & @xmath142 & @xmath210 & @xmath211 & + 654 & @xmath212 & 8.5 & @xmath112 & @xmath93 & @xmath213 & @xmath214 & + 704 & @xmath215 & 5.9 & @xmath190 & @xmath216 & @xmath217 & @xmath218 & + 704 & @xmath219 & 5.9 & @xmath220 & @xmath221 & @xmath222 & @xmath223 & + 925 & @xmath224 & 8.3 & @xmath68 & @xmath225 & @xmath226 & @xmath227 & + 951 & @xmath228 & 11.5 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & + 1512 & @xmath233 & 9.6 & @xmath207 & @xmath234 & @xmath235 & @xmath236 & + 1627 & @xmath237 & 12.9 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & + 1866 & @xmath242 & 12.7 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & + 2867 & @xmath247 & 13.4 & @xmath248 & @xmath249 & @xmath250 & @xmath251 & + 7335 & @xmath252 & 17.0 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & + 68216 & @xmath257 & 16.3 & @xmath258 & & @xmath259 & @xmath260 & + 68216 & @xmath257 & 16.3 & @xmath261 & & @xmath262 & @xmath263 & + 164121 & @xmath264 & 16.2 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & + himalia & @xmath269 & 7.9 & @xmath78 & @xmath78 & @xmath270 & @xmath271 & + 2005 cr37 & @xmath272 & 18.9 & @xmath129 & @xmath207 & @xmath273 & @xmath274 & + 2008 ev5 & @xmath275 & 19.7 & @xmath276 & @xmath277 & @xmath278 & @xmath279 & + phoebe & @xmath280 & 6.6 & @xmath142 & @xmath281 & @xmath282 & @xmath283 & + this figure illustrates the need for the color corrections given by @xcite in order to capture the shift in zero point and effective wavelength as a function of effective temperature : the wise bandpasses are broad , particularly @xmath5 .
the wise bandpasses are shown as shaded bars , and normalized fluxes are plotted for a range of different blackbody temperatures , illustrating the necessity for a color correction that varies as a function of an object s effective temperature . , width=288 ] wise observations as a function of modified julian date as well as a spherical neatm model fit are shown for a typical calibrator object , ( 208 ) lacrimosa .
neowise detections of this asteroid span @xmath117 hours .
the @xmath3 points are shown in dark blue ; @xmath4 in cyan , @xmath5 in green , and @xmath6 in red . ,
width=576 ] with wise , we have observed objects with diameters that have been measured either with radar observations , _ in situ _ spacecraft visits , or by stellar occultations ( see table 1 ) . after applying the color corrections specified in @xcite , we can compare the observed wise magnitudes to the fluxes predicted by a thermophysical model for bands @xmath3 through @xmath6 ( figures ( a)-(d ) , respectively ) .
the dashed line shows the weighted mean value of all the points ; the dotted lines are the 1-@xmath26 errors
. objects with wise measurement errors that are less than 0.1 magnitudes are shown as red dots ; objects with errors greater than 0.1 magnitudes are shown as black dots.,width=576 ] the distribution of peak - to - peak amplitudes in @xmath5 for 117 objects with independently measured diameters peaks at @xmath10.25 magnitudes .
objects with peak - to - peak amplitudes larger than 0.3 magnitudes were excluded from our analysis.,width=288 ] the correlation between subsolar temperature and effective temperature is well - described by a linear relationship .
@xmath284 detections are shown in blue circles ; @xmath4 as cyan triangles ; @xmath5 as green squares ; and @xmath6 as red inverted triangles
. each object can contribute up to four points on the plot , depending on the number of bands in which it was detected.,width=288 ] the beaming parameter vs. phase angle .
the best linear fit to our data is @xmath286 and is plotted as a red dashed line .
the relationship from @xcite is given by @xmath287 and is shown as a blue dash - dot line .
future work will examine the relationship between @xmath20 and @xmath17 using the full wise dataset , which includes many more objects over a wide range of phase angles .
, width=288 ] the ratio of the albedo at @xmath3 and @xmath4 ( we assume the albedo is the same in both of these bands ) compared to the albedo at visible wavelengths as a function of subsolar temperature .
the dashed line indicates the weighted mean value of @xmath288.,width=288 ] | with the _ wide - field infrared survey explorer _ ( wise ; * ? ? ?
* ) , we have observed over 157,000 minor planets @xcite . included in these
are a number of near - earth objects , main belt asteroids , and irregular satellites which have well - measured physical properties ( via radar studies and _ in situ _ imaging ) such as diameters .
we have used these objects to validate models of thermal emission and reflected sunlight using the wise measurements , as well as the color corrections derived in @xcite for the four wise bandpasses as a function of effective temperature .
we have used 50 objects with diameters measured by radar or in situ imaging to characterize the systematic errors implicit in using the wise data with a faceted spherical neatm model to compute diameters and albedos . by using the previously measured diameters and @xmath0 magnitudes with a spherical neatm model ,
we compute the predicted fluxes ( after applying the color corrections given in * ? ? ?
* ) in each of the four wise bands and compare them to the measured magnitudes .
we find minimum systematic flux errors of 5 - 10% , and hence minimum relative diameter and albedo errors of @xmath110% and @xmath120% , respectively . additionally , visible albedos for the objects are computed and compared to the albedos at 3.4 @xmath2 m and 4.6 @xmath2 m , which contain a combination of reflected sunlight and thermal emission for most minor planets observed by wise . finally , we derive a linear relationship between subsolar temperature and effective temperature , which allows the color corrections given in @xcite to be used for minor planets by computing only subsolar temperature instead of a faceted thermophysical model .
the thermal models derived in this paper are not intended to supplant previous measurements made using radar or spacecraft imaging ; rather , we have used them to characterize the errors that should be expected when computing diameters and albedos of minor planets observed by wise using a spherical neatm model . |
up to the energy scale we can test experimentally , the standard model ( sm ) well describes the forces between matter particles except for the gravity .
almost all the candidates for the theory of the gravity at planck scale @xmath8 are defined with the supersymmetry ( susy ) in more than 4-dimensional ( 4d ) space - time , such as supergravity or superstring theory .
it is considered that the sm particles lie in higher dimensional space - time and extra dimensions are compactified smaller than the size we can detect in the low energy experiments or lie on the 4d subspace in higher dimensional space - time .
recent interesting suggestion is that sm may be realized on a ` brane ' which stands for the lower dimensional object in the higher dimensional space - time .
one of the candidates of such object is d - brane or domain wall .
we call such scenario ` brane world model . '
the brane world models now casting new ideas on physics beyond the sm .
for example recent studies in the superstring theory provide us some concrete examples realizing sm in the d - brane systems @xcite .
basis of the sm is a spontaneous gauge symmetry breaking that results in the existence of the massive gauge bosons , w and z. we need at least one ( doublet ) scalar field , called higgs field which develops a vacuum expectation value ( vev ) to give a masses to the w and z gauge bosons and breaks the sm gauge symmetry .
these gauge boson masses are related to the electroweak scale @xmath9 at which the sm gauge symmetry is broken as @xmath10 .
the mass scale of the w and z gauge bosons shows that @xmath9 is of the order tev scale , while the gravitational scale , namely planck scale @xmath8 is of the order @xmath11 gev .
if we consider the unification of sm and gravity , we should explain the hierarchical structure between @xmath8 and @xmath9 .
since the mass of the scalar field is not protected by any symmetries , the radiative correction brings the higgs mass to the fundamental scale @xmath8 without a fine tuning .
this is so called ` hierarchy ' or ` fine tuning ' problem .
there are two remarkable brane world models solving this hierarchy problem .
one is the large extra dimension scenario @xcite and the other is the warped extra dimension scenario in randall - sundrum ( rs ) brane world @xcite .
susy enables to stabilize the scalar mass against the radiative correction of @xmath12 because it is related to the mass of the fermionic superpartner that can be protected by appropriate symmetry .
so the susy permits the light scalar field like higgs field , however if we introduce it around the tev scale in order to keep the higgs mass @xmath13 tev , all the fermions in the sm have their scalar partners with their masses around tev .
these extra scalar fields have not been observed in any experiments yet .
thus we need some mechanisms in the susy breaking process to control these extra light scalars not to appear inside the region of the present experimental observation . however , the supertrace theorem predicts lighter scalar partners than the fermions in sm that is not consistent with the experimental results . to avoid this problem
the theory should have the ` hidden sector ' where susy is broken .
a certain mediation mechanism to communicate it to the ` visible sector ' is required at the loop level through renormalizable interactions or at the tree level through unrenormalizable interactions .
recently some people consider that the hidden sector setup merges into the brane world picture , that separates the hidden and visible sector by spatiality in the extra dimension .
they use the bulk ( moduli ) fields to communicate the susy breaking signals from the hidden to the visible sector @xcite .
the hidden sector is the minimal requirement from the supertrace theorem and we need more severe conditions to the susy breaking from the experiment .
one of them is from the supersymmetric flavor problem , that is the masses of the sm fermions are disjointed each other while the masses of their scalar partners are almost degenerate . as we see above
almost all the problems in the tev scale susy come from too much light extra fields in addition to the sm one .
we need complicated setup about susy breaking ( and its mediation ) mechanism to avoid these problems .
remember that susy itself is needed for the consistency of the quantum theory of gravity , while ` tev scale susy ' is required in order to stabilize the electroweak scale @xmath9 , and to realize gauge coupling unification .
however , in a recent brane world picture , there is a possibility that the strong and electroweak gauge group come from the different brane . in this case
the unification of sm and gravity is able to occur directly without grand unification and the gauge coupling unification in mssm may be accidental .
we can also consider the case that gauge coupling unification is not depend on the tev scale susy ( mssm ) and it happens by the other mechanism , e.g. extra dimensional effect @xcite . if there is another mechanism to stabilize mass of the higgs field , it is possible to break susy at higher energy scale , even around the planck scale .
such a mechanism can set susy free from any problems at tev scale .
in this paper we notice the recent suggestions that the electroweak ( chiral ) symmetry is broken down dynamically by the gauge interaction in more than 4d space - time @xcite .
the higher dimensional gauge theory may realize the stabilization of a light higgs field through the composite higgs scenario .
this is based on the idea that the gauge theory in compact extra dimension has kaluza - klein ( kk ) massive gauge bosons and they act as a binding force between fermion and anti - fermion that results in the composite scalar field .
it was pointed out in refs .
@xcite and @xcite a negative curvature and a finite size effects enhance dynamical symmetry breaking .
if sm is embedded in a brane world , the dynamical mechanism of electroweak symmetry breaking can be realized in a certain bulk gauge theory ( or sm itself in the bulk . )
sm in the bulk has a possibility to realize top quark condensation scenario @xcite in the brane world .
it is very interesting because it may have no extra field ( elementary higgs , techni - color gauge boson , techni - fermion etc . )
except for the sm contents ( without elementary higgs ) in order to break the electroweak symmetry .
we launched a plan to study the dynamical symmetry breaking in the brane world models in detail .
there is no enough knowledge about the ( non - perturbative ) dynamics of the higher dimensional gauge theory in brane world . in this paper
we study the basic properties of dynamical symmetry breaking in the bulk gauge theory couples to a fermion on the brane . based on the effective theory of the bulk yang - mills theory that is defined below the _ reduced _ cutoff scale on the brane
, we formulate the ladder schwinger - dyson ( sd ) equation of the fermion propagator .
it is numerically solved by using the improved ladder approximation with power low running of the effective gauge coupling .
we show how the dynamics of the bulk gauge field affects the chiral symmetry breaking on the brane .
we give a kk reduced lagrangian of our bulk - brane system in @xmath14[sec : lag ] .
in @xmath14[sec : effsd ] we derive the effective theory on the four - dimensional brane and formulate sd equation about the fermion propagator on the brane .
the numerical analysis of the sd equation is performed , within the improved ladder approximation , for the case of yang - mills theory on the brane , in the flat bulk space - time and in the warped bulk space - time in @xmath14[sec : ila ] .
in this paper we analyze dynamical symmetry breaking on the four dimensional brane with bulk gauge theory in various brane world models .
the bulk gauge theory is , of course , more than four dimensional theory and ill defined in ultraviolet ( uv ) region and we need some regularization about it . in this paper
we use effective lagrangian of it defined below the _ reduced _ cutoff scale on the brane . in this section
we derive the kk reduced 4d lagrangian of the bulk gauge theory in the brane worlds , @xmath15 type space - time
. we can easily extend it to the case of higher dimensional bulk space - time , @xmath16 , it is shown in the next section . because one of our goal is to analyze
the dynamically induced mass scale on the brane in the randall - sundrum space - time , i.e. the slice of ads@xmath17 , we take account for the lagrangian and mode function in a curved extra dimension .
the original motivation to introduce a curved extra dimension by randall and sundrum is to produce the weak - planck hierarchy from the exponential factor in the space - time metric .
the factor is called ` warp factor ' , and the extra space ` warped extra dimension ' .
a non - factorizable geometry with the warp factor distinguishes the rs brane world from the others . in the rs model
we consider the fifth dimension @xmath18 which is compactified on an orbifold , @xmath19 of radius @xmath20 and two 3-branes at the orbifold fixed points , @xmath21 and @xmath22 .
requiring the bulk and boundary cosmological constants to be related , einstein s equation in five - dimension leads to the solution @xcite , @xmath23 where @xmath24 = @xmath25 , @xmath26 and @xmath27 .
@xmath28 is the ads curvature scale with the mass dimension one , and @xmath29 describes the flat @xmath15 space - time . in the following
we study a bulk vector field , @xmath30 in this background . substituting the metric @xmath31 defined in eq .
( [ eq : rsmetric ] ) and taking @xmath32 gauge , the lagrangian of a bulk gauge field in the rs brane world is written by @xmath33 where @xmath34 includes the gluon three and four point self interaction in the bulk .
we drop the explicit symbol of the trace operation in terms of the yang - mills index in the lagrangian throughout this paper .
we perform the kk mode expansion as @xmath35 where @xmath36 and @xmath37 are the @xmath38 even and odd mode function of @xmath39-th kk excited mode @xmath40 and @xmath41 , respectively .
the @xmath38 orbifold condition thus means @xmath42 for @xmath43 , i.e. dropping all the odd modes .
after kk expansion , we integrate the lagrangian ( [ eq : bulkgauge ] ) over the y - direction and obtain @xmath44 \nonumber \\ & & + \sum_{n=1}^\infty \left [ -\frac{1}{4}\tilde{f}^{(n)\mu \nu}(x)\tilde{f}_{\mu \nu}^{(n)}(x ) + \frac{1}{2 } m_n^2 \tilde{a}^{(n)\mu } \tilde{a}_{\mu}^{(n ) } \right ] , \label{eq : bulkgaugekinmass}\end{aligned}\ ] ] where the kk mode function @xmath45 and mass eigenvalue @xmath46 satisfy @xmath47 and @xmath48 obeys the same equations .
the boundary conditions are given by @xmath49 eq .
( [ eq : bulkgaugekinmass ] ) is the 4d effective lagrangian of the bulk gauge theory and eqs .
( [ eq : kkeigeneq])-([eq : boundarycond ] ) determine the form of the mode function and the eigenvalue @xmath46 .
below we solve eq .
( [ eq : kkeigeneq ] ) with the boundary condition ( [ eq : boundarycond ] ) and the normalization ( [ eq : normcond ] ) for the flat ( @xmath29 ) and warped ( @xmath50 ) extra dimensions .
we call the case with @xmath29 ` flat brane world ' . in this case
the solution of eqs .
( [ eq : kkeigeneq])-([eq : boundarycond ] ) is given by @xmath51 where @xmath52 .
these are the kk mode function and mass eigenvalue of the bulk gauge field in the flat brane world with @xmath15
. we can easily extend it to higher dimensional bulk space - time , @xmath16 , that is shown latter .
@xmath50 with orbifold condition , @xmath42 for @xmath43 , gives the rs brane world .
the solution of eqs .
( [ eq : kkeigeneq])-([eq : boundarycond ] ) reads @xcite @xmath53 \quad ( n=1,2,\ldots ) , \label{eq : rskkwf } \end{aligned}\ ] ] where the coefficient and the normalization factor are given by @xmath54 and @xmath55 ^ 2 $ ] respectively and @xmath56 and @xmath57 are the bessel functions . the kk mass eigenvalue @xmath46 is obtained as the solution of @xmath58 . for the case with @xmath59 and @xmath60 the asymptotic form of the kk mass eigenvalue
is simplified to @xmath61 where @xmath62 .
we find that the spectrum of the excited mode is shifted by the factor @xmath63 and the mass difference between neighbour modes is suppressed by the warp factor @xmath64 . in the same asymptotic limit @xmath39-dependence of the mode function vanishes on the @xmath2 brane .
it reduces to @xmath65 the profile of the mode functions @xmath45 are shown in fig .
[ fig : wf ] . from the figure
we know that the kk excited gauge bosons are localized in the vicinity of the @xmath2 brane in the rs brane world due to the finite curvature scale @xmath28 of the extra dimension , while the zero mode is flat in the direction of the extra dimension .
we will see that the localized kk excited modes enhance ( suppress ) the dynamical symmetry breaking on the @xmath2 ( @xmath66 ) brane .
next we define the coupling of the bulk gauge field to a fermion on the brane . by decomposing the metric as @xmath67 where @xmath68
, we define @xmath69 where the row and column of the matrix correspond to the indices @xmath70 and @xmath71 respectively .
@xmath72 gives the coupling of a fermion on the brane and the bulk gauge field .
that is described as @xmath73 where @xmath74 , @xmath75 , @xmath76 is the five dimensional gauge coupling and @xmath77 is the position of the brane . in the second line
we take @xmath32 gauge and rescale @xmath78 field as @xmath79 in order to canonically normalize the kinetic term of @xmath78 . by integrating ( [ eq : bbcoup ] ) over the extra dimensional coordinate we obtain the 4d effective gauge coupling written by @xmath80 where @xmath81 where @xmath82 is the ordinary four dimensional gauge coupling .
@xmath83 is defined by replacing @xmath84 with @xmath85 in eq .
( [ eq : gn ] ) . in the case
@xmath86 , i.e. @xmath87 ( see eq .
( [ eq : boundarycond ] ) ) , @xmath41 decouples in eq .
( [ eq:4dgcterm ] ) .
we notice that the coupling of the gauge boson excited modes @xmath88 are enhanced / suppressed by the value of the mode function @xmath89 on the @xmath90 brane , compared with it of zero mode .
thus we obtain larger effect from the kk excited gauge boson on the brane they localize .
in the previous section we have derived the lagrangian of the bulk gauge field and its coupling to a fermion on the brane . putting them together we obtain the effective lagrangian for the system of the bulk gauge theory couples to fermion on the @xmath91 brane as @xmath92 a_\nu^{(n ) } \nonumber \\ & & + \frac{1}{2 } \sum_{n=1}^{n_{\rm kk } } \tilde{a}_\mu^{(n ) } \left [ \eta^{\mu \nu } \left ( \partial^2 + m_n^2 \right ) -(1-\xi ) \partial_\mu \partial_\nu \right ] \tilde{a}_\nu^{(n ) } + { \cal l}_{\rm si } \nonumber \\ & & + \bar{\psi } \left ( i\partial_\mu + g a_\mu^{(0 ) } \right ) \gamma^\mu \psi + g \sum_{n = 1}^{n_{\rm kk } } \frac{\chi_n(y^\ast)}{\chi_0(y^\ast ) } \bar{\psi } a_\mu^{(n ) } \gamma^\mu \psi , \label{eq : efflag}\end{aligned}\ ] ] where @xmath93 represents the self interaction ( and mixing ) terms of @xmath94 and @xmath95 , and @xmath96 stands for the cutoff of the kk summation .
it should be noted that we introduce the gauge fixing parameter @xmath97 on the brane .
( see appendix in ref .
@xcite for more details . )
@xmath97 is chosen so that a qed like ward - takahashi identity is hold in the latter analysis .
we can easily extend the above system to higher dimensional bulk space - time with @xmath16 by replacing @xmath46 and @xmath89 as @xmath98 where @xmath99 correspond to the indices of the kk excited mode for each extra dimension @xmath100 respectively . @xmath101 and @xmath102 @xmath103
are given in eqs .
( [ eq : kkwf ] ) and ( [ eq : kkmasstorus ] ) respectively .
in addition we should interpret @xmath104 as @xmath105 with a cut - off condition , @xmath106 . in the following discussion
we regard the kk reduced lagrangian ( [ eq : efflag ] ) as an effective one below the bulk fundamental scale @xmath107 .
thus we cutoff the momentum of the loop integral , including only the propagator of bulk fields , at the scale @xmath107 . on the other hand we cutoff the momentum at the _ reduced _ cutoff scale @xmath108 on the brane in the case that the loop integral includes the brane fields .
loop integrals in the ladder schwinger - dyson equation is only the latter case . in the flat brane
world @xmath108 should be the same order of @xmath107 , while in the rs brane world we have an large difference between @xmath108 and @xmath107 due to the warp factor .
we also define @xmath96 as the number of the kk modes with their mass below the _ reduced _ cutoff scale @xmath108 . in our numerical analyses , however , we cutoff the kk mode summation at @xmath109 in order to avoid the sharp threshold effect . from the lagrangian ( [ eq : efflag ] ) we construct a schwinger - dyson equation for the fermion propagator , @xmath110 s(q ) \left [ -ig_n t^a \gamma^n \right ] d_{mn}^{(n ) } ( p - q ) , \label{eq : orgsdeq}\end{aligned}\ ] ] where @xmath111 is the gauge group representation matrix of the fermion , and @xmath112 , @xmath113 and @xmath114 are the free fermion propagator , full fermion propagator and full @xmath39-th excited kk gauge boson propagator respectively . in this paper
we use the ( improved ) ladder approximation in which @xmath115 is replaced by the free propagator , @xmath116 .
\end{aligned}\ ] ] by writing the full fermion propagator as @xmath117 the sd equation ( [ eq : orgsdeq ] ) becomes the simultaneous integral equation of @xmath118 and @xmath119 , @xmath120 where @xmath121 , \label{eq : kerl } \\
k_\xi ( p^2,q^2;m_n,\alpha_n ) & = & \frac{\alpha_n}{4\pi } \bigg [ 4f_{m_n^2 } ( p^2,q^2 ) + \frac{p^2+q^2}{m_n^2 } \left\ { f_{m_n^2 } ( p^2,q^2)- f_{\xi m_n^2 } ( p^2,q^2 ) \right\ } \nonumber \\ & & \hspace{1 cm } - \frac{p^2q^2}{m_n^2 } \left\ { f_{m_n^2}^2 ( p^2,q^2)- f_{\xi m_n^2}^2 ( p^2,q^2 ) \right\ } \bigg ] , \label{eq : kerk}\end{aligned}\ ] ] @xmath122 @xmath123 since our analysis is based on the effective theory on the four - dimensional brane below the _ reduced _ cutoff , the explicit uv cutoff @xmath108 in terms of the loop momentum _ on the brane _ appears . @xmath124 and @xmath125 are the wave function normalization factor and the mass function of the fermion respectively .
the mass function @xmath125 is the oder parameter of chiral symmetry breaking on the brane .
the chiral symmetry is broken down for @xmath126 . by solving eqs .
( [ eq : sda ] ) and ( [ eq : sdb ] ) we obtain the behavior of the dynamical fermion mass on the brane .
in this section we numerically solve the sd equation and analyze dynamical brane fermion mass induced by the bulk gauge theory . in the case of the bulk abelian gauge theory with
@xmath127 we can use the point vertex approximation and the results are shown in @xcite . in this section
we extend the analysis in @xcite to the bulk yang - mills theory . to solve the sd equation of yang - mills theory we use the improved ladder approximation where the running coupling is imposed to the vertex function .
the running coupling is derived from the truncated kk approach proposed in ref .
we use the asymptotic form ( [ eq : kkmassasym ] ) of the kk spectrum to evaluate the running coupling in rs brane world .
first we review the process of analyzing sd equation in usual 4d qcd @xcite , and then we extend it to the brane world bulk yang - mills theory .
for convenience we choose bulk qcd ( @xmath128 yang - mills theory with @xmath129 flavor fermion on the brane ) as a concrete example because its results can be compared with the usual 4d qcd .
we expect that the essential point of the results is not changed in the other bulk yang - mills theory such as techni - color .
taking @xmath130 in eqs .
( [ eq : sda ] ) and ( [ eq : sdb ] ) , the sd equation for usual qcd in four dimensional space - time ( i.e. qcd on the brane ) is obtained . in this case the integration kernels ( [ eq : kerl ] ) and ( [ eq : kerk ] ) become @xmath131 where @xmath132
we usually use landau gauge @xmath133 that results in @xmath134 , i.e. @xmath135 in the sd equation . in that case
the sd equation becomes single equation only with @xmath136 .
@xmath135 satisfies the qed like ward - takahashi identity .
we put the running coupling on the vertex function in the sd equation to compensate for neglecting three and four point gluon self interactions @xmath93 in the ladder approximation .
that is so called improved ladder approximation .
we use the 1-loop perturbative running coupling , @xmath137 where @xmath138 and @xmath139 .
we define @xmath140 by @xmath141 and @xmath142 $ ] where @xmath143 and @xmath144 stand for the gauge group and the fermion representation under the group respectively .
@xmath145 , @xmath146 and @xmath147 are defined by @xmath148 , @xmath149 and @xmath150 . in the case
that @xmath151 and @xmath144 is fundamental representation , these values are given by @xmath152 we also define @xmath153 , the dynamical scale of qcd , as @xmath154 and obtain @xmath155 following the analysis in ref .
@xcite we introduce ir cutoff @xmath156 in the running coupling as @xmath157 , \end{aligned}\ ] ] and we take the following approximation @xmath158 in the integration kernel ( [ eq : qcdkernell ] ) and ( [ eq : qcdkernelk ] ) .
it is called higashijima - miransky approximation .
this is due to the consideration that the mean value of the @xmath159-dependent part in @xmath160 where @xmath161 has only a negligible effect after the angle integration in the sd equation)$ ] , but we do nt use it here . ] .
it is known that the approximation ( [ eq : hmapprox ] ) violates the chiral ward - takahashi identity , we can keep the chiral ward - takahashi identity to use a non - local gauge . ] . in the present paper
we choose the gauge parameter @xmath97 to nearly hold the chiral ward - takahashi identity . in the following numerical analyses we use the experimental value @xmath162 and choose @xmath163 for the ir cutoff and @xmath156 @xcite .
in this paper we do nt touch the detailed value of the results but the order of them . ] .
the composite nambu - goldstone field corresponds to the composite higgs field in the dynamical electroweak symmetry breaking scenario .
its decay constant is obtained by the pagels - stokar approximation @xcite , @xmath164 here we show the numerical results of the improved ladder sd equation in 4d qcd .
we set the cutoff scale @xmath165 tev in order to compare with the result of the brane world models . by using the replacement ( [ eq : hmapprox ] )
we solve the sd equation ( [ eq : sda ] ) , ( [ eq : sdb ] ) and calculate the mass function @xmath125 .
the behavior of the mass function @xmath125 is shown as the thin solid line in fig . [
fig : sdb ] .
the scale of the fermion mass function is a little bit smaller than the qcd scale , @xmath166 . substituting the obtained mass function @xmath125 to eq .
( [ eq : psformula ] ) we obtain the decay constant of the composite scalar field . as is shown with the thin solid line in fig .
[ fig : fpi ] , the decay constant of 4d qcd is also smaller than the qcd scale , @xmath167 . since it is the result of the ordinary 4d qcd , the decay constant is too small as higgs field to break electroweak symmetry .
now we analyze the dynamical fermion mass induced by the bulk yang - mills theory in flat brane world , @xmath16 .
for simplicity we set the same radii for all extra directions , i , e , universal extra dimensions .
we study @xmath168 bulk yang - mills theory ( qcd ) with @xmath169 flavor fermion as a concrete example . for qcd extended in the compact extra dimension with the radius @xmath20 ,
the running coupling @xmath160 has the power low behavior @xcite in the region @xmath170 , @xmath171 , \end{aligned}\ ] ] where @xmath52 and @xmath172 and @xmath173 is the beta - function coefficient stimulating from the bulk fields
. thus we obtain the running coupling of the bulk qcd with @xmath129 flavor fermion on the brane , @xmath174 \theta ( z-\mu_r^2 ) } , \label{eq : powerlow}\end{aligned}\ ] ] where @xmath175 using the running coupling ( [ eq : powerlow ] ) , we numerically solve the improved ladder sd equation and find the behaviors of the fermion wave function @xmath124 and the mass function @xmath125 on the brane .
we analyze the sd equation on the @xmath176 @xmath177 brane , with @xmath178 tev , @xmath179 mev and @xmath180 as typical cases , where @xmath181 and @xmath108 are the bulk fundamental scale and the reduced cutoff scale on the brane respectively . for @xmath182 and @xmath183 , the behavior of @xmath125 and @xmath124
are drawn in figs .
[ fig : sdb ] and [ fig : sda ] respectively . for @xmath184
the fermion mass function behaves like 4d qcd , i.e. @xmath166 .
for @xmath1 the situation is dramatically changed .
the fermion mass function develops a value near the _ reduced _ cutoff scale , @xmath185 . using the pagels - stokar formula ( [ eq : psformula ] )
, we calculate the decay constant of the composite scalar field . as is shown in fig .
[ fig : fpi ] the scale of the decay constant is summarized as @xmath186 for @xmath184 and @xmath187 for @xmath1 .
therefore if the bulk space - time has seven or more dimensions , the bulk yang - mills theory induces the tev scale decay constant on the brane .
there is a possibility that the bulk qcd or techni - color can provide a composite higgs field in the flat brane world . in this section
we take @xmath188 tev .
we comment that this setup merges into the scenario of ` large extra dimension ' @xcite in which the fundamental scale is around tev and we assume the existence of a large extra dimension where only the gravity propagates . here
we study the bulk qcd in the rs brane world . in the warped brane world
a mass scale @xmath108 on the brane at @xmath90 is suppressed by the warp factor , i.e. @xmath189 , where @xmath190 is the fundamental scale of the theory and @xmath191 is the warp factor . as is discussed in the previous subsection ,
it is natural to consider that the effective theory of the brane fermion should be regularized by the _ reduced _ fundamental scale _ on the brane_. hence we cutoff the loop momentum in the sd equation at the _ reduced _ cutoff scale @xmath108 on the brane , @xmath192 for @xmath193 uv cutoff scale @xmath108 is suppressed by @xmath194 on the @xmath195 brane .
randall and sundrum obtain the weak - planck hierarchy form this warp suppression factor in the original rs model @xcite .
the behavior of the running coupling is non - trivial in the rs space - time . in this paper
we evaluate it in terms of the 4d effective theory below the cutoff scale @xmath108 by using the truncated kk approach . in the region @xmath59 and
@xmath60 , kk mass spectrum of the gauge boson reduces to the asymptotic form ( [ eq : kkmassasym ] ) . comparing it with the torus case ( [ eq : kkmasstorus ] ) , we find that the running coupling in the rs space - time is obtained by replacing @xmath196 with @xmath197 and shifting the kk lowest threshold by @xmath198 .
in addition we must take the orbifold condition into account .
it drops all the @xmath38 odd modes .
only a half contribution comes from the kk excited modes .
thus we replace @xmath199 in eq .
( [ eq : powerlow ] ) .
the running coupling reads @xmath200 \theta ( z-(3/4)^2\mu_{kr}^2 ) } , \label{eq : rspowerlow}\end{aligned}\ ] ] where @xmath119 and @xmath201 are given by eq .
( [ eq : bandtb ] ) . applying the truncated running coupling ( [ eq : rspowerlow ] )
we evaluate the sd equation on the @xmath2 brane , with @xmath202 tev , @xmath203 tev and @xmath179 mev as typical cases .
the solution of @xmath125 and @xmath124 are plotted in figs .
[ fig : sdb ] and [ fig : sda ] respectively . calculating the eq .
( [ eq : psformula ] ) we obtain the decay constant of the composite scalar as is shown in fig .
[ fig : fpi ] . in the rs brane world
both scales of the fermion mass function and the decay constant are near the _ reduced _ cutoff scale , @xmath204 , @xmath187 @xmath13 @xmath205 tev .
therefore the bulk yang - mills theory induces tev scale decay constant on the brane .
it means that the bulk qcd has possibility to realize the composite higgs scenario of electroweak symmetry breaking .
the original theory has only the planck scale , it has no tev scale source but the fermion mass and decay constant of the composite scalar are generated at tev scale through the warp suppression of the _ reduced _ cutoff scale on the @xmath2 brane .
kk excited modes of the gauge field are localized at the @xmath2 brane in rs model . it enhances the contribution from the kk modes and realize the tev scale composite scalar on the brane with only one extra dimension .
this is the difference from the ( 4 + 1)-dimensional flat brane world , @xmath206 .
the bulk gauge theory with only planck scale in rs brane world can acquire a tev scale dynamically due to a fermion pair condensation on the @xmath2 brane through the effect of the gauge boson kk excited modes localized at the brane .
the warp suppression mechanism in the rs model is realized dynamically .
the extension of above results to the _ bulk sm gauge fields _ will make us to propose two types of primary models of dynamical electroweak symmetry breaking .
we call one of them _
( 1,2)-3 model _ in which the first and second generations in the sm are confined on the @xmath66 brane , and the third generation is on the @xmath2 brane .
the kk excited mode is not localized on the @xmath66 brane shown in fig .
[ fig : wf ] .
the coupling between the kk excited gluons and the brane fermion are suppressed by the mode function @xmath207 on the @xmath66 brane .
therefore we expect that the top quark - pair is condensed on the @xmath2 brane , while the up and charm quark are not condensed . the top mode sm @xcite can be realized in rs brane world @xcite .
we note that the zero modes function , @xmath208 , are constant , so that the fermions on the both brane equally couples to the massless gauge filed ( see also fig . [
fig : wf ] ) .
we can construct the other model , _
( 1,2,3)-4 model _ , which is defined as follows .
all the sm fermions lie in @xmath66 brane and the fourth generation ( or the techni - fermion ) lies in the @xmath2 brane . in this model
the fourth generation can develops tev scale mass function and decay constant by kk excited gluons ( or techni - gluons ) .
it is expected that these models can provide composite higgs and the tev scale physics simultaneously .
although some problems are remained , e.g. how to dynamically realize the detailed mass relation ( dynamical yukawa couplings ) in sm , or flavor breaking etc .. the behaviors of @xmath125 shown in fig .
[ fig : sdb ] are divided into two pieces .
qcd in the flat bulk space - time with @xmath206 and @xmath209 is classified into _ yang - mills type _
, i.e. the contribution from the kk modes is extremely small and @xmath125 has a similar profile to the ordinary qcd ( qcd on the brane ) . on the other hand ,
flat bulk space - time with @xmath210 , @xmath183 and the rs brane world are _ njl type _ , i.e. kk modes contribution dominates the symmetry breaking . in figs .
[ fig : sdb ] , [ fig : sda ] and [ fig : fpi ] , we consider only the case @xmath211 .
finally we analyze the @xmath20- and @xmath28-dependence of the wave function @xmath124 and the mass function @xmath125 on the brane embedded in ( 4 + 1)-dimensional bulk space - time .
the behaviors of @xmath124 and @xmath125 are shown in figs .
[ fig : sdbr ] , [ fig : sdar ] and [ fig : fpir ] .
there are little @xmath20- and @xmath28-dependence .
only a small @xmath96-dependence is observed in the behavior of @xmath125 .
( [ eq : kkmasstorus ] ) or ( [ eq : kkmassasym ] ) show that the number of the kk mode @xmath96 increases as @xmath20 or @xmath212 increases .
the influence of the kk mode propagation seems to be enhanced for larger @xmath20 or @xmath212 . in fig .
[ fig : sdbr ] we clearly observe that the mass function @xmath125 grows up as @xmath20 or @xmath212 increases . in eqs .
( [ eq : powerlow ] ) or ( [ eq : rspowerlow ] ) the power low running terms contain the step function @xmath213 or @xmath214 .
since the kk mode mass scale @xmath196 and @xmath197 are proportional to @xmath215 and @xmath28 respectively , the contribution from these terms is enhanced as @xmath20 or @xmath212 increases .
the contribution from the power low running suppresses dynamical symmetry breaking .
for example , if we drop the power low running term in @xmath216-dimensional flat bulk space - time , @xmath125 blows up from the _ yang - mills type _ to the _ njl type _ as @xmath96 increases .
thus we conclude that the bulk gluon self interactions @xmath93 ,
i.e. the power low running coupling , act as a suppression factor for the dynamical mass on the brane .
we have studied the basic structure of dynamical symmetry breaking on the four dimensional brane in the bulk yang - mills theory ( qcd ) . using the 4d effective theory of the bulk qcd and the improved ladder sd equation , we have found that the dynamical mass scale can be affected by the _ reduced _ cutoff scales on the brane .
starting from the yang - mills theory in @xmath16 space - time which couples to a fermion on the four dimensional brane we derive four dimensional effective theory by kk reduction .
we take @xmath217 @xmath218 gauge in extra directions but leave the gauge fixing parameter free parallel to the brane .
our system needs a proper regularization because it is a theory beyond four dimensional space - time where the gauge theory is unrenomalizabele .
we impose the cutoff regularization at the uv _ reduced _ scale @xmath108 on the brane fermion lives in . for the flat brane world
the reduced scale should be equal to the fundamental scale in the original bulk yang - mills theory , i.e. @xmath219 . the rs brane world , however , has a reduced scale suppressed by the warp factor , that is @xmath220 which depends on the position in the extra direction
. based on the 4d effective theory we derive the sd equation for the fermion propagator on the brane . from the four dimensional point of view
the equation corresponds to the simultaneous integral equation .
the loop integral consists of one massless and @xmath96 massive vector field where @xmath96 is the number of the kk modes under the reduced cutoff scale . using the iteration method
we numerically solve it for typical values of radius , @xmath180 , and the ( ads ) curvature scale , @xmath221 . in order to make the analysis consistent with the qed like ward - takahashi identity
, we choose the appropriate value of the gauge fixing parameter for each analysis .
the results of the numerical analysis is as follows . for a flat extra dimension @xmath206 the dynamical mass of the brane fermion is the same scale as the usual 4d qcd result ,
i.e. there is little contribution from the kk excited gauge boson .
we also study the case of the flat brane world with more than one extra dimension .
the numerical analysis of the sd equation shows that the number of the extra dimension must be no less than three , @xmath1 , to generate the tev scale dynamical fermion mass on the brane .
therefore it is obtained in the case of the yang - mills theory in seven or more dimensional bulk space - time .
the rs brane world has one extra dimension which corresponds to @xmath206 case but it is curved . because the gauge boson kk excited modes localized at the @xmath2 brane due to the curvature effect ( see fig.[fig : wf ] ) , we have a significant result .
the dynamical mass of the brane fermion at the @xmath2 is the order of the _ reduced _ cutoff scale that is warp suppressed from the fundamental scale .
we also estimate the decay constant of the composite nambu - goldstone scalar field by using the pagels - stokar approximation .
the decay constant is also warp suppressed on the @xmath2 brane in the rs brane world . if the chiral symmetry is gauged as in the sm case , the composite scalar absorbed in the longitudinal component of the gauge boson .
this decay constant is directly related to the mass of the gauge boson .
the result in the rs brane world provides a possibility to realize the situation that electroweak symmetry breaking is triggered by the condensation of the composite higgs field on the @xmath2 brane .
such a mechanism is caused by the bulk yang - mills theory , for example the techni - color or the qcd itself in the sm .
therefore we can solve the weak - planck hierarchy from the warp factor , that is realized , in a sense , dynamically .
this is one of the dynamical realization of so called randall - sundrum model .
our analyses are based on the 4d effective theory of the bulk gauge theory with an explicit uv cutoff in the loop momentum integration .
we can also check the regularization dependence of our model by comparing the results of the lattice regularized ( deconstructed ) version .
the deconstructed bulk field theory is given in @xcite on the rs background .
the results of our analysis are also important in a theoretical sense .
we consider that the sd equation includes certain non - perturbative effect of the theory .
the non - trivial results obtained here implies that the bulk gauge theory may have rich non - perturbative dynamics that depends on the configurations of the extra space . in the brane world
we should take account of such effect in addition to the perturbative analysis .
our results imply that the bulk gauge dynamics can induce various types of the _ dynamical _ symmetry breaking in the brane world , depending on the bulk - brane configuration .
for instance , we may obtain the dynamical breaking of the grand unification gauge symmetry @xcite ( or even supersymmetry ? ) , in addition to the possibility of dynamical electroweak symmetry breaking emphasized in this paper .
we are also interested in the consistency of these models with the realistic d - brane system in the string theory .
the authors would like to thank kenji fukazawa , michio hashimoto , tatsuo kobayashi and masaharu tanabashi for fruitful discussions and correspondences .
we also thank kensaku ohkura for stimulating discussions .
l. e. ibanez , f. marchesano and r. rabadan , `` getting just the standard model at intersecting branes , '' jhep * 0111 * , 002 ( 2001 ) [ hep - th/0105155 ] ; r. blumenhagen , b. kors , d. lust and t. ott , `` the standard model from stable intersecting brane world orbifolds , '' nucl . phys .
b * 616 * , 3 ( 2001 ) [ hep - th/0107138 ] ; m. cvetic , g. shiu and a. m. uranga , `` three - family supersymmetric standard like models from intersecting brane worlds , '' phys .
lett . * 87 * , 201801 ( 2001 ) [ hep - th/0107143 ] ; c. kokorelis , `` exact standard model compactifications from intersecting branes , '' hep - th/0206108 .
i. antoniadis , `` a possible new dimension at a few tev , '' phys .
b * 246 * , 377 ( 1990 ) ; n. arkani - hamed , s. dimopoulos and g. r. dvali , `` the hierarchy problem and new dimensions at a millimeter , '' phys .
b * 429 * , 263 ( 1998 ) [ hep - ph/9803315 ] .
l. randall and r. sundrum , `` a large mass hierarchy from a small extra dimension , '' phys .
* 83 * , 3370 ( 1999 ) [ hep - ph/9905221 ] .
a. brignole , l. e. ibanez and c. munoz , `` towards a theory of soft terms for the supersymmetric standard model , '' nucl .
b * 422 * , 125 ( 1994 ) [ erratum - ibid .
b * 436 * , 747 ( 1994 ) ] [ hep - ph/9308271 ] ; g. f. giudice and r. rattazzi , `` theories with gauge - mediated supersymmetry breaking , '' phys . rept . * 322 * , 419 ( 1999 ) [ hep - ph/9801271 ] ; l. randall and r. sundrum , `` out of this world supersymmetry breaking , '' nucl . phys .
b * 557 * , 79 ( 1999 ) [ hep - th/9810155 ] ; d. e. kaplan , g. d. kribs and m. schmaltz , `` supersymmetry breaking through transparent extra dimensions , '' phys .
d * 62 * , 035010 ( 2000 ) [ hep - ph/9911293 ] ; t. kobayashi and k. yoshioka , `` kaluza - klein mediated supersymmetry breaking , '' phys .
lett . * 85 * , 5527 ( 2000 ) [ hep - ph/0008069 ] ; k. r. dienes , e. dudas and t. gherghetta , `` extra spacetime dimensions and unification , '' phys .
b * 436 * , 55 ( 1998 ) [ hep - ph/9803466 ] ; k. r. dienes , e. dudas and t. gherghetta , `` grand unification at intermediate mass scales through extra dimensions , '' nucl .
b * 537 * , 47 ( 1999 ) [ hep - ph/9806292 ] .
l. randall and m. d. schwartz , `` unification and the hierarchy from ads5 , '' phys .
lett . * 88 * , 081801 ( 2002 ) [ hep - th/0108115 ] .
b. a. dobrescu , `` electroweak symmetry breaking as a consequence of compact dimensions , '' phys .
b * 461 * , 99 ( 1999 ) [ hep - ph/9812349 ] ; h. c. cheng , b. a. dobrescu and c. t. hill , `` electroweak symmetry breaking and extra dimensions , '' nucl .
b * 589 * , 249 ( 2000 ) [ hep - ph/9912343 ] ; n. arkani - hamed , h. c. cheng , b. a. dobrescu and l. j. hall , `` self - breaking of the standard model gauge symmetry , '' phys .
d * 62 * , 096006 ( 2000 ) [ hep - ph/0006238 ] .
n. rius and v. sanz , `` dynamical symmetry breaking in warped compactifications , '' phys .
d * 64 * , 075006 ( 2001 ) [ hep - ph/0103086 ] .
m. hashimoto , m. tanabashi and k. yamawaki , `` top mode standard model with extra dimensions , '' phys .
d * 64 * , 056003 ( 2001 ) [ hep - ph/0010260 ] ; v. gusynin , m. hashimoto , m. tanabashi and k. yamawaki , `` dynamical chiral symmetry breaking in gauge theories with extra dimensions , '' phys .
d * 65 * , 116008 ( 2002 ) [ hep - ph/0201106 ] .
h. abe , h. miguchi and t. muta , `` dynamical fermion masses under the influence of kaluza - klein fermions in extra dimensions , '' mod .
a * 15 * , 445 ( 2000 ) [ hep - ph/0002212 ] ; h. abe , t. inagaki and t. muta , `` dynamical fermion masses under the influence of kaluza - klein fermions in randall - sundrum background , '' in _ fluctuating paths and fields _ , eds . w. janke , a. pelster , h .- j .
schmidt , and m. bachmann ( world scientific , singapore , 2001 ) [ hep - ph/0104002 ] .
h. abe , k. fukazawa and t. inagaki , `` chiral phase transition of bulk abelian gauge theories in the randall - sundrum brane world , '' prog .
* 107 * , 1047 ( 2002 ) [ hep - ph/0107125 ] .
t. inagaki , t. muta and s. d. odintsov , `` nambu - jona - lasinio model in curved space - time , '' mod .
lett . a * 8 * , 2117 ( 1993 ) [ hep - th/9306023 ] ; t. inagaki , `` curvature induced phase transition in a four - fermion theory using the weak curvature expansion , '' int . j. mod .
a * 11 * , 4561 ( 1996 ) [ hep - th/9512200 ] ; t. inagaki , t. muta and s. d. odintsov , `` dynamical symmetry breaking in curved spacetime : four - fermion interactions , '' prog .
suppl . * 127 * , 93 ( 1997 ) [ hep - th/9711084 ] .
t. inagaki , s. mukaigawa and t. muta , `` a soluble model of four fermion interactions in de sitter space , '' phys .
d * 52 * , 4267 ( 1995 ) [ hep - th/9505058 ] ; k. ishikawa , t. inagaki and t. muta , `` curvature induced dynamical symmetry restoration in einstein universe @xmath222 , '' mod .
a * 11 * , 939 ( 1996 ) ; k. ishikawa , t. inagaki , k. yamamoto and k. fukazawa , `` dynamical symmetry breaking in compact flat spaces , '' prog .
phys . *
99 * , 237 ( 1998 ) .
i. l. buchbinder , t. inagaki and s. d. odintsov , `` supersymmetric nambu - jona - lasinio model in an external gravitational field , '' mod .
lett . a * 12 * , 2271 ( 1997 ) [ hep - th/9702097 ] ; t. inagaki , s. d. odintsov and y. i. shilnov , `` dynamical symmetry breaking in the external gravitational and constant magnetic fields , '' int .
j. mod .
a * 14 * , 481 ( 1999 ) [ hep - th/9709077 ] .
v. a. miransky , m. tanabashi and k. yamawaki , `` dynamical electroweak symmetry breaking with large anomalous dimension and t quark condensate , '' phys .
b * 221 * , 177 ( 1989 ) ; v. a. miransky , m. tanabashi and k. yamawaki , `` is the t quark responsible for the mass of w and z bosons ?
, '' mod .
lett . a * 4 * , 1043 ( 1989 ) .
s. chang , j. hisano , h. nakano , n. okada and m. yamaguchi , `` bulk standard model in the randall - sundrum background , '' phys .
d * 62 * , 084025 ( 2000 ) [ hep - ph/9912498 ] ; t. gherghetta and a. pomarol , `` bulk fields and supersymmetry in a slice of ads , '' nucl . phys .
b * 586 * , 141 ( 2000 ) [ hep - ph/0003129 ] .
k. i. aoki , m. bando , t. kugo , m. g. mitchard and h. nakatani , `` calculating the decay constant f(pi ) , '' prog .
* 84 * , 683 ( 1990 ) .
v. a. miransky , `` on the quark effective mass in qcd .
( in russian ) , '' sov .
j. nucl .
phys . *
38 * , 280 ( 1983 ) [ yad .
fiz . * 38 * , 468 ( 1983 ) ] .
k. higashijima , `` dynamical chiral symmetry breaking , '' phys .
d * 29 * , 1228 ( 1984 ) .
h. pagels and s. stokar , `` the pion decay constant , electromagnetic form - factor and quark electromagnetic selfenergy in qcd , '' phys .
d * 20 * , 2947 ( 1979 ) .
l. randall and m. d. schwartz , `` quantum field theory and unification in ads5 , '' jhep * 0111 * , 003 ( 2001 ) [ hep - th/0108114 ] .
k. w. choi , h. d. kim and y. w. kim , `` gauge coupling renormalization in orbifold field theories , '' hep - ph/0202257 .
k. agashe , a. delgado and r. sundrum , `` gauge coupling renormalization in rs1 , '' hep - ph/0206099 . | the dynamically generated fermion mass is investigated in the flat brane world with @xmath0-dimensional bulk space - time , and in the randall - sundrum ( rs ) brane world .
we consider the bulk yang - mills theory interacting with the fermion confined on a four - dimensional brane .
based on the effective theory below the _ reduced _ cutoff scale on the brane , we formulate the schwinger - dyson equation of the brane fermion propagator . by using the improved ladder approximation
we numerically solve the schwinger - dyson equation and find that the dynamical fermion mass is near the _ reduced _ cutoff scale on the brane for the flat brane world with @xmath1 and for the rs brane world . in rs
brane world kk excited modes of the bulk gauge field localized around the @xmath2 brane and it enhances the dynamical symmetry breaking on the brane .
the decay constant of the fermion and the anti - fermion composite operator can be taken to be the order of the electroweak scale much smaller than the planck scale . therefore electroweak mass scale can be realized from only the planck scale in the rs brane world due to the fermion and the anti - fermion pair condensation .
that is a dynamical realization of randall - sundrum model which solves the weak - planck hierarchy problem .
kuns-1794 + hupd-0203 + hiroyuki abe@xmath3 and tomohiro inagaki@xmath4 @xmath5_department of physics , kyoto university , kyoto 606 - 8502 , japan _
+ @xmath6_department of physics , hiroshima university , hiroshima 739 - 8526 , japan _ + @xmath7_information media center , hiroshima university , hiroshima 739 - 8521 , japan _ |
all rings in this paper are assumed commutative and noetherian , and all modules unital .
recent years have seen a renewed interest in the properties of symbolic powers of prime ideals , especially in regular local rings .
for instance , see @xcite . often these properties , or conjectured properties ,
are related to the behavior of the symbolic powers @xmath7 of a prime ideal @xmath1 with respect to the the maximal ideal @xmath8 .
these properties frequently emerge from seemingly unrelated questions in algebra and geometry .
for instance , we have the following .
( `` sp '' stands for `` symbolic powers '' . )
[ conj : sp ] let @xmath0 be a regular local ring with prime ideals @xmath1 and @xmath2 such that @xmath3 and @xmath4 .
@xmath5 for all @xmath9 .
@xmath10 for all @xmath11 .
( sp-1 ) is a conjecture of kurano - roberts @xcite , motivated by their work on serre s positivity conjecture ; we briefly discuss the connection between these two conjectures in section [ sec : back ] below .
kurano and roberts verify ( sp-1 ) in @xcite for rings containing a field , and the current author gives a different proof in @xcite .
( sp-2 ) was conjectured and verified for rings containing a field by the current author in @xcite .
each conjecture is open in mixed characteristic . in our main theorem
we show that the ramified case of each conjecture follows from the unramified case .
the current author ( op .
has also made the following conjecture , which is fundamentally related to conjecture [ conj : sp ] .
first , we explain a notational convention .
in general , regular rings are denoted by @xmath12 , and their prime ideals by fraktur characters , e.g. , @xmath1 and @xmath2 .
rings that need not be regular are denoted by @xmath13 and their prime ideals by roman characters , e.g. , @xmath14 and @xmath15 ; the main exception being the maximal ideal @xmath16 .
we give the definitions of technical terms in the conjecture after the statement .
( `` i d '' stands for `` intersection dimension '' . )
[ conj : id ] let @xmath17be a quasi - unmixed local ring with prime ideals @xmath14 and @xmath15 such that @xmath18 .
if @xmath19 and @xmath20 is analytically unramified , then @xmath21 if @xmath22 and both @xmath20 and @xmath23 are analytically unramified , then @xmath21 in the statement , @xmath24 is the hilbert - samuel multiplicity of the local ring @xmath13 with respect to its maximal ideal .
a ring @xmath13 is _ quasi - unmixed _ if the completion @xmath25 is equidimensional , i.e. , for every minimal prime ideal @xmath26 of @xmath25 we have @xmath27 .
a theorem of ratliff @xcite says that @xmath13 is quasi - unmixed if and only if it is equidimensional and universally catenary . in particular , when @xmath13 is excellent , then @xmath13 is quasi - unmixed if and only if it is equidimensional .
the quotient ring @xmath20 is _ analytically unramified _
if @xmath28 is reduced , i.e. , the ideal @xmath29 is an intersection of prime ideals of @xmath25 .
this is automatic if @xmath13 is excellent . for rings that contain a field ,
the current author has verified both ( id-1 ) @xcite and ( id-2 ) @xcite . as with conjecture [ conj : sp ] , each conjecture ( id-@xmath30 ) is open in mixed characteristic . in our main theorem
below , we show that the general case of each conjecture follows from the case of a hypersurface over an unramified regular local ring .
some relations between our conjectures are straightforward to verify .
we summarize them in the following diagram @xmath31 \ar@{=>}[d ] & \text{(sp-2 ) } \ar@{=>}[d ] \\
\text{(id-1 ) } \ar@{=>}[r ] & \text{(sp-1 ) } } \end{gathered}\ ] ] and defer explanation to section [ sec : back ] . in our main theorem , which we now state , we show that the converses of the horizontal implications in this diagram hold . *
main theorem . * here we summarize the contents of this paper . in section [ sec : back ] we give some background material . in subsection
[ subsec : history ] we discuss the history of our conjectures . in subsection
[ subsec : imply ] we verify some implications between the conjectures . in subsection [ subsec : tools ] we include definitions and tools relevant to our results . in subsection
[ sec : ta ] we discuss the technical assumptions in the conjectures . in section [ sec : main ] , we prove our main theorem and verify a weak version of ( id-2 ) in mixed characteristic .
finally , in section [ sec : ex ] we give an example to show that a potential generalization of ( sp-1 ) does not hold .
we begin this section by discussing the connection between serre s positivity conjecture and ( sp-1 ) , and the other evidence supporting our conjectures .
let @xmath0 be a regular local ring of dimension @xmath32 , and @xmath33 prime ideals of @xmath12 such that @xmath3 .
serre ( * ? ? ?
* thorme 3 of chapitre v ) proved that @xmath34 we shall refer to this result as serre s intersection theorem .
serre defined the _ intersection multiplicity _ of @xmath35 and @xmath36 by the formula @xmath37 where @xmath38 is the length of the @xmath12-module @xmath39 . for @xmath12 unramified , serre proved the following : ( i ) @xmath40 , and ( ii ) @xmath41 if and only if @xmath42 .
he conjectured that these results hold even when @xmath12 is ramified .
gillet - soul ( * ? ? ?
* thorme 1 ) and roberts ( * ? ? ?
* theorem 1 ) independently proved the vanishing conjecture : if @xmath42 , then @xmath41 .
gabber @xcite , @xcite , @xcite proved the nonnegativity conjecture : @xmath40 .
the positivity conjecture is the converse of the vanishing conjecture , and is still open .
the following theorem of kurano - roberts @xcite shows the connection between the positivity conjecture and ( sp-1 ) .
essentially , it says that , for rings that are equicharacteristic or ramified , the positivity conjecture implies ( sp-1 ) .
it is proved by applying gabber s methods to the positivity conjecture .
[ thm : kr ] let @xmath0 be a regular local ring with prime ideals @xmath1 and @xmath2 such that @xmath3 and @xmath4 . if @xmath12 is ramified or contains a field and @xmath43 , then @xmath5 for all @xmath9 .
this theorem shows that ( sp-1 ) holds in equal characteristic , because the same is true of the positivity conjecture .
because kurano and roberts expect the positivity conjecture to hold , this theorem motivated them to formulate their conjecture , ( sp-1 ) .
it is worth noting that kurano and roberts @xcite verify the conjecture directly for rings containing a field of characteristic 0 , with no reference to positivity .
the current author @xcite has verified ( sp-1 ) and ( sp-2 ) when @xmath12 contains a field of arbitrary characteristic , also with no reference to positivity .
this is accomplished by first proving that ( id-2 ) holds for rings containing a field , c.f .
then one verifies that ( id-2 ) implies ( sp-2 ) , as follows .
if @xmath12 is a regular local ring , @xmath1 a prime ideal of @xmath12 and @xmath44 a nonzero element of @xmath1 ,
then @xmath45 if and only if @xmath46 . if @xmath1 and @xmath2 are as in the statement of ( sp-2 ) and @xmath47 , then we may apply ( id-2 ) to the hypersurface @xmath48 to arrive at a contradiction .
the evidence in support of ( id-@xmath30 ) is the fact that it holds in equal characteristic .
it is worth noting that ( id-@xmath30 ) generalizes serre s intersection theorem because , if @xmath13 is regular , then @xmath49 .
one can take this as support for the conjecture as well . in this subsection
we discuss the implications in diagram ( [ diagram ] ) .
we have already seen how ( id-2 ) implies ( sp-2 ) , and a similar argument shows that ( id-1 ) implies ( sp-1 ) .
clearly , ( sp-2 ) implies ( sp-1 ) .
it is not immediately clear that ( id-2 ) implies ( id-1 ) , because @xmath23 need not be analytically unramified in ( id-1 ) .
however , ( id-2 ) implies ( id-1 ) for complete domains , and the proof of our main theorem shows that we can reduce ( id-1 ) to the case of a complete domain .
it follows that ( id-2 ) implies ( id-1 ) . at this time
, we do not know if the converses of the vertical implications in diagram ( [ diagram ] ) hold . with theorem [ thm : kr ] in mind
, we remark that we do not know whether serre s positivity conjecture implies any of our conjectures , except ( sp-1 ) in the ramified case . also , we do not know if any of our conjectures implies positivity . in this subsection ,
we give the definition of the hilbert - samuel multiplicity and catalog the properties we will use in the proof of our results .
let @xmath17 be a local ring , @xmath50 a nonzero , finitely generated @xmath13-module of dimension @xmath32 , and @xmath51 is an ideal of @xmath13 such that @xmath52 has finite length .
the hilbert function @xmath53 agrees with a polynomial @xmath54 $ ] of degree @xmath32 , for @xmath55 .
the hilbert - samuel multiplicity is the positive integer @xmath56 such that @xmath57 .
when there is no ambiguity , we shall write @xmath58 .
if @xmath59 , we write @xmath60 or @xmath61 .
the following is a list of properties satisfied by the hilbert - samuel multiplicity that are needed for the proof of our main theorem .
let @xmath13 be a ring and @xmath50 a finite @xmath13-module .
[ prop:1 ] _ ( additivity formula ) _ let @xmath13 be a local ring
. then @xmath62 where the sum is taken over all primes @xmath14 of @xmath13 such that @xmath63 .
this sum is finite because we need only take the sum over such primes which are also in the support of @xmath50 .
( c.f . , bruns - herzog @xcite . )
[ prop:2 ] let @xmath0 be a local subring of @xmath13 such that the extension @xmath64 is module - finite .
then @xmath13 is a semilocal ring such that @xmath65 .
let @xmath66 be the set of maximal ideals of @xmath13 such that @xmath67 and @xmath68
. then @xmath69 e_{a_{{{\mathfrak{n}}}_i}}\!({{\mathfrak{m}}}a_{{{\mathfrak{n}}}_i},a_{{{\mathfrak{n}}}_i}).\ ] ] ( c.f . ,
nagata ( * ? ? ?
* ( 23.1 ) ) . )
[ prop:3 ] let @xmath12 be a subring of @xmath13 such that the extension @xmath64 is module - finite , and let @xmath70 be a prime ideal of @xmath12 .
then there are finitely many prime ideals of @xmath13 that contract to @xmath70 in @xmath12 .
let @xmath71 be the set of prime ideals of @xmath13 such that @xmath72 and @xmath73
. then @xmath74 e_{a_{s_i}}\!({{\mathfrak{s}}}a_{s_i},a_{s_i})\ ] ] where @xmath75 and @xmath76 are the residue fields of @xmath77 and @xmath78 , respectively .
( apply [ prop:2 ] to the extension @xmath79 . )
let @xmath80 be a flat , local homomorphism of local rings @xmath17 and @xmath81 such that @xmath82 . then @xmath83 .
( c.f . , herzog @xcite . )
[ prop:5 ] let @xmath13 be a local ring and @xmath51 an ideal of @xmath13 such that @xmath84 has finite length .
if @xmath50 is an @xmath13-module of positive rank @xmath85 , then @xmath86 .
( c.f . , @xcite ) [ prop:6 ]
if @xmath13 is local with infinite residue field , then there is an ideal @xmath51 generated by a system of parameters for @xmath13 such that @xmath87 .
northcott - rees ( * ? ? ?
* theorem 1 of 6 ) . ) [ prop:7 ] assume that @xmath88 is local and equidimensional , and contains an excellent local domain @xmath0 such that the extension @xmath89 is module - finite .
let @xmath14 and @xmath15 be prime ideals of @xmath13 such that @xmath18 , and let @xmath90 and @xmath91 . if @xmath92 , then @xmath3 .
( c.f . , @xcite .
) from the proof of our main theorem below , one can see why we make certain assumptions in ( sp-@xmath30 ) and ( id-@xmath30 ) .
some of these assumptions are necessary ; we are not sure of the necessity of others . in this subsection , we give some explanation and historical precedence for some of these assumptions .
straightforward examples show that the following assumptions in ( sp-@xmath30 ) are necessary : ( i ) @xmath12 is regular , ( ii ) @xmath3 , and ( iii ) @xmath4 .
when we assume ( i ) and ( ii ) , serre s intersection theorem implies that @xmath93 .
that is , in ( iii ) we are requiring the sum @xmath94 to achieve its maximal value .
similar examples show that we must assume that @xmath18 in ( id-@xmath30 ) . in ( id-@xmath30 ) , we assume that @xmath13 is quasi - unmixed .
there are straightforward examples of complete local rings that are not equidimensional for which the ( id-@xmath30 ) both fail , so the requirement that @xmath13 be equidimensional is certainly necessary . since our arguments involve passing to the completion @xmath25 , we must assume that @xmath25 is equidimensional .
we do not know whether the full strength of the quasi - unmixedness assumption is needed .
however , if one is inclined to assume that one s rings are excellent , then `` quasi - unmixed '' is equivalent to `` equidimensional '' , as we noted above , and is therefore necessary . in ( id-@xmath30 )
we assume that @xmath20 is analytically unramified .
the purpose of this assumption is to guarantee that the multiplicity @xmath95 is well - behaved under passing to the completion @xmath25 . to place this in an historical context , note that our proof is very similar to nagata s proof of the following result .
[ thm : nagata ] ( ( * ? ? ? * ( 40.1 ) ) ) _ let @xmath14 be a prime ideal of a local ring @xmath13 . if @xmath20 is analytically unramified and if @xmath96 , then @xmath97 .
_ regarding the analytically unramified assumption in this theorem , nagata ( ( * ? ? ?
* a2 ) ) writes the following : `` it is not yet known to the writer s knowledge whether or not ( 40.1 ) is true without assuming that @xmath14 is analytically unramified . ''
if it is shown that this condition can be omitted from the statement of theorem [ thm : nagata ] , then one will probably be able to show that the corresponding conditions should be omitted from the statements ( id-@xmath30 ) .
the following theorem includes the main theorem announced in section [ sec : intro ] .
[ thm : reduce ] with the notation of conjectures [ conj : sp ] and [ conj : id ] , fix @xmath98 or @xmath99
. then the following conditions are equivalent . 1 . _
( sp-@xmath30 )
_ holds for all regular local rings .
[ item:1 ] 2 . _
( sp-@xmath30 )
_ holds for all complete , unramified regular local rings of mixed characteristic .
[ item:2 ] 3 . for every prime number @xmath100 , every complete @xmath100-ring @xmath101 and every @xmath102 , _
( sp-@xmath30 )
_ holds for @xmath103_{(p , x_1,\ldots , x_d)}$ ] .
[ item:3 ] 4 . _
( id-@xmath30 )
_ holds for all quasi - unmixed local rings .
[ item:4 ] 5 . _
( id-@xmath30 )
_ holds for all hypersurfaces over an arbitrary complete , unramified regular local ring of mixed characteristic .
[ item:5 ] 6 . for every prime number @xmath100 , every complete @xmath100-ring @xmath101 and every @xmath102 , _ ( id-@xmath30 )
_ holds for all hypersurfaces over @xmath103_{(p , x_1,\ldots , x_d)}$ ] .
[ item:6 ] in the statement , a _
@xmath100-ring _ is a discrete valuation ring @xmath104 whose maximal ideal is generated by the prime number @xmath100 .
we summarize the implications we shall prove in the following diagram .
@xmath105 & ( \ref{item:2 } ) \ar@{<=>}[r ] \ar@{<=>}[d ] & ( \ref{item:3 } ) \ar@{<=>}[d ] \\ ( \ref{item:4 } ) \ar@{=>}[u ] \ar@{<=>}[r ] & ( \ref{item:5 } ) \ar@{=>}[r ] & ( \ref{item:6 } ) } \ ] ] the implications `` ( [ item:1 ] ) @xmath106 ( [ item:2 ] ) '' and `` ( [ item:4])@xmath106([item:5 ] ) '' are obvious .
the implication `` ( [ item:2])@xmath106([item:3 ] ) '' follows by passing to the completion @xmath107 of @xmath12 ; use the faithful flatness of @xmath108 as in the first step of `` ( [ item:3])@xmath106([item:2 ] ) '' . the implication `` ( [ item:5])@xmath106([item:6 ] ) '' also follows by passing to the completion
; argue as in the proof of @xcite . as noted in the introduction ,
if @xmath12 is a regular local ring with prime ideal @xmath1 and @xmath44 is a nonzero element of @xmath1 , then @xmath46 if and only if @xmath45 . from this the implications `` [ item:4 ] @xmath106 [ item:1 ] '' , `` ( [ item:2])@xmath109([item:5 ] ) '' and `` ( [ item:3])@xmath109([item:6 ] ) '' follow easily . `` ( [ item:5])@xmath106([item:4 ] ) '' .
we verify this implication in the case @xmath110 ; the case @xmath98 is similar .
let @xmath17 be a quasi - unmixed , local ring with prime ideals @xmath14 and @xmath15 such that @xmath20 and @xmath23 are analytically unramified , @xmath18 , and @xmath111 .
we now show that we may assume without loss of generality that @xmath13 is a complete local domain of mixed characteristic with infinite residue field .
pass to the ring @xmath112_{{{\mathfrak{n}}}a[x]}$ ] to assume that the residue field of @xmath13 is infinite .
let @xmath113 , @xmath114 and @xmath115 .
then @xmath116 is a local ring with maximal ideal @xmath117 .
multiplicities are preserved by [ prop:4 ] .
that is , @xmath118 , @xmath119 and @xmath120 . both rings @xmath121 and @xmath122
are analytically unramified by ( * ? ? ?
* ( 36.8 ) ) . also , @xmath116 is quasi - unmixed : this is equivalent to it being universally catenary and equidimensional by @xcite , and since @xmath13 satisfies these properties , the same is true of @xmath116 .
if the result holds in @xmath116 , then @xmath123 as desired .
pass to the completion @xmath124 to assume that @xmath13 is complete and equidimensional with infinite residue field .
let @xmath26 be a prime ideal of @xmath25 that is minimal over @xmath29 such that @xmath125 , and similarly for @xmath126 . since @xmath20 is analytically unramified , @xmath127 .
the fact that the extension @xmath128 is flat therefore implies that @xmath129 by [ prop:4 ] .
similarly , @xmath130 .
if the result holds for @xmath25 , then it holds for @xmath13 , as in the previous step .
pass to the quotient @xmath84 for a suitably chosen minimal prime @xmath51 , to assume that @xmath13 is a complete domain with infinite residue field .
to make this reduction , it suffices to verify that there is a minimal prime @xmath51 of @xmath13 contained in @xmath131 such that @xmath132 .
first , we show how this gives the desired reduction .
let @xmath133 with prime ideals @xmath134 , @xmath135 and @xmath136 .
since @xmath13 is equidimensional , @xmath137 , and since @xmath138 , it follows that @xmath139 and @xmath140 .
therefore , as in step 1 , we may pass to @xmath141 .
now we prove that such a minimal prime @xmath51 exists .
let @xmath142 .
suppose that @xmath143 for every @xmath144 such that @xmath145 .
( this supposition includes the hypothetical possibility that no @xmath146 is contained in @xmath131 . ) by ( * ? ? ? * ( 40.1 ) ) , @xmath147 for every @xmath146 contained in @xmath14 , and similarly for @xmath146 contained in @xmath15 .
the additivity formula then implies that @xmath148{\mathrm{len}}(a_{i_j } ) \\ & \leq\sum_{\substack{i_j\subseteq p \\ i_j\not\subseteq q } } e(a / i_j){\mathrm{len}}(a_{i_j } ) + \sum_{\substack{i_j\subseteq q \\ i_j\not\subseteq
p } } e(a / i_j){\mathrm{len}}(a_{i_j } ) \\ & \quad + \sum_{i_j\subseteq p\cap q } e(a / i_j){\mathrm{len}}(a_{i_j } ) \\ & \leq\sum_j e(a / i_j){\mathrm{len}}(a_{i_j } ) \\ & = e(a).\end{aligned}\ ] ] this clearly contradicts the assumption that @xmath111 .
thus , there is a minimal prime @xmath51 of @xmath13 such that @xmath138 and @xmath132 , as claimed .
since ( id-@xmath30 ) has been verified for rings containing a field , we assume without loss of generality that @xmath13 has characteristic 0 and that @xmath149 has characteristic @xmath150 .
let @xmath104 be a coefficient ring for @xmath13 , which is a complete @xmath100-ring . since the residue field of @xmath13 is infinite , [ prop:6 ]
implies that there exists system of parameters @xmath151 of @xmath13 such that the ideal @xmath152 satisfies @xmath153 .
let @xmath154\!]$ ] which is a complete domain contained in @xmath13 such that the extension @xmath89 is module finite , and the induced map on residue fields is an isomorphism .
let @xmath155 , @xmath90 and @xmath91 . since @xmath156
, we have @xmath157 .
therefore , @xmath158=e_r({{\mathfrak{m}}},a)={\mathrm{rank}}_r(a ) e(r)\ ] ] by [ prop:2 ] . since @xmath111 , [ prop:7 ] implies that @xmath3 . if @xmath12 is a regular local ring , then serre s intersection theorem implies that @xmath159 as desired .
therefore , we assume that @xmath12 is not regular .
let @xmath160 be indeterminates and let @xmath161\!]$ ] .
then @xmath162 surjects onto @xmath12 via the natural homomorphism @xmath163 which sends @xmath164 to @xmath165 .
since @xmath166 and @xmath12 is a domain , the kernel of @xmath163 is a height 1 prime of @xmath162 and is therefore principal .
thus , @xmath12 is a hypersurface over @xmath162 .
if we can show that @xmath167 , then our assumptions imply that @xmath168 , completing the proof .
let @xmath169 so that @xmath170 .
we claim that @xmath171 .
as before , @xmath172 \\ \intertext{(where the sum is taken over all primes $ p_i$ of $ a$ contracting to $ { { \mathfrak{p}}}$ in $ r$ ) } & \geq e_{a_p}({{\mathfrak{p}}}a_p , a_p ) \\ & \geq e_{a_p}(pa_p , a_p)\\ & = e(a_p).\end{aligned}\ ] ] similarly , @xmath173 .
thus , @xmath174 and since @xmath175 , we have @xmath167 , as desired . `` ( [ item:3])@xmath106([item:2 ] ) '' .
the main ideas for this proof come from hochster @xcite and dutta @xcite , each of which relies heavily on artin approximation @xcite and on the work of peskine - szpiro @xcite .
let @xmath104 be a complete @xmath100-ring and @xmath102 , and suppose that there were a counterexample to ( sp-2 ) in @xmath176\!]$ ] .
that is , there exist prime ideals @xmath33 in @xmath12 and an element @xmath177 such that @xmath3 , @xmath178 and @xmath47 .
since our conjectures are easily verified for rings of dimension less than 2 , we have @xmath179 . we may assume without loss of generality that @xmath180 is algebraically closed , as follows .
let @xmath181 denote the algebraic closure of @xmath182 .
by matsumura @xcite there exists a @xmath100-ring @xmath183 such that @xmath184 and a flat ring homomorphism @xmath185 .
replace @xmath183 by its completion to assume that @xmath183 is a complete @xmath100-ring .
let @xmath186\!]$ ] , which is a regular local ring of dimension @xmath187 with maximal ideal @xmath188 .
the induced ring homomorphism @xmath189 is flat and local , so we may fix prime ideals @xmath190 such that @xmath191 and @xmath192 .
it follows that @xmath193 and @xmath194 , since going - down holds between @xmath12 and @xmath195 , and each ring is a catenary domain ; apply @xcite . for @xmath196 and @xmath197
we have @xmath198 , and it follows that @xmath199 . since @xmath200 , there exists @xmath201 such that @xmath202 . since @xmath191 , it follows that @xmath203 , and the fact that @xmath204 implies that @xmath205 .
similarly , @xmath206 . finally , if @xmath207 , then @xmath208 by faithful flatness , a contradiction . thus , the counterexample in @xmath12 ascends to a counterexample in @xmath195 , and we may assume that @xmath209 is algebraically closed . let @xmath210_{(p , x_1,\ldots , x_d)}$ ] with maximal ideal @xmath211 .
we shall show that the counterexample in @xmath12 gives rise to a counterexample in @xmath212 .
more specifically , we shall show that there exist prime ideals @xmath213 in @xmath212 and @xmath214 such that @xmath215 , @xmath216 , @xmath217 , and @xmath218 .
first , we show how the given counterexample gives rise to a counterexample in the henselization @xmath219 of @xmath212 , and therefore in a pointed tale neighborhood @xmath220 of @xmath212 .
this is accomplished by showing that the essential data describing the counterexample is determined by a finite number of polynomial equations ( necessarily in a finite number of variables ) with coefficients in @xmath212 .
artin s approximation theorem will then yield a solution to these equations in @xmath219 . furthermore , given an integer @xmath221 we can guarantee that the solutions in @xmath219 will be congruent to the original solutions modulo @xmath222 .
since there are finitely many variables , the solution will lie in a pointed tale neighborhood @xmath220 of @xmath212 .
note that @xmath219 and each pointed tale neighborhood @xmath220 of @xmath212 is an unramified regular local ring of dimension @xmath223 with regular system of parameters @xmath224 .
let @xmath225 be a minimal set of generators of @xmath1 and @xmath226 a minimal set of generators of @xmath2 .
a finite number of equations determine the minimal free resolutions of @xmath35 and @xmath36 ; c.f .
let @xmath227 be the variables that `` keep track '' of the elements @xmath228 .
applying artin approximation will yield ideals @xmath229 in a pointed tale neighborhood @xmath230 of @xmath212 that are generated by the solutions in @xmath212 that are substituted for the variables @xmath227 .
a finite number of equations determine the fact that @xmath3 .
there exists @xmath231 such that @xmath232 for @xmath233 , so we may write @xmath234 for @xmath233 and @xmath235 . using variables @xmath236 for the coefficients , and the variables @xmath237 already designated , we can express this information in @xmath223 equations .
a finite number of equations determine the dimension of @xmath35 and @xmath36 .
this is a result of hochster @xcite .
a proof can be found in @xcite .
see @xcite for a different proof .
a finite number of equations determine the fact that a given sequence @xmath238 is a system of parameters for @xmath35 .
we can keep track of @xmath239 .
we can keep track of the fact that the @xmath240 are elements of @xmath8 and that the sequence has @xmath239 elements .
also we can keep track of the fact that @xmath241 has dimension 0 .
it follows that we can keep track of whether a given element @xmath242 is part of a system of parameters for @xmath35 by extending it to a full system of parameters .
. a finite number of equations determine the fact that @xmath200 .
there are two cases .
case 1 . if @xmath243 , then there exists @xmath244 such that @xmath202 .
that is , we can write @xmath245 where the sum is finite and the coefficients are in @xmath12 . since @xmath244 , @xmath246 is part of a system of parameters for @xmath35 ; we have already noted that we can keep track of this fact with a finite number of equations . by using variables @xmath247 we
can also keep track of the fact that @xmath248 .
let @xmath249 be the variables we use to keep track of the elements @xmath250 .
if @xmath251 , then we can keep track of the fact that @xmath251 with a single equation , as in case 1 . using @xmath252 and
the equation @xmath253 keeps this case consistent with case 1 . in our application of artin approximation
, we may then conclude that there are elements @xmath254 such that @xmath255 and either @xmath256 or @xmath257 is part of a system of parameters for @xmath258 .
similarly , there is an element @xmath259 such that @xmath260 and either @xmath261 or @xmath262 is part of a system of parameters for @xmath263 .
choosing @xmath264 , we require that the solutions in @xmath220 are congruent to the original solutions in @xmath12 modulo @xmath265 .
in particular , since @xmath266 , it follows that @xmath267 and , therefore , that @xmath268 . to summarize
, we have shown the existence of the following : * a pointed tale neighborhood @xmath230 of @xmath269 and an element @xmath270 ; * ideals @xmath271 such that @xmath272 , @xmath273 and @xmath274 ; * an element @xmath275 such that @xmath255 and either @xmath256 or @xmath257 is part of a system of parameters for @xmath258 ; * an element @xmath259 such that @xmath260 and either @xmath261 or @xmath262 is part of a system of parameters for @xmath263 .
in particular , @xmath220 is a regular local ring of dimension @xmath223 , essentially of finite type and smooth over @xmath104 , @xmath276 , and @xmath277 .
let @xmath278 be a minimal prime of @xmath258 such that @xmath279 , and let @xmath280 be a minimal prime of @xmath263 such that @xmath281
. then @xmath282 and @xmath283 .
since @xmath257 is either part of a system of parameters on @xmath258 or @xmath252 , it follows that @xmath284 . since @xmath285
, it follows that @xmath286 .
similarly , @xmath287 .
thus , we have a counterexample in @xmath220 , as desired .
next , we show that we may assume without loss of generality that @xmath288 is irreducible in our conjecture .
since @xmath12 is a unique factorization domain , write @xmath289 with each @xmath290 irreducible . for @xmath291 let @xmath292 be the unique integers such that @xmath293 and @xmath294 .
in other words , let @xmath295 and @xmath296 .
assuming that @xmath297 , we claim that there exists an integer @xmath30 such that @xmath298 and @xmath299 .
this will show that our counterexample gives rise to an irreducible counterexample .
suppose that each @xmath300 .
( if @xmath301 or @xmath302 , this is automatic by ( * ? ? ?
* ( 38.3 ) ) . ) from this it follows that @xmath303 the additivity formula implies that @xmath304 or , in other words , @xmath305 since @xmath306 , it follows that @xmath307 . similarly , @xmath308 .
therefore , @xmath309 , and it follows that @xmath310 which contradicts our hypothesis .
we also note that @xmath311 , as follows .
suppose not . since @xmath312 , the fact that @xmath313 implies that @xmath314 is a regular local ring of dimension @xmath32 with prime ideals
@xmath315 and @xmath316 such that @xmath199 , and @xmath317 this contradicts serre s intersection theorem . observe that @xmath318 because @xmath288 is irreducible and @xmath319 .
finally , we show how the following theorem of dutta @xcite gives rise to a counterexample in @xmath212 , thus completing our proof .
let @xmath320 be a regular local ring of dimension @xmath223 , essentially of finite type and smooth over an excellent discrete valuation ring @xmath321 such that @xmath322 is separably generated over @xmath323 .
let @xmath324 be such that @xmath325 ( i.e. , @xmath326 form an @xmath13-sequence ) .
then there exists a regular local ring @xmath327 such that \1 .
the ring @xmath162 is a localization of a polynomial ring @xmath328 $ ] at a maximal ideal of the type @xmath329 where @xmath330 is a monic irreducible polynomial in @xmath331 $ ] and @xmath332 is an excellent discrete valuation ring contained in @xmath13 ; moreover , @xmath13 is an tale neighborhood of @xmath162 .
there exists an element @xmath333 in @xmath334 such that @xmath335 is an isomorphism .
furthermore @xmath336 .
we apply this theorem to the ring @xmath337 and the element @xmath338 .
the discrete valuation ring @xmath339 whose existence is guaranteed by the theorem is constructed by choosing elements @xmath340 such that their residues modulo @xmath50 form a transcendence basis of @xmath322 over @xmath323 and then setting @xmath341_{\pi v[y_1,\ldots , y_t]}$ ] . the fact that @xmath342 implies that we may take @xmath343 in the conclusion .
since @xmath209 is algebraically closed , the maximal ideal @xmath329 will be of the form @xmath344 for some element @xmath345 ; therefore , @xmath346 .
without loss of generality , we may assume that @xmath347 .
let @xmath348 and @xmath349 . then @xmath350 . since the induced map @xmath351 is an isomorphism , we see that @xmath352 and @xmath353 .
furthermore , under this isomorphism , @xmath354 , which implies that @xmath355 . also , @xmath356 , so that @xmath357 ; in particular , @xmath358 . similarly , @xmath359 and @xmath360
thus , our counterexample in @xmath220 gives rise to a counterexample in @xmath212 , completing the proof .
using methods similar to those employed in the proof of theorem [ thm : reduce ] , we verify the following weaker version of ( id-2 ) .
[ thm : weak(id-2 ) ] let @xmath17 be a quasi - unmixed local ring with prime ideals @xmath14 and @xmath15 such that @xmath18 . if @xmath111 , e.g. , if @xmath19 , then @xmath361 .
let @xmath362 be as in the statement of the theorem .
as in the proof of the implication `` ( [ item:5])@xmath106([item:4 ] ) '' in theorem [ thm : reduce ] we may assume that @xmath13 is a hypersurface over the ring @xmath161\!]$ ] , with surjection @xmath363 .
let @xmath364 which is the maximal ideal of @xmath162 , @xmath365 and @xmath366 .
then @xmath367 is @xmath368-primary , so by serre s intersection theorem , @xmath369 as desired .
with conjecture [ conj : sp ] in mind , it is natural to ask , `` what is the correct conjecture to make for symbolic powers in nonregular rings ? ''
this is a tricky business , as symbolic powers can behave badly in general .
for example , in the ring @xmath370\!]/(xy)=k[\![x , y]\!]$ ] , the prime ideal @xmath371 satisfies @xmath372 for all @xmath9 . however , a few interesting suggestions have been offered in this direction .
for instance , we have question [ q ] below , which is motivated by the work of hochster - huneke @xcite . there
, the authors use tight closure methods to investigate symbolic powers .
for an ideal @xmath51 in a ring @xmath12 of prime characteristic @xmath150 , let @xmath373 denote the tight closure of @xmath51 .
a thorough introduction to the theory of tight closure can be found in ( * ? ? ? * chapter 10 ) the monograph of huneke @xcite . [ q ]
let @xmath182 be an algebraically closed field of prime characteristic @xmath150 and @xmath374/(f)=k[x_1,\ldots , x_d]\ ] ] for some prime element @xmath375 and @xmath376 .
let @xmath33 prime ideals of @xmath12 such that @xmath377 and @xmath378 .
does the containment @xmath379 hold for all @xmath380 ? fix an integer @xmath383 , and let @xmath384/(xy(z+u)-u^sz)=k[x , y , z , u].\ ] ] applying eisenstein s criterion to @xmath385[x]$ ] shows that @xmath44 is prime .
let @xmath386 , @xmath387 and @xmath388
. then @xmath377 and @xmath389 . by hochster - huneke (
* ( 6.2 ) ) , every element in the jacobian ideal has a power that is a test element . in this example , the jacobian ideal is @xmath390 fix an integer @xmath231 such that @xmath391 is a test element and @xmath392 is divisible by @xmath393 .
set @xmath394 we claim that @xmath395 .
next , we show that @xmath400 . as @xmath391 is a test element , it suffices to show that @xmath401 .
expanding @xmath402 yields @xmath403.\ ] ] the quantity in brackets is in @xmath404 by our choice of @xmath6 .
it remains to show that @xmath405 .
this is straightforward and can be verified by the interested reader . | let @xmath0 be a regular local ring with prime ideals @xmath1 and @xmath2 such that @xmath3 and @xmath4 .
it has been conjectured by kurano and roberts that @xmath5 for all positive integers @xmath6 .
we discuss this conjecture and related conjectures .
in particular , we prove that this conjecture holds for all regular local rings if and only if it holds for all localizations of polynomial algebras over complete discrete valuation rings .
in addition , we give an example showing that a certain generalization to nonregular rings does not hold .
american mathematical society |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.